The Pharmacology and Physiology of Orphan G Protein-Coupled Receptor

The pharmacology and physiology of orphan

G protein-coupled receptor, GPR37L1

James Leonard John Coleman

A thesis in fulfilment of the requirements for the degree of

Doctor of Philosophy

St Vincent’s Clinical School, Faculty of Medicine,

The University of New South Wales

Molecular Cardiology & Biophysics Division,

The Victor Chang Cardiac Research Institute

February 2018

Primary Supervisor: Dr Nicola J. Smith

Joint Supervisor: Prof. Robert M. Graham

iii

Abstract

G protein-coupled receptors (GPCRs) are integral membrane proteins, exquisitely evolved to allow the eukaryotic cell to detect and respond to ligands - discrete extracellular stimuli, commonly small molecules or peptides. The ~800 GPCRs in the human genome are not only fundamental to our physiology, but comprise a wealth of pharmacological targets in the treatment of disease. However, approximately 100 GPCRs are yet to be matched with endogenous ligands, and are termed orphans.

One such orphan, GPR37L1, was previously linked to blood pressure regulation, suggesting the pharmacological targeting of this receptor may represent a new strategy in the treatment of hypertension – a disease that is often refractory to existing treatments. Unfortunately, this previous finding was superficially assessed, and the cardiovascular effects of GPR37L1 remained unconfirmed. Further, fragments of the GPR37L1 N-terminus have been identified in human spinal fluid, suggesting the receptor undergoes novel post-translational modification. Accordingly, I tested two main hypotheses in this thesis: that GPR37L1 (i.) undergoes, and is regulated by, N-terminal proteolysis and (ii.) contributes to cardiovascular control. Using in vitro and ex vivo approaches, I report that GPR37L1 signal s G protein in absence of ligand, and is subject to metalloprotease-mediated Ns-terminal through truncation, the Gα which abrogates this activity. I next defined the receptor’s expression profile, using mouse tissue, and observed that expression of GPR37L1 was restricted to the central nervous system (CNS). I then investigated the cardiovascular phenotype of mice lacking GPR37L1 (GPR37L1KO/KO), compared to wild-type controls, and discovered intriguing sex differences. Blood pressure was increased in female, but not male, GPR37L1KO/KO mice.

However, when challenged with an angiotensin II infusion, only male GPR37L1KO/KO mice developed a heart failure-like phenotype. The experiments herein show GPR37L1 to be regulated by a novel metalloprotease-dependent mechanism, and that the receptor contributes to cardiovascular regulation in a sexually-dimorphic manner. Further, the CNS-restricted expression of GPR37L1 suggests it to be a hitherto unknown contributor to central cardiovascular control.

Table of Contents Preface ...... 1

Publications arising from this thesis ...... 1

Additional publications during completion of this thesis ...... 2

Presentations arising from this thesis ...... 3

Awards ...... 5

Contributions of other researchers ...... 6

Acknowledgements ...... 7

List of abbreviations ...... 10

List of figures ...... 13

List of tables ...... 14

List of appendices ...... 14

Chapter One: General Introduction ...... 15

1.1. G protein-coupled receptors ...... 16

1.2. GPCR activation and regulation ...... 17

1.3. The GPCR N-terminus ...... 24

1.4. Orphan GPCRs ...... 28

1.5. GPR37 and GPR37L1 ...... 29

1.6. The physiology of hypertension ...... 33

1.7. Physiological consequences of hypertension ...... 40

1.8. The renin-angiotensin system ...... 42

1.9. Treating hypertension ...... 47

1.10. Mouse models of hypertension and left ventricular hypertrophy ...... 49

1.11. Rationale and aims of PhD project ...... 53

Chapter Two: Methods ...... 55

2.1. Ethics statement ...... 56

2.2. Materials and reagents ...... 56

2.2.1. General reagents ...... 56

2.2.2. DNA constructs ...... 56

2.3. In vitro experiments and cell maintenance ...... 57

2.3.1. Stable HEK293 cells ...... 57

2.3.2. Primary astrocyte cultures ...... 59

2.3.3. Mouse cerebellar slice cultures ...... 61

2.3.4. Saccharomyces cerevisiae ...... 63

2.4. In vivo experiments ...... 65

2.4.1. Mouse line generation ...... 65

2.4.2. Mouse line maintenance ...... 67

2.4.3. Blood pressure measurement ...... 68

2.4.4. Angiotensin II infusion ...... 69

2.4.5. Tissue morphometry ...... 69

2.4.6. Microscopy and imaging ...... 69

2.5. Molecular biology ...... 72

2.5.1. Western blotting and in-gel imaging ...... 72

2.5.2. Verification of TIMP activity ...... 75

2.5.3. Quantitative PCR ...... 75

2.5.4. RNA sequencing ...... 76

2.6. Statistics and data analysis ...... 77

2.6.1. In vitro ...... 77

2.6.2. In vivo ...... 77

Chapter Three: GPR37L1 is a constitutively active orphan G protein-coupled receptor subject to regulation by metalloprotease-mediated N-terminal cleavage ...... 78

