The Dynamic Nature of Hypertrophic and Fibrotic Remodeling of the Fish Ventricle

Thermal Remodelling of the Ectothermic Heart

A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Biology, Medicine and

Health Doctoral Academy.

2016

Adam Nicholas Keen

Thermal remodelling of the ectothermic heart

Page | 2

Thermal remodelling of the ectothermic heart

Contents

ABSTRACT ...... 7

Declaration ...... 9

About the Author ...... 13

Acknowledgements ...... 15

Organisation of Thesis ...... 17

List of Abbreviations ...... 21

1. INTRODUCTION ...... 23

1. 1. THE FUNCTION OF THE HEART ...... 23

1. 1. 1. Pressure-Volume Relationships During the Cardiac Cycle ...... 23

1. 1. 2. Cardiac Muscle Excitation and Contraction ...... 26

1. 1. 3. Contractile Proteins ...... 28

1. 1. 4. Length-Tension Relationships ...... 29

1. 1. 5. Length-Dependent Myofilament Ca2+ Sensitivity ...... 31

1. 1. 6. Passive tension ...... 32

1. 1. 6. 1. Cardiac wall thickness ...... 32 1. 1. 6. 2. The cardiac extracellular matrix ...... 32 1. 1. 6. 3. Intracellular structural proteins ...... 34

1. 2. CARDIAC REMODELLING ...... 34

1. 2. 1. Pressure and Volume Overload Cardiac Remodelling ...... 34

1. 2. 1. 1. Cardiac hypertrophy ...... 35 1. 2. 1. 2. Signalling pathways in hypertrophic growth ...... 37 1. 2. 1. 3. Remodelling of cellular energetics ...... 38 1. 2. 1. 4. Remodelling of connective tissue ...... 39 1. 2. 1. 5. Intracellular structural remodelling ...... 40

1. 3. CARDIAC REMODELLING IN ECTOTHERMS...... 40

1. 3. 1. The Effects of Temperature on Ectotherms ...... 40

1. 4. CARDIAC REMODELLING IN FISH ...... 41

1. 4. 1. Anatomy and Physiology of the Fish Heart ...... 41

1. 4. 1. 1. The sinus venosus ...... 42 1. 4. 1. 2. The atrium ...... 42

Page | 3

Thermal remodelling of the ectothermic heart

1. 4. 1. 3. The ventricle ...... 43 1. 4. 1. 4. The outflow tract (OFT) ...... 43 1. 4. 2. Cardiac Function in Fish with Acute Temperature Change...... 44

1. 4. 2. 1. Effects on myofilaments ...... 45 1. 4. 2. 2. Effects on ion channel flux and the action potential ...... 45 1. 4. 2. 3. Acute effects on the passive properties of the heart ...... 47 1. 4. 3. Cold Acclimation and Remodelling of Fish Heart ...... 47

1. 4. 3. 1. Remodelling of calcium handling...... 48 1. 4. 3. 2. Remodelling of myofibrils ...... 49 1. 4. 3. 3. Cardiac morphology ...... 50 1. 4. 3. 4. The extracellular matrix ...... 52 1. 4. 3. 5. Length-dependent changes in force generation ...... 52 1. 4. 3. 6. Remodelling of cellular energetics ...... 52

1. 5. CARDIAC REMODELLING IN TURTLES ...... 53

1. 5. 1. Anatomy and Physiology of the Turtle Heart ...... 53

1. 5. 1. 1. The atria ...... 53 1. 5. 1. 2. The ventricle ...... 53 1. 5. 2. Cardiac Function in Turtles with Acute Temperature Change ...... 55

1. 5. 3. Cold Acclimation and Remodelling of the Turtle Heart ...... 56

1. 5. 3. 1. Remodelling contractile force ...... 56 1. 5. 3. 2. Remodelling of the passive properties of the heart ...... 57 1. 5. 3. 3. Remodelling of cellular energetics ...... 57

1. 6. SUMMARY ...... 58

1. 7. AIMS ...... 58

2. GENERAL METHODS ...... 61

2. 1. IN VIVO CARDIAC FUNCTION ...... 61

2. 2. WHOLE HEART ...... 62

2. 2. 1. Isolation of the Fish Heart ...... 62 2. 2. 2. Ex Vivo Pressure-Volume Curves ...... 62

2. 3. TISSUE TECHNIQUES ...... 65

2. 3. 1. Tissue Histology ...... 65 2. 3. 1. 1. Haematoxylin and Eosin (H & E) ...... 65 2. 3. 1. 2. Picro-Sirus Red ...... 66 2. 3. 1. 3. Miller’s Elastic Stain ...... 68 2. 3. 1. 4. Oil Red O Stain ...... 69

Page | 4

Thermal remodelling of the ectothermic heart

2. 3. 1. 5. Periodic Acid Schiff (PAS) Stain ...... 69 2. 3. 2. Fourier Transform Infrared (FTIR) Spectroscopy ...... 70 2. 3. 2. 1. Background to FTIR ...... 70 2. 3. 2. 2. Recording a Spectrum ...... 72 2. 3. 2. 3. FTIR Imaging Spectroscopy ...... 73 2. 3. 3. Atomic Force Microscopy (AFM) ...... 76 2. 3. 4. In Situ Zymography ...... 77 2. 4. 1. SDS-PAGE Gelatin Zymography ...... 78 2. 4. 2. Real-Time Quantitative PCR (RT-qPCR)...... 78 3. The Dynamic Nature of Hypertrophic and Fibrotic Remodelling of the Fish Ventricle...... 79

4. Metabolic and biochemical remodelling of the thermally acclimated fish ventricle using Fourier transform infra-red spectroscopy with high magnification optics...... 81

5. Macro- and micro-mechanical remodelling in the fish atrium is associated with regulation of collagen 1 alpha 3 chain expression...... 83

6. Remodelling of compliance, structure and connective tissue in the fish outflow tract with temperature acclimation...... 85

7. Cardiovascular function, compliance and connective tissue remodeling in the turtle, Trachemys scripta, following thermal acclimation...... 87

8. Cold acclimation increases ventricular micromechanical stiffness and collagen content in the freshwater turtle, Trachemys scripta...... 89

9. Metabolic and biochemical remodelling of the thermally acclimated freshwater turtle ventricle using Fourier transform infra-red imaging spectroscopy...... 91

10. GENERAL DISCUSSION ...... 93

10. 1. CARDIAC REMODELLING WITH THERMAL ACCLIMATION IN FISH ...... 93

10. 2. CARDIAC REMODELLING WITH THERMAL ACCLIMATION IN TURTLES ...... 99

10. 3. PERSPECTIVES ...... 104

11. REFERENCES ...... 107

12. APPENDIX ...... 131

Page | 5

Thermal remodelling of the ectothermic heart

Page | 6

Thermal remodelling of the ectothermic heart

PhD Adam N. Keen University of Manchester Thermal Remodelling of the Ectothermic Heart 15.04. 2016 ABSTRACT Chronic changes in cardiac load can cause the vertebrate heart to remodel. For ectotherms, ambient temperature can directly alter cardiac load. Therefore, long- term ambient temperature change can initiate a dynamic cardiac remodelling response to preserve cardiac function. The aims of my PhD thesis were to study the effects of chronic temperature change on the ectothermic heart and cardiovascular system, using the cold-active rainbow trout and the cold-dormant freshwater turtle. In contrast to the majority of previous studies, my experiments focused on the passive, rather than active, properties of the heart. In results chapters 3, 4, 5 and 6, I studied the effects of thermal remodelling on the rainbow trout heart. Chronic cold caused a global increase in chamber stiffness, both at the whole chamber and micromechanical level, with an associated myocardial fibrosis. In the ventricle and atrium there was an up-regulation of collagen promoting genes. In the ventricle, I found cold-induced hypertrophy of the spongy myocardium with an up-regulation of hypertrophic growth factors, which was associated with an increase in tissue lipid suggesting an increase in fatty acid oxidation (FAO). In the atrium, there was no hypertrophy, but there was an increase in extra-bundular sinus, suggesting chronic dilation. Chronic warming initiated an opposite response, with increased cardiac compliance associated with an up- regulation of collagen degrading genes in the ventricle and atrium. In the outflow tract (OFT) and atrium, this increased activity of matrix metalloproteinase (MMPs) and in the OFT abundance of MMPs was increased. The warmed ventricle showed atrophy of the spongy myocardium with a decrease in lipid and an increase in glycogen suggesting a switch in cellular energetics from FAO to glycolytic pathways. In chapters 7, 8 and 9, I studied the effects of thermal remodelling on the freshwater turtle heart. I found an in vivo decrease in systemic resistance causing an increased right to left cardiac shunt flow, associated with an increased elastin content of the major outflow vessels. Cold acclimation increased cardiac sensitivity to preload as well as whole chamber passive stiffness and micromechanical stiffness of tissue sections, associated ventricular fibrosis and increased collagen coherency. In addition, chronic cold decreased the gelatinase activity of MMPs and increased mRNA expression of a tissue inhibitor of MMPs. Furthermore, chronic cold was associated with a decrease in tissue lipid and phosphates, but an increase in tissue protein, glycogen and lactate. These changes in tissue biochemistry suggest a switch in cellular energetics from FAO to glycolytic pathways, likely due to the decreased oxygen availability associated with winter inactivity. Overall, the chambers of the ectothermic heart show distinct remodelling phenotypes, which likely reflect their in cardiac function. Thermal remodelling of the fish ventricle serves both cardio-protection, from the haemodynamic strain of changes in cardiac preload and afterload, as well as compensation for the direct effects of temperature. In the turtle, changes in compliance and cellular energetics of the ventricle suggest a cardio-protective mechanism preparing the heart for increased haemodynamic stress and hypoxic or anoxic conditions during inactive winter hibernation.

Page | 7

Thermal remodelling of the ectothermic heart

Page | 8

Thermal remodelling of the ectothermic heart

Declaration

The quantitative real-time PCR performed on the fish atrium and ventricle has already been submitted in support of Dr Andrew Fenna’s PhD thesis to the University of Manchester. It has been included in this thesis because his work had not yet been published and the work in my thesis has followed on from his resulting in combined manuscripts (one of which has already been published, the other that is ready for submission). No other portion of work in this thesis has been submitted in support of an application for another degree or qualification at this or any other university or institute of learning.

Signed:______Date: ______

Page | 9

Thermal remodelling of the ectothermic heart

Page | 10

Thermal remodelling of the ectothermic heart

i. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes.

ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made.

iii. The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions.

iv. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property University IP Policy (see http://documents.manchester.ac.uk/display.aspx?DocID=24420), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.library.manchester.ac.uk/about/regulations/) and in The University’s policy on Presentation of Theses

Signed:______Date: ______

Page | 11

Thermal remodelling of the ectothermic heart

Page | 12

Thermal remodelling of the ectothermic heart

About the Author

Education and Employment

2016-Present – University of Oxford, Post-doc, Cardiovascular medicine. 2012-2016 – University of Manchester, PhD, Cardiac Physiology with Analytical Chemistry. 2008-2012 – University of Manchester, B.Sc (Hons), Zoology with Industrial Experience.

Publications

Keen, A. N., Shiels, H. A., Crossley, D. A. (2016). The effects of cold acclimation on in vivo cardiovascular function in the freshwater slider turtle, Trachemys scripta. American Journal of Physiology – Regulatory, Integrative and Comparative Physiology 331, R133-R143; 10.1152/ajpregu.00510.2015

Keen, A. N., Fenna, A. J., McConnell, J. C., Sherratt, M. J., Gardner, P., Shiels, H. A. (2016). The dynamic nature of hypertrophic and fibrotic remodelling in the fish ventricle. Frontiers in Physiology – Integrative Physiology 6: 427; doi: 10.3389/fphys.2015.00427

Lea, J. M. D., Keen, A. N., Nudds, R. L., Shiels, H. A. (2016). Kinematics and energetics of swimming performance during acute warming in brown trout Salmo trutta. Journal of Fish Biology 88: 403-17; doi: 10.1111/jfb.12788

Nudds, R. L., John, E. L., Keen, A. N., Shiels, H. A. (2014). Rainbow trout provide the first experimental evidence for adherence to a distinct Strouhal number during animal oscillatory propulsion. Journal of Experimental Biology 217, 2244-2249; doi: 10.1242/jeb.102236

Keen, A. N., Gamperl, A. K. (2012). Blood oxygenation and cardiorespiratory function in steelhead trout (Oncorhynchus mykiss) challenged with acute temperature increase and zatebradine-induced bradycardia. Journal of Thermal Biology 37, 201-210; doi: 10.1016/j.jtherbio.2012.01.002

International Conference Presentations

2015 - Society of Experimental Biology Annual Meeting (Prague, CZ) 2014 – International Congress of Fish Biology (Edinburgh, UK) 2014 – Society of Experimental Biology Annual Meeting (Manchester, UK) 2013 – International Union of Physiologists (Birmingham, UK) 2013 – Society of Experimental Biology Annual Meeting (Valencia, ES) 2012 – Physiology (Edinburgh, UK)

Page | 13

Thermal remodelling of the ectothermic heart

Page | 14

Thermal remodelling of the ectothermic heart

Acknowledgements

I have many people to thank for their help and input during my time as a PhD student and in bringing together my thesis. I would like to thank my supervisors Dr Holly Shiels and Prof Peter Gardner for their guidance, support and expertise, and for pushing me to achieve as much as possible throughout my PhD. Thank you to my advisor, Prof David Eisner, for his help and guidance. Thank you to all of the members of the Shiels lab for their help knowledge and feedback, and particularly help regarding animal care. Thank you to all of the members of the Gardner lab for helping me get to grips with FTIR, helping with data analysis and helpful discussions of my results.

Thank you to Dr Dane Crossley II for inviting me to his laboratory at the University of North Texas, and guiding me through experiments, data analysis and manuscript writing. Thank you to Dr Michael Sherratt and Dr James McConnell for all of their work regarding AFM, as well as helpful discussion and feedback of my work and manuscripts. Thank you to Peter Walker for his help with histology.

Thank you to my examiners Dr Jason Bruce and Prof Matti Vornanen for their time assessing my work and their comments, which have improved my thesis.

Page | 15

Thermal remodelling of the ectothermic heart

Page | 16

Thermal remodelling of the ectothermic heart

Organisation of Thesis

This thesis has been submitted for examination in ‘Alternative Format’. The primary difference between a thesis submitted in alternative format compared to traditional format is that the results chapters are presented as either published scientific papers or as manuscripts that have either been submitted or are pending submission. Throughout my PhD my intention has been to publish my experimental work, so I have focused on writing up studies in a format where that is possible. I feel that working in this way has enriched my experience as a PhD student as it has given me the chance to work and collaborate with a large number of people. It has also meant that I have been lucky enough to have a wide range of input and feedback on my work. I have written each results chapter in the style of a manuscript, each one has a self-contained abstract, introduction, materials and methods, results, discussion and references section, which is specific to the experiments carried out in that chapter. In addition to the results chapters I have written an overall abstract, an extended introduction, a general methods section and a general discussion.

My thesis begins with a review of previous literature relating to heart function and features of cardiac remodelling in mammals, fish and turtles (Chapter 1). It is important to be clear, upfront, that although I give a brief overview of general fish physiology, this thesis primarily concentrates on salmonids and, therefore, much of the discussion is specific to the trout/salmon heart and may not be representative of other fish species. In Chapter 1, I introduce many of the fundamental ideas that contribute to the rationale behind the aims of my thesis, the experiments I conducted and the interpretation of my results. By a thorough review of the literature I cover the background to my overall PhD questions and the individual hypotheses for each of the results chapters.

In Chapter 2 I describe the general methodology. This section contains methodology that is not contained in the individual results sections. For some techniques this may be additional details and for some techniques this may be rationale/justification for using this technique over another.

Chapter 3 is the first results chapter, which investigates the effect of thermal remodelling on the morphology, structure and passive stiffness of the rainbow trout ventricle. This study followed on from the work Dr Andy Fenna conducted during his PhD. He is a named author on this manuscript and conducted the RT-qPCR and some of the stereology during his PhD. I conducted some further stereology, the histological staining and generated the ex vivo pressure volume curves. Following the result of the pressure volume curves, we decided to collaborate with Drs Michael Sherratt and James McConnell. Their expertise lie in tissue biomechanics so they added to the study by assessing micromechanical stiffness of tissue

Page | 17

Thermal remodelling of the ectothermic heart sections by atomic force microscopy (AFM). I wrote the manuscript and the study has resulted in a published paper:

For my PhD I have had two supervisors, Dr Holly Shiels in FLS and Prof Peter Gardner in CEAS. Chapter 4 is the first results chapter from the work I have conducted in Peter’s lab and investigates the effect of thermal remodelling on the tissue biochemistry of the rainbow trout ventricle. I used Fourier transform infrared (FTIR) imaging spectroscopy of ventricle tissue cryosections to assess changes in tissue biochemistry following thermal acclimation. Analysis of FTIR data is complex and requires a high level of computer programming skills in MATLAB (MathWorks, USA). As such, Dr Alex Henderson gave me a lot of help getting to grips with software and I used a number of his algorithms to analyse my data. In addition, due to his understanding of the sophisticated mathematics used, Alex was also involved with the interpretation of the data. For these reasons I have included him as an author on this manuscript. While working on this manuscript I also had an undergraduate student, John Marrin, come and work with me for a summer to gain some lab experience. While in the lab I taught him histological techniques and he conducted the periodic acid Schiff (PAS) staining under my supervision while I conducted the oil red O staining. For this reason he is also included as an author on the manuscript.

In Chapter 5 I switch to investigating the fish atrium following thermal acclimation. This work, again, follows on from the work Dr Andy Fenna conducted in his PhD and, again, he conducted the RT-qPCR. In this study I generated the ex vivo pressure volume curves, performed all of the stereology and the tissue histology, using methodology based on chapter 3. I also assessed the exogenous gelatinase activity of matrix metalloproteinase (MMPs), using in situ zymography. Again, we collaborated with Drs Michael Sherratt and James McConnell to perform AFM. During this study, I was also trained on the atomic force microscope and both James and I collected the data. This work has been submitted for publication in the Journal of Physiology.

In Chapter 6 I turn my investigation to the fish outflow tract (OFT). I generated ex vivo pressure volume curves, performed tissue histology and in situ zymography based on methodology of chapters 3 and 5, and then assessed the collagen coherency of the tissue. We decided to collaborate with Dr John Mackrill who conducted the SDS-PAGE gelatin zymography for us and sent me the data. For this reason he is also included as an author on the manuscript. This work has been submitted for publication in the Journal of Experimental Biology.

Page | 18

Thermal remodelling of the ectothermic heart

Chapter 7 is the first chapter to investigate the effects of thermal remodelling on the freshwater turtle. For this study I travelled to Dr Dane Crossley’s lab at the University of North Texas. Once there, Dane taught me the specialist in vivo techniques to perform the study and supervised me as I collected the in vivo data. On sacrificing the animals I generated ex vivo pressure volume curves for the turtle ventricle, based on the methodology in chapter 3. We then arranged for the tissue to be transported to the University of Manchester and I conducted the tissue histology. This work has been accepted to the American Journal of Physiology – Integrative, Comparative and Regulatory Physiology and we are currently waiting for the proofs to be sent to us.

In chapter 8 I focused on the properties of the turtle ventricle tissue following thermal acclimation. Again, we collaborated with Drs Michael Sherratt and James McConnell to perform AFM and the data was collected by James and myself. I then conducted the tissue histology, the in situ zymography and assessed the collagen coherency of the tissue, based on the methodology in chapters 3, 5 and 6. Again, we sent tissue to Dr John Mackrill who performed SDS-PAGE gelatin zymography for us. I also left some of the turtle ventricle tissue, frozen in RNAlater, in Dr Dane Crossley’s lab. His lab has expertise in RT-qPCR and Dr Janna Crossley performed RT-qPCR on these samples for me, on the genes of interest and using primers that I had designed.

Chapter 9 is the final results chapter. In this chapter I assess the changes in tissue biochemistry of the turtle ventricle using FTIR and tissue histology, as in chapter 4. Again, John Marrin conducted the PAS staining under my supervision and Dr Alex Henderson helped me with FTIR data analysis and interpretation.

To conclude the thesis I have written a general discussion, which makes Chapter 10. As each results section has its own self-contained discussion, in chapter 10 I discuss the relevance of the findings as a whole. By this, I discuss how the results as a whole have added to the field and how the results of each chapter relate to each other. The reference list in Chapter 11 of the thesis relates to the introduction, general methodology and general discussion sections.

I have also included an appendix in Chapter 12. In this section I have included 3 additional manuscripts that I have been involved in working on during my PhD. The first was a continuation of work I conducted during my undergraduate final year project. I was, therefore, very involved in the experiments and writing the manuscript. The work resulted in a published paper:

Page | 19

Thermal remodelling of the ectothermic heart

The second manuscript is from an undergraduate project that I was involved with supervising. I helped set up the experiments, taught the experimental techniques and how to analyse the data and then was involved in writing and editing the manuscript. This work has resulted in a published paper:

The final manuscript that I have included is a review on the thermal acclamation of the fish ventricle that I have written with Dr Jordan Klaiman, Dr Holly Shiels and Dr Todd Gillis. I wrote the sections relating to the passive properties of the ventricle while Jordan, primarily, wrote the sections relating to the active properties of the heart. I was also involved in editing the entire manuscript and many of the figures used were created by me.

Page | 20

Thermal remodelling of the ectothermic heart

List of Abbreviations

AFM – Atomic force microscopy ADP – Adenosine diphosphate ANP – Atrial natriuretic peptide AP – Action potential ATP – Adenosine triphosphate BNP – Brain natriuretic peptide Ca2+ – Calcium ion cAMP – Cyclic adenosine monophosphate COL1α3 – Collagen I alpha 3 cMyBP-C – Cardiac myosin binding protein C cTnC – Cardiac Ca2+ binding cTnI – Cardiac troponin inhibitory cTnT– Cardiac tropomyosin binding E-C coupling – Excitation-contraction coupling ECM – Extracellular matrix EDV – End-diastolic volume ESV – End-systolic volume F-actin – Fibrous actin FAO – Fatty acid oxidation FTIR – Fourier transform infrared spectroscopy GATA4 – GATA 4 binding protein FFPE – Formalin fixed paraffin embedded fH – Heart rate

FMax – Maximal force of contraction H & E – Haematoxylin and Eosin Hct – Haematocrit IMS – Industrial methylated spirit LTCC – L-type calcium channel MHC – Myosin heavy chain MLC – Myosin light chain MLP – Muscle LIM protein MMP – Matrix metalloproteinase mRNA – messenger RNA NCX – Sodium calcium exchanger NFAT – Nuclear factor of activating T cells OFT – Outflow tract PAS – Periodic acid Schiff PBS – Phosphate buffered saline Page | 21

Thermal remodelling of the ectothermic heart

PCNA – Proliferating cell nuclear antigen

Pi – Inorganic phosphate PKA – Phosphokinase A Q – Cardiac output

Q10 – Rate of change over a 10 °C temperature change RCAN – Regulator of calcineurin RT-qPCR – real-time quantitative PCR SAN – Sino atrial node SR – Sarcoplasmic recticulum SERCA – Sarcoplasmic recticulum calcium ion ATPase SMLC – Small myosin light chain

SV – Stroke volume TGF-β – Transforming growth factor TIMP – tissue inhibitor of matrix metalloproteinase Tn – Troponin VEGF – Vascular endothelin growth factor VMHC – Ventricular myosin heavy chain

Page | 22

Thermal remodelling of the ectothermic heart

1. INTRODUCTION

1. 1. THE FUNCTION OF THE HEART

The heart is a chambered, muscular organ with at least one atrium and one ventricle. It lies in the thoracic cavity and is covered by a non-fibrous sack called the pericardium (Katz, 2006; Halper and Kjaer, 2014a). The atria are thin walled chambers that receive blood from the major veins of the body and then pump blood, at relatively low pressure, through valves into the ventricles (Katz, 2006). The ventricles have thick muscular walls and pump blood at high pressure out of the heart to the lungs, or gills, to be oxygenated and to the rest of the body (Katz, 2006; Katz, 2008).

The heart is composed of a number of cell types including cardiac myocytes, cardiac fibroblasts and endothelial cells (Xin et al., 2013). Cardiac myocytes are contractile cells that form the muscular walls (Gerdes and Wang, 2003; Xin et al., 2013). They are composed of repeating sections of sarcomeres, so appear faintly striated (Katz, 2006; Li and Hwang, 2015). In the human heart, cardiac myocytes account for only around a quarter of the total number of cells, but three quarters of the total heart mass (Gerdes and Wang, 2003; Biernacka and Frangogiannis, 2011; Xin et al., 2013).

Cardiac function is strongly influenced by the autonomic nervous system (Brodde et al., 2001; Olshansky et al., 2008). The rhythmic contraction of a heart beat is initiated and controlled by electrical impulses, generated by the spontaneous depolarization of specialized pacemaker cells in the sino-atrial node (SAN) (Saffitz and Corradi; Keith and Flack, 1907; Katz, 2006).

Heart rate and force of contraction are increased by sympathetic stimulation of β1 and β2- adrenoceptors, by adrenalin or noradrenaline via adrenergic nerves, which increase intracellular cyclic AMP (cAMP) (Brodde et al., 2001). Sympathetic innervation can cause vasodilation through stimulation of β2-adrenoceptors in vascular smooth muscle. Sympathetic nervous innervation and noradrenaline are inhibited by the parasympathetic nervous system via the vagus nerve (Olshansky et al., 2008). Acetylcholine release from parasympathetic nerve terminals activates cellular M2-muscarinic receptors, which directly depress heart rate by atrial hyperpolarisation and indirectly reduce force of contraction by inhibiting ventricular adenylyl cyclase (Brodde et al., 2001; Olshansky et al., 2008). Vasoconstriction can occur by parasympathetic innervation of muscarinic receptors in vascular smooth muscle (Olshansky et al., 2008).

1. 1. 1. Pressure-Volume Relationships During the Cardiac Cycle

Each heart beat can be divided into 2 alternating sections, known as diastole and systole (Katz and Rolett, 2016). During diastole ventricles are relaxed and, therefore, fill with blood. During Page | 23

Thermal remodelling of the ectothermic heart systole the ventricles contract, forcing blood into the aorta and pulmonary artery (Katz, 2006). The length and tension of the cardiac myocytes that make up the heart wall alters the intra- chamber pressure and, therefore, the volume of blood pumping in each heart beat (Katz and Rolett, 2016). Therefore, each heart beat can be described by a 4 phase cycle of pressure and volume relationships (Figure 1. 1.). The end-diastolic pressure-volume relationship describes myocardial lusitropic, or relaxation, properties of the chamber (Katz and Rolett, 2016). This relationship can be actively influenced by; the amount and speed of calcium uptake by the sarcoplasmic reticulum, the calcium affinity of troponin (Tn) and the dissociation of contractile proteins once calcium has dissociated from Tn (Arai et al., 1994; Katz and Lorell, 2000; Katz, 2006; Stienen, 2015; Katz and Rolett, 2016). The end-systolic pressure-volume relationship describes myocardial ionotropic, or contractile, properties (Katz and Rolett, 2016). This relationship is actively affected by the same factors that influence actin and myosin cross- bridge formation (Katz and Lorell, 2000; Katz, 2006; Katz and Rolett, 2016). Both of these relationships can be altered, respectively, by changing the lustropic or inotropic state. Together they control end-systolic volume (ESV) and end-diastolic volume (EDV) (the volume of blood remaining in the ventricle at end of the previous cardiac cycle, and, the volume of blood in ventricles at the moment they begin to contract, respectively) and, thus, stroke volume (SV), i.e. the volume of blood ejected during each cardiac cycle (Katz and Rolett, 2016). These pressure-volume relationships are critical for correct cardiac function and determine cardiac compliance, or stiffness of the heart.

Pressure and volume determine cardiac afterload and cardiac preload, respectively, and, therefore, influence contraction and relaxation of myocytes. Ventricular preload is mainly produced by return of venous blood, with a small contribution form atrial systole (Calderone et al., 1995; Toischer et al., 2010; Katz and Rolett, 2016). Ventricular afterload is produced by the systemic pressure that blood is ejected against (Calderone et al., 1995; Toischer et al., 2010; Katz and Rolett, 2016). Changes in either preload or afterload can alter cardiac stroke volume.

Regulation of pressure and volume throughout the cardiac cycle is dependent on a number of active and passive properties of the heart (Katz, 2008; Katz and Rolett, 2016). Although my thesis focuses primarily on the passive properties of the heart, I will now review the key aspects that determine the active and passive properties of the heart and, thus, cardiac function.

Page | 24

Thermal remodelling of the ectothermic heart

Figure 1. 1. A schematic pressure volume loop from a mammalian left ventricle showing the 4 phases of the cardiac cycle. (1) diastolic filling, increase in volume while pressure remains constant, pulmonary flow gives ventricular preload. (2) Isovolumeric contraction during diastole, increased tension of myocytes increases pressure while the mitral valve is closed. (3) Ejection during systole, reduction in volume and plateau in intraventricular pressure. (4) isovolumeric relaxation at the start of next diastole, the afterload pressure causes the aortic valve to close so volume remains constant and pressure decreases as muscles relax. End- diastolic volume (EDV) occurs at the end of phase 1 and end-systolic volume (ESV) occurs at the end of stage 3. The difference between EDV and ESV is stroke volume (SV).

Page | 25

Thermal remodelling of the ectothermic heart

1. 1. 2. Cardiac Muscle Excitation and Contraction

Excitation-contraction (E-C) coupling describes how depolarisation and excitation of the cardiac myocyte membrane translates to the initiation of contraction of the myofilaments. In the mid-20th century A. V. Hill made the observation that simple diffusion was too slow to activate the contractile machinery (Hill, 1949; Katz, 2006). He predicted that cardiac contraction must rely on a process either faster than diffusion or a diffusible activator. As it transpired, in the mammalian heart there are two membrane systems that are brought together in E-C coupling to overcome the limitation of diffusion time; the transverse tubular system and the sarcoplasmic reticulum (SR). These two systems give rise to the intracellular Ca2+ cycle (Katz, 2006).

The action potential causes depolarisation of the sarcolemmal membrane, which activates a small number L-type Ca2+ channels (LTCCs) and brings a small increase in intracellular Ca2+ (Bers, 2002b; Gorski et al., 2015; Shiels and Sitsapesan, 2015). The trans-sarcolemmal influx of Ca2+ bind to ryanodine receptors on the surface of the SR, which initiates massive release of SR stored Ca2+, giving a rapid rise in intracellular Ca2+ (Keef et al., 2001; Bers, 2002a, b; Bers and Despa, 2006; Gorski et al., 2015). Additional Ca2+ may enter the cell via the Na+/Ca2+ exchanger (NCX) (Bers, 2002b; Shiels and Galli, 2014; Shattock et al., 2015). Ca2+ bind with Tn enabling actin and myosin binding sites so that cross-bridges can form, initiating muscle contraction (Figure 1. 2) (Bers, 2002b; Wilson and Lucchesi, 2014; Gorski et al., 2015). For relaxation to occur the intracellular concentration of Ca2+ must decrease so that Ca2+ dissociates from the Tn complex (Bers, 2002b; Shiels and Galli, 2014). There are 4 mechanisms that remove cytosolic Ca2+; uptake of the SR via SR Ca2+-ATPase (SERCA), sarcolemmal NCX, sarcolemmal Ca2+-ATPase and the mitochondrial Ca2+ uniporter (Bers, 2002b).

Page | 26

Thermal remodelling of the ectothermic heart

Figure 1. 2. A schematic representation of excitation contraction coupling. Calcium enters the cell via L-type calcium channels (LTCC) or the sodium calcium exchanger (NCX), which initiates calcium release from the sarcoplasmic reticulum (SR). Together these mechanisms bring a high enough intracellular calcium concentration to activate contraction of the myofilaments. Taken with permission from Shiels and Galli. 2014.

Page | 27

Thermal remodelling of the ectothermic heart

1. 1. 3. Contractile Proteins

The contractile proteins of a cardiac myocyte form thick and thin overlapping filaments, which interact by sliding over one another to produce contractions (Huxley, 1969; Ruegg, 1990; Katz, 2006; Li and Hwang, 2015). They are arranged into highly polarised units known as sarcomeres (Li and Hwang, 2015) (Figure 1. 3).

Figure 1. 3. The sarcomere. Repeating sections of overlapping actin and myosin filaments, anchored at the z-line. Muscle contractions arise from interactions of these filaments as they slide over one another, shortening the sarcomere. The giant sarcomeric protein, titin, determines overall stretch of the myocyte. Adapted with permission from Shiels and White, 2008

The thin filament is composed of three main proteins; fibrous actin (F-actin), tropomyosin and Tn (Ebashi et al., 1964; Ebashi et al., 1967). F-actin is formed from two chains of globular actin molecules in a double stranded macromolecular helix, which forms the backbone of the filament (Rayment et al., 1993; Li and Hwang, 2015). Each actin molecule has an active site where myosin can attach (Stienen, 2015). Tropomyosin is formed of two α-helical chains, polarised from head to tail, and regulates interaction of actin and myosin (Li and Hwang, 2015). Cardiac Tn is formed of 3 proteins, cardiac Tn inhibitory (cTnI), cardiac Tn tropomyosin-binding (cTnT) and cardiac Tn Ca2+ binding (cTnC) (Li and Hwang, 2015; Omland et al., 2015). cTnI modulated actin-myosin interaction by inhibiting actomyosin adenosine triphosphate (ATP) activity, cTnT binds tropomyosin and cTnC binds Ca2+, which accounts for the Ca2+ of the

Page | 28

Thermal remodelling of the ectothermic heart sarcomere and allows interaction of actin and myosin in the presence of Ca2+ (Ebashi, 1963; Omland et al., 2015).

The thick filaments are composed of ~300 myosin molecules. Cardiac myosin is a motor protein, made of two heavy chains and four light chains (Bouvagnet et al., 1984; Yin et al., 2015). The heavy chains coil around each other to form tail and head regions of the filament (Rayment et al., 1993; Yin et al., 2015). The tail regions form the axis of the thick filament while the head region and light chains interact with the actin molecules of the thin filament to form cross-bridges (Ruegg, 1990; Katz, 2006). Each heavy chain contains a single actin-activated ATPase, which provides energy needed for contraction.

Cardiac muscle contraction occurs by the thin filament sliding along the thick filament, shortening the sarcomere (Rayment et al., 1993; Li and Hwang, 2015). The contraction cycle involves a series of interaction between myosin filaments and actin monomers, known as cross-bridge cycling (Li and Hwang, 2015). Ca2+, released from the SR during E-C coupling, binds to the Ca2+-specific binding site of the cTnC (Ebashi et al., 1967; Solaro and Rarick, 1998; Li and Hwang, 2015). The binding of Ca2+ to cTnC allows cTnC to interact with cTNT releasing the inhibitory action of cTnI on the actin binding site (Tripet et al., 1997; Solaro and Rarick, 1998; Li and Hwang, 2015). Removal of the inhibitory region of cTnI moves the troponin-tropomyosin complex from the ‘blocked’ position, allowing the myosin cross-bridges to interact with the actin of the thin filament and the muscle to contract (Tripet et al., 1997; Solaro and Rarick, 1998; Li and Hwang, 2015). Removal of the Ca2+ from cTnC returns tropomyosin to cover the binding site and the muscle relaxes (Solaro and Rarick, 1998; Li and Hwang, 2015).

1. 1. 4. Length-Tension Relationships

Early experiments by Otto Frank (1895) and Ernest Starling (1918) showed a physiological relationship between EDV and systolic pressure (Figure 1. 4) (Starling, 1918). This relationship, now commonly known as the Frank-Starling mechanism, shows the importance of the pressure-volume relationships within the heart as myocyte length directly affects contractility and, therefore, EDV directly alters in vivo SV (Figure 1. 4) (Allen and Kentish, 1985; Katz, 2002, 2008; Shiels and White, 2008).

Although Starling receives the majority of the credit for this finding, this law is based on length- tension relationships in cardiac and skeletal muscle, which can be explained by the degree of overlap between the thick and thin filaments (Figure 1. 3) (Gordon et al., 1966; Stienen, 2015). At cellular level, stretch can be measured by sarcomere length (SL), i.e. distance between striations (Figure 1. 5.). The greater the potential for cross-bridges to occur, and the greater the number can occur simultaneously, results in greater active tension (Stienen, 2015). There

Page | 29

Thermal remodelling of the ectothermic heart are two parts to the Starling curve, known as the ascending and descending limbs (Figure 1. 4). The ascending limb describes normal conditions of the working heart. Under a normal range of physiological EDVs, the capability of the heart to generate systolic pressure increases linearly (Katz, 2008). At short SLs isomeric tension is decreased due to overlap of actin filaments with opposing polarity and increased folding of myosin filaments (Stienen, 2015). Myofilament overlap is optimal near the resting length of the sarcomere (Stienen, 2015). The result is that increasing preload enhances the ability of the heart to eject blood, allowing the heart to respond to high end-diastolic pressure by increases in SV (Katz, 2008). Further, increases in stretch progressively reduce the potential for cross-bridges and, therefore, cause a decrease in contractile force (Stienen, 2015). This describes the descending limb with increases in chamber volume decreasing the heart’s ability to eject blood (Figure 1. 4) (Katz, 2008). Low cardiac compliance and protection from a stiff pericardium are vital to ensure that filling pressure does not exceed a level that the heart can efficiently eject blood, leading to dilated cardiomyopathy, ‘runaway dilation’ and heart failure (Katz and Rolett, 2016).

Figure 1. 4. A schematic starling curve. (A) shows the ascending limb where increases in ventricular filling give increases in cardiac output. (B) shows the descending limb where further increases in ventricular filling decreases cardiac output.

Page | 30

Thermal remodelling of the ectothermic heart

1. 1. 5. Length-Dependent Myofilament Ca2+ Sensitivity

The cellular length tension relationship is not only due to myofilament overlap in cardiac myocytes. Unlike skeletal muscle, an increase in force can still occur where there is no change in myofilament overlap (Figure 1. 5). The reasons for this increase in force is unclear, but may be due to a length dependent change in myofilament Ca2+ sensitivity as the force of cardiac contraction is dependent upon concentration of intracellular Ca2+. The Frank-Starling response shows an increase in response to stretch without a change in intracellular Ca2+ concentration (Allen and Kentish, 1985; Shiels et al., 2006). Therefore, there must be a corresponding change in Ca2+ sensitivity of contractile machinery, which is likely due to cross-bridge formation. In a strongly bonded cross-bridge the position of tropomyosin is shifted further into the actin groove, which increases the probability of further strong binding cross-bridges forming between actin and myosin (Fitzsimons and Moss, 1998; Konhilas et al., 2002). Strong binding cross-bridges can also increase the affinity of TnC for Ca2+, which indicates coupling between the TnC Ca2+ regulatory sites and cross-bridge interactions on the thin filament (Gordon and Ridgeway, 1993; Fukuda and Granzier, 2006).

Figure 1. 5. The celular length-tension relationship in sarcomeres. As sarcomere length increases there is an increase in tension, until the maximum tension the sarcomere can produce. At progressive sarcomere lengths tension is decreased. Progressive sarcomeric stretch is shown in A-D with the overlap between filament shown by brackets. Taken with permission from Shiels and White, 2008.

Page | 31

Thermal remodelling of the ectothermic heart

1. 1. 6. Passive tension

To maintain correct diastolic function the ventricle must be compliant enough to allow sufficient filling, but strong enough to withstand haemodynamic biomechanical stress. Passive tension describes the resistance of a cardiac chamber to diastolic filling and, therefore, plays a role in the Frank-Starling response of the heart (Shiels and White, 2008). There are a number of elements of the myocardium that contribute to passive stiffness. Historically, ventricular wall thickness and connective tissue content were thought to be the dominating factors, however, there is now evidence to suggest important contributing roles for many extracellular and intracellular mechanisms.

1. 1. 6. 1. Cardiac wall thickness

The law of Laplace states that a larger chamber radius, will give a higher wall tension and a greater internal pressure. This relationship is also dependent on chamber shape, with a sphere experiencing only half the wall tension of a cylinder. Laplace’s law is often applied to the heart, demonstrating that the ability of the heart to pump is greatly affected by the size and shape of its chambers (Maenner et al., 2009; Wilson and Lucchesi, 2014). Ventricular walls are thick and muscular to generate sufficient pressure to pump blood around the body. Therefore, the law of Laplace has to be modified to account for the effect of wall thickness on wall tension and cavity pressure, 푃 × 푅 푇훼 ℎ Where T = wall tension, P = cavity pressure, R = chamber radius and h = wall thickness (Katz, 2006). In the heart, individual myocytes contract linearly to develop tension, which translates to wall stress across an area, and wall stress generates pressure at right angles to the wall within the intra-chamber cavity. The law of Laplace demonstrates that intra-chamber pressure is higher with greater dilation, but increased wall thickness will reduce the amount of tension experienced by each muscle fibre at any particular pressure (Wong and Rautaharju, 1968; Katz, 2006).

1. 1. 6. 2. The cardiac extracellular matrix

The elastic elements of the connective tissue are central to determining the overall passive tension of the ventricle as they provide structure and support to the chamber walls (Weber et al., 1993; Katz, 2006). The connective tissue in the cardiac extracellular matrix (ECM) surrounds individual myocytes, muscle bundles and blood vessels forming a complex structural network of interstitial matrix and basement membrane (Sanchez-Quintana et al., 1995; Fomovsky et al., 2010). The main components of the ECM are the interstitial fibrous proteins, collagen and elastin, and glycosaminoglycans (which connect to ECM proteins to form proteoglycans) (Cleutjens and Creemers, 2002; Fomovsky et al., 2010; Halper and Kjaer, Page | 32

Thermal remodelling of the ectothermic heart

2014a). Proteoglycans and and hyaluronic acid (a component of synovial fluid) have roles in determining water and plasma distribution in the interstitial space, affecting fluid pressure and mechanical properties of the matix (Negrini et al., 2008). The ECM also contains fibroblast, macrophages and proteases (Halper and Kjaer, 2014a).

Collagen is the most abundant structural protein in the ECM (Fomovsky et al., 2010). The predominant fibrillar collagen in cardiac tissue is collagen I followed by collagen III (Eghbali and Weber, 1990; Segura et al., 2012). Fibrillar collagen molecules are made by super-coiling 3 alpha amino acid chains into an alpha helix. In mammals, collagen I is composed of two type 1 (α1) and one type 2 (α2) subunits (Saito et al., 2001). It forms stiff fibres that store energy to aid ventricular compression as well as support and maintain alignment of myocytes by bearing wall stress during systole (Figure 1. 6) (Biernacka and Frangogiannis, 2011). During diastole, the fibres uncoil and help re-lengthen myocytes to allow filling (Fomovsky et al., 2010). Once the chambers are full, collagen fibres straighten and become stiff to resist further expansion and damage to myocytes (Figure 1. 6) (Fomovsky et al., 2010).

Figure 1. 6. The structure of collagen in the extracellular matrix and contribution to myocardial mechanics. Myocytes are connected by endoysial collagen struts and are aligned by perimysial fibrillar collagen. In systole the perimysial collagen coils, during diastole the perimysial collagen provides support to cardiac myocytes during passive uniaxial stretch.

Page | 33

Thermal remodelling of the ectothermic heart

Elastin is an extensible, elastic fibrous protein in the ECM of cardiac tissue (Halper and Kjaer, 2014a). Its main function is to accommodate for stretch and compression caused by blood flow through the heart and cardiovascular system (Serafini-Fracassini et al., 1978; Halper and Kjaer, 2014a). Elastin is composed of lysine derived amino acids which allow cross-linking of peptide chains within the protein (Halper and Kjaer, 2014a). It is these protein cross-links that provide the elastic properties, allowing for deformation and conformational changes (Serafini- Fracassini et al., 1978; Halper and Kjaer, 2014a).

1. 1. 6. 3. Intracellular structural proteins

In addition to the contribution of the cardiac ECM, titin plays a key role in the passive tension of the myocardium (Horowits et al., 1989; Granzier et al., 1996; Watanabe et al., 2002; Shiels and White, 2008; Stienen, 2015). Titin is a giant sacromeric protein that runs from the Z-line, at the edge, to the M-line, at the centre, of the sarcomere similar to a myofilament backbone (Figure 1. 3) (Helmes et al., 1996; Wu et al., 2000; Peng et al., 2007; Linke, 2008). The primary function of titin is to generate passive force during diastole, however, it also has important roles in mechanosensitivity and myocyte stress/stretch signalling (Linke, 2008).

The compliance of the titin molecule is regulated by multiple segments at the I-band region of the sarcomere (Shiels et al., 2003). There are two titin isoforms that exist in the heart; a shorter and stiffer N2B isoform and a longer and more compliant N2BA isoform (Cazorla et al., 2000; Prado et al., 2005; Patrick et al., 2009; Stienen, 2015). The ratio of these two isoforms result in different levels of titin based passive tension (Cazorla et al., 2000; Trombitas et al., 2001; Linke, 2008). In addition to isoform, phosphorylation of the N2B element by PKA or PKG can decrease passive force (Linke, 2008).

1. 2. CARDIAC REMODELLING

The heart displays a high degree of plasticity. Chronic changes in cardiac load can initiate a cardiac remodelling response. Cardiac remodelling may involve a combination of changes in morphology, cellular energetics, biochemistry, sensitivity to Ca2+ or muscle force production. It may involve genetic alterations, relating to changes in gene expression, protein synthesis and protein metabolism. Here, I will review a number of common features of cardiac remodelling with reference, primarily, to studies conducted in mammals. In the next section I will fully review cardiac remodelling in ectothermic species.

1. 2. 1. Pressure and Volume Overload Cardiac Remodelling

The main driver of cardiac remodelling is a change in protein expression as part of a compensatory strategy to counteract the adverse effects of chronic stimuli on cardiac function.

Page | 34

Thermal remodelling of the ectothermic heart

The stimulus may be physiological or pathological, but usually will alter work rate of cardiac muscle, thus, the heart remodels to normalize cardiac power (Russell et al., 2000; Katz, 2006). Therefore, alterations in pressure or volume that impact on cardiac function are likely to cause a profound cardiac remodelling response.

1. 2. 1. 1. Cardiac hypertrophy

Cardiac hypertrophy is an abnormal increase in heart muscle (Russell et al., 2000; Dorn et al., 2003; Dorn, 2007). Pressure and/or volume overload can cause an elevated biomechanical strain on myocytes, increasing myocyte stretch and cardiac wall tension (Chen et al., 2007; Katz and Rolett, 2016). The increased biomechanical strain increases mRNA production and protein synthesis leading to cellular hypertrophy, where existing myocytes increase in size to increase wall thickness (Bishop, 1990; Nadal-Ginard et al., 2003). Typically, cardiac hypertrophy does not involve cellular hyperplasia (increase in cell numbers), oedema (cellular swelling) and infiltration (accumulation of extracellular components). In accordance with the law of Laplace, the increased wall thickness reduces tension. Cardiac hypertrophy, therefore, can be seen as a compensatory measure to reduce the strain on each individual myocyte by dispersing a given pressure across a larger amount of sarcomeres (Russell et al., 2000; Katz, 2006).

The morphological form of hypertrophic growth can be dependent upon whether the stimulus is a pressure or volume overload (Figure 1. 7) (Calderone et al., 1995; Toischer et al., 2010). Volume overloads are usually associated with an increase in aerobic physiological requirement, which increases cardiac preload (Bernardo et al., 2010; Ellison et al., 2012). In these situations hypertrophic growth occurs eccentrically, where heart mass increases by longitudinal and lateral growth of myocytes (Gerdes, 2002). Relative wall thickness is independent, but contractile force increases with chamber size to accommodate pumping a greater blood volume (Figure 1. 7) (Dorn, 2007; Bernardo et al., 2010). The overall effect is an increase in cardiac pumping capacity and, therefore, cardiac hypertrophy associated with volume overload is commonly called ‘physiological hypertrophy’ (Bernardo et al., 2010). With a stimulus that causes a pressure overload, such as with hypertension or myocardial disease hypertrophic growth occurs concentrically (Olivetti et al., 1996b; Nadal-Ginard et al., 2003; Dorn, 2007). In this case, both cardiac mass and relative wall thickness increase by lateral growth in mycoytes, but there is minimal associated change in chamber volume (Grant et al., 1965; Gerdes, 2002; Dorn, 2007; Katz and Rolett, 2016). Therefore, this form of hypertrophy does not increase overall pumping capacity and is commonly referred to as ‘pathological hypertrophy’ (Figure 1. 7) (Dorn, 2007; Bernardo et al., 2010).

Page | 35

Thermal remodelling of the ectothermic heart

Figure 1. 7. Differential remodelling with pressure or volume overload. ‘Pathological’ hypertrophy is triggered by pressure overload causing concentric hypertrophy where myocytes grow laterally. Wall thickness increases but chamber volume is variable. ‘Physiological hypertrophy is triggered by volume overload causing eccentric hypertrophy where myocytes grow longitudinally and laterally. In this case, wall thickness increases in line with chamber volume.

Page | 36

Thermal remodelling of the ectothermic heart

Cardiac hypertrophy is often associated with coronary capillary growth to supply the increased cardiac tissue demand for oxygen and nutrition (Bernardo et al., 2010; Ma et al., 2014; Daskalopoulos et al., 2016). Increased capillary density is particularly evident in hearts that have experienced concentric hypertrophy as elevated stress in the endocardial tissue brings a consequent rise in energy expenditure (Katz, 2006). Secretion of angiogenic factors, such as vascular endothelial growth factor (VEGF) and angiopoietins, are increased to stimulate capillary growth (Bernardo et al., 2010). The expression of these angiogenic factors is controlled by transcription factors, such as hypoxia-inducible factor-1 and GATA binding protein 4 (GATA4) (Weber and Janicki, 1989). However, persistent stressful stimuli can result in a decrease in angiogenic factors. An imbalance between oxygen supply and demand can cause myocardial hypoxia, cardiomyocyte apoptosis and necrosis, and contractile dysfunction, all of which are signs of imminent heart failure (Katz, 2008).

1. 2. 1. 2. Signalling pathways in hypertrophic growth

Increases in ventricular muscle mass rely on increased protein synthesis. Therefore, cardiac hypertrophy can alter gene transcription, translation and post-translational control as well as protein trafficking and degradation (Calderone et al., 1995; Frey and Olson, 2003). Pathological hypertrophy is triggered by angiotensin or endothelin I (ET-1), which activates activating protein kinases, map kinases and the calcineurin-nuclear factor of activating T cells (NFAT) signalling cascade (Frey and Olson, 2003). During pathological hypertrophy a number of genes normally expressed at high levels in the adult heart are down-regulated, such as α- myosin heavy chain (α-MHC) and SERCA (Arai et al., 1994; Calderone et al., 1995; McMullen and Izumo, 2004) and instead there is up-regulation of the foetal gene program, including atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP), α-skeletal actin, atrial myosin light chain kinase (MLC-1) and β-myosin heavy chain (β-MHC) (Schiaffino et al., 1989; Calderone et al., 1995; Frey and Olson, 2003; de Bold and de Bold, 2005; Woodard and Rosado, 2008; Bernardo et al., 2010).

Physiological hypertrophy is initiated by peptide growth factors, such as transforming growth factor beta 1 (TGF-β1), and activates the phosphatidylinositol-4,5-bisphosphate 3-kinase (PI3K) pathway (Calderone et al., 1995; Molkentin et al., 1998; Frey and Olson, 2003; McMullen et al., 2004; Bernardo et al., 2010). These signalling pathways are coupled with transcription factors, which translocate to the nucleus and regulate long-term changes in gene expression (Molkentin et al., 1998; Bernardo et al., 2010). Up-regulation of the foetal gene program is not usually associated with physiological hypertrophy (Calderone et al., 1995), although some evidence suggests increased mRNA expression of ANP (Lattion et al., 1986). Instead the gene regulator, GATA4, can directly up-regulate hypertrophic growth genes (McMullen and Izumo, 2004; McMullen et al., 2004; Oka et al., 2006; Perrino and Rockman, 2006).

Page | 37

Thermal remodelling of the ectothermic heart

1. 2. 1. 3. Remodelling of cellular energetics

In a normal adult heart, the primary energy production pathway is fatty acid oxidation (FAO) by the mitochondria to produce ATP (Allard et al., 1994; Ingwall, 2009; Doenst et al., 2013). Energy is released as ATPase cleaves the terminal phosphate ion of the ATP molecule, leaving adenosine diphosphate (ADP) and inorganic phosphate (Pi) (Ingwall, 2009). The heart consumes more energy than any other organ, with ATP needed for all cellular functions including contraction, synthesis and degradation of molecules and active transport of ions across the cell membrane (Neubauer, 2007; Ingwall, 2009; Kolwicz et al., 2013). Under normal conditions, the heart’s fatty acyl-CoA and pyruvate provide the fuel for sufficient ATP production (Kolwicz et al., 2013). However, the cardiac metabolic network is highly flexible and during acute increases in demand glycogen and phosphocreatine (the hearts energy store) are used to provide alternative pathways for ATP production (Ingwall, 2009; Abel and Doenst, 2011).

During cardiac hypertrophy, ATP demand of the myocardium increases to support myocyte growth (Abel and Doenst, 2011). In physiological hypertrophy, both FAO and glucose oxidation may increase or remain unchanged, but importantly ATP production remains sufficient (Abel and Doenst, 2011). In pathological hypertrophy, energy demand outweighs energy supply, due to high wall tension in endocardial layers and hypoxic conditions, and phosphocreatine levels fall (Neubauer, 2007). ATP levels may remain constant, but ADP rises which can inhibit the function of a number of intracellular enzymes, reducing the hearts contractile ability (Neubauer, 2007). A reduction in mitochondrial FAO and a reduction in expression of genes involved in FAO leads to an increased dependence on glycolysis, without an increase in glucose oxidation, and increased pyruvate oxidation (Allard et al., 1994; Shen et al., 1999; Akki et al., 2008; Ingwall, 2009; Abel and Doenst, 2011; Kolwicz et al., 2012). To maintain the tricarboxcylic acid (TCA) cycle, intermediate anaplerosis reactions are also increased by glycolytic pyruvate and malic enzymes, therefore, bypassing energy-yielding reactions and reducing energetic efficiency (Sorokina et al., 2007; Pound et al., 2009; Abel and Doenst, 2011; Kolwicz et al., 2012).

Switching cellular energetics to glycolysis is a state typical of the foetal heart (Kolwicz et al., 2013). Under hypertrophic conditions, anaerobic glycolysis does improve short-term rate of ATP production, as myocardial oxygen efficiency per mole of ATP produced is greater than during FAO (Ingwall, 2009). However, ATP production by glycolysis is inefficient for supplying the high-energy demands of a cardiac myocyte in the long-term as although the relative contribution of glucose is increased, it fails to fully compensate for the decline in ATP due to decreased FAO (Ingwall, 2009; Abel and Doenst, 2011). As a result, this change in energy pathways is often associated with decreased myocardial energetics, impaired cardiac function

Page | 38

Thermal remodelling of the ectothermic heart and an exhaustion of contractile reserves due to ATP deficiency (Herrmann and Decherd, 1939; Ingwall et al., 1985; Jung et al., 1998; Jung et al., 2000; Ingwall and Weiss, 2004; Ventura-Clapier et al., 2004).

1. 2. 1. 4. Remodelling of connective tissue

Chronic pressure overload is often associated with formation of excess connective tissue in an organ or tissue, known as interstitial fibrosis, due to a reactive process or repair (Bing et al., 1971; Weber et al., 1989; Weber et al., 1993; Liao et al., 2005; Collier et al., 2012a; Collier et al., 2012b; Daskalopoulos et al., 2016). This maladaptive state can cause inflammation, changes in matrix metalloproteinase activity, fibroblast activation, myofibroblast formation, and the development of cardiac fibrosis (Collier et al., 2012b). Fibrosis occurs during hypertrophy because the structure and abundance of collagen fibres are under numerous controlling factors including; hormonal and growth factors, regulatory proteins and cytokines, but are also affected by haemodynamic stress (Cleutjens and Creemers, 2002; Horn et al., 2012). During hypertrophy, the levels of collagen fibres cross-linking has been shown to increase in addition to an up-regulation of type I, II and IV mRNA after pressure overload (Chapman et al., 1990; Badenhorst et al., 2003).

The immediate effect of fibrosis is to improve tensile strength and transduction of myocardial force, which provides support to the cardiac wall to counteract the increased haemodynamic stress during high cardiac load (Collier et al., 2012a). This adaptive response increases systolic stiffness and preserves force-generating capacity of the myocardium. If cell loss by necrosis occurs, fibrosis takes a reparative role (Bernardo et al., 2010). However, excessive stiffening of the myocardium can reduce diastolic filing, which can impede maximum SV, and in severe cases lead to diastolic dysfunction (Collier et al., 2012; Zile and Brutsaert, 2002). In addition, coronary blood flow can be reduced and electrical transduction hindered, increasing the chance of arrhythmias (Cleutjens and Creemers, 2002).

Due to its important structural role in the cardiac wall, connective tissue content is tightly regulated across multiple levels of biological organization (Nagase et al., 2006; Löffek et al., 2011). Within tissues, matrix metalloproteinase (MMPs) regulate degradation of several component of the ECM, including collagen (Galis and Khatri, 2002; Löffek et al., 2011). In turn, MMP activity is regulated by tissue inhibitors of matrix metalloproteinase (TIMPs) (Löffek et al., 2011). Hence, MMP activity plays a key role in both physiological and pathological remodelling of the cardiovasculature (Creemers et al., 2001; Galis and Khatri, 2002; Antonio et al., 2013).

Page | 39

Thermal remodelling of the ectothermic heart

1. 2. 1. 5. Intracellular structural remodelling

Intracellular diastolic properties (stiffness and stretch) are modulated by titin (Horowits et al., 1989; Granzier et al., 1996; Watanabe et al., 2002), which shows a large degree of plasticity during development and a diversity of isoforms (Cazorla et al., 2000). The ratios of these compliant (N2BA) versus stiff (N2B) isoforms determines the titin based passive force (Trombitas et al., 2001; Stienen, 2015). Dramatic alterations are seen in the ratios of these two isoforms in response to the altered haemodynamic conditions during cardiac hypertrophy, although little is known about the mechanism (Linke, 2008). Interestingly, total percentage of the compliant N2BA isoform in cardiac tissue has been shown to increase with the elevated ECM stiffness of fibrosis, suggesting a compensatory shift (Neagoe et al., 2002). However, a disadvantage may be a reduced titin spring activity and, therefore, impairment of systolic function due to an inability of myocyte stretch sensing and the Frank-Starling mechanism (Krüger et al., 2006; Linke, 2008).

1. 3. CARDIAC REMODELLING IN ECTOTHERMS

1. 3. 1. The Effects of Temperature on Ectotherms

The phylum Chordata is almost entirely vertebrates. Of the ~64,000 currently described vertebrate species, a wide and diverse proportion of these are exclusively ectothermic including most fishes, amphibians and reptiles (Raske et al., 2012). Ectotherms, commonly referred to as cold-blooded animals, cannot regulate their own body temperature by metabolism, as endothermic animals do, and instead conform to the ambient environmental temperature. Allowing internal body temperature to fluctuate with ambient temperature change has profound implications on ectotherm physiology as biochemical and metabolic reactions are directly altered by temperature (Seebacher and Franklin, 2005). As such, true ectothermy (where animals are completely reliant on ambient temperature) is only found at the far end of the ectothermy to endothermy spectrum and, instead, many ectotherms have devised a number of strategies to provide some degree of thermoregulation and allow for appropriate physiological function (McBride and Hernandez-Divers, 2004). Ectotherms may partially thermoregulate by behaviour, such as basking, or by physiology, such as heart rate hysteresis (‘time-lag’), alterations in peripheral resistance, respiratory cooling and metabolically produced heat (James and Mrosovsky, 2004; Seebacher and Franklin, 2005; Tattersall et al., 2006; Shiels et al., 2011; Raske et al., 2012).

Animals living in temperate areas can experience large daily and seasonal temperature fluctuation (Gamperl and Farrell, 2004; Schulte, 2007). It is, therefore, important that ectothermic animals have a large capacity for adaptation. A capacity for phenotypic variation to long-term changes in temperature is known as thermal acclimatisation (or acclimation when

Page | 40

Thermal remodelling of the ectothermic heart controlled within a laboratory setting), during which many physiological processes are altered to compensate for the direct (i.e. Q10; rate of change over 10 °C) effect of temperature. While a change in temperature will affect the function of all organs, the output of the heart is especially important to support active biological process due to its role in transporting oxygen, metabolic substrates and metabolic by-products around the body. Thus, fish employ a number of mechanisms to preserve cardiac function over short- and long-term temperature changes.

1. 4. CARDIAC REMODELLING IN FISH

1. 4. 1. Anatomy and Physiology of the Fish Heart

The fish circulatory system forms a single circuit in which blood passes through a minimum of two sets of capillaries. Under normal conditions the mass of the fish heart is around 0.1% relative to the body mass of the fish. It is composed of four chambers (Figure 1. 8.), which blood passes through in series; the sinus venosus, atrium, ventricle and the bulbus arteriousus or outflow tract (OFT).

Figure 1. 8. The fish heart. The fish heart is composed of 4 chambers; the sinus venous, one atrium, one ventricle and the bulbus arteriosus, which lie in a semi-rigid pericardium. Venous blood from the ductus cuvea and hepatic veins passes through each chamber in series, arrows show blood flow.

Page | 41

Thermal remodelling of the ectothermic heart

The fish heart lies within a semi-rigid pericardium which can influence cardiac filling (Figure 1. 8) by altering intra-chamber pressure (Farrell et al., 1988c; Farrell, 1991). In addition, to vis-a- tergo filling of heart, where venous pressure from systolic contraction and the arterial ‘windkessel effect’ pushed blood into the atrium, the fish heart fills by vis-a-fonte mechanism, whereby suction from the atrium pulls blood into the chamber (Farrell et al., 1988c; Mendonca et al., 2007). The suction caused by changes in intra-chamber pressure due to the pericardium also allows for cardiac filling at sub-ambient pressures (Farrell, 1991)

Most fish hearts, including salmonids, modulate cardiac output (Q) predominantly by SV rather than heart rate (fH) (Farrell, 1991; Forster and Farrell, 1994). This is because although the fish heart can only achieve modest increases in fH (~2-fold) it has capacity for large increases in

SV (~3-fold).

1. 4. 1. 1. The sinus venosus

The sinus venosus is a thin walled chamber (~60-90 μm) that collects deoxygenated, venous blood from the ductus cuvier, jugular and hepatics veins (Laurent et al., 1983; Icardo, 2006). The chamber walls are formed primarily of connective tissue, with a varying ratio of cardiac muscle to venous tissue dependent upon species. A ring of specialized myocardial pacemaker cells are found within the sinus venosus, which make the SAN (Haverinen and Vornanen, 2007). These cells are activated by pooling blood and responsible for initialising and controlling the heart beat. The electrical pulse is transmitted to the atrium and the rest of the heart.

1. 4. 1. 2. The atrium

Blood passes from the sinus venosus into the atrium through the sino-atrial valve. The atrium is the first of the two actively pumping chambers and has thin, highly trabeculated walls. The trabeculae are thin strands of muscle that extend from the sino-atrial valve and form a web- like structure which aid contractions by pulling the walls and roof inwards. Although delicate structures, atrial force development has been shown to be greater than ventricular per mg of tissue and a faster rate of isomeric contraction (Aho and Vornanen, 1999).

The atrium has the greatest maximum capacity of any chamber in the heart and can modulate stroke volume by acting as a volume reservoir for end-diastolic volume (Forster and Farrell, 1994). Some studies have suggested that atrial contraction may be the sole determinant in ventricular filling and modulating ventricular end-diastolic volume (Johansen and Burggren, 1980; Farrell and Jones, 1992). Although this is debated, the atrial contribution to ventricular filling is generally considered to be of greater importance in fish to mammals (Lai et al., 1998; Aho and Vornanen, 1999). Blood passes from the atrium through the atrio-ventricular ostium into the ventricle.

Page | 42

Thermal remodelling of the ectothermic heart

1. 4. 1. 3. The ventricle

The ventricle is a highly muscular triangular shaped chamber. Depending on the species either the whole or the majority myocardium is organised as a highly trabeulated tissue called the spongy myocardium. The spongy myocardium covers almost the entire inner space of the ventricle, leaving only a very small lumen, and is present in all species of fish (Agnisola and Tota, 1994). It is formed of a highly organized system of muscular trabeculae and loosely connected trabecular sheets (Pieperhoff et al., 2009). The spongy myocardium can also be used to increase end-diastolic volume by acting as a reserve to hold blood (Klaiman et al., 2011).

The ventricle of some fish has another distinct layer, known as the compact myocardium. This dense, outer layer encases the spongy myocardium and is composed of circumferentially arranged cardiac myocytes (Pieperhoff et al., 2009). The compact myocardium is only found in relatively active fish, such as salmonids, where it makes up around 30% of the myocardial mass. This layer supplies a coronary blood flow as well as providing extra strength to allow for greater pressure (Franklin and Davie, 1992).

The relative amount of compact and spongy myocardium determines whether the teleost heart functions as a ‘pressure pump’ or a ‘volume pump’ (Agnisola and Tota, 1994). Active teleosts have ventricles with different compact layer thickness (Farrell and Jones, 1992). The compact layer allow these hearts to act as pressure pumps, moving small volumes of blood at a relatively high fH and pressure. For example, the ‘athletic’ tuna heart has the highest relative mass and proportion of the compact layer (40-70 %) among fishes (Agnisola and Tota, 1994). In contrast, the spongy layer of the ventricle allows for pumping large volumes of blood (Agnisola and Tota, 1994; Pieperhoff et al., 2009). An extreme example is the white blooded Antarctic icefish, whose heart is characterized by a thick spongy layer with a minimal compact layer. This type of heart allows the icefish to displace large systolic volumes at a low rate and relatively low pressure, and allows for the large ventricular fillings (high ventricular compliance) necessary for this to occur (Tota et al., 1988; Agnisola and Tota, 1994). The icefish heart, therefore, functions as a volume pump which is specialised to move large stroke volumes at a low fH, increasing cardiac output (Q) (Tota et al., 1988; Agnisola and Tota, 1994).

1. 4. 1. 4. The outflow tract (OFT)

The outflow tract (OFT) is contains a specialised chamber, called the bulbus arterious (or conus arteriosus in more primitive species) and bulbo-ventricular valves (Icardo, 2006). It is unclear whether it is of cardiac or vascular origin (Licht and Harris, 1973; Priede, 1976; Benjamin et al., 1983; Clark and Rodnick, 1999; Braun et al., 2003; Duran et al., 2008; Seth et al., 2014), however, recent evidence suggest that cardiac progenitor cells form the bulbus

Page | 43

Thermal remodelling of the ectothermic heart arteriosus through a complex regulation process by the extracellular environment and mechanotransduction (Moriyama et al., 2016). The large, high-pressure, chamber has no pumping capacity and thick walls composed of layers of smooth muscle, collagen and elastin (Icardo et al., 1999; Braun et al., 2003; Icardo and Colvee, 2011). The layers of collagen and elastin make the OFT both strong and compliant (Icardo and Colvee, 2011). However, they are organised in a novel amorphous arrangement, which gives the chamber specialised inflation properties (Benjamin et al., 1983; Braun et al., 2003; Icardo, 2006). As the fish cardiovascular system has no pulmonary circulation, blood is pumped at high pressure directly to the gills via a short ventral aorta. To prevent damage of the delicate gill capillaries the bulbus arteriosus smooths high pressure pulsatile blood flow, performing the ‘windkessel’ action of the whole mammalian arterial tree (Stevens et al., 1972; Licht and Harris, 1973; Jones et al., 1974; Priede, 1976; Farrell, 1979; Watson and Cobb, 1979; Bushnell et al., 1992; Jones et al., 1993).

1. 4. 2. Cardiac Function in Fish with Acute Temperature Change

Fish are eurythermal ecotherms so ambient water temperature directly alters cardiac physiology (Aho and Vornanen, 2001; Shiels et al., 2011). Cold temperature causes bradycardia (slowing of fH) (Aho and Vornanen, 2001; Kalinin et al., 2009; Lurman et al., 2012), reduces contractility (Vornanen et al., 2002a; Shiels et al., 2003; Kalinin et al., 2009) and maximal force (FMax) (Shiels et al., 2000). The result is a drop in Q and with decreasing temperature (Kalinin et al., 2009; Lurman et al., 2012; Lee et al., 2016). Although overall metabolic rate is also reduced at low temperature (Costa et al., 2013), many fish remain active, therefore, cardiac function needs to be defended to maintain aerobic power and swimming ability and, therefore, stroke volume is maintained or increased during acute cooling (Overgaard et al., 2004; Lee et al., 2016). Acute warming increases metabolic rate and, therefore, cardiac output which is met by an increase in fH (Farrell et al., 1996; Gamperl and

Farrell, 2004; Gollock et al., 2006; Keen and Gamperl, 2012). However, increased fH is responsible, at least in part, for decreased force and shortened contraction time at high temperatures (Aho and Vornanen, 2001; Shiels et al., 2002a).

In addition, as temperature falls blood viscosity increases (Graham and Farrell, 1989). Blood viscosity is determined by haematocrit (Hct; the % red blood cells per unit volume of blood), red blood cell stiffness and plasma viscosity, with the Hct being most critical. Low temperature is generally associated with reduced levels of haemoglobin and Hct. Changes in blood viscosity have implications on the blood flow and vascular resistance in accordance with Poiseuille’s law; Q = πΔΡr4/8ηl where Q = flow, π = 3.14, ΔP = pressure gradient, r = vessel radius, η = viscosity and l = length of vessel. This equation can be rearranged so that;

Page | 44

Thermal remodelling of the ectothermic heart

R = 8ηl/πr4

Therefore, increases in blood viscosity will increase vascular resistance at any given vessel radius, which will increase cardiac preload, cardiac afterload or both.

1. 4. 2. 1. Effects on myofilaments

Acute low temperature reduces the overall ability of the heart to generate force and the rate of isomeric contraction (Churcott et al., 1994; Aho and Vornanen, 1999). In both cardiac and red muscle low temperature impairs contractile function of the thin filament by reducing its sensitivity to Ca2+ (Harrison and Bers, 1990; Churcott et al., 1994). The reduced strength of contraction has been attributed to a decrease in the Ca2+ affinity of cTnC, the Ca2+ activated trigger of the contractile element (Gillis et al., 2000). As a result, force generation is impaired as is the rate of cross-bridge cycling (Gillis et al., 2000). The negative inotropic effect of low temperature has been shown in a wide range of animals including frogs, mice, rats, rabbits, ferrets and ground squirrels (Harrison and Bers, 1990; Churcott et al., 1994). Trout cardiac muscle behaves similarly, however, the myofilaments of the trout heart have several characteristics that allow them to remain functional at low temperatures and over a range of physiological temperatures. Churcott et al. (1994) demonstrated that Ca2+ sensitivity of cardiac actin-myosin ATPase is higher in trout than rats, when compared at their respective physiological temperatures and pH (7 °C versus 37 °C; for trout and rat, respectfully), and trout cardiac muscle preparations require only ~10 % of the Ca2+ concentration required by rat cardiac muscle to reach half maximal tension when tested at the same experimental temperature (Churcott et al., 1994). Thus, high Ca2+ sensitivity of the trout cardiac muscle is considered an important mechanism to help offset the cardioplegic effects of cold.

1. 4. 2. 2. Effects on ion channel flux and the action potential

Acute changes in ambient temperature can alter the rate of opening and closing of ion channels and the catalytic rate of ion pumps (Vornanen, 2016). Cold-induced bradycardia is a result of decreased rate of diastolic depolarisation and increased action potential (AP) duration (Shiels et al., 2000; ; Haverinen and Vornanen, 2007; Galli et al., 2009; Ballesta et al., 2012; Vornanen, 2016) (Figure 1. 9). The reduced contractility and increased AP duration of the heart with acute cold temperatures is reflected by a reduced flux of Ca2+ into the myocyte through voltage gated Ca2+ channels and reduced Na2+ current, and reduced rate of repolarising K+ currents (Shiels et al., 2000, 2002a, b; Vornanen et al., 2002b; Shiels et al., 2003; Vornanen, 2016) (Figure 1. 9). The increased AP duration can compensate for the reduced rate of Ca2+ influx by allowing more time for Ca2+ influx during the AP plateau, which can maintain intracellular Ca2+ flux during acute cold (Shiels et al., 2000). For some species, such as bluefin tuna, the reduced rate of Ca2+ influx during cooling is not completely

Page | 45

Thermal remodelling of the ectothermic heart compensated for by the prolongation of the AP duration. In these hearts, adrenaline, released during cold water dives, augments Ca2+ influx through voltage gated ion channels (Shiels et al., 2015). These increases in Ca2+ influx combine with a prolonged AP duration to restore a force generating Ca2+ flux into the myocytes, across acute (~10 °C) temperature changes (Shiels et al., 2015).

Figure 1. 9. Changes in Ca2+ flux with acute temperature. (A) temperature reduces Ca2+ flux through L-type Ca2+ channels, but (B) increases the duration of the action potential. Taken with permission from Shiels et al. 2000 and Galli et al. 2008.

Page | 46

Thermal remodelling of the ectothermic heart

1. 4. 2. 3. Acute effects on the passive properties of the heart

The passive properties of the heart are also affected by a change in acute temperature. Reductions in ambient temperature increase the passive stiffness of the muscle, which may impair the ability of the muscle to fill during diastole leading to diastolic dysfunction and a loss of cardiac output. Conversely, acute warming decreases the passive tension of inactive muscle fiber bundles, by decreasing the contribution of viscous tension, viscoelastic tension, and elastic tension to muscle stiffness (Mutungi and Ranatunga, 1998).

1. 4. 3. Cold Acclimation and Remodelling of Fish Heart

There are two opposing strategies used by fish species over winter, those that remain cold active and those that become cold dormant. For cold dormant species, such as common carp (Cyprinus carpio), crucian carp (Carassius carassius), burbot (Lota lota) and cunner (Tautogolabrus adspersus), cold temperature triggers large depressions in metabolic rate (Aho and Vornanen, 1998; Tiitu and Vornanen, 2001; Tiitu and Vornanen, 2002; Stecyk and Farrell, 2006; Stecyk et al., 2008; Costa et al., 2013). Metabolic energy saving strategies can normally be characterised by a decrease in Q (Guderley and St-Pierre, 2002; Stecyk and Farrell, 2002, 2006). Metabolic depression can reduce tissue levels of mitochondrial enzymes and mitochondrial proton leak rates (Guderley and St-Pierre, 2002). In response to a decrease in environmental temperature the heart inversely compensates; an active suppression of cardiac function in addition to those directly associated with temperature (Tiitu and Vornanen, 2001). Together these changes prepare these species for months of reduced energy supply during winter (Tiitu and Vornanen, 2001; Stecyk et al., 2008).

Other species remain active during prolonged cold allowing them to forage, continue to grow and evade predators. These include salmonid species like rainbow trout, Oncorhynchus mykiss, and members of the minnow family like the zebrafish, Danio rerio. Marine species like tunas, also experience seasonal temperature changes associated with oceanic migrations. Although the optimal temperature range for salmonids is 15-18 °C (Gamperl and Farrell, 2004), winter temperature can drop to < 4 °C and summer temperatures can reach ~25 °C (Threader and Houston, 1983), therefore, these fish need the ability to maintain maximum Q and maximum power output over a wide range of temperatures to supply oxygen and metabolic fuel to maintain swimming performance (Driedzic et al., 1996; Taylor et al., 1996). To compensate for reduced cardiac function at low temperature, these fish exhibit a beneficial and reversible cardiac remodelling response to prolonged cold (Kent and Prosser, 1985; Farrell et al., 1988b; Graham and Farrell, 1989; Aho and Vornanen, 1999, 2001; Klaiman et al., 2011; Klaiman et al., 2014). The results are; an in vivo increase in fH (Keen et al., 1993;

Aho and Vornanen, 2001; Haverinen and Vornanen, 2007; Lurman et al., 2012), maximum SV (Graham and Farrell, 1989; Farrell, 1991; Driedzic et al., 1996; Lurman et al., 2012; Lee et al.,

Page | 47

Thermal remodelling of the ectothermic heart

2016), maximum power output (Bailey and Driedzic, 1990; Lurman et al., 2012) and maximum Q (Lurman et al., 2012). However, the cost of these compensatory responses is a decrease in the thermal tolerance of the heart (Aho and Vornanen, 2001). This review will focus primarily on the compensatory cardiac remodelling response of cold active fish species such as rainbow trout, which are cold active, and are a species used in my PhD.

1. 4. 3. 1. Remodelling of calcium handling

Under normal situations, the fish heart is dependent upon extracellular Ca2+ for contraction, via Ca2+ channel flux or NCX, with little contribution of the SR to activating intracellular Ca2+ levels (Hove-Madsen and Gesser, 1989; Shiels and Farrell, 1997; Aho and Vornanen, 1999). Despite increased AP duration with acute cold allowing longer Ca2+ influx time (Figure 1. 9); this mechanism is only effective in the short-term and is less efficient during chronic temperature change. As such, cold acclimation is associated with altered kinetics of LTCCs (Vornanen, 1998). However, cold active species, such as rainbow trout and perch, have been shown to increase the Ca2+-handling efficiency of the SR and the Ca2+ uptake of the SR, following cold acclimation (Bowler and Tirri, 1990; Keen et al., 1994; Aho and Vornanen, 1998). In fact, thermal acclimation alters the rate of Ca2+ sequestration and Ca2+ affinity of SR Ca2+-ATPase providing increased utilization of SR Ca2+ stores to activate contraction following cold acclimation (Aho and Vornanen, 1999; Shiels et al., 2011; Korajoki and Vornanen, 2012). Cold acclimation also shortens the mechanical refractory period and prolongs diastole, as with bradycardia, increasing Ca2+ stores in the cold acclimated atrium and possibly the ventricle also (Aho and Vornanen, 1999). Interestingly, in cold dormant species, such as the crucian carp, the limited Ca2+-handling capacity is depressed after acclimation to low ambient temperature (Aho and Vornanen, 1999).

Increase in basal fH, with cold acclimation, indicates an increase in rate of diastolic depolarisation and the rate of AP discharge in primary pacemaker cells of the SAN (Aho and Vornanen, 2001; Haverinen and Vornanen, 2007). It is likely a partial compensation for reduced inotropic effects of acute cold, increasing rate of contraction and relaxation with cold acclimation (Bailey and Driedzic, 1990; Aho and Vornanen, 1999, 2001). Rate of AP discharge is determined by a number of ion channels, cation pumps and SR Ca2+ release (Aho and Vornanen, 2001). It is possible that chronic temperature change may alter isoform expression of channels (Baker et al., 1997), pumps and exchangers or lipid environment of the membrane (Aho and Vornanen, 2001). With cold acclimation there is an associated remodelling of K+ and Na2+ channels expression, which alters the repolarising K+ currents and, therefore, the shape of the AP, which shortens AP duration in periods of prolonged cold temperature (Vornanen et al., 2002; Haverinen and Vornanen, 2004, 2009). This is probably driven by impingement on electrical restitution.

Page | 48

Thermal remodelling of the ectothermic heart

1. 4. 3. 2. Remodelling of myofibrils

At the subcellular level there are also changes in myofilament protein function following thermal acclimation. The kinetics of isomeric contraction are determined by attachment and detachment rates of actin-myosin cross-bridges. As such, alterations in myosin light chain and heavy chain composition can affect myosin ATPase activity (Bottinelli et al., 1995). Cold acclimation gives increased activity of myofibrillar Ca2+/Mg2+-ATPase, particularly in atrial tissue (Aho and Vornanen, 1999), and increased maximal rates of the cardiac actomyosin- ATPase compared to warm-acclimated trout, when tested at a common temperature (Yang et al., 2000; Klaiman et al., 2011). At the tissue level, cold acclimation can increase in the Ca2+ sensitivity of skinned cardiac trabeculae, without altering the kinetics of contraction (Klaiman et al., 2014). The increased Ca2+ sensitivity combined with the increase in actomyosin-ATPase would allow for an increase in the force of contraction across the physiological Ca2+ range of the force- Ca2+ relationship.

Myofilament function can be modulated by two mechanisms; changes in the phosphorylation state of the cardiac regulatory proteins and differential isoform expression of myofilament proteins. The level of phosphorylation in the contractile regulatory proteins decreases in the trout heart with cold acclimation (Klaiman et al., 2011; Klaiman et al., 2014), which likely contributes to the increase in activity rate of cardiac actomyosin-ATPase (Klaiman et al., 2011), and the increase in Ca2+ sensitivity of skinned trabeculae (Klaiman et al., 2014). These findings agree with studies on the effects of dephosphorylation on mammalian cardiac muscle. In the mammalian heart, dephosphorylation of the contractile regulatory proteins typically results in an increase in several functional properties of the muscle (increased actomyosin- ATPase and Ca2+ sensitivity) (Yang et al., 2000). Together, these studies suggest that altering the phosphorylation state of the contractile proteins is a regulatory mechanism for maintaining the contractile properties of the trout heart during thermal acclimation.

In mammals, protein kinase A (PKA) treatment of muscle typically increases the phosphorylation state of cTnI and cardiac myosin binding protein C (cMyBP-C) (Shaffer and Gillis, 2010; Gillis and Klaiman, 2011). In turn, this can act to reduce the Ca2+ sensitivity of the muscle and the rate of force redevelopment across the physiological range of the force-pCa curve. However, PKA phosphorylation of trout cardiac muscle causes a 25 % reduction in maximal force and a 46 % reduction in the rates of force re-development at maximal force, but not at submaximal Ca2+ concentrations (Gillis and Klaiman, 2011). These differences between mammalian and trout cardiac muscle may be due to a lack of PKA phosphorylation sites on the cardiac TnI and cMyBP-C in trout (Shaffer and Gillis, 2010; Kirkpatrick et al., 2011), which may explain the absence of increases in the amount of phosphorylated contractile proteins (Klaiman et al., 2011; Klaiman et al., 2014). These findings suggests that β-adrenergic stimulation does not have a direct effect on the contractile element, or at least not an effect

Page | 49

Thermal remodelling of the ectothermic heart that is comparable to that reported in mammals. The observed effects of increased β- adrenergic stimulation following cold acclimation may, therefore, be isolated to the Ca2+ handling proteins in trout cardiac muscle.

Another mechanism for regulating contractile function is differential expression of protein isoforms better suited to a particular physiological condition. The trout heart shows expression of both a slow skeletal-like TnT and a cardiac-like TnT (ssTnT and cTnT, respectively) (Klaiman et al., 2014). However, the expression pattern of these two isoforms did not change following thermal acclimation (Klaiman et al., 2014). Moreover, there have been three isoforms of TnI identified at the protein level in the trout heart and gene transcripts for seven different TnI isoforms expressed in the trout heart (Alderman et al., 2012). Cold acclimation causes four of these to increase (Alderman et al., 2012). In addition, recent work has identified transcripts for two isoforms of TnC in the trout heart and has shown that the expression of these is altered in response to cold acclimation (Genge et al., 2013). Together, these studies demonstrate that differential isoform expression of TnC and TnI may be involved in regulating cardiac contractile function in the trout following cold acclimation. With the expression of multiple isoforms of different Tn subunits there is the possibility that changes in the combination of the subunits within the Tn complex may contribute to changes in function. For example, in mammalian cardiac tissue, the replacement of cTnI with ssTnI within the complex results in a change in Ca2+ sensitivity of the muscle (Wolska et al., 2002). Therefore, the combination of different Tn subunits may have significant impact on function.

1. 4. 3. 3. Cardiac morphology

In addition to the changes at the subcellular and tissue level, thermal acclimation has been shown to affect both the size and morphology of the trout heart (Farrell et al., 1988b; Driedzic et al., 1996; Gamperl and Farrell, 2004; Klaiman et al., 2011; Klaiman et al., 2014). Cardiac hypertrophy can alter both the active and passive properties of the heart. A number of studies have shown increased atrial and ventricular mass in fish following chronic cold, which is likely reversed following chronic warm (Farrell et al., 1988b; Kent et al., 1988; Driedzic et al., 1996; Aho and Vornanen, 1998, 2001; Vornanen et al., 2005; Klaiman et al., 2011), goldfish (Hiroko et al., 1985), carp (Goolish, 1987), common carp (Young and Egginton, 2011). The increased ventricular mass is mainly attributed to an increase in myocyte size, suggesting it is a classic hypertrophic response (Driedzic et al., 1996; Aho and Vornanen, 1998; Vornanen et al., 2005; Klaiman et al., 2011). However, some studies also suggest growth myocyte hyperplasia (increase in cell number) in addition to hypertrophy (Farrell et al., 1988b; Sun et al., 2009). The genes involved in protein synthesis are up-regulated as a compensatory measure to hypertrophy of myocytes (Driedzic et al., 1996; Vornanen et al., 2005).

Page | 50

Thermal remodelling of the ectothermic heart

An increase in heart mass increases the amount of muscle mass (force generating units per cross section of the heart), thereby increasing the pressure generating capacity of the ventricle (Graham and Farrell, 1989; Tiitu and Vornanen, 2001). However, even in the absence of cold- induced cardiac hypertrophy cold acclimation of trout results in an increase in pressure generating capability of the ventricles across the pressure volume curve (Klaiman et al., 2014). This represents an increase in contractile force of the ventricle, which can be attributed to changes in ventricle morphology and the contractile element (Klaiman et al., 2014).

The dual myocardial layers of the fish heart mean changes in morphology can occur in the absence of overall cardiac hypertrophy (Driedzic et al., 1996; Klaiman et al., 2014). Following cold acclimation rainbow trout have been shown to have a reduced compact myocardium compared to warm-acclimated trout (Klaiman et al., 2011). A decrease in compact layer thickness without a significant change in relative ventricular mass suggests an increase in the proportion of spongy myocardium in the heart of cold-acclimated trout. It is hypothesized that this change in myocardial layer composition (increased spongy layer and decreased compact layer) may alter the functionality of the heart; changing the heart from a pressure pump at warm temperatures to a volume pump at cold temperatures. This change in ventricular morphology may help to generate more force per contraction because the spongy myocardium is made up of trabecular sheets that have been proposed to act as “contractile girders”, helping to pull the compact myocardium inward during contraction (Pieperhoff et al., 2009). In addition, the presence of the spongy myocardium is thought to be responsible for the extremely high ejection fraction of the trout heart (~80 %) compared to mammals (50-60 %), which do not contain a spongy myocardial layer. Together, the changes in myofibril function (described above) and changes in cardiac morphology following cold acclimation would enable the trout heart to contract more forcefully. This would be beneficial for a ‘volume pump’ type heart to maintain high ejection fractions and, thus, cardiac output. Changes in cardiac morphology will also affect the passive properties of the fish heart. According to the law of LaPlace, cardiac hypertrophy can be viewed as a compensatory measure to bring the wall tension back to a normal level and reduce the stress on each individual myocyte, compared to a non- hypertrophied heart developing the same pressure. Wall thickness can affect passive stiffness of the ventricle, therefore, hypertrophy or atrophy of the ventricle may influence diastolic filling.

Interestingly, Aho and Vornanen (2001) show a greater increase in atrial mass with cold- acclimation than ventricular mass. As hypertrophy can be due to stretch of myocardial wall, as shown in mammals (MacKenna et al., 1994), it is possible that increased cardiac preload associated with prolonged cold cause greater stretch of the atrium than the ventricle (Aho and Vornanen, 1999). The role of the atrium as a volume reservoir modulates diastolic filling may lead to overall cardiac enlargement (Aho and Vornanen, 1999).

Page | 51

Thermal remodelling of the ectothermic heart

1. 4. 3. 4. The extracellular matrix

The ECM matrix proteins contribute to the passive elastic properties of cardiovascular contraction, providing the shift from active force producing myofibrils to the passive force of the ECM (MacKenna et al., 1994). Collagen increases the passive stiffness of the chamber wall so excessive fibrosis of the myocardium can reduce chamber compliance (i.e. the change in pressure for a given change in volume) and chamber distensibilty (i.e. the fold change in cardiac compliance), which can have implications for diastolic filling (Collier et al., 2012b). In rainbow trout and carp, myocardial fibrillar collagen content has been shown to increase following cold acclimation (Pelouch and Vornanen, 1996; Klaiman et al., 2011), which is likely to protect the myocardium from the increased haemodynamic stress of pumping cold viscous blood. However, the opposite response has been observed in zebrafish (Johnson et al., 2014). This may be because an acute effect of cold temperature is to stiffen collagen fibrils. Therefore, in some fish a reduction in myocardial collagen fibrils may occur to maintain compliance of the ventricle at low temperatures.

1. 4. 3. 5. Length-dependent changes in force generation

An increase in end-diastolic volume can increase systolic contraction, and therefore SV, via the Frank-Starling response. At the cellular level, changes in the resting length of the sarcomere can affect the strength of contraction and, thus, the pressure generating capacity of the ventricle as sensitivity to Ca2+ ions increases with myocyte stretch (Asnes et al., 2006; Shiels and White, 2008). This causes a rise in central venous pressure by amplifying ventricular dilation (Shiels and White, 2008). Interestingly, Klaiman et al. (2014) demonstrated that the difference in developed pressure at higher balloon volumes between cold- and warm- acclimated samples was greater than at smaller balloon volumes, but did not find any differences between acclimation groups in the diastolic phase. One possible explanation for this result is that the cardiac muscle of thermally acclimated fish may respond differently to stretch. It is possible that length-dependent activation is more prominent in the trout heart following acclimation to cold temperatures. However, this hypothesis requires future investigation.

1. 4. 3. 6. Remodelling of cellular energetics

During cold-induced cardiac hypertrophy in fish, it is likely that FAO remains the primary energy production pathway, which is increased to meet the high energy requirement (Sephton and Driedzic, 1991; Bailey and Driedzic, 1993; Driedzic and Gesser, 1994; Sephton and Driedzic, 1995; Driedzic et al., 1996). There is an associated accretion of lipids, which increase the rate of complex lipid biosynthesis and aerobic based fatty acid metabolism (Dreidzic et al., 1996). However, there is also evidence to suggest a shift in the predominant energy production

Page | 52

Thermal remodelling of the ectothermic heart pathway to glycogen metabolism (Vornanen et al., 2005). The genes involved in protein synthesis and intermediary metabolism are up-regulated as a compensatory measure to hypertrophy of myocytes (Driedzic et al., 1996; Vornanen et al., 2005). Increased mitochondrial density in the oxidative capacities of individual mitochondria and adjustments of ADP affinities, metabolic depression can reduce tissue levels of mitochondrial enzymes and mitochondrial proton leak rates (Guderley and St-Pierre, 2002). However, metabolic processes to supply ATP appear adequate, and there is no increase in mitochondrial volume density or mitochondrial marker enzymes (Driedzic et al., 1996).

1. 5. CARDIAC REMODELLING IN TURTLES

1. 5. 1. Anatomy and Physiology of the Turtle Heart

The turtle heart forms part of a closed circulatory system and has three chambers; two atria and one undivided ventricle made primarily of trabeculated myocardium , that functions at low pressure (Bettex et al., 2014) (Figure 1. 10). The two atria provide a pulmonary flow, however, the undivided ventricle allows for oxygenated and deoxygenated blood to mix within the ventricle (Burggren and Pinder, 1991).

1. 5. 1. 1. The atria

The right atrium receives deoxygenated, venous blood from the body and preferentially distributes it to the pulmonary circulation. The left atrium receives oxygenated blood from the lungs and preferentially distributes it to the systemic circulation (Wang et al., 2002; Bettex et al., 2014). Initiation of each heart beat comes from the SAN, which lies in the right atrium and both atria contain fast conducting tissue (Dimond, 1959; James and Sherf, 1971; Valentinuzzi and Hoff, 1972; Yamauchi et al., 1974). The atria are subject to both sympathetic and vagal innervation, and have been shown to have different sensitivity to catecholamines (Heinbecker, 1931; Gilson, 1932, 1939; Dimond, 1959).

1. 5. 1. 2. The ventricle

The ventricle is undivided, with only partial separation between pulmonary and systemic flow by the ventricular septum (Burggren and Pinder, 1991; Bettex et al., 2014). The ventricle contains three sections, called cava; the cavum venosum, cavum ateriosum and cavum pulmonale (Figure 1. 10) (Bettex et al., 2014). The structure of the ventricle allows for a number of physiological features of the turtle heart (Kik and Mitchell, 2005). Firstly, blood can mix, with the possibility of transferring between pulmonary and systemic circulation. Secondly, variations in pulmonary and systemic resistance can alter pulmonary to systemic flow ratio, which can result in right to left (net systemic) or left to right (net pulmonary) cardiac shunt flows

Page | 53

Thermal remodelling of the ectothermic heart

(Hicks and Malvin, 1992; Hicks et al., 1996). Above the intraventricular septum lies a muscular ridge, which reduces the ability for ventricular blood to mix during systole (Bettex et al., 2014). The majority of the ventricle is composed of trabeculated myocardium, but has a coronary flow (Tillmanns et al., 1974) and there is compact myocardium around the cavum arteriousum, which may allow it to generate greater pressure (Bettex et al., 2014). Apnoea during diving cause’s bradycardia and pulmonary blood flow is decrease by increases in pulmonary resistance, resulting in a net systemic shunt flow (Hicks et al., 1996; Dominique et al., 2014). At rest, pulmonary resistance is reduced, allowing a net pulmonary cardiac shunt flow (Hicks and Malvin, 1992).

Figure 1. 10. The turtle heart. The turtle heart has two atria and one undivided ventricle. The right atrium receives deoxygenated venous blood and the left atrium receives oxygenated pulmonary blood. The ventricle contains 3 sections called cava; the cavum venosum, the cavum pulmonale and the cavum arteriosum. Deoxygenated and oxygenated blood can mix in the ventricle, arrows show blood flow, with red signifying oxygenated, blue signifying deoxygenated and purple showing mixing.

Page | 54

Thermal remodelling of the ectothermic heart

1. 5. 2. Cardiac Function in Turtles with Acute Temperature Change

Decreases in ambient temperature slows all physiological function, for freshwater turtles meaning they enter a state of fasting as temperature falls below 15 °C. As temperature drops, metabolic rate decreases by a Q10 of 2-3 (Herbert and Jackson, 1985a; Galli and Richards, 2012). However, turtles also show an exaggerated drop in metabolic rate as temperature continues to decrease due to an active suppression of metabolic rate (Rotermund and Priviter, 1972; Beall and Priviter, 1973; Overgaard et al., 2005; Galli and Richards, 2012). This inverse thermal compensation can result in a 12-fold decrease in metabolic rate (Gatten, 1974) and a 10-fold reduction in ATP demand, while aerobic metabolism may be suppressed by as much as 20-fold (to around 5 % of its original value) in the painted turtle (Herbert and Jackson, 1985a; Jackson, 2002). The huge reductions in aerobic capacity are possible thanks to the extreme anoxia tolerance of freshwater turtles (Jackson, 2002).

Decreases in metabolic rate means that only a small amount of oxygen is required by the tissues to continue aerobic respiration. As such, cold temperatures (> 5 °C) are associated with significant decreases in Q (Bethea, 1972; Herbert and Jackson, 1985b; Jackson, 1987; Hicks and Farrell, 2000a; Stecyk et al., 2004; Stecyk and Farrell, 2007; Stecyk et al., 2007a). Decreased Q is almost exclusively due to profound bradycardia at acute cold temperatures, with fH dropping from ~30 beats per minute at 25 °C to ~5-6 beats per hour at 3 °C (Bethea, 1972; Herbert and Jackson, 1985a). Low temperature causes a reduction in calcium sensitivity of the contractile apparatus. This negatively impacts on contractility and generation by reducing the number of actin-myosin cross bridges formed in the muscle. Ventricular twitch force and power output are also negatively affected by temperatures (Stecyk et al., 2007a).

However, SV is maintained with acute cold, despite its negative inotropic effects, which is critical to deliver the necessary metabolites and low levels of oxygen needed to maintain cellular functions (Farrell et al., 1994; Hicks and Farrell, 2000a; Stecyk et al., 2004). The maintenance of SV is partly possible due to decreases in systemic pressure and resistance, which reduce afterload pressure (Stecyk et al., 2007a). In addition there is increased sensitivity of the heart to cardiac load with acute cold, allowing preservation of SV via the Frank-Starling mechanism (Farrell et al., 1994).

Another direct effect of cold temperature is an increase in blood viscosity, which is consistent with other vertebrates (Rand et al., 1964; Langille and Crisp, 1980; Graham and Fletcher, 1983; Clarke and Nicol, 1993; Saunders and Patel, 1998). Blood viscosity is influenced by temperature, shear rate, haematocrit, red blood cell volume and plasma proteins and is particularly influential to blood flow rate (Chien et al., 1975; Saunders and Patel, 1998). For each 1 °C drop in temperature, blood viscosity has been shown to generally increase by around 3%. Blood acts as a non-Newtonian fluid, so the low shear rates (the velocity that blood interacts with another layer of blood or the vessel wall) brought about by bradycardia and low

Page | 55

Thermal remodelling of the ectothermic heart blood pressure at low temperature can act to further increase blood viscosity (Koutsouris et al., 1985; Saunders and Patel, 1998).

1. 5. 3. Cold Acclimation and Remodelling of the Turtle Heart

Freshwater turtles spend winter in water in a state of periodic inactivity (Ultsch, 2006). Cold temperature is the primary trigger for this cold dormancy and prepares the cardiac muscle for winter hibernation conditions (Stecyk et al., 2007c). Heart mass and relative ventricular mass either do not increase with cold acclimation, or if they do increase the change is only minimal (Overgaard et al., 2005).

1. 5. 3. 1. Remodelling contractile force

Autonomic control is lost with cold acclimated turtles (Hicks and Farrell, 2000b; Stecyk et al., 2004; Stecyk et al., 2008). Therefore, increases in cardiac force production are due to intrinsic electrophysiology of the myocytes, such as changes in E-C coupling, substantial modifications to cardiac APs and reduced ventricular density of LTCCs (Stecyk et al., 2007a; Stecyk et al., 2008). The AP duration is extended, which increases the amount of time Ca2+ can enter the cell and activate the contractile machinery (Driedzic and Gesser, 1985; Bers, 1991; Overgaard et al., 2005). Cold temperature has been shown to slow rate of contraction by ~10-fold and the rate of relaxation by ~15-fold (Driedzic and Gesser, 1994; Shi and Jackson, 1997; Hicks and Farrell, 2000a; Overgaard et al., 2005; Stecyk et al., 2007a). The increased duration for calcium influx broadly compensates for reduced sensitivity and, therefore, maximal twitch force is only moderately affected by the decreased temperature (Driedzic and Gesser, 1994). Thus, overall force production at low temperatures is relatively similar to those at high temperature; however, force development is much slower (Overgaard et al., 2005). The reduced rate of relaxation might be caused by a reduced clearance of cytosolic calcium ions via active pumps following contractions (Overgaard et al., 2005). A reduction in active transport of Ca2+ would be consistent with the reduced metabolic rate and suggest that the contractile apparatus of the myocyte does not compensate to prolonged cold temperature (Overgaard et al., 2005). The energy consumed for active transport of Ca2+ ions is reduced at low temperature, and may be due to an associated reduction in membrane leakage of calcium ions (Overgaard et

+ + al., 2005). With cold acclimation the density of Na and K pumps are also reduced per unit of tissue in the ventricle (Overgaard et al., 2005; Stecyk et al., 2007a), which results in a reduction in Na+ and K+ pumping activity and alterations in the cellular concentrations of the ions (Overgaard et al., 2005). In addition to the slower AP upstroke and longer duration, resting membrane potential is less negative in cold acclimated turtles (by ~ 18 to 29 mV).

Following cold acclimation twitch force is greater at cold temperatures than after acute warming (Overgaard et al., 2005). It, therefore, appears that the ventricle remodels with

Page | 56

Thermal remodelling of the ectothermic heart temperature so that maximum twitch force can be achieved at acclimation temperature (Overgaard et al., 2005). As a result maximal isometric forces of ventricular strips are similar independent of temperature (Shi and Jackson, 1997).

During winter, the freshwater bodies that turtles hibernate in often freeze over, reducing oxygen availability to the animals. Overwintering for turtles is often associated with chronic anoxic conditions in addition to chronic cold. Anoxia causes an increase in plasma pH, K+

+ - - concentration, Na concentration, HCO3 concentration, Cl concentration, lactate concentration and glucose remain constant, as do osmolarity and haematocrit (Overgaard et al., 2005). Hyperkalaemia gives a strongly negative inotropic effect and irregular contractions (57-97% reductions in twitch force). If the tissue is made anoxic, twitch force is again reduced (14-38%), similar to acidosis (15-50%), but is restored by adding adrenaline (5-19%). Increase in twitch force when extracellular calcium ion concentration is boosted, is minimal to none for cold acclimated animals at cold temperatures; however, it does cause an increase at warm temperatures (Overgaard et al., 2005).

1. 5. 3. 2. Remodelling of the passive properties of the heart

It is likely that during chronic cold conditions there are large increases in biomechanical stress on the heart due to slow, viscous blood flow. Very little is known about remodelling of passive properties of the reptile heart. Saunders and Patel (1998) found that whole blood viscosity of red-eared slider turtles, whether warm or cold acclimated, tended to be lower than in other vertebrate species when compared under similar conditions. The authors suggest that an innate lower blood viscosity in turtle species may compensate for the potentially damaging effects of increased blood viscosity and decreased shear rates experienced during winter hibernation (Saunders and Patel, 1998).

1. 5. 3. 3. Remodelling of cellular energetics

Overwintering for turtles is often associated with chronic anoxic conditions in addition to chronic cold. Together, these cause profound alterations in cellular energetics shifting the predominant energy producing pathways from aerobic FAO to anaerobic glycolysis (Beall and Priviter, 1973; Jackson, 2002)

In vivo 31P-NMR spectroscopy has shown that temperature-dependent differences in high- energy phosphate metabolism in cardiac tissue (Wasser et al., 1990; Jackson et al., 1995). Following cold acclimation, studies have shown increases in phosphocreatine, phosphomonoester, phosphodiester, free ADP (ADPf), and free adenosine monophosphate compared to warm acclimated animals (Stecyk et al., 2009). Furthermore, it is likely that cardiac ATP content is around 50 % lower in cold acclimated turtles (Stecyk et al., 2009).

Page | 57

Thermal remodelling of the ectothermic heart

2+ Increase in intracellular phosphate (Pi) causes a decrease in contraction force and Ca sensitivity (Stecyk et al., 2009). However, this effect is counteracted by increases in ADPf, which may explain the large increase in ADPf in cold-acclimated turtles. Similarly, a decrease in ATP hydrolysis will weaken cardiac performance as cardiac myocytes require a high potential for phosphorylation (by ratio of ATP to ADPf and Pi) to drive ATPase dependent reactions (Stecyk et al., 2009).

1. 6. SUMMARY

The rainbow trout and freshwater turtle adopt different strategies to seasonal temperature changes resulting in distinct cardiac remodelling phenotypes. The rainbow trout remains active throughout winter and, therefore, cardiac remodelling to prolonged cold temperature provides compensation for decreased contractile force and cardio-protection from the increased haemodynamic stress of pumping cold viscous blood. The freshwater turtle becomes inactive during winter cold and actively reduces metabolic processes. Cardiac remodelling, therefore, prepares the heart for periods of winter hypoxia or anoxia, while defending the necessary low levels of cardiac function and protecting the heart from the increased haemodynamic strain of high blood viscosity.

1. 7. AIMS

The overall aims of my PhD were to investigate the effects of thermal remodelling on the heart and cardiovasuclulature of two ectotherminc species; the rainbow trout (Onchorhychus mykiss) and freshwater the red-eared slider turtle (Trachemys scripta). These two animals were chosen due to their opposite overwintering strategies, with the trout remaining cold-active and the turtle becoming cold-dormant. Many of the active properties of the cardiovascular system of each has been previously studied, with interesting differences between them. However, the thermal remodelling of the passive properties of the heart has received less attention, providing an interesting avenue for new research and comparisons. To fully test the effect of thermal acclimation for rainbow trout we had three acclimation temperatures; chronic cold (5 °C), control (10 °C) and chronic warm (18 °C). For freshwater turtles we used two acclimation temperatures; chronic cold (5 °C) and control (25 °C). Unlike many of the previous studies of thermal acclimation of the cardiovascular system in these species, the aims of my studies were to focus on the passive properties of the heart across multiple layers of organisation. This work follows, primarily, from Dr Andy Fenna’s PhD (2009-2014), which he completed with Dr Holly Shiels as I began in the lab, and from the work by Klaiman et al. 2011.

In chapter 3 our aim was to investigate the effects of thermal remodelling on the rainbow trout ventricle, from the whole organ level down to mRNA expression. We assessed the morphological changes and changes in connective tissue content of the ventricle. To

Page | 58

Thermal remodelling of the ectothermic heart understand the functional consequences of tissue remodelling we generated ex vivo pressure volume curves in isolated ventricles and used atomic force microscopy (AFM) to assess tissue micromechanics.

In chapter 4 our aim was to investigate the alterations in ventricular tissue biochemistry with thermal acclimation in the rainbow trout. We used Fourier transform infrared (FTIR) imaging spectroscopy and histological tissue staining to assess changes in tissue biochemistry following chronic cooling and chronic warming, with particular interest to tissue metabolites.

In chapter 5 our aim was to investigate the effects of thermal remodelling on the rainbow trout atrium, from the whole organ level down to mRNA expression. We assessed the functional consequences of tissue remodelling, we generated ex vivo pressure volume curves in isolated atria and used AFM to assess tissue micromechanics. To understand the mechanism of these changes in function, we assessed changes in morphology, connective tissue content and the endogenous gelatinase activity of matrix metalloproteinase (MMPs) in the thermally acclimated atrium.