3.1. Introduction ...... 79

3.2. Results ...... 81

3.2.1. GPR37L1 is subject to N-terminal processing ...... 81

vii

3.2.2. The N-terminus of GPR37L1 is subject to metalloprotease-mediated cleavage ...... 85

3.2.3. Mutation of the GPR37L1 N-terminus does not abrogate proteolysis ...... 89

3.2.4. s in Saccharomyces cerevisiae ...... 92

3.2.5. ProsaptideGPR37L1 constitutively does not activate couples GPR37L1 to chimeric ...... Gα ...... 95

3.2.6. Cerebrospinal fluid peptides do not activate GPR37L1 ...... 97

3.2.7. GPR37L1 predominates in vivo as a proteolytically processed species ...... 99

3.2.8. GPR37L1 is not expressed in primary cortical astrocyte cultures ...... 101

3.2.9. GPR37L1 is processed by a metalloprotease in organotypic cerebellar cultures .. 103

3.2.10. GPR37L1 increases cAMP in organotypic mouse cerebellar cultures ...... 104

3.3. Discussion ...... 106

Chapter Four: Characterising GPR37L1 tissue expression ...... 114

4.1. Introduction ...... 115

4.2. Results ...... 117

4.2.1. -galactosidase reporter mice ...... 117

4.2.2. Generation of GPR37L1 knockout and β -galactosidase reporter

GPR37L1gene expression is abundant ...... in the cerebellum...... as assessed...... by β ...... 120 4.2.3. -

GPR37L1galactosidase is widely reporter expressed gene expression in the central ...... nervous system...... as assessed...... by β 123 4.2.4. GPR37L1 protein is undetectable in the mouse heart ...... 127

4.2.5. GPR37L1 protein is undetectable in the mouse kidney ...... 130

4.2.6. GPR37L1 transcription is efficiently abrogated in EUCOMM-derived knockout mice ...... 133

4.3. Discussion ...... 136

Chapter Five: The cardiovascular physiology of GPR37L1 ...... 140

5.1. Introduction ...... 141

5.2. Results ...... 143

5.2.1. GPR37L1 regulates baseline blood pressure in female but not male mice...... 143

viii

5.2.2. GPR37L1 protects against AngII-induced cardiovascular stress in male mice ...... 148

5.2.3. GPR37L1 deletion protects against AngII-induced cardiac fibrosis ...... 157

5.2.4. GPR37L1 abundance does not differ between sexes ...... 160

5.3. Discussion ...... 163

Chapter Six: General Discussion ...... 168

6.1. Major findings of this PhD project ...... 171

6.1.1. s is the cognate G protein for GPR37L1 signalling ...... 171

6.1.2. GαGPR37L1 is subject to metalloprotease-mediated N-terminal cleavage ...... 173

6.1.3. Constitutive activity of GPR37L1 is regulated by the N-terminus ...... 175

6.1.4. Cleaved GPR37L1 is the predominant species in mouse cerebellum ...... 177

6.1.5. The GPR37L1 protein is abundant in the CNS but not detectable in the heart and kidney ...... 180

6.1.6. Deletion of GPR37L1 increases blood pressure in female mice ...... 181

6.1.7. Male mice lacking GPR37L1 have an exacerbated response to AngII-induced hypertension ...... 185

6.2. Conclusion ...... 189

Chapter Seven: Appendices ...... 190

Chapter Eight: Bibliography ...... 194

Preface

Publications arising from this thesis Coleman JLJ, Mouat MA, Wu J, Jancovski N, Bassi JK, Chan AY, Humphreys DT, Mrad N, Yu ZY,

Ngo T, Iismaa S, dos Remedios CG, Feneley MP, Allen AM, Graham RM, Smith NJ (2018) "Orphan receptor GPR37L1 contributes to the sexual dimorphism of central cardiovascular control". Biology of Sex Differences, 9 (1), 14.

Coleman JLJ, Ngo T, Smith NJ (2017) “The G protein-coupled receptor N-terminus and receptor signalling: N-tering a new era”. Cellular signalling, 33: 1-9.

Coleman, JLJ*, Ngo T*, Schmidt J, Mrad N, Liew CK, Graham RM, Smith NJ (2016)

“Metalloprotease cleavage of the N terminus of orphan G protein-coupled receptor GPR37L1 reduces its constitutive activity”. Science Signaling. 9(43):ra36 (with accompanying editorial, * indicates shared first authorship).

Coleman JLJ, Brennan K, Ngo T, Balaji P, Graham RM, Smith NJ (2015) “Rapid knockout and reporter mouse line generation and breeding colony establishment using EUCOMM conditional- ready embryonic stem cells: A case study”. Frontiers in Endocrinology. 6:105.