In chapter 6 our aim was to investigate the effects of thermal acclimation on the compliance and connective tissue content, regulation and organisation of the rainbow trout outflow tract. We assessed the functional effect on compliance by generating ex vivo pressure volume curves and then assessed connective tissue content and coherency, using tissue histology. To understand the change in connective tissue regulation we characterised the abundance and activation of specific MMPs, using SDS-PAGE gelatin zymography, and the endogenous activity of MMPs, using in situ zymography.

In chapter 7 our aim was to investigate the effects of thermal acclimation on the in vivo cardiovascular function and response to a cardiac volume load in freshwater turtles. We investigated in vivo cardiovascular function in anaesthetised cold-acclimated and control freshwater turtles at a common temperatures to differentiate the direct effects of acute temperature change from a remodelling of the cardiovasculature, both at baseline levels and in response to an in vivo volume load. We then generate ex vivo pressure volume curves in isolated ventricles and used tissue histology to assess connective tissue content.

In chapter 8 our aim was to investigate the effects of thermal remodelling on the passive properties of the freshwater turtle ventricle. We investigated changes in tissue micromechanics using AFM. We then assessed tissue collagen content and coherency using tissue histology. Finally, we assessed changes in connective tissue regulation by SDS-PAGE gelatin zymography, in situ zymography and RT-qPCR.

Page | 59

Thermal remodelling of the ectothermic heart

In chapter 9 our aim was to investigate the alterations in ventricular tissue biochemistry with thermal acclimation in the freshwater turtle. We used Fourier transform infrared (FTIR) imaging spectroscopy and histological tissue staining to assess changes in tissue biochemistry following chronic cooling and chronic warming, with particular interest to tissue metabolites.

The specific hypotheses and approaches used to test them are detailed in the subsequent results chapters.

Page | 60

Thermal remodelling of the ectothermic heart

2. GENERAL METHODS

As this thesis is presented in alternative format, each results chapter contains a detailed explanation of the specific methods used. However, I will use this general methods section to discuss the rationale for why these methods were chosen and provide any additional information that was not included in the corresponding results chapters. The methods are grouped into the techniques used in vivo, on the whole heart, in tissue sections and molecular techniques.

2. 1. IN VIVO CARDIAC FUNCTION

In chapter 5 I used in vivo techniques to assess cardiovascular function in anaesthetised freshwater turtles. The chapter contains a comprehensive description of the surgical techniques and protocol, so in this section I will discuss the rationale for the experiments.

The fundamental difference between this study and prior studies was that cardiovascular function was assessed at a common temperature to distinguish the effects of thermal acclimation from those of acute temperature (i.e. Q10 effects). We decided that using the control temperature as the common, ‘experimental’ temperature would be more appropriate than studying the effects at the cold acclimation temperature due to the body of literature on cardiovascular function in anaesthetised turtles at this temperature (Comeau and Hicks, 1994; Hicks and Comeau, 1994; Hicks et al., 1996; Crossley et al., 1998; Crossley et al., 2000; Overgaard et al., 2002; Galli et al., 2004; Joyce and Wang, 2014; Crossley et al., 2015) and to make the study more directly comparable to our studies in fish.

Another novel feature of this study was volume loading the heart by bolus injection directly into the jugular vein. To do this, we occlusively cannulated the external jugular vein, which is possible as the turtle has an internal and external jugular vein, so can maintain normal blood flow. We calculated the bolus injection to increase the venous return volume by 5-fold, which correlates with the large increase in stroke volume these animals may experience during breath-hold whilst diving (Burggren et al., 1997). The bolus injection allowed us to calculate in vivo sensitivity to cardiac preload, showing the same increase in sensitivity as shown with acute decreases in temperature using an in situ preparation (Farrell et al., 1994).

Page | 61

Thermal remodelling of the ectothermic heart

2. 2. WHOLE HEART

2. 2. 1. Isolation of the Fish Heart

Fish were killed by a cranial blow and severance of the spinal cord, and then placed, ventral side up, on a bespoke operating table. An incision was made along the abdomen and opened dorsally until the hepatic vein could be seen. The pericardium was exposed and opened via two dorsal incisions (one on either side of the heart, close to the body cavity wall) and the heart exposed without disturbing its blood supply. The chest was spread so that the whole heart could be visualized, and the dorsal aorta cut releasing blood pressure. The heart was then perfused with a physiological extracellular saline solution containing heparin (100 unit). The heart excision was completed by carefully cutting along the base of the sinus venosus, taking care not to cause any damage to the delicate atrium or any of the junction regions. The heart was then placed immediately into extracellular saline solution, so that any remaining blood could be ejected.

2. 2. 2. Ex Vivo Pressure-Volume Curves

In chapters 3, 5, 6 and 7, I generated pressure volume relationships during passive filling of the heart. The most common approach for accessing pressure volume relationships is the use of a Langendorff preparation. A balloon is fed into the left ventricle of a relaxed heart, which can then be filled with an incompressible fluid to modify volume and, hence, pressure in the heart. As a consequence of increasing volume, end-diastolic pressure is increased, causing a curve in the pressure recording that represents the end-diastolic pressure-volume relationship. In turn, this effect modifies the pressure developed by the heart during systole. Using a balloon means that volume can be altered easily allowing the preparation to be used in an isolated working heart. As the heart is still functioning, changes in the pressure-volume relationship for both systole and diastole can be mapped. However, there is controversy as to whether this preparation is appropriate due to the architecture of the fish heart (Figure 2. 1). In a mammalian ventricle, as the preparation was designed for, the myocardium is completely compact with a large lumen in the centre of the chamber. However, in the fish ventricle there are two distinct myocardial tissue types; compact and spongy myocardium. The spongy myocardium covers the majority of the fish ventricle, with only a small lumen present (Figure 2.1). As a result it is difficult to position correctly and altering the volume of the balloon may not accurately represent blood filling the ventricle. For these reasons, we adopted a more classical approach by simply filling the ventricle at a known rate. This method simulates diastolic filling of the heart, however, the limitation is that the systolic pressure volume relationship is lost. However, as my work primarily concentrates on changes in the passive properties of the heart, this approach was deemed an acceptable despite this limitation.

Page | 62

Thermal remodelling of the ectothermic heart

Figure 2.1. A SEM image of a dissected teleost heart permitting the two distinct types of myocardium to be seen. The inner spongy myocardium covers the majority of the inner ventricle, with only a small lumen. The spongy myocardium is covered by a dense compact myocardial layer. AT = atrium; AV = atrioventricular valves; BV = bulboventricular valves; BA = bulbus arteriosus Taken from Pieperhoff et al. (2002).

Each results chapter contains a full description of the specific methods and techniques used within it. However, for the general results section I have included a diagram of the experimental apparatus, which consisted of a syringe placed into a calibrated syringe pump, connected to a network of capillary tubes, with an attached pressure transducer at a right angle to the cannula to be placed into the heart (Figure 2. 2). The entire network was maintained at a constant height with the pressure transducer at the same level as where the heart would be positioned. The pressure transducer was calibrated against a vertical column of water, within the physiological ranges experienced by the heart. The cannula for the heart was fire-polished then fed into a bespoke made organ bath. The syringe and attached capillary tubes were filled with extracellular saline solution (NaCl, KCl, MgSO4, NaH2PO4, HEPES, Glucose and BDM) containing blue food colouring (Silverspoon, UK) (Figure 2. 3). The organ bath was also filled with extracellular saline solution (without food colouring) and temperature controlled at 10 °C.

Page | 63

Thermal remodelling of the ectothermic heart

Figure 2. 2. Pressure-volume curve apparatus and set-up. A pressure controlled syringe, containing physiological saline and blue dye, is connected to an isolated heart via cannulation into atrium via the sinous venosus. The heart is held in a temperature controlled water bath in physiological saline. The increasing pressure of the system is recorded by a pressure transducer connected to a computer by PowerLab software until rupture of the chamber. Taken with permission from Samuel (2010).

Figure 2. 3. Each chamber of the heart was filled, in series, with extracellular saline solution

(NaCl, KCl, MgSO4, NaH2PO4, HEPES, Glucose and BDM) containing blue food colouring. Filling was controlled by a syringe pump at a constant rate of 0.05 ml min-1.

Page | 64

Thermal remodelling of the ectothermic heart

2. 3. TISSUE TECHNIQUES

2. 3. 1. Tissue Histology

Formalin fixed tissue underwent a custom designed tissue processing stage (Micron spin, Fisher Thermo Scientific, UK) and was embedded in paraffin. Snap frozen tissue was stored in OCT at -80 °C.

2. 3. 1. 1. Haematoxylin and Eosin (H & E)

Haematoxylin and eosin are two stains that make up one of the most widely used histological staining techniques; H & E. Together these two stains provide a good overall representation of tissue, so I used this stain in chapters 4, 5 and 9. Haematoxylin is a natural stain obtained from logwood which must be oxidized to haematein and combined with a metallic salt, known as a mordant, to make it basic and able to stain tissue (Marshall and Horobin, 1972). For this reason, it is often combined with aluminium alum to form the tissue-mordant-haematoxylin linkage (Clark, 1974). Haematoxylin is a basic dye and, therefore, is used to stain acid (or basophilic) structures blue/purple. The acidic structures within a cell are DNA (heterochromatin and nucleolus) and the RNA in the ribosomes and the rough endoplasmic reticulum. During H & E staining the samples are counterstained with eosin, which is a negatively charged acidic dye. Most proteins within the cytoplasm are basic (or acidophilic) so eosin is used to stain these structures red/pink.

Protocol – 1. Paraffin embedded sections were de-waxed in xylene and hydrated in sequentially reduced concentrations of industrial methylated spirit (IMS) until distilled water. 2. Sections were then treated with haematoxylin, 5 % acetic acid and blueing agent before increasing concentrations of ethanol to 100 %. 3. Sections were then counter stained with eosin alcoholic 4. Sections were fixed with 100% ethanol, cleared with xylene and mounted in resinous medium.

Myocyte cross sectional area was analysed using ImageJ. The scale was set to match the scale bar of the image. Then the perimeter of circular myocyte bundles was highlighted using the polygon section tool and the area and diameter recorded. Only myocyte bundles that were circular in appearance were analysed as these were considered to lie in the correct orientation, i. e. a cross-section. 20 myocyte bundles were analysed per image, with 10 images per section on 6 sections through the chamber, for each individual.

Extra-bundular sinus was also analysed using ImageJ. Each image was converted to 16-bit and then black and white, making each pixel binary (Figure 2. 4). The image was then adjusted

Page | 65

Thermal remodelling of the ectothermic heart

so that the tissue area was white and the back ground black (Figure 2. 4), and the black area was quantified using a voxel counter plugin. For images at the edge of the tissue, where the background outside of the tissue area was visible, this area was discarded prior to image analysis.

Figure 2. 4. A representative image of fish atrial tissue used to quantify extra-bundular sinus. The image has been converted to binary with the tissue area in white and the extra-bundular sinus in black.

2. 3. 1. 2. Picro-Sirus Red

Picro-sirus red selective stains fibrillar collagen, which I performed in chapters 3, 5, 6, 7 and 8. Staining with this technique can only be performed on formalin-fixed paraffin embedded (FFPE) sections and not on frozen samples. When viewed under a bright-field microscope, collagen is selectively stained red on a yellow background. However, the main benefit of this stain is that when viewed under cross-polarized light large collagen fibres are bright yellow or orange and thinner ones, including reticular fibres, are green (Puchtler et al., 1973; Junqueira et al., 1979). The improved birefringency occurs because the elongated sirus red stain attaches to the collagen fibres in parallel (Junqueira et al., 1979). With alternative stains historically used to visualize collagen, such as van Gieson or Masson’s trichrome, thin reticular fibres may be either missed or obscured (Rich and Whittaker, 2005).

Protocol – 1. Paraffin embedded sections were de-waxed in xylene and hydrated in sequentially reduced concentrations of IMS until distilled water.

Page | 66

Thermal remodelling of the ectothermic heart

2. Samples were stained in picro-sirus red for at least one hour, at which point equilibrium is reached. 3. When removed, samples were given two washes in acetic acid (0.5 %) then dried. 4. Sections were rapidly fixed with 100% ethanol, cleared with xylene and mounted in resinous medium.

Collagen percentage of total tissue was analysed using ImageJ. Each image was captured in both bright-field and polarised light for the purpose of analysis. For both images the colour channels were split. For the bright-field image blue channel was kept and for the polarised image the red channel was kept. For both images a background image was subtracted, then the images were converted to black and white with the sample black and the background white (Figure 2. 5).

Figure 2. 5. A schematic diagram for quantification of picro-sirus red for one tissue image, using both a bright-field and polarised micrograph.

Page | 67

Thermal remodelling of the ectothermic heart

The black area of both images was quantified using a voxel counter plugin on ImageJ. The bright-field image was used to quantify total tissue area and the polarised light image used to quantify fibrillar collagen content. For fish, the compact and spongy myocardium were considered separately and any of the image outside the tissue area was discarded before analysis. For each sample, 3 image montages were taken along 3 cross-sectional transects across the tissue. 3 sections were averaged for each individual.

2. 3. 1. 3. Miller’s Elastic Stain

Miller’s elastic stain can be used to stain the elastic fibres of the extracellular matrix (ECM). Elastin is not found in the ventricle, but may be found in the atria and elastic components of the vessels. Therefore, I have shown Miller’s elastic staining in chapters 6 and 7. I chose this stain as other stains, such as acid fuchsin, had previously been tried by members of our lab, but were not very successful in fish cardiac tissue. The Miller’s elastic staining protocol I used in my thesis stains elastin fibres dark blue/black and collagen red on a yellow/brown background of the tissue.

Protocol - 1. Paraffin embedded sections were de-waxed in xylene and hydrated in sequentially reduced concentrations of IMS until distilled water. 2. Sections were oxidised in 1 % aqueous potassium permanganate for 5 minutes. 3. Sections were rinsed in tap water for 2 minutes. 4. The permanganate stain was removed by treatment with 5 % oxalic acid until the sections were bleached. Although the permanganate staining is not visible on the final stained sections it improves sharpness and intensity of the staining. 5. Sections were rinsed in tap water for 5 minutes. 6. Sections were rinsed in IMS for 2 minutes. 7. Miller’s elastin stain was filtered and slides were stained in Millers elastic for 2-3 hours in a fume hood. 8. Samples were rinsed in IMS for 30 seconds, until all stain was removed. 9. Samples were rinsed in water for 2 minutes. 10. Sections were counterstained with picro-sirus red stain for 30 minutes 11. Without washing in water, samples were rapidly dehydrated with 100 % ethanol, cleared in xylene and mounted with resinous medium.

Elastin fibres were quantified in a similar way to collagen fibres, however, only bright-field images were taken. When the colour channels were split the blue channel was used for total tissue area and the red channel used for elastin fibre content, in the same was as with the picro-sirus red stain.

Page | 68

Thermal remodelling of the ectothermic heart

2. 3. 1. 4. Oil Red O Stain

Oil red O staining was used in chapters 4 and 9 to assess tissue lipid content of tissue cryosections. The stain selectively stains lipid droplets red. A negative control slide was stained for each positive tissue section to allow semi-quantitative analysis of tissue lipid content.

Protocol – 1. Stock oil red O solution was made by adding 0.5 g of oil red O to isopropanol 2. Oil red O working solution was made by diluting 30 ml of the stock solution with 20 ml of distilled water and filtering into a Coplin jar. 3. Frozen sections were air-dried and briefly fixed in formalin. 4. Sections were washed in running tap water for 5 minutes. 5. Sections were rinsed in 60 % isopropanol. 6. Sections were stained in oil red O working solution for 15 minutes. 7. Sections were rinsed in 60 % isopropanol. 8. Sections were rinsed in distilled water. 9. Samples were mounted with aqueous mountant. 10. The edges of the slides were coated in nail polish to avoid sections drying out.

2. 3. 1. 5. Periodic Acid Schiff (PAS) Stain

Periodic acid Schiff (PAS) staining was used in chapters 4 and 9 to assess tissue glycogen content of FFPE tissue sections. The stain selectively stains glycogen droplets pink/purple. A negative control slide was stained for each positive tissue section to allow to semi- quantitatively analysis of tissue glycogen content.

Protocol – 1. Paraffin embedded sections were de-waxed in xylene and hydrated by reduced concentrations of IMS until distilled water. 2. Saliva was collected (preferably two people). 10 ml of saliva is enough for approximately 20 slides. 3. Saliva was diluted with distilled water to a 4:1 ratio (e.g. 40 ml water: 10 ml saliva) and placed saliva mix in water bath at 37 °C to maintain enzyme peak activity. 4. Filtered saliva was added to the negative control slides and incubated for 30 minutes. 5. Negative controls were washed in tap water for 3 minutes. 6. All samples were placed in 1 % Periodic acid for 10 minutes. 7. Samples were washed in tap water for 2-3 minutes. 8. Samples were placed in Schiff’s reagent for 10 minutes (samples will appear pink). 9. Samples were washed in running tap water for 2-3 minutes.

Page | 69

Thermal remodelling of the ectothermic heart

10. Samples were ounterstain with haematoxylin for 1.5 minutes. 11. Samples were ‘Blued’ in tepid water for 2-3 minutes. 12. Samples were dehydrated in progressively increasing concentrations of ethanol, clear in xylene and mount coverslips.

2. 3. 2. Fourier Transform Infrared (FTIR) Spectroscopy

I have used Fourier transform infrared (FTIR) imaging spectroscopy in both chapters 4 and 9 to analyse the biochemistry of tissue sections. In both of these chapters I have thoroughly detailed the methods and analyses used. I will use this section to give some further technical

details of the technique. Samples were mounted onto 10 mm diameter calcium fluoride (CaF2) transmission FTIR slides. Glass slides cannot be used as they absorbs infrared light. After mounting FFPE samples were kept in a bench top desiccator.

2. 3. 2. 1. Background to FTIR

Infrared radiation was discovered in 1800 by Sir William Herschel and occupies a region of electromagnetic spectrum ranging from a wavelength of 1 mm - 1.1 μm (12820 - 10 cm-1). This range can be split into 3 discrete zones named after their respective distance from the visible light region of the electromagnetic spectrum; near infrared, (4000-12500 cm-1), mid infrared (400 - 4000 cm-1) and far infrared (< 400 cm-1).

The energy of atoms or molecules within a system can be represented by quantized discrete energy levels and each atom or molecule must exist at one of these states in agreement with quantum theory. When molecules are irradiated by an infrared source the energy is absorbed.

With the exception of homonuclear diatomic compounds (such as O2 and N2), when a molecule absorbs infrared light it can increase its energy state and cause a net change molecular dipole, which is associated vibration and rotation activity of the molecule. Rotation about the bond axis requires little energy so FTIR is primarily based on molecular vibration. Molecules vibrate when IR radiation is absorbed with a frequency that matches one of the natural vibration frequencies within the molecule and each bond will have its own characteristic vibration based on its chemical constituents, in accordance with Hooke’s law;

1 휅 휐 = 휋√ 2 휇

where 푚 푚 휇 = 1 푚 + 푚2 where 휐 = frequency, 휅 =the force constant, and 휇 = reduced mass. The equation for Hook’s law shows that frequency will increase with bond stiffness, but is decreased by increasing

Page | 70

Thermal remodelling of the ectothermic heart mass. It follows that molecules with a lower atomic mass will absorb IR radiation at a high frequency than molecules with a greater atomic mass.

When molecules abide by Hook’s law, and behave like a perfect harmonic oscillator, they must remain at one of the evenly spaced discrete harmonic oscillation energy states and move stepwise between them. However, in reality molecular behaviour is more complex and is affected by intrinsic properties such as internuclear distance and dissociation energy. These additional features allow for anharmonic oscillation and overtone transitions to occur. During anharmonic oscillations, unlike harmonic, the gap between energy levels progressive decrease as the energy reaches dissociation energy.

These molecular vibrations may be stretching, whereby they alter specific bond lengths and change the inter-atomic distance, or they may be bending, whereby they alter the bond angle causing an atom to move out of its present plane. For diatomic molecules (such as O2, N2, CO or HCl), there are 3 degrees of translational freedom and two degrees of rotational freedom. However, there is only possibility for stretching or compression of the bond, so there is only one degree of vibrational freedom. O2 and N2 lack dipole moment and are, therefore, IR inactive, but CO and HCl give strong vibrational signature. Polyatomic molecules contain more than one atom (N) and so have 3N degrees of freedom. Triatomic molecules (three atoms) can be in linear or non-linear in form. For a linear molecule (e.g. CO2), in addition to the three degrees of translational freedom, there are two degrees of rotational freedom along the two axes. For the total degrees of freedom, the translational and rotational degrees are subtracted from the original 3N = 3N – 5. A linear triatomic molecule, therefore, has 4 modes of vibration; one IR active vibration from anti-symmetric stretch and one IR inactive vibration from symmetric stretch, as well as two doubly degenerate (same energy) bending vibrations (Figure 2. 6).

Figure 2. 6. The bending and stretching vibrational modes of CO2. Taken from University of Colorado (2001).

Page | 71

Thermal remodelling of the ectothermic heart

For a non-linear triatomic molecule (e.g. H2O) there are three degrees of rotational freedom, which account for movement around the x, y and z axis of the Cartesian plane (Figure 2. 7). Therefore, when the translational and rotational degrees are subtracted from the 3N = 3N – 6. A non-linear triatomic molecule, therefore, only has 3 modes of vibration, two of which are stretching motions and the other is a bending motion, all of which are IR active.

Figure 2. 7. The bending and stretching vibrational modes of H2O. Taken from University of Colorado (2001).

In more complex molecule, particularly those of common CH2 and CH3 groups, bending vibrations are common which can be further categorised as scissoring, rocking wagging and twisting (Figure 2. 8).

Figure 2. 8. The bending and stretching vibrations of a CH2 molecule. Taken from University of Colorado (2001).

2. 3. 2. 2. Recording a Spectrum

Infrared spectroscopy takes advantage the interaction between molecules and infrared radiation for chemical analysis. Every chemical bond has a unique frequency of vibration and, therefore, absorbs infrared radiation at a slightly different frequency. The wavelength of Page | 72

Thermal remodelling of the ectothermic heart molecular absorption, which occurs due to these different vibrational modes of chemical bonds, is measured in order to give information about molecular structure.

Spectra are commonly generated by transmission spectroscopy, where infrared light is passed through the sample and the detector records the remaining light. The amount of light absorbed can be calculated by; 퐼 퐴 = −푙표𝑔 퐼0

Where A is the absorbance and I is the intensity of light transmitted, making the ratio I/I0 tranmittance, T. The generation of a spectrum by FTIR is a two-step process. First, an interferogram must be generated; second, the interferogram must be converted to the time domain by a Fourier transform.

The interferogram is the recording of the intensities of wavelengths of infrared light that reach the detector. The device most commonly used to create and measure an infrared spectrum is a Michelson interferometer (Figure 2. 9), which uses the interference properties of photons to measure their frequency. Light from a polychromatic infrared source, commonly a blackbody radiator, called a Globar is collimated and directed to a beam splitter. The beam is split so that 50 % of the radiation is reflected to a fixed mirror, and is reflected back, while the remaining 50 % is transmitted to a moveable mirror and reflected back. Splitting the infrared beam gives a difference in the optical path length of the two beams before they recombine and pass through to the detector. The detector records a voltage at time points related to the mirror position giving a time dependent signal of cosine waves. When a sample is present, its absorption bands alter the background interferogram. However, it is not possible to interpret the spectrum at this stage. The second step is to convert the interferogram from the frequency domain to the time domain by a Fourier transform.

The Fourier transform is a mathematical operation that allows for interpretation of the spectra. In FTIR the interferogram is converted into the frequency domain, which enables the resolution of a wave into its frequency components. The Fourier transform allows a signal to be represented in the time domain (seconds), or in FTIR as wavenumbers (cm-1).

2. 3. 2. 3. FTIR Imaging Spectroscopy

Single point spectroscopy is the first and most simple method of obtaining a spectrum from a sample. The beam is focused to a point on the sample and a single point infrared beam is irradiated to gain a spectrum. The limitation of this technique is that the experimenter has to choose the area to be scanned, which may lead to bias or areas of importance being missed. However, spectrometers with a focal plane array (FPA) are able to create hyperspectral infrared maps (or images). The FPA is a 2 dimensional detector (normally 64 x 64, 128 x 128 or 256 x 256 elements) which allows multiple spectra to be captured in the x and y coordinated. Page | 73

Thermal remodelling of the ectothermic heart

It is in spectral mapping that the real power of the FTIR technique can be used, as it is possible to sample large areas of tissue and accurately compare spectra from areas of interest.

Figure 2. 9. A schematic of the Michelson interferometer. Infrared radiation is emitted from the source, commonly a blackbody radiator, and collimated and directed to a beam splitter. The beam is split so that 50 % of the radiation is reflected to a fixed mirror, and is reflected back, while the remaining 50 % is transmitted to a moveable mirror and reflected back. Splitting the infrared beam gives a difference in the optical path length of the two beams before they recombine and pass through to the detector. The detector records a voltage at time points related to the mirror position giving a time dependent signal of cosine waves. When a sample is present, its absorption bands alter the background interferogram.

2. 3. 2. 4. Interpretation of spectra

When considering a complex sample such as a biological tissue, the spectrum comprised the more simple vibrational modes of the individual molecular groups, as described above, and more complex skeletal vibrations that are generally associated with the molecular ‘back bone’ and involve significant vibration of the whole molecule. Skeletal vibrations are much more difficult to interpret and generally make up much of the so called fingerprint region. As a result Page | 74

Thermal remodelling of the ectothermic heart spectra can be analysed for either specific absorptions related to individual functional groups or by looking at the characteristic fingerprint the overall spectrum.

An infrared spectrum (Figure 2.10) displays frequency of absorption (wavelength or number) along the x axis, against intensity of absorbance (or transmittance; T) on the y axis.

Figure 2. 10. Representative spectra from 3 separate biological samples shown in red, green and blue. Although biological samples are complex, and therefore composed of many million vibrations, areas typically characteristic of overall lipid, protein and glycogen have been highlighted. The wavenumber region between 1800-2700 cm-1 has been removed as this region is not relevant in biological tissue.

The spectra obtained from a recording of mid-IR absorbance can be divided into 4 main regions. In each of which certain functional groups dominate the vibrational frequency. First is the 4000-2500cm-1 region. This area is dominated by stretching motion of hydrogen bonded functional groups (such as C-H, N-H and O-H). Generally, lighter atoms and stronger bonds vibrate at a higher frequency. This region is dominated by these bonds due to the light weight of the hydrogen atom compared to the relatively heavy atom and high strength of the bond. The higher frequency vibration means that a higher wavenumber and high energy are needed for vibrational transition. Second is the 2500-2000cm-1 region, dominated by triple bonds (such as C-C or C-N). The bonds are high strength, but vibrate at lower frequency than the previous group due to their increased mass. Third is the 1900-1500cm-1 region, dominated by double bonds, and finally the1400-400cm-1 region, dominated by single bonds due to their low bond strength.

Page | 75

Thermal remodelling of the ectothermic heart

For biological molecules, the mid-IR region (4000-600cm-1) is the most useful. In generating a spectrum the chemical information of all components of the sample are received simultaneously by the detector and averaged. This means that the spectra show an overview of numerous individual peaks superimposed onto each other to display the IR active components. Due to all of these individual bonds and functional groups vibrating at different wavelengths, a molecule with several will have a complex infrared spectrum and coupling of vibrations.

2. 3. 3. Atomic Force Microscopy (AFM)

I used atomic force microscopy (AFM) to assess tissue micromechanics in chapters 3, 5 and 7. AFM is a nanoscale resolution scanning probe microscopy. A cantilever spring tip is placed onto the tissue section (Figure 2. 4). The cantilever is moved by piezo-elements in the z- direction with nanometer resolution (Gautier et al., 2015). A laser beam is focused onto and reflected by the cantilever, then reflected by a mirror to be detected by the photodiode (Figure 2. 4). As the cantilever exerts force on the sample the laser is deflected which is detected by the photodiode. The magnitude of deflection is proportional to the force exerted by the tissue (Gautier et al., 2015).

Figure 2. 11. A schematic of the AFM apparatus. The cantilever is moved my piezo-elements in the z-direction with nanometer resolution. A laser beam is reflected by the cantilever and reflected by a mirror to be detected by the photodiode. As the cantilever exerts force on the sample the laser is deflected which is detected by the photodiode. The magnitude of deflection is proportional to the force exerted by the tissue.

Page | 76

Thermal remodelling of the ectothermic heart

AFM can be used for a number of applications with either a sharp tip, which is often used to measure tissue topography and ultrastructure, or a blunt tip, which is often used for nanoindentation, record tissue force curves. Each force curve shows the pressure required to make the nano-indentation, the resistance of the tissue and the elasticity (or recoil) of the tissue. We used a blunt tip to measure force curves of tissue cryosections using a nano-indentation protocol where we measured three 50 x 50 μm regions of the tissue, at an indent frequency of 1 Hz and lateral spacing of 2.5 μm. This protocol meant that we generated 400 force curved per 50 x 50 μm area. The data was then quality tested, whereby any force curves falling more than 2 standard deviations from the mean were discarded in order to account for failed indents or indents were the substrate was incorporated. To analyse the data, a baseline correction was applied to each force curve, before a force fit was applied which denotes the force required for that particular indentation. Mean force values were compared by GLM, with statistical significance considered at P < 0.005. By using AFM to measure tissue force curves we were able to measure fine differences on micromechanics of the tissue ultrastructure.

2. 3. 4. In Situ Zymography

I used in situ zymography to assess the localised gelatinase activity of MMPs in tissue cryosections in chapters 5, 6 and 8. By applying a fluorescently-quenched substrate to tissue sections it is possible to detect cleavage of gelatinase by endogenous MMPs. A fluorescent signal is produced which can be detected and semi-quantitatively analysed using fluorescence microscopy.

Protocol – 1. Stock agarose was made by dissolving 1 g of Low Gelling Temperature Agarose (Sigma- Aldrich, UK) into 100 ml of phosphate buffered saline (PBS) in a water bath at 80 °C until the solution became clear. The mixture was then decanted into 5 ml air-tight vials and stored at 4 °C. 2. Stock gelatine was made by dissolving 1 g of DQ gelatine (Porcine; Invitrogen, UK) into 1 ml

of dH2O and stored at 4 °C. 3. Stock DAPI solution was made by adding 5 mg of DAPI to 1 ml of PBS. 4. On the day of experiments 5 ml of stock agarose was melted in a water bath at 60 °C and then cooled to 37 °C. 5. Tissue cryosections were brought to room temperature over 30 minutes and washed in PBS to remove OCT. 6. 1 μl of stock DAPI solution was added to the 5 ml of agarose solution. 7. 0.5 ml of stock DQ gelatin solution was added to the agarose/DAPI solution, to make a 1:10 dilution. 8. Approximately 40 μl of agarose/DAPI/DQ gelatin solution was added to each cryosection and a coverslip placed on top.

Page | 77

Thermal remodelling of the ectothermic heart

9. Slides were incubated in the dark at 4 °C for 1 hour to allow gelling of the agarose. 10. For the reaction to proceed samples were incubated at room temperature in the dark for 18 hours. 11. After 18 hours samples were visualised and captured immediately, to avoid degradation of the fluorescence signal, using a fluorescence microscope with a green filter. 12. For negative controls, sections were treated with 4 % neutral buffered formalin for 15 minutes prior to adding gelatinase, and 1 mM of 1,10-phenanthroline was added to the agarose/DAPI/DQ gelatin solution. 13. To check for auto-fluorescence of tissue, agarose containing DAPI only was applied to sections.

2. 4. MOLECULAR TECHNIQUES

I did not perform either of the molecular techniques described in my thesis. Therefore, I will not provide any further technical details for these. However, I will provide a brief rationale for why these techniques were used.

2. 4. 1. SDS-PAGE Gelatin Zymography

We used SDS-PAGE gelatin zymography in chapters 5, 6 and 8 to characterise the abundance and activation of specific MMPs. MMPs are separated by their molecular weight in a gel, allowing characterisation of specific MMPs and whether they are activated or pro-domain forms to be determined. As a non-reducing buffer is used, the MMPs do not become inactivated in this technique. Therefore, during incubation they cleave the gel at their location. Staining with Coomassiee blue allows the areas that the MMPs have cleaved to be visualised and quantified.

2. 4. 2. Real-Time Quantitative PCR (RT-qPCR)

We used real-time quantitative PCR (RT-qPCR) in chapters 3, 5 and 8 to quantify mRNA expression of specific genes of interest. RT-qPCR uses fluorescently labelled mRNA primers so that mRNA replication can be quantified by fluorescence intensity. In chapters 3 and 5 we used ‘relative’ RT-qPCR, where mRNA expression was standardised to a stable housekeeping gene, in that case β-actin. In chapter 8 we used ‘absolute’ RT-qPCR, where absolute mRNA expression was measured by standardising the quantity of total mRNA to 5 ng.

Page | 78

Thermal remodelling of the ectothermic heart

3. The Dynamic Nature of Hypertrophic and Fibrotic

Remodelling of the Fish Ventricle.

This chapter is presented as a reprint of the following paper published in

Frontiers in Physiology;

Keen, A. N., Fenna, A. J., McConnell, J. C., Sherratt, M. J., Gardner, P., Shiels, H. A. (2016).

The dynamic nature of hypertrophic and fibrotic remodelling in the fish ventricle. Frontiers in

Physiology – Integrative Physiology. 6: 427; doi: 10.3389/fphys.2015.00427.

Page | 79

Thermal remodelling of the ectothermic heart

Page | 80

ORIGINAL RESEARCH published: 21 January 2016 doi: 10.3389/fphys.2015.00427

The Dynamic Nature of Hypertrophic and Fibrotic Remodeling of the Fish Ventricle

Adam N. Keen 1, Andrew J. Fenna 1, James C. McConnell 2, Michael J. Sherratt 2, Peter Gardner 3 and Holly A. Shiels 1*

1 Faculty of Life Sciences, University of Manchester, Manchester, UK, 2 Faculty of Medical and Human Sciences, Centre for Tissue Injury and Repair, University of Manchester, Manchester, UK, 3 School of Chemical Engineering and Analytical Science, Manchester Institute of Biotechnology, University of Manchester, Manchester, UK

Chronic pressure or volume overload can cause the vertebrate heart to remodel. The hearts of fish remodel in response to seasonal temperature change. Here we focus on the passive properties of the fish heart. Building upon our previous work on thermal-remodeling of the rainbow trout ventricle, we hypothesized that chronic cooling would initiate fibrotic cardiac remodeling, with increased myocardial stiffness, similar to that seen with pathological hypertrophy in mammals. We hypothesized that, in contrast to pathological hypertrophy in mammals, the remodeling response in fish would be plastic and the opposite response would occur following chronic warming. Rainbow trout held at 10◦C (control group) were chronically (>8 weeks) exposed to cooling (5◦C) or Edited by: ◦ Ovidiu Constantin Baltatu, warming (18 C). Chronic cold induced hypertrophy in the highly trabeculated inner layer Camilo Castelo Branco University, of the fish heart, with a 41% increase in myocyte bundle cross-sectional area, and an Brazil up-regulation of hypertrophic marker genes. Cold acclimation also increased collagen Reviewed by: deposition by 1.7-fold and caused an up-regulation of collagen promoting genes. In Milan Stengl, Charles University, Czech Republic contrast, chronic warming reduced myocyte bundle cross-sectional area, expression of Bjarke Jensen, hypertrophic markers and collagen deposition. Functionally, the cold-induced fibrosis and University of Amsterdam, Netherlands hypertrophy were associated with increased passive stiffness of the whole ventricle and *Correspondence: Holly A. Shiels with increased micromechanical stiffness of tissue sections. The opposite occurred with [email protected] chronic warming. These findings suggest chronic cooling in the trout heart invokes a hypertrophic phenotype with increased cardiac stiffness and fibrosis that are associated Specialty section: with pathological hypertrophy in the mammalian heart. The loss of collagen and increased This article was submitted to Integrative Physiology, compliance following warming is particularly interesting as it suggests fibrosis may a section of the journal oscillate seasonally in the fish heart, revealing a more dynamic nature than the fibrosis Frontiers in Physiology associated with dysfunction in mammals. Received: 05 November 2015 Accepted: 27 December 2015 Keywords: compliance, fibrosis, heart, stiffness, temperature acclimation, phenotypic plasticity Published: 21 January 2016 Citation: Keen AN, Fenna AJ, McConnell JC, Sherratt MJ, Gardner P and Shiels HA Abbreviations: AFM, atomic force microscopy; ANP, atrial natriuretic peptide; BNP, brain natriuretic peptide; Col1a1, (2016) The Dynamic Nature of collagen I alpha 1; Col1a2, collagen I alpha 2; Col1a3, collagen I alpha 3; ECM, extracellular matrix; GLM, general linear Hypertrophic and Fibrotic Remodeling model; MLP, muscle LIM protein; MMP2, matrix metalloproteinase 2; MMP9, matrix metalloproteinase 9; MMP13, matrix of the Fish Ventricle. metalloproteinase 13; NFAT, nuclear factor of activating T; PNCA, proliferating cell nuclear antigen; RCAN1, regulator Front. Physiol. 6:427. of calcineurin 1; RT-qPCR, real-time quantitative PCR; SMLC2, small myosin light chain 2; TIMP2, tissue inhibitor of doi: 10.3389/fphys.2015.00427 metalloproteinase 2; VEGF, vascular endothelial growth factor; VMHC, ventricular myosin heavy chain.

Frontiers in Physiology | www.frontiersin.org 1 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

INTRODUCTION of myocardial force (Collier et al., 2012). However, fibrosis can also cause excessive stiffness, which reduces diastolic and Chronic changes in pressure or volume load can cause the systolic function and increases the chance of arrhythmias vertebrate heart to change in size, form and function (Clark (Chapman et al., 1990). Fibrosis is largely absent in mammalian and Rodnick, 1999; Opie et al., 2006). The heart remodels in physiological hypertrophy (Bernardo et al., 2010). an attempt to ensure appropriate output, commonly achieved Pathological and physiological hypertrophy are also through hypertrophy which can be defined as an enlargement of differentiated by markers of myocardial stretch, including a part or the whole of anorgandue to anincrease in thesize of its atrial natriuretic peptide (ANP) and brain natriuretic peptide constituent cells (Dorn, 2007). Hypertrophy of the mammalian (BNP), and the activation of the signaling pathways that promote left ventricle can improve cardiac performance to meet increased growth, such as the fetal gene program (de Bold and de Bold, demands such as those occurring with pregnancy or following 2005; Bernardo et al., 2010). Up-regulation of genes associated exercise training (Mone et al., 1996). This “physiological” with mammalian pathological hypertrophy are also found hypertrophy increases ventricular wall thickness in line with in hypertrophic fish hearts following chronic cold or stress chamber radius causing both stroke volume and systolic pressure (Vornanen et al., 2005a; Johansen et al., 2011). Thus, thermal to increase, improving overall cardiac output (Dorn, 2007; remodeling of the fish heart appears to exhibit characteristics Bernardo et al., 2010). Importantly, under most conditions, of both physiological and pathological hypertrophy of the this remodeling is transient and regresses when the stimulus is mammalian heart. removed (Bernardo et al., 2010). Hypertrophy can also occur Here, using acclimation temperatures to simulate seasonal in response to chronic pathological stressors like hypertension. temperature change, we investigate the effects of chronic cooling Here, ventricular wall thickness increases, but the luminal (from 10 ± 1◦C to 5 ± 1◦C) and chronic warming (from volume is reduced (Dorn, 2007) allowing systolic pressure to be 10 ± 1◦C to 18 ± 1◦C) on the rainbow trout ventricle maintained, but at the expense of stroke volume which can lead to to determine the extent of temperature-induced connective both systolic and diastolic dysfunction. Pathological remodeling tissue remodeling and the passive properties of the ventricle is persistent and is associated with various cardiomyopathies across multiple levels of organization. This approach extends including myocardial infarction, arrhythmia, and sudden death our previous study which described thermal remodeling (Bernardo et al., 2010). of the active properties of the salmonid heart (Klaiman The hearts of non-mammalian vertebrates also remodel to et al., 2011). We hypothesized that chronic cooling would meet changing systemic demands. For example, the ventricular increase myocardial stiffness, fibrosis and up-regulation of mass of the Burmese python, Python molurus, can increase factors involved in pathological hypertrophy in mammals. To by 40% after feeding and then return to “normal” following understand functional consequences of remodeling we used digestion (Andersen et al., 2005). The hearts of many fish atomic force microscopy (AFM), to determine micromechanical also show intermittent remodeling in response to seasonal ventricular stiffness, and generated ex vivo pressure-volume temperature change with chronic cooling during winter curves, to determine ventricular chamber compliance. As trout triggering ventricular hypertrophy (Farrell et al., 1988; Graham experience intermittent temperature change, we were particularly and Farrell, 1989; Tervonen et al., 2001; Klaiman et al., 2011; interested in variable remodeling following both warming and Shiels et al., 2011). The hypertrophic trigger in this model is cooling. We found chronic cold induced a remodeling phenotype ◦ thought to be the increased viscosity of blood at cold (<6 C) with aspects analogous to pathological hypertrophy in mammals, temperatures and the hemodynamic stress of pumping this particularly relating to chamber stiffness and collagen deposition. viscous blood (Graham and Farrell, 1989; Clark and Rodnick, The opposite response was found following chronic warming 1999). Importantly, the cold-induced hypertrophic phenotype suggesting a reversible phenotype. The fish heart could thus is thought to regress during warming in summer (Klaiman provide a model to investigate the regression of fibrotic cardiac et al., 2011). The temperature-induced cardiac remodeling in hypertrophy. fish has, therefore, generally been considered analogous to the physiological cardiac remodeling that occurs in mammals following non-pathological hypertrophic stimuli. MATERIALS AND METHODS Stimuli known to trigger pathological hypertrophic remodeling in mammals may also arise in fish following Ethical Approval chronic cold. The acute effect of low temperature is to decrease All husbandry and housing conditions were in accordance with contractile function (Aho and Vornanen, 1999), through direct the University of Manchester handling protocols and adhere to ◦ (i.e., Q10; rate of reaction change over a 10 C temperature the UK Home Office legislation. All experimental procedures change) effects on the ion pumps and channels underlying were approved by the University of Manchester ethical review cellular excitation-contraction coupling (Vornanen et al., 2002), committee. and to reduce the Ca2+ sensitivity of contractile elements (Gillis et al., 2000) which must be compensated for. Furthermore, there Experimental Animals is evidence for ventricular fibrosis in fish following chronic cold Sexually mature female rainbow trout (Klaiman et al., 2011). In mammals, increased collagen fibril (Onchorynchus mykiss; n = 47; morphometric data in Table 1) density can strengthen chamber walls and improve transduction were purchased from Dunsop Bridge Trout Farm (Clitheroe,

Frontiers in Physiology | www.frontiersin.org 2 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

TABLE 1 | The gross morphological parameters for cold acclimated (5◦C), sectioned at 10 µm (Leica CM3050S cryostat, Leica, Wetzlar, ◦ ◦ control (10 C) and warm acclimated (18 C) rainbow trout. Germany), mounted onto glass slides (Super frost plus, Thermo

Cold Control Warm Fisher Scientific, Waltham, MA, USA) and stained using acclimated acclimated Masson’s trichrome (see Klaiman et al., 2011, for details). Previously, work on trout spongy tissue showed that cross- Mass (g) 480.8 ± 64.3 526.1 ± 42.8 524.8 ± 60.2 sections of single ventricular myocytes were not visible with Heart mass (g) 0.94 ± 0.09 1.15 ± 0.12 0.99 ± 0.11 Masson’s trichrome staining. Rather, the visible structures were RHM 0.0021 ± 0.00010 0.0022 ± 0.000086 0.0019 ± 0.000079 bundles of myocytes comprised of ∼10–15 individual cells (gmass−1) (see Klaiman et al., 2011). The myocyte bundle cross-sectional Ventricular 0.59 ± 0.075 0.68 ± 0.11 0.62 ± 0.073 area can, therefore, be used as a proxy for myocyte cross- mass (g) sectional area. Compact thickness, the cross-sectional area of RVM 0.0013 ± 0.000066 0.0012 ± 0.00011 0.0012 ± 0.000052 myocyte bundles and extra-bundular sinus in the spongy layer −1 (gmass ) were quantified using ImageJ software (Schneider et al., 2012). RHM, relative heart mass; RVM, relative ventricular mass. Values given are mean ± S. E. Masson’s trichrome stains amorphous collagen bluish/purple Significance was determined by GLM with Tukey post-hoc test for comparision between (Figures 1A,B) and this was semi-quantified using ImageJ and the groups (P < 0.05), n = 47 for each group. the “threshold colour” plugin in binary mode (Klaiman et al., 2011). For morphometric analysis of compact layer thickness, UK), housed on a 12h light: 12h dark cycle in ∼500 L re- myocyte bundle cross-sectional area and extra-bundular sinus circulated aerated fresh water tanks at 10 ± 1◦C and fed to eight sections were analyzed per individual fish. For compact satiation 3 times per week. Water quality was ensured with 30% layer thickness, 20 measurements were taken per image. For water changes 3 times per week and regular tests for temperature, measurement of cross-sectional area of myocyte bundle area, on pH, nitrates and nitrites. Fish were held under these conditions each tissue section three separate image montages were taken for a minimum of 2 weeks before being randomly assigned to along transects across the full diameter of the cross section. In one of three acclimation groups; cold (5 ± 1◦C), control (i.e., no each image trabeculations were chosen for measurement only if change; 10 ± 1◦C), or warm (18 ± 1◦C). These temperatures they were in the transverse plane, i.e., the image showed a cross- are based on previous literature which describes the cardiac section of the trabeculations making it circular in appearance. remodeling response in salmonids (Klaiman et al., 2011). Water For extra-bundular sinus the non-tissue area of each image was temperature of the warm and cold acclimation groups was measured. changed by 1◦C per day until desired temperature was reached Fibrillar collagen and elastin content were analyzed semi- and then held at that temperature for a minimum of 8 weeks quantitatively following Graham et al. (2011). Briefly, formalin- before experimentation. The photoperiod for the cold acclimated fixed tissue samples were processed, embedded in paraffin wax, animals was changed to 8h light: 16h dark cycle to simulate sectioned at 5 µm (Leica RM2255 microtome, Leica, Wetzlar, winter (Graham and Farrell, 1989). Germany) and mounted onto glass slides. Serial sections from each sample were stained with picro-sirus red for collagen Tissue Processing (Junqueira et al., 1979) and Miller’s elastic stain for elastin (Miller, Fish were killed by a blow to the head followed by severance of the 1971). Picro-sirus red images were quantified using polarized spinal cord and destruction of the brain. The heart was excised, light microscopy and Miller’s elastic images were quantified rinsed in phosphate buffered saline and weighed. The trout using bright-field microscopy. Mean fibrillar collagen and elastin ventricle is composed of two distinct myocardial layers, a highly contents were expressed as a percentage of total tissue cross trabeculated inner spongy layer which makes up the majority of sectional area, excluding the epicardial surface, determined using the organ (>80% in adult fish) and a thinner outer compact layer ImageJ. Three tissue sections were considered for each individual (Farrell and Jones, 1992; Pieperhoff et al., 2009). The apex of to ensure consistency in measurements. On each tissue section each ventricle was removed, the spongy tissue “scooped out” of three separate image montages were taken along transects across the compact and stored at −80◦C for quantitative real-time PCR the full diameter of the cross section. All histological analysis was (RT-qPCR). The remainder of the ventricle was bisected down conducted blind to the acclimation group and in all cases these the sagittal plane with one half snap frozen in OCT (Thermo tissue sections were taken from the central 50% of the ventricle. Fisher Scientific, Waltham, MA, USA) by immersion in liquid nitrogen cooled 2-methylbutane (Sigma-Aldrich, St. Louis, MO, Quantitative Real-Time PCR USA) and stored at -80◦C. The other half was fixed in 10% neutral As previous work has shown that cold-induced hypertrophy in buffered formalin solution (Sigma-Aldrich, St. Louis, MO, USA) fish occurs primarily in the spongy layer of the heart (Klaiman and embedded in paraffin wax so that sections would be cut in et al., 2011) and the compact layer was thin (∼550 µm; making the transverse/axial plane. up ∼17% of the ventricular area in the hearts used in this study; Poupa et al., 1974) all qRT-PCR was performed on Histology spongy myocardial tissue only. Transcript abundance of genes All sections were cut to include both the compact and spongy associated with muscle growth (ventricular myosin heavy chain; layer so that differential remodeling between the ventricular VMHC, muscle LIM protein; MLP, and small myosin light layers could be evaluated histologically. Frozen tissue was chain 2; SMLC2), hyperplasia (proliferating cell nuclear antigen;

Frontiers in Physiology | www.frontiersin.org 3 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

FIGURE 1 | Ventricular remodeling. Representative Masson’s trichrome stained cryo-sections for (A) cold (5◦C) and (B) warm (18◦C) acclimated rainbow trout show remodeling of the compact (c) and spongy (s) myocardium. Quantification of (C) compact layer thickness, (D) spongy myocyte bundle cross sectional area and (E) spongy layer extra-bundular sinus show remodeling with temperature acclimation. These differences are supported by (F) mRNA expression of gene markers of muscle growth (VMHC, MLP and SMLC2), hyperplasia (PCNA) and angiogenesis (VEGF) in the spongy myocardium. (G) mRNA expression of hypertrophic gene markers (ANP and BNP) and the pro-hypertrophic NFAT signaling pathway (RCAN1) in the spongy myocardium with cold (5◦C; blue), control (10◦C; green) and warm (18◦C; red) acclimation (n = 7 fish for each acclimation group; 3 replicates for each animal were averaged for both histology and qPCR). Values presented are mean ± S. E. Significance was assessed by GLM with Holm-Sidak post-hoc test. Significance between groups is shown by dissimilar letters (P < 0.05).

PCNA), angiogenesis (vascular endothelial growth factor; and heart failure (ANP and BNP) and pro-hypertrophic nuclear VEGF), collagen I (Col1a1, Col1a2, and Col1a3), connective factor of activating T (NFAT) signaling mediator (regulator tissue regulators (MMP2, MMP9, MMP13, and TIMP2), stretch of calcineurin; RCAN1) were quantiﬁed in the ventricles of

Frontiers in Physiology | www.frontiersin.org 4 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

fish from cold, control and warm acclimated groups (n = 7 Ex vivo Passive Pressure-Volume Curves ventricles for each temperature). RNA was extracted from 5 mg of The intact isolated heart was placed into an organ bath containing spongy tissue (RNeasyMicrokit, Qiagen, Venlo, NL) and amount Ringers solution [(in mM) 150 NaCl, 5.4 KCl, 2.0 CaCl2, 1.5 and quality was determined (NanoDrop ND-1000, NanoDrop, MgSO4, 0.4 NaH2PO4, 10 HEPES, 10 Glucose at a pH of Wilmington, DE, USA). An RNA concentration of 200 ± 50 ng pH7.7 with NaOH at room temperature] at 10 ± 1◦C to µl−1 was used to make cDNA with SuperScript III First Strand which 20 mM BDM (2, 3 butanedione monoxime) was added Synthesis System (Invitrogen, Carlsbad, CA, USA). SYBR Green to prevent active cross-bridge cycling. Pressure-volume curves I pre-mixed chemo-technology was used for qPCR. qPCR was from ventricles from each acclimation group were generated carried out in a 7900 HT sequence detection system (Applied at a common temperature, of 10 ± 1◦C, to isolate the effects Biosystems, Carlsbad, CA, USA) and cycle threshold (Ct) values of chronic remodeling on myocardial stiffness from the acute generated by the qPCR machine were multiplied by primer effects of temperature. A cannula was fed through the atrium efficiency (Pe) to determine gene expression levels (expression into the ventricular lumen and secured at the atrial-ventricular (e) = 1/(Ct ∗ Pe). All primers were taken from published work junction, using 0–0 silk thread (Harvard Apparatus, Holliston, (e.g., Johansen et al., 2011) or were designed using Primer 3 from MA, USA). An atraumatic clamp was placed at the bulbus- mRNA sequences available on PUBMED. Specific marker genes ventricular junction making the ventricle a sealed chamber with and primers are in Table 2. All expression levels were normalized the cannula inside. The cannula was connected to a syringe to housekeeping gene β-actin to determine absolute expression pump (INFORS AG, Bottmingen, CHE), in series with a pressure levels for comparison at each acclimation temperature. We tested transducer, containing 10 ± 1◦C Ringer solution with BDM and three housekeeping genes; β-actin, GAPDH and DNAJ1, and a small amount of blue food coloring (Silverspoon, London, β-actin had most stable expression in relation to temperature UK). The pressure transducer was calibrated daily against a static acclimation, as found previously (e.g., Johansen et al., 2011). water column and recorded at 1000 Hz (Chart5, PowerLab, ADI

TABLE 2 | The speciﬁc marker genes with primers used for quantitative real-time PCR.

Gene Primer pair GenBank accession Function/marker number

VMHC 5′ – TGCTGATGCAATCAAAGGAA – 3′ AY009126.1 Cardiomyocyte hypertrophy 3′ – GGAACTTGCCCAGATGGTT – 3′ MLP 5′ – AGTTCGGGGACTCGGATAAG – 3′ NC007118.6 (Danio rerio) Cardiomyocyte hypertrophy 3′ – CGCCATCTTTCTCTGTCTGG – 5′ SMLC2 5′ – GACAAGTTCA – 3′ NM001124678.1 Cardiomyocyte hypertrophy 3′ – GGTTCTTGTAGTCC – 3′ VEGF 5′ – AGTGTGTCCCCACGGAAA – 3′ AJ717301.1 Angiogenesis 3′ –TGCTTTAACTTCTGGCTTTGG – 5′ PCNA 5′ – AGCAATGTGGACAAGGAGGA – 3′ EZ763721.1 Cardiomyocyte hyperplasia 3′ – GGGCTATCTTGTACTCCACCA′ Col1a1 5′ – GCTTTTGGCAAGAGGACAAG – 3′ NM001124177.1 Fibrosis 5′ – GCAGATAACTTCGTCGCACA – 3′ Col1a2 5′ – GGCTGATCGGCTCTGTACTC – 3′ NM001124207.1 Fibrosis 3′ – TGGCTCTGCTGGTATCACTG – 3′ Col1a3 5′ – CCCTGCTTTTTATGGTTGGA – 3′ NM001124206.1 Fibrosis 3′ – GCAGGGTTCTGGTTTCCATA – 5′ MMP2 5′ – TGTATTGGGCAACATCAGGA – 3′ NC007118.6 (Danio rerio) Inhibit fibrosis 3′ – CCCAGGAGACGATAGTCCAA – 5′ MMP9 5′ – GGTCCAGTTTTCGTCATCGT – 3′ NM001124370.1 Inhibitfibrosis 3′ – AGACATGGGAGCCTCTCTGA – 5′ MMP13 5′ – TCTGATGTGGTTTGCTGCTC – 3′ NC007121.6 (Danio rerio) Inhibit fibrosis 3′ – CAGATAAGCCCGACCCTACA – 5′ TIMP2 5′ – CAGGCCATCCACCTACTGTT – 3′ NC007123.6 (Danio rerio) Inhibit MMPs 3′ – TGTTGCTCTCTTGCATACGG – 5′ ANP 5′ – CCACAGAGGCTCTCAGACG – 3′ NM001124211.1 Stretch/heart failure 3′ – ATGCGGTCCATCCTAGATC – 5′ BNP 5′ – TGGCCTTGTTCTCCTGTTCT – 3′ NM001124226.1 Stretch/heart failure 3′ – GGAGACTCGCTCAACCTCAC – 5′ RCAN1 5′ – AGTTTCCGGCGTGTGAGA – 3′ BC076439.1 (Danio rerio) NFAT-activity/cardiomyocyte hypertrophic signaling 3′ – GGGGACTGCCTATGAGGAC – 5′ β-actin 5′ – AGAGCTACGAGCTGCCTGAC – 3′ NM001124235.1 Control/housekeeping 3′ – GTGTTGGCGTACAGGTCCTT – 5′

Frontiers in Physiology | www.frontiersin.org 5 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

Instruments, Dunedin, New Zealand). Ringer solution with BDM P < 0.05, except for atomic force curves where signiﬁcance was was pumped into the ventricle at 0.05ml min−1 until maximum considered at P < 0.005. Values are presented as mean ± S. E. volume was achieved, determined by visual leak of the saline throughout except for atomic force curves where values are mean containing blue dye and a drop in the pressure trace. ± S. D. Statistical details are provided in the ﬁgure legends.

Atomic Force Microscopy (AFM) RESULTS Frozen ventricular tissue was sectioned at 5 µm (Leica CM3050S cryostat, Leica, Wetzlar, Germany) and mounted onto calcium Thermal Remodeling of Ventricular Muscle fluoride (CaF2) slides. Excess OCT was removed with distilled Thermal acclimation caused differential and opposite remodeling water and the slides were left to dry for ∼12 h. This methodology of the two myocardial layers within the fish ventricle is consistent with previous work (Kemp et al., 2012; Wallace et al., (Figures 1A–C) which is in line with previous findings (Gamperl 2012) which describes how tissue sections are best preserved and Farrell, 2004; Klaiman et al., 2011, 2014). Compared to dehydrated, with rehydration performed when nanomechanical controls, the compact layer (Figures 1A,B) was 4% thinner after measurements are required. Micro-indentation was carried cooling and 9% thicker after warming (P < 0.05; Figure 1C). out using a Bioscope Catalyst AFM (Bruker, Coventry, UK) However, there was no difference in total ventricular mass or mounted onto an Eclipse T1 inverted optical microscope ventricular mass relative to body mass (RVM) between the three (Nikon, Kingston, UK) fitted with a spherically tipped cantilever temperature acclimation groups (Table 1). Changes in compact (nominal radius and spring constant of 1 µm and 3 Nm−1 layer thickness with no change in overall ventricular mass respectively, Windsor Scientific Ltd., Slough, UK) running indicates compensative remodeling of the spongy myocardial Nanoscope Software v8.15 (Bruker, Coventry, UK). The local layer, which was detected as an increase in the cross-sectional reduced modulus was determined for each of 400 points in a area of the myocyte bundles that make up the spongy trabeculae 50 × 50 µm region, indented at a frequency of 1 Hz with lateral (Klaiman et al., 2011). Cross-sectional area of cold-acclimated spacing of 2.5 µm. The extend curve was used in conjunction myocyte bundles was 83% greater than warm-acclimated bundles with a contact point based model to calculate the reduced (P < 0.05; Figure 1D) and correlated with a reduced sinus space modulus for each indentation (Crick and Yin, 2007). For each between bundles (P < 0.05; Figure 1E). Cold-induced spongy biological sample, 400 force curves were collected at three distinct hypertrophy and warm-induced spongy atrophy were supported 50 µm2 regions. Once all 400 force curves had been generated, by changes in mRNA expression of muscle-specific growth quality control was applied whereby any force values falling more genes (Figure 1F) with VMHC 10.3-fold higher, MLP 6.1- than two standard deviations away from the mean value were fold higher, and SMLC2 7.4-fold higher in cold- compared discarded in order to account for failed indents. Data loss at this with warm-acclimated spongy myocardium (P < 0.05, stage was less than 10% (data not shown). Figure 1F). In addition, PCNA, a marker for hyperplasia, was 2.6-fold higher, in the cold- compared with the warm- Statistical Analysis acclimated spongy myocardium (Figure 1F). VEGF, a marker for Chamber filling volume was calculated from filling time by the angiogenesis, was also higher (7.3-fold) in cold compared with equation: warm myocardium. Spongy layer and myocyte bundle hypertrophy requires 0.05 increased protein synthesis, which can occur as a result of volume(ml) = time(µs) × × 1000 60 hypertrophic signaling cascades. In mammals, RCAN1 is a target gene of the NFAT signaling pathway and promotes pathological The effect of temperature acclimation on the pressure-volume hypertrophic growth via the fetal gene program, which has also relationship was assessed by a general linear model (GLM) with recently been shown in fish (Wilkins et al., 2004; Bernardo et al., pressure as the dependent variable, volume and acclimation 2010; Johansen et al., 2011; Shih et al., 2015). We found RCAN1 group as fixed factors and body mass as the covariate, with a mRNA expression was 9.5-fold higher in cold- compared with Tukey post-hoc test for differences between groups using R (R warm-acclimated spongy myocardium (P < 0.05; Figure 1G). Core Team, 2013). The calculations were performed on all data ANP and BNP are released by cardiomyocytes in response to below 2 kPa, which approximates the maximum physiological myocardial stretch induced by pressure and volume overload pressures experienced by this species (Forster and Farrell, (Kinnunen et al., 1993). mRNA expression of ANP was 3.2-fold 1994). Differences in compact myocardial thickness, myocyte higher and BNP was 4.2-fold higher in cold- compared with bundle cross sectional area, extra-bundular sinus, collagen warm-acclimated animals (P < 0.05; Figure 1G). deposition and transcript abundance were assessed by GLM with Holm-Sidak post-hoc test for differences between groups Thermal Remodeling of Connective Tissue using SigmaPlot 11.0 (SYSTAT Statistics, San Jose, CA, USA). Myocyte remodeling is associated with remodeling of the Post-hoc analyses of AFM force curves were performed using extracellular matrix (ECM) in both fish and mammals (Chapman Nanoscope Analysis v1.40 (Bruker, Coventry, UK), whereby a et al., 1990; Klaiman et al., 2011). Similar to our earlier work baseline correction was applied to each curve before a force fit (Klaiman et al., 2011) we found amorphous collagen deposition was applied using a Herzian (spherical) model and a maximum was greater in both compact and spongy layers with cold force fit of 70%. For all analyses significance was considered to be acclimation (P < 0.05; Figures 2A,B). We then extended this

Frontiers in Physiology | www.frontiersin.org 6 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

finding using picro-sirus red to determine fibrillar collagen. 2010), we used AFM indentation of ventricular cryo-sections to Figure 2C shows a representative bright-field ventricular section assess whether the temperature-dependent collagen remodeling stained with picro-sirus red and Figure 2D shows the same was associated with a change in local tissue stiffness (Figure 4A). section visualized under plane polarized light. The polarized Cold-acclimated ventricular tissue was stiffer than warm- light images were used to quantify collagen as a percentage of acclimated tissue (P < 0.005; Figure 4B). Mean reduced either compact or spongy tissue area. In agreement with results modulus (Er; and hence localized tissue stiffness) was strongly from Masson’s trichrome staining (Figures 2A,B), compact and significantly correlated with temperature in both spongy myocardial collagen content was 1.7-fold higher in cold- (R2 = 0.99, P < 0.0001) and compact myocardium (R2 = acclimated than control and warm-acclimated animals (P < 0.05; 1.00, P < 0.0001; Figure 4B), with the effects of temperature Figure 2E). Despite a trend, we found no statistical differences being more pronounced in spongy tissue compared with compact between acclimation groups for fibrillar collagen in the spongy tissue. Furthermore, the modulus frequency distribution in both layer (Figure 2F). We did not detect elastin in the fish ventricular tissue types (Figures 4C,D) suggests that mechanical remodeling myocardium except in coronary vessels (not shown). following temperature acclimation is due to homogenous The changes in collagen content were supported by differential structural and/or compositional remodeling of the whole tissue, expression of the collagen I gene, Col1a3. In mammals, collagen rather than isolated or specific regions of the tissue. I accounts for ∼80% of total collagen in the myocardium and is the main collagen in cardiac fibrosis (Medugorac, 1982). DISCUSSION Mammalian collagen I is composed of type 1 (α1) and type 2 (α2) alpha-helical chains; fish also have an additional type Seasonal changes in temperature trigger remodeling of the fish 3 (α3) chain (Saito et al., 2001). We found expression of this heart (Farrell et al., 1988; Graham and Farrell, 1989; Pelouch Col1a3 gene was 1.4-fold higher in the cold- compared with the and Vornanen, 1996; Aho and Vornanen, 1999, 2001; Gamperl warm-acclimated ventricles (P < 0.05; Figure 2G). Expression and Farrell, 2004; Vornanen et al., 2005b; Hassinen et al., 2008; of the Col1a1 and Col1a2 mRNA was lower in both cold and Korajoki and Vornanen, 2009, 2012; Klaiman et al., 2011, 2014; warm fish compared with controls, suggesting temperature- Johnson et al., 2014). Here, we focused on the passive properties independent remodeling is also occurring. Total collagen content of the rainbow trout ventricle across multiple levels of biological is a balance between deposition and degradation by matrix organization following chronic warming and chronic cooling. metalloproteinases (MMPs) (Nagase et al., 2006). MMP2 and Our principle and novel findings are: (1) Cold acclimation MMP13 were 3.4- and 1.9-fold higher, respectively, in the warm- increases fibrillar collagen deposition, the expression collagen compared to cold-acclimated ventricle (P < 0.05; Figure 2H). promoting genes and markers for mammalian pathological MMP9 did not differ between groups. MMP activity is regulated remodeling including those associated with the fetal gene by tissue inhibitors of MMPs (TIMPs). TIMP activity inhibits program (RCAN1, ANP, BNP). Each of these observations collagen degradation by MMPs and is, thus, associated with changed in the opposite direction following chronic warming. increased collagen deposition. Expression of the TIMP2 gene was (2) Cold-induced fibrosis and hypertrophy were associated 5.1-fold higher in the cold- compared with the warm-acclimated with increased passive stiffness of the whole ventricle and ventricle (P < 0.05; Figure 2H). increased micromechanical stiffness in tissue sections. Again, the opposite response occured with chronic warming. These Thermal Remodeling of Ex vivo Chamber findings provide support for our hypothesis that chronic cooling Compliance in the fish heart invokes a hypertrophic phenotype with increased Changes in myocardial thickness and fibrosis are known to cardiac stiffness, fibrosis and hypertrophic markers analogous to influence chamber compliance in mammals (Bing et al., 1971). pathological hypertrophy in the mammalian heart. Importantly, To assess the functional effects of cardiac remodeling on the we report a suppression of growth genes and hypertrophic passive properties of the thermally acclimated fish ventricle we markers following chronic warming. This is novel and provides generated ex vivo passive filling curves from freshly isolated intact a potential new investigative route for understanding processes ventricles treated with BDM at a common test temperature of that regulate regression of pathological hypertrophy. We thus 10◦C. Figure 3 shows mean data for each temperature within propose the trout heart as a potential vertebrate model for the range of physiologically relevant filling pressures experienced investigating regression of fibrotic cardiac hypertrophy. by rainbow trout in vivo (Forster and Farrell, 1994). Thermal acclimation altered the pressure-volume relationship during Cold-Induced Hypertrophy in the Spongy 2 filling [R = 0.60, F(2,15,575) = 1447.0, P < 0.001] revealing Myocardium greater stiffness in the cold and greater compliance in the warm The trabecular nature of the spongy myocardium in fish is compared to controls (t ratio = 54.2; Figure 3). important as it increases the surface area and reduces the distance for diffusion of gases and nutrients between the myocardium Micromechanical Ventricular Stiffness and the venous blood that transits the ventricular lumen. Collagen is an important mediator of tissue tensile strength Fish have a single circulation; the heart sends deoxygenated and stiffness, and is arranged into networks that support blood to the gills to be oxygenated, which travels around the cardiomyocytes. As alterations in collagen content are known body before returning to the heart. Thus, the spongy layer to affect cardiac micromechanical properties (Fomovsky et al., of the fish heart receives oxygen-poor venous blood. The

Frontiers in Physiology | www.frontiersin.org 7 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

FIGURE 2 | Ventricular connective tissue remodeling. Semi-quantification of amorphous collagen from Masson’s trichrome histology for compact tissue (A) and spongy tissue (B). Representative (C) bright-field and (D) polarized micrographs of sections of ventricular tissue stained with picro-sirus red, showing the spongy (s) and compact (c) myocardium. Semi-quantitative analysis of picro-sirus red stained sections (E) compact and (F) the spongy myocardium. In (A,B,E,F), collagen content is expressed as a percentage of total tissue for either the compact or spongy layer. The corresponding mRNA expression of (G) collagen genes and (H) collagen regulatory genes (TIMP2, up-regulation; MMP2, MMP9, and MMP13, down-regulation) in the spongy myocardium of cold (5◦C; blue), control (10◦C; green), and warm (18◦C; red) acclimated rainbow trout (n = 7 fish for each acclimation group; 3 replicates for each animal were averaged for both histology and qPCR). Values presented are mean ± S. E. Significance was assessed by GLM with a Tukey, or Holm-Sidak for multiple comparisons, post-hoc test. Significance between groups is shown by dissimilar letters (P < 0.05).

Frontiers in Physiology | www.frontiersin.org 8 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

normal growth of rainbow trout (Farrell et al., 1988) and in exercise trained Atlantic salmon (Castro et al., 2013). The fish myocardium is also known for its regenerative ability, which requires cell proliferation and hyperplasia (Sun et al., 2009). We cannot, therefore, exclude hyperplasia of myocytes or other cell types, such as fibroblasts, in response to chronic cooling. However, the changes in gene expression of hypertrophic-related genes (VMHC, MLP and SMLC2) were between 6.1-fold and 10.3-fold higher in the cold compared with the warm, whereas those of the hyperplasia marker (PCNA) changed by 2.6-fold. Hypertrophic growth follows activation of downstream hypertrophic signaling cascades. Mitogen-activated protein kinases (MAPKs) and the calcineurin-NFAT pathway are central to pathological hypertrophic growth in mammals (Wilkins et al., 2004; Bernardo et al., 2010). RCAN1 can increase calcineurin-NFAT signaling and enhance hypertrophic growth of myocytes (Liu et al., 2009). We found an up-regulation of RCAN1 in the cold- and a down regulation in warm-acclimated FIGURE 3 | Mean ventricular passive pressure-volume relationships within the physiological relevant pressure range of <2 kPa for cold trout ventricle suggesting components of the pathological (5◦C; blue), control (10◦C; green), and warm (18◦C; red) acclimated hypertrophic signaling cascade in mammals are also present rainbow trout (n = 8). Values are mean ± S. E., curves show all mean data in fish. The calcineurin-NFAT signaling system is largely for n > 3. Pressure has been standardized to start at 0 kPa for graphical unexplored in the adult fish heart, although RCAN1 has representation. Significance of changes in compliance with temperature acclimation was assessed by GLM with volume as the dependent variable, been linked to the hypertrophic response to stress in rainbow treatment and pressure as the fixed factors and chamber mass as the trout (Johansen et al., 2011). Calcineurin-NFAT signaling has covariate (P < 0.05) and is shown by dissimilar letters. also been shown to be important in heart development of zebrafish, indicating a physiological role under non-hypertrophic conditions (Armstrong et al., 2003). However, it is important thickness of the spongy myocardium is, therefore, a compromise to note that, although fetal and adult cardiac growth pathways between minimizing the distance for oxygen diffusion and in zebrafish show a high level of conservation with those maximizing the cross-sectional area for tension development of mammals (Shih et al., 2015), the adult phenotype of the (Davie and Farrell, 1991). We have shown spongy layer cardiac fish heart likely more closely resembles a mammalian neonate hypertrophy following cold acclimation, and atrophy following than a mammalian adult (Tibbits et al., 2002) as some genes warm acclimation, indirectly as reciprocal changes in myocyte characteristic of the mammalian fetal gene program, such as ANP, bundle cross-sectional area in the trabeculated region of the heart are intrinsically expressed in the adult fish heart (Sun et al., 2009; and by changes in the expression of hypertrophic genes. Similar Jensen et al., 2012). molecular responses have been observed in fish with stress- We did not measure gene expression or myocyte size in the induced cardiac hypertrophy (Johansen et al., 2011) and chronic compact myocardial layer, which had inverse thermal remodeling cooling (Vornanen et al., 2005a), but the reciprocal responses to the spongy layer. This limits interpretation of transmural with warming have not, to our knowledge, been previously remodeling across layers in our study. Unlike the spongy layer, reported. It should be noted that we did not measure the cross- which receives venous blood, the compact layer of the fish section area or volume of the single myocytes in our study, which ventricle has a coronary blood supply from the gill (Farrell and is the base unit of hypertrophy. However, previous work in this Jones, 1992). The cold-induced atrophy of the compact layer species with similar temperatures (e.g., Vornanen, 1998; Clark may reflect reduced reliance on coronary circulation due to and Rodnick, 1999; Vornanen et al., 2005a) do report single cell an increased oxygen carrying capacity of water and the blood, hypertrophy, associated with changes in temperature and cardiac and a decreased oxygen demand from cardiac muscle at cold load, and importantly correlate this with the changes in gene temperatures, as suggested by Farrell and Clutterham (2003). expression (VHMC, SMLC2, and MLP; Vornanen et al., 2005a) This idea is supported by recent work showing increased VEGF we show here. Our interpretation of cold-induced hypertrophy expression in the compact layer of warm-acclimated Atlantic is in line with our previous work (Klaiman et al., 2011) and salmon (Jørgensen et al., 2014). The reduction in compact also a recent mathematical model of the spongy layer of the thickness after chronic cold exposure could also increase cardiac carp heart (Cyprinus carpio) which shows increasing trabecular compliance following cold-induced stiffening (Johnson et al., tissue volume increases stroke volume, stroke work and ejection 2014; Klaiman et al., 2014). fraction (Kochová et al., 2015). Our novel finding of increased mRNA expression of PCNA Connective Tissue Remodeling suggests hyperplasia may also contribute to the growth of the We found amorphous collagen increased in both myocardial myocyte bundles in the spongy layer during cold acclimation. layers following chronic cold which confirms our earlier findings Cardiac hyperplasia has been previously reported during (Klaiman et al., 2011). We extended this work by showing that

Frontiers in Physiology | www.frontiersin.org 9 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

FIGURE 4 | Micromechanical testing of compact and spongy myocardium by AFM indentation. (A) White light microscope image of control ventricle showing six 50 × 50 µM scan areas used for AFM indentation, closed for compact (c), and dashed for spongy (s). (B) Mean reduced modulus (Er) of spongy (open circles) and compact (closed circles) myocardium by acclimation temperature. (C) Accumulative frequency curves of each individual Er from the spongy myocardium and (D) the corresponding frequency curves for the compact myocardium with cold (5◦C; blue), control (10◦C; green) and warm (18◦C; red). Values presented are mean ± S. D. Signiﬁcance was assessed by GLM and is shown between groups by dissimilar letters and by * between tissue types (P < 0.005).

fibrillar collagen is also increased following cooling. Collagen is 1999; Gamperl and Farrell, 2004; Klaiman et al., 2011, 2014). In the primary support protein in the ECM and is associated with our study, it is likely that the winter photoperiod, in addition to pathological remodeling in mammals. Accordingly, we report 8 week chronic cold, played an important role in the onset of an increase in the fish-specific Col1a3 gene expression in the spongy layer and myocyte bundle hypertrophy. spongy layer from the cold-acclimated ventricle. Cold-induced fibrosis is thought to maintain mechanical cardiac performance Ventricular Compliance under elevated hemodynamic stress of pumping cold, highly Temperature acclimation significantly altered passive pressure- viscous blood (Cerra et al., 2004). The pro-collagen regulatory volume relationships in the ventricle. Cold-acclimated tissue was enzyme TIMP2 increased with cold acclimation and the anti- stiffer and warm-acclimated tissue was more compliant than collagen regulatory MMP genes decreased with cold acclimation control tissue when all were tested at a common temperature. supporting this remodeling. In fish, MMP13 catalyses the Although diastolic filling curves have previously been generated hydrolysis of collagen, degrading it to gelatin (Hillegass et al., for fish hearts (Forster and Farrell, 1994; Mendonça et al., 2007), 2007), and MMP2 and MMP9 digest the gelatin into removable this is the first study where they have been used to probe waste products (Kubota et al., 2003). Although the picro-sirus remodeling. Klaiman et al. (2014) used a Langendorff preparation red histological analysis did not show a statistically resolvable to generate working pressure-volume loops in the trout ventricle increase in spongy layer collagen, we show up-regulation of and found that temperature acclimation affected the systolic, collagen regulatory genes which may pre-empt the fibrotic but not diastolic phase of the cardiac cycle. However, this study response. Interestingly, the reverse has been shown with cooling did not find cold-induced fibrosis which could explain stasis zebrafish (Johnson et al., 2014) and Atlantic salmon (Jørgensen of diastolic stiffness. Our histological and gene expression data et al., 2014) where a reduction in ventricular collagen and suggest that although collagen deposition/regression is a strong collagen-related genes expression has been documented. These driver in the changes in functional compliance following thermal differences may be species-specific; however, varying myocardial acclimation, other factors must also be at play. It is likely that responses to thermal acclimation in fish are commonly reported other temperature dependent components of the myocardium (Sephton and Driedzic, 1995; Johnson et al., 2014; Klaiman et al., are affected, such as the actin cytoskeleton (Qiu et al., 2010) which 2014). Sex, maturation and circannual rhythms are thought to could alter intrinsic stiffness of the cardiomyocytes. We have influence the degree of overall hypertrophy (Clark and Rodnick, previously shown that titin isoform expression in trout ventricle

Frontiers in Physiology | www.frontiersin.org 10 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle underlies the compliance of the fish myocyte (Patrick et al., determine whether chronic cold induces remodeling beyond that 2009). How titin changes with thermal remodeling is currently necessary for accommodating the changes in blood viscosity. To unknown, but could be crucial in understanding tissue and organ understand whether temperature pushes the fish heart past the level compliance. compensation/adaptation phase into the true pathology, future The strong negative correlation between temperature and studies could combine approaches known to induce pathology micromechanical stiffness in both the spongy and compact in mammalian hearts (e.g., outflow constriction) in fish hearts. myocardium suggests that the changes in tissue ultrastructure Our finding of opposite cardiac remodeling in chronically and matrix organization are exerting a functional effect on the warmed fish, compared with control fish and compared with mechanical competency of the myocardium. Previous in situ chronically cooled fish is intriguing as it suggests fibrosis may studies show increased stroke volume and in vitro studies have oscillate seasonally, revealing a more dynamic nature than shown increased Ca2+ sensitivity of force generation following fibrosis associated with dysfunction in mammals. This finding cold acclimation in salmonid hearts (Graham and Farrell, 1989; builds upon previous work to suggest an opposite remodeling Franklin and Davie, 1992). Our work on the passive properties phenotype in fish following chronic warming compared to builds from these studies on the active properties by suggesting chronic cooling (e.g., Klaiman et al., 2011, 2014). However, that the temperature-mediated changes in cardiac force and/or to directly test whether the cardiac remodeling response is cardiac output must compensate for the altered micromechanical “reversible,” cardiac function would have to be assessed in the stiffness. Remodeling of the passive components of the fish heart same animals following acclimation to both temperatures (i.e., with cold results in a ventricle more capable of generating the cooling then warming or warming then cooling). Moreover, forces required to maintain blood pressure at cold temperatures, the collagen I gene that appeared most responsive to thermal whilst being protected from overstretch from pumping viscous remodeling in the current study is fish-specific (Col1a3). blood. Clearly, the cold-induced fibrosis in the fish heart is not Collagen chains containing this domain have been shown to have pathological and the fibrotic phenotype is more plastic in fish greater susceptibility to heat denaturation and degradation by than in mammals. MMP13 than collagen chains without it (Saito et al., 2001), which may explain its malleability with chronic temperature change. If Relevance of Thermal Remodeling to this collagen domain is driving the fish fibrotic response to cold, Mammalian Hypertrophy it may explain why a typically pathological response in mammals Here, we have compared our findings of cardiac remodeling occurs transiently in fish. in the fish ventricle to those in physiological and pathological remodeling of the mammalian ventricle. However, it should be AUTHOR CONTRIBUTIONS noted that the architecture and development of the fish heart differs from the mammalian heart. The mature mammalian heart AK, AF, PG, and HS are responsible for the concept and design is almost entirely compact with integrated trabeculae (Weiford of the research. AK, AF, and JM performed experiments and et al., 2004; Risebro et al., 2006), whereas, the trout ventricle analysed data in HS and MS laboratories. AK, AF, JM, MS, and is formed of two distinct layers; an outer wall of compact HS interpreted the results of the experiments. AK, JM, and HS myocardium and an inner layer of trabecular myocardium drafted the manuscript. AK, AF, JM, MS, PG, and HS revised and (Pieperhoff et al., 2009). Throughout this study we have focused edited the manuscript. AK, AF, JM, MS, PG, and HS approve the on the inner spongy layer for our comparison due to the final version of the manuscript submitted for publication. cold-induced spongy layer hypertrophy. Future studies should investigate the effect of temperature acclimation on the thermal FUNDING remodeling of the outer compact layer of the fish heart which may be more relevant to mammalian heart remodeling. However, AK and AF were supported by studentships from the BBSRC. The although many of the genes and signaling pathways that regulate Shiels lab is supported by the Leverhulme Trust (240613). MJS is cardiac development and morphogenesis are shared between all funded by the Medical Research Council UK (grant G1001398). vertebrates (Harvey, 2002; Luxán et al., 2013; Samsa et al., 2013) the compact layer of a fish heart is not directly analogous to that ACKNOWLEDGMENTS of adult mammals (Gupta and Poss, 2012). We thank Dr. Margaux Horn and Dr. Sanjoy Chowdhury for Perspectives helpful discussions throughout this project. We thank Hamid Hypertrophy and stiffening of the ventricle, with associated Rizvi and Devmalya Sarkar for instructive preliminary work fibrosis, are characteristics of mammalian pathological on this topic; and Nathan Thavarajah, Jenna Legg, Dr. Alex remodeling. In fish, it is likely that reactive fibrosis occurs Henderson and Dr. Robert Nudds for help with data analysis to provide support for increasing muscle mass and to prevent and Peter Walker for help with histology and data analysis. over stretching of myocytes under increased hemodynamic Histology was performed in the University of Manchester stress in the cold. This increased stiffness during cold Histology Facility. Atomic Force Microscopy was carried out in acclimation is likely protective, despite fibrosis. We did not the University of Manchester BioAFM Facility.

Frontiers in Physiology | www.frontiersin.org 11 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

REFERENCES Gamperl, A. K., and Farrell, A. P. (2004). Cardiac plasticity in fishes: environmental influences and intraspecific differences. J. Exp. Biol. 207, 2539–2550. doi: Aho, E., and Vornanen, M. (1999). Contractile properties of atrial and ventricular 10.1242/jeb.01057 myocardium of the heart of rainbow trout Oncorhynchus mykiss: effects of Gillis, T. E., Marshall, C. R., Xue, X. H., Borgford, T. J., and Tibbits, G. F. thermal acclimation. J. Exp. Biol. 202, 2663–2677. (2000). Ca2+ binding to cardiac troponin C: effects of temperature and pH Aho, E., and Vornanen, M. (2001). Cold acclimation increases basal heart rate on mammalian and salmonid isoforms. Am. J. Physiol. Regul. Integr. Comp. but decreases its thermal tolerance in rainbow trout (Oncorhynchus mykiss). Physiol. 279, R1707–R1715. J. Comp. Physiol. B Biochem. Syst. Environ. Physiol. 171, 173–179. doi: 10.1007/ Graham, H. K., Akhtar, R., Kridiotis, C., Derby, B., Kundu, T., Trafford, A. W., s003600000171 et al. (2011). Localised micro-mechanical stiffening in the ageing aorta. Mech. Andersen, J. B., Rourke, B. C., Caiozzo, V. J., Bennett, A. F., and Hicks, J. W. Ageing Dev. 132, 459–467. doi: 10.1016/j.mad.2011.07.003 (2005). Postprandial cardiac hypertrophy in pythons. Nature 434, 37–38. doi: Graham, M. S., and Farrell, A. P. (1989). The effect of temperature-acclimation 10.1038/434037a and adrenaline on the performance of a perfused trout heart. Physiol. Zool. 62, Armstrong, J. D., Kemp, P. S., Kennedy, G. J. A., Ladle, M., and Milner, N. J. (2003). 38–61. Habitat requirements of Atlantic salmon and brown trout in rivers and streams. Gupta, V., and Poss, K. D. (2012). Clonally dominant cardiomyocytes direct heart Fish. Res. 62, 143–170. doi: 10.1016/S0165-7836(02)00160-1 morphogenesis. Nature 484, 470–484. doi: 10.1038/nature11045 Bernardo, B. C., Weeks, K. L., Pretorius, L., and McMullen, J. R. (2010). Molecular Harvey, R. P. (2002). Patterning of the vertebrate heart. Nat. Rev. Genet. 3, 544–556. distinction between physiological and pathological cardiac hypertrophy: doi: 10.1038/nrg843 experimental findings and therapeutic strategies. Pharmacol. Ther. 128, Hassinen, M., Haverinen, J., and Vornanen, M. (2008). Electrophysiological 191–227. doi: 10.1016/j.pharmthera.2010.04.005 properties and expression of the delayed rectifier potassium (ERG) channels Bing, O. H. L., Matsushi, S., Fanburg, B. L., and Levine, H. J. (1971). Mechanical in the heart of thermally acclimated rainbow trout. Am. J. Physiol. Regul. Integr. properties of rat cardiac muscle during experimental hypertrophy. Circ. Res. 28, Comp. Physiol. 295, R297–R308. doi: 10.1152/ajpregu.00612.2007 234–245. doi: 10.1161/01.RES.28.2.234 Hillegass, J. M., Villano, C. M., Cooper, K. R., and White, L. A. (2007). Matrix Castro, V., Grisdale-Helland, B., Helland, S. J., Torgersen, J., Kristensen, T., metalloproteinase-13 is required for zebra fish (Danio rerio) development Claireaux, G., et al. (2013). Cardiac molecular acclimation mechanisms in and is a target for glucocorticoids. Toxicol. Sci. 100, 168–179. doi: response to swimming induced exercise in Atlantic salmon. PLoS ONE 10.1093/toxsci/kfm192 8:e55056. doi: 10.1371/journal.pone.0055056 Jensen, B., Boukens, B. J. D., Postma, A. V., Gunst, Q. D., van den Hoff, Cerra, M. C., Imbrogno, S., Amelio, D., Garofalo, F., Colvee, E., Tota, B., et al. M. J. B., Moorman, A. F. M., et al. (2012). Identifying the evolutionary (2004). Cardiac morphodynamic remodelling in the growing eel (Anguilla building blocks of the cardiac conduction system. PLoS ONE 7:e44231. doi: anguilla L.). J. Exp. Biol. 207, 2867–2875. doi: 10.1242/jeb.01119 10.1371/journal.pone.0044231 Chapman, D., Weber, K. T., and Eghbali, M. (1990). Regulation of fibrillar Johansen, I. B., Lunde, I. G., Røsjø, H., Christensen, G., Nilsson, G. E., collagen type-I and type-III and basement-membrane type-IV collagen gene- Bakken, M., et al. (2011). Cortisol response to stress is associated with expression in pressure overloaded rat myocardium. Circ. Res. 67, 787–794. doi: myocardial remodeling in salmonid fishes. J. Exp. Biol. 214, 1313–1321. doi: 10.1161/01.RES.67.4.787 10.1242/jeb.053058 Clark, R. J., and Rodnick, K. J. (1999). Pressure and volume overloads are Johnson, A. C., Turko, A. J., Klaiman, J. M., Johnston, E. F., and Gillis, T. E. (2014). associated with ventricular hypertrophy in male rainbow trout. Am. J. Physiol. Cold acclimation alters the connective tissue content of the zebrafish (Danio Regul. Integr. Comp. Physiol. 277, R938–R946. rerio) heart. J. Exp. Biol. 217, 1868–1875. doi: 10.1242/jeb.101196 Collier, P., Watson, C. J., Van Es, M. H., Phelan, D., McGorrian, C., Tolan, Jørgensen, S. M., Castro, V., Krasnov, A., Torgersen, J., Timmerhaus, G., Hevrøy, E. M., et al. (2012). Getting to the heart of cardiac remodeling; how collagen M., et al. (2014). Cardiac responses to elevated seawater temperature in Atlantic subtypes may contribute to phenotype. J. Mol. Cell. Cardiol. 52, 148–153. doi: salmon. BMC Physiol. 14:2. doi: 10.1186/1472-6793-14-2 10.1016/j.yjmcc.2011.10.002 Junqueira, L. C. U., Bignolas, G., and Brentani, R. R. (1979). Picrosirius Crick, S. L., and Yin, F. C. P. (2007). Assessing micromechanical properties staining plus polarization microscopy, a specific method for collagen of cells with atomic force microscopy: importance of the contact point. detection in tissue-sections. Histochem. J. 11, 447–455. doi: 10.1007/ Biomech. Model. Mechanobiol. 6, 199–210. doi: 10.1007/s10237-00 BF01002772 6-0046-x Kemp, A. D., Harding, C. C., Cabral, W. A., Marini, J. C., and Wallace, J. M. (2012). Davie, P. S., and Farrell, A. P. (1991). The coronary and luminal circulations of the Effects of tissue hydration on nanoscale structural morphology and mechanics myocardium of fishes. Can. J. Zool. 69, 1993–2001. doi: 10.1139/z91-278 of individual type I collagen fibrils in the Brtl mouse model of Osteogenesis de Bold, A. J., and de Bold, M. L. K. (2005). Determinants of natriuretic peptide Imperfecta. J. Struct. Biol. 180, 428–438. doi: 10.1016/j.jsb.2012.09.012 production by the heart: basic and clinical implications. J. Invest. Med. 53, Kinnunen, P., Vuolteenaho, O., and Ruskoaho, H. (1993). Mechanisms of atrial 371–377. doi: 10.2310/6650.2005.53710 and brain natriuretic peptide release from rat ventricular myocardium - effect Dorn, G. W. II. (2007). The fuzzy logic of physiological cardiac hypertrophy. of stretching. Endocrinology 132, 1961–1970. Hypertension 49, 962–970. doi: 10.1161/HYPERTENSIONAHA.106.079426 Klaiman, J. M., Fenna, A. J., Shiels, H. A., Macri, J., and Gillis, T. E. (2011). Cardiac Farrell, A. P., and Clutterham, S. M. (2003). On-line venous oxygen tensions in remodeling in fish: strategies to maintain heart function during temperature rainbow trout during graded exercise at two acclimation temperatures. J. Exp. change. PLoS ONE 6:e24464. doi: 10.1371/journal.pone.0024464 Biol. 206, 487–496. doi: 10.1242/jeb.00100 Klaiman, J. M., Pyle, W. G., and Gillis, T. E. (2014). Cold acclimation increases Farrell, A. P., Hammons, A. M., Graham, M. S., and Tibbits, G. F. (1988). Cardiac cardiac myofilament function and ventricular pressure generation in trout. growth in rainbow-trout, salmo-gairdneri. Can. J. Zool. 66, 2368–2373. doi: J. Exp. Biol. 217, 4132–4140. doi: 10.1242/jeb.109041 10.1139/z88-351 Kochová, P., Cimrman, R., Štengl, M., Oštádal, B., and Tonar, Z. (2015). A Farrell, A. P., and Jones, D. R. (1992). “The heart,” in Fish Physiology: Vol. 12, Part mathematical model of the carp heart ventricle during the cardiac cycle. A The Cardiovascular System, eds W. S. Hoar, D. J. Randall, and A. P. Farrell J. Theor. Biol. 373, 12–25. doi: 10.1016/j.jtbi.2015.03.014 (San Diego, CA: Academic Press Inc.), 1–87. Korajoki, H., and Vornanen, M. (2009). Expression of calsequestrin in atrial and Fomovsky, G. M., Thomopoulos, S., and Holmes, J. W. (2010). Contribution of ventricular muscle of thermally acclimated rainbow trout. J. Exp. Biol. 212, extracellular matrix to the mechanical properties of the heart. J. Mol. Cell. 3403–3414. doi: 10.1242/jeb.031617 Cardiol. 48, 490–496. doi: 10.1016/j.yjmcc.2009.08.003 Korajoki, H., and Vornanen, M. (2012). Expression of SERCA and phospholamban Forster, M. E., and Farrell, A. P. (1994). The volumes of the chambers of the trout in rainbow trout (Oncorhynchus mykiss) heart: comparison of atrial and heart. Comp. Biochem. Physiol. A Physiol. 109, 127–132. doi: 10.1016/0300- ventricular tissue and effects of thermal acclimation. J. Exp. Biol. 215, 9629(94)90318-2 1162–1169. doi: 10.1242/jeb.065102 Franklin, C. E., and Davie, P. S. (1992). Sexual maturity can double heart mass and Kubota, M., Kinoshita, M., Takeuchi, K., Kubota, S., Toyohara, H., and Sakaguchi, cardiac power output in male rainbow trout. J. Exp. Biol. 171, 139–148. M. (2003). Solubilization of type I collagen from fish muscle connective tissue

Frontiers in Physiology | www.frontiersin.org 12 January 2016 | Volume 6 | Article 427 Keen et al. Hypertrophic and Fibrotic Remodeling in the Fish Ventricle

by matrix metalloproteinase-9 at chilled temperature. Fish. Sci. 69, 1053–1059. Schneider, C. A., Rasband, W. S., and Eliceiri, K. W. (2012). NIH image to doi: 10.1046/j.1444-2906.2003.00726.x imagej: 25 years of image analysis. Nat. Methods 9, 671–675. doi: 10.1038/nmet Liu, Q. H., Busby, J. C., and Molkentin, J. D. (2009). Interaction between TAK1- h.2089 TAB1-TAB2 and RCAN1-calcineurin defines a signalling nodal control point. Sephton, D. H., and Driedzic, W. R. (1995). Low temperature acclimation Nat. Cell Biol. 11, 154–161. doi: 10.1038/ncb1823 decreases rates of protein synthesis in rainbow trout (Oncorhynchus mykiss) Luxán, G., Casanova, J. C., Martínez-Poveda, B., Prados, B., D’Amato, G., heart. Fish Physiol. Biochem. 14, 63–69. doi: 10.1007/BF00004291 MacGrogan, D., et al. (2013). Mutations in NOTCH pathway regulator MIB1 Shiels, H. A., Di Maio, A., Thompson, S., and Block, B. A. (2011). Warm fish cause left ventricular noncompaction cardiomyopathy. Nat. Med. 19, 193–201. with cold hearts: thermal plasticity of excitation-contraction coupling in bluefin doi: 10.1038/nm.3046 tuna. Proc. R. Soc. B Biol. Sci. 278, 18–27. doi: 10.1098/rspb.2010.1274 Medugorac, I. (1982). Characterization of intramuscular collagen in Shih, Y. H., Zhang, Y. J., Ding, Y. H., Ross, C. A., Li, H., Olson, T. M., et al. the mammalian left ventricle. Basic Res. Cardiol. 77, 589–598. doi: (2015). Cardiac transcriptome and dilated cardiomyopathy genes in zebrafish. 10.1007/BF01908312 Cir. Cardiovasc. Genet. 8, 261–269. doi: 10.1161/CIRCGENETICS.114.0 Mendonça, P. C., Genge, A. G., Deitch, E. J., and Gamperl, A. K. (2007). 00702 Mechanisms responsible for the enhanced pumping capacity of the in situ Sun, X., Hoage, T., Bai, P., Ding, Y., Chen, Z., Zhang, R., et al. (2009). Cardiac winter flounder heart (Pseudopleuronectes americanus). Am. J. Physiol. Regul. hypertrophy involves both myocyte hypertrophy and hyperplasia in anemic Integr. Comp. Physiol. 293, R2112–R2119. doi: 10.1152/ajpregu.00202.2007 zebrafish. PLoS ONE 4:e6596. doi: 10.1371/journal.pone.0006596 Miller, P. J. (1971). Elastin stain. Med. Lab. Technol. 28, 148–149. Tervonen, V., Kokkonen, K., Vierimaa, H., Ruskoaho, H., and Vuolteenaho, O. Mone, S. M., Sanders, S. P., and Colan, S. D. (1996). Control mechanisms (2001). Temperature has a major influence on cardiac natriuretic peptide in for physiological hypertrophy of pregnancy. Circulation 94, 667–672. doi: salmon. J. Physiol. Lond. 536, 199–209. doi: 10.1111/j.1469-7793.2001.00199.x 10.1161/01.CIR.94.4.667 Tibbits, G. F., Xu, L., and Sedarat, F. (2002). Ontogeny of excitation–contraction Nagase, H., Visse, R., and Murphy, G. (2006). Structure and function of coupling in the mammalian heart. Comp. Biochem. Physiol. A Mol. Integr. matrix metalloproteinases and TIMPs. Cardiovasc. Res. 69, 562–573. doi: Physiol. 132, 691–698. doi: 10.1016/S1095-6433(02)00128-9 10.1016/j.cardiores.2005.12.002 Vornanen, M. (1998). L-type Ca2+ current in fish cardiac myocytes: effects of Opie, L. H., Commerford, P. J., Gersh, B. J., and Pfeffer, M. A. (2006). Controversies thermal acclimation and beta-adrenergic stimulation. J. Exp. Biol. 201, 533–547. in cardiology 4 - Controversies in ventricular remodelling. Lancet 367, 356–367. Vornanen, M., Hassinen, M., Koskinen, H., and Krasnov, A. (2005a). Steady-state doi: 10.1016/S0140-6736(06)68074-4 effects of temperature acclimation on the transcriptome of the rainbow trout Patrick, S. M., Shiels, H., and Cazorla, O. (2009). Comparison of length-dependent heart. Am. J. Physiol. Regul. Integr. Comp. Physiol. 289, R1177–R1184. doi: calcium activation of cardiac myofilaments between the rat and the rainbow 10.1152/ajpregu.00157.2005 trout. J. Physiol. Sci. 59, 536–536. Vornanen, M., Haverinen, J., Hassinen, M., Koskinen, H., and Krasnov, A. (2005b). Pelouch, V., and Vornanen, M. (1996). Effects of thermal acclimation on ventricle Plasticity of cardiac function in thermally acclimated rainbow trout. Comp. size, protein composition, and contractile properties of crucian carp heart. Biochem. Physiol. A Mol. Integr. Physiol. 141, S355. J. Therm. Biol. 21, 1–9. doi: 10.1016/0306-4565(95)00013-5 Vornanen, M., Shiels, H. A., and Farrell, A. P. (2002). Plasticity of excitation- Pieperhoff, S., Bennett, W., and Farrell, A. P. (2009). The intercellular organization contraction coupling in fish cardiac myocytes. Comp. Biochem. Physiol. A Mol. of the two muscular systems in the adult salmonid heart, the compact Integr. Physiol. 132, 827–846. doi: 10.1016/S1095-6433(02)00051-X and the spongy myocardium. J. Anat. 215, 536–547. doi: 10.1111/j.1469- Wallace, J. M., Harding, C. C., and Kemp, A. D. (2012). Effects of hydration on 7580.2009.01129.x nanoscale structural morphology and mechanics of individual type I collagen Poupa, O., Gesser, H., Jonsson, S., and Sullivan, L. (1974). Coronary-supplied fibrils. MRS Proc. 1465, 428–438. doi: 10.1557/opl.2012.926 compact shell of ventricular myocardium in salmonids - growth and enzyme Weiford, B. C., Subbarao, V. D., and Mulhern, K. M. (2004). Noncompaction of pattern. Comp. Biochem. Physiol. 48, 85–95. doi: 10.1016/0300-9629(74) the ventricular myocardium. Circulation 109, 2965–2971. doi: 10.1161/01.CIR. 90856-1 0000132478.60674.D0 Qiu, H., Zhu, Y., Gansner, M., Depre, C., Peppas, A., Resuello, R., et al. (2010). Wilkins, B. J., Dai, Y. S., Bueno, O. F., Parsons, S. A., Xu, J., Plank, D. Increases in vascular smooth muscle stiffness with aging [Abstract]. Faseb J. 24. M., et al. (2004). Calcineurin/NFAT coupling participates in pathological, R Core Team (2013). R: A Language and Environment for Statistical Computing. but not physiological, cardiac hypertrophy. Circ. Res. 94, 110–118. doi: Vienna: R Foundation for Statistical Computing. Available online at: http:// 10.1161/01.RES.0000109415.17511.18 www.R-project.org/ Risebro, C. A., Smart, N., Dupays, L., Breckenridge, R., Mohun, T. J., and Riley, Conflict of Interest Statement: The authors declare that the research was P. R. (2006). Hand1 regulates cariomyocyte proliferation versus differentiation conducted in the absence of any commercial or financial relationships that could in the developing heart. Development 133, 4595–4606. doi: 10.1242/dev. be construed as a potential conflict of interest. 02625 Saito, M., Takenouchi, Y., Kunisaki, N., and Kimura, S. (2001). Complete Copyright © 2016 Keen, Fenna, McConnell, Sherratt, Gardner and Shiels. This primary structure of rainbow trout type I collagen consisting of alpha is an open-access article distributed under the terms of the Creative Commons 1(I)alpha 2(I)alpha 3(I) heterotrimers. Eur. J. Biochem. 268, 2817–2827. doi: Attribution License (CC BY). The use, distribution or reproduction in other forums 10.1046/j.1432-1327.2001.02160.x is permitted, provided the original author(s) or licensor are credited and that the Samsa, L. A., Yang, B., and Lui, J. (2013). Embryonic cardiac chamber maturation: original publication in this journal is cited, in accordance with accepted academic trabeculation, conduction, and cardiomyocyte proliferation. Am. J. Med. Genet. practice. No use, distribution or reproduction is permitted which does not comply 163, 157–168. doi: 10.1002/ajmg.c.31366 with these terms.

Frontiers in Physiology | www.frontiersin.org 13 January 2016 | Volume 6 | Article 427 Thermal remodelling of the ectothermic heart

Thermal remodelling of the ectothermic heart

4. Metabolic and biochemical remodelling of the thermally

acclimated fish ventricle using Fourier transform infra-red

spectroscopy with high magnification optics.

This chapter is presented in manuscript format to be submitted to the

Journal of Biological Chemistry.

Page | 81

Thermal remodelling of the ectothermic heart

Page | 82

Metabolic and biochemical remodelling of the fish heart

Title: Metabolic and biochemical remodelling of the thermally acclimated fish ventricle using Fourier transform infra-red spectroscopy with high magnification optics.

Adam N. Keen1, 2, Holly A. Shiels1, John Marrin1, Alex Henderson2, Peter Gardner2*

1Faculty of Life Sciences, University of Manchester, Manchester, UK 2School of Chemical Engineering and Analytical Science, Manchester Institute of Biotechnology, University of Manchester, UK

*Author of correspondence: [email protected]

Short title: Metabolic and biochemical remodelling of the fish heart.

Text: 12,390 Figures: 15 Tables: 2 References: 82

Keywords: Compliance, Stiffness, Temperature Acclimation, Collagen, Matrix Metalloproteinase, Extracellular Matrix

List of Abbreviations: ATP, adenosine triphosphate; FAO, fatty acid oxidation; FFPE, formalin-fixed, paraffin embedded; FTIR, Fourier transform infrared; H & E, haematoxylin and eosin; PBS, phosphate buffered saline; PC, principal component; PCA, principal component analysis

Metabolic and biochemical remodelling of the fish heart

Abstract

Ambient temperature changes affect numerous physiological processes in ectothermic species. In the rainbow trout temperature directly influences metabolic rate, increasing cardiac demand at high temperatures. Ambient temperature can also cause changes in cardiac load and, as rainbow trout remain active throughout the cold winter months, can initiate a remodelling of the cardiovascular system. In either case, energy demand of the cardiac myocytes and respiring tissue may be altered. Here, we used Fourier transform infrared spectroscopy and histological staining to assess changes in tissue biochemistry with prolonged temperature change in rainbow trout. Following cold acclimation we found an increase in infrared absorption in regions of the spectrum characteristic of lipids and a decrease in tissue glycogen content, which agreed with our histological staining. Following warm acclimation we saw the opposite, a decrease in tissue lipid, but an increase in infrared absorption in regions of the spectrum characteristic of glycogen. Together our results suggest that, during periods of prolonged cold the increased energy demand required to increase cardiac muscle mass may be met by an increase in aerobic fatty acid oxidation without, or with very little, contribution of glycolysis. However, during periods of prolonged warm temperature the opposite occurs which may suggest a metabolic shift to an increase in anaerobic glycolytic energy production and a decrease in pathways for aerobic lipid oxidation. This switch in energy production pathways is likely maladaptive, as energy production by glycolysis is less efficient than by fatty acid oxidation, and can result in energy deficiencies in the myocardium. Changes in cardiac energetics may explain reduced cardiac function and survival of fish species exposed to chronic warming.

Metabolic and biochemical remodelling of the fish heart

Introduction

Ambient temperature is described as a ‘master regulator’ for ectothermic physiology as it directly determines body temperature, influencing the rate of biochemical reactions and cellular processes. Thus, ambient temperature is directly correlated to metabolic rate and, therefore, tissue oxygen consumption which is met by increases in cardiac output (1-8). In addition, for cold-active species, such as rainbow trout, seasonal changes in ambient temperature can cause chronic changes in cardiac load, triggering a cardiac remodelling response (9-11). Increased cardiac load associated with low temperature can result in cardiac hypertrophy to compensate for increased haemodynamic stress and reduced contractility (9- 15).

Cardiac hypertrophy can be energetically expensive. In a normal adult heart, fatty acid oxidation (FAO) by the mitochondria produces sufficient energy for cellular processes, in the form of adenosine triphosphate (ATP) (16). However, under conditions of high energy demand, glycolysis and phosphotransferase reactions catalyzed by creatine kinase and adenylate kinase can provide additional pathways for ATP production (17). In mammals experiencing cardiac hypertrophy, the increased energy demands and hypoxic conditions can cause an increased dependence on glycolysis for energy production and a reduction in FAO (16,18,19). This metabolic state, typical of a foetal heart, involves increased anaplerotic flux to maintain the tricarboxcylic acid (TCA) cycle (18,20). Switching metabolic reliance to anaerobic glycolysis can improve short-term myocardial efficiency, as glycolysis requires less oxygen per mole of ATP produced, but glycolysis cannot compensate for the reduced ATP production by FAO long-term (17). Therefore, if this metabolic shift to glycolysis persists for a prolonged period of time it is associated with ATP deficiency and impaired cardiac function, leading to a number of cardiac myopathies (21,22). During cardiac hypertrophy in fish, it is likely that FAO remains the primary energy production pathway, which is increased to meet the high energy requirement (12,13,23-25). However, there is also evidence to suggest a shift in the predominant energy production pathway to glycogen and up-regulation of the foetal gene program (9,26). Although generally considered an adaptive and compensatory response, there is mounting evidence to suggest cardiac remodelling in fish also involves stressful and maladaptive aspects (9,10).