Ngo T, Coleman JLJ, Smith NJ (2015) “Using constitutive activity to define appropriate high- throughput screening assays for orphan G protein-coupled receptors”. Methods in Molecular

Biology. 1272: 91-106.

Additional publications during completion of this thesis Mouat MA, Coleman JLJ, Smith NJ (in press) “GPCRs in context: sexual dimorphism in the cardiovascular system”. British Journal of Pharmacology.

Ngo T, Ilatovskiy AV, Stewart AG, Coleman JLJ, McRobb FM, Riek RP, Graham RM, Abagyan R,

Kufareva I, Smith NJ (2017) “Orphan receptor ligand discovery by pickpocketing pharmacological neighbours”. Nature Chemical Biology. 13, 235-242.

Chowdhury MK, Wu LE, Coleman JLJ, Smith NJ, Morris MJ, Shepherd PR, Smith GC (2016)

-catenin in primary rat hepatocytes via“Niclosamide PKA signalling”. blocks Biochemical glucagon phosphorylation Journal. May 1;473(9):1247 of Ser552 on-55. β

Ngo T, Kufareva I, Coleman JLJ, Graham RM, Abagyan R, Smith NJ (2016) “Identifying ligands at orphan GPCRs: current status using structure-based approaches”. British Journal of Pharmacology.

173(20), 2934-2954.

Presentations arising from this thesis Coleman JLJ. (2017) “Orphan receptor GPR37L1: Home is not where the heart is”. Cincinnati

Children’s Hospital, Ohio, USA (Invited seminar).

Coleman JLJ, Mouat MA, Wu J, Jancovski N, Stefani M, dos Remedios CG, Feneley MP, Allen AM,

Graham RM, Smith NJ (2016) “GPR37L1: an orphan G protein-coupled receptor contributing to sex differences in blood pressure regulation”. ASCEPT-MPGPCR Scientific Meeting, Melbourne,

Australia (Oral presentation, Garth McQueen Prize Winner).

Coleman JLJ, Mouat MA, Wu J, Jancovski N, Humphreys DT, Stefani M, dos Remedios CG, Feneley

MP, Allen AM, Graham RM, Smith NJ (2016) “Blood pressure is increased in female mice lacking orphan G protein-coupled receptor GPR37L1”. American Heart Association Scientific Sessions, New

Orleans, USA (Poster presentation).

Coleman JLJ, Mouat MA, Wu J, Jancovski N, Stefani M, dos Remedios CG, Feneley MP, Allen AM,

Graham RM, Smith NJ. (2016) “Orphan G protein-coupled receptor GPR37L1 regulates blood pressure in female mice”. St Vincent’s Campus Research Symposium, Darlinghurst, Australia (Oral presentation).

Coleman JLJ, Ngo T, Holman S, Jones NM, Graham RM, Smith NJ (2015) “Cardiovascular orphan receptor GPR37L1 predominates in vivo as an inactivated product of metalloprotease-dependent N- terminal truncation”. St Vincent’s Campus Research Symposium, Darlinghurst, Australia (Poster presentation).

Coleman JLJ, Ngo T, Schmidt J, Mrad N, Liew CK, Graham RM, Smith NJ (2014) “Cardiovascular orphan receptor GPR37L1 predominates in vivo as an inactivated product of metalloprotease- dependent N-terminal truncation”. St Vincent’s Campus Research Symposium, Darlinghurst,

Australia (Oral presentation, People’s Choice Prize Winner).

Coleman JLJ, Ngo T, Schmidt J, Mrad N, Liew CK, Graham RM, Smith NJ (2014) “GPR37L1: an orphan G protein-coupled receptor with constitutive activity modulated by novel metalloprotease- mediated N-terminal cleavage”. ASCEPT-MPGPCR Scientific Meeting, Melbourne, Australia (Oral presentation).

Coleman JLJ, Ngo T, Schmidt J, Mrad N, Liew CK, Graham RM, Smith NJ (2014) “Constitutive activity of orphan G protein-coupled receptor, GPR37L1, is modulated by novel metalloprotease- mediated N-terminal cleavage”. UNSW Drug Discovery and Medicinal Chemistry Symposium,

Darlinghurst, Australia (Poster presentation).

Coleman JLJ, Ngo T, Schmidt J, Mrad N, Liew CK, Graham RM, Smith NJ (2014) “GPR37L1: an orphan G protein-coupled receptor with constitutive activity modulated by novel metalloprotease- mediated N-terminal cleavage”. Victor Chang Cardiac Research Institute Symposium, Darlinghurst,

Australia (Poster presentation).

Awards People’s Choice Prize for best oral presentation (2016), Paul Korner Seminar Series, Victor

Chang Cardiac Research Institute, Darlinghurst, Australia.

Garth McQueen Prize for best oral presentation by postgraduate student (2016), ASCEPT-

MPGPCR Scientific Meeting, Melbourne, Australia.

People’s Choice Prize for best Fast Forward oral presentation (2014), 22nd St. Vincent's

Campus Research Symposium, Darlinghurst, Australia.

Australian Postgraduate Award (2014-2017), University of New South Wales, Sydney, Australia.

Simon and Michal Wilkenfeld PhD Top-Up Scholarship (2014-2016), Victor Chang Cardiac

Research Institute, Darlinghurst, Australia.

Contributions of other researchers Ms Nadine Mrad and Drs Nicola J. Smith, Chu Kong Liew and Johannes Schmidt generated DNA constructs as indicated in Chapter 2.

Dr Nicola J. Smith performed experiments in Figures 3.1 (panels B, C & D), 3.7 and 4.5 (panel E).

Dr Tony Ngo performed experiments on the following GPR37L1 mutants in Figure 3.3: 115Ala117,

118Ala120, 121Ala123.

Mrs Jaspreet Bassi performed all experiments in Figure 4.4.

All blood pressure measurement by pressure-sensing catheter (Chapter 5) was performed by Dr

Jianxin Wu and analysed by myself and Ms Margaret A. Mouat. All blood pressure measurement by radiotelemetry (Chapter 5) was performed by Dr Nikola Jancovski and analysed by Dr Jancovski and myself.

Ms Margaret A. Mouat assisted with tissue collection and analysis of blood pressure data throughout angiotensin II studies, as well as performing cardiomyocyte counting in Figure 5.2, all experiments in Figure 5.3, and staining and quantification of fibrosis in Figure 5.5.

Dr David T. Humphreys performed RNA sequencing and analysis of Genotype-Tissue Expression consortium data in Figure 5.6 (panel D).

I wish to offer my sincere gratitude to all researchers above for their contributions to the experiments that made possible the completion of this thesis.

Acknowledgements In the lead up to the completion of this thesis, I happened upon a video of famed athlete, actor, and politician, Arnold Schwarzenegger, addressing the 2017 graduating class of the University Of

Houston. He told them:

“Now, on your diplomas, there will only be one name on it, and this is yours.

But I hope that doesn’t confuse you and that you think, maybe, that you made it that far by

yourself.

No, you didn’t. It took a lot of help.”

These words stayed in my mind, as they boldly put what I already knew: this thesis carries my name, but represents the efforts of many names more, and it would not have been completed without the support of my mentors, colleagues, friends and family. So without further delay, I wish to begin the intimidating task of thanking them all!

Firstly, to my Mum and Dad, Rosie and John: it is never far out of my mind, just how hard you both have always worked to ensure I’ve been able to receive the best education possible, throughout all my years. Not only in providing me access to a fantastic school and university, but in investing your own time into raising me with the skills I needed to succeed in both environments.

Further, I am beyond grateful for your personal support at home during my PhD studies, which enabled me to focus my attention on my research, and to minimise the stresses of my working life.

Mum, if there is a part of me that can brag about working with attention to detail and with pride, I owe that to you. Dad, you taught me to ask challenging questions and to think critically about their answers, and gifted me with the sense of enquiry through which I see the world every day. I will always be grateful to you both for your influence and enduring love and support.

To my dearest friends, I thank you for always being there to share in some laughter and good times. Whilst I wanted to avoid naming names (for fear of leaving anyone out!), it would be remiss of me not to mention my most enduring friends: Dave, Dawesie, Joe, Josh and Tristan. All the karaoke nights, guitar jams and gym sessions have been essential to me finishing this thesis with my sanity intact (well, as much as it ever was).

To my wonderful partner, Tracey: I’m aware that as my PhD has progressed, I’ve been increasingly distracted with thoughts of work and less fun to be around. Yet, you’ve looked past this, with support and understanding. For that, I am grateful. The thought of seeing you next has often been my ‘light at the end of the tunnel’ of a difficult working week, and I am fortunate to have had you by my side, ready to lift my spirits and put a smile back on my face.

To the many members of the Victor Chang Cardiac Research Institute (VCCRI) community, I thank you all. More specifically, thank you to my fellow students, for your friendship and inspiration, and thank you to our BioCORE staff, for ensuring the highest standard of care for the animals involved in experiments described in this thesis. To all the staff at the VCCRI (and not just the scientific staff), who are only too willing to lend a helping hand, I have appreciated all the assistance you’ve given me, no matter how minor my request may have been. Further, the VCCRI could not exist without the kind contributions of its many benefactors, and I specifically wish to thank Simon and Michal Wilkenfeld, for their generous contribution of a PhD scholarship.

To my colleagues in the Graham and Smith laboratories, both current and former: whether it has been lending me your reagents or your ears, your technical expertise or some sage advice, I am indebted. The successes that I have found working alongside you all stand to demonstrate how remarkable you all are as scientists and friends. I wish to offer further gratitude to Ms Meg Mouat and Dr Jianxin Wu for their substantial contributions to the experiments that made possible the completion of this thesis.