Fourier transform infrared (FTIR) spectroscopy is a powerful technique for studying molecular structures. A beam of infrared light is radiated through a sample, transferring energy which excites the molecules and causes them to vibrate. The molecular vibrations are determined by molecular dipoles, which cause specific molecular modes of vibration and rotation (27). The discrete fundamental modes of vibration mean that each molecule will absorb infrared radiation at a specific wavelength and, therefore, form a particular part of the spectrum (27). In a complex molecule, many individual bonds and functional groups vibrate at different

Metabolic and biochemical remodelling of the fish heart wavelengths to form a complex spectrum. The power of FTIR to identify biochemical molecules means it is becoming a prominent technique in biomedical fields (28,29). In particular, there is interest in developing methods of disease diagnosis using FTIR by looking at the fingerprint region or specific markers, such as shifts in lipid and glycogen levels in cancer biopsies, (30- 35) and assessing remodelling of the myocardium following heart failure (36-41).

Our aim was to assess changes in tissue biochemistry that occur with thermal remodelling of the fish ventricle, with particular interest in the predominant metabolic substrate present in the tissue. To simulate seasonal temperature change we acclimated rainbow trout to a control (10 °C), chronic cold (5 °C) or chronic warm (18 °C) temperature. Our hypothesis was that chronic cooling would increase energy demand and production and, therefore, cause an accumulation of tissue lipid and glycogen to be used as fuel. We hypothesized that following chronic warming the metabolite content of the tissue would be reduced, reflecting a reduction in cardiac muscle growth. We used Fourier transform infrared (FTIR) imaging spectroscopy to assess changes in the biochemistry of ventricular tissue cryosections and tissue that had been formalin-fixed and paraffin embedded. To compliment these experiments we semi-quantified lipid and glycogen content of ventricular tissue sections using histologically stains. Following cold acclimation, FTIR spectroscopy and histological staining showed increases in overall ventricular lipid content with a decrease in overall glycogen content compared to controls and warm-acclimated fish. Following warm acclimation, ventricular tissue showed an increase in infrared absorption at wavelengths associated with glycogen and a decrease in absorption in lipid bands. Histological analysis also suggested an increase in tissue glycogen content following chronic warming. Together, these results suggest that the increased ATP requirement of tissue following cold acclimation might be met by an increase in FAO. Following warm acclimation the opposite response occurs which may suggest a decrease in FAO and a reduction in glycolysis. If this is the case, this response would likely be due to the decreased oxygen carrying capacity of water and the increased metabolic rate of ectotherms with increasing temperatures.

Experimental Procedures

Ethical Approval

All husbandry and housing conditions were in accordance with the local handling protocols and adhere to the UK Home Office legislation. All experimental procedures were approved by the University of Manchester’s ethical review committee.

Metabolic and biochemical remodelling of the fish heart

Experimental animals

Sexually mature female rainbow trout (Onchorynchus mykiss, n = 24; mean body mass 511.7 ± 69.34) were purchased from Dunsop Bridge Trout Farm (Clitheroe, UK), housed on a 12 hr light: 12 hr dark cycle in ~500 L re-circulated aerated fresh water tanks at 10 ± 1 °C and fed to satiation 3 times per week. Water quality was ensured by 30 % water changes 3 times per week and regular tests for temperature, pH, nitrates and nitrites. Fish were held under these conditions for a minimum of 2 weeks before being randomly assigned to one of three acclimation groups; cold (5 ± 1 °C), control (i.e. no change; 10 ± 1 °C) or warm (18 ± 1 °C). These temperatures were based on previous literature describing the cardiac remodelling response in salmonids (9,10). Water temperature of the warm and cold acclimation groups was changed by 1 °C per day until desired temperature was reached and then held at that temperature for a minimum of 8 weeks before experiments. The photoperiod for the cold- acclimated animals was changed to 8 hr light: 16 hr dark cycle to simulate winter (11).

Tissue processing

Fish were killed by a blow to the head followed by severance of the spinal cord and destruction of the brain. The heart was excised, rinsed in phosphate buffered saline and weighed. The chambers of the heart were dissected from each other, weighed and bisected down the sagittal plane with one half snap frozen in OCT (Thermo Fisher Scientific, Waltham, MA, USA) by immersion in liquid nitrogen cooled 2-Methylbutane (Sigma-Aldrich, St. Louis, MO, USA) and stored at -80 °C. The other half was fixed in 10 % neutral buffered formalin solution (Sigma- Aldrich, St. Louis, MO, USA) before being processed and embedded in paraffin wax so that sections would be cut in the transverse/axial plane.

Fourier transform infrared (FTIR) imaging spectroscopy

To assess changes in the biochemistry of fish ventricular myocardium following thermal acclimation we used FTIR imaging spectroscopy of both frozen and formalin-fixed, paraffin embedded (FFPE) tissue. Ventricular tissue was sectioned at 5 µm (Leica CM3050S cryostat or Leica RM2255 microtome, Leica, Wetzlar, Germany), mounted on calcium fluoride (CaF2) slides. Slides of frozen tissue were stored at -80 °C and thawed in a vacuum desiccator for 60 minutes, prior to being placed in the purge box of the spectrometer for measurements. Slides of FFPE tissue were stored in a bench top desiccator and FTIR imaging was conducted with samples in wax. Transmission mode FTIR imaging spectroscopy was performed on a Varian 670-IR spectrometer coupled with a Varian 620-IR imaging microscope (Agilent Technologies, CA, USA) equipped with a 128 x 128 pixel liquid nitrogen cooled Mercury-Cadmium-Telluride focal plane array (FPA). Data was collected in the 1000-3800 cm-1 range, at a spectral resolution of 4 cm-1, with the co-addition of 64 scans for sample spectra and 128 scans for the

Metabolic and biochemical remodelling of the fish heart background spectra. At a spatial resolution of 1.1 µm, a tissue sampling area of ~1200 x 1200 µm was captured by hyperspectral image.

Tissue Histology

To determine overall tissue morphology and semi-quantitatively assess tissue lipid and glycogen we used histological staining techniques of tissue sections. For overall tissue morphology, the serial tissue sections to those used for FTIR analysis were stained with haematoxylin and eosin (H & E). To assess tissue lipid content, frozen ventricular tissue was sectioned at 10 µm (Leica CM3050S cryostat, Leica, Wetzlar, Germany), mounted onto glass slides (Super frost plus, Thermo Fisher Scientific, Waltham, MA, USA) and stained with oil red O. A negative control was prepared for each section by taking serial sections and treating them with acetone prior to the oil red O protocol to remove all lipids. To assess tissue glycogen content, formalin-fixed tissue samples were processed, embedded in paraffin wax, sectioned at 5 μm (Leica RM2255 microtome, Leica, Wetzlar, Germany) mounted onto glass slides and stained with Periodic acid-Shiff (PAS) stain for glycogen. A negative control was prepared for each section by taking serial sections and digesting the glycogen in amylase before the PAS staining protocol. Images were analysed using bright-field microscopy (Leica Wetzlar, Germany) and ImageJ software (42). Lipid and glycogen content of the tissues were determined by pixel count, compared to their corresponding negative controls, and then expressed as a percentage increase compared to the specific corresponding negative control for that sections. Three tissue sections were considered for each individual to ensure consistency in measurements. On each tissue section 3 separate image montages were taken along transects across the full diameter of the cross section. All histological analysis was conducted blind to the acclimation group.

Calculations and statistical analyses

Infrared spectral data was imported into MATLAB (MATLAB 2014a, Mathworks, Natick, MA, USA) and quality tested by the amide I region (1597-1738 cm-1). The absorbance values to determine which spectra were accepted or rejected were determined separately for each hyperspectral image by reference to an image of the tissue section. The band associated with ambient gas-phase CO2 was removed, as was any data outside of the specified wavenumber range. Despite all tissue sections being cut to a nominal thickness, due to tissue morphology there may be some variation in tissue thickness across the sampled area. To account for this, data was vector normalized and subjected to a PCA noise reduction (43,44). In addition, tissue cryosections show varying ‘roughness’ of edges across the sample area which may cause scattering artifacts. To correct for this, data from frozen tissue was subjected RMieS-EMSC correction with 100 iterations using a MatrigelTM spectrum as the initial reference point (45). The scatter correction was not necessary for the FFPE tissue as the wax gives a uniform

Metabolic and biochemical remodelling of the fish heart thickness. Regions of interest were taken to ensure even sections of the sample were analysed and that the full tissue biochemistry was captured. The region of interest data was then subjected to a K-fold algorithm which randomized the spectra and then condensed them into 1000 average spectra for each individual. All subsequent analyses were conducted on these K-folded spectra. Due to the infrared absorption of paraffin wax some parts of the spectra for FFPE tissue were discarded. The region analysed was, therefore, in the 1000-1800 cm-1 wavenumber range, but with the 1427-1487 cm-1 wavenumber region removed.

Lipid and glycogen content of tissue sections were determined by pixel count and then expressed as a percentage increase compared to the corresponding negative control of each section. Significance was considered at P < 0.05, determined by general linear model (GLM) (SPSS Statistics, IBM, Armonk, NY, USA). For lipid and glycogen content values are expressed as mean ± S. E. Statistical analyses are detailed in figure legends.

Results

Biochemical homogeneity of the compact and spongy myocardium

The rainbow trout ventricle is composed of two distinct myocardial tissue layers known as the compact and spongy myocardium (46). The spongy myocardium is highly trabeculated and covers almost the entire inner part of the ventricle, leaving only a very small lumen (47). This layer is encased with a dense compact layer which also provides a coronary circulation (46). These two layers are generally thought of as distinct tissues and considered separately. We used FTIR imaging spectroscopy to determine biochemical differences between these myocardial tissue layers. Figure 1 A, B & C show representative ventricular tissue sections for cold-acclimated, control and warm-acclimated rainbow trout, stained with H & E and imaged under bright-field light, showing the compact and spongy myocardium. FTIR hyperspectral, heat-map images were generated based on the absorption of overall protein (the intensity of amide I at 1595-1695 cm-1), lipid (the intensity of 2830-3030 cm-1) and glycogen (the intensity of 1035-1045 cm-1) (Figure 1D, E & F), and show the tissue to be reasonably biochemically homogenous across both the compact and spongy myocardium. Next we used principal component analysis (PCA) to determine the important features of the tissue biochemistry and then used the principal component (PC) scores to perform a K-means cluster analysis of the tissue section, as previously performed by Hughes et al. (35). Figure 2 shows a representative control ventricular section with clusters based on PC 1 and 2 (Figure 2A), PC 1-3 (Figure 2B), PC 1-4 (Figure 2C) and PC 1-5 (Figure 2D). We then extracted the spectra form the areas of tissue that contributed to each cluster and their spectral profiles suggested a high degree of heterogeneity between clusters, with distinct spectral profiles (Figure 2E). Again, there appeared to be a high level of homogeneity across tissue sections and no biochemical difference between the compact and spongy myocardium, as clusters were evenly distributed.

Metabolic and biochemical remodelling of the fish heart

Finally, we took 3 discrete regions of interest (ROIs) from the spongy myocardium and 3 ROIs from the compact myocardium (Figure 3A). We then plotted the mean spectra of each of these ROIs (each offset by an absorbance value of 0.01) (Figure 3B) and used PCA to determine separation of these regions by their biochemistry. We did not find separation by PC 1, which accounted for 69.8 % of the variation, but found that the compact ROIs were separated from the spongy ROIs by PC 2, which accounted for 11.6 % of the variation (Figure 4A). The main features of the PCs are shown by their corresponding loadings plot. The key wavenumber regions that separate the biochemistry of the tissue by PC 2 show that the compact tissue has higher absorption of lipids, amide I, amide II and amino acid side chains, but lower absorption in the amide A, amide B, lipid ester stretching and glycogen bands (Figure 4B). There was no separation based on the variation accounted for by PC 3 or PC 4 (Figure 4B, C & D).

Metabolic and biochemical remodelling of the fish heart

Figure 1. Biochemical homogeneity of the compact and spongy myocardium. Representative bright-field micrograph of (A) cold-acclimated (CA), (B) control and (C) warm-acclimated (WA) rainbow trout ventricle, stained with haematoxylin and eosin (H & E), showing the spongy (s) and compact (c) myocardial layers. Hyperspectral images produced based on the overall (D) protein (1595-1695 cm-1), (E) lipid (2830-3030 cm-1) and (F) glycogen (1035-1045 cm-) absorption profiles of the tissue.

Metabolic and biochemical remodelling of the fish heart

Figure 2. K-means cluster analysis of rainbow trout ventricular cryosections. Following data acquisition by FTIR imaging spectroscopy principal component analysis (PCA) was performed on the data to determine the regions of key spectral significance. The principal component (PC) scores were used to perform a K-means cluster analysis on the tissue. (A) shows a representative tissue cryosection with the K-means cluster analysis for PC 1 (blue) & 2 (red). (B) shows the same section but with K-means cluster analysis including PC 1-3 (blue, red, green), (C) PC 1-4 (blue, red green, magenta) and (D) PC 1-5 (blue, red, green, magenta, cyan). The spectra for each area were then extracted from the cluster analysis. (E) shows the mean the spectrum for each cluster (or PC) in the 3700-1000 cm-1 range. Each spectra off-set by an infrared absorption value of 0.01 and the spectra of each PC is drawn in the colour it appears on the tissue image.

Metabolic and biochemical remodelling of the fish heart

Figure 3. A comparison of the rainbow trout compact and spongy myocardium by regions of interest (ROIs). (A) a FTIR hyperspectral tissue image with three ROIs were generated for the compact myocardium (ROI C1, blue; ROI C2, red; ROI C3, green) and three ROIs were generated for the spongy myocardium (ROI S1, cyan; ROI S2, magenta; ROI S3, black). (B) the mean spectra for each ROI in the 3700-1000 cm-1 wavenumber range, each offset by an infrared absorption value of 0.01.

Metabolic and biochemical remodelling of the fish heart

Figure 4. Principal component analysis (PCA) for FTIR spectra, for regions of interest on a control rainbow trout ventricular cryosection. (A) principal component (PC) scores plot for PC 1 and PC 2 for compact and spongy regions of interest (ROIs), with compact ROI 1 (ROI C1; blue), ROI C2 (red), ROI C3 (green) and spongy ROI 1 (ROI S1; cyan), ROI S2 (magenta), ROIS3 (black). (B) the corresponding PC loadings plot for PC 1 and PC 2. (C) PC scores plot for PC 3 and PC 4 from the same analysis. (D) the corresponding PC loadings plot for PC 1 and PC 2. In the scores plot, positive values show correlation with that particular PC and negative values show differences to that particular PC. In the loadings plot, positive values show the key characteristics that are increased in that particular PC and negative values show the key characteristics that are reduced in that particular PC.

Metabolic and biochemical remodelling of the fish heart

Changes in infrared spectra of ventricular tissue sections with thermal acclimation

To assess the biochemical changes in ventricular tissue following thermal acclimation we performed FTIR imaging spectroscopy of tissue cryosections and FFPE tissue sections. Mean spectral absorptions for each group at specific peaks of interest are given for the compact myocardium in Table 1 and for the spongy myocardium in Table 2. To analyse absorption patterns in the 2700-3700 cm-1 wavenumber range we used tissue cryosections. The cold- acclimated ventricle showed increases in bands characteristic of lipid, including the

-1 -1 asymmetric stretch of O-H, at 3420 cm , the asymmetric stretch of CH3, at 2959 cm , and

-1 -1 -1 CH2, at 2924 cm , and the symmetric stretch of CH3, at 2872 cm , and CH2, at 2853 cm of lipids, triglycerides and fatty acids in both the compact and spongy myocardium, compared to warm-acclimated animals (Figure 5A & B). Infrared absorption of the amide B band at 3067 cm-1, linked to the secondary structure of proteins, was reduced following warm acclimation compared to cold-acclimated animals (Figure 5A & B). Infrared absorption of the control ventricle was similar to the cold-acclimated ventricle within the 2700-3700 cm-1 wavenumber range. However, the absorption of the amide A band, at 3308 cm-1, and the symmetric stretch

-1 of CH3, at 2872 cm , in both the compact and spongy myocardium and absorption in the

-1 -1 asymmetric stretch of CH3, at 2959 cm , and CH2, at 2924 cm , was higher in control animals than cold- or warm-acclimated, suggesting temperature independent effects in protein and lipid bands (Figure 5A & B).

The infrared absorption of cryosections in the 1000-1800 cm-1 region showed increased absorption of bands characteristic of lipids with cold-acclimation. The C=O stretch of lipid

-1 -1 - esters, at 1740 cm , the CH2 and CH3 bands, at 1453 cm , the symmetric stretch of COO in amino acids and fatty acids were all increased in both the compact and spongy layer of the ventricle (Figure 5A & B). Collagen side chains show district infrared absorption at 1338 cm-1, which was increased following cold acclimation, particularly in the compact myocardium (Figure 6). In addition, the amide III, at 1308 cm-1, the asymmetric and symmetric stretch of phosphates, at 1236 cm-1 and 1082 cm-1, and the symmetric stretch of CO in serine, therosine and tyrosine were increased in both the compact and spongy myocardium following cold compared to warm acclimation (Figure 5A & B). Conversely, the absorbance of protein shown by the amide I, at 1655 cm-1, and the amide II, at 1545 cm-1, were reduced in the cold- acclimated group compared to controls, as was the asymmetric stretch of glycogen, at 1152 cm-1, and the COH deformation of glycogen, at 1040 cm-1, in both the compact and spongy myocardium (Figure 5A & 5B). In this region the control group, again, followed the cold- acclimated group but was more pronounced, showing lower absorption than the cold- acclimated group when the cold-acclimated group was lower than the warm-acclimated group and higher than the cold-acclimated group at points where the cold-acclimated group was higher than the warm-acclimated group (Figure 5A &B).

Metabolic and biochemical remodelling of the fish heart

Table 1. Differences in the mean infrared absorption spectra of the cold-acclimated (5 °C), control (10 °C) and warm-acclimated (18 °C) rainbow trout compact myocardium at various peaks of interest, given by wavenumber in cm-1, and their peak assignment.

Metabolic and biochemical remodelling of the fish heart

Table 2. Differences in the mean infrared absorption spectra of the cold-acclimated (5 °C), control (10 °C) and warm-acclimated (18 °C) rainbow trout spongy myocardium at various peaks of interest, given by wavenumber in cm-1, and their peak assignment.

Metabolic and biochemical remodelling of the fish heart

Figure 5. Mean spectra for cold-acclimated (5 °C; blue) control (10 °C; green) and warm- acclimated (18 °C; red) rainbow trout (A) compact and (B) spongy ventricle cryosections (n = 6 cold, 5 control, 6 warm). Spectra has been quality tested, been subjected to a noise reduction, vector normalization and RMieS scatter correction. Numbers denote spectral bands at key wavenumbers of interest, which are detailed in Table 1 and 2. 1, 3420 cm-1; 2, 3308 cm- 1; 3, 3067 cm-1; 4, 2959 cm-1; 5, 2924 cm-1; 6, 2872 cm-1; 7, 2853 cm-1; 8, 1740 cm-1; 9, 1655 cm-1; 10, 1545 cm-1; 11, 1453 cm-1; 12, 1391 cm-1; 13, 1308 cm-1; 14, 1236 cm-1; 15, 1173 cm- 1; 16, 1152 cm-1; 17, 1082 cm-1; 18, 1040 cm-1.

Metabolic and biochemical remodelling of the fish heart

Figure 6. Thermal remodelling of ventricular collagen. Hyperspectral images based on the absorption of a specific collagen side chain, mean centred at 1338 cm-1, for (A) cold-acclimated (5 °C), control (10 °C) and warm-acclimated (18 °C) rainbow trout ventricle cryosections.

Metabolic and biochemical remodelling of the fish heart

PCA did not effectively separate the groups based on acclimation for either the compact (Figure 7) or spongy myocardium (Figure 8). However, PC 1 separated 3 control animals and 2 cold-acclimated animals from the others in both the compact (Figure 7A) and spongy myocardium (Figure 8A). From the corresponding loadings plot it appears that PC 1 is separating the animals that have high amide I and II to those that have low amide I and II, but higher amide A, lipid bands and phosphate bands (Figure 7B & 8B). PC 3 and PC 4 did not separate the acclimation groups in either the compact of spongy myocardium (Figure 7C, 7D, 8C & 8D). Indeed, when the mean spectra for each individual animal are plotted together, there is variation and what appears to be two separate spectral profiles within the cold- acclimated and control groups (Supplementary figure 1).

Metabolic and biochemical remodelling of the fish heart

Figure 7. Principal component analysis of FTIR spectra for cold-acclimated (5 °C; blue) control (10 °C; green) and warm-acclimated (18 °C; red) compact layer of the rainbow trout ventricle cryosections (n = 6 cold, 5 control, 6 warm). (A) principal component (PC) scores plot for PC 1 and PC 2 for cold-acclimated, control and warm-acclimated tissue. (B) the corresponding PC loadings plot for PC 1 and PC 2. (C) PC scores plot for PC 3 and PC 4 for cold-acclimated, control and warm-acclimated tissue. (D) the corresponding PC loadings plot for PC 3 and PC 4. In the scores plot, positive values show correlation with that particular PC and negative values show differences to that particular PC. In the loadings plot, positive values show the key characteristics that are increased in that particular PC and negative values show the key characteristics that are reduced in that particular PC. Spectra from each individual fish have been subjected to a K-folding algorithm, which reduces the spectral number to 1000 mean spectra per individual.

Metabolic and biochemical remodelling of the fish heart

Figure 8. Principal component analysis of FTIR spectra for cold-acclimated (5 °C; blue) control (10 °C; green) and warm-acclimated (18 °C; red) spongy layer of the rainbow trout ventricle cryosections (n = 6 cold, 5 control, 6 warm). (A) principal component (PC) scores plot for PC 1 and PC 2 for cold-acclimated, control and warm-acclimated tissue. (B) the corresponding PC loadings plot for PC 1 and PC 2. (C) PC scores plot for PC 3 and PC 4 for cold-acclimated, control and warm-acclimated tissue. (D) the corresponding PC loadings plot for PC 3 and PC 4. In the scores plot, positive values show correlation with that particular PC and negative values show differences to that particular PC. In the loadings plot, positive values show the key characteristics that are increased in that particular PC and negative values show the key characteristics that are reduced in that particular PC. Spectra from each individual fish have been subjected to a K-folding algorithm, which reduces the spectral number to 1000 mean spectra per individual.

Metabolic and biochemical remodelling of the fish heart

To validate these results we took two approaches, we re-imaged a number of samples, with the same result (Supplementary figure 2B) and we imaged FFPE ventricular tissue. Due to the infrared absorption of paraffin wax, the usable spectra for these samples lies in the 1000-1800 cm-1 wavenumber region. Although there was still a high degree of variation in absorption intensity of FFPE samples (Supplementary figure 3), the mean spectral profiles showed that following cold acclimation infrared absorption was higher across the whole wavenumber range than control or warm-acclimated animals in both the compact and spongy myocardium (Figure 9). In the compact myocardium, the infrared absorption of the control animals in the amide I and amide II bands of protein were lower than that of the warm-acclimated animals, but increased from 1000 to ~1200 cm-1. In the spongy myocardium the infrared absorption intensity of the amide I band was similar to that of the warm-acclimated animals. At the peak of the amide II band the absorption of the control group was higher than that of the warm- acclimated group and it remained higher for the remainder of the region. PCA did not effectively separate the biochemistry of the FFPE compact (Figure 10) or spongy (Figure 11) myocardium between acclimation groups. However, overall the FFPE tissue did not show the two spectral profiles as in the frozen tissue. With reference to previous spectra and a MatrigelTM spectrum (Supplementary figure 2C) it appears that the high amide I and II, with low phosphates is the more ‘normal’ spectral profile. To determine the effects of these low amide, high phosphate spectra we re-analysed the data with these animals removed.

Metabolic and biochemical remodelling of the fish heart

Figure 9. Mean spectra for cold-acclimated (5 °C; blue) control (10 °C; green) and warm- acclimated (18 °C; red) rainbow trout. (A) compact and (B) spongy formalin-fixed, paraffin embedded ventricular tissue sections on calcium fluoride (CaF2) slides (n = 3 cold, 4 control, 4 warm). Spectra were measured in the 1000-1800 cm-1 wavenumber range and data in the 1427-1487 cm-1 wavenumber region removed. Spectra were quality tested, subjected to a noise reduction and vector normalization. Numbers denote spectral bands at key wavenumbers of interest, which can be cross-referenced to Table 1 & 2. 1, 1655 cm-1; 2, 1545 cm-1; 3, 1391 cm-1; 4, 1236 cm-1; 5, 1040 cm-1.

Metabolic and biochemical remodelling of the fish heart

Figure 10. Principal component analysis of FTIR spectra for cold-acclimated (5 °C; blue) control (10 °C; green) and warm-acclimated (18 °C; red) compact layer of formalin-fixed, paraffin embedded rainbow trout ventricular tissue (n = 3 cold, 4 control, 4 warm). (A) principal component (PC) scores plot for PC 1 and PC 2 for cold-acclimated, control and warm- acclimated tissue. (B) the corresponding PC loadings plot for PC 1 and PC 2. (C) PC scores plot for PC 3 and PC 4 for cold-acclimated, control and warm-acclimated tissue. (D) the corresponding PC loadings plot for PC 3 and PC 4. In the scores plot, positive values show correlation with that particular PC and negative values show differences to that particular PC. In the loadings plot, positive values show the key characteristics that are increased in that particular PC and negative values show the key characteristics that are reduced in that particular PC. Spectra from each individual fish have been subjected to a K-folding algorithm, which reduces the spectral number to 1000 mean spectra per individual.

Metabolic and biochemical remodelling of the fish heart

Figure 11. Principal component analysis of FTIR spectra for cold-acclimated (5 °C; blue) control (10 °C; green) and warm-acclimated (18 °C; red) of formalin-fixed, paraffin embedded rainbow trout spongy ventricular tissue (n = 3 cold, 4 control, 4 warm). (A) principal component (PC) scores plot for PC 1 and PC 2 for cold-acclimated, control and warm-acclimated tissue. (B) the corresponding PC loadings plot for PC 1 and PC 2. (C) PC scores plot for PC 3 and PC 4 for cold-acclimated, control and warm-acclimated tissue. (D) the corresponding PC loadings plot for PC 3 and PC 4. In the scores plot, positive values show correlation with that particular PC and negative values show differences to that particular PC. In the loadings plot, positive values show the key characteristics that are increased in that particular PC and negative values show the key characteristics that are reduced in that particular PC. Spectra from each individual fish have been subjected to a K-folding algorithm, which reduces the spectral number to 1000 mean spectra per individual.

Metabolic and biochemical remodelling of the fish heart

The re-analysis of the mean spectra showed a very similar overall profile to the original analysis (Figure 12). The overall spectral profile of the compact myocardium was an increase in lipid bands and a decrease in glycogen bands (Figure 12A). However, the increase in amide A and phosphates was removed, and the cold-acclimated ventricle showed increased absorption in the amide II band (Figure 12A). The warm-acclimated animals retained the higher levels of amide I and glycogen bands, however, the absorption between cold- and warm-acclimated animals in the amide B band became similar (Figure 12A). High infrared absorption of the control group in the amide A, lipid ester and phosphate bands remained, and absorption remained lower in the in the amide B, amide I and amide II bands. In the spongy myocardium, the difference between cold- and warm-acclimated animals were reduced (Figure 12B). The key differences that remained in the FFPE tissue in agreement with the frozen tissue were; increased lipid and phosphate bands and reduced glycogen bands for the cold-acclimated compared to the warm-acclimated group. The spectrum of the control animals remained different to both the cold- and warm-acclimated suggesting that temperature independent variation remains. PCA did not effectively separate the biochemical variation in either the compact (Figure 13) or the spongy (Figure 14) myocardium, following thermal acclimation.

Metabolic and biochemical remodelling of the fish heart

Figure 12. Re-analysis of mean spectra for cold-acclimated (5 °C; blue) control (10 °C; green) and warm-acclimated (18 °C; red) rainbow trout (A) compact and (B) spongy ventricle cryosections with outliers removed (n = 4 cold, 2 control, 6 warm). Spectra has been quality tested, been subjected to a noise reduction, vector normalization and RMieS scatter correction. Numbers denote spectral bands at key wavenumbers of interest which can be cross-referenced to Tables 1 & 2. 1, 3420 cm-1; 2, 3308 cm-1; 3, 3067 cm-1; 4, 2959 cm-1; 5, 2924 cm-1; 6, 2872 cm-1; 7, 2853 cm-1; 8, 1740 cm-1; 9, 1655 cm-1; 10, 1545 cm-1; 11, 1453 cm- 1; 12, 1391 cm-1; 13, 1308 cm-1; 14, 1236 cm-1; 15, 1173 cm-1; 16, 1152 cm-1; 17, 1082 cm-1; 18, 1040 cm-1.

Metabolic and biochemical remodelling of the fish heart

Figure 13. Principal component analysis of FTIR spectra for cold-acclimated (5 °C; blue) control (10 °C; green) and warm-acclimated (18 °C; red) compact layer of the rainbow trout ventricle cryosections with outliers removed (n = 4 cold, 2 control, 6 warm). (A) principal component (PC) scores plot for PC 1 and PC 2 for cold-acclimated, control and warm- acclimated tissue. (B) the corresponding PC loadings plot for PC 1 and PC 2. (C) PC scores plot for PC 3 and PC 4 for cold-acclimated, control and warm-acclimated tissue. (D) the corresponding PC loadings plot for PC 3 and PC 4. In the scores plot, positive values show correlation with that particular PC and negative values show differences to that particular PC. In the loadings plot, positive values show the key characteristics that are increased in that particular PC and negative values show the key characteristics that are reduced in that particular PC. Spectra from each individual fish have been subjected to a K-folding algorithm, which reduces the spectral number to 1000 mean spectra per individual.

Metabolic and biochemical remodelling of the fish heart

Figure 14. Principal component analysis of FTIR spectra for cold-acclimated (5 °C; blue) control (10 and control (10 °C; green) and warm-acclimated (18 °C; red) spongy layer of the rainbow trout ventricle cryosections with outliers removed (n = 4 cold, 2 control, 6 warm). (A) principal component (PC) scores plot for PC 1 and PC 2 for cold-acclimated, control and warm- acclimated tissue. (B) the corresponding PC loadings plot for PC 1 and PC 2. (C) PC scores plot for PC 3 and PC 4 for cold-acclimated, control and warm-acclimated tissue. (D) the corresponding PC loadings plot for PC 3 and PC 4. In the scores plot positive values show correlation with that particular PC and negative values show differences to that particular PC. In the loadings plot positive values show the key characteristics that are increased in that particular PC and negative values show the key characteristics that are reduced in that particular PC. Spectra from each individual fish have been subjected to a K-folding algorithm, which reduces the spectral number to 1000 mean spectra per individual.

Metabolic and biochemical remodelling of the fish heart

Histological analysis of tissue sections following thermal acclimation

To complement the results from the FTIR analysis we used histological stains to assess tissue content of lipid and glycogen. We used oil red O stain, which selectively stains lipid droplets and neutral triglycerides in red (48). Figure 15A shows a representative micrograph of control fish ventricle stained with oil red O. The lipid content of the compact myocardium was increased by 5.1-fold, and lipid content of the spongy myocardium increased by 4.8-fold following cold acclimation, compared to controls (Figure 15B & C). Figure 15D shows a representative micrograph of control fish ventricle stained with periodic acid Schiff (PAS) stain which selectively stains glycogen in purple. Following cold acclimation glycogen content of the compact myocardium was 2.4-fold lower, and the spongy myocardium 1.9-fold lower, than controls (Figure 15E & F). The opposite response occurred following warm acclimation with compact glycogen content 2-fold higher, and the spongy myocardium 1.9-fold higher than controls (Figure 15E & F).

Metabolic and biochemical remodelling of the fish heart

Figure 15. Histological staining for lipid and glycogen in the rainbow trout ventricle. (A) A representative micrograph of a control rainbow trout ventricular cryosection stained with oil red O, which stains lipid droplets red. Semi-quantitative analysis of oil red O staining for (B) the compact and (C) the spongy myocardium for cold-acclimated (5 °C; blue), control (10 °C; green) and warm-acclimated (18 °C; red) rainbow trout ventricle cryosections (n = 5 for each group). (D) A representative micrograph of a formalin-fixed control rainbow trout ventricular section stained with periodic acid Schiff (PAS), which stains glycogen purple. Semi- quantitative analysis of PAS staining for (E) the compact and (F) the spongy myocardium for cold-acclimated (5 °C; blue), control (10 °C; green) and warm-acclimated (18 °C; red) rainbow trout ventricle sections (n = 5 for each group). Significant differences between acclimation groups was determined by general linear model with a Tukey post hoc test for differences between groups and indicated by dissimilar letters. Values are presented as mean ± S.E.

Metabolic and biochemical remodelling of the fish heart

Discussion

The fish heart displays a remodelling response to seasonal temperature change. With prolonged cold temperature there is an associated hypertrophy of the ventricle by increase in size of its constituent myocytes, causing an increase in protein synthesis (9,10,12,26). With prolonged warm temperature protein synthesis slows, however, there is an increase in whole animal metabolic rate, which directly increases cardiac demand (1-8). Here, we investigated the effect of thermal remodelling on the biochemistry of the fish ventricle, with particular interest to changes in metabolic substrates of the tissue. Our results suggest an accumulation of lipid following cold acclimation, with a decrease in glycogen in both the spongy and compact myocardium. Following warm acclimation we saw the opposite response, with reductions in lipid content of the ventricular tissue, but an increase in tissue glycogen content in both the compact and spongy myocardium. Although further experiments are required, these results may suggest that with cold acclimation the increased energetic demands of the cardiac myocytes are met by increases in FAO. However, with warm acclimation there may be a decreased reliance on energy production by FAO and a metabolic shift to energy production by glycolysis, indicated by a decrease in tissue glycogen stores, which is likely due to the decreased availability of oxygen at high temperature.

Differences in FTIR absorption profiles due to sample preparation

FITR imaging spectroscopy can be performed on tissue sections that have been snap frozen or chemically fixed in formalin, processed and embedded in paraffin wax (formalin-fixed, paraffin embedded; FFPE). There a number of advantages and considerations required for using either method. Processing of FFPE tissue preserves the secondary structure of proteins and some membrane bound lipids by formalin and ethanol reacting with C=C double bonds (49). However, both formalin fixing and ethanol cause coagulation of the globular proteins in the cytoplasm, which leads to a loss of integrity of intracellular organelles, e.g. mitochondria. Ethanol also denatures the tertiary structure of proteins and causes lipids that are not membrane bound or were not fixed by the formalin to precipitate out of the tissue. Furthermore, chemical fixation and paraffin embedding causes a shift in the amide I and II bands due to changes in protein secondary structure and causes loss of the lipid ester peak (C=O) at 1740 cm-1 due to the removal of cellular lipids (50). Finally, using FFPE tissue means that there are either large infrared absorption bands due to paraffin wax, or the paraffin wax has to be removed from the samples, using hexane, alcohol or xylene (51,52). At present, it is unclear if dewaxing samples causes any further changes in tissue biochemistry.

Snap freezing of tissue avoids the use of organic solvents and, therefore, gives superior preservation of soluble lipids, enzymes and antigens, and there is no shift in the amide peaks (53). However, it can be associated with a reduction in the intensity of the Amide I and Amide

Metabolic and biochemical remodelling of the fish heart

II bands, suggesting a change in conformation of the proteins. In addition, Shim and Wilson (1996) showed an increase in the lipid: protein ratio with drying, suggesting perturbations in protein vibrational modes with dehydration perturbations of cellular chemistry making spectra hard to interpret. Furthermore, the increased temperature needed for sectioning can cause ice crystals to form, which may damage the tissue. Finally, the validity of the chemical data over time has been questioned as Stitt et al. (54) reported appreciable oxidation of lipids over a 90 hr period.

As we had hypothesized that thermal acclimation would cause a shift in cellular energetics we wanted to be able to use as much of the spectrum as possible for interpretation. Frozen tissue provides the best way to assess levels of metabolites as the lipid regions remain intact and there are no other changes to tissue biochemistry due to treating them with organic solvents (49). However, the variation in spectral profiles between samples, even in the same group, prompted us to analyze FFPE tissue also. The differences in the present study between the spectra from frozen and FFPE samples is unclear. We expected that there would be some differences due to removal of biochemistry by the formalin fixation and paraffin embedding process, as unbound molecules are often washed out of the cell, especially lipids (50). However, there were also changes in the ratios of protein bands and the strong infrared absorption of the phosphate band in frozen tissue disappeared when analyzing FFPE tissue.

Our data showed two distinct spectral profiles in the cold-acclimated and control cryosectioned tissue, the reasons for which are unclear. When interpreting FTIR data it is important to reduce or remove a number of artifacts of the technique due to the physics of light when using FTIR imaging spectroscopy, such as scattering effects at the edge of tissue (45,55). When applying FTIR imaging spectroscopy to thawed cryosections, as we did in this study, tissue thickness varies and, in porous tissue such as the heart, scattering at holes can cause large changes in the spectra. We used a number of complex mathematical algorithms to account for artifacts including quality test for each section based on the amide I band to remove any tissue that did not have a sufficient level of protein, a noise reduction of the spectra and a vector normalization. We also used the RMieS scatter correction algorithm created by Bassan et al. (45), which works by applying peak and baseline corrections at certain points in the spectra to shift known peaks and baseline sections to their correct position. The correction appeared to work well for this tissue, removing the scattering, but it was created for dewaxed FFPE sections rather than frozen, which are known to have a number of key peaks at slightly different wavenumbers due to the processing (50,53). Finally, correct sample storage is important when using frozen tissue. Stitt et al. (54) showed that oxidation of polyunsaturated fatty acids and the lipid carbonyl group occurs rapidly when stored at room temperature. For these reasons samples were stored in a freezer at -80 °C until the day of experimentation. On the day of experiments, tissue was removed from the freezer, thawed in a vacuum desiccator for 1 hr and then place into the purge box of the spectrometer for 1 hr before being imaged. This

Metabolic and biochemical remodelling of the fish heart protocol meant that each measurement began > 2.5 hrs after removing the sample from the freezer. Although the thaw and purge times delayed the start of experiments they are necessary to ensure that the tissue is in a stable condition and that water vapour does not interfere with the sample spectrum. In any case, > 2.5 hrs should be a short enough time not to see any appreciable oxidation, and the marker bands used by Stitt et al. (54) (at ~3012 cm- 1 and 1745 cm-1) were present in our spectra. Overall, therefore, this study highlights further implications of sample preparation shows that considerations should be made when planning and interpreting the results of FTIR experiments.

Biochemical homogeneity of the compact and spongy myocardium

The fish ventricle is composed of two distinct myocardial tissue layers. A highly trabeculated ‘spongy’ myocardium that covers the inner part of the ventricle is encased by a thin layer of dense ‘compact’ myocardium (46,47). Although both layers are formed of cardiac myocytes these layers are generally considered separately and perform different functions. By comparing the H & E stained sections and the hyperspectral images based on the infrared absorption of protein, lipid and carbohydrate the tissue sections appeared homogeneous both across the spongy and compact myocardium and between temperature acclimation groups. However, the K-means cluster analysis based on the PCA loadings of the tissue revealed heterogeneity of tissue sections. The reasons for this heterogeneity may be due to the different cells types in cardiac tissue. Ventricular cardiomyocytes form the muscular wall, cardiac fibroblasts regulate the extracellular matrix, endothelial cells form the endocardium, and there are numerous extracellular matrix components (56). It is possible that the infrared absorption profiles of clusters reflect these differences in tissue. Indeed, Wood et al. (58) found large spectral differences between cell types with fibroblasts having strong bands due to overlapping collagen and phosphate and endothelial cells having reduced phosphate and glycogen bands. Similarly, differences have been shown in fibroblasts during their role in remodelling of connective tissue through cancer progression (57). In addition, the presence and/or differing levels of blood cells in the tissue may have influenced overall spectra, with different blood cell types also showing characteristic spectral profiles (58). Although we washed all hearts in PBS immediately after they were excised, we cannot discount a potential effect of residual blood.

These clusters were consistent across both the compact and spongy myocardium suggesting that there is still a high degree of biochemical homogeneity across these two tissue types. However, we found a difference in biochemistry of the spongy and compact myocardium when we generated ROIs for each of the tissue types. PCA separated the key element of tissue biochemistry along PC 2, which showed compact tissue to have higher absorption of lipids, amide I, amide II and amino acid side chains, but lower absorption in the amide A, amide B, lipid ester stretching and glycogen bands. Separation along PC 2 accounted for only 11.6 %

Metabolic and biochemical remodelling of the fish heart of the total biochemical variation in the tissue which suggests that, although they are highly different in organization, the biochemistry is only separated by secondary variation.

Changes in cellular energetics following thermal acclimation

For contractile function the heart needs a continuous and sufficient supply of ATP to power active transport of ions as well as regulate the intracellular environment (17). Hypertrophic cardiac growth increases the energy requirement of the myocardium to synthesize the new proteins that increase size and strength of the cell. Under normal conditions, the adult heart generates ATP by aerobic mitochondrial FAO (16). However, in times of high ATP demand an increased reliance on ATP production by glycolysis can occur (16,18,19). Following cold acclimation of the fish heart, previous studies have shown an up-regulation of genes involved in both FAO and glycolysis (12,13,26).

In this study, the FTIR absorption profile and histological staining with oil red O suggest an increase in tissue lipid following cold acclimation. There was a large increase in staining of lipid droplets and neutral triglycerides with oil red O, which agrees with the overall increase in the spectral profile of triglycerides and cholesterol esters. An accumulation of tissue lipid may suggest an increased reliance on FAO following cold acclimation, rather than an energetic shift to glycolysis, which agrees with previous research on thermally acclimated fish (12,13,23- 25). However, the overall increase in lipid FTIR absorption bands was only modest suggesting that polar lipid molecules, which are not stained by oil red O, may decrease following cold acclimation and reduce overall lipid absorption (48). The CH stretching region of the spectrum (from ~2700 - 3100 cm-1) can be used to estimate overall lipid content (41) and absorption in this region was increased following cold acclimation compared to the warm-acclimated group. This is a very complex region of the spectrum, created by a large number of nearly identical vibrations, super-imposed to form a broad band. Within the broad band there are two regions characteristic of saturated long chain hydrocarbons, showing the symmetric and asymmetric

-1 -1 stretching modes of CH2 (~2926 cm and 2855 cm ) (41,59) which were increased in cold- compared to warm-acclimated fish. There is another pair of distinct regions which show the

-1 -1 symmetric and asymmetric stretching modes of CH3 (2953 cm and 2874 cm ) (41), which were, again increased in cold- compared to warm-acclimated fish. We also saw an increase in the C=O stretching of cholesterol esters and triglycerides (~1740 cm-1) following cold acclimation compared to warm acclimation. Due to the difference in shape of the shoulder it is likely that the difference is due to an increase in tissue cholesterol esters rather than an increase triglycerides as the carbonyl stretching for triglycerides is more typically ~1745 cm-1 (37), at which point the spectra are closer. Although further experiments are required to assess changes in metabolic enzymes following temperature acclimation, together these results may suggests aerobic metabolism by FAO is increased to power cardiac functions at low temperatures and give the necessary energy for hypertrophic growth, with a decrease in

Metabolic and biochemical remodelling of the fish heart glycolytic pathways. Although currently speculative, these results agrees with previous research on heart and skeletal muscle function following chronic cold (12,13,23,24,60-62). The ability to maintain FAO would likely due to a sufficient oxygen supply to maintain aerobic metabolism with cold acclimation, as the oxygen carrying capacity of water is increased in the cold (63).

Interestingly, we found the opposite response in the hearts of warm-acclimated fish, with a decrease in tissue lipid content and an increase in tissue glycogen content. We saw an increase in the infrared absorption at 1152 cm-1 and 1034, which are indicative of glycogen, glycolipids and glycoproteins (37,40,58,64). The fish heart has the ability to switch the predominant metabolic pathway to glycolysis under hypoxic conditions, providing a more oxygen efficient way of producing ATP (65,66). However, glycolysis is less efficient than FAO and, therefore, if ATP production by glycolysis is relied on for prolonged periods it can lead to deficiencies in myocardial ATP. Glycogen accumulation in the present study may indicate an increase in reliance on anaerobic glycolysis following warm acclimation, which supports previous data suggesting altered mitochondrial function and increased dependence on anaerobic pathways in temperate wrasse (Notolabrus celidotus) and cod (Gadus morhua) following warm acclimation (67,68). Again, data to assess the metabolic enzyme activity in the rainbow trout ventricle following thermal acclimation is needed to support this hypothesis. Switching to glycolytic pathways would likely be maladaptive and increases overall energy cost with warm temperature, suggesting species distribution may be related to the thermal limits of mitochondrial stability and oxidative function (69). The high metabolic rate of fish at high temperatures may lead to rapid depletion of myocardial ATP stores reducing cardiac function and, therefore, ability to tolerate high temperatures (2,3,5-7,63,70).

The metabolic profile shown by cold- and warm-acclimated tissue was not reflected in the control tissue. The FTIR absorbance profile of control tissue suggested increased tissue lipid, however, the histological result did not agree. Clearly, there are temperature independent effects altering tissue biochemistry, which likely occur in all groups but are particularly evident in the control group. Previous studies have suggested that the metabolic profile of tissue is not determined by temperature alone and that is likely that a number of additional environmental factors, such as photoperiod-dependent hormonal status and activity, play a role in nature of metabolic acclimatization in fish (60). During acclimation in this study, we reduced the photoperiod of the cold-acclimated fish to reflect reduced daylight hours in winter, however, control and warm animals remained on a 12 hr light, 12 hr dark cycle.

Changes in protein content of the myocardium following thermal acclimation

Cold acclimation of the fish heart is often associated with an increase in cardiac mass, generally thought to be due to an increased physiological requirement and increased blood

Metabolic and biochemical remodelling of the fish heart viscosity (11-13,23,24). The increase in cardiac mass occurs due to increase in size of constituent myocytes (26). As rainbow trout remain active throughout the year, the hypertrophic response of the heart is necessary to defend cardiac function during winter cold, providing additional support of the cardiac wall and additional contractile units to maintain cardiac function, similar to compensatory hypertrophy of the mammalian heart (15,71-74). In the cold-acclimated ventricle we saw increases in the symmetric stretch of protein O-H, increases in amide II and increase in phosphorylated proteins at amide III. However, there were reductions in protein absorptions that contribute to the amide B and amide I bands. All of these bands are, again, composed of many vibrational bands super-imposed to make the broad appearance of the band. It is possible to perform a deconvolution of these bands by converting the spectra to the second derivative. The deconvolution allows for some of the smaller peaks to be resolved, which can suggest differences in secondary confirmations of proteins (37,75). However, when the second derivative was performed on this data there was too much noise in the spectrum to resolve these peaks.

Collagen is a complex fibrillar molecule and, therefore, there are numerous spectral bands that have been assigned to collagen including 1206 cm-1, 1238 cm-1, 1280 cm-1 and 1338 cm- 1 (38,76). In the fish myocardium there are generally low levels of collagen, however, recent studies have suggested an increased fibrosis of the compact myocardium following prolonged cold exposure (9,10). Our hyperspectral image based on the peak at 1338 cm-1 showed increased intensity of infrared absorption in the cold-acclimated group compared to the control or warm acclimated groups. This result is consistent with previous studies which suggest a fibrosis of the compact myocardium following cold acclimation (9,10).

Changes in phosphate macromolecules following thermal acclimation

Following warm acclimation we saw decreases in the infrared absorption of bands at ~1086 cm-1 and 1240 cm-1 compared to cold, which are where the main symmetric and asymmetric phosphate group vibrations appear (77). These bands contain the vibrations of phosphate molecules in DNA, RNA, phospholipids and phosphorylated proteins (64,78-82), which makes it difficult to determine the contribution of each molecule. It is also possible that if ATP stores are depleted due to higher myocardial ATP demand than supply during glycolysis, the reduced ATP molecules would influence this spectral region. In addition (25) showed decreased rates of protein synthesis in rainbow trout following cold acclimation, which may be reflected by a reduction in infrared absorption of mRNA, shown by a decrease in these bands.

Perspectives and significance

Both acute and chronic temperature change have profound effects on the physiology of the fish heart. Here, we have used the novel technique of FTIR imaging spectroscopy to assess

Metabolic and biochemical remodelling of the fish heart the effects of chronic temperature changes that may be experienced seasonally by rainbow trout. As a cold-active species, it is important that these fish can maintain appropriate cardiac function all year around, which they achieve by a seasonal compensatory remodelling of the myocardium (10,12,14). With prolonged cold temperature there is an increase in cardiac muscle mass. Our results show an increase in ventricular lipid which may suggest that the necessary energy is supplied by an increase in FAO of the cardiac myocytes. As low temperature increased the oxygen carry capacity of water, there is sufficient oxygen supply to maintain aerobic energetics, even with high ATP demand. However, with prolonged warm temperature the oxygen carrying capacity of water is reduced while at the same time metabolic rate increases (8,63). To supply the body with sufficient oxygenated blood to maintain aerobic respiration, the cardiac output and, therefore, the work rate of the heart must increase (3,6,7). As fish rely primarily on venous return blood to oxygenate the cardiac muscle these conditions may supply insufficient oxygen to the heart to maintain aerobic FAO (46). Our results show an accumulation of ventricular which may suggest that cardiac myocytes have to switch their predominant energetic pathways to glycolysis to provide ATP. While this energetic state, typical of the foetal heart, increases efficiency of ATP production per mole of oxygen requires, it is insufficient to sustain high cardiac energy demand long-term (21,22). Thus, this switch in metabolic state highlights a maladaptive remodelling response of the heart to chronic warm temperatures. In the present study, warm-acclimated fish were kept at 18 °C, which is lower than the critical temperature of cod reported by Rodnick et al. (67). However, the clear response we report may go some way to explaining the reduced function and even heart failure shown by fish at high temperatures. In our era of climate change, this maladaptive switch in cellular energetics may determine species distribution and fish survival at prolonged high temperatures.

Acknowledgements

Conflicts of interest

The authors declare no conflicts of interest.

REFERENCES

1. Brett, J. R. (1971) Energetic responses of salmon to temperature. A study of some thermal relations in physiology and freshwater ecology of sockeye salmon (Oncorhynchus nerka). American Zoologist 11, 99-113

Metabolic and biochemical remodelling of the fish heart

2. Portner, H. O., and Knust, R. (2007) Climate change affects marine fishes through the oxygen limitation of thermal tolerance. Science 315, 95-97

3. Keen, A. N., and Gamperl, A. K. (2012) Blood oxygenation and cardiorespiratory function in steelhead trout (Oncorhynchus mykiss) challenged with an acute temperature increase and zatebradine-induced bradycardia. Journal of Thermal Biology 37, 201-210

4. Farrell, A. P., Gamperl, A. K., Hicks, J. M. T., Shiels, H. A., and Jain, K. E. (1996) Maximum cardiac performance of rainbow trout (Oncorhynchus mykiss) at temperatures approaching their upper lethal limit. Journal of Experimental Biology 199, 663-672

5. Wang, T., and Overgaard, J. (2007) The heartbreak of adapting to global warming. Science 315, 49-50

6. Eliason, E. J., and Farrell, A. P. (2016) Oxygen uptake in Pacific salmon Oncorhynchus spp.: when ecology and physiology meet. Journal of Fish Biology 88, 359-388

7. Steinhausen, M. F., Sandblom, E., Eliason, E. J., Verhille, C., and Farrell, A. P. (2008) The effect of acute temperature increases on the cardiorespiratory performance of resting and swimming sockeye salmon (Oncorhynchus nerka). Journal of Experimental Biology 211, 3915-3926

8. Farrell, A. P., Eliason, E. J., Sandblom, E., and Clark, T. D. (2009) Fish cardiorespiratory physiology in an era of climate change. Canadian Journal of Zoology-Revue Canadienne De Zoologie 87, 835-851

9. Keen, A. N., Fenna, A. J., McConnell, J. C., Sherratt, M. J., Gardner, P., and Shiels, H. A. (2016) The dynamic nature of hypertrophic and fibrotic remodelling in the fish ventricle. Frontiers in Physiology - Intergrative Physiology 6

10. Klaiman, J. M., Fenna, A. J., Shiels, H. A., Macri, J., and Gillis, T. E. (2011) Cardiac Remodeling in Fish: Strategies to Maintain Heart Function during Temperature Change. PLoS ONE 6, e24464

11. Graham, M. S., and Farrell, A. P. (1989) The effect of temperature acclimation and adrenaline on the performance of a perfused trout heart. Physiological Zoology 62, 38-61

12. Driedzic, W. R., Bailey, J. R., and Sephton, D. H. (1996) Cardiac adaptations to low temperature non-polar teleost fish. Journal of Experimental Zoology 275

13. Driedzic, W. R., and Gesser, H. (1994) Energy Metabolism and Contractility in Ectothermic Vertebrate Hearts: Hypoxia, Acidosis, and Low Temperature. Physiological Reviews 74, 221-258

14. Farrell, A. P., Hammons, A. M., Graham, M. S., and Tibbits, G. F. (1988) Cardiac growth in rainbow trout, Salmo gairdneri. Canadian Journal of Zoology-Revue Canadienne De Zoologie 66, 2368-2373

15. Klaiman, J. M., Pyle, W. G., and Gillis, T. E. (2014) Cold acclimation increases cardiac myofilament function and ventricular pressure generation in trout. Journal of Experimental Biology 217, 4132-4140

16. Allard, M. F., Schonekess, B. O., Henning, S. L., English, D. R., and Lopaschuk, G. D. (1994) Contribution of oxidative metabolism and glycolysis to ATP production in hypertrophied hearts. American Journal of Physiology 267

Metabolic and biochemical remodelling of the fish heart

17. Ingwall, J. S. (2009) Energy metabolism in heart failure and remodelling. Cardiovascular Research 81

18. Kolwicz, S. C., Olson, D. P., Marney, L. C., Garcia-Menendez, L., Synovec, R. E., and Tian, R. (2012) Cardiac-Specific Deletion of Acetyl CoA Carboxylase 2 Prevents Metabolic Remodeling During Pressure-Overload Hypertrophy. Circulation Research 111, 728-+

19. Akki, A., Smith, K., and Seymour, A.-M. L. (2008) Compensated cardiac hypertrophy is characterised by a decline in palmitate oxidation. Molecular and Cellular Biochemistry 311

20. Pound, K. M., Sorokina, N., Ballal, K., Berkich, D. A., Fasano, M., LaNoue, K. F., Taegtmeyer, H., O'Donnell, J. M., and Lewandowski, E. D. (2009) Substrate-Enzyme Competition Attenuates Upregulated Anaplerotic Flux Through Malic Enzyme in Hypertrophied Rat Heart and Restores Triacylglyceride Content Attenuating Upregulated Anaplerosis in Hypertrophy. Circulation Research 104

21. Ingwall, J. S., and Weiss, R. G. (2004) Is the failing heart energy starved? On using chemical energy to support cardiac function. Circulation Research 95

22. Neubauer, S. (2007) Mechanisms of disease. The failing heart - An engine out of fuel. New England Journal of Medicine 356

23. Sephton, D. H., and Driedzic, W. R. (1991) Effect of acute and chronic temperature transition on enzymes of cardiac metabolism in white perch (Morone americana), yellow perch (Perca flavescens), and smallmouth bass (Micropterus dolomieui). Canadian Journal of Zoology 69, 258-262

24. Bailey, J. R., and Driedzic, W. R. (1993) Influence of low temperature acclimation on fate of metabolic fuels in rainbow trout (Oncorhynchus mykiss) hearts. Canadian Journal of Zoology 71, 2167-2173

25. Sephton, D. H., and Driedzic, W. R. (1995) Low temperature acclimation decreases rates of protein synthesis in rainbow trout (Oncorhynchus mykiss) heart. Fish Physiology and Biochemistry 14, 63-69

26. Vornanen, M., Hassinen, M., Koskinen, H., and Krasnov, A. (2005) Steady-state effects of temperature acclimation on the transcriptome of the rainbow trout heart. American Journal of Physiology-Regulatory Integrative and Comparative Physiology 289, R1177-1184

27. Stuart, B. (1997) Biological Applications of Infrared Spectroscopy, John Wiley & Sons, Chichester

28. Ellis, D. I., and Goodacre, R. (2006) Metabolic fingerprinting in disease diagnosis: biomedical applications of infrared and Raman spectroscopy. Analyst 131

29. Baker, M. J., Trevisan, J., Bassan, P., Bhargava, R., Butler, H. J., Dorling, K. M., Fielden, P. R., Fogarty, S. W., Fullwood, N. J., Heys, K. A., Hughes, C., Lasch, P., Martin-Hirsch, P. L., Obinaju, B., Sockalingum, G. D., Sulé-Suso, J., Strong, R. J., Walsh, M. J., Wood, B. R., Gardner, P., and Martin, F. L. (2014) Using Fourier transform IR spectroscopy to analyze biological materials. Nat. Protocols 9, 1771- 1791

30. Gazi, E., Baker, M., Dwyer, J., Lockyer, N. P., Gardner, P., Shanks, J. H., Reeve, R. S., Hart, C. A., Clarke, N. W., and Brown, M. D. (2006) A correlation of FTIR spectra derived from prostate cancer biopsies with Gleason grade and tumour stage. European Urology 50

Metabolic and biochemical remodelling of the fish heart

31. Gazi, E., Gardner, P., Hart, C., Shanks, J., Baker, N., Dwyer, J., Lockyer, N., Miyan, J., Brown, M., and Clarke, N. (2005) Diagnosing prostate tumour biopsies using infrared microspectroscopy. European Urology Supplements 4

32. Harvey, T. J., Henderson, A., Gazi, E., Clarke, N. W., Brown, M., Faria, E. C., Snook, R. D., and Gardner, P. (2007) Discrimination of prostate cancer cells by reflection mode FTIR photoacoustic spectroscopy. Analyst 132

33. Baker, M. J., Gazi, E., Brown, M. D., Shanks, J. H., Gardner, P., and Clarke, N. W. (2008) FTIR-based spectroscopic analysis in the identification of clinically aggressive prostate cancer. British Journal of Cancer 99

34. Baker, M. J., Gazi, E., Brown, M. D., Shanks, J. H., Clarke, N. W., and Gardner, P. (2009) Investigating FTIR based histopathology for the diagnosis of prostate cancer. Journal of Biophotonics 2

35. Hughes, C., Gaunt, L., Brown, M., Clarke, N. W., and Gardner, P. (2014) Assessment of paraffin removal from prostate FFPE sections using transmission mode FTIR-FPA imaging. Analytical Methods 6, 1028-1035

36. Kuwahara, M., Bannai, K., Segawa, H., Miyamoto, K.-i., and Yamato, H. (2014) Cardiac remodeling associated with protein increase and lipid accumulation in early- stage chronic kidney disease in rats. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease 1842, 1433-1443

37. Staniszewska, E., Malek, K., and Baranska, M. (2014) Rapid approach to analyze biochemical variation in rat organs by ATR FTIR spectroscopy. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 118, 981-986

38. Cheheltani, R., Rosano, J. M., Wang, B., Sabri, A. K., Pleshko, N., and Kiani, M. F. (2012) Fourier transform infrared spectroscopic imaging of cardiac tissue to detect collagen deposition after myocardial infarction. Journal of Biomedical Optics 17, 056014

39. Zheng, N., Yang, T., Liang, M., Zhang, H., Li, L., Sunnassee, A., and Liu, L. (2010) Characterization of protein in old myocardial infarction by FTIR micro-spectroscopy. Journal of Huazhong University of Science and Technology [Medical Sciences] 30, 546-550

40. Toyran, N., Turan, B., and Severcan, F. (2007) Selenium alters the lipid content and protein profile of rat heart: An FTIR microspectroscopic study. Archives of Biochemistry and Biophysics 458, 184-193

41. Gough, K. M., Zelinski, D., Wiens, R., Rak, M., and Dixon, I. M. C. (2003) Fourier transform infrared evaluation of microscopic scarring in the cardiomyopathic heart: Effect of chronic AT1 suppression. Analytical Biochemistry 316, 232-242

42. Schneider, C. A., Rasband, W. S., and Eliceiri, K. W. (2012) NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9, 671-675

43. Reddy, R. K., and Bhargava, R. (2010) Accurate histopathology from low signal-to- noise ratio spectroscopic imaging data. Analyst 135, 2818-2825

44. Bhargava, R. (2012) Infrared Spectroscopic Imaging: The Next Generation. Applied Spectroscopy 66, 1091-1120

45. Bassan, P., Sachdeva, A., Kohler, A., Hughes, C., Henderson, A., Boyle, J., Shanks, J. H., Brown, M., Clarke, N. W., and Gardner, P. (2012) FTIR microscopy of biological cells and tissue: data analysis using resonant Mie scattering (RMieS) EMSC algorithm. Analyst 137, 1370-1377 42

Metabolic and biochemical remodelling of the fish heart

46. Farrell, A. P., and Jones, D. R. (1992) The Heart. In Fish Physiology, Volume XII, Part A The Cardiovascular System,

47. Pieperhoff, S., Bennett, W., and Farrell, A. P. (2009) The intercellular organization of the two muscular systems in the adult salmonid heart, the compact and the spongy myocardium. Journal of Anatomy 215, 536-547

48. Mehlem, A., Hagberg, C. E., Muhl, L., Eriksson, U., and Falkevall, A. (2013) Imaging of neutral lipids by oil red O for analyzing the metabolic status in health and disease. Nature Protocols 8, 1149-1154

49. Lyng, F., Gazi, E., and Gardner, P. (2011) Preparation of Tissues and Cells for Infrared and Raman Spectroscopy and Imaging,

50. Faolain, E. O., M. B. Hunter, J. M. Byrne, P. Kelehan, M. McNamara, H. J. Byrne and F. M. Lyng. (2005) A study examining the effects of tissue processing on human tissue sections using vibrational spectroscopy. Vibrational Spectroscopy 38, 121– 127

51. Fernandez, D. C., Bhargava, R., Hewitt, S. M., and Levin, I. W.,. (2005) Infrared spectroscopic imaging for histopathological recognition. Nature Biotechnology 23, 469–474

52. Sahu, R. K., Argov, S., Salman, A., Zelig, U., Huleihel, M., Grossman, N., Gopas, J., Kapelushnik, J., and S. Mordechai, S. (2005) Can Fourier transform infrared spectroscopy at higher wavenumbers (mid IR) shed light on biomarkers? Journal of Biomedical Optometry 10, 05017–05027

53. Shim, M. G., and Wilson, B. C. (1996) The effects of ex vivo handling procedures on the near-infrared Raman spectra of normal mammalian tissues. Photochemistry and Photobiology 63, 662–671

54. Stitt, D. M., Kastyak-Ibrahim, M. Z., Liao, C. R., Morrison, J., Albensi, B. C., and Gough, K. M. (2012) Tissue acquisition and storage associated oxidation considerations for FTIR microspectroscopic imaging of polyunsaturated fatty acids. Vibrational Spectroscopy 60, 16-22

55. Bassan, P. (2011) Light scattering during infrared spectroscopic measurements of biomedical samples. PhD, University of Manchester

56. Xin, M., Olson, E. N., and Bassel-Duby, R. (2013) Mending broken hearts: cardiac development as a basis for adult heart regeneration and repair. Nat Rev Mol Cell Biol 14, 529-541

57. Holton, S. E., Walsh, M. J., and Bhargava, R. (2011) Subcellular localization of early biochemical transformations in cancer-activated fibroblasts using infrared spectroscopic imaging. Analyst 136, 2953-2958

58. Wood, B. R., Quinn, M. A., Tait, B., Ashdown, M., Hislop, T., Romeo, M., and McNaughton, D. (1998) FTIR microspectroscopic study of cell types and potential confounding variables in screening for cervical malignancies. Biospectroscopy 4, 75- 91

59. Snyder, R. G. (1979) Vibrational correlation splitting and chain packing for the crystalline n‐alkanes. The Journal of Chemical Physics 71, 3229-3235

60. Cordiner, S., and Egginton, S. (1997) Effects of seasonal temperature acclimatization on muscle metabolism in rainbow trout, Oncorhynchus mykiss. Fish Physiology and Biochemistry 16, 333-343

Metabolic and biochemical remodelling of the fish heart

61. Kleckner, N. W., and Sidell, B. D. (1985) Comparison of maximal activities of enzymes from tissues of thermally acclimated and naturally acclimatized chain pickerel (Esox niger). Physiological Zoology 58, 18-28

62. Swimmer, Y., McNaughton, L., Moyes, C., and Brill, R. (2005) Metabolic biochemistry of cardiac muscle in three tuna species (bigeye, Thunnus obesus; yellowfin, T. albacares; and skipjack, Katsuwonus pelamis) with divergent ambient temperature and oxygen tolerances. Fish Physiology and Biochemistry 30, 27-35

63. Farrell, A. P. (2009) Environment, antecedents and climate change: lessons from the study of temperature physiology and river migration of salmonids. Journal of Experimental Biology 212, 3771-3780

64. Saravanakumar, M., Manivannan, J., Sivasubramanian, J., Silambarasan, T., Balamurugan, E., and Raja, B. (2011) Molecular metabolic fingerprinting approach to investigate the effects of borneol on metabolic alterations in the liver of nitric oxide deficient hypertensive rats. Molecular and Cellular Biochemistry 362, 203-209

65. Overgaard, J., Stecyk, J. A. W., Gesser, H., Wang, T., and Farrell, A. P. (2004) Effects of temperature and anoxia upon the performance of in situ perfused trout hearts. Journal of Experimental Biology 207, 655-665

66. Becker, T. A., DellaValle, B., Gesser, H., and Rodnick, K. J. (2013) Limited effects of exogenous glucose during severe hypoxia and a lack of hypoxia-stimulated glucose uptake in isolated rainbow trout cardiac muscle. Journal of Experimental Biology 216, 3422-3432

67. Rodnick, K. J., Gamperl, A. K., Nash, G. W., and Syme, D. A. (2014) Temperature and sex dependent effects on cardiac mitochondrial metabolism in Atlantic cod (Gadus morhua L.). Journal of Thermal Biology 44, 110-118

68. Iftikar, F. I., Morash, A. J., Cook, D. G., Herbert, N. A., and Hickey, A. J. R. (2015) Temperature acclimation of mitochondria function from the hearts of a temperate wrasse (Notolabrus celidotus). Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology 184, 46-55

69. Iftikar, F. I., MacDonald, J. R., Baker, D. W., Renshaw, G. M. C., and Hickey, A. J. R. (2014) Could thermal sensitivity of mitochondria determine species distribution in a changing climate? Journal of Experimental Biology 217, 2348-2357

70. Clark, T. D., Sandblom, E., Cox, G. K., Hinch, S. G., and Farrell, A. P. (2008) Circulatory limits to oxygen supply during an acute temperature increase in the Chinook salmon (Oncorhynchus tshawytscha). American Journal of Physiology- Regulatory Integrative and Comparative Physiology 295, R1631-R1639

71. Mone, S. M., Sanders, S. P., and Colan, S. D. (1996) Control mechanisms for physiological hypertrophy of pregnancy. Circulation 94, 667-672

72. Bishop, S. P., Allard, M. F., Anderson, P. G., Bugaisky, L. B., Doss, L. K., and Sreenan, C. M. (1990) Myocardial hypertrophy, fibrosis and failure due to progressive pressure-overload in the rat. Circulation 82, 300

73. Bernardo, B. C., Weeks, K. L., Pretorius, L., and McMullen, J. R. (2010) Molecular distinction between physiological and pathological cardiac hypertrophy: Experimental findings and therapeutic strategies. Pharmacology & Therapeutics 128, 191-227

74. Frey, N., and Olson, E. N. (2003) Cardiac hypertrophy: The good, the bad and the ugly. Annual Review of Physiology 65, 45-79

Metabolic and biochemical remodelling of the fish heart

75. Kretlow, A., Wang, Q., Kneipp, J., Lasch, P., Beekes, M., Miller, L., and Naumann, D. (2006) FTIR-microspectroscopy of prion-infected nervous tissue. Biochim Biophys Acta 1758, 948-959

76. Wang, Q., Sanad, W., Miller, L. M., Voigt, A., Klingel, K., Kandolf, R., Stangl, K., and Baumann, G. (2005) Infrared imaging of compositional changes in inflammatory cardiomyopathy. Vibrational Spectroscopy 38, 217-222

77. Whelan, D. R., Bambery, K. R., Heraud, P., Tobin, M. J., Diem, M., McNaughton, D., and Wood, B. R. (2011) Monitoring the reversible B to A-like transition of DNA in eukaryotic cells using Fourier transform infrared spectroscopy. Nucleic Acids Research 39, 5439-5448

78. Krafft, C., Kirsch, M., Beleites, C., Schackert, G., and Salzer, R. (2007) Methodology for fiber-optic Raman mapping and FTIR imaging of metastases in mouse brains. Analytical and Bioanalytical Chemistry 389, 1133-1142

79. Krafft, C., Codrich, D., Pelizzob, G., and Sergoa, V. (2009) Raman and FTIR imaging of lung tissue: bronchopulmonary sequestration. Journal of Raman Spectroscopy, 595–603

80. Palombo, F., Shen, H., Benguigui, L. E. S., Kazarian, S. G., and Upmacis, R. K. (2009) Micro ATR-FTIRspectroscopic imaging of atherosclerosis: an investigation of the contribution of inducible nitric oxide synthase to lesion composition in ApoE-null mice. Analyst 134, 1107-1118

81. Severcan, F., Gorgulu, G., Gorgulu, S. T., and Guray, T. (2005) Rapid monitoring of diabetes-induced lipid peroxidation by Fourier transform infrared spectroscopy: Evidence from rat liver microsomal membranes. Analytical Biochemistry 339, 36-40

82. Hackett, M. J., McQuillan, J. A., El-Assaad, F., Aitken, J. B., Levina, A., Cohen, D. D., Siegele, R., Carter, E. A., Grau, G. E., Hunt, N. H., and Lay, P. A. (2011) Chemical alterations to murine brain tissue induced by formalin fixation: implications for biospectroscopic imaging and mapping studies of disease pathogenesis. Analyst 136, 2941-2952

Metabolic and biochemical remodelling of the fish heart

Supplementary figure 1. Mean spectra for each individual fish in (A) the cold-acclimated (CA), (B) the control and (C) the warm-acclimated group (WA).

Metabolic and biochemical remodelling of the fish heart

Supplementary figure 2. The mean spectral profile of (A) control fish 1, (B) control fish 2 after the first area scanned (black) and the second are scanned (red) and (C) a MatrigelTM spectrum.

Supplementary figure 3. The mean spectra for formalin-fixed, paraffin embedded tissue for each individual fish. (A) cold-acclimated compact myocardium, (B) cold-acclimated spongy myocardium, (C) control compact myocardium, (D) control spongy myocardium, (E) warm- acclimated compact myocardium, (F) warm-acclimated spongy myocardium.

Thermal remodelling of the ectothermic heart

5. Macro- and micro-mechanical remodelling in the fish atrium is associated with regulation of collagen 1 alpha 3 chain expression.

This chapter is presented in format of the manuscript that has been submitted to the

Journal of Physiology.

Page | 83

Thermal remodelling of the ectothermic heart

Page | 84

Thermal remodeling of the fish atrium

Title: Macro- and micro-mechanical remodelling in the fish atrium is associated with regulation of collagen 1 alpha 3 chain expression.

Adam N. Keen1, Andrew J. Fenna1, James C. McConnell2, Michael J. Sherratt2, Peter Gardner3, Holly A. Shiels1*

1Faculty of Life Sciences, University of Manchester, Manchester, UK 2Centre for Tissue Injury & Repair, Faculty of Medical and Human Sciences, University of Manchester, UK 3School of Chemical Engineering and Analytical Science, Manchester Institute of Biotechnology, University of Manchester, UK

*Author of correspondence: [email protected] +44 161 275 5092

Short title: Thermal remodelling of the fish atrium

Keywords: Compliance, Heart, Stiffness, Temperature Acclimation, Phenotypic Plasticity

Text: 9,545 Figures: 5 Tables: 2 References: 67

List of Abbreviations: AF, atrial fibrillation; AFM, atomic force microscopy; ANP, atrial natriuretic peptide; BNP, brain natriuretic peptide; Col1a1, collagen I alpha 1; Col1a2, collagen I alpha 2; Col1a3, collagen I alpha 3; DAPI, 4’, 6’-diamidino-2-phenylindole; EBS, extra-bundular sinus, ECM, extracellular matrix; GLM, general linear model; H & E, haematoxylin and eosin; MLP, muscle LIM protein; MMP2, matrix metalloproteinase 2; MMP9, matrix metalloproteinase 9; MMP13, matrix metalloproteinase 13; NFAT, nuclear factor of activating T; PNCA, proliferating cell nuclear antigen; RCAN1, regulator of calcineurin 1; RT-qPCR, real-time quantitative PCR; SMLC2, small myosin light chain 2; TIMP2, tissue inhibitor of metalloproteinase 2; VEGF, vascular endothelial growth factor; VMHC, ventricular myosin heavy chain

Thermal remodeling of the fish atrium

Abstract

Numerous pathological states lead to remodelling of the mammalian heart, often associated with fibrosis. Recent work in fish has shown that fibrotic connective tissue remodelling of the ventricle is ‘reversible’. The atrial response to varying cardiac load is less understood in non- mammalian vertebrates and has not been investigated in fish. Here we have used prolonged temperature exposure (>8 weeks at either 10 °C (control), cold (5 °C) or warm (18 °C)) to initiate the cardiac remodelling response in the rainbow trout heart. We found that cold acclimation increased passive stiffness of the whole atrium and increased micromechanical stiffness of tissue sections. The opposite connective tissue remodelling response was seen with chronic warming; the atrium became more compliant. Collagen deposition contributed to the increased tissue stiffness in chronic cold and was associated with an up-regulation of collagen promoting genes and changes in gelatinase activity of collagen degrading matrix metalloproteinase (MMPs). Again, we saw the opposite response in the atrium following chronic warming. Cooling also reduced the mRNA expression of cardiac growth factors and hypertrophic markers, which were increased following chronic warming. Together, these findings suggest that chronic cooling of trout causes atrial dilation and increased myocardial stiffness analogous to pathological states in mammalian atria. The increased compliance and changes in tissue morphology following chronic warming are particularly interesting as they suggest that typically pathological remodelling in mammalian atria may oscillate seasonally in the fish atrium, revealing a more dynamic and plastic remodelling response.

Thermal remodeling of the fish atrium

Introduction Chronic changes in pressure or volume load can cause the vertebrate heart to change in size, form and function (Opie et al., 2006; Kumar et al., 2012). This cardiac remodelling response is often compensatory and maintains optimal cardiac function under conditions of increased haemodynamic preload or afterload. Studies in mammals show chronic pressure or volume overload is associated with remodelling of both the atria and ventricle, however, atrial remodelling is less well understood (Sanfilippo et al., 1990; Pellman et al., 2010; Kumar et al., 2012). Chronic atrial dilation (or atrial enlargement) is a form of cardiomegaly, which may be associated with cellular hypertrophy, myocardial fibrosis, angiogenesis, apoptosis and myolysis (Bauer et al., 2004; Polovkova et al., 2013). The increase in non-conducting extracellular matrix (ECM) can delay and/or impair electrical conduction between cardiomyocytes and may initiate alternate conduction pathways (Schotten et al., 2004; Pellman et al., 2010). This can contribute to atrial arrhythmias, atrial fibrillation (AF) and loss of contractility (Nattel et al., 2008). Symptoms can be self-perpetuating with atrial fibrillation contributing to atrial dilation and the occurrence and maintenance of fibrosis (Sanfilippo et al., 1990; Schotten et al., 2003; Knackstedt et al., 2008). Enlargement of the atria may also occur with increased load on the heart that occurs with increased physiological demand, such as with exercise training (Hauser et al., 1985; Pelliccia et al., 2005; D'Andrea et al., 2010). However, this kind of ‘physiological’ remodelling gives a compensatory increase in myocardial wall thickness, which regresses when the stimulus is removed; it is not associated with arrhythmias or fibrosis as seen with ‘pathological’ remodelling conditions (Pelliccia et al., 2005).

In mammals, the role of the atria in ventricular filling is relatively small with atrial systole providing ~20-30 % of end-diastolic volume (depending on age) with the remainder modulated directly by venous pressure (Guyton, 1981). The fish heart has just one atrium and one ventricle (Farrell & Jones, 1992). The atrium is the largest chamber of the fish heart with thin but highly trabeculated walls which form a web-like structure and aid contractions by pulling the walls and roof inwards (Farrell & Jones, 1992; Forster & Farrell, 1994). The fish atrium directly modulates ventricular stroke volume by acting as a volume reservoir for end-diastolic volume (Forster & Farrell, 1994). Although debated (Lai et al.,1998; Aho & Vornanen, 1999), it has been suggested that atrial systole is the primary mechanism for ventricular filling and, therefore, crucial for determining filling volume, strength of contraction and total stroke volume (Johansen & Burggren, 1980; Farrell & Jones, 1992).

The rainbow trout remains active throughout the year despite the direct effects of temperature on contractile force and on blood viscosity, which alters cardiac load (Driedzic et al., 1996; Keen et al., 2016). Many studies have investigated the ventricular remodelling response, which maintains optimum cardiac output during seasonal temperature change (Graham & Farrell, 1989; Klaiman et al., 2011; Klaiman et al., 2014; Farrell et al., 1988a; Gamperl & Farrell, 2004).

Thermal remodeling of the fish atrium

However, very few studies have investigated the role of the atrium in thermal remodelling in fish. Proteins involved in excitation-contraction coupling like SERCA, phospholamban, Ca2+ binding proteins and the NCX-exchanger show similar changes in gene expression following cold acclimation in the atrium compared with the ventricle (Korajoki & Vornanen, 2009, 2012; Jayasundara et al., 2013). Furthermore, certain proteins such, as FK506-binding protein which regulates sarcoplasmic reticulum Ca2+ release, are only up-regulated in the atrium and remain unchanged in the ventricle following chronic cooling (Korajoki & Vornanen, 2014). Finally, there is evidence of partial positive thermal compensation, as atrial contraction kinetics improves following cold acclimation (Aho & Vornanen, 1999). Whether atrial fibrosis occurs in fish following chronic cooling is currently unknown; however, significant ventricular fibrosis has been shown with cold acclimation (Klaiman et al., 2011; Keen et al., 2016). We know of no studies to directly investigate atrial remodelling following both warming and cooling in any fish.

Here we investigate the effects of chronic cooling (from 10 ± 1 °C to 5 ± 1 °C) and chronic warming (from 10 ± 1 °C to 18 ± 1 °C) on the rainbow trout atrium. These temperatures reflect those trout experience seasonally. We focus our study on changes in the passive and mechanical properties of the atrium, across multiple levels of organization. Based on our recent work on the trout ventricle (Keen et al., 2016), we hypothesized that chronic cooling would increase atrial stiffness, fibrosis and up-regulate growth factors associated with pathological remodelling in mammals. We further hypothesize that the opposite would occur following chronic warming. To understand the functional consequences of thermal acclimation we used atomic force microscopy (AFM) to test micromechanical atrial stiffness, generated ex vivo atrial pressure-volume curves to test whole chamber compliance, and used in situ zymography to assess gelatinase activity of matrix metalloproteinases (MMPs). To determine structural remodelling of the tissue we used histological stains to assess ECM proteins. We then used quantitative real-time PCR (RT-qPCR) to examine growth factors, collagen isoform expression, connective tissue regulators and hypertrophic markers following prolonged temperature exposures. As trout experience intermittent temperature change, we were particularly interested in variable remodelling between chronic warming and chronic cooling. Our findings suggest that chronic cooling increases cardiac preload, which causes chronic dilation of the atrium leading to a functional change in diastolic function. Overall, we found the opposite remodelling response following chronic warming, suggesting that atrial remodeling is dynamic and may be reversible in the fish heart.

Materials and Methods Ethical Approval All husbandry and housing conditions were in accordance with the local handling protocols and adhere to the UK Home Office legislation. All experimental procedures were approved by the University of Manchester’s ethical review committee.

Thermal remodeling of the fish atrium

Experimental animals Sexually mature female rainbow trout (Onchorynchus mykiss; n = 47; morphometric data in Table 1) were purchased from Dunsop Bridge Trout Farm (Clitheroe, UK), housed on a 12 hr light: 12 hr dark cycle in ~500 L re-circulated aerated fresh water tanks at 10 ± 1 °C and fed to satiation 3 times per week. Water quality was ensured with 30 % water changes 3 times per week and regular tests for temperature, pH, nitrates and nitrites. Fish were held under these conditions for a minimum of 2 weeks before being randomly assigned to one of three acclimation groups; cold (5 ± 1 °C), control (i.e. no change; 10 ± 1 °C) or warm (18 ± 1 °C). These temperatures were based on previous literature which describes the cardiac remodelling response in salmonids (Klaiman et al., 2011; Keen et al., 2016). Water temperature of the warm and cold acclimation groups was changed by 1 °C per day until desired temperature was reached and then held at that temperature for a minimum of 8 weeks before experiments. The photoperiod for the cold-acclimated animals was changed to 8 hr light: 16 hr dark cycle to simulate winter (Graham & Farrell, 1989).

Before experiments, fish were killed by a blow to the head followed by severance of the spinal cord and destruction of the brain. The heart was excised, rinsed in phosphate buffered saline and weighed. Atria were used immediately for the ex vivo pressure-volume curves. Atria to be used for RT-qPCR were snap frozen and stored at -80 °C. Atria to be used for histological analysis and in situ zymography were bisected down the sagittal plane with one half snap frozen in OCT (Thermo Fisher Scientific, Waltham, MA, USA) by immersion in liquid nitrogen cooled 2-Methylbutane (Sigma-Aldrich, St. Louis, MO, USA) and stored at -80 °C. The other half was fixed in 10 % neutral buffered formalin solution (Sigma-Aldrich, St. Louis, MO, USA) before being processed and embedded in paraffin wax.

Ex vivo passive pressure-volume curves Whole chamber compliance was tested by generating ex vivo pressure-volume curves. The intact isolated heart was placed into an organ bath containing Ringer’s solution [(in mM) 150

NaCl, 5.4 KCl, 2.0 CaCl2, 1.5 MgSO4, 0.4 NaH2PO4, 10 HEPES, 10 Glucose at a pH of pH 7.7 with NaOH at room temperature] at 10 ± 1 °C to which 20 mM BDM (2, 3 butanedione monoxime) was added to prevent active cross-bridge cycling. Pressure-volume curves from atria from each acclimation group were generated at a common temperature, of 10 ± 1 °C, to isolate the effects of chronic remodelling on myocardial stiffness from the acute effects of temperature. A cannula was fed through the sinus venosus into the atrial lumen and secured at the sino-atrial junction, using 0-0 silk thread (Harvard Apparatus, Holliston, MA, USA). An atraumatic clamp was placed at the atrial-ventricular junction making the atrium a sealed chamber with the cannula inside. The cannula was connected to a syringe pump (INFORS AG, Bottmingen, CHE), in series with a pressure transducer, containing 10 ± 1 °C Ringer’s solution with BDM and a small amount of blue food colouring (Silverspoon, London, UK). The pressure transducer was calibrated daily against a static water column and measurements

Thermal remodeling of the fish atrium

recorded at 1000 Hz (Chart5, PowerLab, ADI Instruments, Dunedin, New Zealand). Ringer’s solution with BDM was pumped into the atrium at 0.05 ml min-1 until maximum volume was achieved, determined by visual leak of the saline containing blue dye and a drop in the pressure trace.

Atomic force microscopy (AFM) Atrial tissue micromechanics were tested using atomic force microscopy (AFM). Frozen atrial tissue was sectioned at 5 μm (Leica CM3050S cryostat, Leica, Wetzlar, Germany) and mounted onto microscope slides. Excess OCT was removed with distilled water and the slides were left to dry for ~12 hrs. This methodology is consistent with previous work (Kemp et al., 2012; Wallace et al., 2012), which shows that tissue sections are best preserved dehydrated with rehydration performed when nanomechanical measurements are required. Micro- indentation was performed using a Bioscope Catalyst AFM (Bruker, Coventry, UK) mounted onto an Eclipse T1 inverted optical microscope (Nikon, Kingston, UK) fitted with a spherically tipped cantilever (nominal radius and spring constant of 1 μm and 3 Nm-1 respectively; Windsor Scientific Ltd., Slough, UK) running Nanoscope Software v8.15 (Bruker, Coventry, UK). The local reduced modulus was determined for each of 400 points in a 50 x 50 μm region, indented at a frequency of 1 Hz with lateral spacing of 2.5 μm. The extend curve was used in conjunction with a contact point based model to calculate the reduced modulus for each indentation (Crick & Yin, 2007). For each biological sample, 400 force curves were collected at three distinct 50 μm2 regions. Once all 400 force curves had been generated, a quality control was applied where any force values falling more than two standard deviations away from the mean value were discarded in order to account for failed indents. Data loss at this stage was less than 10 % (data not shown).

Tissue Histology Tissue morphology was assessed histologically. Formalin-fixed and paraffin embedded atrial tissue was sectioned at 5 μm using a microtome (Leica RM2255, Leica, Wetzlar, Germany), mounted onto glass slides (Super frost plus, Thermo Fisher Scientific, Waltham, MA, USA) and stained using haematoxylin and eosin (H & E). Previous studies have used myocyte bundle cross-sectional area as a proxy for myocyte cross-sectional area as single fish myocytes are too narrow (diameter of 3-6 μm) to resolve in cross-section with this method (Klaiman et al., 2011; Keen et al., 2016). Cross-sectional bundle area and extra-bundular sinus (EBS) space were quantified using ImageJ software (Schneider et al., 2012). For morphometric analysis of myocyte bundle cross-sectional area and EBS, 8 sections were analyzed per individual fish. For measurement of cross-sectional area of myocyte bundles, three separate image montages were taken along transects across the full diameter of the cross-section on each tissue section. In each image trabeculations were chosen for measurement only if they were in the transverse plane, i.e. the image showed a cross-section

Thermal remodeling of the fish atrium

of the trabeculations making it circular in appearance. For EBS the non-tissue area of each image was measured.

Fibrillar collagen and elastin content were analysed semi-quantitatively following previously published methodology (Graham et al., 2011; Keen et al., 2016). Briefly, formalin-fixed paraffin embedded atrial tissue was sectioned at 5 μm (Leica RM2255 microtome, Leica, Wetzlar, Germany) and mounted onto glass slides. Serial sections from each sample were stained with picro-sirus red for collagen (Junqueira et al., 1979) and Miller’s elastic stain for elastin (Miller, 1971). Picro-sirus red images were quantified using polarised light microscopy and Miller’s elastic images were quantified using bright-field microscopy. Mean fibrillar collagen content was expressed as a percentage of total tissue cross sectional area, excluding the epicardial surface, determined using ImageJ. Three tissue sections were considered for each individual to ensure consistency in measurements. On each tissue section 3 separate image montages were taken along transects across the full diameter of the cross section. All histological analysis was conducted blind to the acclimation group.

In situ MMP gelatin zymography The activity of endogenous MMP gelatinase was semi-quantitatively analysed by in situ zymography of tissue cryosections, following previously published methodology (Mook et al., 2003; Akhtar et al., 2014). Frozen tissue was sectioned at 10 μm (Leica CM3050S cryostat, Leica, Wetzlar, Germany) and mounted onto glass slides (Super frost plus, Thermo Fisher Scientific, Waltham, MA, USA). Low temperature gelling agarose (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in phosphate buffered saline (to a final concentration of 10 mg ml-1) in an 80 °C water bath and then cooled to 37 °C. DQ gelatin (porcine; Invitrogen, Thermo

-1 Fisher Scientific, Waltham, MA, USA) was dissolved in dH2O (to a concentration of 1 mg ml ) and diluted 1:10 in the agarose solution. Lastly, 4’, 6’-diamidino-2-phenylindole (DAPI) was added to the agarose/DQ gelatin mixture (at a concentration of 1 μg ml-1). During this time tissue sections were brought to room temperature and washed in PBS to remove excess OCT. Approximately 40 μl of agarose/DAPI/DQ gelatin was added to each tissue section and a coverslip placed on the slide to ensure even film thickness across the sample section. All samples were incubated in the dark for 1 hr at 4 °C and then 18 hrs at room temperature. Following incubation, the samples were imaged immediately using a fluorescent microscope with a green filter (Leica, Wetzlar, Germany). To account for tissue auto-fluorescence, negative control slides were used to determine the microscope settings for each section. Three tissue sections were considered for each individual to ensure consistency in measurements. On each tissue section 3 separate image montages were taken along transects across the full diameter of the cross section. Following background subtraction, mean fluorescence intensity was calculated for each image, analysed using ImageJ. All histological analysis was conducted blind to the acclimation group and in all cases these tissue sections were taken

Thermal remodeling of the fish atrium

from a central portion of the atrial wall, away from either the sino-atrial or atrio-ventricular junctions.

Quantitative real-time PCR Transcript abundance of genes associated with muscle growth (ventricular myosin heavy chain; VMHC, muscle LIM protein; MLP, and small myosin light chain 2; SMLC2), hyperplasia (proliferating cell nuclear antigen; PCNA), angiogenesis (vascular endothelial growth factor; VEGF), collagen I (Col1a1, Col1a2 and Col1a3), connective tissue regulators (MMP2, MMP9, MMP13 and TIMP2), stretch and heart failure (ANP and BNP) and pro-hypertrophic nuclear factor of activating T (NFAT) signalling mediator (a regulator of calcineurin; RCAN1) were quantified in the atria of fish from cold-, control and warm-acclimated groups (n = 7 atria for each temperature). RNA was extracted from 5 mg of snap frozen tissue (RNeasyMicrokit, Qiagen, Venlo, NL) and amount and quality was determined (NanoDrop ND-1000, NanoDrop, Wilmington, DE, USA). An RNA concentration of 200 ± 50 ng μl-1 was used to make cDNA with SuperScript III First Strand Synthesis System (Invitrogen, Carlsbad, CA, USA). SYBR Green I pre-mixed chemo-technology was used for qPCR. qPCR was carried out in a 7900 HT sequence detection system (Applied Biosystems, Carlsbad, CA, USA). All primers were the same as those used in Keen et al. (2016), which were designed using Primer 3 from mRNA sequences available on PUBMED. All expression levels were normalised to housekeeping gene β-actin to determine absolute expression levels for comparison at each acclimation temperature. Three housekeeping genes were tested, β-actin, GAPDH and DNAJ1, and β- actin had most stable expression in relation to temperature acclimation, as found previously (e.g. Johansen et al., 2011; Keen et al., 2016).

Statistical analysis Chamber filling volume was calculated from filling time by the equation:

0.05 푣표푙푢푚푒 (푚푙) = 푡푖푚푒 (µs) × × 1000 60

The effect of temperature acclimation on the pressure-volume relationship was assessed by a general linear model (GLM) with pressure as the dependent variable, volume and acclimation group as fixed factors and body mass as the covariate, with a Tukey post hoc test for differences between groups, using R. Differences in myocyte bundle cross-sectional area, extra-bundular sinus space, collagen deposition, gelatinase activity and transcript abundance were assessed by GLM with Holm-Sidak post hoc test for differences between groups using Prism v6 (GraphPad Software, Inc., La Jolla, CA, USA). Post hoc analyses of AFM force curves were performed using Nanoscope Analysis v1.40 (Bruker, Coventry, UK), whereby a baseline correction was applied to each curve before a force fit was applied using a Herzian (spherical) model and a maximum force fit of 70 %. For all analyses significance was

Thermal remodeling of the fish atrium

considered to be P < 0.05, except for atomic force curves where significance was considered at P < 0.005. Values are presented as mean ± S. E. throughout except for atomic force curves where values are mean ± S. D. Statistical details are provided in the figure legends.

Results

Ex vivo chamber compliance Chronic changes in cardiac load can influence chamber compliance (Bing et al., 1971; Keen et al., 2016). To assess the functional effects of cardiac remodelling on the passive properties of the thermally acclimated fish atrium we generated ex vivo passive filling curves from freshly isolated intact atria, treated with BDM, at a common test temperature of 10 °C. Atrial pressure increased exponentially with filling volume for each temperature acclimation group. Figure 1 shows mean data for each acclimation temperature from 0 to the maximum physiological filling pressures experienced by rainbow trout in vivo (< 1 kPa) (Farrell et al., 1986; Graham & Farrell, 1989; Forster & Farrell, 1994; Hansen et al., 2002). Thermal acclimation altered the pressure-

2 volume relationship during atrial filling (R = 0.63, F2, 20335 = 1328.3, P < 0.001) showing increased stiffness after chronic cold exposure, across all filling volumes, compared to controls. Conversely, chronic warming caused increased compliance compared to controls, particularly at high filling volumes (t ratio = 40.5; Figure 1). The details and interactions of the statistical model are given in Table 2.

Table 1. The gross morpholgical parameters of thermally acclimated rainbow trout.

Thermal remodeling of the fish atrium

Figure 1. Ex vivo atrial passive filling pressure-volume relationships within the physiological relevant pressure range of < 0.5 kPa for cold (5 °C; blue squares), control (10 °C; green circles) and warm (18 °C; red triangles) acclimated rainbow trout (n = 8). Values are mean ± S. E., at all points on the curve n > 3. Pressure has been standardised to start at 0 kPa for graphical representation. Significance differences in compliance between acclimation groups were assessed by GLM with volume as the dependent variable, treatment and pressure as the fixed factors, and chamber mass as the covariate (P < 0.05), shown by dissimilar letters. See Table 2 for the statistical output.

Thermal remodeling of the fish atrium

Table 2. The statistical output table from the general linear model used to assess whole chamber complaince in the atrium.

Micromechanical tissue stiffness Tissue tensile strength and stiffness is mediated by multiple factors, including the abundance and organization of collagen fibrils. As alterations in tissue stiffness affect cardiac micromechanical properties (Fomovsky et al., 2010), we used AFM nano-indentation of atrial cryosections to determine if environmental temperature induces changes in local tissue stiffness. Figure 2A shows a bright-field micrograph of a representative control section of fish atrium, with black boxes to show the size and location of experimental test areas. Mean reduced modulus (Er); was significantly higher following chronic cold exposure, showing that cold-acclimated atrial tissue was stiffer than control and warm-acclimated tissue (P < 0.005;

Figure 2B). The Er frequency showed a standard distribution for each temperature (Figure 2C), suggesting that tissue micromechanics remain homogenous with thermal acclimation.

Thermal remodeling of the fish atrium

Figure 2. Atrial tissue micromechanics. (A) Bright-field microscope image of an unstained control atrial cryosection with black boxes (50 x 50 µm) demarking measurement areas. (B)

Mean reduced modulus (Er) was influenced by acclimation temperature. (C) Frequency distribution of Er, in cold (5 °C; blue), control (10 °C; green) and warm (18 °C; red) acclimation conditions (n = 3 animals per group, 3 regions per animal, 400 force curves per region). Values presented are mean ± S. D. Significance was assessed by GLM and is shown between groups by * (P < 0.005).

Thermal remodeling of the fish atrium

The atrial extracellular matrix Changes in cardiac stiffness/compliance is associated with remodelling of the ECM in both fish and mammals (Chapman et al., 1990; Klaiman et al., 2011; Keen et al., 2016). We used picro-sirus red to assess fibrillar collagen content (Figure 3A & B). Figure 3A shows a representative control section of fish atrium, stained with picro-sirus red, imaged under bright- field light and figure 3B shows the same section under plane polarised light. The highly collagenous epicardial surface of the tissue was excluded from the analysis. The polarised light images were used to quantify collagen as a percentage of whole tissue area. Percent collagen was low and although there was a trend towards increased collagen content in cold- acclimated animals approaching significance (P = 0.06) this was not statistically resolvable (Figure 3C). Although we were able to detect elastin in vessels using Miller’s elastic histological staining, we did not detect any in the atrial muscle (not shown).

In mammals, collagen I accounts for ~80 % of total collagen in the myocardium and is the main collagen in cardiac fibrosis (Medugorac, 1982). Mammalian collagen I is composed of type 1 (α1) and type 2 (α2) alpha-helical chains; fish also have an additional type 3 (α3) chain (Saito et al., 2001). mRNA expression of Col1a3 was 5.7-fold higher in the cold- compared to the warm-acclimated atrium (P < 0.05; Figure 3D). Expression of the Col1a1 and Col1a2 mRNA was down-regulated in both cold and warm fish compared with controls, suggesting temperature-independent remodelling is also occurring.

Matrix metalloproteinases (MMPs) are important regulators of ECM proteins and, therefore, tissue collagen content (Nagase et al., 2006). We assessed gelatinase activity of MMPs in tissue sections by in situ zymography. Figure 3E shows a representative fluorescent micrograph for cold-acclimated atria, and figure 4F for warm acclimated atria, treated with DQ gelatin to show gelatinase activity of MMPs in green and cell nuclei in blue. Semi-quantification of fluorescence intensity showed a 1.8-fold increase in gelatinase activity following warm acclimation compared to controls (P < 0.05), however, there was no difference between cold- acclimated and control groups (Figure 3G). Most MMPs are gelatinases and our mRNA showed MMP2, MMP9 and MMP13 were 2.2, 3.2 and 1.7-fold lower in the atrium of cold- acclimated animals than of warm-acclimated (P < 0.05; Figure 3H) suggesting these MMPs may contribute to the decreased protease activity following cooling. MMP activity is regulated by tissue inhibitors of MMPs (TIMPs) and thus, increased TIMP activity is associated with increased collagen deposition in tissue with active turnover. However, expression of the trout atrial TIMP2 gene was not affected by temperature acclimation (P < 0.05; Figure 3E).

Thermal remodeling of the fish atrium

Figure 3. Atrial connective tissue remodelling. A representative control atrial tissue micrograph, imaged under (A) bright-field and (B) polarised light, stained with picro-sirus red. (C) Semi-quantitative analysis of collagen content from polarized light images expressed as a percentage of total tissue. The epicardial surface of the tissue has been excluded from the analysis. There is a trend toward an increase in collagen in the cold (P=0.063). (D) mRNA expression of collagen genes (Col1a1, Col1a2, Col1a3). (E) Representative fluorescent micrographs of tissue treated with DQ gelatin to show gelatinase activity of matrix metalloprotinases (MMPs) in green and DAPI to show cell nuclei in blue for cold-acclimated and (F) warm-acclimated atria. (G) Semi-quantitative analysis of gelatinase enzyme activity. (H) mRNA expression of collagen regulatory genes (TIMP2 = up-regulation; MMP2, MMP9 and MMP13 = down-regulation) for cold (5 °C; blue), control (10 °C; green) and warm (18 °C; red) acclimated rainbow trout (n = 10 fish per acclimation group; 3 replicates for each animal were averaged for both histology and qPCR). Values presented are mean ± S. E. Significance was assessed by GLM with a Tukey, or Holm-Sidak for multiple comparisons, post hoc test. Significance between groups is shown by dissimilar letters (P < 0.05).

Thermal remodeling of the fish atrium

Atrial muscle morphology Chamber size and wall thickness can alter myocardial compliance in accordance with the law of Laplace (Katz & Rolett, 2016). Temperature acclimation did not affect total atrial mass or atrial mass relative to body mass (RAM) (Table 1). The fish atrium is a highly trabeculated and thin walled chamber. Figure 4A & B show representative tissue micrographs of cold- and warm-acclimated fish atrium, respectively, stained with H & E. Although there was no significant difference in myocyte bundle cross-sectional area with temperature acclimation (Figure 4C), mRNA expression of cardiac muscle-specific growth genes was down-regulated following cold acclimation and up-regulated following warm acclimation. Ventricular myosin heavy chain (VMHC) and small myosin light chain 2 (SMLC2), were 2.8- and 2.9-fold lower in the cold- than warm-acclimated group, respectively (P < 0.05; Figure 4D). Expression of muscle LIM protein (MLP) was not different between cold- and warm-acclimated animals (Figure 4D). A marker of angiogenesis, vascular endothelial growth factor (VEGF), was 8.8- fold lower in the cold-acclimated atrium compared to warm-acclimated (P < 0.05; Figure 4D) but, there was no difference in the expression of a marker of hyperplasia, proliferating cell nuclear antigen (PCNA; Figure 4D). In addition, extra-bundular sinus (EBS) space between myocyte bundles was increased by 32 % with cold acclimation and decreased by 29 % with warm acclimation compared to control animals (P < 0.05; Figure 4E).

Regulator of calcineurin (RCAN1) activates the calcineurin-NFAT signaling cascade which promotes hypertrophic growth (Rothermel et al., 2001; Vega et al., 2003; Wilkins et al., 2004; Johansen et al., 2011; Keen et al., 2016). RCAN1 mRNA expression was 2.7-fold lower in the cold- compared to the warm-acclimated atrium (P < 0.05; Figure 4F). Atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP) are released by cardiomyocytes in response to stretch caused by chronic pressure or volume overload and associated with activation of the foetal gene program (Kinnunen et al., 1993; Shih et al., 2015). ANP mRNA expression following chronic cooling was 6.4-fold lower than that in warm-acclimated atria (P < 0.05; Figure 4F). Temperature acclimation did not alter mRNA expression of BNP (Figure 4F).

Thermal remodeling of the fish atrium

Figure 4. Atrial tissue remodelling. Representative haematoxylin and eosin (H & E) stained atrial tissue sections for (A) cold-acclimated and (B) warm-acclimated rainbow trout. Quantification of (C) myocyte bundle cross sectional-area and (D) mRNA expression of markers of muscle growth (VMHC, MLP and SMLC2), hyperplasia (PNCA) and angiogenesis (VEGF). (E) extra-bundular sinus with temperature acclimation. (F) hypertrophic markers (ANP and BNP) and regulator of the pro-hypertrophic NFAT signalling pathway (RCAN1), with cold (5 °C; blue), control (10 °C; green) and warm (18 °C; red) acclimation (n = 10 fish per acclimation group; 3 replicates for each animal were averaged for both histology and qPCR). Values presented are mean ± S. E. Significance was assessed by GLM with Holm-Sidak post hoc test. Significance between groups is shown by dissimilar letters (P < 0.05).

Thermal remodeling of the fish atrium

Discussion

The fish heart remodels with seasonal temperature change (Graham & Farrell, 1989; Klaiman et al., 2011; Klaiman et al., 2014; Driedzic et al., 1996; Farrell et al., 1988a; Gamperl & Farrell, 2004) (Korajoki & Vornanen, 2009, 2012; Jayasundara et al., 2013; Keen et al., 2016). Here, we focused on the passive properties of the rainbow trout atrium across multiple levels of biological organization following chronic cooling and warming. Functionally, cold acclimation increased passive stiffness of the whole atrium and micromechanical stiffness of tissue sections. Increased stiffness was associated with an up-regulation of collagen promoting genes. Conversely, chronic warming gave an increase in whole atrial compliance, an up- regulation of collagen degrading genes and increased gelatinase activity of collagen degrading MMPs. Cold acclimation increased EBS and reduced mRNA expression of muscle specific and hypertrophic growth markers. The opposite responses occurred following warming. Together, these findings suggest that increased cardiac preload associated with chronic cooling causes dilation of the atrium and leads to a functional change in the passive properties, and upon chronic warming the opposite response occurs. Aspects of this remodelling phenotype reflect mammalian pathological atrial remodelling, which is often associated with chronic dilation and increased myocardial stiffness (Verheule et al., 2003; Pellman et al., 2010). Therefore, the trout heart may be an interesting vertebrate model for investigating the reversibility of chronic dilation and stiffening of the atrium.

Thermal remodelling of atrial compliance Temperature acclimation significantly altered the pressure-volume relationships in the atrium. Cold-acclimated tissue was stiffer and warm-acclimated tissue was more compliant than control tissue when all were tested at a common temperature over a physiological range of filling pressures (with normal filling pressure at ~0.01 kPa and greatest efficiency ~0.15 kPa; Farrell et al., 1986; Graham & Farrell, 1989; Hansen et al., 2002). Although atrial pressure- volume filling curves have previously been generated for fish (Forster & Farrell, 1994; Mendonca et al., 2007), this is the first study where they have been used to probe atrial remodelling. Micromechanical stiffness of the atrium assessed via AFM also increased with cold, however, no difference was found between the control and warm-acclimated groups. The increased micromechanical stiffness suggests that cold temperature leads to a remodelling of tissue ultrastructure and/or matrix organisation, which is exerting a functional effect on the myocardium. Furthermore, the standard distribution of Er accumulative frequency curves suggests that mechanical remodelling following temperature acclimation is due to homogenous structural and/or compositional remodelling of the whole atrial tissue, rather than isolated or specific regions of the tissue. This would suggest that cold-induced remodelling does not result in excess stiff material but rather has a lower proportion of compliant material or changes in intrinsic myocyte stiffness (Qiu et al., 2010).