To my friend and fellow Smith laboratory PhD candidate-turned-alumni, Dr Tony Ngo: I have so much to thank you for. Whether it be for teaching me that structural biology is pretty cool after all, laughing in my face when I get too excited about n=1 experiments, or simply being there to share some beverages and a debrief at the Green Park, I consider myself immensely fortunate to have had a friend like you alongside for the ups and downs of the past few years.

To my joint supervisor, Professor Bob Graham; thank you for welcoming me into your lab and for being a truly involved supervisor and mentor. Your willingness to devote significant time to helping even the most junior members of your team (myself included), in the face of your overwhelming schedule, speaks volumes of your genuine investment in their development and education. Further, your readiness to draw from all disciplines of knowledge (even those outside of science!) is very inspiring, and I’m certain every budding academic would do well to follow your example.

And finally, having saved my greatest ‘thanks’ for last, it remains for me to acknowledge my primary supervisor, Dr Nicola Smith. Nic, I met you at a time in my life when I had not touched a pipette for over two years and would have been hard-pressed to define ‘G protein-coupled receptor’. For whatever mad reason, you took me on as your student, and it is no overstatement to say that this was a life-changing decision for me. You’ve given me the opportunity to spend my days doing work that I’m truly proud of and excited by. You've encouraged me to step up to new challenges, to strive for ambitious goals, and to have a great time whilst doing so. This PhD has been one of the most happy and fulfilling times of my life, and the greatest share of the credit for that belongs to you.

So to everyone who’s shared this journey with me, I cannot put it any better than these simple words: thank you!

List of abbreviations Abbreviation Definition 1K1C One kidney, one clip 2K1C Two kidney, one clip -SkA -skeletal actin ACE Angiotensin-converting enzyme ACTHα Adrenocorticotropicα hormone ADAM A disintegrin and metalloprotease ADH Antidiuretic hormone AFU Arbitrary fluorescence units AGT Angiotensinogen AngI Angiotensin I AngII Angiotensin II AngIII Angiotensin III AngIV Angiotensin IV ANP Atrial natriuretic peptide APMA p-aminophenylmercuric acid ARB Angiotensin receptor blocker -2 microglobulin -gal -galactosidase β2M-MHC β BBβ β-blockers BNPβ Brainβ myosin natriuretic heavy chain peptide BPM Beatsβ per minute BSA Bovine serum albumin cAMP Cyclic adenosine monophosphate CCB Calcium channel blocker CMV Human cytomegalovirus CNS Central nervous system CPM Counts per million CRE cAMP response element cDNA Complementary deoxyribonucleic acid cRNA Complementary ribonucleic acid CSF Cerebrospinal fluid CVD Cardiovascular disease CVLM Caudal ventrolateral medulla DABCO 1,4-diazabicyclo[2.2.2]octane DAG Diacylglycerol DAPI Diamidino-2-phenylindole dihydrochloride DHP Dihydropyridines DMEM Dulbecco’s modified Eagle’s medium DNA Deoxyribonucleic acid DOCA Deoxycorticosterone acetate DP Diastolic pressure EDP End-diastolic pressure EDTA Ethylenediaminetetraacetic acid EGTA Ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid EndoH Endoglycosidase H ELISA Enzyme-linked immunosorbent assay

ER Endoplasmic reticulum ERK Extracellular-signal regulated kinases ES Embryonic stem (cells) ES2M Embryonic Stem Cell-to-Mouse initiative (Monash University) EUCOMM European conditional mouse mutagenesis program eYFP Enhanced yellow fluorescent protein FBS Fetal bovine serum FDG Fluorescein di- -galactopyranoside Flp-In Flp-In™ T-REx™ cells FlpE Enhanced Saccharomycesβ cerevisiae FLP1 recombinase FRT Flippase recognition target GAPDH Glyceraldehyde 3-phosphate dehydrogenase GDP Guanosine diphosphate GEF Guanine nucleotide exchange factor GFAP Glial fibrillary acidic protein GFP Green fluorescent protein GPCR G protein-coupled receptor GPR37 G protein-coupled receptor 37 GPR37L1 GPR37-like 1 receptor GRK G protein-coupled receptor kinase GTEx Genotype-tissue expression consortium GTP Guanosine triphosphate HBSS Hank’s balanced salt solution HEK Human embryonic kidney HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) HPA Hypothalamic pituitary axis HR Heart rate HRP Horseradish peroxidase HW Heart weight IBMX 3-isobutyl-1-methylxanthine IKMC International knockout mouse consortium IRES Internal ribosome entry site kD Kilodaltons KO Knockout lacZ -galactosidase LV Left ventricle LVH LeftGene ventricular encoding β hypertrophy LVW Left ventricle weight MAP Mean arterial pressure mmHg Millimetres of mercury MMP Matrix metalloprotease mRNA Messenger ribonucleic acid miRNA Micro ribonucleic acid NADH Nicotinamide adenine dinucleotide phosphate NMDA N-methyl-D-aspartate NMDAR N-methyl-D-aspartate receptor NTS Nucleus tractus solitarius PAGE Polyacrylamide gel electrophoresis PAR Protease-activated receptor PBN Parabrachial nuclei