Thermal remodeling of the fish atrium

Thermal remodelling of the atrial extracellular matrix In mammals, pathology driven remodelling of the atria is often associated with atrial fibrosis (Verheule et al., 2003). Although cold-induced cardiac remodelling cannot be considered a pathology, there is mounting evidence to suggest a number of consistent characteristics between seasonal remodeling in fish and pathological remodeling in mammals, including myocardial fibrosis (Klaiman et al., 2011; Keen et al., 2016). Picro-sirus red selectively stains fibrillar collagen. A trend (P = 0.063) for increased collagen deposition with cold acclimation was observed histologically which is supported by significant changes in mRNA expression of the fish-specific collagen gene, Col1a3. Col1a3 was strongly up-regulated following chronic cooling and down-regulated following chronic warming. Increased myocardial stiffness through fibrosis can help maintain mechanical cardiac performance and protect against over stretch of myocytes under the increased haemodynamic stress of pumping cold, highly viscous blood. The picro-sirus red analysis did not allow us to resolve perimysial collagen bundles in the fish atrium similar to that seen in mammals, meaning remodelling of connective tissue may only occur in very fine epimysial collagen fibres that surround the individual myocytes, which may have been missed by our histological analysis. We also did not find any evidence of elastin in the rainbow trout atrium using Miller’s elastic stain. This result differs from mammalian atrial anatomy (Smorodinova et al., 2015) and some previous data from the goldfish atrium where elastin fibres were visualised using Orecin stain (Garofalo et al., 2012). In addition, non-ECM components of the myocardium may alter atrial compliance following temperature acclimation, such as the titin (Patrick et al., 2010) and the actin cytoskeleton (Qiu et al., 2010), which could alter intrinsic stiffness of the cardiac myocytes.

The increased gelatinse activity of MMPs following warm acclimation suggests a reduction in collagen degradation, which was supported by increased mRNA expression of MMP9. Moreover, mRNA expression of MMP2, MMP9 and MMP13 was reduced in the cold- acclimated animals compared to control. In fish, MMP13 catalyses the hydrolysis of collagen, degrading it to gelatin (Hillegass et al., 2007) and MMP2 and MMP9 digest the gelatin into removable waste products (Kubota et al., 2003). We found no difference in the mRNA expression of the pro-collagen regulatory enzyme TIMP2 following thermal acclimation which differs from our previous finding in the ventricle (Keen et al., 2016). TIMPs other than TIMP2 may regulate the ECM in the fish atria.

Thermal remodelling of atrial myocardium We found an increase in EBS following cold acclimation and a decrease in EBS following warm acclimation. The change in EBS occurred without a significant change in myocyte cross- sectional area suggesting it is due to increased cardiac preload increasing inflation of the atrium. This result is supported by our mRNA expression data that shows muscle specific growth factors (MLP and SMLC2) and a marker of angiogenesis (VEGF) were reduced following cold acclimation, while SMLC2 and VEGF were increased following chronic warming.

Thermal remodeling of the fish atrium

Increased VEGF expression with warm acclimation may be required to increase blood supply as water oxygen content is reduced as water temperature rises, and angiogenesis is also associated with hypertrophic growth in mammals (Weber & Janicki, 1989). In mammals, chronic dilation increases EBS which can be further supplemented by apoptosis, reducing myocyte number during pathology (Verheule et al., 2003; Goudis et al., 2012; Xu et al., 2013). While we did not probe for apoptosis in this study, we believe that it is unlikely to occur due to the physiological nature of temperature-dependent remodelling, which occurs seasonally and transiently in fish. However, apoptosis has been shown in the fish heart following remodelling induced by exposure to Angiotensin II (Imbrogno et al., 2013). Atrial dilation in mammals can lead to AF (Nattel et al., 2008), however, we did not address whether fish exhibit AF in vivo following cold acclimation.

Up-regulation of ANP mRNA expression in the warm-acclimated atrium suggests an increase in stretch or myocyte hypertrophy which may appear inconsistent with our suggestion of atrial dilation following cold-induced volume overload (Raskin et al., 2009). However, the in vivo effect of ANP on the trout cardiovascular system is vasodilatory, reducing venous return blood to the heart by increasing venous compliance, which in turn decreases cardiac output and arterial pressure (Olson et al., 1997). Due to the long acclimation period in this study it is possible that atrial dilation with cold acclimation reduces the need for the vasodilatory properties of ANP. Mitogen activated protein kinases (MAPKs) and the calcineurin-NFAT pathway are central to pathological hypertrophic growth in mammals (Wilkins et al., 2004; Bernardo et al., 2010). In some mammalian pathologies of the atria, hypertrophic signaling cascades are activated to promote the increased protein synthesis required for hypertrophic muscle growth (Lin et al., 2004; Saygili et al., 2009; Tan et al., 2013). Ventricular expression of RCAN1 can increase calcineurin-NFAT signalling and enhance hypertrophic growth of myocytes in mammals and fish (Liu et al., 2009; Johansen et al., 2011; Keen et al., 2016) (Figure 6), however, it has also been shown to have cardio-protective, anti-hypertrophic effects (Rothermel et al., 2001). RCAN1 gene expression in the atrium during cardiac remodelling is relatively unexplored. The reasons for high levels of RCAN1 mRNA with warm acclimation, in this study, are unclear. It may be due to other stresses on the atrium (Rothermel et al., 2003) or due to the high ANP mRNA levels (Tokudome et al., 2005).

Differences in ventricular and atrial remodelling The remodelling response of the fish atrium shares some but not all characteristics with ventricular remodelling (Figure 6). The increase in EBS with cold acclimation is the opposite response to that seen in the ventricle (Klaiman et al., 2011; Keen et al., 2016) (Figure 6). Ventricular myocyte cross-sectional area increases and EBS decreases in the rainbow trout with cold acclimation (Klaiman et al., 2011; Keen et al., 2016) suggesting a spongy layer driven ventricular hypertrophy (Farrell et al., 1988a; Graham & Farrell, 1989) (Figure 6). The differential response of the atrium and ventricle highlights the different stresses imposed by

Thermal remodeling of the fish atrium

increased cardiac preload on the two chambers and the response required to maintain cardiac function. The ventricle shows a significant increase in collagen deposition in the compact layer, but similar to the atrium we did not find this difference in the spongy layer (Keen et al., 2016) (Figure 6). The mRNA expression data for collagen regulation agrees with that previously determined in the rainbow trout ventricle (Keen et al., 2016) (Fiigure 6). Despite only small levels of collagen, both the atrium and the spongy myocardium showed increased stiffness of the whole chamber and micromechanical stiffness in tissue sections (Keen et al., 2016) (Figure 6). This result shows chronic cold to have the same functional effect on the atrium as we have previously shown in the ventricle. We suggest the overall reason for remodelling in both cases is to increase ventricular pumping capacity, via the Frank-Starling mechanism, as well as protect cardiac myocytes from over inflation while pumping high volumes of viscous blood (Graham & Farrell, 1989; Shiels & White, 2008; Keen et al., 2016).

Perspectives and significance Chronic dilation and stiffening of the atria, with associated fibrosis, are hallmarks of pathological remodelling in mammals (Nattel et al., 2008; Pellman et al., 2010). In fish, atrial stiffening appears to occur without the significant fibrosis seen during aging or pathological remodelling in the mammalian heart. Therefore, increased stiffness with cold acclimation is likely to protect the myocardial wall from over inflation with increased haemodynamic stress of viscous blood and high preload. It may also protect the fish heart from atrial fibrillation. The fish atrium is a volume variable reservoir responsible for altering stroke volume via the Frank- Starling mechanism. At cold temperatures, blood is viscous and stroke volume is high (Graham & Farrell, 1989). The atrium must resist the increased haemodynamic stress, but distend further to hold a larger volume of blood to increase ventricular filling. Interestingly, Hansen et al. (2002) showed the effect of increased preload to be more metabolically costly than increased afterload in the fish heart. Due to the role of the atrium as a volume reservoir, chronic dilation may occur to store venous blood preventing ventricular preload rising above maximal efficiency (Hansen et al., 2002). Atrial filling is further increased by vis-a-fronte filling during ventricle contraction in some fish, where a pressure void in the pericardium distends the atrium (Farrell et al., 1988b). We did not determine whether chronic cold induces remodelling beyond that necessary for accommodating the changes in blood viscosity or whether this response is associated with other pathologies such as atrial fibrilation. To understand whether temperature pushes the fish heart past the compensation/adaptation phase into the true pathology, future studies could combine approaches that increase afterload pressure, known to induce pathology in mammalian hearts, (e.g. outflow constriction) in fish hearts.

Thermal remodeling of the fish atrium

Figure 5. An overview of atrial and ventricular remodelling in rainbow trout exposed to chronic cold (5 °C) and chronic warm (18 °C) temperatures.

Thermal remodeling of the fish atrium

The results of this study suggest opposite cardiac remodelling responses occur in chronically warmed fish compared with chronically cooled fish. However, to directly test whether the cardiac remodelling response is ‘reversible’, cardiac function would have to be assessed in the same animals following acclimation to both temperatures (i.e. cooling then warming and/or warming then cooling). In addition, the collagen I gene that appeared most responsive to thermal remodelling in the current study is fish-specific (Col1a3). Collagen chains containing this domain have been shown to have greater susceptibility to heat denaturation and degradation by MMP13 than collagen chains without it (Saito et al., 2001), which may explain its malleability with chronic temperature change. If this collagen domain is driving changes in chamber compliance, it may explain why a typically pathological response in mammals occurs transiently in fish.

Additional information

Competing interests The authors declare that we have no competing interests

Authors contributions A.N.K., A.J.F. and H.A.S. are responsible for the concept and design of the research. A.N.K., A.J.F. and J.C.M. performed experiments and analysed data in H.A.S. and M.J.S. laboratories. A.N.K., A.J.F., J.C.M., M.J.S., P.G. and H.A.S. interpreted the results of the experiments. A.N.K., J.C.M. and H.A.S. drafted the manuscript. A.N.K., A.J.F., J.C.M., M.J.S., P.G., and H.A.S. revised and edited the manuscript. A.N.K., A.J.F., J.C.M., M.J.S., P.G., and H.A.S. approve the final version of the manuscript submitted for publication.

Funding. A.N.K. and A. J. F. were supported by studentships from the BBSRC. M.J.S. is funded by the Medical Research Council UK (grant G1001398). The Shiels lab is supported by the Leverhulme Trust (240613).

Acknowledgements We thank Dr Margaux Horn and Dr Sanjoy Chowdhury for helpful discussions throughout this project. We thank and, Dr Alex Henderson and Dr Robert Nudds for help with data analysis and Peter Walker for help with histology. Histology was performed at the University of Manchester Histology Facility. Atomic Force Microscopy was carried out in the University of Manchester BioAFM Facility.

Thermal remodeling of the fish atrium

References

Aho E & Vornanen M. (1999). Contractile properties of atrial and ventricular myocardium of the heart of rainbow trout Oncorhynchus mykiss: Effects of thermal acclimation. Journal of Experimental Biology 202, 2663-2677.

Akhtar R, Cruickshank JK, Zhao X, Walton LA, Gardiner NJ, Barrett SD, Graham HK, Derby B & Sherratt MJ. (2014). Localized micro- and nano-scale remodelling in the diabetic aorta. Acta Biomaterialia 10, 4843-4851.

Bauer A, McDonald AD & Donahue JK. (2004) Pathophysiological findings in a model of persistent atrial fibrillation and severe congestive heart failure. Cardiovascular Research 61, 764-770.

Bernardo BC, Weeks KL, Pretorius L & McMullen JR. (2010). Molecular distinction between physiological and pathological cardiac hypertrophy: Experimental findings and therapeutic strategies. Pharmacology & Therapeutics 128, 191-227.

Bing OHL, Matsushi.S, Fanburg BL & Levine HJ. (1971). Mechanical properties of rat cardiac muscle during experimental hypertrophy. Circulation Research 28, 234-245.

Chapman D, Weber KT & Eghbali M. (1990). Regulation of fibrillar collagen type-I and type-III and basement-membrane type-IV collagen gene-expression in pressure overloaded rat myocardium. Circulation Research 67, 787-794.

Crick SL & Yin FCP. (2007). Assessing micromechanical properties of cells with atomic force microscopy: importance of the contact point. Biomechanics and Modeling in Mechanobiology 6, 199-210.

D'Andrea A, Riegler L, Cocchia R, Scarafile R, Salerno G, Gravino R, Golia E, Vriz O, Citro R, Limongelli G, Calabro P, Di Salvo G, Caso P, Russo MG, Bossone E & Calabro R. (2010). Left atrial volume index in highly trained athletes. American Heart Journal 159, 1155-1161.

Driedzic WR, Bailey JR & Sephton DH. (1996). Cardiac adaptations to low temperature non- polar teleost fish. Journal of Experimental Zoology 1996, 275.

Farrell AP, Hammons AM, Graham MS & Tibbits GF. (1988a). Cardiac growth in rainbow trout, Salmo gairdneri. Canadian Journal of Zoology-Revue Canadienne De Zoologie 66, 2368-2373.

Farrell AP, Johansen JA & Graham MS. (1988b). The role of the pericardium in cardiac performance of the trout (Salmo gairdneri). Physiological Zoology 61, 213-221.

Farrell AP & Jones DR. (1992). The Heart. In Fish Physiology, Volume XII, Part A The Cardiovascular System.

Farrell AP, Macleod KR & Chancey B. (1986). Intrinsic mechanical properties of the perfused rainbow-trout heart and the effects of catecholamines and extracellular calcium under control and acidotic conditions. Journal of Experimental Biology 125, 319-345.

Fomovsky GM, Thomopoulos S & Holmes JW. (2010). Contribution of extracellular matrix to the mechanical properties of the heart. Journal of Molecular and Cellular Cardiology 48, 490-496.

Forster ME & Farrell AP. (1994). The volumes of the chambers of the trout heart. Comparative Biochemistry and Physiology a-Physiology 109, 127-132.

Thermal remodeling of the fish atrium

Gamperl AK & Farrell AP. (2004). Cardiac plasticity in fishes: environmental influences and intraspecific differences. Journal of Experimental Biology 207, 2539-2550.

Goudis CA, Kallergis EM & Vardas PE. (2012). Extracellular matrix alterations in the atria: insights into the mechanisms and perpetuation of atrial fibrillation. Europace 14, 623- 630.

Graham HK, Akhtar R, Kridiotis C, Derby B, Kundu T, Trafford AW & Sherratt MJ. (2011). Localised micro-mechanical stiffening in the ageing aorta. Mechanisms of Ageing and Development 132, 459-467.

Graham MS & Farrell AP. (1989). The effect of temperature acclimation and adrenaline on the performance of a perfused trout heart. Physiological Zoology 62, 38-61.

Guyton AC. (1981). Textbook of Medical Physiology. Saunders Company, Philadelphia.

Hansen J, Gesser H & Altimiras J. (2002). Mechanical efficiency of the trout heart during volume and pressure loading: metabolic implications of the stiffness of the ventricular tissue. Journal of Comparative Physiology B-Biochemical Systemic and Environmental Physiology 172, 477-484.

Hauser AM, Dressendorfer RH, Vos M, Hashimoto T, Gordon S & Timmis GC. (1985). Symmetric cardiac enlargement in highly trained endurance athletes: A two- dimensional echocardiographic study. American Heart Journal 109, 1038-1044.

Hillegass JM, Villano CM, Cooper KR & White LA. (2007). Matrix metalloproteinase-13 is required for zebra fish (Danio rerio) development and is a target for glucocorticoids. Toxicological Sciences 100, 168-179.

Imbrogno S, Garofalo F, Amelio D, Capria C & Cerra MC. (2013). Humoral control of cardiac remodeling in fish: Role of Angiotensin II. General and Comparative Endocrinology 194, 189-197.

Jayasundara N, Gardner LD & Block BA. (2013). Effects of temperature acclimation on Pacific bluefin tuna (Thunnus orientalis) cardiac transcriptome. American Journal of Physiology-Regulatory Integrative and Comparative Physiology 305, R1010-R1020.

Johansen IB, Lunde IG, Rosjo H, Christensen G, Nilsson GE, Bakken M & Overli O. (2011). Cortisol response to stress is associated with myocardial remodeling in salmonid fishes. Journal of Experimental Biology 214, 1313-1321.

Johansen K & Burggren W. (1980). Cardiovascular function in the lower vertebrates. In Hearts and Heart-Like Organs, ed. Bourne GH, pp. 61-118. Academic Press, Inc.

Junqueira LCU, Bignolas G & Brentani RR. (1979). Picrosirius staining plus polarization microscopy, a specific method for collagen detection in tissue-sections. Histochemical Journal 11, 447-455.

Katz AM & Rolett EL. (2016). Heart failure: when form fails to follow function. European Heart Journal.

Keen AN, Fenna AJ, McConnell JC, Sherratt MJ, Gardner P & Shiels HA. (2016). The dynamic nature of hypertrophic and fibrotic remodelling in the fish ventricle. Frontiers in Physiology - Intergrative Physiology 6.

Kemp AD, Harding CC, Cabral WA, Marini JC & Wallace JM. (2012). Effects of tissue hydration on nanoscale structural morphology and mechanics of individual Type I collagen fibrils in the Brtl mouse model of Osteogenesis Imperfecta. Journal of Structural Biology 180, 428-438.

Thermal remodeling of the fish atrium

Kinnunen P, Vuolteenaho O & Ruskoaho H. (1993). Mechanisms of atrial and brain natriuretic peptide release from rat ventricular myocardium - effect of stretching. Endocrinology 132, 1961-1970.

Klaiman JM, Fenna AJ, Shiels HA, Macri J & Gillis TE. (2011). Cardiac Remodeling in Fish: Strategies to Maintain Heart Function during Temperature Change. PLoS ONE 6, e24464.

Klaiman JM, Pyle WG & Gillis TE. (2014). Cold acclimation increases cardiac myofilament function and ventricular pressure generation in trout. Journal of Experimental Biology 217, 4132-4140.

Knackstedt C, Gramley F, Schimpf T, Mischke K, Zarse M, Plisiene J, Schmid M, Lorenzen J, Frechen D, Neef P, Hanrath P, Kelm M & Schauerte P. (2008). Association of echocardiographic atrial size and atrial fibrosis in a sequential model of congestive heart failure and atrial fibrillation. Cardiovascular Pathology 17, 318-324.

Korajoki H & Vornanen M. (2009). Expression of calsequestrin in atrial and ventricular muscle of thermally acclimated rainbow trout. Journal of Experimental Biology 212, 3403- 3414.

Korajoki H & Vornanen M. (2012). Expression of SERCA and phospholamban in rainbow trout (Oncorhynchus mykiss) heart: comparison of atrial and ventricular tissue and effects of thermal acclimation. Journal of Experimental Biology 215.

Korajoki H & Vornanen M. (2014). Species- and chamber-specific responses of 12 kDa FK506- binding protein to temperature in fish heart. Fish Physiology and Biochemistry 40, 539- 549.

Kubota M, Kinoshita M, Takeuchi K, Kubota S, Toyohara H & Sakaguchi M. (2003). Solubilization of type I collagen from fish muscle connective tissue by matrix metalloproteinase-9 at chilled temperature. Fisheries Science 69, 1053-1059.

Kumar S, Teh AW, Medi C, Kistler PM, Morton JB & Kalman JM. (2012). Atrial remodeling in varying clinical substrates within beating human hearts: Relevance to atrial fibrillation. Progress in Biophysics & Molecular Biology 110, 278-294.

Lai NC, Graham JB, Dalton N, Shabetai R & Bhargava V. (1998). Echocardiographic and hemodynamic determinations of the ventricular filling pattern in some teleost fishes. Physiological Zoology 71, 157-167.

Lin C-C, Lin J-L, Lin C-S, Tsai M-C, Su M-J, Lai L-P & Huang SKS. (2004). Activation of the Calcineurin-Nuclear Factor of Activated T-Cell Signal Transduction Pathway in Atrial Fibrillation. Chest 126, 1926-1932.

Liu QH, Busby JC & Molkentin JD. (2009). Interaction between TAK1-TAB1-TAB2 and RCAN1-calcineurin defines a signalling nodal control point. Nature Cell Biology 11, 154-161.

Medugorac I. (1982). Characterization of intramuscular collagen in the mammalian left ventricle. Basic Research in Cardiology 77, 589-598.

Mendonca PC, Genge AG, Deitch EJ & Gamperl AK. (2007). Mechanisms responsible for the enhanced pumping capacity of the in situ winter flounder heart (Pseudopleuronectes americanus). American Journal of Physiology-Regulatory Integrative and Comparative Physiology 293, R2112-R2119.

Miller PJ. (1971). Elastin stain. Medical Laboratory Technology 28, 148-149.

Thermal remodeling of the fish atrium

Mook ORF, Overbeek CV, Ackema EG, Maldegem FV & Frederiks WM. (2003). In Situ Localization of Gelatinolytic Activity in the Extracellular Matrix of Metastases of Colon Cancer in Rat Liver Using Quenched Fluorogenic DQ-gelatin. Journal of Histochemistry & Cytochemistry 5, 821-829.

Nagase H, Visse R & Murphy G. (2006). Structure and function of matrix metalloproteinases and TIMPs. Cardiovascular Research 69, 562-573.

Olson KR, Conklin DJ, Farrell AP, Keen JE, Takei Y, Weaver L, Smith MP & Zhang YT. (1997). Effects of natriuretic peptides and nitroprusside on venous function in trout. American Journal of Physiology-Regulatory Integrative and Comparative Physiology 273, R527- R539.

Opie LH, Commerford PJ, Gersh BJ & Pfeffer MA. (2006). Controversies in Cardiology 4 - Controversies in ventricular remodelling. Lancet 367, 356-367.

Pelliccia A, Maron BJ, Di Paolo FM, Biffi A, Quattrini FM, Pisicchio C, Roselli A, Caselli S & Culasso F. (2005). Prevalence and clinical significance of left atrial remodeling in competitive athletes. Journal of the American College of Cardiology 46, 690-696.

Pellman J, Lyon RC & Sheikh F. (2010). Extracellular matrix remodeling in atrial fibrosis: Mechanisms and implications in atrial fibrillation. Journal of Molecular and Cellular Cardiology 48, 461-467.

Polovkova OG, Makeeva OA, Lezhnev AA, Goncharova IA, Kulish EV, Shipulin VM & Puzyrev VP. (2013) Expression profiles of calcineurin pathway genes in myocardium in relation to ischemic heart remodeling in humans. Molecular Biology 47, 382-388.

Qiu H, Zhu Y, Sun Z, Trzeciakowski JP, Gansner M, Depre C, Resuello RRG, Natividad FF, Hunter WC, Genin GM, Elson EL, Vatner DE, Meininger GA & Vatner SF. (2010). Vascular Smooth Muscle Cell Stiffness As a Mechanism for Increased Aortic Stiffness With Aging. Circulation Research 107, 615-619.

R Core Team. (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/.

Raskin AM, Hoshijima M, Swanson E, McCulloch AD & Omens JH. (2009). Hypertrophic Gene Expression Induced by Chronic Stretch of Excised Mouse Heart Muscle. Molecular & cellular biomechanics : MCB 6, 145-159.

Rothermel BA, McKinsey TA, Vega RB, Nicol RL, Mammen P, Yang J, Antos CL, Shelton JM, Bassel-Duby R, Olson EN & Williams RS. (2001). Myocyte-enriched calcineurin- interacting protein, MCIP1, inhibits cardiac hypertrophy in vivo. Proceedings of the National Academy of Sciences of the United States of America 98, 3328-3333.

Rothermel BA, Vega RB & Williams RS. (2003). The role of modulatory calcineurin-interacting proteins in calcineurin signaling. Trends in Cardiovascular Medicine 13, 15-21.

Saito M, Takenouchi Y, Kunisaki N & Kimura S. (2001). Complete primary structure of rainbow trout type I collagen consisting of alpha 1(I)alpha 2(I)alpha 3(I) heterotrimers. European Journal of Biochemistry 268, 2817-2827.

Sanfilippo AJ, Abascal VM, Sheehan M, Oertel LB, Harrigan P, Hughes RA & Weyman AE. (1990). Atrial enlargement as a consequence of atrial fibrillation. A prospective echocardiographic study. Circulation 82, 792-797.

Saygili E, Rana OR, Meyer C, Gemein C, Andrzejewski MG, Ludwig A, Weber C, Schotten U, Krüttgen A, Weis J, Schwinger RHG, Mischke K, Rassaf T, Kelm M & Schauerte P. (2009). The angiotensin–calcineurin–NFAT pathway mediates stretch-induced up-

Thermal remodeling of the fish atrium

regulation of matrix metalloproteinases-2/-9 in atrial myocytes. Basic Research in Cardiology 104, 435-448.

Schneider CA, Rasband WS & Eliceiri KW. (2012). NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9, 671-675.

Schotten U, de Haan S, Neuberger HR, Eijsbouts S, Blaauw Y, Tieleman R & Allessie M. (2004). Loss of atrial contractility is primary cause of atrial dilatation during first days of atrial fibrillation. American Journal of Physiology-Heart and Circulatory Physiology 287, H2324-H2331.

Schotten U, Neuberger HR & Allessie MA. (2003). The role of atrial dilatation in the domestication of atrial fibrillation. Progress in Biophysics & Molecular Biology 82, 151- 162.

Shiels HA & White E. (2008). The Frank-Starling mechanism in vertebrate cardiac myocytes. Journal of Experimental Biology 211, 2005-2013.

Shih YH, Zhang YJ, Ding YH, Ross CA, Li H, Olson TM & Xu XL. (2015). Cardiac Transcriptome and Dilated Cardiomyopathy Genes in Zebrafish. Circulation- Cardiovascular Genetics 8, 261-269.

Tan N, Chung MK, Smith JD, Hsu J, Serre D, Newton DW, Castel L, Soltesz E, Pettersson G, Gillinov AM, Van Wagoner DR & Barnard J. (2013). A Weighted Gene Co-Expression Network Analysis of Human Left Atrial Tissue Identifies Gene Modules Associated with Atrial Fibrillation. Circulation Cardiovascular genetics 6, 362-371.

Tokudome T, Horio T, Kishimoto I, Soeki T, Mori K, Kawano Y, Kohno M, Garbers DL, Nakao K & Kangawa K. (2005). Calcineurin-nuclear factor of activated T cells pathway- dependent cardiac remodeling in mice deficient in guanylyl cyclase A, a receptor for atrial and brain natriuretic peptides. Circulation 111, 3095-3104.

Vega RB, Rothermel BA, Weinheimer CJ, Kovacs A, Naseem RH, Bassel-Duby R, Williams RS & Olson EN. (2003). Dual roles of modulatory calcineurin-interacting protein 1 in cardiac hypertrophy. Proceedings of the National Academy of Sciences of the United States of America 100, 669-674.

Verheule S, Wilson E, Everett T, Shanbhag S, Golden C & Olgin J. (2003). Alterations in atrial electrophysiology and tissue structure in a canine model of chronic atrial dilatation due to mitral regurgitation. Circulation 107, 2615-2622.

Wallace JM, Harding CC & Kemp AD. (2012). Effects of Hydration on Nanoscale Structural Morphology and Mechanics of Individual Type I Collagen Fibrils. MRS Proceedings 1465.

Weber KT & Janicki JS. (1989). Angiotensin and the remodeling of the myocardium. British Journal of Clinical Pharmacology 28.

Wilkins BJ, Dai YS, Bueno OF, Parsons SA, Xu J, Plank DM, Jones F, Kimball TR & Molkentin JD. (2004). Calcineurin/NFAT coupling participates in pathological, but not physiological, cardiac hypertrophy. Circulation Research 94, 110-118.

Xu GJ, Gan TY, Tang BP, Chen ZH, Mahemuti A, Jiang T, Song JG, Guo X, Li YD, Miao HJ, Zhou XH, Zhang Y & Li JX. (2013). Accelerated fibrosis and apoptosis with ageing and in atrial fibrillation: Adaptive responses with maladaptive consequences. Experimental and Therapeutic Medicine 5, 723-729.

Thermal remodeling of the fish atrium

Thermal remodelling of the ectothermic heart

6. Remodelling of compliance, structure and connective tissue in the fish outflow tract with temperature acclimation.

This chapter is presented in format of the manuscript that has been submitted to the

Journal of Experimental Biology.

Page | 85

Thermal remodelling of the ectothermic heart

Page | 86

Thermal remodelling of the fish outflow tract

Title: Remodelling of compliance, structure and connective tissue in the fish outflow tract with temperature acclimation.

Adam N. Keen1*, John J. Mackrill2, Peter Gardner3, Holly A. Shiels1*

1Faculty of Life Sciences, University of Manchester, Manchester, UK 3Department of Physiology, University College Cork, Ireland 4School of Chemical Engineering and Analytical Science, Manchester Institute of Biotechnology, University of Manchester, UK

*Author of correspondence: [email protected] [email protected] +44 161 275 5092

Short title: Thermal remodelling of the fish outflow tract

Text: 8,124 Figures: 5 Tables: 1 References: 52

Keywords: Pressure-volume, Stiffness, Temperature Acclimation, Collagen, Elastin, Matrix Metalloproteinase

List of abbreviations: DAPI, 4’, 6’-diamidino-2-phenylindole; GLM, general linear model; MMP, matrix metalloproteinase; OFT, outflow tract; SDS-PAGE, sodium dodecyl sulfate polyacrylamide electrophoresis.

Summary statement Chronic cold increases stiffness of the fish outflow tract, increasing ventricular afterload. The opposite occurs following chronic warming. The outflow tract may be central thermal remodelling of the fish heart.

Thermal remodelling of the fish outflow tract

ABSTRACT

To protect the delicate gill capillary network from high pressure blood flow the fish outflow (OFT) tract is equipped with a specialized structure called the bulbus arteriosus. This chamber is highly compliant and distensible to act as a windkessel for systolic blood flow. Chronic temperature change alters fish cardiac physiology, causing ventricular hypertrophy and increasing stroke volume. Therefore, the biomechanics of the bulbus arteriosus may be central to a cardiac remodelling response. Rainbow trout (Oncorhynchus mykiss) were held at 10 °C or chronically (> 8 weeks) exposed to cooling (5 °C) or warming (18 °C). Chronic cold increased whole chamber passive stiffness and chronic warm increased chamber compliance. However, despite the chamber being 1.9-fold more compliant in warm- compared to cold- acclimated animals there was no change in chamber distensibility. These changes in chamber compliance were associated with changes in the connective tissue content of the bulbus wall, with the fibrillar collagen to elastin ratio increasing 1.4-fold following chronic cooling and decreasing 2.3-fold following chronic warming compared to control animals. Degradation of fibrillar collagen is regulated by matrix metalloproteinases (MMPs). Gelatin zymography revealed a 39.3-fold higher MMP-2 abundance in warm- compared to cold-acclimated tissue, which meant the proMMP-2 to MMP-2 ratio was 2-fold higher than controls in warm-acclimated tissue and 6.2-fold lower than controls in cold-acclimated tissue. Finally, in situ gelatin zymography showed that localised endogenous MMP activity was 4.1-fold higher in bulbular tissue from warm-acclimated animals compared with controls. These findings indicate that chronic warming increases compliance of the trout OFT by decreasing the collagen to elastin ratio of the bulbus wall. Conversely, chronic cooling increases stiffness of the OFT by increasing the collagen to elastin ratio. Remodelling of the OFT may cause changes in cardiac afterload with thermal remodelling of the fish heart and contribute to ventricular hypertrophy following cold acclimation.

Thermal remodelling of the fish outflow tract

INTRODUCTION

During ventricular contraction, blood leaves the heart at high pressure to circulate the body. In mammals, to ensure blood flow does not damage capillary vessels systemic blood flow is smoothed by the elastic aorta and the arterial tree (Girerd et al., 1991). Pulmonary blood flow is subject to a second, low pressure circulation from the right side of the heart to protect the delicate capillaries of the lungs. The fish circulatory system, however, forms a single circuit in which blood passes through a minimum of two sets of capillaries as it circulates the body (Farrell and Jones, 1992). As blood is ejected from the heart it passes directly to the gills via a short ventral aorta. Therefore, to prevent haemodynamic damage of the capillary network at the gill, the teleost outflow tract (OFT) has a highly specialized structure with a bulbus arteriosus and bulbo-ventricular valves between the ventricle and ventral aorta (Icardo, 2006). Although the terms ‘OFT’ and ‘bulbus arteriousus’ are often used interchangeably, for clarity we defined the bulbus arteriosus as the chamber and the OFT as the whole outflow section. The role of the bulbus is to smooth high pressure, systolic blood, providing a steady flow across capillary networks and throughout the rest of the body (Braun et al., 2003a; Farrell, 1979). The bulbus arteriosus is highly compliant and able to serve the same ‘windkessel’ function as the whole mammalian arterial tree (Bushnell et al., 1992; Farrell, 1979; Jones et al., 1993; Jones et al., 1974; Licht and Harris, 1973; Priede, 1976; Stevens et al., 1972; Watson and Cobb, 1979).

Whether the bulbus arteriousus is developmentally of cardiac or arterial origin is open to debate (Benjamin et al., 1983; Braun et al., 2003a; Clark and Rodnick, 1999; Duran et al., 2008; Licht and Harris, 1973; Priede, 1976; Seth et al., 2014). Unlike the rest of the heart, it has no actively contractile cardiac tissue, so no pumping ability, and is structurally similar to an artery with three morphological layers (Braun et al., 2003a; Icardo and Colvee, 2011). The inner, endocardial layer consists of bulbular ridges lined with endothelial cells. The inner media is predominantly an elastic matrix with inter-dispersed smooth muscle cells. However, unlike an artery, the elastin fibres are not organized into lamellae but are fibrillar (Braun et al., 2003a). The fibrillar arrangement and high elastin to collagen ratio make the bulbus hugely compliant and gives specialized ‘r-shaped’ inflation properties, compared to the typical ‘J-shaped’ inflation of an artery (Benjamin et al., 1983; Braun et al., 2003a; Icardo, 2006; Icardo et al., 2000; Licht and Harris, 1973; Serafini-Fracassini et al., 1978). The outer media integrates subepicardial layers of collagen fibres surrounded by mesothelial cells, designed to prevent damage from high pressure blood flow. The specialized elastic matrix of the bulbus arteriosus make it both strong and compliant; allowing it to withstand high hemodynamic strain and enabling its function of regulating blood flow (Icardo and Colvee, 2011).

Chronic changes in pressure and volume load can cause a dynamic remodelling response of the fish heart (Keen et al., 2016; Klaiman et al., 2011; Klaiman et al., 2014). Rainbow trout experience seasonal fluctuations in ambient temperature, but remain active throughout the

Thermal remodelling of the fish outflow tract year. Low winter temperature directly causes bradycardia, decreases contractile function and increases blood viscosity (Graham and Farrell, 1989; Graham and Fletcher, 1983). To maintain appropriate cardiac output, ventricular mass increases to be able to pump high volumes of viscous blood and this ventricular hypertrophy has been linked to elevated central venous pressure, end-diastolic volume and, thus, increased stroke volume (Clark and Rodnick, 1999; Convertino et al., 1991; Franklin and Davie, 1992). With these changes in stroke volume it is necessary to actively adjust the vessels leading to the gill capillary network to maintain an optimal flow and pressure (Clark and Rodnick, 1999; Farrell, 1979; Seth et al., 2010). Therefore, with seasonal temperature change the fish OFT may have to undergo morpho-functional adaptations to alter compliance and/or distensibility (Clark and Rodnick, 1999).

By simulating seasonal temperature change, our aim was to investigate the effects of chronic cooling (from 10 ± 1 °C to 5 ± 1 °C) and chronic warming (from 10 ± 1 °C to 18 ± 1 °C) on the compliance and connective tissue structure of the rainbow trout OFT. Following our recent studies which suggest an increase in stiffness and collagen deposition with chronic cooling, and the opposite response with chronic warming, in the fish ventricle (Keen et al., 2016) our hypothesis was that chronic cooling would increase stiffness and collagen deposition in the bulbus arteriosus, while chronic warming would increase compliance and decrease collagen content. We generated ex vivo pressure-volume curves from freshly isolated fish OFT and calculated chamber compliance and distensibilty. We assessed connective tissue content by histological staining and orientation of fibrillar collagen. Connective tissue deposition and degradation is regulated by matrix metalloproteinase (MMPs). We assessed abundance and activation of specific MMPs by gelatin zymography and spatial activity of MMPs by in situ gelatin zymography. As rainbow trout remain active with seasonal temperature change, we were particularly interested in the variable remodelling following both cold and warm acclimation. We found that chronic cooling increased passive stiffness of the bulbus arteriosus. Following chronic warming, chamber compliance was increased compared to cold-acclimated and control animals. Cold acclimation resulted in an increase in the collagen to elastin ratio of the bulbus arteriousus. Conversely, warm acclimation resulted in a decrease in the collagen to elastin ratio of the bulbus arteriosus and increased abundance of collagen degrading enzyme, MMP-2. These results suggest that the bulbus arteriosus may play a key role in remodelling of the fish heart with temperature as changes in compliance alter cardiac afterload pressures.

Thermal remodelling of the fish outflow tract

MATERIALS AND METHODS

Ethical approval All husbandry and housing conditions were in accordance with the local handling protocols and adhere to the UK Home Office legislation. All experimental procedures were approved by the University of Manchester’s ethical review committee.

Experimental animals Sexually mature female rainbow trout (Onchorynchus mykiss; n = 47; morphometric data in Table 1) were purchased from Dunsop Bridge Trout Farm (Clitheroe, UK), housed on a 12 hr light: 12 hr dark cycle in ~500 L re-circulated aerated, de-chlorinated fresh water tanks at 10 ± 1 °C and fed to satiation 3 times per week. Water quality was ensured with 30 % water changes 3 times per week and regular tests for temperature, pH, nitrates and nitrites. Fish were held under these conditions for a minimum of 2 weeks before being randomly assigned to one of three acclimation groups; cold (5 ± 1 °C), control (i.e. no change; 10 ± 1 °C) or warm (18 ± 1 °C). These temperatures are based on previous literature which describes the cardiac remodelling response in salmonids (Keen et al., 2016; Klaiman et al., 2011). Water temperature of the warm and cold acclimation groups was changed by 1 °C per day until desired temperature was reached and then held at that temperature for a minimum of 8 weeks before experimentation. The photoperiod for the cold acclimated animals was changed to 8 hr light: 16 hr dark cycle to simulate winter (Graham and Farrell, 1989).

Table 1. Body mass, heart mass and outflow tract (OFT) mass for cold-acclimated (5 °C), control (10 °C) and warm-acclimated (18 °C) rainbow trout.

Thermal remodelling of the fish outflow tract

Tissue processing Fish were killed by a blow to the head followed by severance of the spinal cord and destruction of the brain. The heart was excised, rinsed in phosphate buffered saline and weighed. The OFT was removed from the ventricle, weighed, and bisected down the sagittal plane with one half snap frozen in OCT (Thermo Fisher Scientific, Waltham, MA, USA) by immersion in liquid nitrogen cooled 2-methylbutane (Sigma-Aldrich, St. Louis, MO, USA) and stored at -80 °C. The other half was fixed in 10 % neutral buffered formalin solution (Sigma-Aldrich, St. Louis, MO, USA) and embedded in paraffin wax so that sections would be cut in the transverse/axial plane. As thickness of the outer media and morphology of the inner media are not consistent throughout the length of the bulbus we were careful to section only the middle 50 % of the bulbus for histological comparison (Braun et al., 2003a).

Connective Tissue Histology Fibrillar collagen and elastin content were analysed semi-quantitatively following the methodology of Keen et al. 2016. Briefly, formalin-fixed and paraffin embedded tissue samples were sectioned at 5 μm (Leica CM3050S cryostat, Leica, Wetzlar, Germany) and mounted onto glass slides (Super frost plus, Thermo Fisher Scientific, Waltham, MA, USA). Serial sections from each sample were stained with picro-sirus red for collagen (Junqueira et al., 1979) and Miller’s elastic stain for elastin (Miller, 1971). Picro-sirus red images were quantified using polarised light microscopy and Miller’s elastic images were quantified using bright-field microscopy (Leica, Wetzlar, Germany). Mean fibrillar collagen and elastin contents were expressed as a percentage of total tissue cross-sectional area, excluding the epicardial surface, determined using ImageJ (Schneider et al., 2012). Quantitative analyses of collagen orientation (coherency) were conducted on picro-sirus red stained tissue using ImageJ with the OrientationJ plug-in, following previously published methodology (McConnell et al., 2016; Rezakhaniha et al., 2012). Three tissue sections were considered for each individual to ensure consistency in measurements. On each tissue section 3 separate image montages were taken along transects across the full diameter of the cross section. All histological analysis was conducted blind to the acclimation group and in all cases these tissue sections were taken from the central 50% of the OFT.

MMP gelatin zymography To characterize the abundance and activation of specific MMPs we used sodium dodecyl sulfate polyacrylamide electrophoresis (SDS-PAGE)-based gelatin zymography. Snap frozen OFT tissue was rinsed with phosphate buffered saline, then protein was extracted in ten volumes per wet weight of 0.05% Brij-35, 10 mM CaCl2, 50 mM Tris-HCl pH7.4 on ice, using three 10 s bursts of an MSE Soniprep150 sonicator (exponential probe, 10 µm amplitude). Extracts were cleared by centrifugation at 10,000g for 10 min and protein content was determined using the Bradford assay with bovine serum albumin as a standard. Equal quantities of protein (1 g/lane) were analysed by gelatin SDS-PAGE, as described by

Thermal remodelling of the fish outflow tract

Lødemel et al. 2005 (PMID: 15050523). Conditioned media from HepG2 cell cultures (100 ng protein/lane) and recombinant active human MMP-2 (1 ng protein/lane, Millipore) were used as positive controls. The abundance of each gelatinase band was measured using the ‘Gel’ function of ImageJ.

In situ MMP gelatin zymography The activity of endogenous MMP gelatinase was semi-quantitatively analysed by in situ zymography of tissue cryosections, following the methodology of Akhtar et al. 2014 (Akhtar et al., 2014). Frozen tissue was sectioned at 10 μm (Leica CM3050S cryostat, Leica, Wetzlar, Germany) and mounted onto glass slides (Super frost plus, Thermo Fisher Scientific, Waltham, MA, USA). Low temperature gelling agarose (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in phosphate buffered saline (to a final concentration of 10 mg ml-1) in an 80 °C water bath and then cooled to 37 °C. DQ gelatin (porcine; Invitrogen, Thermo Fisher Scientific,

-1 Waltham, MA, USA) was dissolved in dH2O (to a concentration of 1 mg ml ) and added to the agarose solution so that it was diluted 1:10. Lastly, 4’, 6’-diamidino-2-phenylindole (DAPI) was added to the agarose/DQ gelatin mixture (at a concentration of 1 μg ml-1). During this time tissue sections were brought to room temperature and washed in PBS to remove excess OCT. Approximately 40 μl of agarose/DAPI/DQ gelatin was added to each tissue section and a coverslip placed on the slide to ensure even film thickness across the sample section. All samples were incubated in the dark for 1hr at 4 °C and then 18 hrs at room temperature. Following incubation, the samples were visualized, using a fluorescent microscope using a green filter (Leica, Wetzlar, Germany), and imaged immediately. To remove any effect of tissue auto-fluorescence, negative control slides were used to determine the microscope settings for each section. Three tissue sections were considered for each individual to ensure consistency in measurements. On each tissue section 3 separate image montages were taken along transects across the full diameter of the cross section. Following background subtraction, mean fluorescence intensity was calculated for each image, analysed using image J. All histological analysis was conducted blind to the acclimation group and in all cases these tissue sections were taken from the central 50% of the OFT.

Ex vivo passive pressure-volume curves Pressure-volume curves were generated following the methodology of Keen et al., 2016. The intact isolated heart was placed into an organ bath containing Ringers solution [(in mM) 150

Thermal remodelling of the fish outflow tract

An atraumatic clamp was placed at the end of the bulbus, where it connects to the ventral aorta, making the OFT a sealed chamber with the cannula inside. The cannula was connected to a syringe pump (INFORS AG, Bottmingen, CHE), in series with a pressure transducer, containing 10 ± 1 °C Ringer solution with BDM and a small amount of blue food colouring (Silverspoon, London, UK). The pressure transducer was calibrated daily against a static water column and recorded at 1000 Hz (Chart5, PowerLab, ADI Instruments, Dunedin, New Zealand). Ringer solution with BDM was pumped into the OFT at 0.05 ml min-1 until maximum volume was achieved, determined by visual leak of the saline containing blue dye and a drop in the pressure trace.

Statistical analyses Chamber filling volume was calculated from filling time by the equation:

0.05 푣표푙푢푚푒 (푚푙) = 푡푖푚푒 (µs) × × 1000 60

The effect of temperature acclimation on the pressure-volume relationship of the OFT was assessed by a general linear model (GLM) with pressure as the dependent variable, volume and acclimation group as fixed factors and body mass as the covariate, with a Tukey post hoc test for differences between groups using R. The model was performed on data below 10 kPa, which approximates the maximum physiological pressures experienced by this species (Forster and Farrell, 1994; Seth et al., 2014). Chamber compliance (the change in volume for a given change in pressure standardised to body mass; ml kg-1 kPa-1) was calculated as the slope of the pressure-volume curve within the normal physiological range of this species (between 2 and 5 kPa; Seth et al., 2014). Chamber distensibility (the fold change in compliance) was calculated as the chamber compliance normalized to the chamber volume at onset of the physiological pressure range (i.e. 2 kPa). Differences in chamber compliance and distensibility, and collagen and elastin tissue content were assessed by GLM with Holm-Sidak post hoc test for differences between groups using Prism v6 (GraphPad, San Jose, CA, USA). For all analyses significance was considered to be P < 0.05. Values are presented as mean ±

S. E. throughout unless otherwise stated. Statistical details are provided in the figure legends.

Thermal remodelling of the fish outflow tract

RESULTS

Morphology There was no effect of thermal acclimation on body mass, heart mass, OFT mass or relative masses (Table 1.)

Thermal remodelling of ex vivo chamber compliance To assess the functional differences in the passive properties of the fish OFT following thermal acclimation we generated ex vivo passive filling curves from freshly isolated intact OFT, treated with BDM, at a common test temperature of 10 °C. The mean pressure-volume trace from each group showed the same r-shaped profile as previously described within the in vivo physiological range for teleost fish (Figure 1A; Braun et al. 2003a). Thermal acclimation altered

2 the pressure-volume relationship during passive filling of the OFT (R = 0.68, F2, 8971 = 160.95, P < 0.001) showing increased stiffness following cold acclimation, and increased compliance after warm acclimation, compared to the controls (t ratio = 26.81; Figure 1A). As a result, the characteristic steep pressure increase at high volumes, as the stiff outer collagenous wall was incorporated, (Braun et al., 2003a) began at a lower volume following cold acclimation and a higher volume following warm acclimation. Analysis of chamber volume at the lower boundary (2 kPa) and the upper boundary (5 kPa) of the physiological pressure range revealed a significantly greater chamber volume following warm acclimation, compared with control and cold-acclimated animals. Chamber compliance within the physiological pressure range was increased by 1.9-fold following warm acclimation compared to control and cold-acclimated

2 animals (R = 0.60, F2, 10 = 7.63, P < 0.001; Figure 1B). However, there was no effect on chamber distensibilty with temperature acclimation (Figure 1C).

Thermal remodelling of the fish outflow tract

Figure 1. Compliance and distensibility of the fish outflow tract (OFT). (A) mean passive pressure-volume relationships generated from the OFT for cold (5 °C; blue), control (10 °C; green) and warm (18 °C; red) acclimated rainbow trout (n = 8). Curves are shown within the physiological pressure range for the chamber (< 10 kPa) and represent mean data for n > 3. Pressure has been standardized to start at 0 kPa for graphical representation. Significant differences in the pressure-volume relationships were assessed by GLM with volume as the dependent variable, acclimation group and pressure as the fixed factors and chamber mass as the covariate. (B) maximum compliance (i.e. the change in volume for a given change in pressure) and (C) maximum distensibility (i.e. fold change in compliance) of the chamber for cold (5 °C; blue), control (10 °C; green) and warm (18 °C; red) acclimated rainbow trout (n = 8). Significant differences in chamber compliance and distensibility were assessed by separate GLM with a Holm-Sidak post hoc test for significance between groups (P < 0.05). Values are presented as mean ± S.E. and significance is indicated by dissimilar letters.

Thermal remodelling of the fish outflow tract

Connective tissue content Cardiac remodeling due to high pressure or volume is often associated with remodelling of the cardiac ECM in both fish and mammals (Chapman et al., 1990; Keen et al., 2016). We stained serial bulbus sections with picro-sirus red and Miller’s elastic stain to assess fibrillar collagen elastin content, respectively (Figure 2). Figure 2A shows a representative micrograph of sectioned OFT stained with picro-sirus red, under bright-field light and Figure 2B shows the same section imaged under plane polarised light for semi-quantification. Despite a trend towards higher collagen deposition following cold acclimation (Figure 2C) no significant difference was found (P = 0.08). Figure 2D shows a representative micrograph of sectioned OFT stained with Miller’s elastic stain visualized in bright-field light. Elastin content of the OFT was not affected by thermal acclimation (Figure 2E). However, the combination of these changes in connective tissue content lead to a collagen to elastin ratio that was 1.4-fold higher in cold- and a 2.3-fold lower in warm-acclimated animals, compared to controls (Figure 2F; P < 0.05).

Figure 2. Connective tissue in the fish outflow tract (OFT). A representative (A) bright-field and (B) polarised light micrograph of fish OFT, stained with picro-sirus red. (C) semiquantitative analysis of fibrillar collagen content, with picro-sirus red, expressed as a percentage of total tissue mass. (D) a representative micrograph of fish OFT stained with Miller’s elastic stain. (E) semi-quantitative analysis of fibrillary elastin content, with Miller’s elastic stain, expressed as a percentage of total tissue mass. (F) the collagen to elastin ratio of the fish OFT cold (5 °C; blue), control (10 °C; green) and warm (18 °C; red) acclimated rainbow trout (n = 8). Significant differences in collagen and elastin content were assessed by separate GLM with a Holm-Sidak post hoc test for significance between groups and significance is indicated by dissimilar letters (P < 0.05). Values are presented as mean ± S.E

Thermal remodelling of the fish outflow tract

Coherency of fibrillar collagen The organization of collagen fibres can effect overall tissue stiffness (McConnell et al., 2016). We probed whether collagen alignment in the bulbus arteriosus was influenced by thermal acclimation by quantitatively measuring collagen coherency. Figure 3 shows a representative OFT section, stained with picro-sirus red and imaged under polarized light, assessed using OrientationJ for cold (Figure 3A), control (Figure 3B) and warm (Figure 3C) acclimated fish. Despite a trend towards increased coherency with cold temperature the coherency of organized fibrillar collagen was not significantly altered by temperature acclimation (P = 0.1; Figure 3D).

Figure 3. Fibrillar collagen orientation in the fish outflow tract (OFT). Representative polarized light micrographs for (A) cold-, (B) control and (C) warm- acclimated OFT sections, stained with picro-sirus red, assessed using OrientationJ. Fibre orientation is shown by colour, with fibres of the same colour showing coherency. (D) quantification of coherency of organized fibrillar collagen for cold (5 °C; blue), control (10 °C; green) and warm (18 °C; red) acclimated rainbow trout (n = 8). Significant differences in collagen coherency were assessed by separate GLM with a Holm-Sidak post hoc test for significance between groups (P < 0.05). Values are presented as mean ± interquartile range and range.

Thermal remodelling of the fish outflow tract

Characterization and endogenous gelatinase activity of MMPs Extracellular matrix content is a balance of deposition and degradation, with degradation regulated by matrix metalloproteinases (MMPs) (Nagase et al., 2006; Pedersen et al., 2015). MMPs are initially synthesized as inactive pro-forms, the pro-domain must be removed to activate the enzyme (Löffek et al., 2011). Gelatin SDS-PAGE zymography showed the presence of MMPs with molecular weights matching those of human proMMP-9, activated MMP-9 and proMMP-2 and activated MMP-2 in trout OFT tissue (Figure 4A). In addition, an 83 kDa gelatinase was detected in some fish, as has been previously described in the heart Atlantic cod (Gadus morhua) (Lødemel et al., 2004). Activated MMP-2 abundance was 39.3- fold higher than controls following warm acclimation. Accordingly, the ratio of proMMP-2 to MMP-2, hence level of MMP-2 activation, was 2-fold higher in warm tissue than controls and 6.2-fold lower in cold-acclimated tissue than controls.

In addition to specific MMP abundance, we investigated the potential effect of endogenous MMP gelatinase activity on remodelling of the fish bulbus arteriosus extracellular matrix by in situ zymography of tissue cryosections. Figure 5A, B & C show representative fluorescent micrographs of cold-acclimated, control and warm-acclimated fish OFT that has been treated with DQ gelatin, demonstrating the localised activity of MMPs in tissue sections. Figure Ai, Bi and Ci show the same section with the blue fluorescence micrograph overlaid to show counterstaining with DAPI, which highlights cell nuclei. Semi-quantitative analysis of the fluorescent micrographs revealed a 4.1-fold increase in MMP activity in the warm-acclimated group, compared to controls (Figure 5D).

Thermal remodelling of the fish outflow tract

Figure 4. Characterization of specific matrix metalloproteinase (MMP) activity in the fish outflow tract (OFT) by gelatin SDS-PAGE zymography. (A) Coomassie R250 stained zymogram, indicating the relative molecular weights and abundances of gelatinases in OFT extracts from cold (5 °C; blue), warm (18 °C; red) and control (10 °C; green) acclimated rainbow trout (n = 4). The positions of pro-MMP-2 and MMP-2 are indicated by arrows on the left-hand side of the gel. (B) abundance of MMP-2 and (C) the ratio of proMMP-2 to MMP-2 abundance for the fish OFT cold (5 °C; blue), control (10 °C; green) and warm (18 °C; red) acclimated rainbow trout (n = 4). Significant differences in MMP abundance were assessed by GLM with a Holm-Sidak post hoc test for significance between groups, indicated by dissimilar letters (P < 0.05). Values are presented as mean ± S.E.

Thermal remodelling of the fish outflow tract

Figure 5. Endogenous matrix metalloproteinase (MMP) activity. (A) semi-quantitative analysis of endogenous MMP activity by in situ gelatinase zymography of fish OFT tissue sections for cold (5 °C; blue), control (10 °C; green) and warm (18 °C; red) acclimated rainbow trout (n = 8). Representative fluorescent micrographs for (B) cold-, (C) control and (D) warm-acclimated OFT imaged with a green filter to show gelatinase activity. The same sections were imaged in with a blue filter and the image imposed to show DAPI fluorescence for (Bi) cold-, (Ci) and (Di) warm-acclimated fish. Significant differences in collagen coherency were assessed by separate GLM with a Holm-Sidak post hoc test for significance between groups and shown by dissimilar letters (P < 0.05). Values are presented as mean ± interquartile range and range.

Thermal remodelling of the fish outflow tract

DISCUSSION

For ectothermic species acute and chronic temperature changes present a significant challenge, altering cardiac load. Therefore, many fish that inhabit temperate waters undergo a seasonal, compensatory cardiac remodelling response to maintain appropriate heart function. Here, we assess the effect of long-term temperature change on the often overlooked fish OFT. We show that chronic cooling causes a stiffening of the outflow tract, which can be attributed to an increase in the collagen to elastin ratio of the chamber. The opposite response is true following chronic warming, with an increase in chamber compliance which is likely due to a decrease in the collagen to elastin ratio of the chamber. The decreases in collagen content are associated with an increase in the activity of gelatinase MMPs with chronic warming. Similarly, MMP activity is reduced following chronic cooling, which is associated with an increase in collagen deposition. Our results show that despite the passive and elastic nature of the bulbus arteriosus, it is subject to a seasonal structural and functional remodelling. Increases in bulbus stiffness may cause increases in cardiac afterload contributing to the ventricular hypertrophy experienced by salmonids exposed to low temperature.

Temperature associated changes in the compliance of the fish outflow tract For the fish OFT to function as a windkessel it must be highly compliant and have specialized inflation properties (Braun et al., 2003a). At high stroke volume, the chamber needs to expand to accommodate the increased volume of blood while maintaining a relatively constant outflow pressure. Equally, at low stroke volumes the bulbus must maintain an appropriate outflow pressure for blood to flow through the gill capillary network and circulate the body (Braun et al., 2003b). For the bulbus to behave similarly at high and low cardiac output, it is most compliant near systolic pressure giving the chamber its characteristic ‘r-shaped’ pressure- volume curve compared to the ‘J-shaped’ stress-strain curve of an artery (Braun et al., 2003a; Braun et al., 2003b; Clark and Rodnick, 1999). Under the law of Laplace, the narrow bulbar lumen requires a steep initial pressure increase to expand (Braun et al., 2003a; Braun et al., 2003b). Once in the physiological range, the high compliance of the chamber causes pressure change to plateau (Braun et al., 2003a; Jones et al., 1993).

Here, temperature acclimation altered the pressure-volume relationship in the bulbus arteriosus, with an increase in chamber stiffness following chronic cold and an increase in chamber compliance following chronic warm. Pressure-volume relationships have previously been generated in the fish OFT (Braun et al., 2003a; Clark and Rodnick, 1999; Forster and Farrell, 1994; Seth et al., 2014); however, this is the first time they have been generated to probe temperature-dependent remodelling. Clark and Rodnick (1999) found no difference in bulbus compliance or distensibility with ventricular hypertrophy over a range of physiological pressures, suggesting that changes in cardiac pressure are not sufficient to trigger remodelling alone. An implication of our finding may be recruitment of the stiff collagen fibres in the

Thermal remodelling of the fish outflow tract adventitia at a lower volume (Braun et al., 2003a). These fibres resist further inflation of the bulbus, but cause steep increases in intra-lumen pressure. Indeed, in our study, at the highest volumes within the physiological pressure range, intra-luminal pressure appeared to rise steeply and this occurred at a lower volume following cold acclimation. However, as the pressure-volume curves were generated ex vivo this limits our interpretation. Braun et al. (2003b) suggest that the position of the bulbus arteriosus in the pericardium makes this third pressure phase of the inflation curve unlikely in vivo. It is likely that the semi-rigid fish pericardium would limit the extent to which the bulbus can expand (Braun et al., 2003a; Braun et al., 2003b), however, it is possible that this restriction causes the final pressure rise (Braun et al., 2003a). In addition, if the bulbus were to swell extensively it may interfere with the atrial function, which would reduce cardiac output and, therefore, bulbus filling (Braun et al., 2003b). Indeed, the bulbus preparations in this study only experienced the high pressure phase at pressures approaching 10 kPa, which is the upper limit of the physiological pressure range suggested by Forster and Farrell (1994) and above the physiological pressure range suggested by Seth et al. (2014), of 2-5 kPa.

Structural remodelling of the fish OFT following temperature acclimation The exceptional compliance and low tissue modulus of the bulbus arteriosus is due to the structure and organization of the chamber’s connective tissue. The elastic fibres that make up ~70% of the chamber are not organized into a continuous elastic lamellae sheet, but a loose fibrillar arrangement of low hydrophobicity, high solubility elastin (Braun et al., 2003a; Braun et al., 2003b; Icardo, 2006; Icardo et al., 2000; Licht and Harris, 1973; Serafini-Fracassini et al., 1978; Watson and Cobb, 1979). This arrangement is distinct from that in mammalian arteries, where it is found as an amorphous elastin component and elastin-associated glycoprotein microfibrils (Braun et al., 2003a; Isokawa et al., 1990). In addition, the fish elastin is chemically and genetically unique (Chow et al., 1989; Serafini-Fracassini et al., 1978; Spina et al., 1979). A greater density of polar amino acids gives a low hydrophobic index which is likely to alter elastic recoil and reduce modulus (Braun et al., 2003a). Here, we saw an increased elastin to collagen ratio following warm acclimation which gave an increase in overall chamber compliance. At warm temperatures the trout heart operates at low or routine stoke volumes. Therefore, it is likely that the important part of the pressure-volume curve is at the low volumes, where high compliance is needed to maintain a physiological pressure range.

Further, a direct (i.e. Q10) effect of warming the fish heart in vivo is an increased heart rate. It is possible a more complaint bulbus aids rate of inflation to accommodate high heart rate.

To maximize bulbar compliance there is very little collagen in the media or along the luminal surface. However, the outer media integrates layers of collagen and the outer adventitia is almost entirely collagen (Benjamin et al., 1983; Braun et al., 2003a; Bushnell et al., 1992; Icardo, 2006) (Figure 2). The collagen provides strength to resist radial and/or longitudinal haemodynamic strain on the bulbar wall (Braun et al., 2003a; Icardo et al., 1999a; Icardo et

Thermal remodelling of the fish outflow tract al., 1999b; Icardo et al., 2000; Priede, 1976). With cold acclimation we found a higher collagen content which increased stiffness of the whole chamber. The increased collagen deposition may be a cardio-protective measure to prevent haemodynamic damage to the bulbus wall under the increased strain of pumping a high stroke volume of viscous blood. With high stoke volume it is likely that the bulbus is operating in the high volume range of the pressure-volume curve. At this point the wall has to be strong enough to withstand high levels of inflation. Fibrillar collagen monomers form fibrils which function as supra-molecular assemblies and in turn form fibres (McConnell et al., 2016). Increased alignment of fibrillar collagen can be associated with increased in overall tissue stiffness (McConnell et al., 2016). Although we saw a slight trend towards increased coherency of fibrillar collagen we cannot attribute changes in compliance of the OFT to collagen alignment.

The changes in collagen content are supported by MMP activity in the tissue. Collagen content of tissue is a balance of synthesis and degradation, regulated by gelatinase MMPs (Nagase et al., 2006). Following chronic warm we found an increase in MMP-2 activity which translated to an increased proMMP-2 to MMP-2 ratio. As MMPs degrade collagen, increased MMP activity is associated with low tissue collagen content. Accordingly, following chronic cooling we found a decrease in MMP activity and a decrease in the proMMP-2 to MMP-2 ratio, which is associated with low levels of collagen degradation. Our gelatin zymography also identified an 83 kDa gelatinase. Lødemel et al. (2004) found a gelatinase with this molecular weight in the heart of Atlantic cod (Gadus morhua); however, they suggested that the band may actually be composed of several gelatin degrading enzymes due to its broad appearance. Nevertheless, it has been shown to have a similar activity to MMP-9 which directly digests gelatin (Kubota et al., 2003; Lødemel et al., 2004).

Implications of changes in compliance of the fish OFT on the ventricle The vertebrate heart is sensitive to load, with chronic changes in cardiac pressure or volume potentially initiating changes in size, form and function (Opie et al., 2006). The ectothermic nature of fish means that temperature change directly alters cardiac function and blood viscosity (Graham and Farrell, 1989; Klaiman et al., 2014). The fish heart will, therefore, remodel with chronic temperature change and many species that inhabit temperate regions which experience a seasonal remodelling of the heart and cardiovascular system. As rainbow trout remain active throughout the winter this remodelling response has generally been viewed as an adaptive response to increased volume load of pumping cold viscous blood. However, recent studies have suggested aspects of the remodelling response are more consistent with a pressure overload, which is generally maladaptive (Keen et al., 2016; Klaiman et al., 2011). Clark and Rodnick (1999) investigated the remodelling phenotype of pressure and volume overloads in the fish heart, showing that mean arterial pressure increase relative ventricular mass (Clark and Rodnick, 1999). Importantly, following increased afterload pressure in fish, ventricular hypertrophy is driven by an increase in myocyte cross-sectional area, as it is in

Thermal remodelling of the fish outflow tract mammals exposed to increased afterload pressure (Clark and Rodnick, 1998; Clark and Rodnick, 1999; Grossman et al., 1975).

Stiffening of the mammalian aorta is often associated with increased cardiac afterload pressure and can cause ventricular fibrosis (Girerd et al., 1991). Under normal circumstances the fish bulbus arteriosus reduces afterload pressure (Bushnell et al., 1992; Priede, 1976; Seth et al., 2014), however, reduced compliance and/or distensibility relative to ventricular output may have implications on the fish heart by increasing cardiac afterload (Clark and Rodnick, 1999). Reduced bulbar compliance by parasite infection or cholecystokinin increases cardiac afterload leading to ventricular hypertrophy and fibrosis (Coleman, 1993; Seth et al., 2014). The effects of chronic cooling on the fish ventricle are hypertrophy by increased mycoyte diameter and fibrosis (Keen et al., 2016; Klaiman et al., 2011; Vornanen et al., 2005), both of which are consistent with a remodelling due, at least in part, to cardiac pressure overload (Bernardo et al., 2010; Clark and Rodnick, 1998; Clark and Rodnick, 1999; Grossman et al., 1975).

Perspectives and significance We have previously shown that cold acclimation increases stiffness of contractile chambers of the fish heart, with a reduced wall stiffness following warm acclimation. However, it was unclear if cold acclimation would affect the highly specialized and compliant OFT, which is made almost entirely of connective tissue. Here, we show that cold acclimation is associated with an increase in stiffness of the OFT, with an increase in the collagen to elastin ratio of connective tissue within the bulbus arteriosus. Warm acclimation was associated with the opposite response; an increase in compliance and a decrease in the collagen to elastin ratio. This study adds to our previous work, suggesting a global increase in myocardial stiffness following cold acclimation and a global increase in myocardial compliance following warm acclimation. Finally, we suggest that increased stiffness of the OFT following cold acclimation contributes to the well documented cold-induced ventricular hypertrophy experienced by fish. It appears that cold-dependent remodelling of the fish ventricle serves both cardio-protection from increased cardiac preload and afterload as well as compensation for reduced contractile function and increased cardiac afterload.

Acknowledgements

We thank Drs Michael Sherratt and James McConnell for helpful discussions regarding experimental procedures and interpretation of results, particularly with regard to collagen alignment and in situ gelatin zymography. We thank Russell Craddock for helpful advice with the in situ gelatin zymography protocol and analysis. Histology was performed in the University of Manchester Histology Facility.

Thermal remodelling of the fish outflow tract

Competing interests The authors declare no competing interests

Author contributions

A.N.K., P.G and H.A.S. are responsible for the concept and design of the research. A.N.K. and J.J.M. performed experiments and analysed data in H.A.S. and J.J.M. laboratories. A.N.K., J.J.M and H.A.S. interpreted the results of the experiments. A.N.K. drafted the manuscript. A.N.K., J.J.M., P.G., and H.A.S. revised and edited the manuscript. A.N.K., J.J.M., P.G., and H.A.S. approve the final version of the manuscript submitted for publication.

Funding A.N.K is funded by a BBSRC Doctoral training partnership. The Shiels lab is supported by the Leverhulme Trust (240613).

REFERENCES

Akhtar, R., Cruickshank, J. K., Zhao, X., Walton, L. A., Gardiner, N. J., Barrett, S. D., Graham, H. K., Derby, B. and Sherratt, M. J. (2014). Localized micro- and nano-scale remodelling in the diabetic aorta. Acta Biomaterialia 10, 4843-4851.

Benjamin, M., Norman, D., Santer, R. M. and Scarborough, D. (1983). Histological, histochemical and ultrastructural studies on the bulbus arteriosus of the sticklebacks, Gasterosteus aculeatus and Pungitius pungitius (Pisces: Teleostei). Journal of Zoology 200, 325-346.

Bernardo, B. C., Weeks, K. L., Pretorius, L. and McMullen, J. R. (2010). Molecular distinction between physiological and pathological cardiac hypertrophy: Experimental findings and therapeutic strategies. Pharmacology & Therapeutics 128, 191-227.

Braun, M. H., Brill, R. W., Gosline, J. M. and Jones, D. R. (2003a). Form and function of the bulbus arteriosus in yellowfin tuna (Thunnus albacares), bigeye tuna (Thunnus obesus) and blue marlin (Makaira nigricans): static properties. Journal of Experimental Biology 206, 3311-3326.

Braun, M. H., Brill, R. W., Gosline, J. M. and Jones, D. R. (2003b). Form and function of the bulbus arteriosus in yellowfin tuna (Thunnus albacares): dynamic properties. Journal of Experimental Biology 206.

Bushnell, P. G., Jones, D. R. and Farrell, A. P. (1992). The arterial system. New York, NY, USA: Academic Press, Inc.

Chapman, D., Weber, K. T. and Eghbali, M. (1990). Regulation of fibrillar collagen type-I and type-III and basement-membrane type-IV collagen gene-expression in pressure overloaded rat myocardium. Circulation Research 67, 787-794.

Chow, M., Boyd, C. D., Iruela-Arispe, M.-L., Wrenn, D. S., Mecham, R. and Helene Sage, E. (1989). Characterization of elastin protein and mRNA from salmonid fish (Oncorhynchus kisutch). Comparative Biochemistry and Physiology Part B: Comparative Biochemistry 93, 835-845.

Thermal remodelling of the fish outflow tract

Clark, R. J. and Rodnick, K. J. (1998). Morphometric and biochemical characteristics of ventricular hypertrophy in male rainbow trout (Oncorhynchus mykiss). Journal of Experimental Biology 201, 1541-1552.

Clark, R. J. and Rodnick, K. J. (1999). Pressure and volume overloads are associated with ventricular hypertrophy in male rainbow trout. American Journal of Physiology- Regulatory Integrative and Comparative Physiology 277, R938-R946.

Coleman, F. C. (1993). Morphological and physiological consequences of parasites encysted in the bulbus-arteriosus of an estuarine fish, the sheepshead minnow, Cyprinodon variegatus. Journal of Parasitology 79, 247-254.

Convertino, V. A., Mack, G. W. and Nadel, E. R. (1991). Elevated central venous- pressure - a consequence of exercise training-induced hypervolemia. American Journal of Physiology 260, R273-R277.

Duran, A. C., Fernandez, B., Grimes, A. C., Rodriguez, C., Arque, J. M. and Sans- Coma, V. (2008). Chondrichthyans have a bulbus arteriosus at the arterial pole of the heart: morphological and evolutionary implications. Journal of Anatomy 213, 597-606.

Farrell, A. P. (1979). The wind-kessel effect of the bulbus arteriosus in trout. Journal of Experimental Zoology 209.

Farrell, A. P. and Jones, D. R. (1992). The Heart. In Fish Physiology, Volume XII, Part A The Cardiovascular System.

Forster, M. E. and Farrell, A. P. (1994). The volumes of the chambers of the trout heart. Comparative Biochemistry and Physiology a-Physiology 109, 127-132.

Franklin, C. E. and Davie, P. S. (1992). Sexual maturity can double heart mass and cardiac power output in male rainbow trout. Journal of Experimental Biology 171, 139-148.

Girerd, X., Laurent, S., Pannier, B., Asmar, R. and Safar, M. (1991). Arterial distensibility and left ventricular hypertrophy in patients with sustained essential hypertension. American Heart Journal 122, 1210-1214.

Graham, M. S. and Farrell, A. P. (1989). The effect of temperature-acclimation and adrenaline on the performance of a perfused trout heart. Physiological Zoology 62, 38-61.

Graham, M. S. and Fletcher, G. L. (1983). Blood and plasma viscosity of winter flounder: influence of temperature, red-cell concentration, and shear rate. Canadian Journal of Zoology-Revue Canadienne De Zoologie 61, 2344-2350.

Grossman, W., Jones, D. and McLaurin, L. P. (1975). Wall stress and patterns of hypertrophy in the human left ventricle. Journal of Clinical Investigation 56, 56-64.

Icardo, J. M. (2006). Conus arteriosus of the teleost heart: Dismissed, but not missed. Anatomical Record Part a-Discoveries in Molecular Cellular and Evolutionary Biology 288A.

Icardo, J. M. and Colvee, E. (2011). The Atrioventricular Region of the Teleost Heart. A Distinct Heart Segment. Anatomical Record-Advances in Integrative Anatomy and Evolutionary Biology 294.

Icardo, J. M., Colvee, E., Cerra, M. C. and Tota, B. (1999a). Bulbus arteriosus of the Antarctic teleosts. I. The white-blooded Chionodraco hamatus. Anatomical Record 254.

Icardo, J. M., Colvee, E., Cerra, M. C. and Tota, B. (1999b). Bulbus arteriosus of the antarctic teleosts. II. The red-blooded Trematomus bernacchii. Anatomical Record 256.

Thermal remodelling of the fish outflow tract

Icardo, J. M., Colvee, E., Cerra, M. C. and Tota, B. (2000). Light and electron microscopy of the bulbus arteriosus of the European eel (Anguilla anguilla). Cells Tissues Organs 167, 184-198.

Isokawa, K., Takagi, M. and Toda, Y. (1990). Ultrastructural and cytochemical study of elastic fibers in the ventral aorta of a teleost, Anguilla japonica. The Anatomical Record 226, 18-26.

Jones, D. R., Brill, R. W. and Bushnell, P. G. (1993). Ventricular and arterial dynamics of anesthetized and swimming tuna. Journal of Experimental Biology 182, 97-112.

Jones, D. R., Langille, B. L., Randall, D. J. and Shelton, G. (1974). Blood-flow in dorsal and ventral aortas of cod, Gadus morhua. American Journal of Physiology 226, 90-95.

Junqueira, L. C. U., Bignolas, G. and Brentani, R. R. (1979). Picrosirius staining plus polarization microscopy, a specific method for collagen detection in tissue-sections. Histochemical Journal 11, 447-455.

Keen, A. N., Fenna, A. J., McConnell, J. C., Sherratt, M. J., Gardner, P. and Shiels, H. A. (2016). The dynamic nature of hypertrophic and fibrotic remodelling in the fish ventricle. Frontiers in Physiology - Intergrative Physiology 6.

Klaiman, J. M., Fenna, A. J., Shiels, H. A., Macri, J. and Gillis, T. E. (2011). Cardiac Remodeling in Fish: Strategies to Maintain Heart Function during Temperature Change. Plos One 6, e24464.

Klaiman, J. M., Pyle, W. G. and Gillis, T. E. (2014). Cold acclimation increases cardiac myofilament function and ventricular pressure generation in trout. Journal of Experimental Biology 217, 4132-4140.

Kubota, M., Kinoshita, M., Takeuchi, K., Kubota, S., Toyohara, H. and Sakaguchi, M. (2003). Solubilization of type I collagen from fish muscle connective tissue by matrix metalloproteinase-9 at chilled temperature. Fisheries Science 69, 1053-1059.

Licht, J. H. and Harris, W. S. (1973). The structure, composition and elastic properties of teleost bulbus arteriousus in carp, Cyprinus carpio. Comparative Biochemistry and Physiology 46, 699-704.

Lødemel, J. B., Mæhre, H. K., Winberg, J.-O. and Olsen, R. L. (2004). Tissue distribution, inhibition and activation of gelatinolytic activities in Atlantic cod (Gadus morhua). Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology 137, 363-371.

Löffek, S., Schilling, O. and Franzke, C.-W. (2011). Biological role of matrix metalloproteinases: a critical balance. European Respiratory Journal 38, 191-208.

McConnell, J. C., O’Connell, O. V., Brennan, K., Weiping, L., Howe, M., Joseph, L., Knight, D., O’Cualain, R., Lim, Y., Leek, A. et al. (2016). Increased peri-ductal collagen micro-organization may contribute to raised mammographic density. Breast Cancer Research 18, 1-17.

Miller, P. J. (1971). Elastin stain. Medical Laboratory Technology 28, 148-149.

Nagase, H., Visse, R. and Murphy, G. (2006). Structure and function of matrix metalloproteinases and TIMPs. Cardiovascular Research 69, 562-573.

Opie, L. H., Commerford, P. J., Gersh, B. J. and Pfeffer, M. A. (2006). Controversies in Cardiology 4 - Controversies in ventricular remodelling. Lancet 367, 356-367.

Thermal remodelling of the fish outflow tract

Pedersen, M. E., Vuong, T. T., Rønning, S. B. and Kolset, S. O. (2015). Matrix metalloproteinases in fish biology and matrix turnover. Matrix Biology 44–46, 86-93.

Priede, I. G. (1976). Functional morphology of bulbus arteriosus of rainbow trout (Salmo gairdneri Richardson). Journal of Fish Biology 9, 209-216.

R-Core-Team. (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/.

Rezakhaniha, R., Agianniotis, A., Schrauwen, J. T. C., Griffa, A., Sage, D., Bouten, C. V. C., Vosse, F. N., Unser, M. and Stergiopulos, N. (2012). Experimental investigation of collagen waviness and orientation in the arterial adventitia using confocal laser scanning microscopy. Biomechanics and Modeling in Mechanobiology 11, 461-473.

Schneider, C. A., Rasband, W. S. and Eliceiri, K. W. (2012). NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9, 671-675.

Serafini-Fracassini, A., Field, J. M., Spina, M., Garbisa, S. and Stuart, R. J. (1978). The morphological organization and ultrastructure of elastin in arterial wall of trout (Salmo gairdneri) and salmon (Salmo salar). Journal of Ultrastructure Research 65, 1-12.

Seth, H., Axelsson, M. and Grans, A. (2014). The peptide hormone cholecystokinin modulates the tonus and compliance of the bulbus arteriosus and pre-branchial vessels of the rainbow trout (Oncorhynchus mykiss). Comparative Biochemistry and Physiology a-Molecular & Integrative Physiology 178, 18-23.

Seth, H., Gräns, A. and Axelsson, M. (2010). Cholecystokinin as a regulator of cardiac function and postprandial gastrointestinal blood flow in rainbow trout (Oncorhynchus mykiss). American Journal of Physiology - Regulatory, Integrative and Comparative Physiology 298, R1240-R1248.

Spina, M., Garbisa, S., Field, J. M. and Serafini-Fracassini, A. (1979). The salmonid elastic fibril. Archives of Biochemistry and Biophysics 192, 430-437.

Stevens, E. D., Shelton, G., Randall, D. J. and Bennion, G. R. (1972). Factors affecting arterial pressures and blood flow from heart in intact, unrestrained lingcod, Ophiodon elongatus. Comparative Biochemistry and Physiology 43, 681-695.

Vornanen, M., Hassinen, M., Koskinen, H. and Krasnov, A. (2005). Steady-state effects of temperature acclimation on the transcriptome of the rainbow trout heart. American Journal of Physiology-Regulatory Integrative and Comparative Physiology 289, R1177-1184.

Watson, A. D. and Cobb, J. L. S. (1979). A comparative study on the innervation and the vascularization of the bulbus arteriosus in teleost fish. Cell and Tissue Research 196, 337- 346.

Thermal remodelling of the fish outflow tract

Thermal remodelling of the ectothermic heart

7. Cardiovascular function, compliance and connective

tissue remodeling in the turtle, Trachemys scripta, following

thermal acclimation.

This chapter is presented as a reprint of the following paper published in

American Journal of Physiology – Integrative, regulatory and comparative physiology:

Adam Nicholas Keen, Holly A Shiels, Dane A Crossley. 2016 Cardiovascular function, compliance and connective tissue remodeling in the turtle, Trachemys scripta, following thermal acclimation. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology 311, R133-R143. DOI: 10.1152/ajpregu.00510.2015.

Page | 87

Thermal remodelling of the ectothermic heart

Page | 88

Cardiovascular function, compliance, and connective tissue remodeling in the turtle, Trachemys scripta, following thermal acclimation Adam N. Keen, Holly A. Shiels and Dane A. Crossley, 2nd Am J Physiol Regul Integr Comp Physiol 311:R133-R143, 2016. First published 13 April 2016; doi: 10.1152/ajpregu.00510.2015

You might find this additional info useful...

This article cites 63 articles, 31 of which you can access for free at: http://ajpregu.physiology.org/content/311/1/R133.full#ref-list-1 Updated information and services including high resolution figures, can be found at: http://ajpregu.physiology.org/content/311/1/R133.full Additional material and information about American Journal of Physiology - Regulatory, Integrative

and Comparative Physiology can be found at: Downloaded from http://www.the-aps.org/publications/ajpregu

This information is current as of July 20, 2016. http://ajpregu.physiology.org/ by guest on July 20, 2016

American Journal of Physiology - Regulatory, Integrative and Comparative Physiology publishes original investigations that illuminate normal or abnormal regulation and integration of physiological mechanisms at all levels of biological organization, ranging from molecules to humans, including clinical investigations. It is published 24 times a year (twice monthly) by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2016 the American Physiological Society. ESSN: 1522-1490. Visit our website at http://www.the-aps.org/. Am J Physiol Regul Integr Comp Physiol 311: R133–R143, 2016. First published April 13, 2016; doi:10.1152/ajpregu.00510.2015.

Cardiovascular function, compliance, and connective tissue remodeling in the turtle, Trachemys scripta, following thermal acclimation

Adam N. Keen,1 Holly A. Shiels,1 and Dane A. Crossley 2nd2 1Faculty of Life Sciences, University of Manchester, Manchester, United Kingdom; and 2Department of Biological Sciences, University of North Texas, Denton, Texas Submitted 4 December 2015; accepted in ﬁnal form 7 April 2016

Keen AN, Shiels HA, Crossley DA 2nd. Cardiovascular function, temperature on blood viscosity may alter hemodynamic load compliance, and connective tissue remodeling in the turtle, Trachemys on the heart (44, 56, 62). If prolonged, these physical and scripta, following thermal acclimation. Am J Physiol Regul Integr functional changes may trigger a dynamic remodeling of the Comp Physiol 311: R133–R143, 2016. First published April 13, 2016; passive and active properties of the cardiovascular system doi:10.1152/ajpregu.00510.2015.—Low temperature directly alters (62, 63).

cardiovascular physiology in freshwater turtles, causing bradycardia, Downloaded from arterial hypotension, and a reduction in systemic blood pressure. At Freshwater slider turtles (Tracheyms scripta) spend winter in the same time, blood viscosity and systemic resistance increase, as water in a state of periodic inactivity. Following cold acclima- does sensitivity to cardiac preload (e.g., via the Frank-Starling re- tion slider turtles exhibit a compensatory increase in cardiac sponse). However, the long-term effects of these seasonal responses muscle twitch force and maximal isometric force, as well as a on the cardiovascular system are unclear. We acclimated red-eared suppression of cholinergic inhibition, slower action potential slider turtles to a control temperature (25°C) or to chronic cold (5°C). upstroke, and longer action potential duration, accompanied by To differentiate the direct effects of temperature from a cold-induced

a dissipation of resting membrane potential (28, 51, 59, 62, 64). http://ajpregu.physiology.org/ remodeling response, all measurements were conducted at the control However, heart and ventricular mass do not increase (51). With temperature (25°C). In anesthetized turtles, cold acclimation reduced the increased sensitivity to cardiac preload (15) and ability for systemic resistance by 1.8-fold and increased systemic blood flow by a large Frank-Starling response (19), it has been suggested that 1.4-fold, resulting in a 2.3-fold higher right to left (R-L; net systemic) cardiac shunt flow and a 1.8-fold greater shunt fraction. Following a VStot has the potential to increase fivefold in diving slider volume load by bolus injection of saline (calculated to increase stroke turtles (6). volume by 5-fold, ϳ2.2% of total blood volume), systemic resistance Cardiac parameters of warm- and cold-acclimated turtles was reduced while pulmonary blood flow and systemic pressure have been previously reported in a number of studies, which increased. An increased systemic blood flow meant the R-L cardiac are in general agreement that following cold acclimation fH, shunt was further pronounced. In the isolated ventricle, passive Qtotal, and Psys are decreased while VStot, systemic resistance stiffness was increased following cold acclimation with 4.2-fold (Rsys) and hematocrit are increased (27, 51, 62, 64). However, greater collagen deposition in the myocardium. Histological sections these studies were performed at the turtle’s acclimation tem- by guest on July 20, 2016 of the major outflow arteries revealed a 1.4-fold higher elastin content perature (i.e., ϳ5 and 25°C for cold and control, respectively). in cold-acclimated animals. These results suggest that cold acclima- While this experimental methodology allows determinations of tion alters cardiac shunting patterns with an increased R-L shunt flow, achieved through reducing systemic resistance and increasing sys- cardiac function at the acclimation temperature, the physical temic blood flow. Furthermore, our data suggests that cold-induced changes to the cardiovascular system caused by cold acclima- cardiac remodeling may reduce the stress of high cardiac preload by tion are difficult to isolate from the direct effect of cold increasing compliance of the vasculature and decreasing compliance temperature. Studies in rainbow trout have shown significant of the ventricle. Together, these responses could compensate for structural remodeling of the cardiovascular system, which reduced systolic function at low temperatures in the slider turtle. persists in the absence of the direct effects of low temperature heart; in vivo; blood flow; blood pressure; cardiac preload; stiffness; (39, 42). Therefore, to differentiate a remodeling response collagen; elastin from the direct effect of temperature, cardiovascular function must be assessed at a common “test” temperature. Our objective was to determine whether prolonged temper- REDUCTIONS IN AMBIENT TEMPERATURE have direct (i.e., Q10; the ature acclimation caused remodeling of the freshwater red- rate of change over 10°C) and immediate effects on the eared slider turtle (Tracheyms scripta) cardiovascular system. ectotherm heart and cardiovascular system. This response to To differentiate the direct effect of temperature, and focus on low temperature is clear in freshwater turtles, resulting in the longer term remodeling response of the cardiovascular decreased heart rate (fH), cardiac twitch force, cardiac output system, animals were acclimated to a cold (5 Ϯ 0.3°C) or (Qtotal), and ventricular power output (15, 27). However, car- control temperature (25 Ϯ 0.3°C), but all experiments were diac stroke volume (VStot) is maintained at low temperatures, conducted at the control temperature (of 25 Ϯ 0.3°C). We partly by reduced end-systolic volume and increased diastolic assessed in vivo cardiac parameters in anesthetized animals filling due to changes in vascular resistance, and the sensitivity after a period of postsurgical stabilization and then in response of the turtle heart to cardiac preload is increased (15, 27). to a bolus injection (mean ϭ 2.2 Ϯ 0.4% of total blood These temperature-induced responses and the direct effect of volume) of saline directly into the jugular vein. Our first hypothesis was that chronic cold would reduce Qtotal, increase Address for reprint requests and other correspondence: H. A. Shiels, Faculty VStot, and decrease Psys. Our second hypothesis was that cold of Life Sciences, Univ. of Manchester, Manchester, UK (e-mail: acclimation would dampen the cardiac response to increased [email protected]). cardiac preload. As turtles have incomplete ventricular sepa- http://www.ajpregu.org Licensed under Creative Commons Attribution CC-BY 3.0 © the American Physiological Society. ISSN 0363-6119. R133 R134 CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES ration of venous and arterial circulations, allowing blood flow using a S-3A/I oxygen analyzer and a CD-3A carbon dioxide analyzer to bypass or partially bypass either systemic or pulmonary (Ametek, Berwyn, PA). circulation (29, 30, 32), we were particularly interested in To expose the central vascular blood vessels, a 5 ϫ 5 cm section of the plastron was cut away using a bone saw. The pectoral muscles and cardiac shunt flow (Qshunt) patterns. We further investigated the physical properties of the ventricle hypothesizing that cold connective tissue were gently loosened from the excised piece and bleeding from small superficial vessels stopped by cauterization. For acclimation would cause an increase in passive stiffness of the measurements of blood flows, 1- to 1.5-cm sections of all of the major ventricle, which would be reflected in an increased collagen outflow vessels of the heart were freed from connective tissue by blunt deposition. We found that Qshunt patterns were altered follow- dissection, taking care not to damage any smaller branching vessels or ing cold acclimation, with a right to left (R-L; net systemic) perforate the pericardium. 2S transit-time ultrasonic blood flow probes Qshunt achieved by reducing Rsys and, therefore, increasing (Transonic Systems, Ithica, NY) were fitted around all of the major Qsys. Furthermore, it appears that cold-induced cardiovascular outflow vessels. The first around the common pulmonary artery (CPa), remodeling increases ventricular stiffness and systemic vascu- the second around the left aorta (LAo), the third around both the left lature compliance, which reduces the stress of high blood carotid and subclavian arteries (LCa and LSCa) and the final around the volume load and may compensate for decreased systolic func- right aortic bundle (RAo) (Fig. 1A shows the major vessels as a histological section) (69). After flow probe placement, a small inci- tion at low temperatures. sion was made in the left side of the neck and the left carotid artery

and exterior jugular vein were exposed by blunt dissection. Both Downloaded from MATERIALS AND METHODS vessels were occlusively cannulated with polyethylene tubing (PE50) containing heparinized (50 units/ml) saline (0.9% NaCl) and the Experimental animals and acclimation. Male and female red-eared incision was sutured closed. Finally, a small incision was made in the sliders (Trachemys scripta, Schoepff; n ϭ 20; mean body mass ϭ pericardium to expose the apex of the heart. During each heartbeat, 1,352 Ϯ 69 g) were obtained from Lake Lewisville, TX, and trans- the separation of the cavum arteriosum and cavum venosum could be ported to the University of North Texas. Here, they were housed in visualized. A small hole was made in the heart wall, into the cavum ϳ50 L plastic containers (dimensions ϳ50 ϫ 50 ϫ 100 cm) contain- arteriosum, using a 28-gauge needle, and polyethylene tubing (PE10) http://ajpregu.physiology.org/ ing freshwater at a temperature of 25 Ϯ 0.3°C on a 12:12-h light-dark containing heparinized saline was inserted into the heart. The same cycle. Water quality was maintained by 100% water changes twice a process was repeated for the cavum venosum. We are confident that week and all animals were fed three times per week on commercial this method allowed for correct cannula placement; however, due to reptile feed (Aquatic turtle diet, Mazuri exotic animal nutrition). After the subsequent passive filling experiments we could not verify this 2 wk, 10 turtles were randomly assigned for cold acclimation where post mortem. All catheters were connected to pressure transducers, ambient temperature was reduced by 1°C per day until 5 Ϯ 0.3°C was which were calibrated daily against a static column of water by a reached and turtles were maintained at this temperature for 8–12 wk two-point calibration at 0 and 1 kPa. Before experimental procedures before experiments. An acclimation period of Ͼ8 wk was chosen to animals were left for a minimum of 1-h stabilization period. After 1 agree with our recent studies on fish (39), where Ͼ8 wk acclimation h all pressures and flows were checked to ensure they had been stable time is necessary to ensure cardiovascular structural remodeling (18, for at least 30 min; if they were not, the animals were left until they 42). Animals selected for cold acclimation were fasted following had. The total time from anesthetic dose to commencing experiments temperature reduction. The acclimation temperatures (control at 25 Ϯ was ϳ4 h, during which cold-acclimated animals warmed to the by guest on July 20, 2016 0.3°C; cold at 5 Ϯ 0.3°C) were chosen based on previous literature control temperature. (27, 28) to simulate summer and winter conditions. All acclimation In vivo cardiovascular measurements. Each catheter was attached temperatures were maintained in a walk in temperature controlled to a disposable pressure transducer (model MLT0699; ADInstru- room (model IR-912L5; Percival Scientific, Perry, IA) and animals ments, Colorado Springs, CO), adjusted to sit at the same level as were maintained in water without an area for basking. All animals the animal’s heart, connected to an amplifier (Quad Bridge Amp; survived the acclimation protocols and there were no signs of poor ADInstruments). Flow probes were connected to T206 dual chan- health in either group. Animal care and surgical preparations adhered nel small animal blood flow meters (Transonic Systems, Ithica, to the University of North Texas animal care and use protocol NY) for instantaneous blood flow rates. The output from the (Institutional Animal Care and Use Committee No. 11-007). transonic meters and the pressure signal were acquired at 40 Hz Anesthesia and surgical procedure. Before study, control animals using a PowerLab 16/35 data recording system (ADInstruments) were fasted for 1 wk. On the day of study, turtles were removed from and LabChart Pro software (v 7.2.5; ADInstruments). acclimation tanks, immediately weighed, and then anesthetized via an All experiments were conducted at the control temperature of 25 Ϯ intramuscular injection of sodium pentobarbital (50 mg/kg; Sigma- 0.3°C. Turtle body temperature was monitored by a cloacal thermistor Aldrich, St. Louis, MO) while still at their acclimation temperature. and a temperature probe (BAT-12 Microprobe Thermometer; Physi- The sodium pentobarbital was purchased as a powder and made as 2 temp, Clifton, NJ) placed in the chest cavity to ensure it remained at ml 100% ethanol, 8 ml propylene glycol, and 10 ml saline with 1,000 25 Ϯ 0.3°C throughout experiments. The control temperature was mg pentobarbital. Animals were then transported to the experimental chosen as the common temperature, instead of 5 Ϯ 0.3°C, as there is chamber (at 25 Ϯ 0.3°C). In most cases the pedal withdrawal response a body of literature on turtle heart function under anesthesia at ϳ25°C ceased within 30–60 min postinjection; however, if it persisted, an (11, 13, 36, 50). After stable pressures and flows were ensured, additional injection (25 mg/kg) was administered. During surgery, and baseline recordings of all pressures and flows were taken for a 5-min throughout experiments, turtles were maintained ventral side up and period. During this time, average VStot was calculated for each artificially ventilated to maintain normoxia via a tube inserted through individual. Following baseline recording, a bolus injection of physi- the glottis into the trachea (model 665; Harvard Apparatus, Holliston, ological saline (125 mM NaCl, 2.5 mM KCl, 2 mM CaCl2,1mM MA) at a rate of 24 breaths/min and a volume of 20 ml as previously MgSO4,1mMNaH2PO4, 10 mM HEPES, and 3 mM glucose at a pH reported for studies of this species (11, 13). Gas composition was of 7.7 with NaOH at room temperature) was administered via the controlled by rotameters (Sho-Rate Brooks Instruments Division, jugular cannula. The bolus volume was calculated to increase the Hatfield, PA) and bubbled into a gas mixer to maintain hydration. average stroke volume by fivefold for each particular animal, mim-

Fractional CO2 (FCO2) was maintained at 3 kPa to mimic partial icking the possible increase in venous return associated with breath pressure of CO2 in arterial blood (PCO2) (11, 22). The ventilated gas hold, ventilation, or movement (6, 60). Based on existing data sug- ϭ ϭ ϳ mixture (FO2 0.21, FCO2 0.03, balance N2) was checked regularly gesting total blood volume is 7% of total body mass in this species

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES R135 A

LCa RAo

LAo

Fig. 1. Flow probe placement. A: representative histological Downloaded from CPa section through the outflow vessels directly above the turtle 1 mm heart, stained with picro-sirus red, around which the flow probes were placed. Flow probes were placed around the common pulmonary artery (CPa), the left aorta (LAo), the left carotid artery (LCa) and left subclavian artery (LSCa), and the right aortic bundle (RAo). B: original representative recordings B bolus 60 secs of blood flow in the right vessel bundle (QRAo), accumulative

) 50 -1 left carotid and left subclavian (QLCa and QLSCa), left aorta http://ajpregu.physiology.org/ (QLAo), and common pulmonary artery (QCPa) taken from one

(ml min anesthetized turtle immediately before and after exogenous

RAo volume load calculated to increase average stroke volume by Q 0 5-fold, via the external jugular vein. The arrow marks the point

) at which the volume load (bolus) was injected. -1 24 (ml min LSc Q

+ 0 LCa Q 50 ) -1 by guest on July 20, 2016 (ml min LAo

Q 0

16 ) -1 (ml min CPa Q 0

(33), and under the assumption that 1 g body mass is equal to 1 ml cold-acclimated group and control group were generated at the same blood volume, the average volume load was 2.2 Ϯ 0.4% of total blood common control temperature, of 25 Ϯ 1°C, to isolate the effects of volume. The immediate change in pressure and flows were recorded chronic remodeling on myocardial stiffness from the acute effects of during this period (ϳ1 min) and then during the following sustained temperature. A cannula containing 25 Ϯ 0.3°C Ringer solution with elevated increase in pressure and flows (ϳ5 min). The animal was BDM was fed through the left aorta into the ventricular lumen and then left for ϳ25 min for pressure and flows to return to preinjection secured, using 0-0 silk thread (Harvard Apparatus), occluding the values, termed the recovery period, before a final 5-min recording was vessel. A second cannula, containing saline solution (0.9% NaCl), was taken (Fig. 1B). After completion of experimental protocol, all can- connected to a pressure transducer and fed through the common nulas and flow probes were removed, the animals were euthanized by pulmonary artery into the ventricle lumen. Again the cannula was intravenous administration of sodium pentobarbital (150 mg/kg), and secured in place with 0-0 silk thread, occluding the vessel. The the heart was excised. pressure transducer was calibrated daily against a static water column Ex vivo ventricular passive pressure-volume curves. The intact and recorded at 50 Hz using a PowerLab 16/35 data recording system isolated heart (free from cannulas) was washed in phosphate-buffered (ADInstruments) and LabChart Pro software (v 7.2.5; ADInstru- saline and placed into an organ bath containing Ringer solution (125 ments). All other inflow and outflow vessels were occluded, using 0-0 mM NaCl, 2.5 mM KCl, 2 mM CaCl2, 1 mM MgSO4,1mM silk thread, making the ventricle a sealed chamber with the two NaH2PO4, 10 mM HEPES, and 3 mM glucose at a pH of 7.7 with cannulas inside. With the use of a calibrated precision syringe pump NaOH at room temperature) at 25 Ϯ 0.3°C to which 60 mM 2,3 (INFORS), Ringer solution with BDM was pumped into the ventricle butanedione monoxime (BDM) was added to prevent active cross- at 0.05 ml/min until maximum volume was achieved, determined by bridge cycling. Pressure-volume curves from ventricles of both the a drop in the pressure trace following the protocol of Keen et al. (39).

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org R136 CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES All passive filling experiments were conducted Ͻ9 h after the original considered at P Ͻ 0.05. Values are presented as mean Ϯ SE through- injection of sodium pentobarbital. The two atria and the ventricle were out unless otherwise stated. Statistical details for each measurement separated, their mass was determined to the nearest milligram, and are given in the figure legends. they were fixed in 4% paraformaldehyde solution before being processed and embedded in paraffin wax. RESULTS Tissue histology. Fibrillar collagen and elastin content were ana- Heart and chamber mass. Body mass (1,438.1 Ϯ 115.4 vs. lyzed semiquantitatively following the methodology of Keen et al. Ϯ Ϯ Ϯ (39). Briefly, Formalin-fixed tissue samples were processed, embed- 1,266.4 80.5 g), heart mass (3.07 0.24 vs. 2.84 0.211 ded in paraffin wax, sectioned at 5 ␮m (Leica RM2255 microtome; g), left atrial mass (0.26 Ϯ 0.034 vs. 0.24 Ϯ 0.034 g), right Leica, Wetzlar, Germany), and mounted onto glass slides (Super frost atrial mass (0.40 Ϯ 0.035 vs. 0.39 Ϯ 0.035 g), ventricular mass plus; Thermo Fisher Scientific, Waltham, MA). Serial sections from (2.24 Ϯ 0.17 vs. 2.07 Ϯ 0.14 g), relative heart mass (0.22 Ϯ each sample were stained with picro-sirus red for collagen (37) and 0.013 vs. 0.22 Ϯ 0.009 g), and relative ventricular mass Miller’s elastic stain for elastin (48). Picro-sirus red images were (0.16 Ϯ 0.007 vs. 0.16 Ϯ 0.007 g) were not different between quantified using polarized light microscopy and Miller’s elastic im- cold-acclimated and control animals, respectively. ages were quantified using bright-field microscopy. Mean fibrillar Effects of thermal acclimation on baseline in vivo cardiac collagen and elastin contents were expressed as a percentage of total function. Cold acclimation resulted in a 1.8-fold reduction in tissue cross-sectional area, excluding the epicardial surface, deter-

systemic resistance (Rsys) compared with control animals Downloaded from mined using ImageJ (57). All histological analysis was conducted 2 ϭ ϭ Ͻ blind to the acclimation group. Three tissue sections were considered [R 0.67, F(1,53) 79.25, P 0.0001; Fig. 2A]. This for each individual to ensure consistency in measurements. On each reduction in Rsys contributed to a 1.4-fold increase in systemic tissue section three separate image montages were taken along tran- blood flow (Qsys) in cold-acclimated compared with control 2 sects across the full diameter of the tissue cross section. animals [R ϭ 0.38, F(1,58) ϭ 8.69, P Ͻ 0.005; Fig. 2B]. As Calculations and statistics. Pulmonary blood flow (Qpul) was there was no effect of temperature acclimation on total cardiac determined directly based on the flow probe output, while total output (Qtotal; Fig. 2C), pulmonary flow (Qpul; Fig. 2D)or systemic blood flow (Qsys) was calculated as the sum of the blood systemic pressure (Psys; Fig. 2E), both groups had a right to left http://ajpregu.physiology.org/ flow recorded from the left aorta, left carotid and subclavian arteries, (R-L; net systemic) cardiac shunt (Qshunt; Fig. 2F). Cardiac and the right bundle. Total blood flow (Qtotal) was the sum of Qpul and shunting occurs in turtles because they have incomplete ven- Qsys. Net cardiac shunt flow (Qshunt) was calculated as the difference tricular separation of venous and arterial circulations, allowing between Qpul and Qsys (Qpul Ϫ Qsys); therefore, a positive Qshunt indicates a left to right (L-R; net pulmonary) shunt and a negative blood flow to bypass or partially bypass either the systemic or value indicates a right to left (R-L; net systemic) shunt (12, 43, 69). pulmonary circulation. These intracardiac shunt flow (Qshunt) All flow data were standardized to body mass (kg). Heart rate (fH) was patterns, i.e., the direction of the blood shunt, are largely calculated on the basis of the instantaneous blood flow profile in the determined by vascular resistance of the systemic and pulmo- left aorta. Total stroke volume (VStot) was calculated as Qtotal/fH and nary circulations (31, 50, 58). However, the combined effect of standardized to body mass (kg). Systemic resistance (Rsys) was the greater Qsys and reduction in Rsys of the cold-acclimated 2 calculated as mean systemic blood pressure relative to systemic blood turtles produced a 2.3-fold higher R-L Qshunt [R ϭ 0.62, by guest on July 20, 2016 flow (Psys/Qsys), under the assumption that atrial pressure is zero, and F(1,58) ϭ 20.69, P Ͻ 0.0001] and a 1.8-fold greater shunt standardized to body mass (kg). Ventricular contractility (dP/dt) was 2 fraction compared with control [R ϭ 0.31, F(1,58) ϭ 7.02, P Ͻ calculated as the maximum rate of pressure increase over six heart- 0.05]. Interestingly, despite changes in blood flow distribution, beats, taken from the cannula positioned in the cavum arteriosum. we did not find an effect of temperature acclimation on max- For all in vivo recordings, and calculations based on in vivo recordings, significant differences between acclimation temperatures imum or minimum intraventricular pressure between two of the and within groups were assessed by two-way repeated measures cava in the heart (Parteriosum and Pvenosum; Fig. 3, A and B) nor ANOVA/General Linear Model (GLM), with the cardiovascular pa- was there a change in ventricular contractility (Fig. 3C). rameter as the dependent variable, and stage of experiment and Finally, there was no difference between resting heart rate (fH) acclimation group as the fixed factors. When significance was found, or total stroke volume (VStot) values between cold-acclimated a Sidak multiple comparisons post hoc test was conducted to assess and control animals (Fig. 4, A and B). significance between acclimation groups at each stage of the experi- In vivo cardiovascular response to volume load. The bolus ment. Differences between means within groups were subsequently injection, designed to give a fivefold increase in venous return assessed by Tukey’s multiple comparisons post hoc test. Mass param- volume, increased VStot in both groups, which remained ele- eters and differences in collagen and elastin deposition were analyzed vated during the sustained pressure period {for cold [R2 ϭ separately by multiple unpaired t-tests for parametric data or Mann 2 0.75, F(3,18) ϭ 18.08, P Ͻ 0.0001] and for control [R ϭ 0.42, Whitney U-tests for nonparametric data, with each parameter as the ϭ Ͻ test variables and acclimation group as the grouping variable. Each of F(3,18) 4.32, P 0.05]; Fig. 4B}. This increase in VStot these statistical analyses were performed using Prism v6.04 (Graph- directly elevated Qtotal in both groups during the bolus injection 2 Pad Software, La Jolla, CA). and sustained pressure period {for cold [R ϭ 0.72, F(3,18) ϭ 2 Chamber filling volume was calculated from filling time by the 15.39, P Ͻ 0.0001] and for control [R ϭ 0.37, F(3,18) ϭ 3.57, equation: P Ͻ 0.05]; Fig. 2C}. The fH was not affected by the volume load (Fig. 4A). The increase in Qtotal translated into an increase 0.05 2 ϭ volume (ml) ϭ time (s) ϫ in Qsys in both temperature groups {for cold [R 0.72, 2 60 F(3,18) ϭ 15.48, P Ͻ 0.0001] and for control [R ϭ 0.37, ϭ Ͻ The effect of temperature acclimation on the pressure-volume rela- F(3,18) 3.57, P 0.0001]; Fig. 2B} and Qpul in control 2 ϭ ϭ Ͻ tionship was assessed by ANCOVA with pressure as the dependent animals only [R 0.39, F(3,18) 4.48, P 0.05; Fig. 2D], 2 variable, volume and acclimation group as fixed factors, and body reducing the net R-L shunt in cold-acclimated animals [R ϭ mass as the covariate, with a Tukey post hoc test for differences 0.52, F(3,18) ϭ 6.58, P Ͻ 0.005; Fig. 2F]. The bolus injection 2 between groups using R (53). For all analyses, significance was caused a decrease in Rsys in both groups {for cold [R ϭ 0.55,

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES R137

Fig. 2. Blood pressures and flows. In vivo measurements of systemic resistance (Rsys; A), systemic blood flow (Qsys; B), pulmonary blood flow (Qpul; C), total cardiac output (Qtotal; D), systemic pressure (Psys; E), and net cardiac shunt flow (Qshunt ϭ Qpul Ϫ Qsys;

F) in cold-acclimated (open circles, dashed Downloaded from line) and control (closed circles, solid line), anesthetized and artificially ventilated, freshwater turtles assessed at 25°C. Values are displayed as means Ϯ SE; n ϭ 10. Flow values have been standardized to mass for graphical representation. *P Ͻ 0.05, significant differences between acclimation groups; ϩ Ͻ P 0.05, significant changes from the http://ajpregu.physiology.org/ baseline value within groups during the experiment [general linear model (GLM)]. by guest on July 20, 2016

2 F(3,18) ϭ 7.38, P Ͻ 0.005] and for control [R ϭ 0.68, F(3,18) ϭ with a greater shunt fraction, in the cold-acclimation compared 12.93, P Ͻ 0.0001]; Fig. 2A}, but an increase in Psys in the with the control animals during both the bolus injection and 2 control group only [R ϭ 0.42, F(3,18) ϭ 4.38, P Ͻ 0.05; Fig. sustained pressure phase (P Ͻ 0.05). Temperature acclimation 2E]. Maximum Parteriosum was increased in the cold-acclimated did not influence Qtotal, Parteriosum,Pvenosum, fH,orVStot during animals; however, this was decreased in the control animals either the bolus injection or sustained pressure phase (Figs. 2C, 2 {for cold [R ϭ 0.50, F(3,18) ϭ 6.134, P Ͻ 0.05] and for control 3, and 4). 2 [R ϭ 0.54, F(3,18) ϭ 8.31, P Ͻ 0.05]; Fig. 3A}. Following the Thermal remodeling of ex vivo ventricular compliance. The bolus injection there was an increase in ventricular contractility maximum Parteriosum following the volume load was increased 2 in both groups {for cold [R ϭ 0.48, F(3,21) ϭ 6.58, P Ͻ 0.05] in the cold-acclimated and reduced in the control group, 2 and for control [R ϭ 0.73, F(3,24) ϭ 22.09, P Ͻ 0.001]; Fig. compared with baseline values. The increased pressure in 3C}. After the recovery period, Rsys remained elevated in the response to volume suggests a decrease in ventricular compli- control group whereas it returned to baseline levels in the ance with cold acclimation. To assess the functional effects of cold-acclimated (Fig. 2A). cardiac remodeling on the passive properties of the thermally Effects of thermal acclimation on the in vivo cardiovascular acclimated ventricle we generated ex vivo passive filling response to volume load. Following the bolus injection, Rsys curves from freshly isolated intact ventricles treated with BDM remained twofold higher in control compared with cold-accli- at a common test temperature of 25 Ϯ 0.3°C. Thermal acclimated animals during both the bolus injection and the sustained mation altered the pressure-volume relationship during filling 2 pressure phase (P Ͻ 0.05; Fig. 2A). During both the bolus [R ϭ 0.75, F(7,379916) ϭ 167200, P Ͻ 0.001] revealing injection and sustained pressure phase Qsys was slightly in- increased stiffness in the cold compared with controls creased, while Qpul and Psys were slightly decreased by cold (Fig. 5A). acclimation compared with control (Fig. 2, B, D, and E). The Thermal remodeling of connective tissue. Increased myocar- R-L Qshunt persisted in both acclimation groups, however, the dial stiffness can be due to a remodeling of the extracellular reduced Rsys,Qpul, and Psys combined with the increased Qsys matrix in both mammals and ectotherms (8, 39, 42). We used meant the R-L Qshunt remained more pronounced (Fig. 2F), picro-sirus red to determine fibrillar collagen deposition in the

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org R138 CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES acclimation compared with controls (Fig. 6, B and C, respectively). We were unable to statistically resolve differences between the acclimation groups for fibrillar collagen in the major outflow arteries (Fig. 6D). Representative bright-field images with their corresponding plane polarized light image are shown following cold acclimation in Fig. 6, E and Ei, respectively, and for controls in Fig. 6, F and Fi, respectively.

DISCUSSION Ectotherms show a wide range of physiological responses to acute and chronic temperature change. Freshwater turtles en- dure large ﬂuctuations in seasonal temperature in their native environments (66). These seasonal temperature changes directly affect many physiological processes, with extreme cold triggering winter long hibernation or brumation (66). Here,

freshwater slider turtles were exposed to chronic cold (ϳ5°C) Downloaded from to simulate a winter phenotype. In vivo cardiovascular function, ex vivo ventricular compliance, and connective tissue content of the ventricle and major outflow vessels were assessed. However, unlike previous studies, all experiments were conducted at a common control temperature (25 Ϯ 0.3°C) to differentiate thermal remodeling of the cardiovascular system from the direct effects of temperature. Our findings indicate http://ajpregu.physiology.org/ cold acclimation increased R-L Qshunt by a reduction in Rsys and an increase in Qsys. Furthermore, cold-induced cardiovascular remodeling increased ventricular stiffness during passive filling. Critique of the methods. The acclimation duration in this study (Ͼ8 wk before the start of experiments) was longer than by guest on July 20, 2016

Fig. 3. Intraventricular pressure and contractility. In vivo measurements of maximum (black) and minimum (grey) pressure in the cavum arteriousm (Parteriosum; A) and the cavum venosum (Pvenosum; B), and ventricular contractility (dP/dt; C) in cold-acclimated (open circles, dashed line) and control (closed circles, solid line), anesthetized and artificially ventilated, freshwater turtles assessed at 25°C. Values are displayed as means Ϯ SE; n ϭ 10. ϩP Ͻ 0.05, significant changes from the baseline value within groups during the experiment (GLM). turtle ventricle and major outflow vessels. Ventricular collagen content was 4.3-fold higher in cold-acclimated animals compared with controls (R2 ϭ 0.69, P Ͻ 0.0001; Fig. 5B), as shown by the higher degree of dark red staining, when visualized under bright-field light, and the increased number of birefrin- gent fibers, when visualized under plane polarized light (Fig. 5, C, Ci, D, and Di). We did not detect elastin in the turtle ventricular myocardium except in coronary vessels (not shown). The major outflow vessels contain a thick layer of connective tissue that provides structural support and elastic recoil. With high cardiac preload and VStot associated with cold acclimation, the balance of collagen and elastin in these vessels Fig. 4. Cardiac performance. In vivo measurements of heart rate (fH; A) and may be critical to vascular function (4, 24, 52). Staining the total stroke volume (VStot; B) in cold-acclimated (open circles, dashed line, vessels with Miller’s elastic stain revealed a 1.4-fold increase striped bar) and control (closed circles, solid line, solid bar), anesthetized and artificially ventilated, freshwater turtles assessed at 25°C. Values are displayed in elastin content in the artery wall following cold acclimation Ϯ ϭ 2 as means SE; n 10. VStot has been standardized to mass for graphical (R ϭ 0.51, P Ͻ 0.05; Fig. 6A), which is visualized by the representation. ϩP Ͻ 0.05, significant changes from the baseline value within increased black staining of the elastic lamellae following cold groups during the experiment (GLM).

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES R139 A B *

Fig. 5. Ventricular stiffness and connective tissue C content. Ex vivo ventricular passive pressure-vol- C 100 m i 100 m ume relationships (A) and semiquantitative analysis of ventricular collagen content (B) in cold- acclimated (open circles, dashed line) and control (closed circles, solid line) freshwater turtles. Rep- resentative micrographs of picro-sirus red stained Downloaded from ventricular tissue, for cold-acclimated freshwater turtles viewed under bright-field (C) and plane polarized light (Ci) and control freshwater turtles viewed under bright-field (D) and plane polarized light (Di). Values are displayed as means Ϯ SE; n ϭ 10. Pressure has been standardized to start at 0 kPa for graphical representation. *P Ͻ 0.05, significant differences between acclimation groups http://ajpregu.physiology.org/ (GLM). D D i 100 m 100 m by guest on July 20, 2016

used in a number of previous studies on freshwater turtles (12, of apneas causing bodily gas stores to fluctuate (22, 46, 58). 27, 62, 63) but was chosen to agree with the timeframe However, fH and overall Qpul are not affected by this contin- required for a cardiac remodeling response to temperature in uous ventilation pattern (46). Finally, the turtles used in this other ectotherms (e.g., the rainbow trout) (39, 42). Second, the study were maintained in water throughout the acclimation majority of previous studies on cold-acclimated turtle heart period in both groups. To adjust buoyancy turtles may retain function were conducted on recovered animals (27, 62, 63). water in their urinary bladder or cloacal bursac when main- Anesthesia with pentobarbital blunts autonomic tone on the tained in water for prolonged periods of time (34, 55). We did cardiovascular system and is, therefore, useful in assessing not find any difference in total body mass between control and hemodynamic effects of central nervous system regulation (13, cold-acclimated turtles, in agreement with previous studies (27, 36, 50). A number of studies have previously assessed cardiac 51). However, as we did not measure body mass before and function while animals remained under anesthesia (Table 1) after acclimation periods we are unsure if it was altered by (10, 11, 13, 21, 31, 36, 50). At warm temperatures anesthesia water retention. with sodium pentobarbital has been shown to cause a L-R (net The effect of thermal acclimation on systemic pressure and pulmonary) shunt as it blocks the cholinergic mediated con- blood flow. Our findings suggest cold acclimation reduced Rsys, striction of the pulmonary artery that is normally associated increased Qsys, and, therefore, increased R-L Qshunt. Previ- with apnoea (51). Third, we also continuously artificially ously, cold acclimation has been reported to reduce both Psys ventilated our animals to prevent hypoxia. Lung ventilation in and systemic conductance (Gsys; 1/Rsys) in turtles (12, 27, recovered animals is associated with an increased fH and Qpul 61–63). Speculatively, this may be due to atrial natriuretic and a reduction in the overall R-L shunt (58, 69, 71), suggested peptide (ANP), which is present in the testudine heart (65). In to be due to vagal and adrenergic tone (27, 30). Using a mammals and reptiles, ANP is released by cardiomyocytes in continuous artificial ventilation protocol has been shown to response to pressure or volume induced myocardial stretch and remove pulmonary CO2 more effectively (46) than the episodic has systemic vasodilatory action (41, 49); however, it is un- ventilation pattern seen in a nonanesthetized and unventilated clear if the action of ANP is blunted by cold temperatures like animal; a series of consecutive breaths interspersed by periods some other endocrine functions (45). Vasodilation is consistent

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org R140 CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES ABC *

100 m 100 m

D Downloaded from 100 m 100 m

Ei Fi http://ajpregu.physiology.org/

100 m 100 m

Fig. 6. Connective tissue content of the major outflow vessels. A: semiquantitative analysis of elastin content in the major outflow (OTF) arteries for cold-acclimated (striped bar) and control (solid bar) freshwater turtles. This result is shown by representative micrographs of a Miller’s elastic stained cold-acclimated (B) and control vessel wall (C). D: semiquantitative analysis of collagen content in the major outflow arteries for cold-acclimated (striped bar) and control (solid bar) freshwater turtles. This result is shown by representative micrographs of a picro-sirus red stained cold-acclimated tissue under bright-field (E) and plane polarized light (Ei) and control tissue under bright-field (F) and plane polarized light (Fi). Values are displayed as means Ϯ SE; n ϭ 10. *P Ͻ 0.05, significant differences between acclimation groups (GLM). by guest on July 20, 2016 with previous studies that show arterial hypotension following or 5°C. Furthermore, we found an increased Qsys when cold- cold acclimation when animals are studied at their acclimation acclimated animals were studied at 25°C, which differs from temperatures (27). However, temperature is also inversely the findings previously reported for animals studied at 5°C (62, related to blood viscosity (44, 54, 56), which leads to higher 63). We also report an increased elastin content found in the viscosity at low temperatures and an overall increase in Rsys major outflow arteries, suggesting increased compliance and (27, 62, 63). The overall decrease in Rsys following cold elastic recoil in these vessels. Indeed, cold acclimation may acclimation, in the current study, suggests either a functional induce decreases in Rsys via a structural change in vasculature change in the vasculature, changes in vascular physical prop- compliance; however, further studies are needed to test this erties, or reduced viscosity at 25°C. However, Saunders and speculation experimentally. Patel (56) report very little difference between blood viscosi- Cardiovascular remodeling with temperature acclimation. ties of red-eared sliders when tested at 25°C, except at very low In both acclimation groups, a volume load of ϳ2.2% of total shear rates, regardless of whether turtles were acclimated to 25 blood volume increased VStot,Psys, and Qtotal in agreement

Table 1. In vivo cardiac parameters in anesthetized freshwater slider turtles (Trachemys scripta) at warm temperatures (ϳ25°C)

Ϫ1 Ϫ1 Ϫ1 Ϫ1 Ϫ1 Ϫ1 Ϫ1 Ϫ1 Ϫ1 Ϫ1 Ϫ1 Qpul, ml·min ·kg Qsys, ml·min ·kg fH, min VStot,kg Rsys, kPa·ml ·min ·kg Qshunt, ml·min ·kg Reference ϳ 16 ϳ45 ϳ26 ϳ2.7 ϳ0.05 ϳϪ29 Joyce and Wang (35) 31.9 31.1 20.7 3.1 0.13 Ϫ0.23 Galli et al. (21) 60.4 28.0 35.6 2.6 0.14 2.34 Overgaard et al. (50) 27.3 22.2 37.5 1.32 1.72 5.1 Crossley et al. (13) 35.1 31.9 38.5 1.73 0.12 3.2 Crossley et al. (11) 42.3 19.5 43.2 1.37 0.15 22 Hicks et al. (31) 57.4 — 44.0 — — — Hicks and Comeau (30) 33.8 31.9 38.5 1.73 0.12 1.9 Comeau and Hicks (10) 5.4 9.2 31.2 0.43 0.25 Ϫ3.2 Present study Data from previous studies is presented alongside baseline cardiac parameters from the present study for control animals (acclimated at 25°C) assessed at 25°C. Qpul, pulmonary blood ﬂow; Qsys, systemic blood blow; fH, heart rate; VStot, total stroke volume; Rsys, systemic resistance; Qshunt, net cardiac shunt ﬂow.

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES R141 with previous studies on this species (14). In control animals and there is a greater degree of blood flow separation (17, 68, this caused both Qpul and Qsys to increase; however, in cold- 70). Interestingly, although there was no change in intraven- acclimated animals only Qsys increased. This finding could tricular pressure, we did see changes in cardiac shunting relate to acclimation-induced vasculature remodeling to in- patterns with an increased R-L Qshunt following cold acclima- crease compliance in cold-acclimated animals, reducing the tion. This finding suggests that cardiac shunting is controlled effect of changes in ejected VStot or Qtotal on Psys (20). This entirely by peripheral resistance rather than modulating pump- idea is supported by our histological analysis of the outflow ing pressure within the ventricle. arteries, which suggests an increase in compliant elastin fibres The effect of thermal acclimation on heart rate, stroke in the cold-acclimated group. Increased vascular compliance volume, and cardiac output. Previous studies report decreased would be able to accommodate the large increases in volume as fH and Qtotal, with a corresponding increase in VStot following well as reduce afterload pressure and systolic wall stress, cold acclimation (15, 62–64). However, we did not find dif- potentially allowing systolic pressure and ejection fraction to ferences in these parameters between temperature groups. be preserved (40). The increased elastin content would also Indeed, our values for these parameters in both groups better improve recoil in the arteries, helping them to more efficiently correlate to previously reported baseline levels under anesthe- smooth blood flow and regain their shape faster in diastole sia at ϳ25°C (Table 1) (10, 11, 13, 21, 30, 31, 36, 50). Our

(25, 67). blood flow values are lower than that of most other studies. The Downloaded from The maximum Parteriosum following volume load was in- basis for this difference was not clear but may relate to creased in the cold-acclimated and reduced in the control intraventricular pressure cannula placement, which is the main group, compared with baseline values. The increased pressure difference between this study and those previously conducted. response to volume suggests a decrease in ventricular chamber As the animals in the present study were anesthetized and compliance with cold acclimation. This finding agrees with the ventilated, fH and Qtotal are likely higher than if the animals result of our ex vivo pressure-volume relationships and ven- were nonanesthetized and apnoeic as anesthesia blunts vagal tricular connective tissue data, which suggest that the cold- tone on the heart and pulmonary artery (11, 50). Our findings http://ajpregu.physiology.org/ acclimated ventricle was stiffer than controls, with a higher suggest that, although cold-acclimated animals show depressed deposition of collagen in the myocardial wall. To our knowl- fH and Qtotal with an increased VStot when studied at 5°C, the edge this is the first time pressure-volume curves have been main driver of these changes is acute cold rather than cold generated in the turtle heart; however, Farrell et al. (16) used an acclimation per se. Indeed, Stecyk et al. (63) came to a similar in situ working heart preparation from slider turtles to show conclusion in regard to fH while working with spontaneously increased sensitivity to filling pressure following acute cold beating whole heart preparations. exposure, which is consistent with our findings. Data from Perspectives and Significance mammals also suggest that a shift in the Frank-Starling curve due to decreased ventricular compliance can preserve systolic Prolonged cold initiates profound alterations in ectotherm function and increase preload recruitable stroke work (1, 3, 23, physiology. In the case of the freshwater turtle, acute cold by guest on July 20, 2016 47). Moreover, ventricular stiffness is correlated with increased triggers large depressions in metabolic rate to initiate winter systolic pressure sensitivity to cardiac preload, and therefore, hibernation/brumation and results in modifications in cardio- increases in central blood volume give greater increases in vascular function (26). The data presented in this study builds systolic developed pressure, even in the absence of cardiac from previous studies in freshwater turtles (15, 27, 28), indi- hypertrophy as shown in these turtles (9, 20). Structural re- cating that cold acclimation alters cardiac shunting patterns modeling causing increased ventricular stiffness can also be with an increased R-L Qshunt, achieved through a reduction in associated with diastolic dysfunction and increased diastolic Rsys and an increase in Qsys. Furthermore, cold acclimation pressure (2, 41). It is unclear whether a reduction in cardiac increased the heart’s sensitivity to an in vivo volume load, compliance with associated fibrosis is beneficial or maladap- which may be relevant during hibernation, when diving turtles tive in the turtle heart following cold acclimation. Diastolic have been suggested to increase VStot by up to fivefold (6). Ex function appears normal at low volumes as minimum Parteriosum vivo passive filling of the ventricle revealed a reduction in and Pvenosum were not affected by thermal acclimation. We did ventricular compliance, which was associated with fibrosis of not find a difference in ventricular contractility following the myocardium. In turn, the major outflow arteries exhibited temperature acclimation. We speculate that the lack of change an increase in elastin content of the elastic lamellae suggesting is explained by thermal remodeling in the turtle heart being a increased outflow vessel compliance following cold acclima- physiological, rather than a pathological, response. tion. These findings suggest that cold-induced structural car- Maximum and minimum intraventricular pressures were diovascular remodeling alters the hemodynamics of freshwater similar, independent of temperature acclimation, which agrees turtles to limit the stress of high blood preload on the heart. It with previous data on most reptiles (5, 35, 58). It is likely that is possible that these structural changes may also compensate differential pressure generation is unnecessary due to the low for decreased systolic function associated with low tempera- Psys in the turtle and, therefore, low risk of pulmonary edema tures. in the pulmonary vasculature (7, 29). In more active reptile species, such as the Burmese python (Python molurus), the ACKNOWLEDGMENTS systemic side of the heart (or cavum arteriousm) generates We thank Dr. Josele Flores-Santin, John Marrin, Oliver Wearing, and higher pressure than the pulmonary side (68). However, in the Taylor Shackelford for help with experiments and animal care. We thank Dr. Jonathan Codd and Dr. Peter Tickle for help with experimental design and python heart the muscular ridge between the cava is larger than set-up. We thank Dr. Jonathan Stecyk for helpful discussions and revisions to that in the turtle heart, the ventricular wall surrounding the earlier versions of the manuscript. Histology was performed in the University systemic side of the heart is thicker than the pulmonary side, of Manchester Histology Facility.

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org R142 CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES

GRANTS 17. Farrell AP, Gamperl AK, Francis ET. Comparative aspects of heart morphology. In: Biology of the Reptilia. Volume 19: Morphology G A. N. Keen was funded by the Biotechonology and Biological Sciences Visceral Organs, edited by Gans C, Gaunt AS. Ithica, NY: Soc. Study Research Council and this research trip was funded by travel awards from the Amphibians Reptiles, 1998, p. 375–424. Fisheries Society of the British Isles and the Royal Society of Biology. D. A. 18. Farrell AP, Hammons AM, Graham MS, Tibbits GF. Cardiac growth Crossley, 2nd was funded by National Science Foundation CAREER Award in rainbow trout, Salmo gairdneri. Can J Zool 66: 2368–2373, 1988. IBN IOS-0845741. H. A. Shiels was supported by a Wellcome Trust ISSF 19. Franklin CE. Intrinsic properties of an in situ turtle heart (Emydura Collaborative Research Visit grant. signata) preparation perfused via both atria. Comp Biochem Physiol A 107: 501–507, 1994. DISCLOSURES 20. Frenneaux M, Williams L. Ventricular-arterial and ventricular-ventric- No conflicts of interest, financial or otherwise, are declared by the author(s). ular interactions and their relevance to diastolic filling. Progr Cardiovasc Dis 49: 252–262, 2007. AUTHOR CONTRIBUTIONS 21. Galli G, Taylor EW, Wang T. The cardiovascular responses of the freshwater turtle, Trachemys scripta, to warming and cooling. J Exp Biol A.N.K. and H.A.S. conception and design of research; A.N.K. and 207: 1471–1478, 2004. D.A.C.I. performed experiments; A.N.K. analyzed data; A.N.K., H.A.S., 22. Glass ML, Boutilier RG, Heisler N. Ventilatory control of arterial Po2 in and D.A.C.I. interpreted results of experiments; A.N.K. prepared figures; the turtle Chrysemys picta bellii: effects of temperature and hypoxia. J A.N.K. drafted manuscript; A.N.K., H.A.S., and D.A.C.I. edited and Comp Physiol 151: 145–153, 1983. revised manuscript; A.N.K., H.A.S., and D.A.C.I. approved final version of 23. Glezeva N, Baugh JA. Role of inflammation in the pathogenesis of heart manuscript. failure with preserved ejection fraction and its potential as a therapeutic Downloaded from target. Heart Fail Rev 19: 681–694, 2014. REFERENCES 24. Graham HK, Akhtar R, Kridiotis C, Derby B, Kundu T, Trafford 1. Arbab-Zadeh A, Dijk E, Prasad A, Fu Q, Torres P, Zhang R, Thomas AW, Sherratt MJ. Localised micro-mechanical stiffening in the ageing JD, Palmer D, Levine BD. Effect of aging and physical activity on left aorta. Mech Ageing Dev 132: 459–467, 2011. ventricular compliance. Circulation 110: 1799–1805, 2004. 25. Halper J, Kjaer M. Basic components of connective tissues and extra- 2. Asrar Ul Haq M, Mutha V, Rudd N, Hare DL, Wong C. Heart failure cellular matrix: elastin, fibrillin, fibulins, fibrinogen, fibronectin, laminin, with preserved ejection fraction–unwinding the diagnosis mystique. Am J tenascins and thrombospondins. In: Progress in Heritable Soft Connective http://ajpregu.physiology.org/ Cardiovasc Dis 4: 100–113, 2014. Tissue Diseases, edited by Halper J. Berlin, Germany: Springer-Verlag, 3. Bhella PS, Hastings JL, Fujimoto N, Shibata S, Carrick-Ranson G, 2014, p. 31–47. Palmer MD, Boyd KN, Adams-Huet B, Levine BD. Impact of lifelong 26. Herbert CV, Jackson DC. Temperature effects on the responses to exercise “dose” on left ventricular compliance and distensibility. JAm prolonged submergence in the turtle Chrysemys picta bellii. II. Metabolic Coll Cardiol 64: 1257–1266, 2014. rate, blood acid-base and ionic changes, and cardiovascular function in 4. Bruel A, Oxlund H. Biosynthetic growth-hormone changes the collagen aerated and anoxic water. Physiol Zool 58: 670–681, 1985. and elastin contents and biomechanical properties of the rat aorta. Acta 27. Hicks JM, Farrell AP. The cardiovascular responses of the red-eared Endocrinol 125: 49–57, 1991. slider (Trachemys scripta) acclimated to either 22 or 5°C. I. Effects of 5. Burggren W. Pulmonary circulation of chelonian reptile: morphology, anoxic exposure on in vivo cardiac performance. J Exp Biol 203: 3765– hemodynamics and pharmacology. J Comp Physiol 116: 303–323, 1977. 3774, 2000. 6. Burggren W, Farrell AP, Lillywhite H. Vertebrate cardiovascular sys- 28. Hicks JM, Farrell AP. The cardiovascular responses of the red-eared tems. In: The Handbook of Physiology. Comparative Physiology, edited by slider (Trachemys scripta) acclimated to either 22 or 5°C. II. Effects of anoxia on adrenergic and cholinergic control. J Exp Biol 203: 3775–3784,

Dantzler WH. Bethesda, MD: Am. Physiol. Soc., 1997, sect. 13, p. by guest on July 20, 2016 215–308. 2000. 7. Burggren WW. Pulmonary blood plasma filtration in reptiles: a wet 29. Hicks JW. Adrenergic and cholinergic regulation of intracardiac shunting. vertebrate lung. Science 215: 77–78, 1982. Physiol Zool 67: 1325–1346, 1994. 8. Chapman D, Weber KT, Eghbali M. Regulation of fibrillar collagen 30. Hicks JW, Comeau SG. Vagal regulation of intracardiac shunting in the type-I and type-III and basement-membrane type-IV collagen gene-ex- turtle Pseudemys scripta. J Exp Biol 186: 109–126, 1994. pression in pressure overloaded rat myocardium. Circ Res 67: 787–794, 31. Hicks JW, Ishimatsu A, Molloi S, Erskin A, Heisler N. The mechanism 1990. of cardiac shunting in reptiles: a new synthesis. J Exp Biol 199: 1435– 9. Chen CH, Nakayama M, Nevo E, Fetics BJ, Maughan WL, Kass DA. 1446, 1996. Coupled systolic-ventricular and vascular stiffening with age implications 32. Hicks JW, Wang T. Functional role of cardiac shunts in reptiles. J Exp for pressure regulation and cardiac reserve in the elderly. J Am Coll Zool 275: 204–216, 1996. Cardiol 32: 1221–1227, 1998. 33. Hutton KE. Blood volume, corpuscular constants, and shell weight in 10. Comeau SG, Hicks JW. Regulation of central vascular blood flow in the turtles. Am J Physiol 200: 1004–1006, 1961. turtle. Am J Physiol Regul Integr Comp Physiol 267: R569–R578, 1994. 34. Jackson DC. Mechanical basis for lung volume variability in turtle 11. Crossley D, Altimiras J, Wang T. Hypoxia elicits an increase in pulmo- Pseudemys scripta elegans. Am J Physiol 220: 754–758, 1971. nary vasculature resistance in anaesthetised turtles (Trachemys scripta). J 35. Johansen K. The influence of temperature on the electrocardiograms of Exp Biol 201: 3367–3375, 1998. some northern reptiles. Acta Physiol Scand 46: 346–357, 1959. 12. Crossley DA, II, Wearing OH, Platzack B, Hartzler LK, Hicks JW. 36. Joyce W, Wang T. Adenosinergic regulation of the cardiovascular system Acute and chronic temperature effects on cardiovascular regulation in the in the red-eared slider Trachemys scripta. Comp Biochem Physiol A 174: red-eared slider (Trachemys scripta). J Comp Physiol B 185: 401–411, 18–22, 2014. 2015. 37. Junqueira LC, Bignolas G, Brentani RR. Picrosirius staining plus 13. Crossley DA, Wang T, Altimiras J. Role of nitric oxide in the systemic polarization microscopy, a specific method for collagen detection in and pulmonary circulation of anesthetized turtles (Trachemys scripta). J tissue-sections. Histochem J 11: 447–455, 1979. Exp Zool 286: 683–689, 2000. 39. Keen AN, Fenna AJ, McConnel JC, Sherratt MJ, Gardner P, Shiels 14. Faraci FM, Shirer HW, Orr JA, Trank JW. Circulatory mechanore- HA. The dynamic nature of hypertrophic and fibrotic remodelling in the ceptors in the pond turtle Pseudemys scripta. Am J Physiol Regul Integr fish ventricle. Front Physiol 6: 2016. Comp Physiol 242: R216–R219, 1982. 40. Kelly RP, Tunin R, Kass DA. Effect of reduced aortic compliance on 15. Farrell AP, Franklin CE, Arthur PG, Thorarensen H, Cousins KL. cardiac efficiency and contractile function of in situ canine left ventricle. Mechanical performance of an in situ perfused heart from the turtle Circ Res 71: 490–502, 1992. chrysemys scripta during normoxia and anoxia at 5°C and 15°C. J Exp 41. Kinnunen P, Vuolteenaho O, Ruskoaho H. Mechanisms of atrial and Biol 191: 207–229, 1994. brain natriuretic peptide release from rat ventricular myocardium–effect 16. Farrell AP, Franklin CE, Arthur PG, Thorarensen H, Cousins KL. of stretching. Endocrinology 132: 1961–1970, 1993. Mechanical performance of an in situ perfused heart from the turtle 42. Klaiman JM, Fenna AJ, Shiels HA, Macri J, Gillis TE. Cardiac Chrysemys scripta during normoxia and anoxia at 5°C and 15°C. J Exp remodeling in fish: strategies to maintain heart function during tempera- Biol 191: 207–229, 1994. ture change. PLoS One 6: e24464, 2011.

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org CARDIOVASCULAR FUNCTION IN COLD-ACCLIMATED TURTLES R143

43. Krosniunas EH, Hicks JW. Cardiac output and shunt during voluntary 57. Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 activity at different temperatures in the turtle, Trachemys scripta. Physiol years of image analysis. Nat Methods 9: 671–675, 2012. Biochem Zool 76: 679–694, 2003. 58. Shelton G, Burggren W. Cardiovascular dynamics of chelonia during 44. Langille BL, Crisp B. Temperature dependence of blood viscosity in apnea and lung ventilation. J Exp Biol 64: 1976. frogs and turtles: effect on heat exchange with environment. Am J Physiol 59. Shi HY, Jackson DC. Effects of anoxia, acidosis and temperature on the Regul Integr Comp Physiol 239: R248–R253, 1980. contractile properties of turtle cardiac muscle strips. J Exp Biol 200: 45. Licht P, Pavgi S, Denver RJ. The role of hormone binding in the cold 1965–1973, 1997. suppression of hormone stimulation of the pituitary, thyroid, and testis of 60. Shiels HA, White E. The Frank-Starling mechanism in vertebrate cardiac the turtle. Gen Comp Endocrinol 80: 381–392, 1990. myocytes. J Exp Biol 211: 2005–2013, 2008. 46. Malte CL, Malte H, Wang T. Episodic ventilation lowers the efficiency 61. Stecyk JA, Bock C, Overgaard J, Wang T, Farrell AP, Portner HO. of pulmonary CO2 excretion. J Appl Physiol 115: 1506–1518, 2013. Correlation of cardiac performance with cellular energetic components in 47. Methawasin M, Hutchinson KR, Lee EJ, Smith JE 3rd, Saripalli C, the oxygen-deprived turtle heart. Am J Physiol Regul Integr Comp Physiol Hidalgo CG, Ottenheijm CA, Granzier H. Experimentally increasing 297: R756–R768, 2009. titin compliance in a novel mouse model attenuates the frank-starling 62. Stecyk JA, Overgaard J, Farrell AP, Wang T. a-Adrenergic regulation mechanism but has a beneficial effect on diastole. Circulation 129: of systemic peripheral resistance and blood flow distribution in the turtle 1924–1936, 2014. Trachemys scripta during anoxic submergence at 5°C and 21°C. J Exp 48. Miller PJ. Elastin stain. Med Lab Technol 28: 148–149, 1971. Biol 207: 269–283, 2004. 49. Nobata S, Ventura A, Kaiya H, Takei Y. Diversified cardiovascular 63. Stecyk JA, Paajanen V, Farrell AP, Vornanen M. Effect of temperature actions of six homologous natriuretic peptides (ANP, BNP, VNP, CNP1, and prolonged anoxia exposure on electrophysiological properties of the CNP3, and CNP4) in conscious eels. Am J Physiol Regul Integr Comp turtle (Trachemys scripta) heart. Am J Physiol Regul Integr Comp Physiol Downloaded from Physiol 298: R1549–R1559, 2010. 293: R421–R437, 2007. 50. Overgaard J, Stecyk JA, Farrell AP, Wang T. Adrenergic control of the 64. Stecyk JA, Stenslokken KO, Nilsson GE, Farrell AP. Adenosine does cardiovascular system in the turtle Trachemys scripta. J Exp Biol 205: not save the heart of anoxia-tolerant vertebrates during prolonged oxygen 3335–3345, 2002. deprivation. Comp Biochem Physiol A 147: 961–973, 2007. 51. Overgaard J, Wang T, Nielsen OB, Gesser H. Extracellular determi- 65. Trajanovska S, Donald JA. Molecular cloning of natriuretic peptides nants of cardiac contractility in the cold anoxic turtle. Physiol Biochem from the heart of reptiles: loss of ANP in diapsid reptiles and birds. Gen Zool 78: 976–995, 2005. Comp Endocrinol 156: 339–346, 2008.

52. Qiu H, Depre C, Ghosh K, Resuello RG, Natividad FF, Rossi F, 66. Ultsch GR. The ecology of overwintering among turtles: where turtles http://ajpregu.physiology.org/ Peppas A, Shen YT, Vatner DE, Vatner SF. Mechanism of gender- overwinter and its consequences. Biol Rev 81: 339–367, 2006. speciﬁc differences in aortic stiffness with aging in nonhuman primates. 67. Wagenseil JE, Ciliberto CH, Knutsen RH, Levy MA, Kovacs A, Circulation 116: 669–676, 2007. Mecham RP. Reduced vessel elasticity alters cardiovascular structure and 53. R Core Team. R: A Language and Environment for Statistical Computing function in newborn mice. Circ Res 104: 1217–U1233, 2009. (Online). Vienna, Austria: R Foundation for Statistical Computing. http:// 68. Wang T, Altimiras J, Klein W, Axelsson M. Ventricular haemodynam- www.R-project.org/ [2013]. ics in Python molurus: separation of pulmonary and systemic pressures. J 54. Rand PW, Austin WH, Lacombe E, Hunt HE. Viscosity of normal Exp Biol 206: 4241–4245, 2003. human blood under normothermic and hypothermic conditions. J Appl 69. Wang T, Hicks JW. Cardiorespiratory synchrony in turtles. J Exp Biol Physiol 19: 117–122, 1964. 199: 1791–1800, 1996. 55. Reese SA, Crocker CE, Jackson DC, Ultsch GR. The physiology of 70. Webb GJ, Heatwole H, Bavay JD. Comparative cardiac anatomy of hibernation among painted turtles: the midland painted turtle (Chrysemys reptilia. II. Critique of literature on Squamata and Rhynchocephalia. J picta marginata). Respir Physiol 124: 2000. Morphol 142: 1–20, 1974. 56. Saunders DK, Patel KH. Comparison of blood viscosity in red-eared 71. White FN, Hicks JW, Ishimatsu A. Relationship between respiratory by guest on July 20, 2016 sliders (Trachemys scripta) adapted to cold and room temperature. J Exp state and intracardiac shunts in turtles. Am J Physiol Regul Integr Comp Zool 281: 157–163, 1998. Physiol 256: R240–R247, 1989.

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00510.2015 • www.ajpregu.org Thermal remodelling of the ectothermic heart

8. Cold acclimation increases ventricular micromechanical

stiffness and collagen content in the freshwater turtle,

Trachemys scripta.

This chapter is presented in format of a manuscript to be submitted to Biophysical Journal.

Page | 89

Thermal remodelling of the ectothermic heart

Page | 90

Thermal remodelling of the turtle ventricle

Title: Cold acclimation increases ventricular micromechanical stiffness and fibrillar collagen content in the freshwater turtle, Trachemys scripta.

Adam N. Keen1*, James C. McConnell2, Janna Crossley3, John J. Mackrill4, Michael J. Sherratt2, Dane A. Crossley II3, Peter Gardner4, Holly A. Shiels1

1Faculty of Life Sciences, University of Manchester, Manchester, UK 2Centre for Tissue Injury & Repair, Faculty of Medical and Human Sciences, University of Manchester, UK 3Department of Biology, University of North Texas, TX, USA 4Department of Physiology, University College Cork, Ireland 5School of Chemical Engineering and Analytical Science, Manchester Institute of Biotechnology, University of Manchester, UK

*Author of correspondence: [email protected] [email protected] +44 161 275 5092

Short title: Thermal remodeling of the turtle ventricle

Text: 8,120 Figures: 5 Tables: 2 References: 67

Keywords: Compliance, Stiffness, Temperature Acclimation, Collagen, Matrix Metalloproteinase, Extracellular Matrix

List of Abbreviations: AGEs, advanced glycation end-products; ECM, Extracellular matrix; MMP2, matrix metalloproteinase 2; MMP9, matrix metalloproteinase 9; mRNA, messenger RNA; PBS, phosphate buffered saline; RT-qPCR, real time quantitative PCR; Er, reduced modulus; TIMP2, tissue inhibitor of matrix metalloproteinase 2;

Thermal remodelling of the turtle ventricle

ABSTRACT

The freshwater turtle, Trachemys scripta, inhabits temperate regions of north America. Due to its ectothermic nature ambient temperature can influence cardiac load, meaning that prolonged temperature change can trigger a cardiac remodelling response. Here, we investigated the effect of thermal acclimation on the micromechanical stiffness of the ventricle and on the regulation of the ventricular extracellular matrix (ECM) proteins. We hypothesized that increased whole chamber stiffness and myocardial fibrosis associated with chronic cooling would increase localized micromechanical stiffness and that the regulation of ECM proteins would be altered to promote increased collagen deposition. Atomic force microscopy nanoindentation showed that tissue micromechanical stiffness increased following cold acclimation, with an increase in coherency of fibrillar collagen fibres. Chronic cold-induced myocardial fibrosis was associated with a decrease in the endogenous gelatinase activity of collagen degrading matrix metalloproteinases (MMPs) and in increase in the mRNA expression of the tissue inhibitor of MMPs, TIMP2. Together, these results suggest that cold acclimation decreases the myocardial activity of collagen degrading MMPs by an increase in inhibition by TIMPs. These changes in connective tissue regulation, in turn, cause increased coherency of collagen fibres and increases in ventricular passive tension in the cold-acclimated turtle. The regulation of the cardiac compliance and connective tissue content in the freshwater turtle is seasonal and, therefore, must be more plastic than the irreversible myocardial fibrosis associated with human cardiovascular disease.

Thermal remodelling of the turtle ventricle

INTRODUCTION

Chronic change in cardiac load can cause a dynamic remodelling of the vertebrate heart, altering size, form and function (1). Increases in cardiac load may be due to an increase in cardiac preload, often termed a volume overload, or increase in afterload pressure, often termed pressure overload (2). To protect the heart from the increased haemodynamic strain of increased load, the heart may undergo a structural remodelling (3, 4). Collagen I is the primary structural protein in the extracellular matrix (ECM) followed by collagen III, both of which are known as fibillar collagen. Fibrillar collagen is formed by monomers that form fibrils and function as supra-molecular assemblies which, in turn, form fibres (5, 6). Chronic increases in cardiac load are often associated with increased fibrillar collagen deposition in the myocardium, which increases passive tension (i.e. the force the chamber wall exerts during filling) (4, 7, 8). Increased fibrillar collagen offers two functions. It can be cardio-protective, providing additional support to cardiac myocytes to counteract increased haemodynamic stress. In addition, increased myocardial collagen content can improve transduction of myocardial force, and thus systolic function, via the Frank-Starling law of the heart (4, 7). However, excessive fibrosis of the myocardium can drastically reduce cardiac compliance (i.e. the change in pressure for a given change in volume), which, in mammals, is often non- reversible and leads to numerous cardiac pathologies associated with diastolic and/or systolic dysfunction (9-12).

For ectothermic animals ambient temperature changes can significantly alter cardiac load and, hence, periods of prolonged temperature change may trigger a cardiac remodelling response (8). The freshwater turtle, Tracheyms scripta, spends the cold winter months in water and enters a state of periodic inactivity. Despite chronic increase in cardiac load there is no hypertrophic response of the turtle heart (8, 13), but we have recently shown a significant fibrosis in the turtle ventricle following chronic cold exposure (8). The result is a decrease in cardiac compliance but an increased sensitivity to cardiac load (8, 14). Interestingly, as changes in ambient temperature occur seasonally, it is likely that this phenotype is plastic and regresses as temperatures rise during spring and summer months. Therefore, regulation of tissue ECM protein content may be crucial to understanding this response in freshwater turtles. Due to its important structural role in the cardiac wall, connective tissue content is tightly regulated across multiple levels of biological organization. Within tissues, matrix metalloproteinases (MMPs) regulate degradation of several component of the ECM, including collagen (15, 16). In turn, MMP activity is regulated by tissue inhibitors of matrix metalloproteinases (TIMPs) (15). Hence, MMP activity plays a key role in both physiological and pathological remodeling of the cardiovasculature (16-18).

Our aim was to investigate the effects of thermal acclimation on the freshwater turtle ventricle by assessing changes in the localized passive stiffness of ventricular tissue and the changes

Thermal remodelling of the turtle ventricle in regulation of collagen content across multiple levels of biological organization. Following our recent study which showed an increase in passive stiffness of the whole ventricle with associated ventricular fibrosis (8), we used atomic force microscopy to probe differences in the localized micromechanical stiffness of ventricular tissue sections. We then used histological stains to investigate the orientation of fibrillar collagen fibres. We hypothesized that following chronic cold exposure the micromechanical stiffness of ventricular tissue sections would increase due to increases in collagen content and alignment within the tissue. To assess changes in the regulation of fibrillar collagen we focused on 3 areas. We assessed abundance and activation of specific MMPs by gelatin zymography. Then, we assessed endogenous gelatinase activity, or compartmentalization, of MMPs by in situ gelatin zymography. Finally, we used quantitative real-time PCR (RT-qPCR) to analyse mRNA expression of specific collagen (COL1α2) and elastin (2-ELA) genes, as well as the mRNA expression of their regulatory enzymes (MMP2, MMP9, TIMP2). Our hypothesis here was that prolonged cold exposure would cause an up-regulation of collagen promoting genes and a down-regulation of collagen degrading genes. We found that cold acclimation caused an increase in the micromechanical stiffness of ventricular tissue with an increase in coherency of collagen fibres. Lower tissue collagen content in control tissue was correlated to an increase in the endogenous gelatinase activity of MMPs, while increased myocardial collagen content following cold acclimation was correlated with an increase in mRNA expression of the MMP inhibiting TIMP2 gene. Together, these results suggest that ventricular passive tension in the turtle is increased following cold acclimation by increased micromechanical stiffness of tissue sections and increased collagen coherency, regulated by a decrease in the activity of collagen degrading MMPs by an increases in inhibition by TIMPs.

MATERIALS AND METHODS

Experimental animals and acclimation

Male and female red-eared sliders (Trachemys scripta, Schoepff; n = 5 in each group; morphometric data in Table 1) were obtained from lake Lewisville, TX, and transported to the University of North Texas. Turtles were housed in ~50 L plastic containers (dimensions ~50 cm x 50 cm x 100 cm) containing freshwater at a temperature of 25 ± 0.3 °C on a 12 hr: 12 hr light dark cycle. Water quality was maintained by 100 % water changes twice a week and all animals were fed three times per week on commercial reptile feed (Aquatic turtle diet, Mazuri exotic animal nutrition). After 2 weeks, 10 turtles were randomly assigned for cold acclimation where ambient temperature was reduced by 1 °C per day until 5 ± 0.3 °C was reached and turtles were maintained at this temperature for a minimum of 8 weeks prior to experiments (8). Animals selected for cold acclimation were fasted following temperature reduction. The acclimation temperatures (control at 25 ± 0.3 °C; cold at 5 ± 0.3 °C) were chosen based on previous literature (19, 20) to simulate summer and winter conditions. All acclimation

Thermal remodelling of the turtle ventricle temperatures were maintained in a walk in temperature control room (model IR-912L5, Percival Scientific, Perry, IA) and animals were maintained in water. All animals survived acclimation protocols and there were no signs of health deterioration in cold-acclimated turtles. Animal care and surgical preparations followed the approved University of North Texas animal care and use protocol (IACUC #11-007).

Table 1. The body, heart and chamber mass values for cold-acclimated (5 °C) and control (25 °C) freshwater slider turtles (n = 5 in each group).

Cold-Acclimated Control

Body mass (g) 1328.5 ± 43.9 1254.2 ± 31.4

Heart mass (g) 3.04 ± 0.09 3.01 ± 0.11

Ventricle (g) 2.16 ± 0.08 2.00 ± 0.03

Left atrium (g) 0.26 ± 0.034 0.24 ± 0.034

Right atrium (g) 0.40 ± 0.035 0.39 ± 0.035

Relative heart mass (RHM) 0.0022 ± 0.00006 0.0023 ± 0.000067 (g body mass1 x 100)

Relative ventricular mass 0.0016 ± 0.000073 0.0016 ± 0.000001 (RVM) (g body mass1 x 100)

Note: Values given are mean ± S. E. There were no significant differences as determined by GLM with Tukey post-hoc test for comarison between the groups (P < 0.05), n = 10 for both groups.

Tissue processing

Animals were euthanized by intravenous administration of sodium pentobarbital (150 mg kg-1) and the heart excised. The heart was rinsed in phosphate buffered saline (PBS) and weighed. The chambers of the heart were dissected from each other, weighed and bisected down the sagittal plane with one half snap frozen in OCT (Thermo Fisher Scientific, Waltham, MA, USA) by immersion in liquid nitrogen cooled 2-Methylbutane (Sigma-Aldrich, St. Louis, MO, USA) and stored at -80 °C. The other half was fixed in 10 % neutral buffered formalin solution 7

Thermal remodelling of the turtle ventricle

(Sigma-Aldrich, St. Louis, MO, USA) and embedded in paraffin wax so that sections would be cut in the transverse/axial plane.

Tissue histology

Collagen orientation (coherency) was quantified following previously published methodology (5, 21). Briefly, formalin-fixed and paraffin embedded tissue samples were sectioned at 5 μm (Leica CM3050S cryostat, Leica, Wetzlar, Germany) and mounted onto glass slides (Super frost plus, Thermo Fisher Scientific, Waltham, MA, USA). Sections were stained with picro- sirus red to show fibrillar collagen (22) and imaged under bright-field and plane polarized light (Leica, Wetzlar, Germany). Quantitative analyses of collagen orientation (coherency) were conducted on picro-sirus red stained polarized micrographs using the OrientationJ plug-in on ImageJ (21, 23). Three tissue sections were considered for each individual to ensure consistency in measurements. On each tissue section 3 separate image montages were taken along transects across the full diameter of the cross section. All histological analysis was conducted blind to the acclimation group and in all cases these tissue sections were taken from the central 50 % of the ventricle.

Quantitative real-time PCR

Transcript abundance of genes associated with collagen I (COL1α2), elastin (2-ELA) matrix metalloproteinases (MMP2 and MMP9) and a tissue inhibitor of metalloproteinases (TIMP2) were quantified in the ventricles of cold-acclimated and control turtles (n = 5 ventricles for each temperature) using quantitative real-time PCR following, the methodology of Eme et al. (24). All primers designed using Primer 3 from mRNA sequences available on PUBMED (specific marker genes and primers are in Table 2) and purchased from Integrated DNA Technologies (Coralville, IA, USA). Standards were made as previously described (25, 26). Standard curves were log-linear over eight orders of magnitude, which allowed quantification of gene expression in attograms of cDNA per 2.5 ng of input RNA. Total RNA was extracted from heart using a Zymo Research Tissue and Insect RNA MicroPrep™ kit (Zymo Research Corporation, Irvine, CA, USA) and treated with RNAse-free DNase. The concentration and purity of total RNA was measured using a NanoDrop™-1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA). All samples had 260/280 absorbance ratios of 1.8–2.0 and displayed discrete 18S and 28S rRNA bands on agarose gels. cDNA was synthesized by reverse transcribing 300 ng of total RNA using iScript™ Reverse Transcription Supermix as per manufacturer’s instructions (Bio-Rad, Hercules, CA, USA). Real-time PCR was performed with SsoFast™ EvaGreen® Supermix according to the manufacturer’s instructions (Bio-Rad, Hercules, CA, USA). Each reaction contained 5 ml of 2X SsoFast™ EvaGreen® Supermix, 0.3 ml of forward primer and 0.3 ml of reverse primer (final primer concentration = 300 nM), 2 ml of heart cDNA (equivalent to 2.5 ng of input RNA), and water to bring the total volume to

Thermal remodelling of the turtle ventricle

10 ml. Reactions were run on white 384-well plates in a CFX384 Real-Time PCR Detection System. The thermal profile was 94 °C for 30 s to activate DNA polymerase followed by 40 cycles of two-step PCR (94 °C for 1 s and 60 °C for 10 s). Negative control reactions (no reverse transcriptase controls and water controls) demonstrated no contamination with genomic DNA or exogenous PCR products. A melt curve at the end of each run verified that a single product was amplified.

Table 2. The specific marker genes with primers used for quantiative real time PCR analysis.

Gene Primer pair GenBank Function assession number COL1α2 5’-AACTTGCCTTCATGCGTCTG-3’ NW_007359883.1 Fibrosis 5’-GGTTGCCAGTTTCCTCATCC-3’ 1-ELA 5’-GCGTTACAACTGGCACCC-3’ NW_007281577.1 Elastin 5’-CGTCCCTCTCCACAGGTG-3’ deposition MMP2 5’-GGTGCCCAAAAGACAACTGC-3’ NC_024229.1 Inhibit fibrosis 5’-TGTTTCAGGCAGCCCAAAGA-3’ MMP9 5’-CGGAGGATGCAGAAGAAGCT-3’ NW_007281435.1 Inhibit fibrosis 5’-TGATCCCACTTGAGGTCTCC-3’ TIMP2 5’-ATAGAGTTAATTTACACAGCTCCCT-3’ NW_007281340.1 Inhibit MMPs 5’-ATATTCCTTCTTCCCGCCGG-3’

MMP gelatin zymography

To characterize the abundance and activation of specific MMPs we used sodium dodecyl sulfate polyacrylamide electrophoresis (SDS-PAGE)-based gelatin zymography. Snap frozen ventricular tissue was rinsed with phosphate buffered saline, then protein was extracted in ten volumes per wet weight of 0.05% Brij-35, 10 mM CaCl2, 50 mM Tris-HCl pH7.4 on ice, using three 10 s bursts of an MSE Soniprep150 sonicator (exponential probe, 10 µm amplitude). Extracts were cleared by centrifugation at 10,000g for 10 min and protein content was determined using the Bradford assay with bovine serum albumin as a standard. Equal quantities of protein (1 µg / lane) were analysed by gelatin SDS-PAGE, as described by Lødemel et al. (27) et al. 2004 (PMID: 15050523). Conditioned media from HepG2 cell cultures (100 ng protein / lane) and recombinant active human MMP-2 (1 ng protein / lane, Millipore) were used as positive controls. The abundance of each gelatinase band was measured using the ‘Gel’ function of ImageJ.

Thermal remodelling of the turtle ventricle

In situ MMP gelatin zymography

The activity of endogenous MMP gelatinase was semi-quantitatively analysed by in situ zymography of tissue cryosections, following previously published methodology (29)(28, 29). Frozen tissue was sectioned at 10 μm (Leica CM3050S cryostat, Leica, Wetzlar, Germany) and mounted onto glass slides (Super frost plus, Thermo Fisher Scientific, Waltham, MA, USA). Low temperature gelling agarose (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in phosphate buffered saline (to a final concentration of 10 mg ml-1) in an 80 °C water bath and then cooled to 37 °C. DQ gelatin (porcine; Invitrogen, Thermo Fisher Scientific, Waltham, MA,

-1 USA) was dissolved in dH2O (to a concentration of 1 mg ml ) and added to the agarose solution so that it was diluted 1:10. Lastly, 4’, 6’-diamidino-2-phenylindole (DAPI) was added to the agarose/DQ gelatin mixture (at a concentration of 1 μg ml-1). During this time tissue sections were brought to room temperature and washed in PBS to remove excess OCT. Approximately 40 μl of agarose/DAPI/DQ gelatin was added to each tissue section and a coverslip placed on the slide to ensure even film thickness across the sample section. All samples were incubated in the dark for 1 hr at 4 °C and then 18 hrs at room temperature. Following incubation, the samples were visualized, using a fluorescent microscope with a green filter (Leica, Wetzlar, Germany), and imaged immediately. To remove any effect of tissue auto-fluorescence, negative control slides were used to determine the microscope settings for each section. Three tissue sections were considered for each individual to ensure consistency in measurements. On each tissue section 3 separate image montages were taken along transects across the full diameter of the cross section. Following background subtraction, mean fluorescence intensity was calculated for each image, analysed using imageJ. All histological analysis was conducted blind to the acclimation group and in all cases these tissue sections were taken from the central 50 % of the ventricle.

Atomic force microscopy (AFM)

Frozen ventricle was sectioned at 5 μm (Leica CM3050S cryostat, Leica, Wetzlar, Germany) and mounted onto glass slides. Excess OCT was removed with distilled water and the slides were left to dry for ~12 hrs. This methodology is consistent with previous work (30, 31) which outlines that tissue sections are best preserved dehydrated, with rehydration performed when nanomechanical measurements are required. Micro-indentation was carried out using a Bioscope Catalyst AFM (Bruker, Coventry, UK) mounted onto an Eclipse T1 inverted optical microscope (Nikon, Kingston, UK) fitted with a spherically tipped cantilever (nominal radius and spring constant of 1 μm and 3 Nm-1 respectively; Windsor Scientific Ltd., Slough, UK) running Nanoscope Software v8.15 (Bruker, Coventry, UK). The local reduced modulus was determined for each of 400 points in a 50 x 50 μm region, indented at a frequency of 1 Hz with lateral spacing of 2.5 μm. The extend curve was used in conjunction with a contact point based model to calculate the reduced modulus for each indentation (32). For each biological sample,

Thermal remodelling of the turtle ventricle

400 force curves were collected at three distinct 50 μm2 regions. Once all 400 force curves had been generated, quality control was applied whereby any force values falling more than two standard deviations away from the mean value were discarded in order to account for failed indents. Data loss at this stage was less than 10 % (data not shown).

Calculations and statistical analysis

Post hoc analyses of AFM force curves were performed using Nanoscope Analysis v1.40 (Bruker, Coventry, UK), whereby a baseline correction was applied to each curve before a force fit was applied using a Herzian (spherical) model and a maximum force fit of 70 %. All mRNA expression levels were compared to a house-keeping 1-GAPDH gene to determine absolute expression levels for comparison at each acclimation temperature. We tested 2 housekeeping genes; 1-18S and 1-GAPDH, and 1-GAPDH had most stable expression in relation to temperature acclimation. Gene expression patterns were assessed by 2-way ANCOVA and a Holm-Sidak post hoc test to account for multiple comparisons. Differences in collagen coherency, MMP activity and MMP abundance were assessed by GLM (students t- test for parametric, Mann-Whitney U for non-parametric), with each parameter as the test variables and acclimation group as the grouping variable, using SPSS Statistics 20 (IBM, Armonk, NY, USA). For all analyses significance was considered to be P < 0.05, except for atomic force curves where significance was considered at P < 0.005. Values are presented as mean ± S. E. throughout except for atomic force curves where values are mean ± S. D. Statistical details are provided in the figure legends.

RESULTS

Morphology and chamber mass

Temperature acclimation did not alter body mass, heart mass, ventricular mass, relative heart mass (RHM) or relative ventricular mass (RVM; Table 1).

Micromechanical ventricular tissue stiffness

To assess the functional differences in the micromechanical stiffness of ventricular tissue sections we used localized AFM nano-indentation. Three 50 x 50 µm areas were tested on each tissue cryosection shown in the representative control section, imaged under bright-field light, in figure 1A. Nano-indentation showed that micromechanical stiffness of cold-acclimated ventricular tissue was greater than control tissue (P < 0.001; Figure 1B). Mean reduced modulus (Er), and hence localized tissue stiffness, was correlated with temperature, giving higher and, therefore, stiffer values in the cold-acclimated compared with the control tissue. The modulus frequency distribution (Figure 1C) is particularly interesting. There is a shift in

Thermal remodelling of the turtle ventricle the peak of the frequency curve to the right (from ~0.4 MPa in control to ~0.6 MPa following cold acclimation), and a difference in the gradient at which the frequency curve tails off (i.e. an increased number of curves between 1.0-1.5 MPa) following cold acclimation. These results suggest that mechanical remodelling following temperature acclimation is not due to homogenous structural and/or compositional remodelling of the tissue, but rather, isolated or specific regions of the tissue are becoming stiffer.

Coherency of ventricular fibrillar collagen

The organization of collagen fibres can effect overall tissue stiffness (5). We explored whether collagen alignment in the turtle ventricle was influenced by thermal acclimation by quantitatively measuring collagen coherency. Figure 2A shows a representative cold- acclimated ventricle section and figure 2B a representative control ventricle section, stained with picro-sirus red and imaged under polarized light, assessed using OrientationJ. Quantitative analysis of fibrillar collagen coherency showed coherency of fibres was increased by 1.8-fold in turtle ventricle following cold acclimation compared to controls (P < 0.05; Figure 2C).

Thermal remodelling of the turtle ventricle

Figure 1. Micromechanics of ventricular tissue. (A) A white light image of a representative control (25 °C acclimated) ventricular cryosection. The 3 square boxes show the 50 x 50 µm areas used for AFM nanoindentation. (B) mean reduced modulus (Er) and (C) the accumulative frequency curves for each individual Er for cold-acclimated (5 °C; blue) and control (25 °C; red) ventricular tissue. The box and whisker plot represents mean ± interquartile range and 2 standard deviations from the mean. Significance differences in mean Er between groups was assessed by GLM and indicated by * (P < 0.005).

Thermal remodelling of the turtle ventricle

Figure 2. Fibrillar collagen orientation in the ventricular myocardium. Representative polarized light micrographs for (A) cold-acclimated and (B) control ventricular sections, stained with picro-sirus red and assessed using OrientationJ. Fibre orientation is shown by colour, with fibres of the same colour showing coherency. (C) quantification of coherency of organized fibrillar collagen for cold (5 °C; blue), control (25 °C; red) acclimated freshwater turtles (n = 5). Values are presented as mean ± interquartile range and range. Significant differences in collagen coherency between groups were assessed by GLM and indicated by * (P < 0.05).

Thermal remodelling of the turtle ventricle

Regulation of ventricular myocardial collagen content

MMPs regulate collagen degradation in the ECM. We first assessed changes in the endogenous gelatinase activity of MMPs by in situ zymography of ventricular cryosections. Figure 3 shows representative green fluorescence micrographs of cold-acclimated (Figure 3A) and control turtle ventricle (Figure 3B), treated with DQ gelatin to show the localized activity of

MMPs in the tissue section. Figure 3Ai & Bi show the same section with a blue fluorescence micrograph overlaid to show counterstaining with DAPI, which highlights cell nuclei. Semi- quantitative analysis of florescent micrographs revealed a 1.7-fold lower gelatinase activity of MMPs following cold acclimation, compared to controls (P < 0.005; Figure 3C).

MMPs are initially synthesized as inactive pro-forms, the pro-domain must be removed to activate the enzyme (15). Gelatin SDS-PAGE zymography showed the presence of MMPs with molecular weights matching those of human activated MMP-9 and human proMMP-2 and human activated MMP-2 in the turtle ventricle (Figure 4A). We did not find any significant differences in the abundance of proMMP-2, MMP2 or MMP9, nor in the ratio of proMMP-2 to MMP2, following cold acclimation compared to controls, however, there was an trend towards decreased abundance of MMP2 and MMP9 following cold accliamtion (Figure 4B & C).

Collagen synthesis and degradation can be regulated by gene expression. We assessed mRNA expression of specific genes associated with changes in connective tissue content and connective tissue regulators using RT-qPCR. Temperature acclimation did not directly alter mRNA expression of the collagen I gene (COL1α2) or the elastin gene (2-ELA; Figure 5A & B). Similarly, there was no significant change in the mRNA expression of MMP2 or MMP9 following temperature, however, both showed a trend for increased expression in the cold compared to control (Figure 5C & D). However, MMP activity is regulated by tissue inhibitors of metalloproteinases (TIMPs). We found a 6.9-fold increase in the mRNA expression of TIMP2 following cold acclimation compared to controls (P < 0.05; Figure 5E). This could be driving the temperature-induced changes in collagen deposition.

Thermal remodelling of the turtle ventricle

Figure 3. Endogenous matrix metalloproteinase (MMP) activity. Representative fluorescent micrographs for (A) cold-acclimated and (B) control ventricle imaged with a green filter to show gelatinase activity. The same sections were imaged in with a blue filter and the image imposed to show DAPI fluorescence for (Ai) cold-acclimated and (Bi) control turtle tissue. (C) semiquantitative analysis of endogenous MMP activity by in situ gelatinase zymography of ventricular sections for cold-acclimated (5 °C; blue) and control (25 °C; red) freshwater turtles (n = 5). Values are presented as mean ± interquartile range and range. Significant differences in collagen coherency were assessed by GLM and indicated by * (P < 0.05).

Thermal remodelling of the turtle ventricle

Figure 4. Characterization of specific matrix metalloproteinase (MMP) activity in the ventricular myocardium by gelatin SDS-PAGE zymography. (A) Coomassie R250 stained zymogram, indicating the relative molecular weights and abundances of gelatinases in ventricle extracts from cold-acclimated (5 °C; blue) and control (25 °C; red) freshwater turtle (n = 5). The positions of MMP9, proMMP-2 and MMP2 are indicated by arrows on the left-hand side of the gel. (B) abundance of MMP9 and (C) abundance of MMP2 for ventricle of cold-acclimated (5 °C; blue) and (25 °C; red) freshwater turtle (n = 5). Values are presented as mean ± S.E. No significant differences in MMP abundance were assessed by GLM.

Thermal remodelling of the turtle ventricle

Figure 5. Regulation of connective tissue genes. Absolute quantitative real time PCR analysis of mRNA expression of (A) collagen I gene (COLα2), (B) elastin gene (2-ELA), collagen degrading (C) matrix metalloproteinase-2 (MMP2) and (D) matrix metalloproteinase-9 (MMP9), (E) tissue inhibitor of matrix metalloproteinases (TIMP2), and (F) house-keeping gene (1-GAPDH) for cold-acclimated (striped) and control (filled) freshwater turtles (n = 5). Values are presented as mean ± S.E. Significant differences in mRNA expression were assessed by GLM and indicated by * (P < 0.05).

Thermal remodelling of the turtle ventricle

DISCUSSION

Prolonged increase in cardiac load associated with low winter temperatures can cause a structural remodelling response of the ectothermic heart (8, 33, 34). For freshwater turtles, cold winter temperatures prompt the start of a winter hibernation or brumation. To compensate for the changes in cardiac load the heart undergoes a structural remodelling with increasing passive stiffness of the ventricle and myocardial collagen content (8). To simulate winter cold we exposed freshwater turtles to chronic cold and assessed ventricular micromechanics, collagen alignment and ECM protein regulation. Our findings suggest that chronic cold increases micromechanical stiffness of ventricular tissue by increasing collagen content and fibril coherency. The changes in connective tissue content are regulated by a reduction in the gelatinase activity of collagen degrading MMPs and increases in MMP regulating TIMPs.

Increases in the micromechanical stiffness of ventricular tissue

We have previously shown that chronic cold temperature increases the passive stiffness of the whole ventricle and causes a fibrosis of the ventricular myocardium in freshwater turtles (8). Here, we build upon this previous study by assessing ventricular micromechanics using AFM indentation. The advantage of using this technique to functionally assess tissue passive stiffness compared to ex vivo pressure-volume curves is that stiffness is measured at the nanoscale level, to assess whether the temperature-dependent collagen remodelling was associated with a change in local tissue stiffness compared to the ventricle as a whole (35). Indeed, we show an increase in overall tissue stiffness with this technique; however, what is particularly interesting is the changes in the accumulative frequency curve of reduced modulus. These findings suggest an increase in the heterogeneity of the tissue following cold acclimation and, as such, tissue stiffness is modulated by the density of stiff fibres in the tissue. Collagen is an important mediator of tissue tensile strength and stiffness. It is arranged into networks that support cardiomyocytes (36) and alterations in collagen content can affect cardiac micromechanical properties (36, 37). By histological staining with picro-sirus red we have recently shown increases in fibrillar collagen following cold acclimation in turtles (8). The nano-scale resolution of AFM means the contribution of smaller perimysial collagen to overall tissue stiffness will also be measured (35). We found a high frequency of force cuvres with a

Mr between 1.0 and 1.5 MPa, which is consistent with Mr of tissue with high collagen fibre content indented at a lower loading rate (5, 38, 39). Therefore, it appears that stiffness of the turtle ventricular myocardium following thermal acclimation is modulated, at least in part, by the relative content of fibrillar collagen in the tissue.

Fibrillar collagen is a complex molecule. Fibrillar collagen monomers form fibrils which function as supra-molecular assemblies, forming fibres (5, 6). Recent studies suggest that organization of collagen fibres can effect overall tissue stiffness (5). We used coherency of collagen fibrils

Thermal remodelling of the turtle ventricle to assess collagen alignment, with coherency specifically assessing the percentage of collagen fibrils at the same angle (21). These changes in fibrillar collagen alignment, along with previously shown changes in fibrillar collagen content (8), can be used to explain increases in tissue stiffness in the cold. However, it is likely that a number of additional temperature dependent components of the myocardium also add to ventricular stiffening. Firstly, cross-linking by advanced glycation end-products (AGEs) can increase tissue stiffness and is common in a number of cardiac pathologies associated with diabetes (40-45). Both glycation and glycosylation are the result of a protein molecule binding to a sugar molecule, with the latter being enzyme regulated, and, thus, occur due to chronic hyperglycemia (41). The turtle heart is known to accumulate glycogen prior to winter, as cardiac myocytes switch cellular metabolism from fatty acid oxidation to glycolysis (46, 47), which may produce a hyperglycemic environment in cardiac tissue. Secondly, the actin cytoskeleton can modulate cell stiffness (48, 49). Finally, the giant sarcomeric protein titin can alter cellular passive tension (50-53). There are two titin isoforms that exist in the human heart; a shorter and stiffer N2B isoform and a longer and more compliant N2BA isoform (54, 55). The ratio of these two isoforms can result in different levels of titin based passive tension (54, 56, 57). It is currently unclear what contribution titin has in determining stiffness of the turtle myocardium, so future experiments are required.

Regulation of myocardial collagen with thermal acclimation

Fibrillar collagen is the primary structural protein of the ECM and responsible for stiffness in the myocardium (6). Therefore, tight and precise regulation of collagen is critical for correct cardiac function (15, 58). Under normal conditions fibrillar collagen content of the freshwater turtle ventricular myocardium is relatively low, however, following cold acclimation there is an increase in fibrillar collagen content (8). Thus, regulation of deposition and degradation of the ECM proteins must be altered by cold acclimation.

MMPs are important regulators tissue homeostasis involved in complex cell signaling pathways within tissues and cells (15). MMPs regulate degradation of the ECM and, therefore, their activity is closely regulated across multiple biological levels (15, 59). Firstly, MMPs are regulated by gene expression. Genetic regulation is primarily achieved by directly altering levels of MMP transcription, but also the stability of post-transcriptional mRNA regulated by cytokines, nitric oxide and micro-RNA (15, 60, 61). Secondly, MMPs are regulated extracellular localization, i.e. which tissues or cells release MMPs. The cells or tissues that release MMPs determines where the MMPs can act, often referred to as compartmentalization (15, 62). Thirdly, MMPs are regulated by activation, i.e. the ratio of inactive to activated enzyme. MMPs are synthesized as pro-forms, such as proMMP-2, often called zymogens. The pro-domain must be cleaved to allow enzymatic activity of the MMP (15, 62). Finally, MMPs are regulated by inhibition of their activity. MMPs have a number of specific inhibitors, such as tissue

Thermal remodelling of the turtle ventricle inhibitors of matrix metalloproteinases (TIMPs), but can also be inhibited by a number of non- specific proteinase inhibitors, such as α2-macroglobulin (15, 58, 59, 63).

At the genetic level, we did not find any changes in the mRNA expression of the collagen I gene (COL1α2), the elastin gene (2-ELA), or either of the MMP genes (MMP2 and MMP9) we tested. However, there was a trend towards increased expression of COL1α2, MMP2 and MMP9 with cold acclimation. This trend may be explained by the involvement of MMPs in degrading many of the components of the ECM so that remodelling can occur (16-18). Despite no change in MMP gene expression we did see a decrease in the endogenous gelatinase activity of MMPs following cold acclimation, suggesting there was less degradation of collagen fibrils (28, 64). A change in MMP activity without an increase in mRNA expression could be due to the level of activation of MMPs in the tissue, i.e. a change in the ratio of proMMP to activated MMP. However, our gelatin zymography showed no change in the abundance or activation of MMP2 or MMP9. Indeed, if anything the trend was in the opposite direction.

We did see a difference in the expression of TIMP2 mRNA, which was increased following cold acclimation suggesting up-regulation of this gene. TIMP2 inhibits the gelatinase activity of MMPs and, therefore, is associated with an increase in tissue collagen content (15). Overall, our findings suggest that following cold acclimation the collagen degrading of activity of MMPs is suppressed by up-regulation and increased activity of TIMPs. However, we did not directly test the activity of TIMPs in the ventricular tissue so further studies need to be conducted to fully test this theory.

Perspectives and Significance

For the ectothermic heart, changes in ambient temperature alter cardiac load, meaning long- term temperature change can cause the heart to remodel. Here, we detail structural changes in the ventricular ECM and its regulation that cause an increase in the micromechanical tissue stiffness following chronic cold. Functionally, an increase in tissue stiffness can help bear wall stress, with stiff collagen fibres providing tensile strength to support myocytes with the increased haemodynamic stress of pumping cold viscous blood (4, 65, 66). However, increased tissue stiffness and collagen content also increase passive tension of the chamber (i.e. the resistance to filling). In accordance with the Frank-Starling law of the heart increased passive tension can increase systolic pressure, which may help to compensate for the direct effect of temperature reducing contractile force of the ventricle (53, 66).

Many freshwater turtle species are native to temperate regions of north America. The red- eared slider inhabits a region from southern USA to southern Canada (67). Therefore, the structural remodelling of the ventricle reported in this study is a seasonal response. In mammals, increased ventricular stiffness and fibrosis are hallmarks of a number of

Thermal remodelling of the turtle ventricle pathological conditions which are often non-reversible (2, 11). Therefore, the regulation of the cardiac compliance and ventricular ECM protein content in the freshwater turtle needs to be more plastic, increasing stiffness in response to cold and decreasing stiffness in response to warm. The freshwater turtle, therefore, may be a potential new model in investigating ventricular ECM regulation in the reversibility of cardiac stiffness and fibrosis.

ACKNOWLEDGEMENTS

We thank Russell Craddock for helpful advice with the in situ gelatin zymography protocol and analysis. We thank Oliver Wearing and Taylor Shackelford for their help with animal care and sample collection. Atomic force microscopy was conducted in the University of Manchester BioAFM Facility. Histology was performed in the University of Manchester Histology Facility.

COMPETING INTERESTS

The authors declare no competing interests

FUNDING

A.N.K is funded by a BBSRC Doctoral training partnership. The Shiels lab is supported by the Leverhulme Trust (240613). MJS is funded by the Medical Research Council UK (grant G1001398).

REFERENCES

1. Opie, L. H., Commerford, P. J., Gersh, B. J., and Pfeffer, M. A. (2006) Controversies in Cardiology 4 - Controversies in ventricular remodelling. Lancet 367, 356-367

2. Bernardo, B. C., Weeks, K. L., Pretorius, L., and McMullen, J. R. (2010) Molecular distinction between physiological and pathological cardiac hypertrophy: Experimental findings and therapeutic strategies. Pharmacology & Therapeutics 128, 191-227

3. Dorn, G. W., II. (2007) The fuzzy logic of physiological cardiac hypertrophy. Hypertension 49, 962-970

4. Chapman, D., Weber, K. T., and Eghbali, M. (1990) Regulation of fibrillar collagen type-I and type-III and basement-membrane type-IV collagen gene-expression in pressure overloaded rat myocardium. Circulation Research 67, 787-794

5. McConnell, J. C., O’Connell, O. V., Brennan, K., Weiping, L., Howe, M., Joseph, L., Knight, D., O’Cualain, R., Lim, Y., Leek, A., Waddington, R., Rogan, J., Astley, S. M., Gandhi, A., Kirwan, C. C., Sherratt, M. J., and Streuli, C. H. (2016) Increased peri- ductal collagen micro-organization may contribute to raised mammographic density. Breast Cancer Research 18, 1-17

6. Lockhart, M., Wirrig, E., Phelps, A., and Wessels, A. (2011) Extracellular matrix and heart development. Birth Defects Research Part A: Clinical and Molecular Teratology 91, 535-550

Thermal remodelling of the turtle ventricle

7. Collier, P., Watson, C. J., van Es, M. H., Phelan, D., McGorrian, C., Tolan, M., Ledwidge, M. T., McDonald, K. M., and Baugh, J. A. (2012) Getting to the heart of cardiac remodeling; how collagen subtypes may contribute to phenotype. Journal of Molecular and Cellular Cardiology 52, 148-153

8. Keen, A. N., Shiels, H. A., and Crossley II, D. A. (2016) Cardiovascular function, compliance and connective tissue remodeling in the turtle, Trachemys scripta, following thermal acclimation. . American Journal of Physiology - Regulatory, Integrative and Comparative Physiology

9. Segura, A. M., Frazier, O. H., and Buja, L. M. (2012) Fibrosis and heart failure. Heart Failure Reviews 19, 173-185

10. Bishop, S. P., Allard, M. F., Anderson, P. G., Bugaisky, L. B., Doss, L. K., and Sreenan, C. M. (1990) Myocardial hypertrophy, fibrosis and failure due to progressive pressure-overload in the rat. Circulation 82, 300

11. Frey, N., and Olson, E. N. (2003) Cardiac hypertrophy: The good, the bad and the ugly. Annual Review of Physiology 65, 45-79

12. Weber, K. T., Brilla, C. G., and Janicki, J. S. (1993) Myocardial fibrosis: functional significance and regulatory factors. Cardiovascular Research 27

13. Overgaard, J., Wang, T., Nielsen, O. B., and Gesser, H. (2005) Extracellular determinants of cardiac contractility in the cold anoxic turtle. Physiological and Biochemical Zoology 78, 976-995

14. Farrell, A. P., Franklin, C. E., Arthur, P. G., Thorarensen, H., and Cousins, K. L. (1994) Mechanical performance of an in situ perfused heart from the turtle chrysemys scripta during normoxia and anoxia at 5 oC and 15 oC. Journal of Experimental Biology 191, 207-229

15. Löffek, S., Schilling, O., and Franzke, C.-W. (2011) Biological role of matrix metalloproteinases: a critical balance. European Respiratory Journal 38, 191-208

16. Galis, Z. S., and Khatri, J. J. (2002) Matrix Metalloproteinases in Vascular Remodeling and Atherogenesis: The Good, the Bad, and the Ugly. Circulation Research 90, 251- 262

17. Creemers, E. E. J. M., Cleutjens, J. P. M., Smits, J. F. M., and Daemen, M. J. A. P. (2001) Matrix Metalloproteinase Inhibition After Myocardial Infarction: A New Approach to Prevent Heart Failure? Circulation Research 89, 201-210

18. Antonio, R. C., Ceron, C. S., Rizzi, E., Coelho, E. B., Tanus-Santos, J. E., and Gerlach, R. F. (2013) Antioxidant effect of doxycycline decreases MMP activity and blood pressure in SHR. Molecular and Cellular Biochemistry 386, 99-105

19. Hicks, J. M. T., and Farrell, A. P. (2000) The cardiovascular responses of the red- eared slider (Trachemys scripta) acclimated to either 22 or 5 oC. I. Effects of anoxic exposure on in vivo cardiac performance. Journal of Experimental Biology 203, 3765- 3774

20. Hicks, J. M. T., and Farrell, A. P. (2000) The cardiovascular responses of the red- eared slider (Trachemys scripta) acclimated to either 22 or 5 oC. II. Effects of anoxia on adrenergic and cholinergic control. Journal of Experimental Biology 203, 3775- 3784

21. Rezakhaniha, R., Agianniotis, A., Schrauwen, J. T. C., Griffa, A., Sage, D., Bouten, C. V. C., Vosse, F. N., Unser, M., and Stergiopulos, N. (2012) Experimental investigation

Thermal remodelling of the turtle ventricle

of collagen waviness and orientation in the arterial adventitia using confocal laser scanning microscopy. Biomechanics and Modeling in Mechanobiology 11, 461-473

22. Junqueira, L. C. U., Bignolas, G., and Brentani, R. R. (1979) Picrosirius staining plus polarization microscopy, a specific method for collagen detection in tissue-sections. Histochemical Journal 11, 447-455

23. Schneider, C. A., Rasband, W. S., and Eliceiri, K. W. (2012) NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9, 671-675

24. Eme, J., Rhen, T., Tate, K. B., Gruchalla, K., Kohl, Z. F., Slay, C. E., and Crossley, D. A. (2013) Plasticity of cardiovascular function in snapping turtle embryos (Chelydra serpentina): chronic hypoxia alters autonomic regulation and gene expression. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology 304, R966-R979

25. Rhen, T., Metzger, K., Schroeder, A., and Woodward, R. (2007) Expression of Putative Sex-Determining Genes during the Thermosensitive Period of Gonad Development in the Snapping Turtle, Chelydra serpentina. Sexual Development 1, 255-270

26. Ream, M., Ray, A. M., Chandra, R., and Chikaraishi, D. M. (2008) Early fetal hypoxia leads to growth restriction and myocardial thinning. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology 295, R583-R595

27. Lødemel, J. B., Mæhre, H. K., Winberg, J.-O., and Olsen, R. L. (2004) Tissue distribution, inhibition and activation of gelatinolytic activities in Atlantic cod (Gadus morhua). Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology 137, 363-371

28. Akhtar, R., Cruickshank, J. K., Zhao, X., Walton, L. A., Gardiner, N. J., Barrett, S. D., Graham, H. K., Derby, B., and Sherratt, M. J. (2014) Localized micro- and nano-scale remodelling in the diabetic aorta. Acta Biomaterialia 10, 4843-4851

29. Mook, O. R. F., Overbeek, C. V., Ackema, E. G., Maldegem, F. V., and Frederiks, W. M. (2003) In Situ Localization of Gelatinolytic Activity in the Extracellular Matrix of Metastases of Colon Cancer in Rat Liver Using Quenched Fluorogenic DQ-gelatin. Journal of Histochemistry & Cytochemistry 51, 821-829

30. Wallace, J. M., Harding, C. C., and Kemp, A. D. (2012) Effects of Hydration on Nanoscale Structural Morphology and Mechanics of Individual Type I Collagen Fibrils. MRS Proceedings 1465

31. Kemp, A. D., Harding, C. C., Cabral, W. A., Marini, J. C., and Wallace, J. M. (2012) Effects of tissue hydration on nanoscale structural morphology and mechanics of individual Type I collagen fibrils in the Brtl mouse model of Osteogenesis Imperfecta. Journal of Structural Biology 180, 428-438

32. Crick, S. L., and Yin, F. C. P. (2007) Assessing micromechanical properties of cells with atomic force microscopy: importance of the contact point. Biomechanics and Modeling in Mechanobiology 6, 199-210

33. Klaiman, J. M., Fenna, A. J., Shiels, H. A., Macri, J., and Gillis, T. E. (2011) Cardiac Remodeling in Fish: Strategies to Maintain Heart Function during Temperature Change. PLoS ONE 6, e24464

34. Keen, A. N., Fenna, A. J., McConnell, J. C., Sherratt, M. J., Gardner, P., and Shiels, H. A. (2016) The dynamic nature of hypertrophic and fibrotic remodelling in the fish ventricle. Frontiers in Physiology - Intergrative Physiology 6

Thermal remodelling of the turtle ventricle

35. Graham, H. K., Hodson, N. W., Hoyland, J. A., Millward-Sadler, S. J., Garrod, D., Scothern, A., Griffiths, C. E. M., Watson, R. E. B., Cox, T. R., Erler, J. T., Trafford, A. W., and Sherratt, M. J. (2010) Tissue section AFM: In situ ultrastructural imaging of native biomolecules. Matrix Biology 29, 254-260

36. Fomovsky, G. M., Thomopoulos, S., and Holmes, J. W. (2010) Contribution of extracellular matrix to the mechanical properties of the heart. Journal of Molecular and Cellular Cardiology 48, 490-496

37. Graham, H. K., Akhtar, R., Kridiotis, C., Derby, B., Kundu, T., Trafford, A. W., and Sherratt, M. J. (2011) Localised micro-mechanical stiffening in the ageing aorta. Mechanisms of Ageing and Development 132, 459-467

38. Baldwin, Samuel J., Quigley, Andrew S., Clegg, C., and Kreplak, L. (2014) Nanomechanical Mapping of Hydrated Rat Tail Tendon Collagen I Fibrils. Biophysical Journal 107, 1794-1801

39. Andriotis, O. G., Manuyakorn, W., Zekonyte, J., Katsamenis, O. L., Fabri, S., Howarth, P. H., Davies, D. E., and Thurner, P. J. (2014) Nanomechanical assessment of human and murine collagen fibrils via atomic force microscopy cantilever-based nanoindentation. Journal of the Mechanical Behavior of Biomedical Materials 39, 9-26

40. Russo, I., and Frangogiannis, N. G. (2016) Diabetes-associated cardiac fibrosis: Cellular effectors, molecular mechanisms and therapeutic opportunities. Journal of Molecular and Cellular Cardiology 90, 84-93

41. Bodiga, V. L., Eda, S. R., and Bodiga, S. (2013) Advanced glycation end products: role in pathology of diabetic cardiomyopathy. Heart Failure Reviews 19, 49-63

42. Yamazaki, K. G., Gonzalez, E., and Zambon, A. C. (2012) Crosstalk Between the Renin–Angiotensin System and the Advance Glycation End Product Axis in the Heart: Role of the Cardiac Fibroblast. Journal of Cardiovascular Translational Research 5, 805-813

43. Willemsen, S., Hartog, J. W. L., Heiner-Fokkema, M. R., Veldhuisen, D. J., and Voors, A. A. (2011) Advanced glycation end-products, a pathophysiological pathway in the cardiorenal syndrome. Heart Failure Reviews 17, 221-228

44. Méndez, J. D., Xie, J., Aguilar-Hernández, M., and Méndez-Valenzuela, V. (2010) Molecular susceptibility to glycation and its implication in diabetes mellitus and related diseases. Molecular and Cellular Biochemistry 344, 185-193

45. Yan, S. F., Ramasamy, R., and Schmidt, A. M. (2010) The RAGE Axis: A Fundamental Mechanism Signaling Danger to the Vulnerable Vasculature. Circulation Research 106, 842-853

46. Beall, R. J., and Priviter, C. A. (1973) Effects of cold exposure on cardiac metabolism of the turtle, Pseudemys (Chrysemys) picta. American Journal of Physiology 224, 435- 441

47. Jackson, D. C. (2002) Hibernating without Oxygen: Physiological Adaptations of the Painted Turtle. The Journal of Physiology 543, 731-737

48. Qiu, H., Zhu, Y., Gansner, M., Depre, C., Peppas, A., Resuello, R., Natividad, F., Hunter, W., Meininger, G., Genin, G., Elson, E., Vatner, D. E., and Vatner, S. F. (2010) Increases in Vascular Smooth Muscle Stiffness with Aging. Faseb Journal 24, 615- 619

49. Meininger, G. A. (2014) The central importance of the cytoskeleton for increased cell stiffness in cardiovascular disease. Focus on “Diabetes increases stiffness of live 25

Thermal remodelling of the turtle ventricle

cardiomyocytes measured by atomic force microscopy nanoindentation”. American Journal of Physiology - Cell Physiology 307, C908-C909

50. Horowits, R., Maruyama, K., and Podolsky, R. J. (1989) Elastic behavior of connectin filaments during thick filament movement in activated skeletal-muscle. Journal of Cell Biology 109

51. Granzier, H., Helmes, M., and Trombitas, K. (1996) Nonuniform elasticity of titin in cardiac myocytes: A study using immunoelectron microscopy and cellular mechanics. Biophysical Journal 70

52. Watanabe, K., Nair, P., Labeit, D., Kellermayer, M. S. Z., Greaser, M., Labeit, S., and Granzier, H. (2002) Molecular mechanics of cardiac titin's PEVK and N2B spring elements. Journal of Biological Chemistry 277

53. Shiels, H. A., and White, E. (2008) The Frank-Starling mechanism in vertebrate cardiac myocytes. Journal of Experimental Biology 211, 2005-2013

54. Cazorla, O., Freiburg, A., Helmes, M., Centner, T., McNabb, M., Wu, Y., Trombitas, K., Labeit, S., and Granzier, H. (2000) Differential expression of cardiac titin isoforms and modulation of cellular stiffness. Circulation Research 86

55. Patrick, S. M., Shiels, H., and Cazorla, O. (2009) Comparison of length-dependent calcium activation of cardiac myofilaments between the rat and the rainbow trout. Journal of Physiological Sciences 59, 536-536

56. Trombitas, K., Wu, Y., Labeit, D., Labeit, S., and Granzier, H. (2001) Cardiac titin isoforms are coexpressed in the half-sarcomere and extend independently. American Journal of Physiology-Heart and Circulatory Physiology 281

57. Linke, W. A. (2008) Sense and stretchability: The role of titin and titin-associated proteins in myocardial stress-sensing and mechanical dysfunction. Cardiovascular Research 77, 637-648

58. Nagase, H., Visse, R., and Murphy, G. (2006) Structure and function of matrix metalloproteinases and TIMPs. Cardiovascular Research 69, 562-573

59. Brew, K., Dinakarpandian, D., and Nagase, H. (2000) Tissue inhibitors of metalloproteinases: evolution, structure and function1. Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology 1477, 267-283

60. Clark, I. M., Swingler, T. E., Sampieri, C. L., and Edwards, D. R. (2008) The regulation of matrix metalloproteinases and their inhibitors. The International Journal of Biochemistry & Cell Biology 40, 1362-1378

61. Yan, C., and Boyd, D. D. (2007) Regulation of matrix metalloproteinase gene expression. Journal of Cellular Physiology 211, 19-26

62. Ra, H.-J., and Parks, W. C. (2007) Control of Matrix Metalloproteinase Catalytic Activity. Matrix biology : journal of the International Society for Matrix Biology 26, 587- 596

63. Ahmed, S. H., Clark, L. L., Pennington, W. R., Webb, C. S., Bonnema, D. D., Leonardi, A. H., McClure, C. D., Spinale, F. G., and Zile, M. R. (2006) Matrix metalloproteinases/tissue inhibitors of metalloproteinases - Relationship between changes in proteolytic determinants of matrix composition and structural, functional, and clinical manifestations of hypertensive heart disease. Circulation 113, 2089-2096

64. Bruggink, A. H., van Oosterhout, M. F. M., de Jonge, N., Cleutjens, J. P. M., van Wichen, D. F., van Kuik, J., Tilanus, M. G. J., Gmelig-Meyling, F. H. J., van den Tweel, 26

Thermal remodelling of the turtle ventricle

J. G., and de Weger, R. A. (2007) Type IV collagen degradation in the myocardial basement membrane after unloading of the failing heart by a left ventricular assist device. Laboratory Investigation 87, 1125-1137

65. Collier, P., Ledwidge, M., and McDonald, K. (2012) Diagnostics and therapeutic interventions in myocardial interstitial disease, a previously neglected pathology. Qjm- an International Journal of Medicine 105, 721-724

66. Glezeva, N., and Baugh, J. A. (2014) Role of inflammation in the pathogenesis of heart failure with preserved ejection fraction and its potential as a therapeutic target. Heart Failure Reviews 19, 681-694

67. Ultsch, G. R. (2006) The ecology of overwintering among turtles: where turtles overwinter and its consequences. Biological Reviews 81, 339-367

Thermal remodelling of the turtle ventricle

Thermal remodelling of the ectothermic heart

9. Metabolic and biochemical remodelling of the thermally acclimated freshwater turtle ventricle using Fourier transform

infra-red imaging spectroscopy.

This chapter is presented in format of a manuscript to be submitted to the

Journal of Biological Chemistry.

Page | 91

Thermal remodelling of the ectothermic heart

Page | 92

Metabolic and biochemical remodelling of the turtle heart

Title: Metabolic and biochemical remodelling of the thermally acclimated freshwater turtle ventricle using Fourier transform infra-red imaging spectroscopy.

Adam N. Keen1, 2, Holly A. Shiels1, John Marrin1, Alex Henderson2, Peter Gardner2*

1Faculty of Life Sciences, University of Manchester, Manchester, UK 2School of Chemical Engineering and Analytical Science, Manchester Institute of Biotechnology, University of Manchester, UK

*Author of correspondence:

Short title: Metabolic and biochemical remodelling of the turtle heart.

Text: 9,470 Figures: 7 Tables: 2 References: 71

Keywords: Fatty Acid Oxidation, Glycolysis, Collagen, Hypoxia, Hibernation

List of Abbreviations: ADP, adenosine diphosphate; ATP, adenosine triphosphate; FAO, fatty acid oxidation; GLM, general linear model; H & E, Haematoxylin and Eosin; PC, principal component; PCA, principal component analysis

Metabolic and biochemical remodelling of the turtle heart

ABSTRACT

The primary energy production pathway for the adult heart is adenosine triphosphate (ATP) production by mitochondrial fatty acid oxidation (FAO). However, under conditions of low oxygen availability reliance on FAO can decrease and the primary energy production pathways switch to glycolysis. This metabolic state, typical of the foetal heart, is a hallmark of pathological hypertrophy. The hypoxia tolerance of the freshwater turtle myocardium makes these animals an interesting potential biomedical model. We investigated the effect of chronic cooling (from 25 °C to 5 °C) on the turtle heart and hypothesized that metabolite content of the myocardial tissue would change in advance of winter anoxia to allow energetic remodeling from FAO to glycolysis. Using Fourier transform infrared (FTIR) imaging spectroscopy, we saw an increase in absorption bands characteristic of protein and glycogen following cold acclimation, with a corresponding decrease in the absorption bands characteristic of lipids and phosphates. We used histological stains to assess tissue content of lipids and glycogen, which again suggested a 2.9-fold increase in glycogen and a 2.2-fold decrease in lipid content of ventricular tissue following chronic cold exposure. Together these results may suggest that the cellular energetics of the turtle ventricle switch from FAO to glycolysis during cold winter temperatures. We propose the freshwater turtle as a potential new biomedical model for investigating a reversible metabolic remodeling response of the myocardium.

Metabolic and biochemical remodelling of the turtle heart

INTRODUCTION

In the adult vertebrate heart, adenosine triphosphate (ATP) production by fatty acid oxidation (FAO) is the primary energy production pathway (Allard et al., 1994). Energy is released when the terminal phosphate ion of the ATP molecule is cleaved by the ATPase, leaving adenosine diphosphate (ADP) and inorganic phosphate (Ingwall, 2009). Under normal situations fatty acid synthesis by mitochondria produces a sufficient and continuous supply of ATP for all myocyte functions, including contraction. However, under conditions of extreme ATP demand glycolysis and phosphotransferase reactions catalyzed by creatine kinase and adenylate kinase can provide alternative pathways for additional ATP production (Ingwall, 2009).

A metabolic shift from mitochondrial FAO to glycolysis and pyruvate oxidation is a common feature of pathological hypertrophy of the mammalian heart; a metabolic state typical of a foetal heart (Allard et al., 1994; Akki et al., 2008; Kolwicz et al., 2012). To maintain the tricarboxcylic acid (TCA) cycle, intermediate anaplerosis reactions (i. e. replenishment of the TCA cycle intermediates) are also increased (Pound et al., 2009; Kolwicz et al., 2012). This switch in cardiac energetics can be preferable when oxygen availability is limited by marginally improving short-term myocardial oxygen efficiency per mole of ATP produced (Ingwall, 2009). However, ATP production by glycolysis is less efficient at powering the high energy demand of cardiac myocytes for prolonged periods. As a result, this change in energy pathway is often associated with ATP deficiency, causing decreased myocardial energetics, impaired cardiac function and an exhaustion of contractile reserves (Ingwall and Weiss, 2004; Neubauer, 2007).

Red-eared slider turtles (Trachemys scripta) inhabit temperature regions of north America and spend winter in a state of periodic inactivity (Ultsch, 2006). For ectothermic animals, reductions in ambient temperature directly reduce metabolic rate and some species of freshwater turtle are able to further suppress metabolic rate via active suppression (Herbert and Jackson, 1985; Jackson, 2002; Galli and Richards, 2012). During winter turtles hibernate or brumate in bodies of freshwater which may freeze over, dramatically reducing availability of oxygen (Stecyk et al., 2008). These animals have high anoxia tolerance, partially due to their ability to switch cardiac energetics from FAO to anaerobic glycolysis during hibernation (Beall and Priviter, 1973; Jackson, 2002; Stecyk et al., 2008). Interestingly, electrophysiology studies indicate that temperature, not hypoxia, drives the remodelling response (Stecyk et al., 2007a). In this way, chronic cold can be considered the trigger that drives the turtle heart to remodel functionally (Keen et al., 2016b) and biochemically to prepare to the onset of hypoxia/anoxia (Stecyk et al., 2007a). For this reason, the freshwater turtle is emerging as a new biomedical model to study myocardial remodeling and hypoxia and anoxia tolerance of the myocardium (Galli and Richards, 2014).

Metabolic and biochemical remodelling of the turtle heart

Fourier transform infrared (FTIR) spectroscopy is a powerful technique for studying molecular structures. As a beam of infrared light is passed through a sample the energy from the radiation excites the molecules and causes them to vibrate. The way molecules vibrate is determined by molecular dipoles, associated with molecular modes of vibration and rotation (Stuart, 1997). The discrete fundamental modes of vibration mean that each molecule will absorb infrared radiation at discrete wavelengths and, therefore, at a particular region of the spectrum (Stuart, 1997). In a complex molecule, many individual bonds and functional groups vibrate at different wavelengths to form a complex spectrum. The power of FTIR to identify biochemical molecules means it is becoming a prominent technique in biomedical fields (Ellis and Goodacre, 2006; Baker et al., 2014). In particular, there is interest in developing methods of disease diagnosis using FTIR by looking at the fingerprint region or specific markers, such as shifts in lipid and glycogen levels in cancer biopsies, (Gazi et al., 2005; Gazi et al., 2006; Harvey et al., 2007; Baker et al., 2008; Baker et al., 2009; Hughes et al., 2014) and assessing remodelling of the myocardium following heart failure (Gough et al., 2003; Toyran et al., 2007; Zheng et al., 2010; Cheheltani et al., 2012; Kuwahara et al., 2014; Staniszewska et al., 2014).

Here, we used temperature acclimation to invoke cardiac remodeling that could occur with seasonal temperature change. We investigated the effects of chronic cold (from 25 °C to 5 °C) on the tissue morphology and biochemistry of the turtle ventricle. We hypothesized that chronic cooling would reduce the energetic reliance on mitochondrial FAO and cause a metabolic shift to glycolysis and pyruvate oxidation, similar to the metabolic shift seen in pathological remodelling of the mammalian heart. We used Fourier transform infrared (FTIR) imaging spectroscopy to assess changes in the biochemistry of ventricular tissue cryosections and then histologically stained tissue sections for lipid and glycogen to validate the FTIR results. Following cold acclimation, FTIR spectroscopy showed increases in overall protein content compared to controls. The cold-acclimated ventricle also showed a decrease in overall lipid content both through the overall infrared spectral profile and through histological analysis with oil red O. In contrast, infrared absorption at wavenumbers characteristic of glycogen and products of glycolysis, such a lactate, show an increase following cold acclimation. In support of this, glycogen content was increased following cold acclimation assessed histologically with periodic acid Schiff (PAS) stain. Previous studies have indicated that chronic cold, rather than low oxygen levels (Stecyk et al., 2007a), are key to trigger remodelling of excitable and contractile properties of the turtle myocardium. Our results support this idea and expand it by suggesting that cold temperature initiates a remodelling of myocardial energetics and fuel deposition in advance of winter hypoxia or anoxia.

Metabolic and biochemical remodelling of the turtle heart

EXPERIMENTAL PROCEDURES

Experimental animals and acclimation

Husbandry of, male and female red-eared sliders (Trachemys scripta, Schoepff; n = 5 for each group; mean body mass = 1352 ± 69 g) have been describe previously in Keen et al. (2016b). Briefly animals were housed in ~50 L plastic containers (dimensions ~50 cm x 50cm x 100 cm) containing freshwater at a temperature of 25 ± 0.3 °C or 5 ± 0.3 ° on a 12 hr: 12 hr light dark cycle for a minimum of 8 weeks prior to experiments (Keen et al., 2016). Animals selected for cold acclimation were fasted following temperature reduction as previous work, and feeding trials showed animals refused food when cooled. The acclimation temperatures (control at 25 ± 0.3 °C; cold at 5 ± 0.3 °C) were chosen based on previous literature (Hicks and Farrell, 2000a; b) to simulate summer and winter conditions. All animals survived the acclimation protocols and there were no signs of poor health in either group. Animal care and surgical preparations adhered to the University of North Texas animal care and use protocol (IACUC #11-007).

Tissue processing

Animals were killed by intravenous administration of sodium pentobarbital (150 mg kg-1) and the heart excised. The heart was rinsed in phosphate buffered saline and weighed. The chambers of the heart were dissected from each other, weighed and bisected down the sagittal plane with one half snap frozen in OCT (Thermo Fisher Scientific, Waltham, MA, USA) by immersion in liquid nitrogen cooled 2-Methylbutane (Sigma-Aldrich, St. Louis, MO, USA) and stored at -80 °C. The other half was fixed in 10 % neutral buffered formalin solution (Sigma- Aldrich, St. Louis, MO, USA), processed and embedded in paraffin wax so that sections would be cut in the transverse/axial plane. Tissue was transported by courier to the University of Manchester.

Fourier transform infrared (FTIR) imaging spectroscopy

To assess changes in the biochemistry of turtle ventricular myocardium following cold acclimation we used FTIR imaging spectroscopy. Frozen ventricle was sectioned at 5 µm

(Leica CM3050S cryostat, Leica, Wetzlar, Germany), mounted on calcium fluoride (CaF2) slides and stored at -80 °C. Frozen tissue was removed from the freezer and thawed in a vacuum desiccator for 60 minutes, prior to being placed in the purge box of the spectrometer. Transmission mode FTIR imaging spectroscopy was performed on a Varian 670-IR spectrometer coupled with a Varian 620-IR imaging microscope (Agilent Technonlgies, CA, USA) equipped with a 128 x 128 pixel liquid nitrogen cooled Mercury-Cadmium-Telluride focal plane array (FPA) detector. Data were collected in the 950-3800 cm-1 range, at a spectral

Metabolic and biochemical remodelling of the turtle heart resolution of 5 cm-1, with the co-addition of 96 scans for sample spectra and 256 scans for the background spectra. At a pixel size of 5.5 µm, a tissue sampling area of ~6000 x 6000 µm was captured by the hyperspectral image.

Tissue Histology

To determine overall tissue morphology and semi-quantitatively assess tissue lipid and glycogen we used histological techniques on tissue sections. For overall tissue morphology, serial tissue sections to those used for FTIR analysis were stained with haematoxylin and eosin (H & E) stain. To assess lipid content of the tissue, frozen ventricular tissue was sectioned at 10 µm (Leica CM3050S cryostat, Leica, Wetzlar, Germany), mounted onto glass slides (Super frost plus, Thermo Fisher Scientific, Waltham, MA, USA) and stained with oil red O stain for lipid. A negative control was prepared for each section by taking serial sections and treating them with acetone prior to the oil red O protocol to remove all lipid. To assess glycogen content of the tissue, formalin-fixed tissue samples were processed, embedded in paraffin wax, sectioned at 5 μm (Leica RM2255 microtome, Leica, Wetzlar, Germany) mounted onto glass slides and stained with Periodic acid-Shiff (PAS) stain for glycogen. A negative control was prepared for each section by taking serial sections and digesting the glycogen in amylase before the PAS staining protocol. Images were analysed using bright-field microscopy (Leica Wetzlar, Germany) and ImageJ software (Schneider et al., 2012). Lipid and glycogen content of the tissues, compared to their corresponding negative controls, were determined by pixel count and then expressed as a percentage increase compared to that sections control. Three tissue sections were considered for each individual to ensure consistency in measurements. On each tissue section 3 separate image montages were taken along transects across the full diameter of the cross section. All histological analysis was conducted blind to the acclimation group.

Calculations and statistical analyses

Infrared spectral data was imported into MATLAB (MATLAB 2014a, Mathworks, USA) and quality tested by the amide I region (1597-1738 cm-1). The absorbance values to determine which spectra were accepted or rejected by this quality test were determined separately for each hyperspectral image by reference to an image of the tissue section. The band associated with ambient gas-phase CO2 was removed, as was any data outside of the specified wavenumber range. Data was subjected to a PCA noise reduction (Reddy and Bhargava, 2010; Bhargava, 2012), vector normalized and subject to RMieS-EMSC correction with 100 iterations using a MatrigelTM spectrum as the initial reference point (Bassan et al., 2012). Following the scatter correction, regions of interest were taken to ensure there were no artifacts due to sample preparation and that we captured the full biochemistry. The region of interest data was then subject to a K-fold algorithm which randomized the spectra and then

Metabolic and biochemical remodelling of the turtle heart condensed them into 1000 average spectra for each individual turtle. All subsequent analyses were conducted on these K-folded spectra. Following analysis of mean spectra, we transformed the data into the 2nd derivative. Transforming the data to the 2nd derivative is commonly used in FTIR spectroscopy to reduce background effects and deconvolute some of the broad peaks of the spectrum, which encompass a number of smaller peaks, i.e. enhance peak resolution. Although a number of previous studies have used the 2nd derivative spectra to interpret their findings, interpretation can be challenging as the 2nd derivative is actually showing the change in rate of infrared absorption. Therefore, although the peaks are in the same position the size of the peak is not necessarily correlated to absorption. Therefore, interpretation of the 2nd derivative spectra should be used with caution and instead it should be used to highlight changes in bands, rather than band intensities.

Lipid and glycogen content of the tissues were determined by pixel count compared to the corresponding negative control section, using ImageJ. Significance was considered at P < 0.05, determined by general linear model (GLM) (SPSS Statistics, IBM, USA). For lipid and glycogen content values are expressed as mean ± S. E. Statistical analyses are detailed in figure legends.

RESULTS

Biochemical homogeneity of turtle ventricular tissue sections

This is the first time that turtle cardiac tissue has been assessed using FTIR. To ensure we included all cell types and areas of the tissue in our analysis, we first assessed the homogeneity of ventricular cryosections. Figure 1 shows representative micrographs of turtle ventricular tissue sections stained with H & E for cold-acclimated (Figure 1A) and control (Figure 1B) animals. To assess the biochemical homogeneity of the turtle ventricle we generated FTIR hyperspectral tissue images based on protein absorption (the intensity of amide I at 1595-1695 cm-1), lipid absorption (the intensity of 2830-3030 cm-1) and glycogen absorption (the intensity of 1035-1045 cm-1) (Figure 1C, D &E). Intensity of absorption is shown by darkness of the colour of the images. Each image suggested homogeneity of the tissue with absorption evenly distributed. Next we used principal component analysis (PCA) to determine the important features of the tissues biochemistry and then used the principal component (PC) scores to perform a K-means cluster analysis of the tissue section, as previously performed by Hughes et al. (2014). Figure 2 shows a representative control ventricular section with clusters based on PC 1 and 2 (Figure 2A), PC 1-3 (Figure 2B), PC 1- 4 (Figure 2C) and PC 1-5 (Figure 2D). We then extracted the spectra from the areas of tissue that contributed to each cluster; their spectral profiles suggested a high degree of heterogeneity between clusters (Figure 2E). Finally, we generated 3 regions of interest from the tissue and PCA to determine separation of these regions by their biochemistry. We found

Metabolic and biochemical remodelling of the turtle heart that the regions showed some separation along PC 1, which described 77.1 % of the variation in the data (Figure 3A). The main features of PC 1 are shown in the corresponding PC loadings plot, which suggests the key variation in tissue biochemistry is between areas with high absorption in bands relating to protein or areas with high absorption in bands relating to lipids and carbohydrates (Figure 3B). We did not find any separation by PC 2, PC 3 or PC4 (Figure 3B, C &D).

Figure 1. Homogeneity of turtle ventricle tissue cryosections. Representative bright-field micrograph of (A) cold-acclimated and (B) control turtle ventricle, stained with haematoxylin and eosin (H & E). Representative hyperspectral images of control ventricle imaged using transmission mode Fourier transform infrared (FTIR) imaging spectroscopy based on the overall (C) protein (1595-1695 cm-1), (D) lipid (2830-3030 cm-1) and (E) glycogen (1035-1045 cm-1) absorption profiles of the tissue.

Metabolic and biochemical remodelling of the turtle heart

Figure 2. K-means cluster analysis of ventricular cryosections. Following data acquisition by FTIR imaging spectroscopy, principal component analysis (PCA) was performed on the data to determine the regions of key spectral significance. The principal component (PC) scores were used to perform a K-means cluster analysis on the tissue. (A) shows a representative tissue cryosection with the K-means cluster analysis for PC 1 (blue) & 2 (red). (B) shows the same section but with K-means cluster analysis including PC 1-3 (blue, red, green), (C) PC 1- 4 (blue, red, green, black) and (D) PC 1-5 (blue, red, green, black, magenta). The spectra for each area were then extracted from the cluster analysis. (E) shows the mean the spectrum for each cluster in the 3800-1000 cm-1 range, of each PC is drawn in the colour it appears on the tissue image.

Metabolic and biochemical remodelling of the turtle heart

Figure 3. Principal component analysis (PCA) for FTIR spectra, for regions of interest (ROI) on the same ventricular cryosection. (A) principal component (PC) scores plot for PC 1 and PC 2 for ROI 1 (blue), ROI 2 (green) and ROI 3 (red). (B) the corresponding PC loadings plot for PC 1 and PC 2. (C) PC scores plot for PC 3 and PC 4 for ROI 1 (blue), ROI 2 (green) and ROI 3 (red). (D) the corresponding PC loadings plot for PC 1 and PC 2. In the scores plot, positive values show correlation with that particular PC and negative values show differences to that particular PC. In the loadings plot, positive values show the key wavenumber bands that are increased in that particular PC and negative values show the key wavenumber bands that are reduced in that particular PC.