PBS Phosphate-buffered saline PCR Polymerase Chain Reaction PFA Paraformaldehyde PKC Protein kinase C PLC- Phospholipase C- PNGaseF Peptide N-glycosidase F PP β Pulse pressure β Pre-GM Preganglionic sympathetic motor neurons PVDF Polyvinylidene difluoride qPCR Quantitative polymerase chain reaction RAS Renin angiotensin system REN Renin RGS Regulator of G protein signalling RIPA Radio-immunoprecipitation buffer RNA Ribonucleic acid RT Room temperature RVLM Rostral ventrolateral medulla RVW Right ventricle weight SC Subcutaneous SDS Sodium dodecyl sulfate SG Sympathetic ganglia Shh Sonic hedgehog protein siRNA Small interfering ribonucleic acid SNS Sympathetic nervous system SRE Serum response element SRP Signal recognition particle SP Systolic pressure TAC Thoracic aortic constriction TAPI Tumour necrosis factor- TBS-T Tris-buffered saline with 0.1% Tween 20. TFB-TBOA (3S)-3-[[3-[[4-(Trifluoromethyl)benzoyl]amino]phenyl]methoxy]α protease inhibitor -L-aspartic acid TH Tyrosine hydroxylase TIMP Tissue inhibitor of metalloprotease TL Tibia length TM Transmembrane WGA Wheat germ agglutinin WT Wild-type YNB Yeast nitrogen base YPD Yeast extract-peptone-dextrose

List of figures

Figure 1.1: The extended ternary complex model of GPCR activation ...... 19

Figure 1.2: Schematic of heterotrimeric G protein-mediated signalling in response to GPCR activation ...... 22

Figure 1.3: Predicted secondary structure of human GPR37L1 and annotated features ...... 32

Figure 1.4: Schematic of key central nervous system pathways of cardiovascular control ...... 37

Figure 1.5: Schematic of the classical renin-angiotensin system ...... 45

Figure 3.1: GPR37L1 is expressed in HEK293 cells as both a full-length and N-terminally truncated isoform ...... 83

Figure 3.2: GPR37L1 is subject to N-terminal proteolysis by an ADAM ...... 87

Figure 3.3: N-terminal proteolysis of GPR37L1 is not interrupted by alanine mutagenesis ...... 91

Figure 3.4: The GPR37L1 N-terminus is necessary for constitutive coupling to chimeric Gαs G proteins ...... 94

Figure 3.5: Prosaptide (TX14A) does not stimulate GPR37L1 signalling ...... 96

Figure 3.6: N-terminal peptides to not modify GPR37L1 Gαs signalling in yeast ...... 98

Figure 3.7: GPR37L1 exists as two species in the mouse cerebellum ...... 100

Figure 3.8: Primary cortical astrocyte cultures do not express GPR37L1 ...... 102

Figure 3.9: GPR37L1 stimulates cAMP synthesis and is cleaved by a metalloprotease in the ex vivo cerebellum ...... 105

Figure 3.10: Schematic contrasting protease-activated and inactivated receptors ...... 113

Figure 4.1: EUCOMM targeting vector design and alleles resulting from crosses with Cre or Flp deleter mice ...... 118

Figure 4.2: Expression profile of β-galactosidase in EUCOMM-derived reporter mouse cerebellum .... 121

Figure 4.3: Expression profile of β-galactosidase in EUCOMM-derived reporter mouse brain ...... 124

Figure 4.4: Expression of β-galactosidase in central nervous system cardiovascular control regions in EUCOMM-derived reporter mouse ...... 126

Figure 4.5: GPR37L1 protein was not detected in the heart ...... 128

Figure 4.6: GPR37L1 protein was not detected in the kidney ...... 131

Figure 4.7: Cerebellar GPR37L1 transcript and protein is efficiently ablated in knockout mice ...... 134

Figure 5.1: Female GPR37L1 knockout mice have elevated blood pressure ...... 145

Figure 5.2: AngII infusion decreased cardiomyocyte density ...... 152

Figure 5.3: Activation of the cardiac fetal gene program in vehicle- and AngII-treated mice as determined by quantitative PCR analysis ...... 153

Figure 5.4: Male GPR37L1KO/KO mice develop pulmonary edema in response to AngII infusion ...... 156

Figure 5.5: GPR37L1KO/KO mice are protected from AngII-induced cardiac fibrosis ...... 158

Figure 5.6: Comparison of male and female GPR37L1 expression in mouse and human ...... 161

List of tables

Table 1.1: Summary of the diverse modalities by which the GPCR N-terminus can directly and indirectly contribute to receptor signalling ...... 27

Table 4.1: Breeding statistics for the first generation of each EUCOMM line ...... 119

Table 5.1: Blood pressure by micromanometry in GPR37L1 knockout mice ...... 147

Table 5.2: Hemodynamic assessment of AngII-treated (2 mg/kg/day) female GPR37L1wt/wt and GPR37L1KO/KO mice by micromanometry under anaesthesia ...... 150

Table 5.3: Hemodynamic assessment of AngII-treated (2 mg/kg/day) male GPR37L1wt/wt and GPR37L1KO/KO mice by micromanometry under anaesthesia ...... 151

List of appendices

Table S1: Expanded statistical analysis (two-way ANOVA) of haemodynamics, morphometry and histology of vehicle- and AngII-treated, GPR37L1wt/wt and GPR37L1KO/KO mice, showing F statistic and P value of interaction, genotype and AngII effects……………………………………………………………………191

Table S2: Expanded statistical analysis (two-way ANOVA) of Figure 5.3 showing F statistic and P value of interaction, genotype and AngII effects on ANP, BNP, β-MHC and α-SkA mRNA abundance as determined by qPCR……...………………………………………………………………...……………………….193

Chapter One: General Introduction

1.1. G protein-coupled receptors With over 800 unique members, G protein-coupled receptors (GPCRs) are the largest protein superfamily in the human genome (Katritch et al., 2013) and are paramount to the eukaryotic cell’s ability to mount an intracellular response to extracellular stimuli (Coleman et al., 2017). Crucial to the ability of GPCRs to transform extracellular stimulation into an intracellular signal is their characteristic structure – a GPCR comprising of a series of seven membrane- -helices, flanked by an extracellular N-terminus and an intracellular C-terminus, and joinedtraversing by alternating α intracellular and extracellular loops (Rosenbaum et al., 2009). Each GPCR serves to detect one or more cognate biological stimuli that may include photons, ions and molecular stimuli referred to as ligands, which include odorants, nucleotides, peptides and proteins. Accordingly, GPCRs may well have been essential to the evolution and diversification of multicellular life itself (de Mendoza et al.,

2014).

Despite sharing common mechanisms of activation and modulation, which will be discussed in more detail below, there exists appreciable diversity within the GPCR family. Of the 826 human

GPCRs, 719 belong to the rhodopsin family (of which 422 are olfactory receptors), 36 belong to the frizzled/taste 2 family, 33 belong to the adhesion family, 22 belong to the glutamate family and 16 belong to the secretin family (Fredriksson et al., 2003; Lv et al., 2016). Some key features that distinguish the families include the following: (i.) rhodopsin; a modest-length N-terminus and transmembrane-based ligand-binding site, (ii.) secretin; a larger N-terminus that forms part of the receptor’s ligand-contacting surfaces and confers ligand selectivity, (iii.) glutamate; obligate dimers with an even larger N-terminus that contains the entirety of the receptor’s ligand binding domain

(iv.) adhesion; atypical GPCRs with vast N-termini containing adhesion domains and being subject to autoproteolysis and (v.) frizzled; GPCRs that are distantly related to the other classes but are highly evolutionarily conserved and carry long N-termini for ligand interaction (Schioth et al., 2005;

Lee et al., 2015). The sheer size of the GPCR family indicates their fundamental importance in so

16 many physiological processes, not limited to those of the central nervous (Krishnan et al., 2015), cardiovascular (Foster et al., 2015), gastrointestinal (Reimann et al., 2012), immune (Zlotnik et al.,

2000) and skeletomuscular (Jean-Baptiste et al., 2005) systems. Owing to both their effectiveness as physiological regulators and their specificity for detection of discrete extracellular stimuli, GPCRs have proved effective targets in the pharmaceutical management of health, with approximately

30% of approved small molecule and biologic drugs targeting this remarkable family (Santos et al.,

2017; Sriram et al., 2018).

1.2. GPCR activation and regulation Whilst the mechanisms of GPCR activation have long been debated, the most favourable in terms of free energy is the “multi-state” model (Kenakin, 1995; Christopoulos et al., 2002): a given

GPCR exists in a constant state of isomerisation between multiple conformations, some of which are favourable for binding by its cognate ligand. Whilst research reveals an ever-increasing number of such possible GPCR states, for the purposes of this thesis, the “extended-ternary complex” model

(Samama et al., 1993) is sufficient in complexity, and is illustrated in Figure 1.1. A ligand that binds and stabilises an active GPCR conformation is termed an agonist, permitting the receptor’s intracellular domain to associate with and modulate signal-transducing heterotrimeric G proteins.