Metabolic and biochemical remodelling of the turtle heart

Changes in infrared spectra of ventricular tissue sections with thermal acclimation

To assess the biochemical changes in ventricular tissue following thermal acclimation we performed FTIR imaging spectroscopy of tissue cryosections. The differences in infrared absorption between acclimation groups across key spectral bands are detailed in Table 1. Following cold acclimation overall protein tissue content was increased, characterized by increases in infrared absorbance peaks in the amide B, mean centered at 3069 cm-1, the amide I band, mean centered at 1655 cm-1 and composed of the stretching of the C=O bond of protein (~80 %) coupled with C-N stretching (~10 %) and N-H bending (~10 %), the amide II band, mean centered at 1545 cm-1 and composed of N-H bending (~60 %) and C-N stretching of proteins (40 %) (Sivakumar et al., 2013; Prakash et al., 2015) (Figure 4). In addition, there was an increase in the band mean centered at 1154 cm-1, which is assigned to the CO-O-C asymmetric stretch of glycogen and nucleic acids (Figure 4). Finally, there was an increase in infrared absorbance in the peak mean centered at 1034 cm-1, which correlates with the COH deformation of glycogen (Figure 4). Conversely, lipid signatures were reduced in ventricular tissue following cold acclimation. The asymmetric and symmetric stretch of the CH3 group of lipids, at peaks centered at 2953 cm-1 and 2874 cm-1 (Prakash et al., 2015), the asymmetric

-1 -1 and symmetric stretch of CH2 of lipids, at 2926 cm and 2855 cm , and the ester C=O stretch

-1 of lipid, mean centered at 1742 cm , were all reduced (Figure 4). However, the CH2 bending of amino acid side chains of peptides and proteins, mean centered at 1451 cm-1, and stretching of amino acid residues, at 1387 cm-1, were also reduced. Finally, the COO- stretching of fatty acids and amino acids, mean centered at 1397 cm-1, and the asymmetric and symmetric stretching modes of nucleic acid, at 1240 cm-1 and 1080 cm-1, were also reduced (Çakmak et al., 2003). PCA showed separation in the data by PC 1, explaining 72.1 % of the variation, and some separation along PC 2, explaining 15 % of the variation (Figure 5A). The predominant spectral profiles in PC 1 suggest that following cold acclimation ventricular tissue has higher levels of protein and glycogen, but lower levels of lipid and phosphates (Figure 5B). The predominant spectral profiles in PC 2 suggest that following cold acclimation tissue there are increases in amide A and glycogen and lactate absorption (Figure 5C).

Metabolic and biochemical remodelling of the turtle heart

Table 1. Differences in the mean infrared absorption spectra of cold-acclimated and control freshwater turtle ventricles at various peaks of interest and their peak assignment.

Metabolic and biochemical remodelling of the turtle heart

Following analysis of mean spectra, we transformed the data into the 2nd derivative. In cold- acclimated animals, the 2nd derivative of spectra showed an overall increase in protein with increases in two regions of the amide A band, at 3302 cm-1 and 3270 cm-1, and at the amide B peak at 3067 cm-1 (Figure 6). There were also changes in the spectra in 3 positions of the amide I band, at 1692 cm-1, 1657 cm-1, and 1645 cm-1, and at 2 positions within the amide II band, at 1547 cm-1 and 1514 cm-1 (Figure 6). The deconvolution also showed an increase in the band mean centered at 1338 cm-1, which is indicative of collagen amino acid side chain vibrations. The 2nd derivative of spectra also suggested an increase in bands indicative of glycolysis as they suggest glycogen and lactate in the tissue, at 1154 cm-1 and 1029 cm-1, and a change in the asymmetric and symmetric stretching modes of phosphodiester of nucleic acid, at 1080 cm-1 (Figure 6). Following cold acclimation the second derivative of mean spectra also suggested an increase in overall lipid with changes in the asymmetric stretch of CH2, at 2920

-1 -1 -1 cm , the CH3, at 2852 cm , and the ester C=O stretch of lipids, at 1742 cm (Diem et al., 1999; Prakash et al., 2015). Additionally, there was a decrease in in asymmetric stretch of

- -1 PO2 , at 1235 cm (Figure 6).

Figure 4. Mean spectra for cold-acclimated (5 °C; blue) and control (25 °C; red) turtle ventricle cryosections (n = 5, for each group). Spectra has been quality tested, been subjected to a noise reduction, vector normalization and RMieS scatter correction. Numbers denote key spectral bands of interest, which are detailed in Table 1. 1, 3069 cm-1; 2, 2953 cm-1; 3, 2924 cm-1; 4, 2872 cm-1; 5, 2853 cm-1; 6, 1736 cm-1; 7, 1655; cm-1 8, 1545 cm-1; 9, 1451 cm-1; 10, 1387 cm-1; 11, 1308 cm-1; 12, 1240 cm-1; 13, 1152 cm-1; 14, 1086 cm-1; 15, 1040 cm-1.

Metabolic and biochemical remodelling of the turtle heart

Figure 5. Principal component analysis of FTIR spectra for cold-acclimated (5 °C; blue) and control (25 °C; red) turtle ventricle cryosections (n = 5, for each group). (A) principal component (PC) scores plot for PC 1 and PC 2 for cold-acclimated and control tissue. (B) the corresponding PC loadings plot for PC 1 and PC 2. (C) PC scores plot for PC 3 and PC 4 for cold-acclimated and control tissue. (D) the corresponding PC loadings plot for PC 3 and PC 4. In the scores plot, positive values show correlation with that particular PC and negative values show differences to that particular PC. In the loadings plot, positive values show the key characteristics that are increased in that particular PC and negative values show the key characteristics that are reduced in that particular PC. Spectra from each individual turtle have been subjected to a K-folding algorithm which reduces the spectral number to 1000 mean spectra per individual.

Metabolic and biochemical remodelling of the turtle heart

Figure 6. The 2nd derivative of mean spectra for cold-acclimated (5 °C; blue) and control (25 °C; red) acclimated turtle ventricle cryosections (n = 5). Spectra has been quality tested, been subjected to a noise reduction, vector normalization and RMieS scatter correction. Numbers denote key spectral bands of interest, which are detailed in Table 1. 1, 3302 cm-1; 2, 3270 cm-1; 3, 3067 cm-1; 4, 2920 cm-1; 5, 2851 cm-1; 6, 1738 cm-1; 7, 1692 cm-1; 8, 1657 cm-1; 9, 1645 cm-1; 10, 1547 cm-1; 11, 1514 cm-1; 12, 1377 cm-1; 13, 1338 cm-1; 14, 1235 cm-1; 15, 1152 cm-1; 16, 1080 cm-1; 17, 1029 cm-1.

Histological analysis of tissue sections following thermal acclimation

To complement the results from the FTIR analysis we used histological stains to assess tissue content of lipid and glycogen. We used oil red O stain, which selectively stains lipid droplets in red. Figure 7A show a representative micrograph of cold-acclimated turtle ventricle and figure 7B shows a representative micrograph of control turtle. The lipid content of the turtle ventricle was decreased by 2.2-fold following cold acclimation (Figure 7C). Figure 7D show a representative micrograph of cold-acclimated turtle ventricle and figure 7E shows a representative micrograph of control turtle ventricle stained with PAS stain which selectively stains glycogen in purple. We found that following cold acclimation glycogen content of the tissue was 2.9-fold greater than controls (Figure 7F).

Metabolic and biochemical remodelling of the turtle heart

Figure 7. Histological staining for lipid and glycogen in the turtle ventricle. A representative micrograph of (A) cold-acclimated and (B) control turtle ventricular cryosections stained with oil red O, which stains lipid droplets red. (C) Semi-quantitative analysis of oil red O staining for cold-acclimated (5 °C; blue) and control (25 °C; red) acclimated turtle ventricle cryosections, compared to their negative control slides (n = 5 for each group). A representative micrograph of (D) cold-acclimated and (E) control turtle formalin fixed ventricular sections stained with periodic acid Schiff (PAS), which stains glycogen purple. (F) Semi-quantitative analysis of PAS staining for cold-acclimated (5 °C; blue) and control (25 °C; red) acclimated turtle ventricle cryosections, compared to their negative control slides (n = 5 for each group). Significant differences between acclimation groups was determined by general linear model and indicated by *. Values are presented as mean ± S.E.

Metabolic and biochemical remodelling of the turtle heart

DISCUSSION

For efficient function the heart needs a continuous and sufficient supply of ATP, which is predominantly achieved by mitochondrial FAO (Allard et al., 1994). A shift in cardiac energetics from FAO to glycolysis is characteristic of early stage pathological remodelling in the human heart, returning to a metabolic state typical of the foetal heart (Allard et al., 1994; Akki et al., 2008; Kolwicz et al., 2012). The freshwater turtle switches predominant cardiac energetics to glycolysis for winter hibernation or brumation, which is thought to explain, at least in part, the extreme anoxia tolerance of these animals (Beall and Priviter, 1973; Jackson, 2002; Stecyk et al., 2008; Galli and Richards, 2012). Here, we used FTIR imaging spectroscopy and histological staining to assess the biochemistry of tissue sections following thermal acclimation, with particular interest in levels of the compounds involved in these cardiac energetic pathways. We found that following cold acclimation overall levels of lipid and phosphates were reduced, while levels of glycogen and markers of glycolysis were increased. Overall tissue protein was increased and the 2nd derivative of infrared absorption profiles suggested changes in secondary confirmation. Together, our results may suggest that chronic cold triggers a metabolic shift to glycolysis. This is particularly interesting in relation to earlier work (Stecyk et al., 2007a), which shows that the turtle heart remodels ion channels in response to a cold prior to a hypoxic or anoxic trigger. The turtle heart may, thus, provide a new model for studying plasticity of cardiac energetics.

Homogeneity of the turtle ventricle tissue biochemistry

At first look the turtle ventricle appears to be a very homogeneous tissue with both the H & E stain and with the hyperspectral images based on the protein, lipid and carbohydrate absorptions (Figure 1). However, it is clear from the K-means cluster analysis and PCA that the tissue is, in fact, markedly heterogeneous in its biochemistry. The spectra extracted from the K-means cluster analysis showed significantly different profiles. Probably the 2 most different were clusters 1 and 2 as together PC 1 and PC 2 explain most of the biochemical variation. Between these two spectra there are large differences in proteins, shown by different absorption profiles in the amide A, amide I and amide II, and in phosphates.

The reasons for such large changes in spectra are unclear. There are a number of different cells types in cardiac tissue. Ventricular cardiomyocytes form the muscular wall, cardiac fibroblasts regulate the extracellular matrix, endothelial cells form the endocardium, and there are numerous extracellular matrix components (Xin et al., 2013). It is possible that the infrared absorption profiles of clusters reflect these differences in tissue. Indeed, Wood et al. (1998) found large spectral differences between cell types with fibroblasts having strong bands due to overlapping collagen and phosphate and endothelial cells having reduced phosphate and

Metabolic and biochemical remodelling of the turtle heart glycogen bands. Similarly, differences have been shown in fibroblasts during their role in remodelling of connective tissue through cancer progression (Holton et al., 2011). In addition, the presence and/or differing levels of blood cells in the tissue may have influenced overall spectra, with different blood cell types also showing characteristic spectral profiles (Wood et al., 1998). To minimize this effect all samples were washed in PBS before fixation, however, residual blood may still have influenced spectra.

However, with FTIR analysis we have to be careful not to include and artifacts of the technique. The tissues tested in this study were thawed cryosections. The advantage of cryosections is that the tissue is not treated with organic solvents so the chemistry should be as close to in vivo as possible but, as a disadvantage, photo-oxidation can occur (Stitt et al., 2012) and dehydration can alter cellular chemistry (Shim and Wilson, 1996). By using formalin-fixed and paraffin embedded (FFPE) tissue some of the chemistry is lost, due to the chemical fixation and processing, but sample preservation is improved (Lyng et al., 2011). Furthermore, there are issues due to the physics of light when using FTIR imaging spectroscopy, such as scattering effects at the edge of tissue, which have to be accounted for. By using samples embedded in wax, the sample thickness becomes consistent and, therefore, scattering is reduced. If you dewax the tissue or use a cryosection, tissue thickness varies and, in porous tissue such as the heart, scattering at holes can cause large changes in the spectra. However, for this study cryosections were preferable as they have not been exposed to organic solvents and, therefore, retain all metabolic substrates. Consequently, to correct for potential artifacts we applied a quality test for each section based on the amide I band to remove any tissue that did not have a sufficient level of protein. We also used the RMieS scatter correction algorithm created by Bassan et al. (2012). This algorithm works by applying peak and baseline corrections at certain points in the spectra to shift known peaks and baseline sections to their correct position. The correction appeared to work well for this tissue, removing the scattering, but it was created for dewaxed FFPE sections rather than frozen, which are known to have a number of key peaks at slightly different wavenumbers due to the processing (Shim and Wilson, 1996; Faolain, 2005).

Changes in cellular energetics following cold acclimation

To maintain contractile function the heart requires a sufficient and continuous supply of ATP, with sarcomeric myosin and ion pumps being the highest consumers (Ingwall, 2009). During cardiac hypertrophy, muscle mass increases due to an increase in size of constituent myocytes. Hypertrophy is commonly associated with increased angiogenesis to improve oxygen delivery to the cardiac muscle. However, it still often causes limited oxygen availability. A shift in cardiac energetics from mitochondrial FAO to glycolysis and pyruvate oxidation can be preferable when oxygen available is limited and is a common early event in the development of hypertrophy (Ingwall et al., 1985; Shen et al., 1999). Although beneficial in the

Metabolic and biochemical remodelling of the turtle heart short-term, by increasing myocardial oxygen efficiency per mole of ATP produced (Ingwall, 2009), in the failing human myocardium ATP supply is reduced to around ~30 % of normal levels (Ingwall and Weiss, 2004; Neubauer, 2007).

During winter hibernation or brumation, freshwater turtles can find themselves in increasingly hypoxic or anoxic conditions (Stecyk et al., 2008). As a result, turtles may use chronic cold as the trigger (Stecyk et al., 2007a) to switch the predominant metabolic substrate of the cardiac myocytes from FAO to anaerobic glycolysis (Beall and Priviter, 1973; Jackson, 2002; Galli and Richards, 2012) and actively suppress metabolic rate to prevent a mismatch in ATP supply to demand (Galli and Richards, 2014). In the present study, turtles were exposed to chronic cold but in normoxic water with access to air; conditions that may correlate to the onset of winter hibernation. Therefore, although quantification of the activity of enzymes involved in respiration is needed, the observations from our results suggest that cardiac energetics may begin to change due to the onset of cold temperature, before hypoxic or anoxic water conditions. Following cold acclimation we saw an overall decrease in the spectral profile of triglycerides and cholesterol esters. The CH stretching region of the spectrum (from ~2700 - 3100 cm-1) can be used to estimate overall lipid content (Gough et al., 2003) and we show absorption in this region was decreased following cold acclimation. This is a very complex region of the spectrum, created by a large number of nearly identical vibrations, super-imposed to form a broad band. Within the broad band there are two regions characteristic of saturated long chain hydrocarbons, showing the symmetric and asymmetric stretching modes of CH2 (~2926 cm-1 and 2855 cm-1) (Snyder, 1979; Gough et al., 2003). There is another pair of

- distinct regions which show the symmetric and asymmetric stretching modes of CH3 (2953 cm

1 -1 and 2874 cm ) (Gough et al., 2003). The ratio between the asymmetric stretch of CH3 and the asymmetric stretch of CH2 can be used to determine the degree of lipid saturation in the tissue, with a higher absorption of CH2 indicating a higher number of C=C double bonds (Gough et al., 2003; Staniszewska et al., 2014). Following cold acclimation the ratio of lipid unsaturation was increased (Table 2). From this region it is also possible to determine the ratio of methyl to methylene groups and, therefore, the degree of branching of lipids (Gough et al., 2003; Staniszewska et al., 2014). We found that following cold acclimation branching was increased, suggesting a higher tissue content of branched rather than long chain fatty acids (Table 2) (Staniszewska et al., 2014). Lastly, it is possible to assess the state of biological membranes by the ratio of asymmetric to symmetric CH2 stretch (Staniszewska et al., 2014). Following cold acclimation we found that the disorder of lipid acyl chains was also increased (Table 2). We also saw a decrease in the C=O stretching of cholesterol esters and triglycerides (~1740 cm-1). Due to the difference in shape of the shoulder it is likely that the difference is due to a reduction in tissue cholesterol esters rather than a decrease triglycerides as the carbonyl stretching for triglycerides is more typically ~1745 cm-1, at which point the spectra are closer (Staniszewska et al., 2014). However, this band is susceptible to rapid oxidation once cryosections have thawed (Stitt et al., 2012). We ensured that samples were

Metabolic and biochemical remodelling of the turtle heart imaged as quickly as possible after removal from the freezer, and as such within a time-frame that oxidation is minimal (Stitt et al., 2012). However, we cannot rule out that oxidation may have affected out data in this wavenumber region.

Table 2. Differences in the infrared absorption ratios of macromolecules in cold-acclimated and control freshwater turtle ventricles.

The mean spectral profiles suggest an increase in glycogen following cold acclimation with increased infrared absorption at 1152 cm-1 and 1034, which are indicative of glycogen, glycolipids and glycoproteins (Wood et al., 1998; Toyran et al., 2007; Saravanakumar et al., 2011; Staniszewska et al., 2014). In addition, following conversion of the mean spectra to the 2nd derivative, the deconvolution of the spectra allowed resolution of peaks at 1080 cm-1 and 1029 cm-1. These peaks also showed differences with thermal acclimation and may be indicative of glycolysis as they further suggest differences in glycogen and differences in lactate, which is a metabolic by-product of glycolysis (Saravanakumar et al., 2011).

Changes in protein content of the myocardium following thermal acclimation

The early stages of cardiac remodelling of the human heart are often compensatory, enabling the heart to maintain optimal cardiac function in a time of increased afterload, reduced contractile function or increase physiological requirement (Mone et al., 1996; Opie et al., 2006; Akki et al., 2008). A common remodelling response is cardiac hypertrophy, where the heart increases in size due to an increase in the size of its constituent myocytes, which increases support of the cardiac wall and provides additional contractile units (Bishop, 1990; Frey and Olson, 2003; Bernardo et al., 2010). During prolonged cold temperature, the rainbow trout shows an increase in ventricular muscle mass, which compensates for the direct effects of

Metabolic and biochemical remodelling of the turtle heart cold temperature on cardiac function (Farrell et al., 1988; Driedzic et al., 1996; Vornanen et al., 2005; Klaiman et al., 2014; Keen et al., 2016a). Ventricular hypertrophy provides an increased number of contractile units, to compensate for the negative ionotropic effects of cold, and gives greater structural strength, to protect the cardiac wall from the increased haemodynamic stress of increased blood viscosity (Aho and Vornanen, 1999; Klaiman et al., 2011; Klaiman et al., 2014; Keen et al., 2016a). Together these features help defend cardiac function at low temperatures (Aho and Vornanen, 1999; Vornanen et al., 2005; Klaiman et al., 2011; Klaiman et al., 2014; Keen et al., 2016a)

As turtles become relatively dormant during winter, their need to defend cardiac function is not as great as they adapt in large part through metabolic suppression (Beall and Priviter, 1973; Jackson, 2002; Stecyk et al., 2008). Therefore, an overall increase in heart mass is not a common feature, however, collagen content of the tissue has been shown to increase (Keen et al., 2016b). Following cold acclimation we saw an increase in overall protein content of the ventricular tissue, shown by increases in the amide B, amide I and amide II bands. The lipid to protein ratio was decreased following cold acclimation (Table 2). The amide bands are, again, composed of many vibrational bands super-imposed to make the broad band they appear as. Therefore, to try and identify specific differences in proteins it is useful to perform a deconvolution of the spectra, which we did by converting it to the 2nd derivative. Once in the 2nd derivative it is possible to determine peaks that suggest different types of secondary conformations. Following cold acclimation we saw changes in bands associated with anti- parallel turns (1692 cm-1), α-helices (1657 cm-1) and β-sheet conformations (1645 cm-1) (Kretlow et al., 2006; Staniszewska et al., 2014). Collagen is a complex fibrillar molecule and, therefore, there are numerous spectral bands that have been assigned to collagen including 1206 cm-1, 1238 cm-1 and 1280 cm-1 (Wang et al., 2005; Cheheltani et al., 2012). As the myocardium generally has relatively low levels of collagen, and due to artifacts of converting the spectra to the 2nd derivative, it was difficult to resolve these spectral bands, but we did see a change at 1202 cm-1. However, Cheheltani et al. (2012) showed that collagen can be resolved well by the peak at 1338 cm-1, which arises due to specific collagen amino acid side chains. We also saw an increase in this band following cold acclimation, which is consistent with previous studies which suggest a fibrosis of the myocardium following cold acclimation (Keen et al., 2016b).

Changes in phosphate macromolecules following thermal acclimation

Following cold acclimation we saw decreases in the infrared absorption of bands at ~1086 cm- 1 and 1240 cm-1, which are where the main symmetric and asymmetric phosphate group vibrations appear (Whelan et al., 2011). These bands contain the vibrations of phosphate molecules in DNA, RNA, phospholipids and phosphorylated proteins (Severcan et al., 2005; Krafft et al., 2007; Krafft et al., 2009; Palombo et al., 2009; Hackett et al., 2011;

Metabolic and biochemical remodelling of the turtle heart

Saravanakumar et al., 2011), which makes it difficult to determine the contribution of each molecule. It is also possible that if ATP stores are depleted due to higher myocardial ATP demand than supply during glycolysis, the reduced ATP molecules would influence this spectral region as was found using 31P-NMR spectroscopy (Stecyk et al., 2009).

Perspectives and conclusions

The anoxia tolerance of the freshwater turtle has long been of interest to physiologists, with the cellular mechanisms behind myocardial tolerance key to understanding this extreme physiological response. The hearts of all but a few vertebrates are highly susceptible to oxygen deprivation with very short survival times and lasting damage caused by acute hypoxia and anoxia (Stecyk et al., 2008). The freshwater turtle, therefore, has potential to become a model species for studying cardiac defense to hypoxic injury after events such as ischemia (Ravingerová, 2007). It is likely that a switch from FAO to glycolysis is an early cardio- protective event to reduce oxygen demand of the myocardium. Indeed, neonatal mammals have been shown to have much greater cardiac tolerance to hypoxia than mammals, likely due to cardiac energetics (Singer, 1999). During cardiomyopathies, this switch has negative connotations due to its inefficiency in generating sufficient ATP and the effect is often irreversible. However, the seasonal nature of turtle hibernation/brumation suggests that in the turtle heart cellular energetics are plastic and can switch to be optimal in different conditions. These features make turtles an exciting model for studying reversibility of cardiac metabolic disease and the benefits of alternative energy production pathways in the heart.

Here, we have used novel FTIR imaging spectroscopy techniques of ventricular cryosections and then histologically stained tissue sections for lipid and glycogen to validate the FTIR results. Following cold acclimation, FTIR spectroscopy showed increase in overall protein content compared to controls. Specific changes in infrared absorption profiles suggest that the changes in protein alter secondary confirmation of proteins and increase the connective tissue content of the ventricular myocardium. The cold-acclimated ventricle also showed a decrease in overall lipid content both through overall infrared spectral profile and through histological analysis with oil red O. In contrast, cold-acclimated tissue showed increased infrared absorption at wavenumbers characteristic of glycogen and products of glycolysis, such as lactate, and showed increased glycogen staining when histologically assessed with periodic acid Schiff (PAS) stain. Our results give further support to the hypothesis that cold temperature is the primary regulator of the turtle hibernation or brumation, triggering a remodelling of myocardial energetics in advance of winter hypoxia or anoxia (Stecyk et al., 2007b; Stecyk et al., 2008). As this response occurs seasonally the energetic remodelling of the turtle heart must be plastic and switch back to FAO upon ambient warming in spring. We suggest the turtle as a potential model for studying a reversibility of energetic remodelling, often associate with human cardiomyopathies, and cardiac defenses to hypoxic injury.

Metabolic and biochemical remodelling of the turtle heart

ACKNOLEDGMENTS

Thank you to Drs Caryn Hughes, Paul Bassan, Melody Hernadez-Jiminez, Graeme Clemens and Mike Pilling for their help with FTIR experiments, analysis and helpful discussions throughout these studies. Thank you to Dr Dane Crossley II for acclimating the animals and supplying samples. Histology was conducted in the University of Manchester Histology Facility and we thank Peter Walker for his expertise and guidance.

REFERENCES

Aho, E., Vornanen, M., 1999. Contractile properties of atrial and ventricular myocardium of the heart of rainbow trout Oncorhynchus mykiss: Effects of thermal acclimation. Journal of Experimental Biology 202, 2663-2677.

Akki, A., Smith, K., Seymour, A.-M.L., 2008. Compensated cardiac hypertrophy is characterised by a decline in palmitate oxidation. Molecular and Cellular Biochemistry 311.

Allard, M.F., Schonekess, B.O., Henning, S.L., English, D.R., Lopaschuk, G.D., 1994. Contribution of oxidative metabolism and glycolysis to ATP production in hypertrophied hearts. American Journal of Physiology 267.

Baker, M.J., Gazi, E., Brown, M.D., Shanks, J.H., Clarke, N.W., Gardner, P., 2009. Investigating FTIR based histopathology for the diagnosis of prostate cancer. Journal of Biophotonics 2.

Baker, M.J., Gazi, E., Brown, M.D., Shanks, J.H., Gardner, P., Clarke, N.W., 2008. FTIR- based spectroscopic analysis in the identification of clinically aggressive prostate cancer. British Journal of Cancer 99.

Baker, M.J., Trevisan, J., Bassan, P., Bhargava, R., Butler, H.J., Dorling, K.M., Fielden, P.R., Fogarty, S.W., Fullwood, N.J., Heys, K.A., Hughes, C., Lasch, P., Martin-Hirsch, P.L., Obinaju, B., Sockalingum, G.D., Sulé-Suso, J., Strong, R.J., Walsh, M.J., Wood, B.R., Gardner, P., Martin, F.L., 2014. Using Fourier transform IR spectroscopy to analyze biological materials. Nat. Protocols 9, 1771-1791.

Bassan, P., Sachdeva, A., Kohler, A., Hughes, C., Henderson, A., Boyle, J., Shanks, J.H., Brown, M., Clarke, N.W., Gardner, P., 2012. FTIR microscopy of biological cells and tissue: data analysis using resonant Mie scattering (RMieS) EMSC algorithm. Analyst 137, 1370-1377.

Beall, R.J., Priviter, C.A., 1973. Effects of cold exposure on cardiac metabolism of the turtle, Pseudemys (Chrysemys) picta. American Journal of Physiology 224, 435-441.

Bernardo, B.C., Weeks, K.L., Pretorius, L., McMullen, J.R., 2010. Molecular distinction between physiological and pathological cardiac hypertrophy: Experimental findings and therapeutic strategies. Pharmacology & Therapeutics 128, 191-227.

Bhargava, R., 2012. Infrared Spectroscopic Imaging: The Next Generation. Applied Spectroscopy 66, 1091-1120.

Bishop, S.P., 1990. The myocardial cell: normal growth, cardiac hypertrophy and response to injury. Toxicologic Pathology 18, 438-453.

Metabolic and biochemical remodelling of the turtle heart

Borchman, D., Giblin, F.J., Leverenz, V.R., Reddy, V.N., Lin, L.-R., Yappert, M.C., Tang, D., Li, L., 2000. Impact of Aging and Hyperbaric Oxygen In Vivo on Guinea Pig Lens Lipids and Nuclear Light Scatter. Investigative Ophthalmology & Visual Science 41, 3061-3073.

Çakmak, G., Togan, I., Uğuz, C., Severcan, F., 2003. FT-IR Spectroscopic Analysis of Rainbow Trout Liver Exposed to Nonylphenol. Applied Spectroscopy 57, 835-841.

Cheheltani, R., Rosano, J.M., Wang, B., Sabri, A.K., Pleshko, N., Kiani, M.F., 2012. Fourier transform infrared spectroscopic imaging of cardiac tissue to detect collagen deposition after myocardial infarction. Journal of Biomedical Optics 17, 056014.

Diem, M., Boydston-White, M.S., Chiriboga, L., 1999. Infrared Spectroscopy of Cells and Tissues: Shining Light on to a Novel Subject. Applied Spectroscopy 53, 148–161.

Driedzic, W.R., Bailey, J.R., Sephton, D.H., 1996. Cardiac adaptations to low temperature non-polar teleost fish. Journal of Experimental Zoology 275.

Ellis, D.I., Goodacre, R., 2006. Metabolic fingerprinting in disease diagnosis: biomedical applications of infrared and Raman spectroscopy. Analyst 131.

Faolain, E.O., M. B. Hunter, J. M. Byrne, P. Kelehan, M. McNamara, H. J. Byrne and F. M. Lyng, 2005. A study examining the effects of tissue processing on human tissue sections using vibrational spectroscopy. Vibrational Spectroscopy 38, 121–127.

Farrell, A.P., Hammons, A.M., Graham, M.S., Tibbits, G.F., 1988. Cardiac growth in rainbow- trout, salmo-gairdneri. Canadian Journal of Zoology-Revue Canadienne De Zoologie 66, 2368-2373.

Frey, N., Olson, E.N., 2003. Cardiac hypertrophy: The good, the bad and the ugly. Annual Review of Physiology 65, 45-79.

Galli, G.L.J., Richards, J.G., 2012. The effect of temperature on mitochondrial respiration in permeabilized cardiac fibres from the freshwater turtle, Trachemys scripta. Journal of Thermal Biology 37, 195-200.

Galli, G.L.J., Richards, J.G., 2014. Mitochondria from anoxia-tolerant animals reveal common strategies to survive without oxygen. Journal of Comparative Physiology B 184, 285-302.

Gazi, E., Baker, M., Dwyer, J., Lockyer, N.P., Gardner, P., Shanks, J.H., Reeve, R.S., Hart, C.A., Clarke, N.W., Brown, M.D., 2006. A correlation of FTIR spectra derived from prostate cancer biopsies with Gleason grade and tumour stage. European Urology 50.

Gazi, E., Gardner, P., Hart, C., Shanks, J., Baker, N., Dwyer, J., Lockyer, N., Miyan, J., Brown, M., Clarke, N., 2005. Diagnosing prostate tumour biopsies using infrared microspectroscopy. European Urology Supplements 4.

Gough, K.M., Zelinski, D., Wiens, R., Rak, M., Dixon, I.M.C., 2003. Fourier transform infrared evaluation of microscopic scarring in the cardiomyopathic heart: Effect of chronic AT1 suppression. Analytical Biochemistry 316, 232-242.

Hackett, M.J., McQuillan, J.A., El-Assaad, F., Aitken, J.B., Levina, A., Cohen, D.D., Siegele, R., Carter, E.A., Grau, G.E., Hunt, N.H., Lay, P.A., 2011. Chemical alterations to murine brain tissue induced by formalin fixation: implications for biospectroscopic imaging and mapping studies of disease pathogenesis. Analyst 136, 2941-2952.

Harvey, T.J., Henderson, A., Gazi, E., Clarke, N.W., Brown, M., Faria, E.C., Snook, R.D., Gardner, P., 2007. Discrimination of prostate cancer cells by reflection mode FTIR photoacoustic spectroscopy. Analyst 132.

Metabolic and biochemical remodelling of the turtle heart

Herbert, C.V., Jackson, D.C., 1985. Temperature Effects on the Responses to Prolonged Submergence in the Turtle Chrysemys picta bellii. II. Metabolic Rate, Blood Acid-Base and Ionic Changes, and Cardiovascular Function in Aerated and Anoxic Water. Physiological Zoology 58, 670-681.

Hicks, J.M.T., Farrell, A.P., 2000a. The cardiovascular responses of the red-eared slider (Trachemys scripta) acclimated to either 22 or 5 oC. I. Effects of anoxic exposure on in vivo cardiac performance. Journal of Experimental Biology 203, 3765-3774.

Hicks, J.M.T., Farrell, A.P., 2000b. The cardiovascular responses of the red-eared slider (Trachemys scripta) acclimated to either 22 or 5 oC. II. Effects of anoxia on adrenergic and cholinergic control. Journal of Experimental Biology 203, 3775-3784.

Holton, S.E., Walsh, M.J., Bhargava, R., 2011. Subcellular localization of early biochemical transformations in cancer-activated fibroblasts using infrared spectroscopic imaging. Analyst 136, 2953-2958.

Hughes, C., Gaunt, L., Brown, M., Clarke, N.W., Gardner, P., 2014. Assessment of paraffin removal from prostate FFPE sections using transmission mode FTIR-FPA imaging. Analytical Methods 6, 1028-1035.

Ingwall, J.S., 2009. Energy metabolism in heart failure and remodelling. Cardiovascular Research 81.

Ingwall, J.S., Kramer, M.F., Fifer, M.A., Lorell, B.H., Shemin, R., Grossman, W., Allen, P.D., 1985. The creatine-kinase system in normal and diseased human myocardium. New England Journal of Medicine 313.

Ingwall, J.S., Weiss, R.G., 2004. Is the failing heart energy starved? On using chemical energy to support cardiac function. Circulation Research 95.

Jackson, D.C., 2002. Hibernating without Oxygen: Physiological Adaptations of the Painted Turtle. The Journal of Physiology 543, 731-737.

Keen, A.N., Fenna, A.J., McConnell, J.C., Sherratt, M.J., Gardner, P., Shiels, H.A., 2016a. The dynamic nature of hypertrophic and fibrotic remodelling in the fish ventricle. Frontiers in Physiology - Intergrative Physiology 6.

Keen, A.N., Shiels, H.A., Crossley II, D.A., 2016b. Cardiovascular function, compliance and connective tissue remodeling in the turtle, Trachemys scripta, following thermal acclimation. . American Journal of Physiology - Regulatory, Integrative and Comparative Physiology 331, R133-R143.

Klaiman, J.M., Fenna, A.J., Shiels, H.A., Macri, J., Gillis, T.E., 2011. Cardiac Remodeling in Fish: Strategies to Maintain Heart Function during Temperature Change. PLoS ONE 6, e24464.

Klaiman, J.M., Pyle, W.G., Gillis, T.E., 2014. Cold acclimation increases cardiac myofilament function and ventricular pressure generation in trout. Journal of Experimental Biology 217, 4132-4140.

Kolwicz, S.C., Olson, D.P., Marney, L.C., Garcia-Menendez, L., Synovec, R.E., Tian, R., 2012. Cardiac-Specific Deletion of Acetyl CoA Carboxylase 2 Prevents Metabolic Remodeling During Pressure-Overload Hypertrophy. Circulation Research 111, 728-+.

Krafft, C., Codrich, D., Pelizzob, G., Sergoa, V., 2009. Raman and FTIR imaging of lung tissue: bronchopulmonary sequestration. Journal of Raman Spectroscopy, 595–603.

Metabolic and biochemical remodelling of the turtle heart

Krafft, C., Kirsch, M., Beleites, C., Schackert, G., Salzer, R., 2007. Methodology for fiber-optic Raman mapping and FTIR imaging of metastases in mouse brains. Analytical and Bioanalytical Chemistry 389, 1133-1142.

Kretlow, A., Wang, Q., Kneipp, J., Lasch, P., Beekes, M., Miller, L., Naumann, D., 2006. FTIR- microspectroscopy of prion-infected nervous tissue. Biochim Biophys Acta 1758, 948-959.

Kuwahara, M., Bannai, K., Segawa, H., Miyamoto, K.-i., Yamato, H., 2014. Cardiac remodeling associated with protein increase and lipid accumulation in early-stage chronic kidney disease in rats. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease 1842, 1433-1443.

Lyng, F., Gazi, E., Gardner, P., 2011. Preparation of Tissues and Cells for Infrared and Raman Spectroscopy and Imaging.

Mone, S.M., Sanders, S.P., Colan, S.D., 1996. Control mechanisms for physiological hypertrophy of pregnancy. Circulation 94, 667-672.

Neubauer, S., 2007. Mechanisms of disease. The failing heart - An engine out of fuel. New England Journal of Medicine 356.

Opie, L.H., Commerford, P.J., Gersh, B.J., Pfeffer, M.A., 2006. Controversies in Cardiology 4 - Controversies in ventricular remodelling. Lancet 367, 356-367.

Palombo, F., Shen, H., Benguigui, L.E.S., Kazarian, S.G., Upmacis, R.K., 2009. Micro ATR- FTIRspectroscopic imaging of atherosclerosis: an investigation of the contribution of inducible nitric oxide synthase to lesion composition in ApoE-null mice. Analyst 134, 1107-1118.

Pound, K.M., Sorokina, N., Ballal, K., Berkich, D.A., Fasano, M., LaNoue, K.F., Taegtmeyer, H., O'Donnell, J.M., Lewandowski, E.D., 2009. Substrate-Enzyme Competition Attenuates Upregulated Anaplerotic Flux Through Malic Enzyme in Hypertrophied Rat Heart and Restores Triacylglyceride Content Attenuating Upregulated Anaplerosis in Hypertrophy. Circulation Research 104.

Prakash, C., Kamboj, V.K., Ahlawat, P., Kumar, V., 2015. Structural and molecular alterations in arsenic-induced hepatic oxidative stress in rats: a FTIR study. Toxicological & Environmental Chemistry 97, 1408-1421.

Ravingerová, T., 2007. Intrinsic defensive mechanisms in the heart: a potential novel approach to cardiac protection against ischemic injury. General Physiology Biophysics 26, 3-13.

Reddy, R.K., Bhargava, R., 2010. Accurate histopathology from low signal-to-noise ratio spectroscopic imaging data. Analyst 135, 2818-2825.

Saravanakumar, M., Manivannan, J., Sivasubramanian, J., Silambarasan, T., Balamurugan, E., Raja, B., 2011. Molecular metabolic fingerprinting approach to investigate the effects of borneol on metabolic alterations in the liver of nitric oxide deficient hypertensive rats. Molecular and Cellular Biochemistry 362, 203-209.

Schneider, C.A., Rasband, W.S., Eliceiri, K.W., 2012. NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9, 671-675.

Severcan, F., Gorgulu, G., Gorgulu, S.T., Guray, T., 2005. Rapid monitoring of diabetes- induced lipid peroxidation by Fourier transform infrared spectroscopy: Evidence from rat liver microsomal membranes. Analytical Biochemistry 339, 36-40.

Shen, W.Q., Asai, K., Uechi, M., Mathier, M.A., Shannon, R.P., Vatner, S.F., Ingwall, J.S., 1999. Progressive loss of myocardial ATP due to a loss of total purines during the development of heart failure in dogs - A compensatory role for the parallel loss of creatine. Circulation 100.

Metabolic and biochemical remodelling of the turtle heart

Shim, M.G., Wilson, B.C., 1996. The effects of ex vivo handling procedures on the near- infrared Raman spectra of normal mammalian tissues. Photochemistry and Photobiology 63, 662–671.

Singer, D., 1999. Neonatal tolerance to hypoxia: a comparative-physiological approach. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology 123, 221-234.

Sivakumar, S., Sivasubramanian, J., Khatiwada, C.P., Manivannan, J., Raja, B., 2013. Determination of aluminium induced metabolic changes in mice liver: A Fourier transform infrared spectroscopy study. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 110, 241-248.

Snyder, R.G., 1979. Vibrational correlation splitting and chain packing for the crystalline n‐ alkanes. The Journal of Chemical Physics 71, 3229-3235.

Staniszewska, E., Malek, K., Baranska, M., 2014. Rapid approach to analyze biochemical variation in rat organs by ATR FTIR spectroscopy. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 118, 981-986.

Stecyk, J.A.W., Bock, C., Overgaard, J., Wang, T., Farrell, A.P., Pörtner, H.-O., 2009. Correlation of cardiac performance with cellular energetic components in the oxygen-deprived turtle heart. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology 297, R756-R768.

Stecyk, J.A.W., Galli, G.L., Shiels, H.A., Farrell, A.P., 2008. Cardiac survival in anoxia-tolerant vertebrates: An electrophysiological perspective. Comparative Biochemistry and Physiology C-Toxicology & Pharmacology 148.

Stecyk, J.A.W., Paajanen, V., Farrell, A.P., Vornanen, M., 2007a. Effect of temperature and prolonged anoxia exposure on electrophysiological properties of the turtle (Trachemys scripta) heart. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology 293, R421-R437.

Stecyk, J.A.W., Paajanen, V., Farrell, A.P., Vornanen, M., 2007b. Effect of temperature and prolonged anoxia exposure on electrophysiological properties of the turtle (Trachemys scripta) heart. American Journal of Physiology-Regulatory Integrative and Comparative Physiology 293, R421-R437.

Stitt, D.M., Kastyak-Ibrahim, M.Z., Liao, C.R., Morrison, J., Albensi, B.C., Gough, K.M., 2012. Tissue acquisition and storage associated oxidation considerations for FTIR microspectroscopic imaging of polyunsaturated fatty acids. Vibrational Spectroscopy 60, 16- 22.

Stuart, B., 1997. Biological Applications of Infrared Spectroscopy. John Wiley & Sons, Chichester.

Toyran, N., Turan, B., Severcan, F., 2007. Selenium alters the lipid content and protein profile of rat heart: An FTIR microspectroscopic study. Archives of Biochemistry and Biophysics 458, 184-193.

Ultsch, G.R., 2006. The ecology of overwintering among turtles: where turtles overwinter and its consequences. Biological Reviews 81, 339-367.

Vornanen, M., Haverinen, J., Hassinen, M., Koskinen, H., Krasnov, A., 2005. Plasticity of cardiac function in thermally acclimated rainbow trout. Comparative Biochemistry and Physiology a-Molecular & Integrative Physiology 141.

Metabolic and biochemical remodelling of the turtle heart

Wang, Q., Sanad, W., Miller, L.M., Voigt, A., Klingel, K., Kandolf, R., Stangl, K., Baumann, G., 2005. Infrared imaging of compositional changes in inflammatory cardiomyopathy. Vibrational Spectroscopy 38, 217-222.

Whelan, D.R., Bambery, K.R., Heraud, P., Tobin, M.J., Diem, M., McNaughton, D., Wood, B.R., 2011. Monitoring the reversible B to A-like transition of DNA in eukaryotic cells using Fourier transform infrared spectroscopy. Nucleic Acids Research 39, 5439-5448.

Wood, B.R., Quinn, M.A., Tait, B., Ashdown, M., Hislop, T., Romeo, M., McNaughton, D., 1998. FTIR microspectroscopic study of cell types and potential confounding variables in screening for cervical malignancies. Biospectroscopy 4, 75-91.

Xin, M., Olson, E.N., Bassel-Duby, R., 2013. Mending broken hearts: cardiac development as a basis for adult heart regeneration and repair. Nat Rev Mol Cell Biol 14, 529-541.

Zheng, N., Yang, T., Liang, M., Zhang, H., Li, L., Sunnassee, A., Liu, L., 2010. Characterization of protein in old myocardial infarction by FTIR micro-spectroscopy. Journal of Huazhong University of Science and Technology [Medical Sciences] 30, 546-550.

Thermal remodelling of the ectothermic heart

10. GENERAL DISCUSSION

Ambient temperature is a ‘master regulator’ for ectotherm physiology. As such, the effects of long- and short-term temperature change on the ectotherm cardiovascular system have long been of interest to comparative physiologists and remain a key area of research to this day (Zwikster and Boyd, 1936; Stullken et al., 1949; Gatten, 1974; Wood et al., 1979; Harrison and Bers, 1990; Kalinin et al., 2009; Crossley et al., 2015; Lee et al., 2016). Before I began my PhD, many studies had shown the compensatory remodelling strategy of the cold-active rainbow trout and the active suppression of physiological function by the freshwater turtle, in response to prolonged cold temperatures (Farrell et al., 1988b; Bailey and Driedzic, 1993; Driedzic and Gesser, 1994; Farrell et al., 1994; Sephton and Driedzic, 1995; Driedzic et al., 1996; Taylor et al., 1996; Hicks and Wang, 1998; Saunders and Patel, 1998; Vornanen, 1998; Aho and Vornanen, 1999; Hicks and Farrell, 2000a, b; Aho and Vornanen, 2001; Vornanen et al., 2005; Haverinen and Vornanen, 2007; Stecyk et al., 2007a; Klaiman et al., 2011; Korajoki and Vornanen, 2012). However, work had predominately focused on understanding the thermal remodelling response of the active properties of the heart and, by comparison, very little was known about the passive properties. For this reason, my aim was to study the cardiovascular responses of thermal acclimation in these species, but with particular interest as to how the passive properties of the heart were affected. I will now discuss the significant findings of each chapter in relation to each other and with other studies in the field. I will start by discussing the fish heart, then move onto discussing the turtle heart before concluding with a general discussion on the cardiovascular response of both species.

10. 1. CARDIAC REMODELLING WITH THERMAL ACCLIMATION IN FISH

As discussed in chapter 1, ambient temperature change directly alters the physiology of the fish heart, primarily due to the direct effect of temperature on biochemical and metabolic processes. Rainbow trout are a cold-active species; therefore, the heart needs to compensate for reduced function at cold temperatures to permit continued swimming performance. Although the active properties of the heart had been well studied, the effects on the passive properties had been largely overlooked. When I began my PhD, collaborative research between our lab and Dr. Todd Gillis’ lab had recently investigated some of the active and passive properties of the fish heart, published in Klaiman et al. (2011). Not only did work from this study provide support for a profound remodelling of the passive properties of the heart following both cold and warm acclimation compared to controls, it suggested that some of these responses may be maladaptive rather than solely compensatory (Klaiman et al., 2011). Historically, cardiac remodelling in fish has generally been considered an adaptive response to increased physiological load (Farrell et al., 1988b; Kent et al., 1988; Graham and Farrell, 1989). Therefore, throughout my PhD studies I aimed to both describe the response to thermal

Page | 93

Thermal remodelling of the ectothermic heart acclimation and put the remodelling phenotype into the context of mammalian cardiac remodelling to suggest if this phenotype is likely adaptive or maladaptive in the rainbow trout.

Chapter 3 builds on the work by Klaiman et al. (2011), and some of the same techniques were used to validate the remodelling phenotype. In chapter 3, we focused on remodelling of the passive properties of the heart. Similar to Klaiman et al. (2011) we showed hypertrophy (cell growth) in the spongy myocardium by increases in myocyte bundle cross-sectional area and an increase in amorphous collagen with cold acclimation (Keen et al., 2016). We found the opposite following chronic warming, with an atrophy of the spongy myocardium by a decrease in the myocyte bundle cross-sectional area and a decrease in amorphous collagen (Figure 10. 1) (Klaiman et al., 2014). We built upon the study of Klaiman et al. (2011) by assessing fibrillar collagen content with picro-sirus red, which is the gold standard for histological collagen assessment, and by investigating mRNA expression of muscle specific growth factors, collagen genes, collagen regulating genes and markers of hypertrophic growth. Muscle specific growth factors (VHMC, SMLC2 and MLP) were strongly up-regulated with cold acclimation and down-regulated with warm acclimation (Figure 10.1), which agrees with a recent transcriptome study by Vornanen et al. (2005). Our results also suggest hyperplasia (cell proliferation) of myocytes, by increased mRNA expression of PCNA, and angiogenesis, by increased mRNA expression of VEGF. The relative contributions of hypertrophy to hyperplasia in cardiac growth have been debated (Farrell et al., 1988b; Vornanen et al., 2005; Sun et al., 2009; Klaiman et al., 2011; Castro et al., 2013), in chapter 3 we suggest that growth is primarily due to myocyte hypertrophy, but there is likely also a smaller contribution of myocyte hyperplasia (Figure 10.1) (Keen et al., 2016).

In chapter 4 we probed the cellular energetics that permit these changes in tissue morphology using FTIR imaging spectroscopy and histological tissue stains. The cardiac growth associated with cold-temperatures in fish is likely to be energetically expensive as it involves increased protein synthesis (Vornanen et al., 2005). Cardiac growth pathways are reviewed extensively in chapter 1. Briefly, in the normal adult heart fatty acid oxidation (FAO) is the primary metabolic pathway for cellular function. Our results in chapter 4 suggest that during chronic cold-induced hypertrophic growth, FAO remains the primary energetic pathway and is increased to meet the high cellular demand, which is consistent with a number of previous studies (Sephton and Driedzic, 1991; Bailey and Driedzic, 1993; Driedzic and Gesser, 1994; Sephton and Driedzic, 1995; Driedzic et al., 1996). However, following warm acclimation there appeared to be a shift in the primary metabolic pathway, with a decreased reliance on FAO and an increase in glycolytic pathways. This result is supported by recent studies in Atlantic cod (Gadus morhua) and temperate wrasse (Notolabrus celidotus) which show altered mitochondrial function due to an increased dependence on anaerobic energetic pathways (Rodnick et al., 2014; Iftikar et al., 2015). The switch in cellular energetics is likely due to decreased oxygen availability and high metabolic rate associated with acute high

Page | 94

Thermal remodelling of the ectothermic heart temperatures (Farrell, 1991; Farrell et al., 1996; Wang and Overgaard, 2007; Eliason and Farrell, 2016). Glycolysis provides higher ATP production per mole of oxygen used, so can increase short-term ATP supply (Allard et al., 1994; Neubauer, 2007; Ingwall, 2009). However, glycolysis is ~20 times slower than FAO and, therefore, the high metabolic rate of the fish heart at high temperatures may lead to rapid depletion of ATP, suggesting it is a maladaptive response (Neubauer, 2007; Portner and Knust, 2007; Wang and Overgaard, 2007; Iftikar et al., 2014).

In Chapter 3 we provide further support to the suggestion that cold-induced cardiac growth may have maladaptive effects on the rainbow trout heart. We found that mRNA expression of regulator of calcineurin (RCAN1), was increased with cold acclimation and decreased with warm acclimation, which suggests activation of the calcineurin-NFAT signalling cascade (Keen et al., 2016). This pathway is central to pathological hypertrophic growth in mammals, suggesting similarities in the cardiac growth phenotype in fish (Frey and Olson, 2003; Liu et al., 2009; Bernardo et al., 2010). Furthermore, we found an up-regulation of mRNA for genes involved in the foetal gene program (ANP and BNP) following cold acclimation and a down- regulation of mRNA of these genes following warm acclimation, which have been shown to be involved in dilated cardiomyopathy in zebrafish (Shih et al., 2015; Keen et al., 2016). In addition, there was a fibrosis of the myocardium following cold acclimation with an up- regulation of collagen I mRNA (COL1α3), with the opposite response following chronic warming (Keen et al., 2016). Fibrosis in mammals is associated with pathological hypertrophy and is usually absent in physiological growth (Chapman et al., 1990; Bernardo et al., 2010). However, COL1α3 is a fish specific collagen I gene which encodes a collagen 1 type 3 subunit (Saito et al., 2001) which, in fish, replaces one of the type 1 or type 2 subunits that make mammalian collagen I (Saito et al., 2001). Interestingly, the α3 chain reduces the denaturation temperature of the collagen I molecule and makes it more susceptible to degradation by MMP13 (Saito et al., 2001), which may explain the transient nature of cardiac fibrosis in fish following chronic temperature exposure.

As in mammals, fish myocardial collagen content is a balance of deposition and degradation (Pedersen et al., 2015). Collagen degradation is regulated by matrix metalloproteinase (MMPs). With cold-induced ventricular hypertrophy and fibrosis in rainbow trout, myocardial expression of MMP2 and MMP13 mRNA has been shown to be down-regulated (Figure 10.1) (Keen et al., 2016). In turn, the gelatinase activity of MMPs is regulated by tissue inhibitors of MMPs (TIMPs) (Pedersen et al., 2015). Increased enzymatic activity of TIMPs inhibits collagen degradation by MMPs and, therefore, TIMP activity is associated with increased collagen deposition. Rainbow trout with cold-induced ventricular fibrosis and decreased mRNA expression of MMPs have an associated up-regulation of TIMP2 in the ventricular myocardium (Figure 10.1) (Keen et al., 2016).

Page | 95

Thermal remodelling of the ectothermic heart

Figure 10.1. Thermal remodelling of the fish heart. An overview of the effects of chronic cooling (5 °C) and chronic warming (18 °C) on the rainbow trout heart.

Page | 96

Thermal remodelling of the ectothermic heart

Finally in chapter 3, we progressed to study the functional effects of tissue morphology and connective tissue remodelling on tissue stiffness. As discussed in chapter 1, changes in myocardial thickness and connective tissue content can alter whole chamber compliance, which we assessed by generating in vivo pressure-volume curves (Biernacka and Frangogiannis, 2011; Segura et al., 2012). We found an increase in whole chamber stiffness with cold acclimation and an increased whole chamber compliance with warm acclimation compared to controls, and these changes in chamber stiffness were also reflected in the micromechanical stiffness of tissue sections (Figure 10.1) (Keen et al., 2016). Unfortunately, we were unable to determine the contribution of tissue level remodelling to these changes in compliance. Our accumulative frequency of force curves showed an even distribution, suggesting that tissue stiffness was increasing evenly across the tissue rather than due to an increase in stiff elements, such as fibrillar collagen (Keen et al., 2016). It is, therefore, likely that there is an additional intracellular contribution to changes in compliance and passive tension of the fish ventricle.

In chapter 5 we turned our attention to remodelling of the fish atrium following thermal acclimation. Interestingly, we saw some features that were consistent with remodelling of the ventricle and some features that were opposite. In the rainbow trout atrium, cold acclimation increase stiffness of the chamber as a whole and the micromechanical stiffness of tissue sections, with the opposite remodelling response following chronic warm (Figure 10.1). Although we did not see a direct increase in the collagen content of atrial tissue following cold acclimation, and a reduction following warm acclimation, mRNA expression of the COL1α3 gene was up-regulated in the atrium following chronic cooling and reduced following chronic warming (Figure 10.1). Accordingly, the mRNA expression of collagen degrading MMPs (MMP2, MMP9 and MMP13) were depressed following cold acclimation and increased following warm acclimation, as was the gelatinase activity of MMPs (Figure 10.1).

However, we did not find any increase in whole chamber mass or myocyte bundle cross- sectional area following thermal acclamation. Instead we saw an increase in myocyte extra- bundular sinus space following cold acclimation, which was reduced following warm acclimation, compared to controls. In addition, we found that mRNA expression of cardiac muscle-specific growth genes (VMHC and SMLC2) were down-regulated following cold acclimation and up-regulated following warm acclimation (Figure 10.1). Together, we suggest that these changes are due to chronic dilation of the fish atrium with chronic cold. Furthermore, our qPCR data suggested that angiogenesis was reduced (by VEGF mRNA expression) with cold acclimation and increased with warm acclimation, as was activation of the calcineurin- NFAT signalling cascade (by RCAN1 mRNA expression) and the foetal gene program (by ANP mRNA expression).

Page | 97

Thermal remodelling of the ectothermic heart

The differences in remodelling of the atrium compared to the ventricle following chronic temperature change are likely due to differences in the stresses imposed on the myocardium and differential compensatory strategies required to maintain their chamber-specific role in cardiac function. The fish heart functions as a volume pump, modulating cardiac output (Q) primarily by changes in stroke volume (SV) (Farrell and Jones, 1992; Forster and Farrell, 1994). The trout heart has high sensitivity to cardiac load and the atrium acts as a volume reservoir, directly altering end-diastolic volume (EDV) and, therefore, SV via the Frank-Starling mechanism (Forster and Farrell, 1994; Shiels and White, 2008). During times of high haemodynamic stress of pumping large volumes of viscous blood, inflation of the atrium may be particularly high (Graham and Farrell, 1989; Aho and Vornanen, 1999). Increased inflation of the atrium, acting as a volume reservoir, may ‘buffer’ ventricular preload, which could be particularly important to maintain ventricular function as large increases in ventricular preload are metabolically costly to the heart (Hansen et al., 2002). Indeed, a similar observation was made by Aho and Vornanen (1999), who report atrial enlargement to be greater than ventricular enlargement following cold acclimation in rainbow trout due to the increased cardiac preload. Although we did not see an overall increase in mass, as these authors did, in mammals atrial enlargement has been shown to occur with and without a corresponding increase in wall thickness, with increased wall thickness considered the more ‘healthy’ or ‘physiological’ phenotype (Verheule et al., 2003; Pelliccia et al., 2005; Pellman et al., 2010).

In chapter 6 we focused on remodelling of the trout outflow tract (OFT) following thermal acclimation. The rainbow trout OFT is composed of a highly specialized structure with a bulbus arteriosus and bulbo-ventricular valves between the ventricle and ventral aorta (Icardo, 2006). Following cold acclimation we found an increase in the passive stiffness of the whole chamber compared to controls, with a decrease in passive stiffness following warm acclimation. These changes in passive stiffness caused an increase in compliance over a physiological pressure range following warm acclimation (Forster and Farrell, 1994; Seth et al., 2014). The changes in chamber stiffness were supported by changes in the connective tissue content of the bulbus arteriosus, with the collagen to elastin ratio increased following cold acclimation and decreased following warm acclimation. With warm acclimation there was also an increase in the abundance and gelatinase activity of collagen degrading MMPs (Kubota et al., 2003; Lødemel et al., 2004; Pedersen et al., 2015).

Changes in OFT stiffness with cold temperature likely have profound implications in thermal remodelling of the ventricle. Under normal circumstances the fish bulbus arteriosus reduces afterload pressure (Priede, 1976; Bushnell et al., 1992; Seth et al., 2014). However, increased stiffness of the OFT with low temperature means that intraluminal pressure is higher at any given volume and as the ventricle ejects blood directly into the OFT this will increase ventricular afterload, which may explain some of the features of the hypertrophic phenotype of the fish heart following thermal acclimation (Girerd et al., 1991; Coleman, 1993; Seth et al.,

Page | 98

Thermal remodelling of the ectothermic heart

2014). In mammals, hypertension increase ventricular afterload and is associated with concentric hypertrophy (i.e. lateral growth of myocytes) and ventricular fibrosis (Grossman et al., 1975; Weber et al., 1989; Nadal-Ginard et al., 2003; Dorn, 2007; Bernardo et al., 2010; Collier et al., 2012a). In the hypertrophic fish ventricle, in chapter 3, we report an increase in myocyte cross-sectional area and ventricular fibrosis, which are inconsistent with a ‘physiological’ response due to increased cardiac preload alone and suggest increased afterload pressure also plays a role (Grossman et al., 1975; Farrell et al., 1988b; Kent et al., 1988; Graham and Farrell, 1989; Clark and Rodnick, 1998, 1999; Bernardo et al., 2010).

Bringing these studies together, it appears that chronic cold-induced remodelling increases global stiffness of the fish heart, in most cases by remodelling of the connective tissue, particularly collagen, in response to increase haemodynamic strain (Graham and Farrell, 1989; Pelouch and Vornanen, 1996; Klaiman et al., 2011; Keen et al., 2016). The contribution of intracellular remodelling remains unclear and warrants further investigations as this is likely to provide further insight into the overall remodelling response. The ventricle increases in size to provide additional passive tension, contractile force and pumping capacity (Klaiman et al., 2011; Klaiman et al., 2014), with the fuel for this growth provided by increases in FAO (Bailey and Driedzic, 1993; Driedzic and Gesser, 1994; Driedzic et al., 1996). We did not assess the alterations in cellular energetics in the other chambers of the heart and we did not see a hypertrophic response. However, the chambers of the heart show distinct remodelling phenotypes, which likely reflect the role of that chamber in cardiac function. Overall, we suggest that remodelling of the fish ventricle serves both cardio-protection, from the haemodynamic strain of increased cardiac preload and afterload, as well as compensation for reduced contractile function and increased cardiac afterload.

10. 2. CARDIAC REMODELLING WITH THERMAL ACCLIMATION IN TURTLES

Ectothermic animals can adopt one of three strategies when faced with environmental temperature change. The can either submit to the direct (Q10; rate of change over 10 °C) effects on their physiological processes, offset the direct effects of temperature by a compensatory remodelling, or enhance the direct effects of temperature and enter a state of hibernation or torpor (Kalinin et al., 2009). Freshwater turtles take the third strategy, actively suppressing metabolic and physiological processes over winter and entering a state of periodic inactivity (Herbert and Jackson, 1985b; Hicks and Farrell, 2000a, b; Jackson, 2002; Stecyk et al., 2007a; Stecyk et al., 2008). When I began my PhD, very little was known about the passive properties of the turtle heart, with or without thermal acclimation. In addition, the studies that had investigated the effect of prolonged temperature change on the turtle heart had conducted experiments at the animal’s acclimation temperature, making it difficult to distinguish the direct effects of temperature from a long-term remodelling response.

Page | 99

Thermal remodelling of the ectothermic heart

For my first experiment of thermal remodelling of the turtle heart, in chapter 7, the aim was to assess in vivo cardiovascular function and sensitivity to cardiac load in thermally acclimated turtles, but at a common temperature so that the direct effects of temperature could be differentiated from a chronic remodelling response. This study was designed to build upon previous work by Farrell et al. (1994), who showed that sensitivity of the turtle heart to cardiac load increased with acute decreases in temperature, and Hicks and Farrell (2000a,b) who recorded in vivo function of the turtle heart following thermal acclimation, at the corresponding acclimation temperatures. The data in chapter 7 suggests that prolonged cold temperature alters cardiac shunting patterns by reducing systemic resistance and increasing systemic flow, which results in a net right to left (increased systemic) shunt flow (Figure 10. 2). Cold acclimation also increased the heart’s sensitivity to an in vivo volume load, administered as a bolus injection of saline into the jugular vein, (Figure 10. 2) which may be relevant during hibernation, when diving turtles have been suggested to increase SV by up to 5-fold (Burggren et al., 1997). To further probe this change in the sensitivity of the heart to cardiac preload, we generated ex vivo pressure volume curves, based on the methodology in chapter 3 (Keen et al., 2016). The pressure volume curves showed an increase in cardiac stiffness following cold acclimation (Figure 10. 2), so we investigated changes in connective tissue content of the ventricle and major outflow vessels using histological tissue stains. Semi-quantitative analysis of the stained tissue sections suggested an increase in collagen content of the ventricular myocardium and an increase in the elastin content of the major outflow arteries following cold acclimation (Figure 10. 2). These results suggest that the stiffness of the ventricle is increased by increases in connective tissue content similar to interstitial fibrosis with pathological hypertrophy in mammals (Grossman et al., 1975; Weber et al., 1989; Nadal-Ginard et al., 2003; Dorn, 2007; Bernardo et al., 2010; Collier et al., 2012a). However, there was no corresponding decrease in in vivo systolic function suggesting the phenotype is more similar to that observed in heart failure with preserved ejection fraction in mammals, where increased myocardial stiffness defends systolic function (Katz and Rolett, 2016). In addition, an increased elastin content of the major outflow arteries may correspond to an increased compliance of the outflow vessels with elastin providing greater stretch and compression in the arteries, increasing elastic recoil (Halper and Kjaer, 2014b). Together these two features of the remodelling response may explain the maintenance of SV at chronic low temperature with increased ventricular stiffness preserving systolic function and increased ventricular compliance reducing ventricular afterload pressure (Frenneaux and Williams, 2007; Katz and Rolett, 2016).

Page | 100

Thermal remodelling of the ectothermic heart

Figure 10. 2. Thermal remodelling of the turtle heart. An overview of the effects of chronic cooling (5 °C) on the freshwater heart and major outflow vessels.

Page | 101

Thermal remodelling of the ectothermic heart

With these measures of ventricular passive stiffness and connective tissue content, chapter 8 followed on directly from chapter 7. The aim of chapter 8 was to investigate the effects of thermal remodelling on the tissue micromechanics of the ventricle and how the changes in connective tissue seen in chapter 7 were regulated. AFM suggested that micromechanical stiffness of tissue cryosections was increased following cold acclimation (Figure 10. 2). In this study, the accumulative frequency distribution of force curves was particularly interesting. By histological staining, with picro-sirus red, we showed increases in ventricular fibrillar collagen following cold acclimation, in chapter 7. The nano-scale resolution of AFM means the contribution of smaller perimysial collagen to overall tissue stiffness is also measured using this technique (Graham et al., 2010). Here, we found a higher frequency of force curves with a reduced modulus between 1.0 and 1.5 MPa following cold acclimation, which is consistent with reduced modulus of tissue with high collagen fibre content indented at a low loading rate (Andriotis et al., 2014; Baldwin et al., 2014; McConnell et al., 2016). These results suggest that mechanical remodelling following temperature acclimation is not due to homogenous structural and/or compositional remodelling of the tissue, but rather, isolated or specific regions of the tissue becoming stiffer. Therefore, it appears that stiffness of the turtle ventricular myocardium following thermal acclimation is modulated, at least in part, by the relative content of fibrillar collagen in the tissue.

In addition to increased collagen content of the ventricular myocardium following cold acclimation, shown in chapter 7, there is an increase in collagen fibre alignment (Figure 10. 2), which we assessed quantitatively by coherency of fibres (Rezakhaniha et al., 2012). McConnell et al. (2016) have recently suggested that organization of collagen fibres can effect overall tissue stiffness, so it is likely that this plays a role in the increased passive tension of the ventricle with prolonged cold temperatures. In chapter 8, we also investigated changes in the regulation of connective tissue following thermal acclimation across multiple layers of organisation. Despite no change in mRNA expression of specific MMP genes, we found a decrease in the endogenous gelatinase activity of MMPs following cold acclimation (Figure 10. 2), suggesting there was less degradation of collagen fibrils (Bruggink et al., 2007; Akhtar et al., 2014). In addition, we found an increase in the mRNA expression of TIMP2 following cold acclamation (Figure 10. 2), which suggests an increased inhibition of gelatinase activity (Löffek et al., 2011). Overall chapter 8 suggests that increased passive stiffness of the turtle ventricle following cold acclimation involves a complex array of changes, involving collagen content, collagen fibril arrangement and connective tissue regulation. In addition, contributions of cellular remodelling, the actin-cytoskeleton and cross-linking of glycation end-products are also likely.

In chapter 9, we examined the tissue biochemistry of the turtle ventricle. We did not see an increase in ventricular muscle mass with cold acclimation, as we did with the fish heart; however, the rationale for this study was the well-documented metabolic shift in the metabolic

Page | 102

Thermal remodelling of the ectothermic heart state of the turtle heart during cold anoxia (Beall and Priviter, 1973; Herbert and Jackson, 1985c; Jackson, 2002; Stecyk et al., 2008). Our animals were maintained in normoxia throughout the acclimation and, therefore, did not experience anoxia. However, the trigger for winter hibernation is reduction in ambient temperature so we wanted to investigate if cold alone was capable of altering cellular energetics (Stecyk et al., 2007b; Stecyk et al., 2008). We used FTIR imaging spectroscopy and histological tissue stains to probe changes in tissue biochemistry of the turtle ventricle following thermal acclimation, with particular interest to tissue metabolites. Our results showed a decrease in tissue lipid content following cold acclimation, which suggests a decreased reliance on FAO, and an increase in tissue glycogen and lactate (Figure 10. 2), which suggests an increased utilisation of glycolytic pathways (Beall and Priviter, 1973; Saravanakumar et al., 2011). Therefore, overall our results for chapter 9 suggest that cold temperature is the primary regulator of the turtle hibernation or brumation, triggering a remodelling of myocardial energetics in advance of winter hypoxia or anoxia (Stecyk et al., 2007b; Stecyk et al., 2008). As this response occurs seasonally, the energetic remodelling of the turtle heart must be plastic and switch back to FAO upon ambient warming in spring (Jackson, 2002; Stecyk et al., 2008). Therefore, it is unlikely to be a maladaptive response as it is in when experienced in pathological remodelling of the human heart (Allard et al., 1994; Neubauer, 2007; Ingwall, 2009).

The combined results of chapters 7, 8 and 9 show that prolonged temperature change has profound effects on the turtle heart and cardiovascular system. Our data suggests that chronic cold alters the structural properties of the vasculature by decreasing the collagen to elastin ratio. The increased elastin likely increases the compliance of the major outflow vessels, which may partly account for the in vivo decreases in systemic pressure and systemic resistance (Halper and Kjaer, 2014b). The decreased systemic pressure and resistance likely compensate for the slow blood flow with acute cold temperatures due to high blood viscosity as well as decreasing afterload pressure to defend SV and cardiac function (Saunders and Patel, 1998; Frenneaux and Williams, 2007). In addition, acute cold increases sensitivity to cardiac load (Farrell et al., 1994) and chronic cold increases ventricular tissue stiffness, which together help to defend SV and cardiac function by shifting the Frank-Starling relationship (Katz and Rolett, 2016). The increases in ventricular tissue stiffness are due, at least in part, to increases in collagen content and collagen alignment in the tissue. The increase in collagen is permitted by decreased MMP activity due to an up-regulation of TIMPs (Löffek et al., 2011). In addition, the cellular energetics of the ventricle switch from predominantly FAO to glycolytic pathways, which is likely a cardio-protective mechanism due to the reduced oxygen availability and prepares the heart for hypoxic or anoxic winter hibernation conditions (Stecyk et al., 2008).

Page | 103

Thermal remodelling of the ectothermic heart

10. 3. PERSPECTIVES

The studies conducted in my PhD have added to the field of comparative physiology by providing further information on the cardiac remodelling response of fish and freshwater turtles, associated with chronic temperature change. During my studies I have focused on adding to the knowledge of the passive properties of the heart. I have also attempted to relate my findings back to mammalian physiology and propose instances where the rainbow trout or freshwater turtle may present an appropriate or interesting biomedical model organism.

The turtle does not show a hypertrophic response to temperature, likely due to its cold-dormant nature reducing cardiac output (Overgaard et al., 2005). However, the cold-active rainbow trout shows a well-documented hypertrophic response to cold (Farrell et al., 1988a; Bailey and Driedzic, 1990; Driedzic et al., 1996; Vornanen et al., 2005; Klaiman et al., 2011). Although generally considered an adaptive response triggered by increased physiological requirement of pumping cold viscous blood, recent studies, and the data presented in this thesis, suggest there are maladaptive components that correlate to mammalian ‘pathological’ remodelling (Kent and Prosser, 1985; Farrell et al., 1988a; Graham and Farrell, 1989; Keen et al., 2016). Clearly, the response of the fish heart is not ‘pathological’ as it is a physiological trigger, which comes and goes seasonally. However, some aspects of the remodelling response that are analogous with ‘pathological’ remodelling may provide a novel system to gain insight inot human cardiomyopathies. The transient nature may prove particularly valuable by providing insight into progression, regulation and regression of disease.

In response to prolonged cold temperature both rainbow trout and freshwater turtles show an increase in ventricular passive tension with an associated fibrosis of the myocardium. In both cases these responses are likely cardio-protective and provide increased structural support to the myocardium for pumping cold viscous blood (Graham and Farrell, 1989; Saunders and Patel, 1998). It is also possible that increased passive stiffness helps defend cardiac function by increasing sensitivity to cardiac preload by the Frank-Starling mechanism. In mammals, increased myocardial stiffness and fibrosis are commonly associated with a maladaptive and irreversible ‘pathological’ remodelling response to increased afterload (Weber, 1989; Weber et al., 1989; Collier et al., 2012b; Daskalopoulos et al., 2016). In both the fish and the turtle this response occurs seasonally and, therefore, must be more plastic than in mammals. In fish it is likely a combination of preload and afterload that causes the remodelling response (Clark and Rodnick, 1999). The transient nature is most likely provided by the fish specific COL1α3 subunit (Saito et al., 2001). In the turtle, it is unclear how the fibrotic response occurs and regresses transiently, but it may be due to high activity of regulatory enzymes. In either case, both of these animals may provide insight into the progression and reversibility of mammalian interstitial fibrosis and, therefore, have the potential to provide important new biomedical models.

Page | 104

Thermal remodelling of the ectothermic heart

Finally, there is an interesting comparison in the different strategies for remodelling of cellular energetics between these two species. In turtles, low oxygen availability at low temperatures triggers a switch in cellular energetics (Beall and Priviter, 1973; Jackson, 2002; Stecyk et al., 2008). Interestingly, the turtle heart remodels cellular energetics from a steady-state of FAO during warm conditions to a steady-state of glycolysis during cold conditions (Stecyk et al., 2008; Stecyk et al., 2009). This ‘steady-state’ response is different to energetic remodelling of the mammalian myocardium, which is compromised by long-term reliance on glycolytic pathways and leads to ATP depletion and heart failure (Neubauer, 2007). Similarly, it is likely that the metabolic shift of the fish heart at warm temperatures is also maladaptive, leading to ATP depletion and, in extreme cases, heart failure (Iftikar et al., 2014). The turtle’s ability to remodel to a steady state glycolytic metabolism is likely due to the low oxygen demand of the cardiac tissue during winter conditions (Stecyk et al., 2008; Galli and Richards, 2012). Nevertheless, this transient and adaptive metabolic remodelling may provide important insight into the progression, regression and reversibility of metabolic cardiac remodelling across species.

Page | 105

Thermal remodelling of the ectothermic heart

Page | 106

Thermal remodelling of the ectothermic heart

11. REFERENCES

Abel, E.D., and T. Doenst. 2011. Mitochondrial adaptations to physiological vs. pathological cardiac hypertrophy. Cardiovascular Research. 90:234-242.

Agnisola, C., and B. Tota. 1994. Structure and function of the fish cardiac ventricle: flexibility and limitations. Cardioscience. 5.

Aho, E., and M. Vornanen. 1998. Ca2+-ATPase activity and Ca2+ uptake by sarcoplasmic reticulum in fish heart: Effects of thermal acclimation. Journal of Experimental Biology. 201:525-532.

Aho, E., and M. Vornanen. 1999. Contractile properties of atrial and ventricular myocardium of the heart of rainbow trout Oncorhynchus mykiss: Effects of thermal acclimation. Journal of Experimental Biology. 202:2663-2677.

Aho, E., and M. Vornanen. 2001. Cold acclimation increases basal heart rate but decreases its thermal tolerance in rainbow trout (Oncorhynchus mykiss). Journal of Comparative Physiology B-Biochemical Systemic and Environmental Physiology. 171.

Akhtar, R., J.K. Cruickshank, X. Zhao, L.A. Walton, N.J. Gardiner, S.D. Barrett, H.K. Graham, B. Derby, and M.J. Sherratt. 2014. Localized micro- and nano-scale remodelling in the diabetic aorta. Acta Biomaterialia. 10:4843-4851.

Akki, A., K. Smith, and A.-M.L. Seymour. 2008. Compensated cardiac hypertrophy is characterised by a decline in palmitate oxidation. Molecular and Cellular Biochemistry. 311.

Alderman, S.L., J.M. Klaiman, C.A. Deck, and T.E. Gillis. 2012. Effect of cold acclimation on troponin I isoform expression in striated muscle of rainbow trout. American Journal of Physiology-Regulatory Integrative and Comparative Physiology. 303.

Allard, M.F., B.O. Schonekess, S.L. Henning, D.R. English, and G.D. Lopaschuk. 1994. Contribution of oxidative metabolism and glycolysis to ATP production in hypertrophied hearts. American Journal of Physiology. 267.

Allen, D.G., and J.C. Kentish. 1985. The cellular basis of the length tension relation in cardiac- muscle. Journal of Molecular and Cellular Cardiology. 17:821-840.

Andriotis, O.G., W. Manuyakorn, J. Zekonyte, O.L. Katsamenis, S. Fabri, P.H. Howarth, D.E. Davies, and P.J. Thurner. 2014. Nanomechanical assessment of human and murine collagen fibrils via atomic force microscopy cantilever-based nanoindentation. Journal of the Mechanical Behavior of Biomedical Materials. 39:9-26.

Antonio, R.C., C.S. Ceron, E. Rizzi, E.B. Coelho, J.E. Tanus-Santos, and R.F. Gerlach. 2013. Antioxidant effect of doxycycline decreases MMP activity and blood pressure in SHR. Molecular and Cellular Biochemistry. 386:99-105.

Arai, M., H. Matsui, and M. Periasamy. 1994. Sarcoplasmic reticulum gene expression in cardiac hypertrophy and heart failure. Circulation Research. 74:555-564.

Ask, J.A., G. Stenelarsen, and K.B. Helle. 1981. Temperature effects on the beta-2- adrenoceptors of the trout atrium. Journal of Comparative Physiology. 143:161-168.

Asnes, C.F., J.P. Marquez, E.L. Elson, and T. Wakatsuki. 2006. Reconstitution of the Frank- Starling mechanism in engineered heart tissues. Biophysical Journal. 91.

Page | 107

Thermal remodelling of the ectothermic heart

Badenhorst, D., M. Maseko, O.J. Tsotetsi, A. Naidoo, R. Brooksbank, G.R. Norton, and A.J. Woodiwiss. 2003. Cross-linking influences the impact of quantitative changes in myocardial collagen on cardiac stiffness and remodelling in hypertension in rats. Cardiovascular Research. 57.

Bailey, J.R., and W.R. Driedzic. 1990. Enhanced maximum frequency and force development of fish hearts following temperature-acclimation. Journal of Experimental Biology. 149.

Bailey, J.R., and W.R. Driedzic. 1993. Influence of low temperature acclimation on fate of metabolic fuels in rainbow trout (Oncorhynchus mykiss) hearts. Canadian Journal of Zoology. 71:2167-2173.

Baker, K., K.S. Warren, G. Yellen, and M.C. Fishman. 1997. Defective “pacemaker” current (I(h)) in a zebrafish mutant with a slow heart rate. Proceedings of the National Academy of Sciences of the United States of America. 94:4554-4559.

Baldwin, Samuel J., Andrew S. Quigley, C. Clegg, and L. Kreplak. 2014. Nanomechanical Mapping of Hydrated Rat Tail Tendon Collagen I Fibrils. Biophysical Journal. 107:1794-1801.

Ballesta, S., L.M. Hanson, and A.P. Farrell. 2012. The effect of adrenaline on the temperature dependency of cardiac action potentials in pink salmon Oncorhynchus gorbuscha. Journal of Fish Biology. 80:876-885.

Beall, R.J., and C.A. Priviter. 1973. Effects of cold exposure on cardiac metabolism of the turtle, Pseudemys (Chrysemys) picta. American Journal of Physiology. 224:435-441.

Benjamin, M., D. Norman, R.M. Santer, and D. Scarborough. 1983. Histological, histochemical and ultrastructural studies on the bulbus arteriosus of the sticklebacks, Gasterosteus aculeatus and Pungitius pungitius (Pisces: Teleostei). Journal of Zoology. 200:325- 346.

Bernardo, B.C., K.L. Weeks, L. Pretorius, and J.R. McMullen. 2010. Molecular distinction between physiological and pathological cardiac hypertrophy: Experimental findings and therapeutic strategies. Pharmacology & Therapeutics. 128:191-227.

Bers, D.M. 1991. Calcium regulation in cardiac muscle. Medicine and Science in Sports and Exercise. 23.

Bers, D.M. 2002a. Calcium and cardiac rhythms: Physiological and pathophysiological. Circulation Research. 90.

Bers, D.M. 2002b. Cardiac excitation-contraction coupling. Nature. 415.

Bers, D.M., and S. Despa. 2006. Cardiac myocytes Ca2+ and Na+ regulation in normal and failing hearts. Journal of Pharmacological Sciences. 100.

Bethea, N.J. 1972. Effects of temperature on heart rate and rates of cooling and warming in (Terrapene ornata). Comparative Biochemistry and Physiology. 41.

Bettex, D.A., R. Prêtre, and P.-G. Chassot. 2014. Is our heart a well-designed pump? The heart along animal evolution. European Heart Journal. 35:2322-2332.

Biernacka, A., and N.G. Frangogiannis. 2011. Aging and Cardiac Fibrosis. Aging and Disease. 2:158-173.

Bing, O.H.L., Matsushi.S, B.L. Fanburg, and H.J. Levine. 1971. Mechanical properties of rat cardiac muscle during experimental hypertrophy. Circulation Research. 28:234-245.

Page | 108

Thermal remodelling of the ectothermic heart

Bishop, S.P. 1990. The myocardial cell: normal growth, cardiac hypertrophy and response to injury. Toxicologic Pathology. 18:438-453.

Bottinelli, R., M. Canepari, V. Cappelli, and C. Reggiani. 1995. Maximum speed of shortening and ATPase activity in atrial and ventricular myocardia of hyperthyroid rats. American Journal of Physiology. 269:C785-790.

Bouvagnet, P., J. Leger, F. Pons, Dechesne, and J.J. Leger. 1984. Fiber Types and Myosin Types in Human Atrial and Ventricular Myocardium. Circulation Research. 55:794- 804.

Bowler, K., and R. Tirri. 1990. Temperature dependence of the heart isolated from the cold or warm acclimated perch (Perca fluviatilis). Comparative Biochemistry and Physiology Part A: Physiology. 96:177-180.

Braun, M.H., R.W. Brill, J.M. Gosline, and D.R. Jones. 2003. Form and function of the bulbus arteriosus in yellowfin tuna (Thunnus albacares), bigeye tuna (Thunnus obesus) and blue marlin (Makaira nigricans): static properties. Journal of Experimental Biology. 206:3311-3326.

Brodde, O.-E., H. Bruck, K. Leineweber, and T. Seyfarth. 2001. Presence, distribution and physiological function of adrenergic and muscarinic receptor subtypes in the human heart. Basic Research in Cardiology. 96:528-538.

Bruggink, A.H., M.F.M. van Oosterhout, N. de Jonge, J.P.M. Cleutjens, D.F. van Wichen, J. van Kuik, M.G.J. Tilanus, F.H.J. Gmelig-Meyling, J.G. van den Tweel, and R.A. de Weger. 2007. Type IV collagen degradation in the myocardial basement membrane after unloading of the failing heart by a left ventricular assist device. Laboratory Investigation. 87:1125-1137.

Burggren, W., A.P. Farrell, and H. Lillywhite. 1997. Vertebrate cardiovascular systems. In The Handbook of Physiology. W.H. Dantzler, editor. Oxford: Oxford University Press, New York. 215-308.

Burggren, W.W., and A.W. Pinder. 1991. Ontogeny of Cardiovascular and Respiratory Physiology in Lower Vertebrates. Annual Review of Physiology. 53:107-135.

Bushnell, P.G., D.R. Jones, and A.P. Farrell. 1992. The arterial system. Academic Press, Inc., New York, NY, USA. 89-139 pp.

Calderone, A., N. Takahashi, N.J. Izzo Jr, C.M. Thaik, and W.S. Colucci. 1995. Pressure- and Volume-Induced Left Ventricular Hypertrophies Are Associated With Distinct Myocyte Phenotypes and Differential Induction of Peptide Growth Factor mRNAs. Circulation. 92:2385-2390.

Castro, V., B. Grisdale-Helland, S.J. Helland, J. Torgersen, T. Kristensen, G. Claireaux, A.P. Farrell, and H. Takle. 2013. Cardiac molecular acclimation mechanisms in response to swimming induced exercise in Atlantic salmon. PLoS ONE. 8:e55056.

Cazorla, O., A. Freiburg, M. Helmes, T. Centner, M. McNabb, Y. Wu, K. Trombitas, S. Labeit, and H. Granzier. 2000. Differential expression of cardiac titin isoforms and modulation of cellular stiffness. Circulation Research. 86.

Chapman, D., K.T. Weber, and M. Eghbali. 1990. Regulation of fibrillar collagen type-I and type-III and basement-membrane type-IV collagen gene-expression in pressure overloaded rat myocardium. Circulation Research. 67:787-794.

Chen, J.A., S.T. Chu, and E. Amaya. 2007. Maintenance of motor neuron progenitors in Xenopus requires a novel localized cyclin. Embo Reports. 8:287-292.

Page | 109

Thermal remodelling of the ectothermic heart

Chien, S., R.G. King, R. Skalak, S. Usami, and A.L. Copley. 1975. Viscoelastic properties of human blood and red cell suspensions. Biorheology. 12:341-346.

Churcott, C.S., C.D. Moyes, B.H. Bressler, K.M. Baldwin, and G.F. Tibbits. 1994. Temperature and pH effects on Ca2+ sensitivity of cardiac myofibrils - a comparison of trout with mammals. American Journal of Physiology. 267:R62-R70.

Clark, G. 1974. Comparison of various oxidants for alum hematoxylin. Stain Technology. 49:225-227.

Clark, R.J., and K.J. Rodnick. 1998. Morphometric and biochemical characteristics of ventricular hypertrophy in male rainbow trout (Oncorhynchus mykiss). Journal of Experimental Biology. 201:1541-1552.

Clark, R.J., and K.J. Rodnick. 1999. Pressure and volume overloads are associated with ventricular hypertrophy in male rainbow trout. American Journal of Physiology- Regulatory Integrative and Comparative Physiology. 277:R938-R946.

Clarke, J., and S. Nicol. 1993. Blood viscosity of the little penguin, Eudyptula-minor and the adelie penguin, Pygoscelis adeliae: effects of temperature and shear rate. Physiological Zoology. 66:720-731.

Cleutjens, J.P.M., and E. Creemers. 2002. Integration of concepts: Cardiac extracellular matrix remodeling after myocardial infarction. Journal of Cardiac Failure. 8:S344-S348.

Coleman, F.C. 1993. Morphological and physiological consequences of parasites encysted in the bulbus-arteriosus of an estuarine fish, the sheepshead minnow, Cyprinodon variegatus. Journal of Parasitology. 79:247-254.

Collier, P., M. Ledwidge, and K. McDonald. 2012a. Diagnostics and therapeutic interventions in myocardial interstitial disease, a previously neglected pathology. Qjm-an International Journal of Medicine. 105:721-724.

Collier, P., C.J. Watson, M.H. van Es, D. Phelan, C. McGorrian, M. Tolan, M.T. Ledwidge, K.M. McDonald, and J.A. Baugh. 2012b. Getting to the heart of cardiac remodeling; how collagen subtypes may contribute to phenotype. Journal of Molecular and Cellular Cardiology. 52:148-153.

Comeau, S.G., and J.W. Hicks. 1994. Regulation of central vascular blood flow in the turtle. American Journal of Physiology. 267:R569-R578.

Costa, I.A.S.F., W.R. Driedzic, and A.K. Gamperl. 2013. Metabolic and Cardiac Responses of Cunner Tautogolabrus adspersus to Seasonal and Acute Changes in Temperature. Physiological and Biochemical Zoology: Ecological and Evolutionary Approaches. 86:233-244.

Creemers, E.E.J.M., J.P.M. Cleutjens, J.F.M. Smits, and M.J.A.P. Daemen. 2001. Matrix Metalloproteinase Inhibition After Myocardial Infarction: A New Approach to Prevent Heart Failure? Circulation Research. 89:201-210.

Crossley, D., J. Altimiras, and T. Wang. 1998. Hypoxia elicits an increase in pulmonary vasculature resistance in anaesthetised turtles (Trachemys scripta). Journal of Experimental Biology. 201:3367-3375.

Crossley, D.A., II, O.H. Wearing, B. Platzack, L.K. Hartzler, and J.W. Hicks. 2015. Acute and chronic temperature effects on cardiovascular regulation in the red-eared slider (Trachemys scripta). Journal of Comparative Physiology B-Biochemical Systemic and Environmental Physiology. 185:401-411.

Page | 110

Thermal remodelling of the ectothermic heart

Crossley, D.A., T. Wang, and J. Altimiras. 2000. Role of nitric oxide in the systemic and pulmonary circulation of anesthetized turtles (Trachemys scripta). Journal of Experimental Zoology. 286:683-689.

Daskalopoulos, E.P., C. Dufeys, L. Bertrand, C. Beauloye, and S. Horman. 2016. AMPK in cardiac fibrosis and repair: Actions beyond metabolic regulation. Journal of Molecular and Cellular Cardiology. 91:188-200. de Bold, A.J., and M.L.K. de Bold. 2005. Determinants of natriuretic peptide production by the heart: Basic and clinical implications. Journal of Investigative Medicine. 53:371-377.

Dimond, M.T. 1959. Responses to phenethylamines and nicotine and histology of turtle atria. American Journal of Physiology. 197:747-751.

Doenst, T., T.D. Nguyen, and E.D. Abel. 2013. Cardiac Metabolism in Heart Failure - Implications beyond ATP production. Circulation research. 113:709-724.

Dorn, G.W., II. 2007. The fuzzy logic of physiological cardiac hypertrophy. Hypertension. 49:962-970.

Dorn, G.W., J. Robbins, and P.H. Sugden. 2003. Phenotyping hypertrophy - Eschew obfuscation. Circulation Research. 92.

Driedzic, W.R., J.R. Bailey, and D.H. Sephton. 1996. Cardiac adaptations to low temperature non-polar teleost fish. Journal of Experimental Zoology. 275.

Driedzic, W.R., and H. Gesser. 1985. Ca2+ protection from the negative inotropic effect of contraction frequency on teleost hearts. Journal of Comparative Physiology B- Biochemical Systemic and Environmental Physiology. 156:135-142.

Driedzic, W.R., and H. Gesser. 1994. Energy Metabolism and Contractility in Ectothermic Vertebrate Hearts: Hypoxia, Acidosis, and Low Temperature. Physiological Reviews. 74:221-258.

Duran, A.C., B. Fernandez, A.C. Grimes, C. Rodriguez, J.M. Arque, and V. Sans-Coma. 2008. Chondrichthyans have a bulbus arteriosus at the arterial pole of the heart: morphological and evolutionary implications. Journal of Anatomy. 213:597-606.

Ebashi, S. 1963. Third Component Participating in the Super precipitation of `Natural Actomyosin'. Nature. 200:1010-1010.

Ebashi, S., F. Ebashi, and A. Kodama. 1967. Troponin as the Ca++-receptive Protein in the Contractile System. Journal of Biochemistry. 62:137-138.

Ebashi, S., F. Ebashi, and K. Maruyama. 1964. A New Protein Factor promoting Contraction of Actomyosin. Nature. 203:645-646.

Eghbali, M., and K.T. Weber. 1990. Collagen and the myocardium: fibrillar structure, biosynthesis and degradation in relation to hypertrophy and its regression. Molecular and Cellular Biochemistry. 96.

Eliason, E.J., and A.P. Farrell. 2016. Oxygen uptake in Pacific salmon Oncorhynchus spp.: when ecology and physiology meet. Journal of Fish Biology. 88:359-388.

Ellison, G.M., C.D. Waring, C. Vicinanza, and D. Torella. 2012. Physiological cardiac remodelling in response to endurance exercise training: cellular and molecular mechanisms. Heart. 98:5-10.

Farrell, A.P. 1979. The wind-kessel effect of the bulbus arteriosus in trout. Journal of Experimental Zoology. 209. Page | 111

Thermal remodelling of the ectothermic heart

Farrell, A.P. 1991. From hagfish to tuna - A perspective on cardiac-function in fish. Physiological Zoology. 64:1137-1164.

Farrell, A.P., C.E. Franklin, P.G. Arthur, H. Thorarensen, and K.L. Cousins. 1994. Mechanical performance of an in situ perfused heart from the turtle Chrysemys scripta during normoxia and anoxia at 5 oC and 15 oC. Journal of Experimental Biology. 191:207- 229.

Farrell, A.P., A.K. Gamperl, J.M.T. Hicks, H.A. Shiels, and K.E. Jain. 1996. Maximum cardiac performance of rainbow trout (Oncorhynchus mykiss) at temperatures approaching their upper lethal limit. Journal of Experimental Biology. 199:663-672.

Farrell, A.P., A.M. Hammons, M.S. Graham, and G.F. Tibbits. 1988a. Cardiac growth in rainbow-trout, salmo-gairdneri. Canadian Journal of Zoology-Revue Canadienne De Zoologie. 66:2368-2373.

Farrell, A.P., A.M. Hammons, M.S. Graham, and G.F. Tibbits. 1988b. Cardiac growth in rainbow trout, Salmo gairdneri. Canadian Journal of Zoology-Revue Canadienne De Zoologie. 66:2368-2373.

Farrell, A.P., J.A. Johansen, and M.S. Graham. 1988c. The role of the pericardium in cardiac performance of the trout (Salmo gairdneri). Physiological Zoology. 61:213-221.

Farrell, A.P., and D.R. Jones. 1992. The Heart. In Fish Physiology, Volume XII, Part A The Cardiovascular System.

Fitzsimons, D.P., and R.L. Moss. 1998. Strong Binding of Myosin Modulates Length- Dependent Ca2+ Activation of Rat Ventricular Myocytes. Circulation Research. 83:602-607.

Fomovsky, G.M., S. Thomopoulos, and J.W. Holmes. 2010. Contribution of extracellular matrix to the mechanical properties of the heart. Journal of Molecular and Cellular Cardiology. 48:490-496.

Forster, M.E., and A.P. Farrell. 1994. The volumes of the chambers of the trout heart. Comparative Biochemistry and Physiology a-Physiology. 109:127-132.

Frank, O. 1985. Zur Dynamik des Herzmuskels. Z Biol. 32:370-447.

Franklin, C.E., and P.S. Davie. 1992. Sexual maturity can double heart mass and cardiac power output in male rainbow trout. Journal of Experimental Biology. 171:139-148.

Frenneaux, M., and L. Williams. 2007. Ventricular-arterial and ventricular-ventricular interactions and their relevance to diastolic filling. Progress in Cardiovascular Diseases. 49:252-262.

Frey, N., and E.N. Olson. 2003. Cardiac hypertrophy: The good, the bad and the ugly. Annual Review of Physiology. 65:45-79.

Fukuda, N., and H.L. Granzier. 2006. Titin/connectin-based modulation of the Frank-Starling mechanism of the heart. Journal of Muscle Research & Cell Motility. 26:319-323.

Galis, Z.S., and J.J. Khatri. 2002. Matrix Metalloproteinases in Vascular Remodeling and Atherogenesis: The Good, the Bad, and the Ugly. Circulation Research. 90:251-262.

Galli, G., E.W. Taylor, and T. Wang. 2004. The cardiovascular responses of the freshwater turtle, Trachemys scripta, to warming and cooling. Journal of Experimental Biology. 207:1471-1478.

Page | 112

Thermal remodelling of the ectothermic heart

Galli, G.L.J., and J.G. Richards. 2012. The effect of temperature on mitochondrial respiration in permeabilized cardiac fibres from the freshwater turtle, Trachemys scripta. Journal of Thermal Biology. 37:195-200.

Galli, G.L.J., H.A. Shiels, and R.W. Brill. 2009. Temperature Sensitivity of Cardiac Function in Pelagic Fishes with Different Vertical Mobilities: Yellowfin Tuna (Thunnus albacares), Bigeye Tuna (Thunnus obesus), Mahimahi (Coryphaena hippurus), and Swordfish (Xiphias gladius). Physiological and Biochemical Zoology. 82:280-290.

Gamperl, A.K., and A.P. Farrell. 2004. Cardiac plasticity in fishes: environmental influences and intraspecific differences. Journal of Experimental Biology. 207:2539-2550.

Gatten, R.E. 1974. Effects of temperature and activity on aerobic and anaerobic metabolism and heart rate in turtles Pseudemys scripta and Terrapene ornata. Comparative Biochemistry and Physiology - Part A: Molecular & Integrative Physiology. 48:619- 648.

Gautier, H.O.B., A.J. Thompson, S. Achouri, D.E. Koser, K. Holtzmann, E. Moeendarbary, and K. Franze. 2015. Chapter 12 - Atomic force microscopy-based force measurements on animal cells and tissues. In Methods in Cell Biology. K.P. Ewa, editor. Academic Press. 211-235.

Genge, C.E., W.S. Davidson, and G.F. Tibbits. 2013. Adult teleost heart expresses two distinct troponin C paralogs: cardiac TnC and a novel and teleost-specific ssTnC in a chamber- and temperature-dependent manner. Physiological Genomics. 45:866-875.

Gerdes, A.M. 2002. Cardiac myocyte remodeling in hypertrophy and progression to failure. Journal of Cardiac Failure. 8:S264-S268.

Gerdes, A.M., and X. Wang. 2003. Structural Remodeling of Cardiac Myocytes in Hypertrophy and Progression to Failure. In Cardiac Remodeling and Failure. P.K. Singal, I.M.C. Dixon, L.A. Kirshenbaum and N.S. Dhalla, editors. Springer US, Boston, MA. 183-193.

Gillis, T.E., and J.M. Klaiman. 2011. The influence of PKA treatment on the Ca2+ activation of force generation by trout cardiac muscle. Journal of Experimental Biology. 214:1989- 1996.

Gillis, T.E., C.R. Marshall, X.H. Xue, T.J. Borgford, and G.F. Tibbits. 2000. Ca2+ binding to cardiac troponin C: effects of temperature and pH on mammalian and salmonid isoforms. American Journal of Physiology-Regulatory Integrative and Comparative Physiology. 279:R1707-R1715.

Gilson, A.S. 1932. Vagal depression of the turtle atrium a study of the integration of effect of nerve impulses American Journal of Physiology. 100:459-473.

Gilson, A.S. 1939. The increased accommodation to electric currents produced by vagal inhibition of the turtle atrium American Journal of Physiology. 127:333-337.

Girerd, X., S. Laurent, B. Pannier, R. Asmar, and M. Safar. 1991. Arterial distensibility and left ventricular hypertrophy in patients with sustained essential hypertension. American Heart Journal. 122:1210-1214.

Gollock, M.J., S. Currie, L.H. Petersen, and A.K. Gamperl. 2006. Cardiovascular and haematological responses of Atlantic cod (Gadus morhua) to acute temperature increase. Journal of Experimental Biology. 209:2961-2970.

Goolish, E.M. 1987. Cold-acclimation increases the ventricle size of carp, cyprinus-carpio. Journal of Thermal Biology. 12:203-205.

Page | 113

Thermal remodelling of the ectothermic heart

Gordon, A.M., A.F. Huxley, and F.J. Julian. 1966. The variation in isometric tension with sarcomere length in vertebrate muscle fibres. The Journal of Physiology. 184:170- 192.

Gordon, A.M., and E.B. Ridgeway. 1993. Cross-bridges affect both TnC structure and calcium affinity in muscle fibers. Plenium Press Div Plenum Publishing Corp, New York.

Gorski, Przemek A., Delaine K. Ceholski, and Roger J. Hajjar. 2015. Altered Myocardial Calcium Cycling and Energetics in Heart Failure—A Rational Approach for Disease Treatment. Cell Metabolism. 21:183-194.

Graham, H.K., N.W. Hodson, J.A. Hoyland, S.J. Millward-Sadler, D. Garrod, A. Scothern, C.E.M. Griffiths, R.E.B. Watson, T.R. Cox, J.T. Erler, A.W. Trafford, and M.J. Sherratt. 2010. Tissue section AFM: In situ ultrastructural imaging of native biomolecules. Matrix Biology. 29:254-260.

Graham, M.S., and A.P. Farrell. 1989. The effect of temperature acclimation and adrenaline on the performance of a perfused trout heart. Physiological Zoology. 62:38-61.

Graham, M.S., and G.L. Fletcher. 1983. Blood and plasma viscosity of winter flounder: influence of temperature, red-cell concentration, and shear rate. Canadian Journal of Zoology-Revue Canadienne De Zoologie. 61:2344-2350.

Grant, C., D.G. Greene, and I.L. Bunnell. 1965. Left ventricular enlargement and hypertrophy. A clinical and angiocardiographic study. American Journal of Medicine. 39:895-904.

Granzier, H., M. Helmes, and K. Trombitas. 1996. Nonuniform elasticity of titin in cardiac myocytes: A study using immunoelectron microscopy and cellular mechanics. Biophysical Journal. 70.

Grossman, W., D. Jones, and L.P. McLaurin. 1975. Wall stress and patterns of hypertrophy in the human left ventricle. Journal of Clinical Investigation. 56:56-64.

Guderley, H., and J. St-Pierre. 2002. Going with the flow or life in the fast lane: contrasting mitochondrial responses to thermal change. Journal of Experimental Biology. 205:2237-2249.

Halper, J., and M. Kjaer. 2014a. Basic Components of Connective Tissues and Extracellular Matrix: Elastin, Fibrillin, Fibulins, Fibrinogen, Fibronectin, Laminin, Tenascins and Thrombospondins. In Progress in Heritable Soft Connective Tissue Diseases. J. Halper, editor. Springer Netherlands, Dordrecht. 31-47.

Halper, J., and M. Kjaer. 2014b. Basic Components of Connective Tissues and Extracellular Matrix: Elastin, Fibrillin, Fibulins, Fibrinogen, Fibronectin, Laminin, Tenascins and Thrombospondins. In Progress in Heritable Soft Connective Tissue Diseases. J. Halper, editor. Springer-Verlag Berlin, Berlin. 31-47.

Hansen, J., H. Gesser, and J. Altimiras. 2002. Mechanical efficiency of the trout heart during volume and pressure loading: metabolic implications of the stiffness of the ventricular tissue. Journal of Comparative Physiology B-Biochemical Systemic and Environmental Physiology. 172:477-484.

Harrison, S.M., and D.M. Bers. 1990. Temperature dependence of myofilament Ca2+ sensitivity of rat, guinea pig, and frog ventricular muscle. American Journal of Physiology. 258:C274-C281.

Haverinen, J., and M. Vornanen. 2004. Temperature acclimation modifies Na+ current in fish cardiac myocytes. Journal of Experimental Biology. 207.

Page | 114

Thermal remodelling of the ectothermic heart

Haverinen, J., and M. Vornanen. 2007. Temperature acclimation modifies sinoatrial pacemaker mechanism of the rainbow trout heart. American Journal of Physiology- Regulatory Integrative and Comparative Physiology. 292:R1023-R1032.

Haverinen, J., and M. Vornanen. 2009. Comparison of sarcoplasmic reticulum calcium content in atrial and ventricular myocytes of three fish species. American Journal of Physiology-Regulatory Integrative and Comparative Physiology 297

Heinbecker, P. 1931. The efffct of fibers of specific types in the vagus and sympathetic nerves on the sinus and atrium of the turtle and frog heart American Journal of Physiology. 98:220-229.

Helmes, M., K. Trombitas, and H. Granzier. 1996. Titin develops restoring force in rat cardiac myocytes. Circulation Research. 79.

Herbert, C.V., and D.C. Jackson. 1985a. Temperature Effects on the Responses to Prolonged Submergence in the Turtle Chrysemys picta bellii. II. Metabolic Rate, Blood Acid-Base and Ionic Changes, and Cardiovascular Function in Aerated and Anoxic Water. Physiological Zoology. 58:670-681.

Herbert, C.V., and D.C. Jackson. 1985b. Temperature effects on the responses to prolonged submergence in the turtle, Chrysemys picta bellii. I. Blood acid-base and ionic changes during and following anoxic submergence. Physiological Zoology. 58:655- 669.

Herbert, C.V., and D.C. Jackson. 1985c. Temperature effects on the responses to prolonged submergence in the turtle, Chrysemys picta bellii. II. Metabolic rate, blood acid-base and ionic changes, and cardiovascular function in aerated and anoxic water. Physiological Zoology. 58:670-681.

Herrmann, G., and G.M. Decherd. 1939. The chemical nature of heart failure. Annals of Internal Medicine. 12.

Hicks, J.M.T., and A.P. Farrell. 2000a. The cardiovascular responses of the red-eared slider (Trachemys scripta) acclimated to either 22 or 5 oC. I. Effects of anoxic exposure on in vivo cardiac performance. Journal of Experimental Biology. 203:3765-3774.

Hicks, J.M.T., and A.P. Farrell. 2000b. The cardiovascular responses of the red-eared slider (Trachemys scripta) acclimated to either 22 or 5 oC. II. Effects of anoxia on adrenergic and cholinergic control. Journal of Experimental Biology. 203:3775-3784.

Hicks, J.W., and S.G. Comeau. 1994. Vagal regulation of intracardiac shunting in the turtle Pseudemys scripta. Journal of Experimental Biology. 186:109-126.

Hicks, J.W., A. Ishimatsu, S. Molloi, A. Erskin, and N. Heisler. 1996. The mechanism of cardiac shunting in reptiles: A new synthesis. Journal of Experimental Biology. 199:1435- 1446.

Hicks, J.W., and G.M. Malvin. 1992. Mechanism of intracardiac shunting in the turtle, Pseudemys scripta. American Journal of Physiology. 262:R986-992.

Hicks, J.W., and T. Wang. 1998. Cardiovascular regulation during anoxia in the turtle: An in vivo study. Physiological Zoology. 71:1-14.

Hill, A.V. 1949. The abrupt transition from rest to activity in muscle. Proceedings of the Royal Society B-Biological Sciences. 136.

Hiroko, T., L. Bunshin, and F. Ken-Ichi. 1985. Pulsation rate and oxygen consumption of isolated hearts of the goldfish, Carassius auratus, acclimated to different

Page | 115

Thermal remodelling of the ectothermic heart

temperatures. Comparative Biochemistry and Physiology Part A: Physiology. 82:281- 283.

Horn, M.A., H.K. Graham, M.A. Richards, J.D. Clarke, D.J. Greensmith, S.J. Briston, M.C.S. Hall, K.M. Dibb, and A.W. Trafford. 2012. Age-related divergent remodeling of the cardiac extracellular matrix in heart failure: Collagen accumulation in the young and loss in the aged. Journal of Molecular and Cellular Cardiology. 53.

Horowits, R., K. Maruyama, and R.J. Podolsky. 1989. Elastic behavior of connectin filaments during thick filament movement in activated skeletal-muscle. Journal of Cell Biology. 109.

Hove-Madsen, L., and H. Gesser. 1989. Force frequency relation in the myocardium of rainbow trout: effects of K+ and adrenaline. Journal of Comparative Physiology B- Biochemical Systemic and Environmental Physiology. 159:61-69.

Huxley, H.E. 1969. The mechanism of muscular contraction. Science. 164:1356-1366.

Icardo, J.M. 2006. Conus arteriosus of the teleost heart: Dismissed, but not missed. Anatomical Record Part a-Discoveries in Molecular Cellular and Evolutionary Biology. 288A.

Icardo, J.M., and E. Colvee. 2011. The Atrioventricular Region of the Teleost Heart. A Distinct Heart Segment. Anatomical Record-Advances in Integrative Anatomy and Evolutionary Biology. 294.

Icardo, J.M., E. Colvee, M.C. Cerra, and B. Tota. 1999. Bulbus arteriosus of the antarctic teleosts. II. The red-blooded Trematomus bernacchii. Anatomical Record. 256.

Iftikar, F.I., J.R. MacDonald, D.W. Baker, G.M.C. Renshaw, and A.J.R. Hickey. 2014. Could thermal sensitivity of mitochondria determine species distribution in a changing climate? Journal of Experimental Biology. 217:2348-2357.

Iftikar, F.I., A.J. Morash, D.G. Cook, N.A. Herbert, and A.J.R. Hickey. 2015. Temperature acclimation of mitochondria function from the hearts of a temperate wrasse (Notolabrus celidotus). Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology. 184:46-55.

Ingwall, J.S. 2009. Energy metabolism in heart failure and remodelling. Cardiovascular Research. 81.

Ingwall, J.S., M.F. Kramer, M.A. Fifer, B.H. Lorell, R. Shemin, W. Grossman, and P.D. Allen. 1985. The creatine-kinase system in normal and diseased human myocardium. New England Journal of Medicine. 313.

Ingwall, J.S., and R.G. Weiss. 2004. Is the failing heart energy starved? On using chemical energy to support cardiac function. Circulation Research. 95.

Jackson, D.C. 1987. Cardiovascular function in turtles during anoxia and acidosis: in vivo and in vitro studies. American Zoologist. 27.

Jackson, D.C. 2002. Hibernating without Oxygen: Physiological Adaptations of the Painted Turtle. The Journal of Physiology. 543:731-737.

Jackson, D.C., S.J. Warburton, E.A. Meinertz, R.G. Lawler, and J.S. Wasser. 1995. The effect of prolonged anoxia at 3 oC on tissue high-energy phosphates and phosphodiesters in turtles - a P31 NMR study. Journal of Comparative Physiology B-Biochemical Systemic and Environmental Physiology. 165.

Page | 116

Thermal remodelling of the ectothermic heart

James, M.C., and N. Mrosovsky. 2004. Body temperatures of leatherback turtles (Dermochelys coriacea) in temperate waters off Nova Scotia, Canada. Canadian Journal of Zoology-Revue Canadienne De Zoologie. 82.

James, T.N., and L. Sherf. 1971. Specialized tissues and preferential conduction in the atria of the heart. American Journal of Cardiology. 28:414-427.

Johansen, K., and W. Burggren. 1980. Cardiovascular function in the lower vertebrates. In Hearts and Heart-Like Organs. G.H. Bourne, editor. Academic Press, Inc. 61-118.

Johnson, A.C., A.J. Turko, J.M. Klaiman, E.F. Johnston, and T.E. Gillis. 2014. Cold acclimation alters the connective tissue content of the zebrafish (Danio rerio) heart. Journal of Experimental Biology. 217:1868-1875.

Jones, D.R., R.W. Brill, and P.G. Bushnell. 1993. Ventricular and arterial dynamics of anesthetized and swimming tuna. Journal of Experimental Biology. 182:97-112.

Jones, D.R., B.L. Langille, D.J. Randall, and G. Shelton. 1974. Blood-flow in dorsal and ventral aortas of cod, Gadus morhua. American Journal of Physiology. 226:90-95.

Joyce, W., and T. Wang. 2014. Adenosinergic regulation of the cardiovascular system in the red-eared slider Trachemys scripta. Comparative Biochemistry and Physiology a- Molecular & Integrative Physiology. 174:18-22.

Jung, W.I., T. Hoess, M. Bunse, S. Widmaier, L. Sieverding, J. Breuer, J. Apitz, O. Schmidt, F. van Erckelens, G.J. Dietze, and O. Lutz. 2000. Differences in cardiac energetics between patients with familial and nonfamilial hypertrophic cardiomyopathy. Circulation. 101.

Jung, W.I., L. Sieverding, J. Breuer, T. Hoess, S. Widmaier, O. Schmidt, M. Bunse, F. van Erckelens, J. Apitz, O. Lutz, and G.J. Dietze. 1998. P-31 NMR spectroscopy detects metabolic abnormalities in asymptomatic patients with hypertrophic cardiomyopathy. Circulation. 97.

Junqueira, L.C.U., G. Bignolas, and R.R. Brentani. 1979. Picrosirius staining plus polarization microscopy, a specific method for collagen detection in tissue-sections. Histochemical Journal. 11:447-455.

Kalinin, A.L., M.J. Costa, F.T. Rantin, and M.L. Glass. 2009. Effects of Temperature on Cardiac Function in Teleost Fish. In Cardio-Respiratory Control in Vertebrates: Comparative and Evolutionary Aspects. L.M. Glass and C.S. Wood, editors. Springer Berlin Heidelberg, Berlin, Heidelberg. 121-160.

Katz, A.M. 2002. Ernest Henry Starling, His Predecessors, and the “Law of the Heart”. Circulation. 106:2986-2992.

Katz, A.M. 2006. Physiology Of The Heart (4th Ed.). Lippincott Williams And Wilkins, London.

Katz, A.M. 2008. The “Modern” View of Heart Failure: How Did We Get Here? Circulation: Heart Failure. 1:63-71.

Katz, A.M., and B.H. Lorell. 2000. Regulation of Cardiac Contraction and Relaxation. Circulation. 102:Iv-69-Iv-74.

Katz, A.M., and E.L. Rolett. 2016. Heart failure: when form fails to follow function. European Heart Journal.

Keef, K.D., J.R. Hume, and J.M. Zhong. 2001. Regulation of cardiac and smooth muscle Ca2+ channels (Ca(V)1.2a,b) by protein kinases. American Journal of Physiology-Cell Physiology. 281. Page | 117

Thermal remodelling of the ectothermic heart

Keen, A.N., A.J. Fenna, J.C. McConnell, M.J. Sherratt, P. Gardner, and H.A. Shiels. 2016. The dynamic nature of hypertrophic and fibrotic remodelling in the fish ventricle. Frontiers in Physiology - Intergrative Physiology. 6.

Keen, A.N., and A.K. Gamperl. 2012. Blood oxygenation and cardiorespiratory function in steelhead trout (Oncorhynchus mykiss) challenged with an acute temperature increase and zatebradine-induced bradycardia. Journal of Thermal Biology. 37:201- 210.

Keen, J.E., D.-M. Vianzon, A.P. Farrell, and G.F. Tibbits. 1994. Effect of temperature and temperature acclimation on the ryanodine sensitivity of the trout myocardium. Journal of Comparative Physiology B. 164:438-443.

Keen, J.E., D.M. Vianzon, A.P. Farrell, and G.F. Tibbits. 1993. Thermal-acclimation alters both adrenergic sensitivity and adrenoceptor density in cardiac tissue of rainbow-trout. Journal of Experimental Biology. 181.

Keith, A., and M. Flack. 1907. The form and nature of the muscular connections between the primary divisions of the vertebrate heart. journal of Anatomy and Physiology. 41:172- 189.

Kent, J., M. Koban, and C.L. Prosser. 1988. Cold acclimation-induced protein hypertrophy in channel catfish and green sunfish. Journal of Comparative Physiology B-Biochemical Systemic and Environmental Physiology. 158:185-198.

Kent, J.D., and C.L. Prosser. 1985. Protein hypertrophy in liver and heart following cold acclimatization and acclimation in channel catfish (ccf). American Zoologist. 25:A29- A29.

Kik, M.J.L., and M.A. Mitchell. 2005. Reptile cardiology: A review of anatomy and physiology, diagnostic approaches, and clinical disease. Seminars in Avian and Exotic Pet Medicine. 14:52-60.

Kirkpatrick, K.P., A.S. Robertson, J.M. Klaiman, and T.E. Gillis. 2011. The influence of trout cardiac troponin I and PKA phosphorylation on the Ca2+ affinity of the cardiac troponin complex. Journal of Experimental Biology. 214:1981-1988.

Klaiman, J.M., A.J. Fenna, H.A. Shiels, J. Macri, and T.E. Gillis. 2011. Cardiac Remodeling in Fish: Strategies to Maintain Heart Function during Temperature Change. PLoS ONE. 6:e24464.

Klaiman, J.M., W.G. Pyle, and T.E. Gillis. 2014. Cold acclimation increases cardiac myofilament function and ventricular pressure generation in trout. Journal of Experimental Biology. 217:4132-4140.

Kolwicz, S.C., D.P. Olson, L.C. Marney, L. Garcia-Menendez, R.E. Synovec, and R. Tian. 2012. Cardiac-Specific Deletion of Acetyl CoA Carboxylase 2 Prevents Metabolic Remodeling During Pressure-Overload Hypertrophy. Circulation Research. 111:728- +.

Kolwicz, S.C., S. Purohit, and R. Tian. 2013. Cardiac Metabolism and Its Interactions with Contraction, Growth, and Survival of the Cardiomyocte. Circulation research. 113:10.1161/CIRCRESAHA.1113.302095.

Konhilas, J.P., T.C. Irving, and P.P. de Tombe. 2002. Frank-Starling law of the heart and the cellular mechanisms of length-dependent activation. Pflugers Archiv-European Journal of Physiology. 445.

Page | 118

Thermal remodelling of the ectothermic heart

Korajoki, H., and M. Vornanen. 2012. Expression of SERCA and phospholamban in rainbow trout (Oncorhynchus mykiss) heart: comparison of atrial and ventricular tissue and effects of thermal acclimation. Journal of Experimental Biology. 215.

Koutsouris, D., E. Delatourhanss, and M. Hanss. 1985. Physicochemical factors of erythrocyte deformability. Biorheology. 22.

Krüger, M., T. Kohl, and W.A. Linke. 2006. Developmental changes in passive stiffness and myofilament Ca2+ sensitivity due to titin and troponin-I isoform switching are not critically triggered by birth. American Journal of Physiology - Heart and Circulatory Physiology. 291:H496-H506.

Kubota, M., M. Kinoshita, K. Takeuchi, S. Kubota, H. Toyohara, and M. Sakaguchi. 2003. Solubilization of type I collagen from fish muscle connective tissue by matrix metalloproteinase-9 at chilled temperature. Fisheries Science. 69:1053-1059.

Lai, N.C., J.B. Graham, N. Dalton, R. Shabetai, and V. Bhargava. 1998. Echocardiographic and hemodynamic determinations of the ventricular filling pattern in some teleost fishes. Physiological Zoology. 71:157-167.

Langille, B.L., and B. Crisp. 1980. Temperature dependence of blood viscosity in frogs and turtles: effect on heat exchange with environment. American Journal of Physiology. 239:R248-253.

Lattion, A.L., J.B. Michel, E. Arnauld, P. Corvol, and F. Soubrier. 1986. Myocardial recruitment during ANF mRNA increase with volume overload in the rat. American Journal of Physiology - Heart and Circulatory Physiology. 251:H890-H896.

Laurent, P., S. Holmgren, and S. Nilsson. 1983. Nervous and humoral control of the fish heart - Structure and function. Comparative Biochemistry and Physiology a-Physiology. 76.

Lee, L., C.E. Genge, M. Cua, X. Sheng, K. Rayani, M.F. Beg, M.V. Sarunic, and G.F. Tibbits. 2016. Functional Assessment of Cardiac Responses of Adult Zebrafish (Danio rerio) to Acute and Chronic Temperature Change Using High-Resolution Echocardiography. PLoS ONE. 11:e0145163.

Li, M.X., and P.M. Hwang. 2015. Structure and function of cardiac troponin C (TNNC1): Implications for heart failure, cardiomyopathies, and troponin modulating drugs. Gene. 571:153-166.

Liao, J., L. Yang, J. Grashow, and M.S. Sacks. 2005. Molecular orientation of collagen in intact planar connective tissues under biaxial stretch. Acta Biomaterialia. 1.

Licht, J.H., and W.S. Harris. 1973. The structure, composition and elastic properties of teleost bulbus arteriousus in carp, Cyprinus carpio. Comparative Biochemistry and Physiology. 46:699-704.

Linke, W.A. 2008. Sense and stretchability: The role of titin and titin-associated proteins in myocardial stress-sensing and mechanical dysfunction. Cardiovascular Research. 77:637-648.

Liu, Q.H., J.C. Busby, and J.D. Molkentin. 2009. Interaction between TAK1-TAB1-TAB2 and RCAN1-calcineurin defines a signalling nodal control point. Nature Cell Biology. 11:154-161.

Lødemel, J.B., H.K. Mæhre, J.-O. Winberg, and R.L. Olsen. 2004. Tissue distribution, inhibition and activation of gelatinolytic activities in Atlantic cod (Gadus morhua). Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology. 137:363-371.

Page | 119

Thermal remodelling of the ectothermic heart

Löffek, S., O. Schilling, and C.-W. Franzke. 2011. Biological role of matrix metalloproteinases: a critical balance. European Respiratory Journal. 38:191-208.

Lurman, G.J., L.H. Petersen, and A.K. Gamperl. 2012. In situ cardiac performance of Atlantic cod (Gadus morhua) at cold temperatures: long-term acclimation, acute thermal challenge and the role of adrenaline. Journal of Experimental Biology. 215:4006-4014.

Ma, Y., L.E. de Castro Brás, H. Toba, R.P. Iyer, M.E. Hall, M.D. Winniford, R.A. Lange, S.C. Tyagi, and M.L. Lindsey. 2014. Myofibroblasts and the extracellular matrix network in post-myocardial infarction cardiac remodeling. Pflugers Archiv. 466:1113-1127.

MacKenna, D.A., J.H. Omens, A.D. McCulloch, and J.W. Covell. 1994. Contribution of collagen matrix to passive left-ventricular mechanics in isolated rat hearts. American Journal of Physiology. 266.

Maenner, J., L. Thrane, K. Norozi, and T.M. Yelbuz. 2009. In Vivo Imaging of the Cyclic Changes in Cross-sectional Shape of the Ventricular Segment of Pulsating Embryonic Chick Hearts at Stages 14 to 17: A Contribution to the Understanding of the Ontogenesis of Cardiac Pumping Function. Developmental Dynamics. 238:3273- 3284.

Marshall, P.N., and R.W. Horobin. 1972. Oxidation-products of hematoxylin and their role in biological staining. Histochemical Journal. 4:493-503.

McBride, M., and S.J. Hernandez-Divers. 2004. Nursing care of lizards. The veterinary clinics of North America. Exotic animal practice. 7.

McConnell, J.C., O.V. O’Connell, K. Brennan, L. Weiping, M. Howe, L. Joseph, D. Knight, R. O’Cualain, Y. Lim, A. Leek, R. Waddington, J. Rogan, S.M. Astley, A. Gandhi, C.C. Kirwan, M.J. Sherratt, and C.H. Streuli. 2016. Increased peri-ductal collagen micro- organization may contribute to raised mammographic density. Breast Cancer Research. 18:1-17.

McMullen, J.R., and S. Izumo. 2004. Mechanisms controlling heart growth in mammals. In Cell growth: Control of cell size. M.N. Hall, M. Raff and G. Thomas, editors. Cold Spring Harbor Laboratory Press, 10 Skyline Drive, Plainview, NY, 11803-2500, USA. 549- 603.

McMullen, J.R., T. Shioi, W.Y. Huang, L. Zhang, O. Tarnavski, E. Bisping, M. Schinke, S. Kong, M.C. Sherwood, J. Brown, L. Riggi, P.M. Kang, and S. Izumo. 2004. The insulin- like growth factor 1 receptor induces physiological heart growth via the phosphoinositide 3-kinase(p110 alpha) pathway. Journal of Biological Chemistry. 279:4782-4793.

Mendonca, P.C., A.G. Genge, E.J. Deitch, and A.K. Gamperl. 2007. Mechanisms responsible for the enhanced pumping capacity of the in situ winter flounder heart (Pseudopleuronectes americanus). American Journal of Physiology-Regulatory Integrative and Comparative Physiology. 293:R2112-R2119.

Molkentin, J.D., J.R. Lu, C.L. Antos, B. Markham, J. Richardson, J. Robbins, S.R. Grant, and E.N. Olson. 1998. A calcineurin-dependent transcriptional pathway for cardiac hypertrophy. Cell. 93:215-228.

Moriyama, Y., Ito, F., Takeda, H., Yano, T., Okabe, M., Kuraku, S., Keeley, F. W., and K. Koshiba- Takeuchi. 2016. Evolution of the fish heart by sub/neofunctionalization of an elastin gene. Nature Communications 7, 10397

Page | 120

Thermal remodelling of the ectothermic heart

Mutungi, G., and K.W. Ranatunga. 1998. Temperature-dependent changes in the viscoelasticity of intact resting mammalian (rat) fast- and slow-twitch muscle fibres. The Journal of Physiology. 508:253-265.

Nadal-Ginard, B., J. Kajstura, A. Leri, and P. Anversa. 2003. Myocyte death, growth, and regeneration in cardiac hypertrophy and failure. Circulation Research. 92:139-150.

Nagase, H., R. Visse, and G. Murphy. 2006. Structure and function of matrix metalloproteinases and TIMPs. Cardiovascular Research. 69:562-573.

Neagoe, C., M. Kulke, F. del Monte, J.K. Gwathmey, P.P. de Tombe, R.J. Hajjar, and W.A. Linke. 2002. Titin isoform switch in ischemic human heart disease. Circulation. 106.

Negrini, D., Passi, A., Moriondo, A. 2008. The role of proteoglycans in pulmonary edema development. Intensive Care Medicine. 34:610-618.

Neubauer, S. 2007. Mechanisms of disease. The failing heart - An engine out of fuel. New England Journal of Medicine. 356.

Oka, T., M. Maillet, A.J. Watt, R.J. Schwartz, B.J. Aronow, S.A. Duncan, and J.D. Molkentin. 2006. Cardiac-specific deletion of Gata4 reveals its requirement for hypertrophy, compensation, and myocyte viability. Circulation Research. 98:837-845.

Olivetti, G., E. Cigola, R. Maestri, D. Corradi, C. Lagrasta, S.R. Gambert, and P. Anversa. 1996a. Aging, cardiac hypertrophy and ischemic cardiomyopathy do not affect the proportion of mononucleated and multinucleated myocytes in the human heart. Journal of Molecular and Cellular Cardiology. 28.

Olivetti, G., E. Cigola, R. Maestri, D. Corradi, C. Lagrasta, S.R. Gambert, and P. Anversa. 1996b. Aging, cardiac hypertrophy and ischemic cardiomyopathy do not affect the proportion of mononucleated and multinucleated myocytes in the human heart. Journal of Molecular and Cellular Cardiology. 28:1463-1477.

Olshansky, B., H.N. Sabbah, P.J. Hauptman, and W.S. Colucci. 2008. Parasympathetic Nervous System and Heart Failure: Pathophysiology and Potential Implications for Therapy. Circulation. 118:863-871.

Omland, T., H. Røsjø, E. Giannitsis, and S. Agewall. 2015. Troponins in heart failure. Clinica Chimica Acta. 443:78-84.

Overgaard, J., J.A.W. Stecyk, A.P. Farrell, and T. Wang. 2002. Adrenergic control of the cardiovascular system in the turtle Trachemys scripta. Journal of Experimental Biology. 205:3335-3345.

Overgaard, J., J.A.W. Stecyk, H. Gesser, T. Wang, and A.P. Farrell. 2004. Effects of temperature and anoxia upon the performance of in situ perfused trout hearts. Journal of Experimental Biology. 207:655-665.

Overgaard, J., T. Wang, O.B. Nielsen, and H. Gesser. 2005. Extracellular determinants of cardiac contractility in the cold anoxic turtle. Physiological and Biochemical Zoology. 78:976-995.

Patrick, S.M., H. Shiels, and O. Cazorla. 2009. Comparison of length-dependent calcium activation of cardiac myofilaments between the rat and the rainbow trout. Journal of Physiological Sciences. 59:536-536.

Pedersen, M.E., T.T. Vuong, S.B. Rønning, and S.O. Kolset. 2015. Matrix metalloproteinases in fish biology and matrix turnover. Matrix Biology. 44–46:86-93.

Page | 121

Thermal remodelling of the ectothermic heart

Pelliccia, A., B.J. Maron, F.M. Di Paolo, A. Biffi, F.M. Quattrini, C. Pisicchio, A. Roselli, S. Caselli, and F. Culasso. 2005. Prevalence and clinical significance of left atrial remodeling in competitive athletes. Journal of the American College of Cardiology. 46:690-696.

Pellman, J., R.C. Lyon, and F. Sheikh. 2010. Extracellular matrix remodeling in atrial fibrosis: Mechanisms and implications in atrial fibrillation. Journal of Molecular and Cellular Cardiology. 48:461-467.

Pelouch, V., and M. Vornanen. 1996. Effects of thermal acclimation on ventricle size, protein composition, and contractile properties of crucian carp heart. Journal of Thermal Biology. 21:1-9.

Peng, J., K. Raddatz, J.D. Molkentin, Y. Wu, S. Labeit, H. Granzier, and M. Gotthardt. 2007. Cardiac hypertrophy and reduced contractility in hearts deficient in the titin kinase region. Circulation. 115:743-751.

Perrino, C., and H.A. Rockman. 2006. GATA4 and the two sides of gene expression reprogramming. Circulation Research. 98:715-716.

Pieperhoff, S., W. Bennett, and A.P. Farrell. 2009. The intercellular organization of the two muscular systems in the adult salmonid heart, the compact and the spongy myocardium. Journal of Anatomy. 215:536-547.

Portner, H.O., and R. Knust. 2007. Climate change affects marine fishes through the oxygen limitation of thermal tolerance. Science. 315:95-97.

Pound, K.M., N. Sorokina, K. Ballal, D.A. Berkich, M. Fasano, K.F. LaNoue, H. Taegtmeyer, J.M. O'Donnell, and E.D. Lewandowski. 2009. Substrate-Enzyme Competition Attenuates Upregulated Anaplerotic Flux Through Malic Enzyme in Hypertrophied Rat Heart and Restores Triacylglyceride Content Attenuating Upregulated Anaplerosis in Hypertrophy. Circulation Research. 104.

Prado, L.G., I. Makarenko, C. Andresen, M. Kruger, C.A. Opitz, and W.A. Linke. 2005. Isoform diversity of giant proteins in relation to passive and active contractile properties of rabbit skeletal muscles. Journal of General Physiology. 126.

Priede, I.G. 1976. Functional morphology of bulbus arteriosus of rainbow trout (Salmo gairdneri Richardson). Journal of Fish Biology. 9:209-216.

Puchtler, H., F.S. Waldrop, and Valentin.Ls. 1973. Polarization microscopic studies of connective-tissue stained with picro-sirius red fba. Beitrage Zur Pathologie. 150:174- 187.

Rand, P.W., W.H. Austin, E. Lacombe, and H.E. Hunt. 1964. Viscosity of normal human blood under normothermic and hypothermic conditions. Journal of Applied Physiology. 19:117-122.

Raske, M., G.A. Lewbart, D.S. Dombrowski, P. Hale, M. Correa, and L.S. Christian. 2012. Body temperatures of selected amphibian and reptile species. Journal of Zoo and Wildlife Medicine. 43:517-521.

Rayment, I., H. Holden, M. Whittaker, C. Yohn, M. Lorenz, K. Holmes, and R. Milligan. 1993. Structure of the actin-myosin complex and its implications for muscle contraction. Science. 261:58-65.

Rezakhaniha, R., A. Agianniotis, J.T.C. Schrauwen, A. Griffa, D. Sage, C.V.C. Bouten, F.N. Vosse, M. Unser, and N. Stergiopulos. 2012. Experimental investigation of collagen waviness and orientation in the arterial adventitia using confocal laser scanning microscopy. Biomechanics and Modeling in Mechanobiology. 11:461-473. Page | 122

Thermal remodelling of the ectothermic heart

Rich, L., and P. Whittaker. 2005. Collagen and picrosirius red staining: A polarized light assessment of fibrillar hue and spatial distribution. Brazilian Journal of Morphological Sciences. 22:97-104.

Rodnick, K.J., A.K. Gamperl, G.W. Nash, and D.A. Syme. 2014. Temperature and sex dependent effects on cardiac mitochondrial metabolism in Atlantic cod (Gadus morhua L.). Journal of Thermal Biology. 44:110-118.

Rotermund, A.J., and C.A. Priviter. 1972. Effects of induced cold torpor on ATPase activity of turtle, Pseudemys (Chrysemys) picta. Comparative Biochemistry and Physiology. 41:511-520.

Ruegg, J.C. 1990. Towards a molecular understanding of contracticity. Cardioscience. 1:163- 168.

Russell, B., D. Motlagh, and W.W. Ashley. 2000. Form follows function: how muscle shape is regulated by work. Journal of Applied Physiology. 88:1127-1132.

Saffitz, J.E., and D. Corradi. The electrical heart: 25 years of discovery in cardiac electrophysiology, arrhythmias and sudden death. Cardiovascular Pathology. 25:149- 157.

Saito, M., Y. Takenouchi, N. Kunisaki, and S. Kimura. 2001. Complete primary structure of rainbow trout type I collagen consisting of alpha 1(I)alpha 2(I)alpha 3(I) heterotrimers. European Journal of Biochemistry. 268:2817-2827.

Sanchez-Quintana, D., V. Garcia-Martinez, V. Climent, and J.M. Hurle. 1995. Morphological analysis of the fish heart ventricle: Myocardial and connective tissue architecture in teleost species. Annals of Anatomy-Anatomischer Anzeiger. 177.

Sandblom, E., and M. Axelsson. 2006. Adrenergic control of venous capacitance during moderate hypoxia in the rainbow trout (Oncorhynchus mykiss): role of neural and circulating catecholamines. American Journal of Physiology-Regulatory Integrative and Comparative Physiology. 291.

Saravanakumar, M., J. Manivannan, J. Sivasubramanian, T. Silambarasan, E. Balamurugan, and B. Raja. 2011. Molecular metabolic fingerprinting approach to investigate the effects of borneol on metabolic alterations in the liver of nitric oxide deficient hypertensive rats. Molecular and Cellular Biochemistry. 362:203-209.

Saunders, D.K., and K.H. Patel. 1998. Comparison of blood viscosity in red-eared sliders (Trachemys scripta) adapted to cold and room temperature. Journal of Experimental Zoology. 281:157-163.

Schiaffino, S., J.L. Samuel, D. Sassoon, A.M. Lompré, I. Garner, F. Marotte, M. Buckingham, L. Rappaport, and K. Schwartz. 1989. Nonsynchronous accumulation of alpha- skeletal actin and beta-myosin heavy chain mRNAs during early stages of pressure- overload--induced cardiac hypertrophy demonstrated by in situ hybridization. Circulation Research. 64:937-948.

Schulte, P.M. 2007. Responses to environmental stressors in an estuarine fish: Interacting stressors and the impacts of local adaptation. Journal of Thermal Biology. 32:152-161.

Seebacher, F., and C.E. Franklin. 2005. Physiological mechanisms of thermoregulation in reptiles: a review. Journal of Comparative Physiology B-Biochemical Systemic and Environmental Physiology. 175.

Segura, A.M., O.H. Frazier, and L.M. Buja. 2012. Fibrosis and heart failure. Heart Failure Reviews. 19:173-185.

Page | 123

Thermal remodelling of the ectothermic heart

Sephton, D.H., and W.R. Driedzic. 1991. Effect of acute and chronic temperature transition on enzymes of cardiac metabolism in white perch (Morone americana), yellow perch (Perca flavescens), and smallmouth bass (Micropterus dolomieui). Canadian Journal of Zoology. 69:258-262.

Sephton, D.H., and W.R. Driedzic. 1995. Low temperature acclimation decreases rates of protein synthesis in rainbow trout (Oncorhynchus mykiss) heart. Fish Physiology and Biochemistry. 14:63-69.

Serafini-Fracassini, A., J.M. Field, M. Spina, S. Garbisa, and R.J. Stuart. 1978. The morphological organization and ultrastructure of elastin in arterial wall of trout (Salmo gairdneri) and salmon (Salmo salar). Journal of Ultrastructure Research. 65:1-12.

Seth, H., M. Axelsson, and A. Grans. 2014. The peptide hormone cholecystokinin modulates the tonus and compliance of the bulbus arteriosus and pre-branchial vessels of the rainbow trout (Oncorhynchus mykiss). Comparative Biochemistry and Physiology a- Molecular & Integrative Physiology. 178:18-23.

Shaffer, J.F., and T.E. Gillis. 2010. Evolution of the regulatory control of vertebrate striated muscle: the roles of troponin I and myosin binding protein-C. Physiological Genomics. 42:406-419.

Shattock, M.J., M. Ottolia, D.M. Bers, M.P. Blaustein, A. Boguslavskyi, J. Bossuyt, J.H.B. Bridge, Y. Chen-Izu, C.E. Clancy, A. Edwards, J. Goldhaber, J. Kaplan, J.B. Lingrel, D. Pavlovic, K. Philipson, K.R. Sipido, and Z.-J. Xie. 2015. Na+/Ca2+ exchange and Na+/K+-ATPase in the heart. The Journal of Physiology. 593:1361-1382.

Shen, W.Q., K. Asai, M. Uechi, M.A. Mathier, R.P. Shannon, S.F. Vatner, and J.S. Ingwall. 1999. Progressive loss of myocardial ATP due to a loss of total purines during the development of heart failure in dogs - A compensatory role for the parallel loss of creatine. Circulation. 100.

Shi, H.Y., and D.C. Jackson. 1997. Effects of anoxia, acidosis and temperature on the contractile properties of turtle cardiac muscle strips. Journal of Experimental Biology. 200:1965-1973.

Shiels, H.A., S.C. Calaghan, and E. White. 2006. The cellular basis for enhanced volume- modulated cardiac output in fish hearts. Journal of General Physiology. 128.

Shiels, H.A., A. Di Maio, S. Thompson, and B.A. Block. 2011. Warm fish with cold hearts: thermal plasticity of excitation-contraction coupling in bluefin tuna. Proceedings of the Royal Society B-Biological Sciences. 278:18-27.

Shiels, H.A., and A.P. Farrell. 1997. The effect of temperature and adrenaline on the relative importance of the sarcoplasmic reticulum in contributing Ca2+ to force development in isolated ventricular trabeculae from rainbow trout. Journal of Experimental Biology. 200:1607-1621.

Shiels, H.A., and G.L.J. Galli. 2014. The Sarcoplasmic Reticulum and the Evolution of the Vertebrate Heart. Physiology. 29:456-469.

Shiels, H.A., G.L.J. Galli, and B.A. Block. 2015. Cardiac function in an endothermic fish: cellular mechanisms for overcoming acute thermal challenges during diving. Proceedings of the Royal Society of London B: Biological Sciences. 282.

Shiels, H.A., and R. Sitsapesan. 2015. Is there something fishy about the regulation of the ryanodine receptor in the fish heart? Experimental Physiology. 100:1412-1420.

Page | 124

Thermal remodelling of the ectothermic heart

Shiels, H.A., M. Vornanen, and A.P. Farrell. 2000. Temperature-dependence of L-type Ca2+ channel current in atrial myocytes from rainbow trout. Journal of Experimental Biology. 203:2771-2780.

Shiels, H.A., M. Vornanen, and A.P. Farrell. 2002a. Effects of temperature on Ca2+ (i) in trout atrial myocytes. Journal of Experimental Biology. 205:3641-3650.

Shiels, H.A., M. Vornanen, and A.P. Farrell. 2002b. Temperature dependence of cardiac sarcoplasmic reticulum function in rainbow trout myocytes. Journal of Experimental Biology. 205:3631-3639.

Shiels, H.A., M. Vornanen, and A.P. Farrell. 2003. Acute temperature change modulates the response of ICa to adrenergic stimulation in fish cardiomyocytes. Physiological and Biochemical Zoology. 76:816-824.

Shiels, H.A., and E. White. 2008. The Frank-Starling mechanism in vertebrate cardiac myocytes. Journal of Experimental Biology. 211:2005-2013.

Shih, Y.H., Y.J. Zhang, Y.H. Ding, C.A. Ross, H. Li, T.M. Olson, and X.L. Xu. 2015. Cardiac Transcriptome and Dilated Cardiomyopathy Genes in Zebrafish. Circulation- Cardiovascular Genetics. 8:261-269.

Solaro, R.J., and H.M. Rarick. 1998. Troponin and tropomyosin - Proteins that switch on and tune in the activity of cardiac myofilaments. Circulation Research. 83.

Sorokina, N., J.M. O’Donnell, R.D. McKinney, K.M. Pound, G. Woldegiorgis, K.F. LaNoue, K. Ballal, H. Taegtmeyer, P.M. Buttrick, and E.D. Lewandowski. 2007. Recruitment of Compensatory Pathways to Sustain Oxidative Flux With Reduced Carnitine Palmitoyltransferase I Activity Characterizes Inefficiency in Energy Metabolism in Hypertrophied Hearts. Circulation. 115:2033-2041.

Starling, E.H. 1918. The Linacre Lecture of the Law of the Heart. Longman, Greens and Co, Ltd, London.

Stecyk, J., and A. Farrell. 2007. Control of turtle (Trachemys scripta) cardiac activity during anoxia: I. Effects of extracellular factors on spontaneous heart rate. Comparative Biochemistry and Physiology: Part A. Molecular & Integrative Physiology. 146.

Stecyk, J.A.W., C. Bock, J. Overgaard, T. Wang, A.P. Farrell, and H.O. Portner. 2009. Correlation of cardiac performance with cellular energetic components in the oxygen- deprived turtle heart. American Journal of Physiology-Regulatory Integrative and Comparative Physiology. 297:R756-R768.

Stecyk, J.A.W., and A.P. Farrell. 2002. Cardiorespiratory responses of the common carp (Cyprinus carpio) to severe hypoxia at three acclimation temperatures. Journal of Experimental Biology. 205:759-768.

Stecyk, J.A.W., and A.P. Farrell. 2006. Regulation of the cardiorespiratory system of common carp (Cyprinus carpio) during severe hypoxia at three seasonal acclimation temperatures. Physiological and Biochemical Zoology. 79.

Stecyk, J.A.W., G.L. Galli, H.A. Shiels, and A.P. Farrell. 2008. Cardiac survival in anoxia- tolerant vertebrates: An electrophysiological perspective. Comparative Biochemistry and Physiology C-Toxicology & Pharmacology. 148.

Stecyk, J.A.W., J. Overgaard, A.P. Farrell, and T. Wang. 2004. Adrenergic regulation of systemic peripheral resistance and blood flow distribution in the turtle Trachemys scripta during anoxic submergence at 5 oC and 21 oC. Journal of Experimental Biology. 207:269-283.

Page | 125

Thermal remodelling of the ectothermic heart

Stecyk, J.A.W., V. Paajanen, A.P. Farrell, and M. Vornanen. 2007a. Effect of temperature and prolonged anoxia exposure on electrophysiological properties of the turtle (Trachemys scripta) heart. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology. 293:R421-R437.

Stecyk, J.A.W., V. Paajanen, A.P. Farrell, and M. Vornanen. 2007b. Effect of temperature and prolonged anoxia exposure on electrophysiological properties of the turtle (Trachemys scripta) heart. American Journal of Physiology-Regulatory Integrative and Comparative Physiology. 293:R421-R437.

Stecyk, J.A.W., K.-O. Stenslokken, G.E. Nilsson, and A.P. Farrell. 2007c. Adenosine does not save the heart of anoxia-tolerant vertebrates during prolonged oxygen deprivation. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology. 147:961-973.

Stevens, E.D., G. Shelton, D.J. Randall, and G.R. Bennion. 1972. Factors affecting arterial pressures and blood flow from heart in intact, unrestrained lingcod, Ophiodon elongatus. Comparative Biochemistry and Physiology. 43:681-695.

Stienen, G.J.M. 2015. Pathomechanisms in heart failure: the contractile connection. Journal of Muscle Research and Cell Motility. 36:47-60.

Stullken, D.E., F.M. White, and W.A. Hiestand. 1949. Effect of cold adaptation on highest and lowest temperatures at which turtle heart continues beating. Federation Proceedings. 8.

Sun, X., T. Hoage, P. Bai, Y. Ding, Z. Chen, R. Zhang, W. Huang, A. Jahangir, B. Paw, Y.-G. Li, and X. Xu. 2009. Cardiac Hypertrophy Involves Both Myocyte Hypertrophy and Hyperplasia in Anemic Zebrafish. PLoS ONE. 4:e6596.

Tattersall, G.J., V. Cadena, and M.C. Skinner. 2006. Respiratory cooling and thermoregulatory coupling in reptiles. Respiratory Physiology & Neurobiology. 154.

Taylor, S.E., S. Egginton, and E.W. Taylor. 1996. Seasonal temperature acclimatisation of rainbow trout: Cardiovascular and morphometric influences on maximal sustainable exercise level. Journal of Experimental Biology. 199:835-845.

Threader, R.W., and A.H. Houston. 1983. Heat tolerance and resistance in juvenile rainbow trout acclimated to diurnally cycling temperatures. Comparative Biochemistry and Physiology Part A: Physiology. 75:153-155.

Tiitu, V., and M. Vornanen. 2001. Cold adaptation suppresses the contractility of both atrial and ventricular muscle of the crucian carp heart. Journal of Fish Biology. 59:141-156.

Tiitu, V., and M. Vornanen. 2002. Regulation of cardiac contractility in a cold stenothermal fish, the burbot Lota lota L. Journal of Experimental Biology. 205:1597-1606.

Tillmanns, H., S. Ikeda, H. Hansen, J.S. Sarma, J.-M. Fauvel, and R.J. Bing. 1974. Microcirculation in the Ventricle of the Dog and Turtle. Circulation Research. 34:561- 569.

Toischer, K., A.G. Rokita, B. Unsöld, W. Zhu, G. Kararigas, S. Sossalla, S.P. Reuter, A. Becker, N. Teucher, T. Seidler, C. Grebe, L. Preuß, S.N. Gupta, K. Schmidt, S.E. Lehnart, M. Krüger, W.A. Linke, J. Backs, V. Regitz-Zagrosek, K. Schäfer, L.J. Field, L.S. Maier, and G. Hasenfuss. 2010. Differential Cardiac Remodeling in Preload Versus Afterload. Circulation. 122:993-1003.

Tota, B., F. Farina, and G. Zummo. 1988. Ultrastructural aspects of functional interest in the ventricular myocardial wall of the antarctic icefish Chaenocephalus aceratus.

Page | 126

Thermal remodelling of the ectothermic heart

Comparative Biochemistry and Physiology Part B: Comparative Biochemistry. 90:561- 566.

Tripet, B., J.E. Van Eyk, and R.S. Hodges. 1997. Mapping of a second actin-tropomyosin and a second troponin C binding site within the C terminus of troponin I, and their importance in the Ca2+-dependent regulation of muscle contraction1. Journal of Molecular Biology. 271:728-750.

Trombitas, K., Y. Wu, D. Labeit, S. Labeit, and H. Granzier. 2001. Cardiac titin isoforms are coexpressed in the half-sarcomere and extend independently. American Journal of Physiology-Heart and Circulatory Physiology. 281.

Ultsch, G.R. 2006. The ecology of overwintering among turtles: where turtles overwinter and its consequences. Biological Reviews. 81:339-367.

Valentinuzzi, M.E., and H.E. Hoff. 1972. The sinus venosus-atrial Wenckebach-Luciani phenomenon. Journal of Electrocardiology. 5:1-14.

Ventura-Clapier, R., A. Garnier, and V. Veksler. 2004. Energy metabolism in heart failure. Journal of Physiology-London. 555.

Verheule, S., E. Wilson, T. Everett, S. Shanbhag, C. Golden, and J. Olgin. 2003. Alterations in atrial electrophysiology and tissue structure in a canine model of chronic atrial dilatation due to mitral regurgitation. Circulation. 107:2615-2622.

Vornanen, M. 1989. Regulation of contractility of the fish (Carassius carassius L.) heart ventricle. Comparative Biochemistry and Physiology Part C: Comparative Pharmacology. 94:477-483.

Vornanen, M. 1998. L-type Ca2+ current in fish cardiac myocytes: Effects of thermal acclimation and beta-adrenergic stimulation. Journal of Experimental Biology. 201:533-547.

Vornanen, M., M. Hassinen, H. Koskinen, and A. Krasnov. 2005. Steady-state effects of temperature acclimation on the transcriptome of the rainbow trout heart. American Journal of Physiology-Regulatory Integrative and Comparative Physiology. 289:R1177-1184.

Vornanen, M., A. Ryokkynen, and A. Nurmi. 2002a. Temperature-dependent expression of sarcolemmal K+ currents in rainbow trout atrial and ventricular myocytes. American Journal of Physiology-Regulatory Integrative and Comparative Physiology. 282.

Vornanen, M., H.A. Shiels, and A.P. Farrell. 2002b. Plasticity of excitation-contraction coupling in fish cardiac myocytes. Comparative Biochemistry and Physiology A-Molecular and Integrative Physiology. 132:827-846.

Vornanen, M. 2016. Temperature-dependence of electrical excitability in fish hearts. Journal of Experimental Biology

Wang, T., J. Altimiras, and M. Axelsson. 2002. Intracardiac flow separation in an in situ perfused heart from Burmese python Python molurus. Journal of Experimental Biology. 205:2715-2723.

Wang, T., and J. Overgaard. 2007. The heartbreak of adapting to global warming. Science. 315:49-50.

Wasser, J.S., K.C. Inman, E.A. Arendt, R.G. Lawler, and D.C. Jackson. 1990. P31 NMR measurements of phi and high-energy phosphates in isolated turtle hearts during anoxia and acidosis. American Journal of Physiology. 259.

Page | 127

Thermal remodelling of the ectothermic heart

Watanabe, K., P. Nair, D. Labeit, M.S.Z. Kellermayer, M. Greaser, S. Labeit, and H. Granzier. 2002. Molecular mechanics of cardiac titin's PEVK and N2B spring elements. Journal of Biological Chemistry. 277.

Watson, A.D., and J.L.S. Cobb. 1979. A comparative study on the innervation and the vascularization of the bulbus arteriosus in teleost fish. Cell and Tissue Research. 196:337-346.

Weber, K.T. 1989. Cardiac interstitium in health and disease - the fibrillar collagen network. Journal of the American College of Cardiology. 13.

Weber, K.T., C.G. Brilla, and J.S. Janicki. 1993. Myocardial fibrosis: functional significance and regulatory factors. Cardiovascular Research. 27.

Weber, K.T., J.E. Jalil, J.S. Janicki, and R. Pick. 1989. Myocardial collagen remodeling in pressure overload hypertrophy - a case for interstitial heart-disease. American Journal of Hypertension. 2.

Weber, K.T., and J.S. Janicki. 1989. Angiotensin and the remodeling of the myocardium. British Journal of Clinical Pharmacology. 28.

Wilson, K., and P.A. Lucchesi. 2014. Myofilament dysfunction as an emerging mechanism of volume overload heart failure. Pflügers Archiv - European Journal of Physiology. 466:1065-1077.

Wolska, B.M., G.M. Arteaga, J.R. Peña, G. Nowak, R.M. Phillips, S. Sahai, P.P. de Tombe, A.F. Martin, E.G. Kranias, and R.J. Solaro. 2002. Expression of Slow Skeletal Troponin I in Hearts of Phospholamban Knockout Mice Alters the Relaxant Effect of β-Adrenergic Stimulation. Circulation Research. 90:882-888.

Wong, A.Y.K., and P.M. Rautaharju. 1968. Stress distribution within left ventricular wall approximated as a thick ellipsoidal shell. American Heart Journal. 75:649-&.

Wood, C.M., P. Pieprzak, and J.N. Trott. 1979. Influence of temperature and anemia on the adrenergic and cholinergic mechanisms controlling heart rate in the rainbow trout. Canadian Journal of Zoology-Revue Canadienne De Zoologie. 57:2440-2447.

Woodard, G.E., and J.A. Rosado. 2008. Natriuretic peptides in vascular physiology and pathology. In International Review of Cell and Molecular Biology, Vol 268. K.W. Jeon, editor. 59-93.

Wu, Y.M., O. Cazorla, D. Labeit, S. Labeit, and H. Granzier. 2000. Changes in titin and collagen underlie diastolic stiffness diversity of cardiac muscle. Journal of Molecular and Cellular Cardiology. 32.

Xin, M., E.N. Olson, and R. Bassel-Duby. 2013. Mending broken hearts: cardiac development as a basis for adult heart regeneration and repair. Nat Rev Mol Cell Biol. 14:529-541.

Yamauchi, A., Y. Fujimaki, and K. Kumagai. 1974. Direct contact between a neuron soma and myocardial cells; an electronmicroscopic finding in the sinus venosus of the turtle heart. The Anatomical Record. 179:491-496.

Yang, H., J. Velema, M.S. Hedrick, G.F. Tibbits, and C.D. Moyes. 2000. Evolutionary and Physiological Variation in Cardiac Troponin C in Relation to Thermal Strategies of Fish. Physiological and Biochemical Zoology. 73:841-849.

Yin, Z., J. Ren, and W. Guo. 2015. Sarcomeric protein isoform transitions in cardiac muscle: A journey to heart failure. Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease. 1852:47-52.

Page | 128

Thermal remodelling of the ectothermic heart

Young, S., and S. Egginton. 2011. Temperature acclimation of gross cardiovascular morphology in common carp (Cyprinus carpio). Journal of Thermal Biology. 36:475- 477.

Zhang, Y.T., L. Weaver, A. Ibeawuchi, and K.R. Olson. 1998. Catecholaminergic regulation of venous function in the rainbow trout. American Journal of Physiology-Regulatory Integrative and Comparative Physiology. 274.

Zile, M.R., and D.L. Brutsaert. 2002. New concepts in diastolic dysfunction and diastolic heart failure: Part I Diagnosis, prognosis, and measurements of diastolic function. Circulation. 105.

Zwikster, G.H., and T.E. Boyd. 1936. Influence of potassium concentration on temperature of cold standstill in the turtle heart. Proceedings of the Society for Experimental Biology and Medicine. 34.

Page | 129

Thermal remodelling of the ectothermic heart

Page | 130

Thermal remodelling of the ectothermic heart

12. APPENDIX

Page | 131

Predicted futurechangesto where, forfish,f relate toefficiency ofmovement. disturbed byincreasingwatertemperature( beat kinematicsofrainbowtrout,Oncorhynchusmykiss,were maintaining kinematicsatasingleoptimum(orpreferred) experimental manipulationthatsupportstheimportanceof kinematics. Thisstudypresentsthefirstevidenceusingan this adequatetoindicateitsimportanceingoverningwingortail falling withinthebroadoptimumrange(0.2< based onobservationaldataandsupportedbytheory, shows concomitant increasesinA et al.,1991),indicatingthatanimals areevolutionarilyconstrained animals (Taylor etal.,2003;Triantafyllou etal.,1993;Triantafyllou appear topeakwithin0.2< frequency, wakewidthandfree-streamvelocity, respectively. affect theincrementalincreaseinmetaboliccostwithincreasing increase in 2244 Received 9January 2014;Accepted20March2014 *Author ([email protected]) for correspondence Manchester, Life Sciences, University of ManchesterFaculty M139PT, of UK. converge onasingle similar organisms moving in dynamicallysimilarwaysshould 2004; Taylor etal.,2003;Triantafyllou etal.,1993).Geometrically dimensionless parameter, theStrouhalnumber( propulsion isthoughttoconverge uponanoptimumrangeofa animals movingataconstantcruisingspeedusingoscillatory The relationshipbetweenmovementandforwardvelocity( flume, Swimming KEY WORDS:Fish,Locomotion,Respirometry, Salmonid,Swim The relationshipbetweentail(orwing)beatfrequency( Robert L.Nudds*,EmmaJohn,AdamN.KeenandHollyA.Shiels propulsion adherence toadistinctStrouhalnumberduringanimaloscillatory Rainbow troutprovidethefirstexperimentalevidencefor RESEARCH ARTICLE © 2014.PublishedbyTheCompanyofBiologistsLtd|JournalExperimentalBiology(2014)217,2244-2249doi:10.1242/jeb.102236 optimum rangeoftheStrouhalnumber( when movingataconstantcruisingspeed,convergesuponan ( 20±1°C. ElevatedT salmonids. the next100 years)maynotpresentmajorlocomotoryproblemsto INTRODUCTION ABSTRACT speed (m stroke amplitudeofthetailtip(m),and any givenU A ) andforwardvelocity(U s T − 1 water ), whicharerepresentativeofvortexshedding was conserved.St tail also increasedbasalmetaboliccosts,butdidnot water is tailbeatfrequency(Hz), St increased and itisintuitivetoexpectthatnumber ) inanimalsusingoscillatorypropulsion, St St<0.4 forbothswimmingandflying , whilstf T water = fA St tail f of lakesandrivers(5–10°Cover U tail is definedas: increased withU ⋅ and decreased tail St= ,(1) was unaffected byU. U f tail is equaltoswimming St<0.4) andconsiders · T A water / U A St) (Nuddsetal., ). Previouswork, is peak-to-peak ) from11±1 to A f tail , whilstSt , drivenby ), amplitude St. Thetail U ) in An U St St at . by temperature,with well establishedthatvertebrate skeletalmusclefunctionisaffected fin) drivenbymuscletoproduce thrustforforwardpropulsion.Itis carangiform swimmersrelying onoscillationsoftheirtail(caudal manipulated. Therefore,whether organism adheringtoagiven al., 1993;Triantafyllou etal.,1991);there is noevidenceforan (Eloy, 2012;Nuddsetal.,2004;Taylor etal.,2003;Triantafyllou et of speeds.Allpreviousworkonanimal 1984; James,2013),and,asfish areectotherms,changesto biomechanical efficiency of swimming. oxygen carryingcapacityofwater, mayhave implicationsforthe aerobic scopeatwarmtemperatures,coupledwithadecreased 2007). Becauselocomotionisenergetically costly, the reduced minimum metabolicrate)(Farrell,2009;Wang andOvergaard, a reducedaerobiccapacity(thedifference betweenmaximumand Elevated water speed(HodgsonandQuinn,2002;Klemetsenetal.,2003). fish copewithchangesinsalinity, watertemperature( (Farrell etal.,2008;Finstad2005).Duringthesemigrations, completion ofareturnmigrationtonatalriverinespawninggrounds variation inthemeasuredvaluesvalid,isnotcertain. of oscillatorypropulsionkinematicsand,hence,theexplanationsfor that eels, Norberg andWinter, 2006).Incontrast,Tytell (Tytell, 2004)found 2005; Webb, 1971),birds(Tobalske etal.,1999)andbats(Lindhe 2004), however, showsthattherangeof look atthedataandsubsequentmeasurements(RohrFish, to anarrowoptimumrangeforhighpropulsiveefficiency. A closer will directlyaffect muscletemperature(Johnston andTemple, 2002; ectotherms, soincreased temperature inexcessof19°C(HodgsonandQuinn,2002).Fish are successful migrationsbeingestablishedbyapopulation at a water over21°C(Farrelletal.,2008),andthereisnoevidence of 2008). Reportsshowthatsalmontemporarilyceasemigration in systems (Karppinenetal.,2002;LaineThorstad et al., having anegativeimpactonmigrationsuccessinsomeriver increasing al., 2003).St also needtoproduceliftasinbirds(Nuddsetal.,2004;Taylor et they aremainlythrustproducers(suchasmostfishandinsects)or et al.,1978).St cetaceans, surprising thatvariationsin organisms, thelessgeometricallysimilartheyare;therefore,itisnot actually alarge target initself.Ofcourse,thebroaderrangeof varies considerably. Indeed, theoptimumrangeof0.2< Rainbow trout,Oncorhynchus mykiss Reproductive successformanysalmonidspeciesdependson T Anguilla rostrata, maintainedaconstant St water U in fish(HunterandZweifel,1971;LauderTytell, appears tochangewithbodysize(Eloy, 2012;Kayan linked toclimatechange,however, isthoughttobe also differs betweenanimalsdependingonwhether also varieswith Q 10 reported torangefrom1.53.0 (Bennett, T water St St elevates metabolicrate,resultingin are evidentacrosstaxa.Infishand St when itskinematicsaredirectly U is reallyanimportantgovernor , generallydecreasingwith St, however, isobservational (Walbaum 1792),aresub- St measured inanimals St across arange T St<0.4 is water ) and T water

The Journal of Experimental Biology slight, however, being0.19forbothtreatmentsat0.68 U individual fish(Table ramnsars l pes(i.1 al 1).Theincreasein 1,Table treatments acrossallspeeds(Fig. 4 as, temperature levelsthroughatrade-off between a changeinfishkinematicswhereby modulated throughanincreasein V basal processesandnotnecessarily swimminglocomotionperse. temperature suggeststhattheincreased metaboliccostisfocusedon 2).Thesameabsolute increaseacrossallspeedsatthewarm (Fig. the warmtemperaturetreatmentthanatcontrol increased, and 0.22 and0.23at1.11 maintain optimumswimmingkinematics. U species andU obviates theaforementionedvariationin individuals ofasinglespeciesinbothtemperaturetreatments temperatures overarangeofincreasing beat kinematicswerequantifiedinrainbowtroutswimmingattwo increased second hypothesiswasthatthemetabolicstresscausedby also dependentuponanobservablechangeinkinematics.The however, andofcourse,whetherthishypothesisistrulytestedwas both treatments(Table St of temperatureuponmigratoryswimmingability. optimum experimentally testingtheimportanceofadherencetoagiven manipulations mayprovideapotentialmechanismfor to beaffected by Rome etal.,1984).Previousstudieshaveshowntailbeatkinematics RESEARCH ARTICLE respectively (Fig. in thecontroltreatment(4.17±0.09 was higherinthewarmtemperaturetreatment(4.70±0.08 among individualfish.Theincreased 0.005±0.001 and 0.003±0.001 more markedathigher, morephysiologicallydemandingspeeds. section wassmallforboth RESULTS ·

O mm) inthecontrolgroupatallspeeds. V In thepresentstudy, A Fore–aft positionalchange(accelerations) withintheworking V V U U T T St M l V f A List of symbols List of would notbeaffected bytemperature,asthefishwouldstriveto 2 based uponthelinesofbestfitfromGLM(Fig. body tail · · increased with water opt m O O U crit differed betweenindividualfish(Table b increased concomitantlywith,andatthesameincrementalrate 2 2 increased withincreasing in bothtreatments(Fig. St, aswellbeingofimportanceinidentifyingtheeffects T water . Thehypothesestestedwerethatthe rate ofoxygenconsumption fish volume maximum aerobicswimmingspeedofthefish water temperature optimum temperature fish bodymass body length tail beatfrequency swim tunnelvolume forward velocity(swimmingspeed) Strouhal number peak-to-peak strokeamplitudeofthetailtip A would leadtoanincreasein , whichwasdecreased. T

1A). water U

1). Incontrast,f , butdidnotdiffer betweentemperature

changes (Stevens,1979).Consequently, 1). V m · O V

2 s · O − (rate ofoxygenconsumption)andtail 2 1 , however, washigheratallspeedsin for controlandwarmtreatments,

1B). U A

T m water . Thewarmtemperatureinduced at thesameincrementalratein

s Hz) (Table − A tail St 2 , however, washigher(by treatments (meanswere was notaffected by was maintainedatcontrol St for controland warm A St

1). that isduetobodysize, St did notdiffer between with U V · may varyacrossU O . Usingthesame

1). 2 , whichwouldbe St f U tail f used atagiven tail , whichwas is therefore also varied

1A) was Hz) than

s St − U 1 , and T , but with water , differ betweenfish, accelerations duringtherecordedswimmingbouts,andtheydid not η increase in evident, but Which variablethetemperature directlyaffected (Fig. disturbed, theyadheredtothe same As hypothesised,whenthelocomotor systemofthetroutwas effect between temperatures, respectively).Furthermore,therewasnointeraction ( Fig. the fishwereswimmingatasteadyandnear-constant the fishremainedincentreofworkingsection,means that trials. This,coupledwiththefactthatdatawereonlycollected when was bothminimalandconsistentthroughoutalloftheexperimental F best-fit calculatedfromtheGLMare model (GLM)are Strouhal number( red circlesandsolidlinesrepresentthewarmtemperaturedata.(A)For experimental (involving amanipulation)evidence foranapparent DISCUSSION temperature data,respectively. (B)Fortailbeatamplitude( predictable relationshipbetweentailbeatfrequency( for coldandwarmtemperaturedata,respectively. (C)Therewasno U p 1,66 ). 2 =0.13,

Blue squaresanddashedlinesrepresentthecoldtemperaturedata, .Scatterplotsofkinematicsparameters againstswimmingspeed 1. =3.34,

f A St

P tail (Hz) (m) The JournalofExperimentalBiology(2014)doi:10.1242/jeb.102236 η 0.02 0.04 0.06 0.08 0.15 0.25 0.35 f =0.204: p 0.1 0.2 0.3 tail St 2 =0.05, 0 0 1 2 3 4 5 6 7 0.6 0.6 0.6 y T St), thelinesofbestfitcalculatedfromgenerallinear =0.128+0.087 and decreasein was maintainedthroughatrade-off betweenan water B A C and T

T P 0.7 0.7 0.7 water water =0.072). Hence,fore–aftpositionalchange U , treatment orwith

( x 0.8 0.9 0.8 0.9 0.8 0.9 F F and y 1,65 1,66 y U (ms =–0.010+0.064 =0.80, =0.909, A =0.131+0.087 . Nonetheless,thisisthefirst St –1 ) η at anygiven p

2 η 1 1 1 <0.01, p 2 =0.01, f x tail x U and y ) andU.

for coldandwarm 1.1 1.1 1.1 (fish, P A =–0.014+0.064 =0.375) onthe ), thelinesof P

U U

1B,C) isnot 1.2 1.2 1.2 =0.344: F . (Fig. 7,66 =1.44, 2245 1A). U x ,

The Journal of Experimental Biology V f previous findings.Earlierworkinfishsuggeststhat al., 2007;Shadwicket2004). Sepulveda, 2005;Bernaletal.,2009;Donley2012; (Altringham andBlock,1997;Bernaletal.,2005; temperatures inanumberofbonyfishandsharkspecies maximum poweroutputofredmuscledecreasedatlower 1984; James,2013),andmusclecontractionkineticsareslowed are widelyagreedtoshowhightemperaturedependence(Bennett, (Fig. 1C). Thecontractilepropertiesofvertebratemusclefunction 2013), withthecentreoftheseboundsincreasingtemperature frequency maybeoptimisedwithinverynarrowbounds(James, 2246 RESEARCH ARTICLE line) andwarm(solid line)temperaturedata,respectively. output are (red circles)treatments. speed (U with Fig. with oscillatory propulsion.Atbothtemperatures,aslightincreasein adherence toapreferred(perhapsoptimum) y Removed termsarepresentedinorderofdeletionfromtheGLMs. T St y temperature. consumption Table all velocities.Invariant (increased) whilst AT 2004). Inbirds,f was alsofoundacrosssevenspeciesofcetaceans(RohrandFish, 1958; Webb, 1971).A similarincreasing although atverylowspeeds Rome etal.,1984;Stevens,1979;Tytell, 2004;Webb, 1971), (Bainbridge, 1958;HunterandZweifel,1971;Romeetal.,1990; tail · , dependentvariable; O 2 An invariant

.Rate ofoxygenconsumption(V 2. ·

–1 –1 U V U O2 (ml min kg ) 1. Statisticaloutputfromthegenerallinearmodels(GLMs)foreachofthreekinematicmeasuresandrateoxygen was drivenbyincreasesin 100 200 300 400 500 600 700 800 [e.g. seetable ) forcoldtemperature(bluesquares) andwarmtemperature y 0 0.6 0.7 =–410.33+977.10 f U T T eoe em FinalGLM Removed terms tail wing water water water water , F and increasing 1,66 St, Strouhalnumber; also appearstoincreasewith × U × U × U × U A =1.47, η 2 inTobalske etal.(Tobalske etal.,1999)]. The linesofbestfitcalculatedfrom the GLM , F , F , F , F remains constantwithincreasing x 1,48 1,65 1,65 1,65

f 0.8 and y tail A =2.90, η =0.79, η =0.07, η =3.21, η p 2 =0.02, P may alsobemodulated(Bainbridge, suggests thatmusclecontraction =–270.27+977.10 U (ms A A ·

O 0.9 p p p p with increasing , whilstf 2 2 2 2 2 ) plottedagainstswimming =0.06, P =0.01, P <0.01, P =0.05, P 020Fish, =0.230 A –1 , tail-beatamplitude; ) f tail and invariant

=0.095 =0.376 =0.791 =0.078 tail St 1 was constantacross x for ananimalusing U for cold(dashed , butA f U tail

1.1 is contraryto is modulated decreases f A tail , tail-beatfrequency; with

1.2 U U U Fish, F T T T T Fish, F Fish, F St U U water water water water , F , F , F 1,49 1,66 1,66 , F , F , F , F F 7,49 7,67 7,66 7,66 =112.29, =38.86, η =7.01, η may bespeciesdependent.Forexample,increased kinematics arelimitedandtheeffects arenotconsistent,whichagain present study. Previousstudiesontemperaturechangeandfishtail (Rohr andFish,2004),whichiscontrarytothefindingsof the was foundtobeindependentof and ~0.41to0.30injackmackerel(HunterZweifel,1971), and with resulting fromstresseffects. the lowSt et al.,1999;Read2003;Triantafyllou etal.,1991).Perhaps a muchreducedpropulsiveefficiency (Andersonetal.,1998;Barrett however, isstillproducedoutsideoftheoptimumrange,albeitwith several valuesfellbelowtheoptimumrange(Fig. propulsive efficiency (Triantafyllou etal.,1993).Nevertheless, 0.20 and0.40,withinthehypothesizedoptimumrangeforhigh differences in relationship between the phaseanglebetweentailheave andpitch)willsubtlyaffect the is likelythattailshapeandother kinematicparameters(forexample, proxy forwakewidth.Although thisisareasonableassumption,it or differences inmethodology. Indeed, be speciesspecific,asaresultof different dataanalysisapproaches are congruent(Stevens,1979).Again,thesedifferences arelikelyto 1971) arecontrarytothefindingsofpresentstudy, thoseforbass Therefore, althoughthepreviousdataforrainbowtrout(Webb, in speedeffects ontail-beatkinematics. is perhapsprematuretospeculateatlengthabouttheincongruence standardised conditions,whichremainsapriorityforfuturework,it thunniform). Therefore,withoutamulti-speciesstudyunder swimming form(i.e.carangiform,sub-carangiform,anguilliformor to habitat(i.e.stillwaterversusflowingspecies)and to expectdifferences intailbeatkinematicsacrossspecies,relating preferred cruisingspeed(RohrandFish,2004).Also,itisintuitive were cajoledintoswimmingatmaximum elliptical pool,inabroadrangeoftemperaturesfrom12to20°Cand example, thecetaceanswereswimmingfreelyaroundalarge electric grid),thecameraresolutionandspeciesstudied.For flume, themethodforpersuadingfishtoswim(lightversus those previouslyconducted:thespecificationanddesignof There are,however, manydifferences betweenthepresentstudyand 1979) andhadnoeffect on f f al., 1990).T 1,49 1,67 1,66 1,66 tail =7.87, η =2.24, η =1.32, η =2.47, η The majorityofcalculated =72.24, η =22.70, η =5.94, η =0.25, η in bass(Stevens,1979),decreased U p , rangingfrom0.45to0.30inrainbowtrout(Webb, 1971) V 2 η · p =0.10, P O 2 p p p p p 2 =0.37, P 2 p p 2 2 2 2 , rateofoxygenconsumption; =0.70, P 2 2 =0.53, P =0.19, P =0.12, P =0.21, P p p =0.08, P <0.01, P The JournalofExperimentalBiology(2014)doi:10.1242/jeb.102236 2 2 water =0.60, P =0.25, P recorded herewereduetounpreferredkinematics St =0.010 <0.001 also didnotaffect calculations between species.Ifhydrodynamics <0.001 <0.001 =0.041 =0.253 =0.026 =0.018 =0.618 <0.001 <0.001 A and wakewidth,whichinturnwould leadto tail St St in carp,Cyprinuscarpio for therainbowtroutwerebetween was previouslyshowntodecrease U A U in sevenspeciesofcetacean , swimmingspeed; f in carp(Romeetal.,1990). tail A in rainbowtrout(Stevens, U is generallychosenasa , whichmaynotbetheir T water T

1). Thrust, water (Rome et increased , water

The Journal of Experimental Biology T changes toT foraging orpredatoravoidance.Furthermore,futurepredictionsof top speedattainable,whichmaybeadvantageousformigration, with fastertopspeedsbeingmeasuredathigher increased Indeed, fishrefusingtoswimat Furthermore, acutetemperaturechanges(increasing et al.,2001;Gamperl2002;Steinhausen2008). (Jain etal.,1997)–isinagreementwithpreviousstudies(Brodeur of 2°C (Fig. least, cancopewiththiseasilyintermsofswimmingbiomechanics. 100 (equal to respirometer (~67.2 St, eventhoughtheirtailbeatkinematicsweredisruptedbya performance, atleastwithinthethermaltolerancezoneoffish. U flumes thatdonotemployelectricgrids(Rodnicketal.,2004). levelling off athigh V adhere toUKHomeOffice legislation. housing conditionswereinaccordance withthelocalhandlingprotocolsand week, butwerefastedfor24 3 were closelymonitoredand30%waterchangesconducted every freshwater tanksat11±1°C. Temperature, pH,ammoniaandnitratelevels housed ona12 mean bodymass=262.20±0.03 Eight sexuallyimmaturefemalerainbowtrout(meanlength=0.29±0.003 University ofManchester. licence (40/3584)heldbyH.A.S.andwereundertheethicalapproval ofthe All experimentalprocedureswerecoveredbyaUKHomeOffice project preferred) importance ofmaintainingkinematicsatasingleoptimum(or change. Thisisthefirstexperimentalevidenceforpotential at higher as significantlyotherparameters,showninthispresentstudy. dominate, however, thenforanindividualfish, RESEARCH ARTICLE T increases inbasalenergy expenditure.Infact,marginal increases in when facedwiththermaladjustmentstomusclefunctionand to maintaintheirmovementpatternsatabiomechanicaloptimum rainbow trout,andperhapsotherfishspecies,possesstheplasticity cost (changesin costs, butcontrarytohypothesised,itdidnotaffect theincremental basal/routine processes.Anincreasein et al.,2002).Here, 1997). Thisfindingsuggeststhatwarmer (Farrell, 2002;Romeetal.,1990;1984;Taylor etal., eight fish(Fig. Fish Ethics statement MATERIALS ANDMETHODS Respirometry This resultmaybeexplainedifthefishwerenotfullyreaching ·

O water water days. Fishwerefedtosatiationoncommercialtroutpelletsthreetimes a crit 2 In conclusion,rainbowtroutatanygiven Swimming performancealsoappearedtobeinfluencedby V · was measuredusingstop-flowrespirometry inaBrett-styleswimflume O years (Sharmaetal.,2007),anditappearsthatrainbowtrout,at – definedasthemaximumaerobicswimmingspeedoffish 2 2). Theincrementalchangein values and,therefore,thehypothesisedgreaterincreasein may actuallyimproveswimmingperformanceintermsofthe increased with h − U 1 T ) leadingtoincreased ) wascalibratedusing ahandheldHFA flow meter (Höntzsch T water St. Anincreasein water water was notobserved.Theincreasedmetaboliccostsat h:12

2), whichisinagreementwithpreviouswork of lakesandriversareonly5–10°Coverthenext appear tobeaconsequenceofanincreasein V · l volume;Loligo Systems,Denmark).Water velocity O U 2 V h dark:lightcycleinaerated500 · U O with (Gamperl etal.,2002;Thorstad2008). 2 and washigheratthe continued torisetowards

h priortoexperimentation.Allhusbandry and U

g) (ChirkTrout Farm,Wrexham, UK)were ) ofswimming.Itwouldappearthat T U water V · O crit 2 V is welldocumented(Altimiras · is frequentlyreportedinswim did increasebasalmetabolic O 2 V with · O T 2 water with increasingU U U adhered tothesame may aidswimming was similaratboth T St water should notvary U T T crit

water l re-circulated water in fiveoutof , insteadof treatment by arate T T U until water water V crit

· m, O 2 , . swimming speedtrial.Duringthisperiod,U of theswimtunnelandconnectedtoaPCrunningOxyView a temperature-calibrateddippingprobe,situatedintherespirometrychamber a fibre-opticoxygenmeter(modelFIBOX3LCD,PreSens,Germany)and (0.17×0.17×0.65 0.0033l two temperaturetreatments. treatments sodonotcompromisethevalidityofcomparisonbetween fish buoyancyorblockageeffects. Ofcourse,thesewerethesameinboth end ofthevideoclip.Nocorrectionsto positional variationswithintheworkingsectionbetweenbeginningand centre oftheworkingsection. framework). Datawereonlycollectedfromfishthatswimminginthe relative toeachindividual’s bodylength( flume respirometer. Fishlengthwasmeasuredsothatspeedincrements the twotemperatures. of therespirometryandkinematicevaluationwerekeptconstantbetween aerobic capacity(Farrell,2009)andaffect swimmingkinematics.Allaspects l 30 (LCDPST3 V1.16).LoggerPro 1 recommencing swimming.Atthe point ofexhaustion, stopped swimming,anddidnot respond toabrightlightstimulusby recorded ateach Technology, USA)wasusedtocalculatetherateofdeclineinO filming. To determine working sectionwasbrieflyilluminatedbyasinglehalogenlampduring section sothatthesensorplanewasparalleltowatersurface.The the fish(DanosandLauder, 2012). 1.0 slight changesinwaterviscositybetweenthetwotemperatures(~1.2and camera filmingat100 Tail beatkinematicswerecapturedusinganHDR-SR8E(Sony, Japan)video GmbH, Germany)and (~1l above optimumtemperature( because 20°Cisbelowthecriticalthermalmaximumforrainbowtrout,but (control) temperatureandthenat20±1°C.Thiswaschosen Experiments wererunonallfishat11±1°C, whichwasalsotheacclimation Experiments wereconductedonindividualfishoverthecourseof3 water) andM respirometer and weighed todetermine (mg Steinhausen etal.,2008).U where in thereservoir. in theswimtunnelwasalwaysmaintainedat>90%byairstonespositioned 12 flume at11±1°C andthewaterwasgraduallyheatedto20±1°Cover respirometer. Forexperimentsat20°C,thefishwasplacedinswim remaining quiescentandtoensurecontinuousmixingwithin the 70–75% ofU approximately 70–75%of actual where then calculatedusingtheformula: using Tracker 4 Kinematic data collection Experimental protocol body l body Fish werenettedfromtheholdingtankandtransferredintoswim Fish werethenheldintherespirometerfor12 For theswimmingtrial,fishwassubjectedtoarampincreasein h periodpriortothestartofswimmingtrial.Water oxygensaturation min untilexhaustion(indicative of mPa / body O U s 2 . U V − body l l 1 U s body − s at11 and20°C,respectively)wereunlikelytohaveanyeffects on is swimtunnelvolume(l), − (0.28 1 and afteranhour ofrecoverythefishwasremoved fromthe 1 min ) toprovideacurrentforthefishorientatetowardswhilst s was increasedfrom1 increments wereusedtokeepindividualfisheffort constant,but − The JournalofExperimentalBiology(2014)doi:10.1242/jeb.102236 1 b crit − until thefishwasswimmingatarateof2.5 m is fishbodymass(kg). 1 ) byfittingalinearregressiontotheO

© m) oftheflume.Water oxygencontent wasmeasuredwith have anear-maximal aerobiccapacity(Lee etal.,2003; s U V −  1 . Exhaustionwasdefinedasthepoint atwhichthefish O video analysissoftware(OpenSourcePhysics,Java 2 to 1.11 A

= frames U , thepositionoftailtipwasdigitizedandtracked aeo dcie(–)60 ) – ( decline Rate of O was measuredinthecentreofworkingsection U was thenincreasedatarateof0.2 m crit T s

s opt − . Studieshaveshownthatfishswimmingat − 1 1 ) andthusshouldcauseareductionintheir U ) wasusedinthedataanalyses.Notethat positioned onatripodabovetheworking ® 2m V was adjustedtoaccountforfore–aft l (Version 3.4;Vernier Softwareand m body is fishvolume(assuming1 M M U s U b b − . were madetoaccountforeither l crit 1 ⋅⋅ body VV ). Kinematicsand was maintainedat0.28 (resting) inincrementsof ; m)couldbecalculatedas h priortothestartof 2 U –time data. ,(2) was decreasedto l body l body s − ® 1 , whichis

V 2 kg=1 s software · O − V content

· 1 weeks. 2 O 2247 m every 2 were was l of s U, − 1

The Journal of Experimental Biology n term rendering thedataunsuitableforarepeated-measuresmodel.Theinteraction values inbothtemperaturetreatments,thenumberofincrementsvaried, measures takenforeachfish.Althoughallthefishswamatarangeof a covariate.IndividualwasincludedintheGLMbecauseofrepeated P (Table this valuethatisnotedinthestatisticaloutputssubsequenttext Satterthwaite’s correctionofthedenominatordegreesfreedomanditis group varianceswereencountered.Inthesecases,SPSSimplementsa and meansaredisplayed±s.e.m.InsomecaseswithintheGLMsunequal GLM. significant, theinteractionterm( relationship betweenthekinematicsvariableand 2248 RESEARCH ARTICLE study, andprovidedfundsforswim-flumeperipherals. A UniversityofManchesterInvestinginSuccessAward toH.A.S.supportedthe and draftingrevisingthemanuscript. and A.N.K.conductedthestudy, andwereinvolvedininterpretationofthefindings interpretation ofthefindings,anddraftingrevisionmanuscript.E.L.J. R.L.N. andH.A.S.wereinvolvedinconception,designexecutionofthestudy, The authorsdeclarenocompetingfinancialinterests. manuscript. two anonymousreviewersfortheirhelpfulcommentsonanearlierversionofthe generous donationoftheswim-flumerespirometer. We wouldalsoliketothank The authorswouldliketothankProf.PatButler, UniversityofBirmingham,forthe normality (Shapiro–Wilk test): The datadistributionofeachthekinematicsvariablesdidnotdiffer from liia,J,Aeso,M,Carax . ernos . ece,C n Farrell, and C. Mercier, C., Lefrancois, G., Claireaux, M., Axelsson, J., Altimiras, T the workingsection)wereinvestigatedusinggenerallinearmodels(GLMs). control) uponthekinematicsvariables(andfore–aftpositionalchangewithin Data analyses References Funding Author contributions Competing interests Acknowledgements lrnhm .D n lc,B.A. Block, and J.D. Altringham, enl .adSplea C.A. Sepulveda, and D. Bernal, nesn .M,Srile,K,Bret .S n ratflo,M.S. Triantafyllou, and D.S. Barrett, K., Streitlien, J.M., Anderson, enl . ye . cilva,D,Dne,J n euvd,C. Sepulveda, and J. Donley, D., McGillivray, D., D.A. Syme, Syme, D., and Bernal, R.E. Shadwick, J.M., Donley, D., Bernal, A.F. Bennett, ao,N n adr G.V. Lauder, and N. R.S. McKinley, Danos, and D.G. Dixon, J.C., Brodeur, and M.A. Grosenbaugh, D.K.P., Yue, M.S., Triantafyllou, D.S., Barrett, anrde R. Bainbridge, oly .M,Splea .A,Ales .A,MGliry .G,Sm,D .and D.A. Syme, D.G., McGillivray, S.A., Aalbers, C.A., Sepulveda, J.M., Donley, D.A. Syme, and C.A. Sepulveda, R.E., Shadwick, J.M., Donley, =76, water =0.185). acclimation temperatures. A. P. ectothermic fish. muscle temperatures?Poweroutputofisolatedfromendothermicand Copeia aerobic swimmingmusculatureofthecommonthreshershark, R217-R229. frequency andamplitudeofthetailbeat. Oscillating foilsofhighpropulsiveefficiency. 183-212. shark. Integr. Comp.Biol. effect oftemperatureonthemusclecontractilepropertiesincommonthresher muscles powerswimminginacold-watershark. M.J. Wolfgang, increasing waterviscosity. output asapredictorofmetabolicratein rainbowtrout. enl D. Bernal, shark andshortfinmakoshark. dependence ofcontractileproperties the aerobiclocomotormuscleinleopard Possible effects ofthetwotemperature( All statisticalanalyseswereperformedusingIBM T was includedasafactor, individualfishasarandomfactorand

P 1). (2002). Cardiorespiratorystatusoftriploidbrowntroutduringswimmingattwo water =0.214), 146-151. (2012). Effects oftemperatureonpoweroutput andcontractionkineticsin × (1984). Thermaldependenceofmusclefunction. U (1958). Thespeedofswimmingfishasrelatedtosizeandthe J. Exp.Biol. was includedtotestfordifferences intheslopeof (1999). Dragreductioninfish-likelocomotion. f tail ( W =0.984, 49, E199. J. FishBiol. J. Exp.Biol. 200 (2012). Challengingzebrafishescape responses by J. Exp.Biol. , 2617-2627. (2005). Evidencefortemperatureelevationinthe St n =76, ( T (1997). Whydotunamaintainelevatedslow W 60, 102-116. water 215 =0.985, J. Exp.Biol. , 1854-1862. P 210 J. FluidMech. × =0.431) and U , 1194-1203. Nature ) wasremovedfromthefinal n T =76, water (2001). Assessmentofcardiac 35, 109-133. J. FishBiol. 437 ) treatments(warmand P ® U 360 =0.535), , 1349-1352. V SPSS · . Ifnotstatistically O , 41-72. 2 (2005). Mammal-like Am. J.Physiol. ( J. FluidMech. W ® Alopias vulpinus 58, 439-452. =0.972, (2007). Thermal Statistics v.20 A ( (2009). The W =0.978, (1998). n U =59, 247 392 as U , . , apr,A . onc,K . as,H . en .C,Bnet .T,Crawshaw, M.T., Bennett, E.C., Venn, H.A., Faust, K.J., Rodnick, A.K., Gamperl, Lapointe, G.T., Crossin, D.A., Patterson, S.J., Cooke, S.G., Hinch, A.P., Farrell, A.P. Farrell, A.P. Farrell, C. Eloy, an,A,Jkvra .adKtpds C. Katopodis, and T. Jokivirta, A., Laine, Steinhausen, isa,A . kad . hrtd .B n egegt T. G. Heggberget, and E.B. Thorstad, F., Okland, A.G., Finstad, ed .A,Hvr .S n ratflo,M.S. Triantafyllou, and F. S. Hover, D.A., Read, A.L. R. Thomas, and G.K. Taylor, R.L., Nudds, onc,K . apr,A . ias .R,Bnet .T,Ruc,R .and R. N. Rausch, M.T., Bennett, K.R., Lizars, A.K., Gamperl, K.J., Rodnick, hra . ako,D . in,C .adSue,B.J. Shuter, and C.K. Minns, D.A., Jackson, S., Sharma, J.M. Donley, and D. A. Syme, D., Bernal, R.E., Shadwick, R.M. Alexander, and R.P. Funke, L.C., Rome, ae,R.S. James, e,C . arl,A . ot,A,Mcut .J,Hnh .G n ely M.C. Healey, and S.G. Hinch, M.J., MacNutt, A., Lotto, A.P., Farrell, C.G., Lee, E.D. Tytell, and G.V. Lauder, Klemetsen, V. E. Pyatetskii, and L.F. Kozlov, V. P., Kayan, A.P. Farrell, and J.C. Hamilton, K.E., Jain, J.R. Zweifel, and J.R. Hunter, T. P. Quinn, and S. Hodgson, ideNreg .M n itr Y. Winter, and U.M. Lindhe Norberg, or .J n ih F. E. Fish, and J. Rohr, oe .C,Luha .T n odpn,G. Goldspink, and P. T. Loughna, L.C., Rome, ontn .A n epe G.K. Temple, and I.A. Johnston, apie,P,Mknn .S,Eknr,J,Ksi,V . akvkj R.V., Sadkovskij, V. V., Kostin, J., Erkinaro, T. S., Makinen, P., Karppinen, (2008). Theeffect ofacutetemperature increases onthecardiorespiratory vulpinus the locomotormuscleofregionallyendothermiccommonthreshershark( ’onl,M .adMresn E. trout Mortensen, and M.F. O’Connell, their lifehistories. H.W. Li, mykiss and M.S. Powell, performance, andtissuebiochemistryofhighdesertredbandtrout( E.R., Keeley, L. I., trutta. J.FishBiol. upriver spawningmigrationofAtlanticsalmon Biochem. Zool. M.T.models andbiotelemetrytopredictthesuccessofspawningmigrations. Mathes, and M. 3771-3780. study oftemperaturephysiologyandrivermigrationsalmonids. Biochem. Physiol. high temperature:insightsintocardiovasculardesignlimitationsinfishes. 30, 205-218. efficiency, fishentranceandenvironmentalfactors. and seatrout, 2052-2064. populations beinhotwaterbecause of climatechange? Biochem. Physiol. stocks. J.Exp.Biol. Zool. into freshwater:adaptationsbypopulationstoprevailingthermalregimes. Biochem. Zool. numbers. feeding bat, for propulsionandmaneuvering. 2071-2076. amplitude relateandscalewithsizeflightspeedinbirds. for highpropulsiveefficiency accuratelypredictshowwingbeatfrequencyandstroke trout populationsinsouth-easternOregon. E.R. Keeley, Lamna ditropis water: contractilepropertiesofmusclefrom theendothermicAlaskansalmonshark, 163-178. on musclevelocityandsustainedperformance inswimmingcarp. R279. odontocete cetaceans. J. Exp.Biol. phenotype inectothermicvertebratesanditssignificanceforlocomotorybehaviour. skeletal muscle. measure criticalswimmingspeedinrainbowtrout( consumption inadultsockeye( (2003). Theeffect oftemperature onswimmingperformanceandoxygen MA: AcademicPress. Fish Physiology the swimmingofcertainaquaticanimals. 30. Fishery Bull.FishWildl. Serv. U.S. amplitude, andsizeinjackmackerel, a functionofswimmingspeedandmuscletemperature. uadn .I n akrna M. of ascendingAtlanticsalmonintheregulatedRiverTuloma. Kaukoranta, and A.I. Lupandin, Salmo trutta 80, 542-555. (2012). OptimalStrouhalnumberforswimminganimals. ssp.): evidenceforphenotypicdifferences inphysiologicalfunction. ). FishPhysiol.Biochem. J. Exp.Biol. . mnsn .A,Dmsn .B,Jnsn . oso,N., Jonsson, B., Jonsson, J.B., Dempson, P. A., Amundsen, A., (2009). Environment,antecedentsandclimatechange:lessonsfromthe (2013). A reviewofthethermalsensitivitymechanicsvertebrate (2002). Cardiorespiratoryperformanceinsalmonidsduringexerciseat 205 The JournalofExperimentalBiology(2014)doi:10.1242/jeb.102236 .F,Snbo,E,Eisn .J,Vril,C n arl,A.P. Farrell, and C. Verhille, E.J., Eliason, E., Sandblom, M. F., Glossophaga soricina (2004). Thermaltoleranceandmetabolicphysiologyamongredband . Integr. Comp.Biol. 81, 697-709. 75, 413-431. , Vol. 23(ed.E.S.RobertandV. L.George),pp.425-468.Waltham, Salmo trutta , 2305-2322. J. Comp.Physiol.B Ecol. Freshwat.Fish L. andArcticcharr 132A 117A 67, 919-930. 206 209 J. Exp.Biol. (2008). Pacificsalmoninhotwater:applyingaerobicscope , 441-444. , 3239-3251. , 797-810. (2004). Strouhalnumbersandoptimizationofswimmingby , 3887-3897. (2002). Thetimingofadultsockeyesalmonmigration L., passageinaregulatednorthernriver–fishway (1971). Swimmingspeed,tailbeatfrequency, tailbeat (2005). Hydrodynamicsofundulatorypropulsion.In Oncorhynchus nerka J. FluidsStructures 38, 1507-1519. 69, 253-266. 44, 746. 207 , flyingatdifferent flight speeds andStrouhal 183 (2002). Thermalplasticityofskeletalmuscle (2002). Migratoryandroute-seekingbehaviour Salvelinus alpinus (2003). Atlanticsalmon 12, 1-59. , 1633-1642. (2006). Wingbeatkinematicsofanectar- , 723-733. Trachurus symmetricus Fluid Dyn. J. FishBiol. (2002). Atlanticsalmon, (1997). Useofarampvelocitytestto (1984). Musclefiberactivityincarpas (1990). Theinfluenceoftemperature (1978). Kinematiccharacteristicsof Salmo salar (2004). Tuning ofStrouhalnumber (2003). Forcesonoscillatingfoils Fish. Manag.Ecol. 13, 641-646. ) andcoho( 17, 163-183. Oncorhynchus mykiss (2002). Metabolism,swimming 64, 310-335. (L.): areviewofaspects Am. J.Physiol. (2004). Hotsharkincold (2007). Willnorthernfish Glob. Chang.Biol. Hydrobiologia Salmo salar and seatrout Proc. Biol.Sci. J. FluidsStructures , andotherfishes. O. kisutch)salmon (2005). Comparing J. Exp.Biol. J. Exp.Biol. Salmo salar Oncorhynchus 9 , 65-77. 247 L., brown ). 483 Physiol. Physiol. , R272- Can. J. Alopias Comp. Comp. Salmo , 23- 212 271 154 13, L., , , ,

The Journal of Experimental Biology RESEARCH ARTICLE hrtd .B,Oln,F,Arsrp .adHgbre,T. G. Heggberget, and K. Aarestrup, F., Okland, E.B., Thorstad, alr .W,Egno,S,Tyo,S .adBte,P. J. Butler, and S.E. Taylor, S., Egginton, E.W., Taylor, E.D. Stevens, alr .K,Nds .L n hms A.L.R. Thomas, and R.L. Nudds, G.K., Taylor, human impacts. affecting thewithin-riverspawningmigrationofAtlanticsalmon,withemphasison Exp. Biol. performance ofrestingandswimmingsockeyesalmon( cruise ataStrouhalnumbertunedforhighpowerefficiency. pp. 105-133.Cambridge:CambridgeUniversityPress. Implications forFreshwaterandMarineFish limit swimmingperformanceatdifferent temperatures.In constant velocity. 211, 3915-3926. (1979). Effect oftemperatureontailbeatfrequencyfishswimmingat Rev. FishBiol.Fish. Can. J.Zool. 57, 1628-1635. 18, 345-371. (ed. D.G.McDonaldandC.M.Wood), (2003). Flyingandswimminganimals (1997). Factorswhichmay Oncorhynchus nerka Nature Global Warming. 425 (2008). Factors , 707-711. ). J. yel E.D. Tytell, ratflo,G . ratflo,M .adGoebuh M.A. Grosenbaugh, and M.S. Triantafyllou, G.S., Triantafyllou, ag .adOegad J. Overgaard, and T. Wang, ratflo,M . ratflo,G .adGplrsnn R. Gopalkrishnan, and G.S. Triantafyllou, M.S., Triantafyllou, K.P. Dial, and W. L. Peacock, B.W., Tobalske, eb P. W. Webb, Structures thrust developmentinoscillatingfoilswithapplicationtofishpropulsion. J. Exp.Biol. mechanics forthrustgenerationinoscillatingfoils. flight inthezebrafinchoverawiderangeofspeeds. cruising speeds. warming. Science (2004). ThehydrodynamicsofeelswimmingII.Effect ofswimmingspeed. 7 (1971). Theswimmingenergeticsoftrout.I.Thrustandpoweroutputat , 205-224. 207 The JournalofExperimentalBiology(2014)doi:10.1242/jeb.102236 , 3265-3279. J. Exp.Biol. 315 , 49-50. (2007). Ecology. Theheartbreakofadaptingtoglobal 55, 489-520. (1999). Kinematicsofflap-bounding Phys. FluidsA 3 J. Exp.Biol. 202 , 2835-2837. (1993). Optimal , 1725-1739. (1991). Wake J. Fluids 2249

The Journal of Experimental Biology Thermal remodelling of the ectothermic heart

Journal of Fish Biology (2016) 88, 403–417 doi:10.1111/jfb.12788, available online at wileyonlinelibrary.com

Kinematics and energetics of swimming performance during acute warming in brown trout Salmo trutta J. M. D. Lea, A. N. Keen, R. L. Nudds and H. A. Shiels*

Faculty of Life Sciences, University of Manchester, Manchester, M13 9NT, U.K.

This study examined how acute warming of water temperature affects the mechanical efficiency of swimming and aerobic capabilities of the brown trout Salmo trutta. Swimming efficiency was assessed using the relationship between swimming kinematics and forward speed (U), which is thought to converge upon an optimum range of a dimensionless parameter, the Strouhal number (St). Swim-tunnel intermittent stopped-flow respirometry was used to record kinematics and measure oxygen consump- ˙ tion (MO2)ofS. trutta during warming and swimming challenges. Salmo trutta maintained St between 0⋅2and0⋅3 at any given U over a range of temperatures, irrespective of body size. The maintenance of St within the range for maximum efficiency for oscillatory propulsion was achieved through an increase in tail-beat frequency (f tail) and a decrease in tail-beat amplitude (A) as temperature increased. Maintenance of efficient steady-state swimming was fuelled by aerobic metabolism, which increased as temperature increased up to 18∘ C but declined above this temperature, decreasing the apparent metabolic scope. As St was maintained over the full range of temperatures whilst metabolic scope was not, the results may suggest energetic trade-offs at any given U at temperatures above thermal optima. © 2015 The Fisheries Society of the British Isles

Key words: aerobic scope; locomotion; respirometry; Strouhal number; tail-beat amplitude; tail-beat frequency.

INTRODUCTION Brown trout Salmo trutta L. 1758 are a cold water salmonid species that can make long migrations to spawning grounds, both potadromous and anadromous, to feed and reproduce (Jonsson & Jonsson, 2011). In the Mediterranean region, populations have declined coincident with warming and climate change scenarios predict that more than half of the suitable habitat could disappear from the lower reaches of S. trutta’s range by 2040 (Almodóvar et al., 2012). Each S. trutta population is likely to respond differently to warming depending on location, habitat and migration route, making management and conservation challenging (Elliott & Elliott, 2010). Recent work on Atlantic salmon Salmo salar L. 1758, however, has shown that thermal tolerances may be less population specific than originally thought (Anttila et al., 2014). There are three recognized speeds that salmonids adopt when swimming: sustained, prolonged and burst (Beamish, 1978). Fishes use a variety of gaits at different swimming speeds, which are supported by different combinations of red and white

*Author to whom correspondence should be addressed. Tel.: +44 161 275 5092; email: [email protected] 403

© 2015 The Fisheries Society of the British Isles 404 J. M. D. LEA ET AL. muscle fibres (Peake & Farrell, 2004). It is likely that the optimal gait for a particular speed corresponds with the most efficient muscle dynamics. The relationship between wing and tail kinematics and forward speed in animals using oscillatory propulsion and moving at a constant speed can be described by the dimensionless parameter, the Strouhal number (St) (Triantafyllou et al., 1993; Taylor et al., 2003; Nudds et al., − 1 2004). For fishes, St (y) is defined as: y = f tailAU , where A is the tail-beat amplitude (m) calculated as the lateral tail-tip excursion, f tail is tail-beat frequency (Hz) and −1 ⋅ ⋅ U (m s ) is the forward speed of the fish. St numbers in the range of 0 2–0 4are considered optimal for propulsive efficiency in terms of the ratio of hydrodynamic power output to mechanical power input and animals appear to operate in this range (Triantafyllou et al., 1993; Taylor et al., 2003; Rohr & Fish, 2004). Similar St may be achieved using different combinations of A, f tail and U. Hence, if one of the tail kinematic variables is disturbed in some way [i.e. by temperature (T)], a fish could alter the other kinematic variable or U and still maintain swimming at an optimum St. Recent work has shown that in rainbow trout Oncorhynchus mykiss (Walbaum, 1792) an acute increase in water T (from 11 to 20∘ C) resulted in a decrease in A and an increase in f tail whilst St remained unaltered at any given U, suggesting the fish maintained kinematics at a single optimum (or preferred) St (Nudds et al., 2014). The size range of the O. mykiss studied in Nudds et al. (2014) was narrow [mean ± s.e. fork ⋅ ⋅ ⋅ ⋅ length (LF) = 0 29 ± 0 00 m, mean ± s.e. body mass = 262 20 ± 0 03 g] and tail-beat kinematics are known to vary with body size (Bainbridge, 1958; Kayan et al., 1978; Webb et al., 1984; Videler & Wardle, 1991; Eloy, 2012). It is therefore possible that fishes of different sizes will operate at slightly different St numbers and perhaps modulate kinematic variables differently in order to maintain their favoured St (van Weerden et al., 2014). Swimming is energetically demanding and the whole-body oxygen requirement needed to perform sustained swimming will change in response to acute changes in T and water speed. In fishes, the oxygen demand of the tissues, and therefore metabolic ˙ rate (MR), can be estimated by measuring the rate of oxygen consumption (MO2) using a swim flume respirometer (Clark et al., 2013). Cech & Brauner (2011) define five categories of aerobic metabolism in fishes: standard, resting-routine, routine, ˙ swimming and active. MO2 measurements during routine metabolism (i.e. routine metabolic rate, RMR) and during swimming metabolism (swimming metabolic rate, swMR) may be more typical of those experienced by a fish in the wild than the ˙ commonly measured MO2 minimum and maximum, known as standard metabolic rate (SMR) and maximum metabolic rate (MMR), respectively. It is unlikely that a fish would swim at its maximum aerobic capacity for prolonged periods of time and would rarely (if at all) experience true minimum MRs in the wild. Here, swMR is defined as the rate of oxygen consumption under maximum steady-state voluntary swimming conditions (i.e. when an electric grid is not used to motivate the fish to swim). RMR is defined as the rate of oxygen consumption in a post-absorptive fish in aflumeatrest (i.e. where active swimming was not required to maintain position in the flume). The difference between RMR and swMR (defined here as the metabolic scope) could more accurately represent the capacity for increasing aerobic energy expenditure during prolonged swimming events than aerobic scope (the difference between SMR and MMR), a variable often measured in swimming performance studies. Metabolic scope will vary with temperature and there may be a point on the thermal gradient where it is largest (an optimum temperature, or Topt) similar to the previously

defined and much discussed optimal temperature for aerobicT scope( optAS) where energy for growth, migration and reproduction is maximal (Farrell, 2009; Khan et al., 2014). The interactions among temperature, aerobic scope and aspects of fish fitness are complex. For example, ToptAS for a pink salmon Oncorhynchus gorbuscha (Walbaum 1792) occurs close to the upper incipient lethal temperature of the species (Clark et al., 2011) so a fish is expected seek out a temperature below ToptAS because other important physiological and biochemical functions are preferentially optimized at lower temper- ≠ atures (i.e. Topt ToptAS). Similarly, recent work on Atlantic halibut Hippoglossus hippoglossus (L. 1758) has shown that ToptAS does not correlate with optimal temperature for growth, but does correlate with optimal cardiac performance (Gräns et al., 2014). This study set out to determine whether acute warming of water T affects the mechanical efficiency of swimming (defined by St) and the aerobic capabilities of S. trutta. The first hypothesis was that S. trutta would strive to maintain a similar St value across all temperatures and this value would lay within the predicted optimal range of 0⋅20–0⋅40. The second hypothesis was that St would decrease with increasing fish size because of the relationship between body size and tail-beat kinematics (van Weerden et al., 2014).

MATERIALS AND METHODS

ETHICS STATEMENT All experimental procedures were covered by a U.K. Home Office project licence (40/3584) held by H.A.S. and were under the ethical approval of the University of Manchester.

EXPERIMENTAL ANIMALS

Ten juvenile S. trutta (body mass range = 36–129 g, mean wet body mass = 77 g, LF range = 15–23 cm, mean LF = 19 cm) were purchased from Chirk Trout Farm (www. chirktroutfarm.co.uk) and transported to the University of Manchester, U.K. Fish were housed in 500 l recirculating freshwater tanks for at least 10 weeks prior to experimental use. Water T in the holding tanks was 11∘ C, range ± 1∘ C, and fish experienced a 12L:12D photoperiod. Ammonia, nitrate and pH levels were monitored and tanks underwent a 30% water change three times a week, at which time fish were fed to satiation using commercial trout pellets. Fish were fasted for at least 36 h prior to experimentation to ensure they were post-absorptive. Individual fish were identified by LF and wet body mass. All husbandry and housing conditions were in accordance with the local handling protocols and adhered to the U.K. Home Office Legislation.

RESPIROMETRY AND KINEMATICS Swim-tunnel intermittent flow-through respirometry was used to record kinematics and ˙ measure MO2 of fish during T and swimming challenges. Fish were placed individually into a custom-built Brett-style swim-tunnel respirometer (70 l) and left for a minimum of 12 h overnight, receiving aerated water (>90% air saturation) at 11∘ C, range ± 1∘ C, circulated at ⋅ −1 ⋅ ⋅ −1 a fixed speed of 0 25 m s (1 09–1 67 LF s ). This speed allowed sufficient mixing of the water but did not require fish to swim actively to maintain position. The section of the tunnel holding the fish (working section) was covered with soft black plastic for the duration ofthe experiment to minimize stress. Plastic was lifted periodically to observe fish. Water speed was calibrated in m s−1 using a hand-held flow-meter (Höntzsch GmbH; www.hoentzsch.com/en). Prior to and after the experimental period, background respiration rate in the swim tunnel was measured over a 24 h period. Each time it was negligible.

To test the effect of T on RMR, T in the tunnel was increased stepwise at a rate of c.4∘ Ch−1 between 11, 14, 18 and 20∘ Candc.2∘ Ch−1 between 20, 22 and 24∘ C. Water speed remained fixed at⋅ 0 25 m s−1 during this time. Fish were held at each new T for 30 min prior to recording RMR. Dissolved oxygen concentration of the water was recorded in the respirometry chamber using a fibre optic oxygen sensor (model FIBOX 3 LCD, PreSens; www.presens.de) and a temperature-calibrated dipping probe interfaced with OxyView software (LCDPST3 1.16; http://oxyview.software.informer.com). The oxygen saturation of the water was taken every second for a measurement period of 20 min at each temperature, during which time the chamber was sealed. Upon completion of the temperature trial, water T was decreased over 60–90 min to 11∘ C and the fish was removed from the tunnel and placed in a separate holding tankat11∘ C. The same fish were used to examine the effect of T on swMR. There was at least 48 h between RMR and swMR measurements, during which time fish were not fed. The protocol for examining the effect of T on swMR was the same as that described above, except that at each T (after stabilization at the new temperature, which took ≥90 min), water oxygen saturation was recorded ⋅ −1 ⋅ −1 ⋅ −1 whilst water speed was increased from c.14 LF s (0 25 m s )toc.03 LF s every minute, until the early stages of burst and glide swimming were observed. At this point, water speed was reduced slightly so that the fish was swimming at the highest speed where only a steady swimming gait was observed. To encourage the fish to swim, the front end of the respirometry chamber was covered in black plastic and a halogen lamp was flashed at the rear. Thus, fish were able to choose to either continue or cease swimming. Each swimming test at each T lasted a maximum of 15 min, during which time the chamber was sealed and oxygen saturation of the water was recorded. swMR was taken as the steepest slope of oxygen saturation decline for any ⋅ −1 3–5 min period during steady-state swimming. Water speed was then reduced back to 1 4 LF s and fully aerated water was flushed through the swim tunnel. Water T was then increased to the next level over a duration of ≥90 min and the procedure was repeated. Oxygen saturation of the water did not fall below 90% at any point during the c. 15 min swim challenge. Tail-beat kinematics were captured at each T simultaneous to swMR measurements via aerial-view slow-motion video recordings at 100 frames s−1 (Sony HDR-SR8E; www.sony. co.uk). Video recordings were analysed using Tracker 4 video analysis software (Open Source Physics, Java framework; www.opensourcephysics.org) to calculate A and f tail. The video is viewed frame by frame and specific points of the fish are selected manually; distances are calibrated using a known length marked within the frame. The distances from the midline of the fish (running through the dorsal fin) to the furthest point that the tip of the tailreached on the left and right tail movements were added together to give A. Videos were taken when the fish was swimming in a steady gait in the centre of the working section at speeds ranging ⋅ ⋅ −1 ⋅ ⋅ −1 from 0 56 to 0 85 m s (c.25–4 5 LF s ), using water speed as a proxy for fish swimming speed. Swimming speeds were subsequently adjusted for any forward or backward movement of the fish (mean ± s.e. adjustment = 0⋅009 ± 0⋅001 m s−1). The distance between the tip of the head and dorsal fin was measured in each video; variation in this measurement in the samefish would indicate a difference in vertical position in the chamber (from floor to ceiling). Where this was found, A was adjusted by scaling to the smallest measurement taken where the fish is closest to the calibration marking on the floor of the chamber. As the cross-sectional area ofthe fish was never greater than 3% of that of the swimming chamber, no adjustments weremade for blocking effects (Jones et al., 1974). During both RMR and swMR protocols, fish were observed for signs of thermal stress. No fish experienced loss of equilibrium during the RMR protocol, one fish lost equilibrium atthe end of the swMR protocol and successfully recovered, and one fish lost equilibrium during the 23–24∘ CswMRT increase and did not survive.

DATA ANALYSES AND STATISTICS Linear regression lines were plotted for water oxygen saturation against time and the gra- ˙ −1 2 ⋅ ⋅ dient was used to calculate MO2 in mg O2 min (mean ± s.e. r for RMR was 0 82 ± 0 03 and was 0⋅72 ± 0⋅02 for swMR). Due to the allometric relationship between fish size and oxygen consumption rates (whereby smaller fish consume relatively more oxygen than do larger ones), a mass exponent was calculated and used to adjust all respirometry data. This allometric relationship can be linearized with a log10 transformation: the log10 of RMR

and swMR were plotted against log10 body mass and a linear regression line fitted for both. The mean slope of these regression lines gives the mass exponent, provided that the regression lines do not significantly differ from one another, as was the case here [mean slope = 0⋅60, mean r2 = 0⋅40, P > 0⋅05; more detailed discussion and methods are given by ˙ Myrick & Cech (2000), Cech & Brauner (2011) and Lucas et al. (2014)]. Therefore, MO2 0⋅60 ˙ 0⋅60 −1 measurements were adjusted to body mass . MO2 (y)inmgO2 kg min was then calculated using the following equation from Clark et al. (2013), modified to include the mass exponent: y = [(V − V )𝛥C ](𝛥tM 0 ⋅ 6)− 1,whereV is the respirometer volume, V r f wo2 f r 𝛥 f is the fish volume (where 1 g of fish is equivalent to 1 ml of water), Cwo2 is the change in oxygen concentration in the respirometer water, 𝛥t is the change in time during which 𝛥 Cwo2 was measured and Mf is the mass of the fish in kg. On three occasions during swMR ˙ measurements fish refused to swim, achieving less than 25% increase in MO2 compared with that observed during RMR. These fish were excluded from further analyses (two fish at 18∘ C and one at 24∘ C). Statistical analyses were conducted in IBM SPSS Statistics 20 (http://www-01.ibm.com/software/uk/analytics/spss/) and means are displayed ±s.e. The effect of temperature on RMR was determined using a repeated-measures general linear model (GLM). swMR and metabolic scope were analysed using GLMs, with individual fish included as a random factor and subsequently removed if not significant. Significance was considered as P < 0⋅05. Kinematic measurements were successfully collected over a range of temperatures and speeds for seven fish; each fish was recorded at one or more speed per temperature increment giving varying sample sizes at each temperature (n = 10, 23, 26, 25, 17, 23 and 15 at 11, 14, 18, 20, 22, 23 and 24∘ C, respectively). Because of the possibility of size effects, results are presented in absolute units (m) and in units relative to fish body size; by dividing tail-beat amplitude and forward speed by fish LF (m) giving AREL and UREL, respectively. The effects of both absolute and relative speeds are discussed, as well as the difference between A and AREL. The effect of T and U or UREL on A, AREL, f tail and St was investigated using GLMs (six full models in total), with T included as a fixed factor with seven levels, speed as a covariate and individual fish as a random factor. The fish swam at a range of U across T treatments, but the number of and increments of U varied, rendering the data unsuitable for a full repeated-measures model. Interactions between all variables were included in the GLMs; if not statistically significant, each term was removed in order of least significance to give the final GLM. Linear regression analyses were used to investigate relationships between A, AREL, f tail, St, U, UREL,RMR,swMR and the metabolic scope, and fish size. To ensure that the assumption of normally distributed data was not violated, the standardized residuals from each statistical model were tested for normality using Kolmogorov–Smirnov tests.

RESULTS

KINEMATICS ⋅ ⋅ Almost all St values were in the range 0 2–0 3, with only 13 out of 139 falling below, ⋅ ⋅ and only two lying above. The St number was 0 23 ± 0 00 and it was not influenced by T (Table I); however, it showed a significant, but slight, decline with both U and U (Fig. 1). At any given U, A decreased with increasing T, being larger at 11 and REL∘ ∘ ∘ 14 C than at 18–24 C [Fig. 2(a)]. In contrast, f tail was higher at 18–24 C than at 11 and 14∘ C at any given U [Fig. 2(b)]. This implies a temperature-dependent trade-off between A and f tail whereby St is maintained. The same temperature effect is seen when A and U are adjusted for differences in body size. When AREL [peak-to-peak stroke amplitude of the tail tip (m) relative to LF of fish] is plotted against UREL [forward speed (m s−1) relative to L of fish], there is less variation around the lines of bestfit F ∘ [Fig. 2(c)]. Interestingly, f tail increased with UREL at all temperatures except at 11 C [Fig. 2(d)], where there is a slight decline in f tail as UREL increased; this is denoted by

Table I. Statistical output from the general linear models (GLM) for the three kinematic variables (expressed as absolute and relative measures), Strouhal number (St) and metabolic rate variables measured from Salmo trutta at 11, 14, 18, 20, 22, 23 and 24∘ C, over a range of steady-state swimming speeds. Removed terms are shown in order of removal y Removed terms Final GLM

⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ > ⋅ St T × fish, F30,83 = 1 02, p = 0 27, T, F6,125 = 0 30, p = 0 01, P 0 05 P > 0⋅05 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ < ⋅ T × U, F6,113 = 1 27, p = 0 06, U, F1,125 = 4 06, p = 0 03, P 0 05 P > 0⋅05 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ U × fish, F6,119 = 1 06, p = 0 05, Fish, F6,125 = 5 00, p = 0 19, P > 0⋅05 P < 0⋅001 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ > ⋅ St T × fish, F30,83 = 0 99, p = 0 26, T, F6,125 = 0 29, p = 0 01, P 0 05 P > 0⋅05 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ UREL × fish, F6,113 = 1 30, p = 0 06, UREL, F1,125 = 4 33, p = 0 03, P > 0⋅05 P < 0⋅05 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ T × UREL, F6,119 = 1 16, p = 0 06, Fish, F6,125 = 4 33, p = 0 17, P > 0⋅05 P = 0⋅001 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ < ⋅ AT× fish, F30,83 = 0 77, p = 0 22, T, F6,125 = 6 43, p = 0 24, P 0 001 P > 0⋅05 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ T × U, F6,113 = 0 69, p = 0 04, U, F1,125 = 16 25, p = 0 12, P > 0⋅05 P < 0⋅001 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ U × fish, F6,119 = 1 77, p = 0 08, Fish, F6,125 = 10 72, p = 0 34, P > 0⋅05 P < 0⋅001 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ < ⋅ AREL T × fish, F30,83 = 0 78, p = 0 22, T, F6,125 = 6 44, p = 0 24, P 0 001 P > 0⋅05 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ T × UREL, F6,113 = 0 87, p = 0 04, UREL, F1,125 = 15 39, p = 0 11, P > 0⋅05 P < 0⋅001 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ UREL × fish, F6,119 = 1 40, p = 0 07, Fish, F6,125 = 8 92, p = 0 30, P > 0⋅05 P < 0⋅001 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ ⋅ f tail T × fish, F30,83 = 0 88, p = 0 24, T, F6,125 = 3 90, p = 0 16, P = 0 001 P > 0⋅05 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ U × fish, F6,113 = 2 01, p = 0 10, U, F1,125 = 45 90, p = 0 27, P > 0⋅05 P < 0⋅001 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ T × U, F6,119 = 1 59, p = 0 07, Fish, F6,125 = 3 25, p = 0 14, P > 0⋅05 P < 0⋅01 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ < ⋅ f tail T × fish, F30,83 = 0 75, p = 0 21, T, F6,119 = 2 87, p = 0 13, P 0 05 P > 0⋅05 ⋅ 𝜂 2 ⋅ ⋅ 𝜂 2 ⋅ U × fish, F6,113 = 1 44, p = 0 07, UREL, F1,119 = 34 36, p = 0 22, P > 0⋅05 P < 0⋅001 ⋅ 𝜂 2 ⋅ Fish, F6,119 = 6 39, p = 0 24, P < 0⋅001 ⋅ 𝜂 2 ⋅ T × UREL, F6,119 = 3 19, p = 0 14, P < 0⋅01 ⋅ 𝜂 2 ⋅ < ⋅ RMR T, F5,45 = 13 00, p = 0 59, P 0 001 ⋅ 𝜂 2 ⋅ ⋅ swMR T, F5,42 = 5 05, p = 0 38, P = 0 001 ⋅ 𝜂 2 ⋅ < ⋅ Fish, F9,42 = 3 29, p = 0 41, P 0 01 ⋅ 𝜂 2 ⋅ > ⋅ ⋅ 𝜂 2 ⋅ < ⋅ Metabolic scope Fish, F9,42 = 1 18, p = 0 20, P 0 05 T, F5,51 = 7 90, p = 0 44, P 0 001

St, Strouhal number; A, peak to peak stroke amplitude of tail tip (m); AREL, peak to peak stroke amplitude of tail tip (m) 0⋅60 −1 relative to fork length of fish (LF); f tail, tail-beat frequency; RMR, routine metabolic rate (mg O2 kg min ); swMR, 0⋅60 −1 ∘ −1 maximum metabolic rate while swimming (mg O2 kg min ); T, temperature ( C); U, forward speed (m s ); UREL, −1 forward speed (m s ) relative to LF.

(a) 0·35 (b)

0·30 t

S 0·25

0·20

0·15 0·5 0·6 0·7 0·8 0·9 2·0 2·5 3·0 3·5 4·04·5 5·0 –1 U (m s–1) UREL (LF s )

Fig. 1. Strouhal number (St) plotted against (a) absolute swimming speed (U) and (b) relative swimming speed (UREL)forSalmo trutta. The lines of best fit calculated from the general linear model (GLM) outputs are (a) y =−0⋅057x + 0⋅259 (r2 = 0⋅06) and (b) y =−0⋅012x + 0⋅254 (r 2 = 0⋅32). Statistical analyses are given in Table I.

the significant interaction between T and UREL (Table I). In all other instances, each kinematic variable increased with increasing U or UREL (Fig. 2). All the kinematic variables (St, A, AREL and f tail) differed between individual fish (Table I). 2 ⋅ ⋅ There was no relationship between f tail or St and fish sizer ( = 0 002, F1,137 = 0 26, > ⋅ 2 ⋅ ⋅ > ⋅ P 0 05 and r = 0 008, F1,137 = 1 09, P 0 05), however as expected, A increased 2 ⋅ ⋅ < ⋅ with increasing LF (r = 0 148, F1,137 = 23 78, P 0 001). Moreover, larger fish had 2 ⋅ ⋅ < ⋅ smaller AREL [r = 0 288, F1,137 = 55 37, P 0 001; Fig. 3(a)], which is associated 2 ⋅ ⋅ < ⋅ with a decline in UREL as fish size increased [r = 0 349, F1,137 = 73 33, P 0 001; Fig. 3(b)].

RESPIROMETRY 0⋅60 −1 RMR (mg O2 kg min ) increased with increasing T [Fig. 4(a) and Table I], ⋅ ⋅ 0⋅60 −1 ∘ and had a Q10 value of 2 83 ± 0 36. SwMR (mg O2 kg min ) was higher at 14 C (4⋅20 ± 0⋅14) and 18∘ C(4⋅48 ± 0⋅17) than at all other temperatures [Fig. 4(b)], and differed with individual fish (Table I). The available metabolic scope (difference between RMR and swMR) was highest between 11 and 18∘ C, declined between 20 and 22∘ C and was lowest at 24∘ C [Fig. 4(c)]. There was no relationship between fish 2 ⋅ ⋅ size and either RMR, swMR or the available metabolic scope (r = 0 002, F1,59 = 0 14, > ⋅ 2 ⋅ ⋅ > ⋅ 2 ⋅ ⋅ < ⋅ P 0 05, r = 0 02, F1,55 = 0 928, P 0 05 and r = 0 001, F1,55 = 0 07, P 0 05).

DISCUSSION

The key findings of this study were firstly that S. trutta maintained the same St at any given U over a range of temperatures, and these values fell between 0⋅2 and 0⋅3. The maintenance of St within the range for maximum efficiency for oscillatory propulsion (Triantafyllou et al., 1993; Taylor et al., 2003) was achieved through an increase in f tail and a decrease in A as T increased (Fig. 2). Second, the energetic demands of swimming at increasing temperatures were met by an increase in RMR and swMR up to 18∘ C meaning the available metabolic scope remained constant (Fig. 4); however, at warmer temperatures, swMR fell and thus, the metabolic scope was reduced. Third,

(a) 0·09 (c) 0·50 0·45 0·08 0·40 ) ) 0·07 0·35 m m ( ( A A 0·06 0·30 0·25 0·05 0·20 0·04 0·15 0·5 0·6 0·7 0·8 0·9 2·0 2·5 3·0 3·5 4·0 4·5 5·0

(b) 4·0 (d) 4·0

3·5 3·5

3·0 3·0 (Hz) (Hz) f 2·5 f 2·5

2·0 2·0

1·5 1·5 0·5 0·6 0·7 0·8 0·9 2·0 2·5 3·0 3·5 4·04·5 5·0 –1 U (m s–1) UREL (LF s )

Fig. 2. The effect of swimming speed on tail-beat kinematics for Salmo trutta. (a, b) The effect of absolute swim- ∘ ming speed (U) on (a) tail-beat amplitude (A) and (b) tail-beat frequency (f tail) at 11 and 14 C( and ) and at 18, 20, 22, 23 and 24∘ C( and ). The lines of best fit calculated from the general linear model (GLM) outputs are (a) blue: y = 0⋅023x + 0⋅050 (r2 = 0⋅01); red: y = 0⋅023x + 0⋅043 (r2 = 0⋅06) and (b) blue: y = 2⋅01x + 0⋅912 (r2 = 0⋅09); red: y = 2⋅01x + 1⋅144 (r2 = 0⋅27). (c, d) The effect of relative ∘ swimming speed (UREL) on (c) relative tail-beat amplitude (AREL) at 11 and 14 C( and )andat ∘ ∘ ∘ 18, 20, 22, 23 and 24 C( and ) and (d) tail-beat frequency (f tail)at11 C( and ); 14 C ( and ) and at 18, 20, 22, 23 and 24∘ C( and ). The lines of best fit calculated from the GLM outputs are (c) dark blue: y = 0⋅023x + 0⋅270 (r2 = 0⋅23); red: y = 0⋅023x + 0⋅239 (r2 = 0⋅23) and (d) purple: y =−0⋅215x + 3⋅019 (r2 = 0⋅26); light blue: y = 0⋅450x + 0⋅704 (r 2 = 0⋅01); red: y = 0⋅450x + 0⋅912 (r 2 = 0⋅20). Statistical analyses are given in Table I.

no effect of fish size was detected on St. Hence, only one of the original hypotheses that related to maintenance of St across temperature treatments was fully supported. ⋅ ⋅ Of the St numbers measured, across all conditions, 89% fell between 0 2 and 0 3 with a mean value similar to that previously found for O. mykiss (Nudds et al., 2014) and for other swimming animals (Triantafyllou et al., 1993), including cetaceans (Rohr & Fish, 2004). The finding that St decreases with U is also congruent with some previous studies (Hunter & Zweifel, 1971; Webb, 1971a). In contrast, St did not vary with U in cetaceans (Rohr & Fish, 2004) and increased slightly with U in O. mykiss (Nudds et al., 2014). The disagreement may be due to the range of U that the studies are conducted over. Perhaps, St decreases initially at low speeds, is constant at intermediate speeds and increases at high speeds. Therefore, although speculative, the true relationship between St and U for fishes may be a curve, similar to that seen across Reynolds numbers (van Weerden et al., 2014). An increase in f tail with increasing U is consistent with most previous studies on fishes (Bainbridge, 1958; Brill & Dizon, 1979; Hunter & Zweifel, 1971;a Webb,1971 ;

(a) 0·55

0·45

REL 0·35 A

0·25

0·15

(b) 5·5

4·5

REL 3·5 U

2·5

1·5 0·14 0·16 0·18 0·20 0·22 0·24

LF (m)

Fig. 3. The relationship between fork length (LF)ofSalmo trutta and (a) relative tail-beat amplitude (AREL) across a range of temperatures and swimming speeds and (b) relative swimming speed (UREL) across a range of temperatures. AREL was sampled repeatedly at varying UREL, giving multiple data points for each fish at ⋅ ⋅ 2 ⋅ different LF. The lines of best fit from the linear regression outputs are(a) y =−1 004x + 0 523 (r = 0 29) and (b) y =−13⋅96x + 6⋅263 (r2 = 0⋅35). Statistical analyses are given in Table I.

Stevens; 1979; Rome et al., 1984, 1990; Nudds et al., 2014). Previous study of Nudds et al. (2014) on O. mykiss, using the same swim flume as here, however, found that f tail was invariant with U and there is no obvious explanation for this. The warmer T used here may promote a faster f tail [Fig. 2(b)], and perhaps faster swimming speeds (Rome et al., 1984). Moreover, it is well established that locomotory muscle is highly T sensitive, with the rate of force generation and maximum shortening speed increasing with increasing T, which increases the speed of movement and force produced, respectively (James, 2013). In yellowfin tuna Thunnus albacares (Bonnaterre 1788), warming has been clearly shown to increase slow-twitch muscle power output and the frequency at which maximal power output can be maintained (Altringham & Block, 1997). If S. trutta can achieve faster and more powerful tail-beats at warmer T [shown by an increase in f tail, Fig. 2(b)], then it is possible that the decrease in A is a passive response. When forced to swim at increasing speed at colder T, f tail may be limited and therefore A is increased in order to achieve mechanically efficient propulsion [Fig. 2(a); Altring- ham & Block, 1997]. The fact that f tail rather than A appears to be increased when water

(a) 2·5 b c b

) b

−1 2·0 b in m

0·60 1·5 g k 2

O 1·0 a mg

0·5 RMR (

0·0 (b) 5·0

) b

−1 b

in 4·5 m

0·60 g k 2 4·0 a a O a ·

mg a 3·5 MR ( sw

3·0 (c) 4·0

3·5

). a a a −1

in 3·0 m cope

0·60 2·5 g

k bb 2

O 2·0 Metabolic mg

( c 1·5

1·0 10 15 20 25 Temperature (° C)

Fig. 4. The effect of temperature on Salmo trutta (a) routine metabolic rate (RMR), (b) maximum recorded MR whilst swimming (swMR) and (c) available metabolic scope under these conditions. Significance was determined by one-way ANOVA and shown between values by different lower-case letters (P < 0⋅05). Values are mean ± s.e.

T permits suggests that a higher f tail to A ratio may give a higher energetic efficiency (conversion of chemical energy to mechanical energy ratio). Work-loop studies with isolated slow-twitch muscle from S. trutta would help clarify these relationships. The direct relationship between muscle contraction (in vivo and in vitro) and whole animal performance, however, is difficult to demonstrate (James, 2013). It is worth noting, however, that although an increasing f tail with T was seen in O. mykiss swimming at

⋅ ⋅ −1 1 0–3 7 body lengths (Lb)s (Nudds et al., 2014) and in largemouth bass Micropterus ⋅ ⋅ −1 salmoides (Lacépède 1802) swimming at 1 0–2 0 LF s (Stevens, 1979), a decreasing ⋅ ⋅ −1 f tail with T has also been found in O. mykiss swimming at 1 0–3 0 LF s (Stevens, 1979) and T had no effect on the f tail of common carp Cyprinus carpio L. 1758 swim- ⋅ ⋅ −1 ming at 1 07–4 09 Lb s (Rome et al., 1990). Thus, responses are not consistent across studies. Here, S. trutta decrease A at warmer T, which agrees with previous work on O. mykiss (Nudds et al., 2014), yet is contrary to what was found for C. carpio (Rome et al., 1990), where no temperature effect on A was evident. Therefore, again the effect of changes in T across U is not necessarily consistent across studies. With regard to the effect of fish size, AREL was smaller in larger fish, which supports the findings of Webb et al. (1984). There was no change in f tail with fish size, but the maximum UREL attained during the swim challenge was lower in larger fish. Videler & Wardle (1991) suggest that AREL is more variable at low speeds; similarly, Webb (1971a) found that AREL reaches maximal values when f tail approaches 5 Hz, which exceeds f tail values seen in this study. The reduction in UREL and AREL for any given flume speed in larger fish seen here results in the maintenance of St in the range for optimal mechanical efficiency. ˙ To understand how temperature and swimming affect the energetics of S. trutta, MO2 was recorded under routine conditions (RMR) and under steady-state swimming conditions (swMR) as temperature and water speed were increased. The terms RMR and swMR were chosen over the more commonly used terms of SMR and MMR as these are more reflective of the swim flume conditions used in this study and are perhaps more relevant to fishes in the wild. Moreover, Clark et al. (2013) have requested that more rigour be applied to the use of the terms SMR and MMR and their difference (aerobic scope) to facilitate greater comparison between studies. The values presented ⋅ 0⋅60 −1 ∘ ⋅ 0⋅60 −1 here for RMR (RMR = 0 67 mg O2 kg min at 11 C and 1 5mgO2 kg min at 14∘ C) are in the same range as those previously reported in salmonids as SMR, ⋅ ⋅ −1 −1 ∘ SMR ≈ 0 1–2 4mgO2 kg min at 10–15 C (Webb, 1971b; Burgetz et al., 1998; Clark et al., 2011; Norin & Malte, 2011; Eliason et al., 2013a; Nudds et al., 2014). ⋅ The Q10 value of 2 83 for RMR is also comparable with previous studies that range from c.2⋅5to3⋅3 (Heath & Hughes, 1973; Altimiras et al., 2002; Clark et al., 2011). ⋅ 0⋅60 −1 ∘ swMR values presented (swMR = 4 35 mg O2 kg min at 18 C) are at the lower end of a rather large MMR range previously reported in salmonids −1 −1 (MMR ≈ 6–20mgO2 kg min at ToptAS) (Altimiras et al., 2002; Clark et al., 2011; Eliason et al., 2013a). The upper end of this range is elevated by wild semelparous salmonid species ready to undertake their upstream spawning migration. The fish in this study were hatchery-reared juveniles that were not exercise trained, and many studies suggest that such animals have lower MR and swim performance (Araki et al., 2008; Van Leeuwen et al., 2011). Indeed, these swMR values are more comparable ⋅ −1 −1 with active MR (RACT,643 O2 kg min ; Altimiras et al., 2002), measured in farmed S. trutta than wild fish. It is also likely that fish in this study were not swimming attheir maximum aerobic swimming speed, as they were not forced to swim. As the swMR measurements were low while RMR was comparable with previously reported, the ⋅ ⋅ 0⋅60 −1 difference between the two (c.15–2 7mgO2 kg min , which has been equated to the available metabolic scope under the conditions of this study) is slightly lower than that previously reported for salmonids over a comparable temperature range ⋅ ⋅ −1 −1 (5 1–16 2mgO2 kg min ) (Altimiras et al., 2002; Clark et al., 2011; Eliason et al., 2013a). Additionally, because RMR was not measured during the incremental

© 2015 The Fisheries Society of the British Isles, Journal of Fish Biology 2016, 88, 403–417 414 J. M. D. LEA ET AL. increases in water temperature at low speeds, the possibility of fatigue contributing to the low swMR values at high temperatures cannot be excluded. No fish, however, were included that refused to swim or which were not able to swim at a steady gait in the analysis. Moreover, the rest period between swimming challenges (≤90 min) is longer than recovery periods commonly used in repeat swim challenges (Farrell et al., 1998, 2003; Jain et al., 1998; MacNutt et al., 2004; Eliason et al., 2013b), where significant ˙ differences in MO2 or critical swimming speed (Ucrit) between swimming challenges were not found. Available metabolic scope was highest between 11 and 18∘ C, declining by half at ∘ 24 C [Fig. (4c)]. This range encompasses previous findings where the Topt for scope for ⋅ ∘ activity (comparable to ToptAS)forS. trutta was 17 8 C (Elliott, 1976), and is slightly higher than the range of preferred T (9⋅5–16⋅3∘ C) reported by Elliott & Allonby (2013). These fish commonly migrate between July and December (Jonsson &Jon- sson, 2011) when river temperatures rarely exceed 18∘ C, with many remaining around 14∘ C (Hammond & Pryce, 2007). Therefore, the T range where available metabolic scope was found to be highest in this study reflects the range of T that are likely to be experienced by S. trutta in the wild. In conclusion, and as hypothesized, when their kinematics are disrupted by a change in water T, S. trutta appear to adhere to the same St at any given U as seen in O. mykiss (Nudds et al., 2014). No effect of fish size on this relationship was found. Thus, the data here support the idea that maintaining kinematics at a single optimum (or preferred) St is desirable in fish ranging from 36 to 129 g. The effect ofboth U and T upon tail kinematics, however, is not consistent across all studies of fishes. Whether the differences are species specific, methodological or a product of individual fish motivation is not clear. Maintenance of mechanically efficient steady-state swimming in S. trutta is fuelled by aerobic metabolism, which increases as temperature increases up to 18∘ C. Above this temperature swMR falls, decreasing the apparent metabolic scope for activity. As St was maintained over the full range of temperatures tested while metabolic scope was not, the results may suggest energetic trade-offs at any given U at temperatures above ToptAS. Thus, S. trutta will be able to maintain efficient locomotion during aquatic warming, despite declines in aerobic capacity. A more detailed study of a broad range of fish species performed under the same conditions and with thesame equipment is required to better understand these relationships and remains a priority for future work.

This work was supported in part by a Fisheries Society of the British Isles summer internship awarded to J.M.D.L. and a University of Manchester Invest in Success grant to H.A.S. The authors would like to thank S. Chowdury and S. Lambert for their help with data collection, and T. D. Clark and R. Antwis, and anonymous reviewers for valuable comments on earlier drafts. The swim-flume respirometer was a generous gift from P. Butler at the University of Birmingham.

References Almodóvar, A., Nicola, G. G., Ayllón, D. & Elvira, B. (2012). Global warming threatens the persistence of Mediterranean brown trout. Global Change Biology 18, 1549–1560. doi: 10.1111/j.1365-2486.2011.02608 Altimiras, J., Axelsson, M., Claireaux, G., Lefrancois, C., Mercier, C. & Farrell, A. P. (2002). Cardiorespiratory status of triploid brown trout during swimming at two acclimation temperatures. Journal of Fish Biology 60, 102–116. doi: 10.1111/j.1095-8649.2002.tb02390

Altringham, J. D. & Block, B. A. (1997). Why do tuna maintain elevated slow muscle temperatures? Power output of muscle isolated from endothermic and ectothermic fish. Journal of Experimental Biology 200, 2617–2627. Anttila, K., Couturier, C. S., Overli, O., Johnsen, A., Marthinsen, G., Nilsson, G. E. & Farrell, A. P. (2014). Atlantic salmon show capability for cardiac acclimation to warm temperatures. Nature Communications 5, 5252. doi: 10.1038/ncomms5252 Araki, H., Berejikian, B. A., Ford, M. J. & Blouin, M. S. (2008). Fitness of hatchery-reared salmonids in the wild. Evolutionary Applications 1, 342–355. doi: 10.1111/j.1752- 4571.2008.00026 Bainbridge, R. (1958). The speed of swimming of fish as related to size and to the frequency and amplitude of the tail beat. Journal of Experimental Biology 35, 109–135. Beamish, F. W. H. (1978). Swimming capacity. In Fish Physiology, Vol. 7 (Hoar, W. S. & Ran- dall, D. J., eds), pp. 101–187. New York, NY: Academic Press Inc. Brill, R. W. & Dizon, A. E. (1979). Red and white muscle fibre activity in swimming skipjack tuna, Katsuwonus pelamis (L.). Journal of Fish Biology 15, 679–685. doi: 10.1111/j.1095-8649.1979.tb03677 Burgetz, I. J., Rojas-Vargas, A., Hinch, S. G. & Randall, D. J. (1998). Initial recruitment of anaerobic metabolism during sub-maximal swimming in rainbow trout (Oncorhynchus mykiss). Journal of Experimental Biology 201, 2711–2721. Cech, J. J. Jr. & Brauner, C. J. (2011). Techniques in whole animal respiratory physiology. In Encyclopedia of Fish Physiology: From Genome to Environment (Farrell, A. P., Cech, J. J. Jr., Richards, J. G. & Stevens, E. D., eds), pp. 846–853. New York, NY: Academic Press. Clark, T. D., Jeffries, K. M., Hinch, S. G. & Farrell, A. P. (2011). Exceptional aerobic scope and cardiovascular performance of pink salmon (Oncorhynchus gorbuscha) may underlie resilience in a warming climate. Journal of Experimental Biology 214, 3074–3081. doi: 10.1242/jeb.060517 Clark, T. D., Sandblom, E. & Jutfelt, F. (2013). Aerobic scope measurements of fishes in an era of climate change: respirometry, relevance and recommendations. Journal of Experimental Biology 216, 2771–2782. doi: 10.1242/jeb.084251 Eliason, E. J., Wilson, S. M., Farrell, A. P., Cooke, S. J. & Hinch, S. G. (2013a). Low cardiac and aerobic scope in a costal population of sockeye salmon (Oncorhynchus nerka) with a short upriver migration. Journal of Fish Biology 82, 2104–2112. doi: 10.1111/jfb.12120 Eliason, E. L., Clark, T. D., Hinch, S. G. & Farrell, A. P. (2013b). Cardiorespiratory performance and blood chemistry during swimming and recovery in three populations of elite swim- mers: adult sockeye salmon. Comparative Biochemistry and Physiology A 166, 385–395. doi: 10.1016/j.cbpa.2013.07.020 Elliott, J. M. (1976). The energetic of feeding, metabolism and growth of brown trout (Salmo trutta L.) in relation to body weight, water temperature and ration size. Journal of Animal Ecology 45, 923–948. Elliott, J. M. & Allonby, J. D. (2013). An experimental study of ontogenetic and seasonal changes in the temperature preferences of unfed and fed brown trout, Salmo trutta. Fresh- water Biology 58, 1840–1848. doi: 10.1111/fwb.12173 Elliott, J. M. & Elliott, J. A. (2010). Temperature requirements of Atlantic salmon Salmo salar, brown trout Salmo trutta and Arctic charr Salvenius alpinus:predict- ing the effects of climate change. Journal of Fish Biology 77, 1793–1817. doi: 10.1111/j.1095-8649.2010.02762 Eloy, C. (2012). Optimal Strouhal number for swimming animals. Journal of Fluids and Struc- tures 30, 205–218. doi: 10.1016/j.jfluidstructs.2012.02.008 Farrell, A. P. (2009). Environment, antecedents and climate change: lessons from the study of temperature physiology and river migration of salmonids. Journal of Experimental Biol- ogy 212, 3771–3780. doi: 10.1242/jeb.023671 Farrell, A. P., Gamperl, A. K. & Birtwell, I. K. (1998). Prolonged swimming, recovery and repeat swimming performance of mature sockeye salmon (Oncorhynchus nerka) exposed to moderate hypoxia and pentachlorophenol. Journal of Experimental Biology 201, 2183–2193. Farrell, A. P., Lee, C. G., Tierney, K., Hodaly, A., Clutterham, S., Healey, M., Hinch, S. & Lotto, A. (2003). Field-based measurements of oxygen uptake and swimming performance with

adult Pacific salmon using a mobile respirometer swim tunnel. Journal of Fish Biology 62, 64–84. doi: 10.1046/j.1095-8649.2003.00010 Gräns, A., Jutfelt, F., Sandblom, E., Jönsson, E., Wiklander, K., Seth, H., Olsson, C., Dupont, S., Ortega-Martinez, O., Einarsdottír, I., Björnsson, D. T., Sundell, K. & Axelsson, M. (2014). Aerobic scope fails to explain the detrimental effects on growth resulting from warming and elevated CO2 in Atlantic halibut. Journal of Experimental Biology 217, 711–717. doi: 10.1242/jeb.096743 Hammond, D. & Pryce, A. R. (2007). Climate Change Impacts and Water Temperature.Bristol: Environment Agency. Heath, A. G. & Hughes, G. M. (1973). Cardiovascular and respiratory changes during heat stress in rainbow trout (Salmo gairdneri). Journal of Experimental Biology 59, 323–338. Hunter, J. R. & Zweifel, J. R. (1971). Swimming speed, tail beat frequency, tail beat amplitude, and size in jack mackerel, Trachurus symmetricus, and other fishes. Fisheries Bulletin NOAA 69, 253–266. Jain, K. E., Britwell, I. K. & Farrell, A. P. (1998). Repeat swimming performance of mature sockeye salmon following a brief recovery period: a sensitive measure of fish health and water quality. Canadian Journal of Zoology 76, 1488–1496. doi: 10.1139/z98-079 James, R. S. (2013). A review of the thermal sensitivity of the mechanics of vertebrate skeletal muscle. Journal of Comparative Physiology B 183, 723–733. doi: 10.1007/s00360-013-0748-1 Jones, D. R., Kiceniuk, J. W. & Bamford, O. S. (1974). Evaluation of the swimming performance of several fish species from the Mackenzie River. Journal of the Fisheries Research Board of Canada 31, 1641–1647. doi: 10.1139/f74-206 Jonsson, B. & Jonsson, N. (2011). Ecology of Atlantic salmon and brown trout: habitat as a template for life histories. In Fish and Fisheries Series 33 (Noakes, D. L. G., ed.), pp. 247–271. Dordrecht: Springer Science + Business Media. Kayan, V.P., Kozlov, L. F. & Pyatetskii, V.E. (1978). Kinematic characteristics of the swimming of certain aquatic animals. Fluid Dynamics 13, 641–646. doi: 10.1007%2FBF01050223 Khan, J. R., Pether, S., Bruce, M., Walker, S. P. & Herbert, N. A. (2014). Optimum temperatures for growth and feed conversion in cultured hapuku (Polyprion oxygeneios)–is there a link to aerobic metabolic scope and final temperature preference? Aquaculture 430, 107–113. doi: 10.1016/j.aquaculture.2014.03.046 Lucas, J., Schouman, A., Lyphout, L., Cousins, X. & Lefrancois, C. (2014). Allometric relationship between body mass and aerobic metabolism in zebrafishDanio ( rerio). Journal of Fish Biology 84, 1171–1178. doi: 10.1111/jfb.12306 MacNutt, M. J., Hinch, S. G., Farrell, A. P. & Topp, S. (2004). The effect of temperature and acclimation period on repeat swimming performance in cutthroat trout. Journal of Fish Biology 65, 342–353. doi: 10.1111/j.0022-1112.2004.00453 Myrick, C. A. & Cech, J. J. Jr. (2000). Temperature influences on California rainbow trout physiological performance. Fish Physiology and Biochemistry 22, 245–254. doi: 10.1023/A:1007805322097 Norin, T. & Malte, H. (2011). Repeatability of standard metabolic rate, active metabolic rate and aerobic scope in young brown trout during a period of moderate food availability. Journal of Experimental Biology 214, 1668–1675. doi: 10.1242/jeb.054205 Nudds, R. L., Taylor, G. K. & Thomas, A. L. R. (2004). Tuning of Strouhal number for high propulsive efficiency accurately predicts how wingbeat frequency and stroke amplitude relate and scale with size and flight speed in birds. Proceedings of the Royal Society B 271, 2071–2076. doi: 10.1098/rspb.2004.2838 Nudds, R. L., John, E. L., Keen, A. N. & Shiels, H. A. (2014). Rainbow trout provide the first experimental evidence for adherence to a distinct Strouhal number during animal oscillatory propulsion. Journal of Experimental Biology 217, 2244–2249. doi: 10.1242/jeb.102236 Peake, S. J. & Farrell, A. P. (2004). Locomotory behaviour and post-exercise physiology in relation to swimming speed, gait transition and metabolism in free-swimming smallmouth bass (Micropterus dolomieu). Journal of Experimental Biology 207, 1563–1575. doi: 10.1242/jeb.00927 Rohr, J. J. & Fish, F. E. (2004). Strouhal numbers and optimization of swimming by odontocete cetaceans. Journal of Experimental Biology 207, 1633–1642. doi: 10.1242/jeb.00948

Rome, L. C., Loughna, P. T. & Goldspink, G. (1984). Muscle fiber activity in carp as a function of swimming speed and muscle temperature. American Journal of Physiology 247, R272–R279. Rome, L. C., Funke, R. P. & Alexander, R. M. (1990). The influence of temperature on muscle velocity and sustained performance in swimming carp. Journal of Experimental Biology 154, 163–178. doi: 10.1.1.319.2664 Stevens, E. D. (1979). Effect of temperature on tail beat frequency of fish swimming at constant velocity. Canadian Journal of Zoology 57, 1628–1635. doi: 10.1139/z79-214 Taylor, G. K., Nudds, R. L. & Thomas, A. L. R. (2003). Flying and swimming animals cruise at a Strouhal number tuned for high power efficiency. Nature 425, 707–711. doi: 10.1038/nature02000 Triantafyllou, G. S., Triantafyllou, M. S. & Grosenbaugh, M. A. (1993). Optimal thrust development in oscillating foils with application to fish propulsion. Journal of Fluids and Structures 7, 205–224. doi: 10.1006/jfls.1993.1012 Van Leeuwen, T. E., Rosenfeld, J. S. & Richards, J. G. (2011). Adaptive trade-offs in juvenile salmonid metabolism associated with habitat partitioning between coho salmon and steelhead trout in coastal streams. Journal of Animal Ecology 80, 1012–1023. doi: 10.1111/j.1365-2656.2011.01841 Videler, J. J. & Wardle, C. S. (1991). Fish swimming stride by stride: speed limits and endurance. Reviews in Fish Biology and Fisheries 1, 23–40. doi: 10.1007%2FBF00042660 Webb, P. W. (1971a). Swimming energetics of trout I. Thrust and power output at cruising speeds. Journal of Experimental Biology 55, 489–520. Webb, P. W. (1971b). Swimming energetics of trout II. Oxygen consumption and swimming efficiency. Journal of Experimental Biology 55, 521–540. Webb, P. W., Kostecki, P. T. & Stevens, E. D. (1984). The effect of size and swimming speed on locomotor kinematics of rainbow trout. Journal of Experimental Biology 109, 77–95. van Weerden, J. F., Reid, D. A. P. & Hemelrijk, C. K. (2014). Meta-analysis of steady undulatory swimming. Fish and Fisheries 15, 397–409. doi: 10.1111/faf.12022

1 Temperature induced cardiac remodeling in fish 2 3 4 Adam N. Keen1, Jordan M. Klaiman2, Holly A. Shiels1 and Todd E. Gillis3,* 5 1Faculty of Life Sciences, University of Manchester, Manchester, UK 6 2Department of Bioengineering, University of Washington, Seattle, WA, USA 7 3Department of Integrative Biology, University of Guelph, ON, Canada. 8 9 10 *Address correspondence to: 11 [email protected] 12 http://comparativephys.ca/gillislab/ 13

14 Summary statement 15 Cold acclimation of some temperate fishes causes extensive remodeling of the heart. The 16 resultant changes to the active and passive properties of the heart represent a highly integrated 17 phenotypic response. 18 19 20 Abstract 21 Cold acclimation can cause the heart of multiple fish species to undergo significant remodeling. 22 This includes changes in connective tissue, in the size of the myocardium as well as in the 23 phosphorylation state of the myofilament proteins. This remodeling response is thought to be 24 responsible for the changes in cardiac stiffness, the Ca2+ sensitivity of force generation and the 25 rate of force generation by the heart following cold acclamation. Such changes to the active and 26 passive properties of the heart help to compensate for the loss of cardiac function caused by a 27 decrease in physiological temperature. Temperature induced cardiac remodeling occurs in fish 28 that remain active following seasonal decreases in temperature, including rainbow trout, 29 Oncorhynchus mykiss, and zebrafish, Danio rerio. The purpose of this paper is to review the 30 current state of knowledge of temperature-induced cardiac remodeling in fish across multiple 31 levels of biological organization, and to examine how such changes result in the modification of 32 the functional properties of the heart. This review is organized around the ventricular phases of 33 the cardiac cycle, specifically diastolic filling, isovolumic pressure generation, and ejection so 34 that the consequence of the remodeling can be fully described. We will also compare the 35 modifications to the fish heart with thermal acclimation to those seen in the mammalian heart 36 in response to cardiac pathologies and exercise. Finally, we will consider how the plasticity of 37 the fish heart may be relevant to survival in a climate change context where seasonal 38 temperature changes could become more extreme and variable. 39

40 Introduction 41 Ectothermic animals living in temperate environments can experience significant, long-term 42 changes in ambient temperature. These seasonal fluctuations influence all levels of biological 43 function due to the universal effect of temperature on molecular interactions. As a result, 44 biochemical, physiological and biomechanical processes are affected. However, a number of 45 ectothermic species, including some fish, remain active across the seasons. These fish species 46 include fresh water salmonids like rainbow trout, Oncorhynchus mykiss, that remain active at 47 temperatures ranging from approximately 4 to 24 oC, and members of the minnow family like 48 the zebrafish, Danio rerio, who experience a 10 oC change in temperature between winter and 49 summer (Lopez-Olmeda and Sanchez-Vazquez, 2011). Marine species, like tunas, also 50 experience seasonal temperature changes associated with seasonal oceanic migrations 51 (Boustany et al., 2010). While a change in temperature will affect the function of all organs, the 52 output of the heart is especially important due to its role in moving oxygen, metabolic 53 substrates and metabolic byproducts around the body and, therefore, supporting active 54 biological processes. Thus, many fish have mechanisms that preserve cardiac function across 55 seasonal temperature changes. 56 57 The purpose of this review is to examine thermal cardiac remodeling in the hearts of selected 58 fish species. We build upon excellent original work (i.e. ‘Steady-state effects of temperature 59 acclimation on the transcriptome of the rainbow trout heart’ (Vornanen et al., 2005) and 60 comprehensive reviews of cardiac plasticity in fish, e.g. (Gamperl and Farrell, 2004)). 61 Importantly, here we review changes in both the active and passive properties of the fish heart 62 following prolonged temperature change and discuss ways in which the remodeling preserves 63 or improves function (physiological remodeling) and ways in which the remodeling causes 64 dysfunction (pathological remodeling). Indeed, one of the interesting aspects of thermal 65 remodeling in the fish heart is that it shows hallmarks of what could be considered physiological 66 as well as pathological remodeling in mammalian species (see (Dorn, 2007; Keen et al., 2016; 67 Klaiman et al., 2011; Klaiman et al., 2014)). 68

69 We have focused our review on ventricular remodeling in two species – rainbow trout and 70 zebrafish. Cardiac remodeling in the trout has been extensively studied, and as a cold-active 71 species its heart develops robust cardiac outputs across a range of temperatures. We also 72 discuss recent work on cardiac remodeling in the zebrafish. This species has become a popular 73 model for understanding the development and regenerative capabilities of the vertebrate heart. 74 With more than 40,000 extant species of fish, the possible remodeling phenotypes are 75 abundant and we do not attempt to cover all of these in this review. However, we do include 76 key studies on other fish species such as tunas, cod, flat fish and carp, where appropriate. A key 77 aim of this review is to show how thermal remodeling of active and passive properties work 78 together to preserve cardiac function across temperatures. For this reason, we have divided the 79 review into 3 main sections that align with the ventricular phases of the cardiac cycle, namely 80 diastolic filling, isovolumic pressure generation, and ejection. We discuss remodeling across 81 multiple levels of biological organization and emphasize the utility of thermal remodeling for 82 maintaining cardiac function throughout the phases of the ventricular cardiac cycle. Through 83 this approach we hope to illustrate how integrated and comprehensive the thermal cardiac 84 remodeling response is. For simplicity, we have structured the article around observations 85 associated with cold acclimation. However, with warming temperatures becoming a global 86 concern, we have added a concluding section that draws together the less studied response of 87 the fish heart to an increase in temperature, and discuss its implications for fish heart function. 88 89 Acute temperature change and cardiac function 90 Effects on the whole heart

91 Acute temperature change directly influences physiological processes in fish through Q10 effects 92 on reaction rates (i.e. the change in rate over a 10 °C temperature difference). As temperature 93 drops, blood viscosity increases, which directly affects vascular resistance and increases cardiac 94 load (Graham and Farrell, 1989; Graham and Fletcher, 1984). At the same time, the heart 95 becomes bradycardic (Keen et al., 1993) which is largely due to a greater diastolic duration, with 96 systolic duration less affected. The larger diastolic duration acts to maintain cardiac output by 97 increasing filling time, which can lead to an increase in stroke volume even though cardiac

98 contractility and force production are reduced (Shiels et al., 2002; Vornanen et al., 2005). 99 Following cold acclimation there are compensatory increases in basal heart rate (Haverinen and

100 Vornanen, 2007; Keen et al., 1993; Lurman et al., 2012) maximum stroke volume (SVmax)

101 (Driedzic et al., 1996; Farrell, 1991; Lurman et al., 2012), maximum power output (PO max) (Bailey

102 and Driedzic, 1990; Lurman et al., 2012), maximum cardiac output (Qmax) (Lurman et al., 2012) 103 and the sensitivity of the β-adrenergic system (Keen et al., 1993). 104 105 Effects on the myofilaments 106 A decrease in the temperature of the vertebrate heart causes an impairment in contractile 107 function as the thin filament loses its sensitivity to Ca2+ resulting in a loss of force generating 108 capacity (Churcott et al., 1994; Harrison and Bers, 1990; Stephenson and Williams, 1985). The 109 decrease in Ca2+ sensitivity in cardiac muscle has been reported in a variety of animals including 110 frogs, mice, rats, rabbits, ferrets and ground squirrels (Churcott et al., 1994; Harrison and Bers, 111 1989; Liu et al., 1993; Liu et al., 1990). Studies by Gillis et al. (Gillis et al., 2005; Gillis et al., 2000; 112 Gillis et al., 2003b; Gillis and Tibbits, 2002) show that this decrease in the Ca2+ sensitivity 113 following an acute decrease in temperature, is due to a decrease in the Ca2+ affinity of cardiac 114 troponin C, which is the Ca2+ activated trigger for the muscle. Trout cardiac muscle behaves 115 similarly to that of mammals in response to a reduction in temperature, however, the 116 myofilaments of the trout heart have several characteristics that allow it to remain functional at 117 low temperatures and over a range of physiological temperatures. Churcott et al. (Churcott et 118 al., 1994) demonstrated that trout cardiac actin-myosin ATPase was more Ca2+ sensitive than 119 that from rats when compared at their respective physiological temperatures and pH (7 °C 120 versus 37 °C; for trout and rat respectfully). Moreover, these authors found that the Ca2+ 121 concentration required by trout cardiac muscle preparations to reach half maximal tension was 122 approximately one-tenth that of rat cardiac tissue when tested at the same experimental 123 temperature (Fig. 1). This higher Ca2+ sensitivity of the trout cardiac tissue is believed to be one 124 mechanism that helps to offset the cardioplegic effects of cold on the trout heart (Blumenschein 125 et al., 2004; Gillis et al., 2003a). 126

127 Effects on ion channel flux and the action potential

2+ 2+ 128 Acute cooling reduces the flux of Ca (ICa, the Ca current) into the myocyte through voltage 129 gated Ca2+ channels, which can directly reduce the contractility of the heart at cold

2+ 130 temperatures (Fig. 2A). This is because ICa is the primary source of the activating Ca that 131 triggers cross-bridge cycling. In fish species which utilize intracellular Ca2+ stores of the 132 sarcoplasmic reticulum (SR) (e.g. rainbow trout, (Hove-Madsen and Tort, 1998; Shiels and 133 White, 2005); burbot (lota Lota), (Shiels et al., 2006b); yellowfin tuna (Thunnus albacares),

134 (Shiels et al., 1999); bluefin tuna (Thunnus orientalis), (Shiels et al., 2011)) the reduction in ICa 2+ 135 has a cascading effect: a reduced amplitude of ICa reduces the trigger for SR Ca release, further 136 reducing Ca2+ available for interacting with the myofilaments. 137 138 Importantly, acute cooling also slows the electrical properties of the heart which can be 139 observed as an increase in the duration of the ventricular action potential (Fig. 2B) (rainbow 140 trout, (Shiels et al., 2000); bluefin tuna, (Galli et al., 2009), and pink salmon (Oncorhynchus 141 gorbuscha), (Ballesta et al., 2012)). This allows more time for Ca2+ influx during the action

142 potential plateau, possibly on the ICa window current (see (Vornanen, 1998)), which can increase 143 Ca2+ influx during cooling. It is important to note that in some species, like bluefin tuna, the drop 144 in Ca2+ influx during cooling is not completely compensated for by the prolongation of the action 145 potential duration. In these hearts, adrenaline, thought to be released during dives into cold 146 water, augments Ca2+ influx through voltage gated ion channels. This increased Ca2+ influx 147 combines with a prolonged action potential duration to restore force generating Ca2+ flux into 148 the myocytes across acute (> 10 oC) temperature changes (Shiels et al., 2015). 149 150 Although this trade-off between action potential duration and Ca2+ influx can suffice for short- 151 term changes in temperature, it is less effective during prolonged thermal acclimation. Indeed, 152 during chronic cold exposure there is a remodeling of potassium (K+) channel expression that 153 serves to re-shorten the action potential duration, probably driven by impingement on electrical 154 restitution. These temperature induced alterations in the ion channels of the fish heart are 155 discussed in detail in a recent review by Vornanen (in press JEB). Together, the negative

156 inotropic effects of an acute decrease in temperature on electrical and mechanical function 157 illustrate the need for temperature-dependent remodeling to preserve the active pumping 158 properties of the fish heart during chronic temperature change. 159 160 Effects on the resting properties of the heart 161 When the temperature of ventricular trabeculae from Atlantic cod was increased from 10 oC to 162 20 oC the amount of work required to lengthen the preparations nearly doubled (Syme et al., 163 2013). The authors suggest that this result was due to an increase in the resting tension of the 164 muscle (Syme et al., 2013). Such a response could be caused by the increase in temperature 165 enhancing the Ca2+ sensitivity of the myofilament, thereby increasing the number of active 166 cross-bridges during diastole (Gillis et al., 2005; Gillis et al., 2000; Gillis et al., 2003b; Gillis and 167 Tibbits, 2002). This effect would stiffen the muscle, impair cardiac filling and potentially limit the 168 ability of the fish to maintain stroke volume as temperature rises (Syme et al., 2013). It logically 169 follows that a decrease in physiological temperature would have an opposite effect on the 170 stiffness of the lengthening muscle. 171 172 Temperature change also influences the resting, non-force generating properties of the heart by 173 affecting the passive properties of the myocardium. For example, an increase in temperature 174 decreases the contribution of viscous tension, viscoelastic tension, and elastic tension to muscle 175 stiffness resulting in decreased passive stiffness (Mutungi and Ranatunga, 1998) and the reverse 176 is true for a decrease in physiological temperature. Together, the changes in the non-force 177 generating properties of the muscle caused by a change in physiological temperature represent 178 a potential challenge for the maintenance of normal cardiac function. It is, therefore, not 179 surprising that factors which contribute to the passive properties of the heart, such as collagen

180 content and composition, are modified in response to thermal acclimation. 181 182 Cardiac Remodeling following chronic temperature change 183 Phase 1- Diastolic filling of the Ventricle

184 The first stage of the cardiac cycle is diastolic filling. As the ventricle relaxes, pressure decreases. 185 When ventricular pressure drops below atrial pressure the atrioventricular valve opens and 186 blood flows from the atrium into the ventricle. This phase of the cardiac cycle is known as 187 isovolumic relaxation and is the duration between the atrioventricular valves opening and 188 closing again. Ventricular pressure and, therefore, diastolic filling volume is largely determined 189 by cardiac preload, which is determined by venous pressure and atrial systole. The sinus 190 venosus and atrium are larger than the ventricle and act as reservoirs by modulating the volume 191 of blood entering the heart (Farrell, 1991). To maintain correct diastolic function the ventricle 192 must be compliant enough to allow sufficient filling, but also be strong enough to withstand the 193 haemodynamic stress of pumping a large volume of blood. Passive tension describes the 194 resistance of a cardiac chamber to diastolic filling and, therefore, plays a role in the Frank- 195 Starling response of the heart (Shiels and White, 2008) where an increase in end-diastolic 196 volume results in an increase in systolic contraction and stroke volume. In rainbow trout, 197 passive stiffness of the whole ventricle increases following cold acclimation, shown by 198 generating ex vivo pressure volume relationships (Fig. 3) (Keen et al., 2016). Functionally, these 199 decreases in chamber compliance may be cardioprotective by providing support to the cardiac 200 wall to counteract the increased haemodynamic stress during high cardiac load. However, 201 excessive stiffening of the myocardium can reduce diastolic filing and in severe cases lead to 202 diastolic dysfunction (Collier et al., 2012). These features are discussed in more detail below. 203 204 Stiffness, compliance and the extracellular matrix 205 The end-diastolic pressure-volume relationship describes myocardial lusitropy, or relaxation. 206 This relationship, and therefore cardiac compliance, is influenced at the organ level by the 207 pericardium and the geometry and thickness of the ventricular walls. In fish, the ratio of spongy 208 to compact tissue is also likely to contribute to cardiac compliance, with compact myocardium 209 being stiffer than spongy myocardium. Historically, ventricular wall thickness and connective 210 tissue content were thought to be the dominating factors driving ventricular compliance, 211 however, there is now evidence to suggest important contributing roles for many extracellular

212 and intracellular mechanisms. In fish hearts, it is likely that a combination of a number of factors 213 determine overall passive stiffness/compliance. 214 215 The main components of the cardiac extracellular matrix (ECM) are the interstitial fibrous 216 proteins, collagen and elastin, and glycosaminoglycans, which connect to ECM proteins to form 217 proteoglycans (Cleutjens and Creemers, 2002; Fomovsky and Holmes, 2010). The ECM also 218 contains fibrils, fibroblast, macrophages, and proteases. The elastic elements of the ECM 219 (collagen and elastin) provide structure and support to the chamber walls, and are therefore 220 central to the overall passive tension of the ventricle (Katz, 2006). Matrix proteins also surround 221 individual myocytes, muscle bundles and blood vessels forming a complex structural network of 222 interstitial matrix and basement membrane (Sanchez-Quintana et al., 1995). Together, this 223 network of proteins help maintain the structural integrity of the heart while also enabling, and 224 controlling, the distensibility of the tissue. 225 226 Collagen is the most common structural protein in the ECM (Fomovsky and Holmes, 2010). It 227 forms stiff fibres that support and maintain alignment of myocytes by bearing wall stress. At 228 high chamber volume the collagen fibres become stiff and straight to resist overexpansion and 229 damage to myocytes (Fomovsky and Holmes, 2010). Chronic increases in cardiac load are often 230 associated with increased myocardial collagen deposition to resist the increased haemodynamic 231 stress. Collagen also increases the passive stiffness of the chamber wall, so excessive fibrosis of 232 the myocardium can reduce chamber compliance (i.e. the change in pressure for a given change 233 in volume) and chamber distensibilty (i.e. the fold change in cardiac compliance), which can 234 have implications for diastolic filling (Collier et al., 2012). In the fish heart, collagen can be 235 identified using picrosirius red staining and is visible in the compact layer and spongy 236 myocardium (Fig. 4A and Fig. 4B). In rainbow trout, myocardial fibrillar collagen content (Keen 237 et al., 2016; Klaiman et al., 2011) and/or connective tissue (Keen et al., 2016; Klaiman et al., 238 2011) has been shown to increase following cold acclimation (Fig. 4C), which is likely to protect 239 the myocardium from the increased haemodynamic stress of pumping cold viscous blood. 240 However, the opposite response has been observed in zebrafish (Fig. 4D) (Johnson et al., 2014).

241 One difference between zebrafish and trout that may help explain this opposite response is the 242 relative size of the ventricle chambers within the heart. According to LaPlace’s law, wall tension 243 (T) of the ventricle is proportional to the product of intraventricular pressure (P) and ventricular 244 radius (r). This means that for a given systolic pressure there is greater tension in the ventricular 245 wall of a trout heart than in that of a zebrafish, and that an increase in systolic pressure, caused 246 by an increase in blood viscosity, would cause a greater net increase in ventricular wall tension 247 in the trout heart. As a result, the increase in connective tissue content in the trout heart would 248 help maintain the structural integrity as described above. In addition, according to LaPlace’s law, 249 greater force is required to inflate a small chamber than a larger chamber. Therefore, an 250 increase in the stiffness of the zebrafish myocardium caused by an acute decrease in 251 temperature would make it more difficult for the zebrafish heart to fill with blood during 252 diastole. The observed decrease in collagen content of the zebrafish heart with cold acclimation 253 may, therefore, help to compensate for this increase in tissue stiffness and as a result, help 254 maintain diastolic function. Recent work by Lee et al. (Lee et al., 2016) using high resolution 255 echocardiography demonstrates that cold acclimation of zebrafish did not alter the early peak 256 velocity / atrial peak velocity (E/A) ratio indicating that there was no loss of diastolic function. 257 This study also demonstrated that cold-acclimated fish had a slower isovolumetric contraction 258 time (IVCT) compared to warm-acclimated fish when measured at 18 oC (Lee et al., 2016). This is 259 suggestive of improved systolic function and that the zebrafish is able to effectively compensate 260 for the influence of low temperature on cardiac function following cold acclimation. 261 262 Myocardial collagen content is a balance of deposition and degradation. Collagen degradation is 263 regulated by matrix metalloproteinases (MMPs) and the gelatinase activity of MMPs is 264 regulated by tissue inhibitors of MMPs (TIMPs). Increased enzymatic activity of TIMPs therefore 265 inhibits collagen degradation by MMPs and is associated with increased collagen deposition. 266 With cold-induced ventricular hypertrophy and fibrosis in rainbow trout, myocardial expression 267 of MMP2 and MMP13 mRNA is downregulated (Keen et al., 2016) and there is an associated 268 upregulation of TIMP2 mRNA transcripts (Fig. 4E) (Keen et al., 2016). Conversely, when cold 269 acclimation of zebrafish caused a decrease in collagen content and in the proportion of thick

270 collagen fibers in the compact myocardium, there was an increase in the gene transcripts for 271 MMP2, and MMP9 in the heart (Fig. 4F) (Johnson et al., 2014). This is suggestive of an increase 272 in collagen turnover that would result in the observed changes in collagen content (Johnson et 273 al., 2014) and is further evidence that MMPs play a role in regulating collagen content in the fish 274 heart during thermal acclimation. 275 276 The predominant fibrillar collagen in cardiac tissue is collagen I, followed by collagen III (Eghbali 277 and Weber, 1990). Fibrillar collagen molecules are made by super-coiling 3 alpha amino acid 278 chains into an alpha helix. In mammals, collagen I is composed of two type 1 (α1) and one type 2 279 (α2) subunits. However, in collagen I of bony fishes, one of the α1 chains is replaced with a type 280 3 (α3) subunit (Saito et al., 2001). This fish-specific α3 chain is upregulated in the cold-induced 281 fibrosis observed in the trout heart (Keen et al., 2016). Interestingly, the α3 chain reduces the 282 denaturation temperature of the collagen I molecule and makes it more susceptible to 283 degradation by MMP13 (Saito et al., 2001), which may explain the transient nature of cardiac 284 fibrosis in trout following thermal acclimation. Comparatively, in mammals an increase in 285 cardiac connective tissue, or in the ratio of Type I:Type III collagen, is considered a pathological 286 condition that stiffens the myocardium, since Type I collagen is less extensible than Type III (Jalil 287 et al., 1988; Jalil et al., 1989; Marijianowski et al., 1995; Pauschinger et al., 1999). Such changes 288 are common, and permanent, in the hearts of patients suffering from cardiac hypertension, 289 dilated cardiomyopathy or chronic congestive heart failure and contribute to the associated 290 diastolic dysfunction and eventual heart failure (Jalil et al., 1988; Jalil et al., 1989; Marijianowski 291 et al., 1995; Pauschinger et al., 1999). The ability of fish species, including the zebrafish and 292 trout, to regulate myocardial collagen content in response to changes in physiological 293 conditions suggests greater phenotypic plasticity. 294 295 Intracellular contribution to stiffness/compliance 296 At the myocyte level, cardiac compliance during diastolic filling is influenced by a number of 297 features. Firstly, the amount and speed of Ca2+ removal from the cytoplasm by the SR and the 298 Na-Ca exchanger alters stiffness/compliance through residual active tension. The Ca2+ affinity of

299 troponin and the dissociation of contractile proteins once Ca2+ has dissociated from troponin 300 (Katz, 2006) influences this relationship. Secondly, passive stiffness of the cytoskeleton and of 301 sarcomeric proteins such as titin plays a large role in overall myocyte stiffness/compliance 302 (Granzier et al., 1996; Horowits et al., 1989; Shiels and White, 2008; Watanabe et al., 2002). 303 Titin is a giant sacromeric protein that runs from the Z-line through to the M-line (Helmes et al., 304 1996; Linke, 2008; Linke et al., 1996; Peng et al., 2007; Wu et al., 2000). Two titin isoforms exist 305 in the vertebrate adult heart; a shorter and stiffer N2B isoform and a longer and more 306 compliant N2BA isoform (Cazorla et al., 2000a; Patrick et al., 2010). The ratio of the two 307 isoforms modulates titin-based passive tension (Cazorla et al., 2000a; Fukuda et al., 2005; Linke, 308 2008; Trombitas et al., 2001). In addition, phosphorylation of the N2B element by protein kinase 309 A (PKA) or protein kinase G (PKG) can decrease passive force (Kruger and Linke, 2009). Cardiac 310 output in the rainbow trout heart can be modulated by up to 3-fold increases in stroke volume. 311 Therefore, it is perhaps unsurprising that rainbow trout ventricular myocytes have a higher ratio 312 of the compliant N2BA isoform compared to a rat myocyte (Patrick et al., 2010). However, 313 passive tension remains higher in a fish myocyte than a rat myocyte due to a lower titin 314 phosphorylation, which may explain the large Frank-Starling response (Patrick et al., 2010). 315 316 Titin specific isoforms show plasticity, with changing haemodynamics in cardiac growth altering 317 titin ratios, but little is known about the mechanism (Linke, 2008). In mammals, the ratios of 318 titin isoforms have been suggested to shift to compensate for cardiac fibrosis by increasing the 319 compliant N2BA isoform (Neagoe et al., 2002). However, increased compliance of titin may 320 reduce systolic function via the Frank-Starling curve as reduced titin spring activity (Linke, 2008). 321 At present the effect of temperature acclimation on the intracellular structure and titin 322 remodeling in the fish heart is not known. However, this is likely to be an important feature of 323 determining the passive properties of the fish heart. Keen et al. (Keen et al., 2016) 324 demonstrated this in rainbow trout by measuring micromechanical stiffness of ventricular tissue 325 sections with atomic force microscopy. The accumulative frequency curves showed an even 326 distribution, suggesting that tissue stiffness was increasing across the tissue rather than due to 327 an increase in stiff elements, such as fibrillar collagen. Future studies should aim to understand

328 the changes in the intracellular structure of the fish myocyte with temperature acclimation and 329 its contribution to the overall changes in passive tension of the fish ventricle. 330 331 Cardiac Hypertrophy 332 Wall thickness is known to affect passive stiffness of the ventricle, therefore hypertrophy 333 (muscle growth) or atrophy (muscle loss) of the ventricle may influence the diastolic filling 334 phase of the cardiac cycle. In mammals, cardiac hypertrophy is initiated by increased cardiac 335 load caused by physiological stressors, including aerobic exercise and pregnancy, or a 336 pathological condition, such as a myocardial infarction or hypertension (Dorn, 2007; Dorn et al., 337 2003). The elevated biomechanical strain of chronic pressure or volume overload causes 338 increased tension of the heart wall, which triggers increased mRNA production and protein 339 synthesis leading to cellular hypertrophy (Bishop, 1990; Nadal-Ginard et al., 2003; Chen et al., 340 2007). Chronic increases in cardiac preload, or ‘volume overload’, cause an eccentric 341 hypertrophy where heart mass increases in line with wall thickness and lumen volume is 342 preserved (Dorn, 2007). Sarcomeres are added in series causing longitudinal growth of 343 myocytes. Chronic increases in cardiac afterload, or ‘pressure overload’, invoke concentric 344 hypertrophy where both cardiac mass and relative wall thickness increase, but there is minimal 345 associated change in chamber volume. Sarcomeres are added in parallel causing lateral myocyte 346 growth (Dorn, 2007). Following concentric hypertrophy the inner layers of ventricular muscle 347 develop greater wall stress than the outer layers, bringing a rise in energy expenditure that 348 means the heart is more vulnerable to energy starvation if demand increases (Katz, 2006). 349 Capillary growth is vital to provide the growing cardiac muscle with a sufficient supply of oxygen 350 and nutrition and secretion of angiogenic factors, such as vascular endothelial growth factor 351 (VEGF) and angiopoietins (Ang-1 and 2), are typically increased in mammals (Weber and Janicki, 352 1989). Cardiac hypertrophy in mammals does not typically involve cellular hyperplasia (increase 353 in cell numbers), oedema (cellular swelling) or infiltration (accumulation of extracellular 354 components), but under pathological conditions hypertrophy occurs concurrently with an 355 increase in fibrotic tissue. 356

357 A number of studies have shown increased ventricular mass in fish following cold acclimation, 358 which is likely reversed following warm acclimation (Aho and Vornanen, 1998; Driedzic et al., 359 1996; Farrell et al., 1988; Kent et al., 1988; Klaiman et al., 2011; Vornanen et al., 2005). The 360 increased ventricular mass is mainly attributed to an increase in myocyte size, suggesting it is a 361 physiological hypertrophic response (Aho and Vornanen, 1998; Driedzic et al., 1996; Keen et al., 362 2016; Klaiman et al., 2011; Vornanen et al., 2005). However, some studies also suggest myocyte 363 hyperplasia (increase in cell number) in addition to hypertrophy A number of studies have 364 shown increased ventricular mass in fish following cold acclimation, (Aho and Vornanen, 1998; 365 Driedzic et al., 1996; Farrell et al., 1988; Kent et al., 1988; Klaiman et al., 2011; Vornanen et al., 366 2005a) and that warm acclimation can have the opposite response (Klaiman et al., 2011). This 367 suggests that the cold-induced cardiac hypertrophy is reversible. The increased ventricular mass 368 is mainly attributed to an increase in myocyte size, suggesting it is a physiological hypertrophic 369 response (Aho and Vornanen, 1998; Driedzic et al., 1996; Keen et al., 2016; Klaiman et al., 2011; 370 Vornanen et al., 2005a). However, some studies also suggest myocyte hyperplasia in addition to 371 hypertrophy (Farrell et al., 1988; Keen et al., 2016; Sun et al., 2009). With cold-induced 372 hypertrophy of the rainbow trout heart, mRNA expression of vascular endothelial growth factor 373 (VEGF) is up regulated (Keen et al., 2016). As VEGF is involved in angiogenesis, this suggests an 374 increased blood supply to the muscle (Keen et al., 2016). These changes occur in parallel with 375 the growth of the spongy myocardium, a tissue that obtains oxygen directly from the blood in 376 the chamber. This hypertrophic response with cold acclimation along with the increase in 377 cardiac connective tissue indicates that changes in physiological conditions can elicit a 378 significant phenotypic response as the heart continues to function. 379 380 Phase 2 - Pressure Generation 381 The second stage of the cardiac cycle is pressure generation. Following ventricle filling and 382 closing of the atrioventricular valve the ventricular myocardium starts to isometrically contract, 383 building up pressure within the ventricle. During this phase of the ventricular cardiac cycle, the 384 strength of contraction is modulated by changes in contractility at a given sarcomere length. If 385 increased filling stretched the ventricle during phase 1 of the cardiac cycle, this would increase

386 the sarcomere length and contribute to greater force generation during phase 2. In this way, an 387 increase in end-diastolic volume results in an increase in systolic contraction and stroke volume, 388 (Frank-Starling response). At the cellular level, an increase in pressure during ventricle loading 389 stretches the myocytes in the ventricle causing increased sarcomere length (SL) and, thus 390 changes in the force of contraction (reviewed in (Shiels and White, 2008)). Mammalian cardiac 391 myocytes show an increase in the force of contraction with an increase in SL until a peak of 392 about 2.2 μm (Gordon et al., 1966); however, Shiels et al. (Shiels et al., 2006a) have 393 demonstrated that the active force of contraction increased in trout cardiac myocytes until an 394 SL of 2.6 μm. Since the trout heart has a high ejection fraction volume (>80%; (Franklin and 395 Davie, 1992)), this would allow the ventricle to be stretched to a greater extent, and as a result, 396 allow for greater diastolic filling, and increased strength of contraction. These factors are critical 397 to the critical role of stroke volume in the regulation of cardiac output in fish (Shiels et al., 398 2006a). 399 400 Myofibril remodeling 401 Force is produced in striated muscle by the cycling of cross-bridges between the actin thin 402 filaments and myosin thick filaments. This reaction is initiated by Ca2+ binding to the thin 403 filament and results in muscle contraction. Phosphorylation of key regulatory proteins – 404 including cardiac troponin I (cTnI), cardiac troponin T (cTnT) and myosin binding protein C 405 (MyBP-C) – can modulate myofilament function in the vertebrate heart (reviewed by (Shaffer 406 and Gillis, 2010). In the mammalian heart, these proteins can be targeted by protein kinase A 407 (PKA) or protein kinase C (PKC) following β-adrenergic or α-adrenergic stimulation, respectively 408 (Shaffer and Gillis, 2010). The resultant functional changes that follow PKA phosphorylation 409 include a decrease in the Ca2+ sensitivity of force generation, increased kinetics of Ca2+ 410 activation and a decrease in force generation (Chandra et al., 1997; Dong et al., 2007). In trout 411 hearts, phosphorylation of myofilaments from the trout heart by PKA results in an increase in 412 the kinetics of Ca2+ activation and a decrease in maximal force generation (Gillis and Klaiman, 413 2011). Interestingly, cold acclimation of trout results in an increase in the maximal rate of the 414 cardiac actomyosin-ATPase (Fig. 5a), an increase in the Ca2+ sensitivity of force generation by

415 skinned cardiac trabeculae (Fig. 5b) as well as an increase in the magnitude and rate of pressure 416 generation by the isolated heart (Fig. 5c) (Klaiman et al., 2014). This indicates that the functional 417 ability of the heart has been increased with cold acclimation (Klaiman et al., 2014). 418 Quantification of protein phosphorylation of the myofilament proteins in the cold-acclimated 419 hearts demonstrates a decrease in the phosphorylation of cTnT, slow skeletal TnT, and MyBP-C. 420 This suggests that the changes in myofilament function are due, at least in part, to post- 421 translational changes in the myofilament regulatory proteins (Klaiman et al., 2011; Klaiman et 422 al., 2014). 423 424 Cold acclimation of trout has also been found to alter the gene transcript levels for different 425 isoforms of cardiac myofilament proteins. More specifically, Genge et al, (Genge et al., 2013) 426 identified transcripts for two isoforms of TnC in the trout heart that are modulated by cold 427 acclimation. Troponin C is the Ca2+ activated trigger that initiates myocyte contraction, and 428 previous studies have demonstrated that manipulation of the isoform working in the muscle can 429 alter contractile function (Gillis et al., 2005). In addition, Alderman et al. (Alderman et al., 2012) 430 demonstrated that the trout heart expresses the gene transcripts for seven different TnI 431 isoforms, and that the abundance of four of these change with cold acclimation. If these 432 changes in transcript abundance translate into changes in the complement of protein isoforms 433 present in the muscle, this would potentially alter the contractile function of the muscle. Such a 434 strategy may be utilized to maintain contractile function in the trout heart with cold 435 acclimation. 436 437 Cardiac morphology 438 The hypertrophic response caused by cold acclimation represents an increase in muscle mass. It 439 is not surprising, therefore, that cold-acclimated trout hearts, exhibiting cardiac hypertrophy, 440 have a greater capacity to generate pressure (Graham and Farrell, 1989). Therefore, cardiac 441 hypertrophy in fish is a strategy to help compensate for the effect of low temperature on the 442 active properties of the muscle (Driedzic et al., 1996; Gamperl and Farrell, 2004; Keen et al., 443 2016; Klaiman et al., 2011). However, recent work by Klaiman et al. (Klaiman et al., 2014)

444 demonstrated that cold acclimation of trout can increase the pressure generating capacity of 445 the heart in the absence of a hypertrophic response (Fig. 5C). This change in function is likely 446 due, at least in part, to alterations of the myofilament (see the Myofibril Remodeling section 447 above). In this study there were also changes to the morphology of the heart (Klaiman et al., 448 2014), including a decrease in the relative proportion of compact myocardium and a reciprocal 449 increase in spongy myocardium (Fig. 6 and Fig. 7). Such a change in cardiac morphology with 450 cold acclimation has been reported in other studies of trout (Farrell et al., 1988; Keen et al., 451 2016; Klaiman et al., 2011) as well as for zebrafish (Johnson et al., 2014). In the fish heart, the 452 spongy myocardium is composed of trabecular sheets that enable the formation of lacunae that 453 fill with blood during diastole. Then, during systole the trabecular sheets act as “contractile 454 girders”, helping to pull the compact myocardium inward during contraction (Pieperhoff et al., 455 2009). This functional organization of the myocardium is thought to enable the extremely high 456 ejection fraction of the trout heart (~80%) compared to the mammalian heart (50-60%), which 457 does not contain spongy myocardium (Franklin and Davie, 1992). The observed increase in 458 spongy myocardium seen in the trout heart with cold acclimation would, therefore, increase the 459 stroke volume of the heart while also increasing the relative proportion of contractile 460 machinery. Such a change would make the heart more of a volume pump, as opposed to a 461 pressure pump. An extreme example of the heart acting as a volume pump is in the Antarctic 462 icefish (Channichthyidae), where the heart has a comparatively high proportion of spongy 463 myocardium and very little compact myocardium. These characteristics allow the icefish heart 464 to displace large systolic volumes at a low rate and relatively low pressure, which requires large 465 ventricular fillings (high ventricular compliance) (Zummo et al., 1995), and this in turn enables 466 adequate cardiac output at near freezing temperatures. 467 468 Length-dependent changes in force generation 469 As discussed above, changes in the resting length of the sarcomere can affect the strength of 470 contraction, and thus the pressure generating capacity of the ventricle. Interestingly, Klaiman 471 et al. (Klaiman et al., 2014) demonstrated that the difference in developed pressure at higher 472 balloon volumes between hearts from cold- and warm-acclimated fish was greater than at

473 smaller balloon volumes. One possible explanation for this result is that the cardiac muscle of 474 thermally acclimated fish may respond differently to stretch. As discussed earlier, rainbow trout 475 cardiac muscle has a larger working range of the Frank-Starling curve compared to rats as well 476 as a longer optimal SL (2.6 vs 2.2) (Patrick, et al., 2010; Shiels et al., 2006; Cazorla et al., 2000). 477 In addition, previous work in mammals has shown that following a physiological stressor such as 478 exercise training, cardiac tissue has a greater response to stretch (known as length dependent 479 activation, LDA) (Diffee and Nagle, 2003). Thus, it is possible that LDA is more prominent in the 480 trout heart following acclimation to cold temperatures. This hypothesis deserves future 481 investigation. 482 483 Phase 3 – Ejection 484 The third stage of the cardiac cycle is ejection. Following pressure generation by the 485 myocardium the bulbo-ventricular valve opens and blood is forced from the ventricle into the 486 bulbus arteriosus in the fish outflow tract and to the rest of the body. In zebrafish, ejection time 487 decreases with acute? reductions in ambient temperature, however, there are no effects 488 following cold acclimation (Lee et al., 2016). Heart rate determines the duration between 489 ejections. Although an acute decrease in temperature slows heart rate (Driedzic and Gesser, 490 1994), cold acclimation results in partial thermal compensation (Aho and Vornanen, 1999; Little 491 and Seebacher, 2014). Conversely, stroke volume is not altered by acute temperature change 492 (Clark et al., 2008; Gollock et al., 2006; Lee et al., 2016; Mendonca et al., 2007; Steinhausen et 493 al., 2008), while during chronic cooling it may remain constant or increase. Although Lee et al., 494 (Lee et al., 2016) showed that stroke volume peaks when ambient temperature matches 495 acclimation temperature, cold acclimation significantly increases systolic function, with 496 increases in ejection fraction and fractional shortening, which is consistent with increases in 497 contractile proteins (as explained above) (Genge et al., 2013). In zebrafish, acute temperature 498 change does not to affect the E/A ratio suggesting that at all temperatures ventricular preload, 499 and therefore ejection fraction, is primarily determined by late diastolic filling, which is atrial 500 contraction dependent (Farrell and Jones, 1992; Lee et al., 2016). 501

502 Influence of warm acclimation on the structure and function of the heart 503 Ectothermic animals living in temperate environments are likely to experience larger 504 fluctuations in ambient temperatures as global climate change progresses, including periods of 505 higher than average temperatures during summer months. The ability of fish species to respond 506 to acute changes in temperature may therefore be critical to their long-term survival. However, 507 so likely is the capacity of fish to remodel their hearts in response to a change in physiological 508 temperature. For example Badr et al. (Badr et al., 2016) demonstrated that warm acclimation 509 increases the temperature at which heart rate become irregular in the loach, Rutilus rutilus. 510 Similarly, our groups demonstrated that warm acclimation affects the morphology, composition 511 and function of the trout heart (Keen et al., 2016; Klaiman et al., 2011; Klaiman et al., 2014). 512 Specifically, acclimation of trout to 17 oC causes an increase in the thickness of the spongy layer, 513 decreases in the area of the spongy myocardium and in connective tissue content, as well as a 514 decrease in ventricular pressure generation and in the rate of pressure development compared 515 to cold acclimated (4 oC) fish when measured at a common experimental temperature of 15 oC 516 (Fig. 6 and Fig. 7). The increase in compact layer thickness and decrease in spongy layer 517 thickness are linked to a functional increase in ventricular compliance (Keen et al., 2016), 518 suggesting that the volume of blood being pumped per beat is reduced. As an increase in 519 physiological temperature increases heart rate in fish (Aho and Vornanen, 2001; Badr et al., 520 2016; Lee et al., 2016), this suggests that the heart functions more as a pressure pump, as 521 opposed to a volume pump, with warm acclimation. While the pressure generation capacity and 522 rate of contraction were lower in the hearts of fish acclimated to 17 oC these functional 523 properties of the warm-acclimated fish hearts were being compared to that of control (11 oC) 524 and cold-acclimated (4 oC) fish at a common experimental temperature of 15 oC (Klaiman et al., 525 2011; Klaiman et al., 2014). Together these results demonstrate that higher temperatures also 526 initiate a cardiac remodeling response in the trout. What is currently unknown, however, is 527 how/if the trout heart can remodel to temperatures above its normal seasonal range, and what 528 the functional consequence of such remodeling is. Finally, future work should investigate how 529 rapidly a fish heart can remodel in response to a change in environmental temperature and the 530 physiological consequence of multiple remodeling events. In addition, all known studies have

531 looked at fixed time points (6 or 8 weeks) of thermal acclimation and not at the time course of 532 remodeling. Such information would be relevant to understanding how stochastic 533 environmental temperatures may impact natural fish populations. 534 535 Conclusions 536 The ability of some fish to remodel their heart in response to changes in environmental 537 temperature has ecological consequences, as it enables them to remain active over a wide 538 range of environmental temperatures. Such plasticity may also provide a better ability to 539 maintain cardiac function as average seasonal temperatures begin to increase with global 540 climate change. Independent of these potential advantages, the ability of fish to remodel their 541 heart in response to changes in environmental conditions is a significant feat that results from . 542 Current and future studies focused on characterizing the molecular and cellular processes that 543 underpin these changes in form and function will increase our understanding of the 544 mechanisms responsible for such plasticity. Such knowledge has significant biomedical 545 application in the development of strategies to control the pathological remodeling response, 546 and will increase our understanding of what limits the ability of the vertebrate heart to remodel 547 in response to a physiological stressor. 548 549 GRANTS 550 ANK is supported by a Doctoral Training Partnership from the Biotechnology and Biosciences 551 Research Council (BBSRC). JMK is supported by a Post Doctoral Fellowship from the Heart and 552 Stroke Foundation of Canada. TEG is supported by the Natural Sciences and Engineering 553 Research Council (NSERC) and the Canadian Foundation for Innovation. 554 555 AUTHOR CONTRIBUTIONS 556 All Authors contributed to the writing and editing of the review. 557 558 ACHNOWLEDGEMENTS 559 The authors thank SA Alderman for editorial comments on an earlier version of the manuscript.

560 561 REFERENCES 562 Aho, E. and Vornanen, M. (1998). Ca2+-ATPase activity and Ca2+ uptake by sarcoplasmic 563 reticulum in fish heart: effects of thermal acclimation. J Exp Biol 201, 525-32. 564 Aho, E. and Vornanen, M. (1999). Contractile properties of atrial and ventricular 565 myocardium of the heart of rainbow trout oncorhynchus mykiss: effects of thermal acclimation. 566 J Exp Biol 202, 2663-77. 567 Aho, E. and Vornanen, M. (2001). Cold acclimation increases basal heart rate but 568 decreases its thermal tolerance in rainbow trout (Oncorhynchus mykiss). J Comp Physiol [B] 171, 569 173-9. 570 Alderman, S. L., Klaiman, J. M., Deck, C. A. and Gillis, T. E. (2012). Effect of cold 571 acclimation on troponin I isoform expression in striated muscle of rainbow trout. Am J Physiol 572 303, R168-76. 573 Badr, A., El-Sayed, M. F. and Vornanen, M. (2016). Effects of seasonal acclimatization on 574 temperature-dependence of cardiac excitability in the roach, Rutilus rutilus. J Exp Biol. 575 Bailey, J. R. and Driedzic, W. R. (1990). Enhanced maximum frequency and force 576 development of fish hearts following temperature acclimation. J Exp Biol 149, 239-254. 577 Ballesta, S., Hanson, L. M. and Farrell, A. P. (2012). The effect of adrenaline on the 578 temperature dependency of cardiac action potentials in pink salmon Oncorhynchus gorbuscha. J 579 Fish Biol 80, 876-85. 580 Blumenschein, T. M., Gillis, T. E., Tibbits, G. F. and Sykes, B. D. (2004). Effect of 581 temperature on the structure of trout troponin C. Biochemistry 43, 4955-63. 582 Boustany, A., Matteson, R., Castleton, M., Farwell, C. and Block, B. A. (2010). 583 Movements of pacific bluefin tuna (Thunnus orientalis) in the Eastern North Pacific revealed 584 with archival tags, . Progress in Oceanography 86, 94-104. 585 Cazorla, O., Freiburg, A., Helmes, M., Centner, T., McNabb, M., Wu, Y., Trombitas, K., 586 Labeit, S. and Granzier, H. (2000a). Differential expression of cardiac titin isoforms and 587 modulation of cellular stiffness. Circ Res 86, 59-67.

588 Cazorla, O., Le Guennec, J. Y. and White, E. (2000b). Length-tension relationships of 589 subepicardial and sub-endocardial single ventricular myocytes from rat and ferret hearts. J Mol 590 Cell Cardiol 32, 735-744. 591 Chandra, M., Dong, W. J., Pan, B. S., Cheung, H. C. and Solaro, R. J. (1997). Effects of 592 protein kinase A phosphorylation on signaling between cardiac troponin I and the N-terminal 593 domain of cardiac troponin C. Biochemistry 36, 13305-11. 594 Churcott, C. S., Moyes, C. D., Bressler, B. H., Baldwin, K. M. and Tibbits, G. F. (1994). 595 Temperature and pH effects on Ca2+ sensitivity of cardiac myofibrils: a comparison of trout with 596 mammals. Am J Physiol 267, R62-70. 597 Clark, T. D., Taylor, B. D., Seymour, R. S., Ellis, D., Buchanan, J., Fitzgibbon, Q. P. and 598 Frappell, P. B. (2008). Moving with the beat: heart rate and visceral temperature of free- 599 swimming and feeding bluefin tuna. Proc Royal Soc B 275, 2841-2850. 600 Cleutjens, J. P. and Creemers, E. E. (2002). Integration of concepts: cardiac extracellular 601 matrix remodeling after myocardial infarction. J Card Fail 8, S344-8. 602 Collier, P., Watson, C. J., van Es, M. H., Phelan, D., McGorrian, C., Tolan, M., Ledwidge, 603 M. T., McDonald, K. M. and Baugh, J. A. (2012). Getting to the heart of cardiac remodeling; how 604 collagen subtypes may contribute to phenotype. J Mol Cell Cardiol 52, 148-53. 605 Diffee, G. M. and Nagle, D. F. (2003). Exercise training alters length dependence of 606 contractile properties in rat myocardium. J App Physiol 94, 1137-44. 607 Dong, W. J., Jayasundar, J. J., An, J., Xing, J. and Cheung, H. C. (2007). Effects of PKA 608 phosphorylation of cardiac troponin I and strong crossbridge on conformational transitions of 609 the N-domain of cardiac troponin C in regulated thin filaments. Biochemistry 46, 9752-61. 610 Dorn, G. W., 2nd. (2007). The fuzzy logic of physiological cardiac hypertrophy. 611 Hypertension 49, 962-70. 612 Dorn, G. W., 2nd, Robbins, J. and Sugden, P. H. (2003). Phenotyping hypertrophy: 613 eschew obfuscation. Circ Res 92, 1171-5. 614 Driedzic, W. R., Bailey, J. R. and Sephton, D. H. (1996). Cardiac Adaptations to low 615 temperature in non-polar teleost fish. J Exp Zool 275, 186-195.

616 Driedzic, W. R. and Gesser, H. (1994). Energy metabolism and contractility in 617 ectothermic vertebrate hearts: hypoxia, acidosis, and low temperature. Physiol Rev 74, 221-58. 618 Eghbali, M. and Weber, K. T. (1990). Collagen and the myocardium: fibrillar structure, 619 biosynthesis and degradation in relation to hypertrophy and its regression. Mol Cell Biochem 96, 620 1-14. 621 Farrell, A. P. (1991). From Hagfish to Tuna - a Perspective on Cardiac-Function in Fish. 622 Physioll Zool 64, 1137-1164. 623 Farrell, A. P., Hammons, A. M., Graham, M. S. and Tibbits, G. F. (1988). Cardiac Growth 624 in Rainbow-Trout, Salmo-Gairdneri. Cana J Zool 66, 2368-2373. 625 Farrell, A. P. and Jones, D. R. (1992). The Heart. In Fish Physiology, The Cardiovascular 626 system. , vol. 12A. London: ACADEMIC PRESS, INC. 627 Fomovsky, G. M. and Holmes, J. W. (2010). Evolution of scar structure, mechanics, and 628 ventricular function after myocardial infarction in the rat. Am J Physiol 298, H221-8. 629 Franklin, C. E. and Davie, P. S. (1992). Dimensional analysis of the ventricle of an in situ 630 perfused trout heart using echocardiography. J Exp Biol 166, 47-60. 631 Fukuda, N., Wu, Y., Farman, G., Irving, T. C. and Granzier, H. (2005). Titin-based 632 modulation of active tension and interfilament lattice spacing in skinned rat cardiac muscle. 633 Pflugers Arch 449, 449-57. 634 Galli, G. L., Shiels, H. A. and Brill, R. W. (2009). Temperature sensitivity of cardiac 635 function in pelagic fishes with different vertical mobilities: yellowfin tuna (Thunnus albacares), 636 bigeye tuna (Thunnus obesus), mahimahi (Coryphaena hippurus), and swordfish (Xiphias 637 gladius). Physiol Biochem Zool : PBZ 82, 280-90. 638 Gamperl, A. K. and Farrell, A. P. (2004). Cardiac plasticity in fishes: environmental 639 influences and intraspecific differences. J Exp Biol 207, 2539-2550. 640 Genge, C. E., Davidson, W. S. and Tibbits, G. F. (2013). Adult teleost heart expresses two 641 distinct troponin C paralogs: cardiac TnC and a novel and teleost-specific ssTnC in a chamber- 642 and temperature-dependent manner. Physiol Genomics 45, 866-75.

643 Gillis, T. E., Blumenschein, T. M., Sykes, B. D. and Tibbits, G. F. (2003a). Effect of 644 temperature and the F27W mutation on the Ca2+ activated structural transition of trout cardiac 645 troponin C. Biochemistry 42, 6418-26. 646 Gillis, T. E. and Klaiman, J. M. (2011). The influence of PKA treatment on the Ca2+ 647 activation of force generation by trout cardiac muscle. J Exp Biol 214, 1989-96. 648 Gillis, T. E., Liang, B., Chung, F. and Tibbits, G. F. (2005). Increasing mammalian 649 cardiomyocyte contractility with residues identified in trout troponin C. Physiol Genomics 22, 1- 650 7. 651 Gillis, T. E., Marshall, C. R., Xue, X. H., Borgford, T. J. and Tibbits, G. F. (2000). Ca(2+) 652 binding to cardiac troponin C: effects of temperature and pH on mammalian and salmonid 653 isoforms. Am J Physiol 279, R1707-15. 654 Gillis, T. E., Moyes, C. D. and Tibbits, G. F. (2003b). Sequence mutations in teleost 655 cardiac troponin C that are permissive of high Ca2+ affinity of site II. Am J Physiol 284, C1176-84. 656 Gillis, T. E. and Tibbits, G. F. (2002). Beating the cold: the functional evolution of 657 troponin C in teleost fish. Comp Biochem Physiol A 132, 763-72. 658 Gollock, M. J., Currie, S., Petersen, L. H. and Gamperl, A. K. (2006). Cardiovascular and 659 haematological responses of Atlantic cod (Gadus morhua) to acute temperature increase. J Exp 660 Biol 209, 2961-70. 661 Graham, M. S. and Farrell, A. P. (1989). The effect of temperature acclimation and 662 adrenaline on the performance of a perfused trout heart. Physiol Zool 62, 38-61. 663 Graham, M. S. and Fletcher, G. L. (1984). On the blood viscosity of a cold-water, marine 664 sculpin: a comparison with the winter flounder. J Comp Physiol B 155, 455-459. 665 Granzier, H., Helmes, M. and Trombitas, K. (1996). Nonuniform elasticity of titin in 666 cardiac myocytes: a study using immunoelectron microscopy and cellular mechanics. Biophys J 667 70, 430-42. 668 Harrison, S. M. and Bers, D. M. (1989). Influence of temperature on the calcium 669 sensitivity of the myofilaments of skinned ventricular muscle from the rabbit. J Gen Physiol 93, 670 411-28.

671 Harrison, S. M. and Bers, D. M. (1990). Temperature dependence of myofilament Ca 672 sensitivity of rat, guinea pig, and frog ventricular muscle. Am J Physiol 258, C274-81. 673 Haverinen, J. and Vornanen, M. (2007). Temperature acclimation modifies sinoatrial 674 pacemaker mechanism of the rainbow trout heart. Am J Physiol 292, R1023-32. 675 Helmes, M., Trombitas, K. and Granzier, H. (1996). Titin develops restoring force in rat 676 cardiac myocytes. Circ Res 79, 619-26. 677 Horowits, R., Maruyama, K. and Podolsky, R. J. (1989). Elastic behavior of connectin 678 filaments during thick filament movement in activated skeletal muscle. J Cell Biol 109, 2169-76. 679 Hove-Madsen, L. and Tort, L. (1998). L-type Ca2+ current and excitation-contraction 680 coupling in single atrial myocytes from rainbow trout. Am J Physiol 275, R2061-9. 681 Jalil, J. E., Doering, C. W., Janicki, J. S., Pick, R., Clark, W. A., Abrahams, C. and Weber, 682 K. T. (1988). Structural vs. contractile protein remodeling and myocardial stiffness in 683 hypertrophied rat left ventricle. J Mol Cell Cardiol 20, 1179-87. 684 Jalil, J. E., Doering, C. W., Janicki, J. S., Pick, R., Shroff, S. G. and Weber, K. T. (1989). 685 Fibrillar collagen and myocardial stiffness in the intact hypertrophied rat left ventricle. Circ Res 686 64, 1041-50. 687 Johnson, A. C., Turko, A. J., Klaiman, J. M., Johnston, E. F. and Gillis, T. E. (2014). Cold 688 acclimation alters the connective tissue content of the zebrafish (Danio rerio) heart. J Exp Biol. 689 Katz, A. M. (2006). Physiology of the heart. Philadelphia: Lippincott Williams & Wilkins. 690 Keen, A. N., Fenna, A. J., McConnell, J. C., Sherratt, M. J., Gardner, P. and Shiels, H. A. 691 (2016). The Dynamic Nature of Hypertrophic and Fibrotic Remodeling of the Fish Ventricle. 692 Front Physiol 6, 427. 693 Keen, J. E., Vianzon, D.-E., Farrell, A. P. and Tibbits, G. F. (1993). Thermal 694 acclimationalters both adrenergic sensitivity and adrenoceptor density in cardiac tissue of 695 rainbow trout. J Exp Biol 181, 27-47. 696 Kent, J., Koban, M. and Prosser, C. L. (1988). Cold-acclimation-induced protein 697 hypertrophy in channel catfish and green sunfish. J Comp Physiol [B] 158, 185-98.

698 Klaiman, J. M., Fenna, A. J., Shiels, H. A., Macri, J. and Gillis, T. E. (2011). Cardiac 699 Remodeling in Fish: Strategies to Maintain Heart Function during Temperature Change. PLoS 700 One 6, e24464. 701 Klaiman, J. M., Pyle, W. G. and Gillis, T. E. (2014). Cold acclimation increases cardiac 702 myofilament function and ventricular pressure generation in trout. J Exp Biol 217, 4132-40. 703 Kruger, M. and Linke, W. A. (2009). Titin-based mechanical signalling in normal and 704 failing myocardium. J Mol Cell Cardiol 46, 490-8. 705 Lee, L., Genge, C. E., Cua, M., Sheng, X., Rayani, K., Beg, M. F., Sarunic, M. V. and 706 Tibbits, G. F. (2016). Functional Assessment of Cardiac Responses of Adult Zebrafish (Danio 707 rerio) to Acute and Chronic Temperature Change Using High-Resolution Echocardiography. PLoS 708 One 11, e0145163. 709 Linke, W. A. (2008). Sense and stretchability: the role of titin and titin-associated 710 proteins in myocardial stress-sensing and mechanical dysfunction. Cardiovasc Res 77, 637-48. 711 Linke, W. A., Ivemeyer, M., Olivieri, N., Kolmerer, B., Ruegg, J. C. and Labeit, S. (1996). 712 Towards a molecular understanding of the elasticity of titin. J Mol Biol 261, 62-71. 713 Little, A. G. and Seebacher, F. (2014). Thyroid hormone regulates cardiac performance 714 during cold acclimation in zebrafish (Danio rerio). J Exp Biol 217, 718-25. 715 Liu, B., Wang, L. C. and Belke, D. D. (1993). Effects of temperature and pH on cardiac 716 myofilament Ca2+ sensitivity in rat and ground squirrel. Am J Physiol 264, R104-8. 717 Liu, B., Wohlfart, B. and Johansson, B. W. (1990). Effects of low temperature on 718 contraction in papillary muscles from rabbit, rat, and hedgehog. Cryobiology 27, 539-46. 719 Lopez-Olmeda, J. F. and Sanchez-Vazquez, F. J. (2011). Thermal biology of zebrafish 720 (Danio rerio). J Therm Biol 36, 91-104. 721 Lurman, G. J., Petersen, L. H. and Gamperl, A. K. (2012). In situ cardiac performance of 722 Atlantic cod (Gadus morhua) at cold temperatures: long-term acclimation, acute thermal 723 challenge and the role of adrenaline. J Exp Biol 215, 4006-14. 724 Marijianowski, M. M., Teeling, P., Mann, J. and Becker, A. E. (1995). Dilated 725 cardiomyopathy is associated with an increase in the type I/type III collagen ratio: a quantitative 726 assessment. J Am College Cardiol 25, 1263-72.

727 Mendonca, P. C., Genge, A. G., Deitch, E. J. and Gamperl, A. K. (2007). Mechanisms 728 responsible for the enhanced pumping capacity of the in situ winter flounder heart 729 (Pseudopleuronectes americanus). Am J Physiol 293, R2112-9. 730 Mutungi, G. and Ranatunga, K. W. (1998). Temperature-dependent changes in the 731 viscoelasticity of intact resting mammalian (rat) fast- and slow-twitch muscle fibres. J Physiol 732 508 ( Pt 1), 253-65. 733 Neagoe, C., Kulke, M., del Monte, F., Gwathmey, J. K., de Tombe, P. P., Hajjar, R. J. and 734 Linke, W. A. (2002). Titin isoform switch in ischemic human heart disease. Circulation 106, 1333- 735 41. 736 Patrick, S. M., Hoskins, A. C., Kentish, J. C., White, E., Shiels, H. A. and Cazorla, O. 737 (2010). Enhanced length-dependent Ca2+ activation in fish cardiomyocytes permits a large 738 operating range of sarcomere lengths. J Mol Cell Cardiol 48, 917-24. 739 Pauschinger, M., Knopf, D., Petschauer, S., Doerner, A., Poller, W., Schwimmbeck, P. L., 740 Kuhl, U. and Schultheiss, H. P. (1999). Dilated cardiomyopathy is associated with significant 741 changes in collagen type I/III ratio. Circulation 99, 2750-6. 742 Peng, J., Raddatz, K., Molkentin, J. D., Wu, Y., Labeit, S., Granzier, H. and Gotthardt, M. 743 (2007). Cardiac hypertrophy and reduced contractility in hearts deficient in the titin kinase 744 region. Circulation 115, 743-51. 745 Pieperhoff, S., Bennett, W. and Farrell, A. P. (2009). The intercellular organization of the 746 two muscular systems in the adult salmonid heart, the compact and the spongy myocardium. J 747 Anat 215, 536-47. 748 Saito, M., Takenouchi, Y., Kunisaki, N. and Kimura, S. (2001). Complete primary 749 structure of rainbow trout type I collagen consisting of alpha1(I)alpha2(I)alpha3(I) 750 heterotrimers. Eur J Biochem 268, 2817-27. 751 Sanchez-Quintana, D., Garcia-Martinez, V., Climent, V. and Hurle, J. M. (1995). 752 Morphological analysis of the fish heart ventricle: myocardial and connective tissue architecture 753 in teleost species. Ann Anat 177, 267-74.

754 Shaffer, J. F. and Gillis, T. E. (2010). Evolution of the regulatory control of vertebrate 755 striated muscle: the roles of troponin I and myosin binding protein-C. Physiol Genomics 42, 406- 756 19. 757 Shiels, H. A., Calaghan, S. C. and White, E. (2006a). The cellular basis for enhanced 758 volume-modulated cardiac output in fish hearts. Journal of General Physiology 128, 37-44. 759 Shiels, H. A., Di Maio, A., Thompson, S. and Block, B. A. (2011). Warm fish with cold 760 hearts: thermal plasticity of excitation-contraction coupling in bluefin tuna. Proc Royal Soc B 761 278, 18-27. 762 Shiels, H. A., Freund, E. V., Farrell, A. P. and Block, B. A. (1999). The sarcoplasmic 763 reticulum plays a major role in isometric contraction in atrial muscle of yellowfin tuna. J Exp Biol 764 202, 881-890. 765 Shiels, H. A., Galli, G. L. and Block, B. A. (2015). Cardiac function in an endothermic fish: 766 cellular mechanisms for overcoming acute thermal challenges during diving. Proc Biol Sci 282, 767 20141989. 768 Shiels, H. A., Paajanen, V. and Vornanen, M. (2006b). Sarcolemmal ion currents and 769 sarcoplasmic reticulum Ca2+ content in ventricular myocytes from the cold stenothermic fish, 770 the burbot (Lota lota). J Exp Biol 209, 3091-100. 771 Shiels, H. A., Vornanen, M. and Farrell, A. P. (2000). Temperature-dependence of L-type 772 Ca(2+) channel current in atrial myocytes from rainbow trout. J Exp Biol 203, 2771-80. 773 Shiels, H. A., Vornanen, M. and Farrell, A. P. (2002). Effects of temperature on 774 intracellular Ca2+ in trout atrial myocytes. J Exp Biol 205, 3641-50. 775 Shiels, H. A. and White, E. (2005). Temporal and spatial properties of cellular Ca2+ flux 776 in trout ventricular myocytes. Am J Physiol 288, R1756-66. 777 Shiels, H. A. and White, E. (2008). The Frank-Starling mechanism in vertebrate cardiac 778 myocytes. J Exp Biol 211, 2005-13. 779 Steinhausen, M. F., Sandblom, E., Eliason, E. J., Verhille, C. and Farrell, A. P. (2008). The 780 effect of acute temperature increases on the cardiorespiratory performance of resting and 781 swimming sockeye salmon (Oncorhynchus nerka). J Exp Biol 211, 3915-26.

782 Stephenson, D. G. and Williams, D. A. (1985). Temperature-dependent calcium 783 sensitivity changes in skinned muscle fibres of rat and toad. J Physiol 360, 1-12. 784 Sun, X., Hoage, T., Bai, P., Ding, Y., Chen, Z., Zhang, R., Huang, W., Jahangir, A., Paw, B., 785 Li, Y. G. et al. (2009). Cardiac hypertrophy involves both myocyte hypertrophy and hyperplasia 786 in anemic zebrafish. PLoS One 4, e6596. 787 Syme, D. A., Gamperl, A. K., Nash, G. W. and Rodnick, K. J. (2013). Increased ventricular 788 stiffness and decreased cardiac function in Atlantic cod (Gadus morhua) at high temperatures. 789 Am J Physiol 305, R864-76. 790 Trombitas, K., Wu, Y., Labeit, D., Labeit, S. and Granzier, H. (2001). Cardiac titin 791 isoforms are coexpressed in the half-sarcomere and extend independently. Am J Physiol 281, 792 H1793-9. 793 Vornanen, M. (1998). L-type Ca2+ current in fish cardiac myocytes: effects of thermal 794 acclimation and be a-adrenergic stimulation. J Exp Biol 201, 533-47. 795 Vornanen, M., Hassinen, M., Koskinen, H. and Krasnov, A. (2005a). Steady-state effects 796 of temperature acclimation on the transcriptome of the rainbow trout heart. Am J Physiol 289, 797 R1177-84. 798 Vornanen, M., Hassinen, M., Koskinen, H. and Krasnov, A. (2005b). Steady-state effects 799 of temperature acclimation on the transcriptome of the rainbow trout heart. Am J Physiol 289, 800 R1177-84. 801 Watanabe, K., Nair, P., Labeit, D., Kellermayer, M. S., Greaser, M., Labeit, S. and 802 Granzier, H. (2002). Molecular mechanics of cardiac titin's PEVK and N2B spring elements. J Biol 803 Chem 277, 11549-58. 804 Wu, Y., Cazorla, O., Labeit, D., Labeit, S. and Granzier, H. (2000). Changes in titin and 805 collagen underlie diastolic stiffness diversity of cardiac muscle. J Mol Cell Cardiol 32, 2151-62. 806 Zummo, G., Acierno, R., Agnisola, C. and Tota, B. (1995). The heart of the icefish: 807 bioconstruction and adaptation. Braz J Med Biol Res 28, 1265-76. 808 809 810

811 FIGURE LEGENDS 812 Figure 1. Comparison of the Ca2+ sensitivity of force generation by skinned ventricular fibers 813 isolated from hearts of: trout, frog, and rat, over a range of temperatures, while pH was

2+ 2+ 814 maintained at 7.0. Ca sensitivity was measured as the pCa at KF1/2, which is the Ca 815 concentration required to generate half-maximum force. When compared at the same 816 temperature, the trout preparations required 10-fold less Ca2+ to generate the same measure of 817 force than those from the mammalian species. Figure is adapted from Churcott et al. 1994 818 819 Figure 2. Changes in Ca2+ flux with acute temperature. (A) temperature reduces Ca2+ flux 820 through L-type Ca2+ channels, but (B) increases the duration of the action potential. Taken with 821 permission from Shiels et al. 2000 and Galli et al. 2008 822 823 Figure 3. Thermal remodelling of ventricular compliance in the rainbow trout. Ex vivo pressure 824 volume relationships (show increased passive stiffness of the whole ventricle following cold 825 acclimation (5 °C; blue) compared to controls (10 °C; green) and increased compliance following 826 warm acclimation (18 °C; red). Figure is adapted from Keen et al. 2016. 827 828 Figure 4. Cardiac contractile properties of trout acclimated to 4 oC, 11 oC and 17 oC. (A) Activity 829 of actomyosin Mg2+-ATPase isolated from ventricles measured at 17oC. The maximal activity was 830 higher in preparations from cold acclimated trout than those from warm acclimated trout. (B)

2+ o 831 Relative Ca -activated force generated by cardiac trabeculae measured at 15 C. pCa50 is pCa at 832 half-maximum force. Different superscript letters denote a significant difference between 833 values (P<0.05). (C) Pressure development by ventricles measured at 15 oC using a Langandorff 834 preparation. Circles indicate ventricular developed pressures while squares indicate diastolic 835 pressures. Developed pressures at balloon volumes greater then baseline were higher for the 4 836 oC acclimated (blue symbols) fish than those for the 11 oC (black symbols) and 17 oC (red

837 symbols) acclimated fish. Figures modified from Klaiman et al. 2011; and Klaiman et al. 2014. 838 839 Figure 5. Thermal remodelling of ventricular collagen in rainbow trout and zebrafish.

840 Representative bright-field (BF) and polarised light (POL) micrographs of control rainbow trout 841 (A) and zebrafish (B) ventricular tissue sections stained with picro-sirus red, which allows semi- 842 quantification of fibrillar collagen content. Cold acclimation causes an increase in ventricular 843 collagen content in rainbow trout (A), but a decrease in zebrafish ventricular collagen content 844 (D). Ventricular collagen content in rainbow trout is associated with an increased mRNA 845 expression of collagen promoting genes (5 °C; blue), compared to control (10 °C; green), while 846 warm acclimation (18 °C; red) is associated with an increase in mRNA expression of collagen 847 degrading genes (E). Following cold acclimation, zebrafish show an increase in mRNA expression 848 of collagen regulatory genes, suggesting increased collagen turnover (F). Figures modified from 849 Johnson et al. 2014 and Keen et al. 2016) 850 851 Figure 6. An overview of the integrated remodeling response of the rainbow trout ventricle in 852 response to prolonged cold temperature, across multiple levels of biological organization. 853 854 Figure 7. Thermal remodelling of the rainbow trout heart. A summary of the effects of chronic 855 cooling (5 °C) and chronic warming (18 °C) compared to control (10 °C) temperature. 1) 856 (Klaiman et al., 2011), 2) (Klaiman et al., 2014), 3) (Keen et al., 2016), 4) (Vornanen et al., 2005), 857 5) (Driedzic et al., 1996), 6) (Driedzic and Gesser, 1994).

31 Figure 1 Figure 2

B Figure 3 Results:15°C$$–$Force$GeneraLon$ Figure 4

100 1.0$1.0 Cold acclimated A Warm acclimated 0.8$ 80 0.8 Control

60 0.6 -ATPase acvity 0.6$ 2+ Mg 40 0.4

Pi/(min/mg protein) 0.4$ nmol ( 0.2$ 20 0.2 Actomyosin 17°C, pH = 7 0 Normalizedforcegeneration $$0$0.0 8.0 7.5 7.0 6.5 6.0 5.5 5.0 6.2$6.2 6.0$6.0 5.8$5.8 pCa 5.6$5.6 5.4$5.4 5.2$5.2 pCapCa$ pCa50 B 1.0 Cold acclimated 5.69 ± 0.03a pH$=$7,$SL$=$2.2$μm$ Warm acclimated 5.61 ± 0.01b Control 5.59 ±0.02b 0.8

0.6

0.4

0.2 Normalized force generaon

15°C, pH = 7, SL = 2.2 μm 0.0 6.2 6.0 5.8 5.6 5.4 5.2 pCa C 70

20 Ventricular Pressure (mmHg) 10

0 0 10 20 30 40 50 60 Balloon Volume (μl)

Figure 6

Gene Expression mRNA of muscle growth genes (Vornanen et al., 2005; Keen et al. 2016) mRNA of hypertrophic markers (Keen et al., 2016) An-sense strand mRNA of collagen promong genes (Keen et al., 2016) mRNA A G G A G G C mRNA of collagen degrading genes (Keen et al., 2016) C C T C C G Sense strand VEGF expression (Keen et al. 2016)

Myoﬁlaments AM-ATPase (Yang et al., 2000)

Δ Gene expression of 4 TnI isoforms (Alderman et al. 2012) and 2 cTnC isoforms (Genge et al. 2013)

Phosphorylaon of TnT (Klaiman et al. 2014)

Calcium Handling Rate of SR Ca2+ release/uptake (Keen et al., 1994; Aho and Vornanen 1998, 1999) SERCA transcript expression (Korajoki and Vornanen 2012)

β-adrenergic receptor density and sensivity (Graham and Farrell, 1989; Keen et al., 1993; Aho and Vornanen, 2001)

~ RyR density and localizaon (Birkedal et al., 2009)

Myocyte

Rate of contracon (intact muscle) (Aho and Vornanen 1999)

Refractoriness (Aho and Vornanen 1999)

Ca2+ sensivity of skinned trabeculae (Klaiman et al. 2014)

Whole Heart Heart size (Birkendal et al., 2009; Keen et al. 2016)

Atrium Sinus Connecve ssue content (Klaiman et al. 2014) venosus Bulbus Fibrillar collagen content (Keen et al. 2016) arteriosus Compact layer thickness (Klaiman et al. 2014)

Heart rate (Aho and Vornanen 2001)

Passive sﬀness (Keen et al. 2016)

Ventricle Magnitude and rate of developed pressure (Klaiman et al. 2014) Figure 7

Ventricle - spongy Ventricle - compact Myocyte bundle hypertrophy(1,2,3) Compact thickness(1,2,3) Extra-bundular sinus(1,3) Fibrillar and amorphous collagen(1,2,3) mRNA of muscle growth genes(3) Whole chamber passive stiffness(3) mRNA of hypertrophic markers(3) Micromechanical stiffness of tissue(3) (3) mRNA of collagen promoting genes Cellular lipid droplets(5)

mRNA of collagen degrading Cellular glycogen(6) genes(3) 17/18 °C Amorphous collagen(1,2,3) Whole chamber passive stiffness(3) Cellular lipid droplets(5) Cellular glycogen(6)

Blood to ventral aorta Bulbus arteriosus Ventricle (compact myocardium)

Ventricle Atrium (spongy myocardium) 10/11 °C

Sinus venosus Venous blood from cardinal and hepatic veins

Ventricle - spongy Ventricle - compact Myocyte bundle hypertrophy(1,2,3) Compact thickness(1,2,3)

Extra-bundular sinus(1,3) Fibrillar and amorphous collagen(3) (3,4) mRNA of muscle growth genes Whole chamber passive stiffness(3) mRNA of hypertrophic markers(3) Tissue micromechanical stiffness(3) (3) mRNA of collagen promoting genes Cellular lipid droplets(5) mRNA of collagen degrading genes(3) Cellular glycogen(6) 4/5 °C Amorphous collagen(1,2,3) Whole chamber passive stiffness(3) Tissue micromechanical stiffness(3) Cellular lipid droplets(5) Cellular glycogen(6)