Of particular relevance to this thesis, the multi-state model of GPCR activation permits the existence of activated receptors in the absence of their ligand (described by factors J and M in Figure 1.1), or constitutive activity (Lefkowitz et al., 1993). That is to say a GPCR does not necessarily require a ligand to adopt an active conformation. This constitutive activity also permits the existence of inverse agonists, which preferentially associate with inactive states of the GPCR and thus, inhibit the baseline signal of constitutively active receptors (Katritch et al., 2013). This contrasts with neutral antagonists, which prevent agonist-induced activation but without affecting the unstimulated receptor equilibrium (Katritch et al., 2013). Further, the agonists, antagonists and inverse antagonists described above are orthosteric modulators, having their effects on the GPCR

17 via interacting with the same topological site as the receptor’s endogenous ligand (Wootten et al.,

2013). Additional complexity is provided by allosteric modulators, which alter GPCR activity by association with the receptor at a site topographically distinct from that of the orthosteric modulators. Fascinatingly, allosteric modulators can also affect the activity mediated by an orthosteric ligand (May et al., 2007; Wootten et al., 2013).

Figure 1.1: The extended ternary complex model of GPCR activation

The extended ternary complex model of GPCR activation as described by, and adapted from,

Samama et al. (1993). A, agonist; R, receptor; R*, activated receptor; E, effector (i.e. heterotrimeric G protein). Constants K and M are independent and describe the reaction of agonist binding to receptor, and effector binding to receptor, respectively. J describes spontaneous isomerisation of

ofthe agonist receptor and to receptor the active affects state that (R to of R*)effector in the absence of agonist. α describes how the association which the receptor activation (R to R*) conversionand receptor,is influenced and byvice agonist versa. binding.β describes the extent to

Even more diverse than the modulators of GPCR activity is the nature of that activity itself. As the name suggests, GPCRs classically exert their effects through coupling to heterotrimeric G proteins – a family that, in humans, is derived from 35 genes (Milligan et al., 2006). These effectors

respectivelyfunction as trimers (Milligan of anet al.α, ,β 2006 and )γ. Thesubunit, designation encoded of by these 16, 5 effectors and 14 different as G proteins, genes abbreviated in humans, from guanine nucleotide-binding proteins, reflects their function as guanosine triphosphate (GTP) binding and hydrolysing proteins. The GPCR serves as a guanine nucleotide exchange factor (GEF)

releasepromoting its bounda conformational guanosine diphosphatechange in the (GDP), α subunit in exchange of the G proteinfor cellular heterotrimer GTP (Oldham that et allows al., 2008 it to) .

This, in turn, induces(Milligan a conformational et al., 2006) change that triggers dissociation of the Gα subunit from the aidedGβ/γ complexby regulators of G protein signalling. The Gα (RGS) subunit proteins hydrolyses which theenhance terminal its GTPase phosphate-activity. of the The GTP, resul

ting GDP remains bound to the Gα subunit, (whichOldham is etreturned al., 2008 to) .a A conformation key tenet of such permissive G protein of - heterotrimer formation with the Gβ/γ subunits proteincoupled subunits, signalling which, is that in each turn, GPCR herald favours different interaction intracellular with oneresponses or more through of the different modulation Gα of effectors that include metabolic enzymes, channels and transporters, and transcriptional machinery

(Figure 1.2) (Neves et al., 2002). These effects may be opposing, with the GTP- s protein increasing intracellular cyclic adenosine monophosphate (cAMP) concentrationsbound through Gα stimulation of adenylate cyclase, whilst GTP- i proteins inhibit adenylate cyclase and, thus, oppose increases in cAMP (Neves et al., 2002bound). Of importance Gα to this thesis, defining the cognate G protein(s) to which a GPCR may couple is necessary for understanding the intracellular consequences of that receptor’s activation. Whilst once considered a passive component in G protein si

gnalling, the Gβ/γ portion of the G protein [which does not commonly dissociate further

(Neves et al., 2002)] is now appreciated as a potent effector in its own right, modulating targets including ion channels, adenylate cyclases and protein kinases (Figure 1.2) (Milligan et al., 2006).

In concert with the machinery that permits GPCRs to modulate intracellular signalling are feedback processes that desensitise the cell to GPCR stimulation following acute or chronic agonism, thus protecting the cell from overstimulation (Ferguson, 2001). GPCR desensitisation can refer to processes that enhance receptor internalisation from the cell surface, enhance receptor degradation, mitigate receptor synthesis, or those that limit the receptor’s ability to couple to G proteins, but has previously been used to refer only to the latter, in order to differentiate mechanisms that alter receptor abundance from those that affect receptor responsiveness

(Ferguson, 2001). Strict desensitisation in this manner occurs through steric obstruction of G protein-coupling by recruitment of arrestin proteins to the GPCR, which is an association greatly enhanced by phosphorylation of intracellular loop and C-terminal domains by second-messenger dependent kinases, but more so by GPCR kinases (GRK) (Ferguson, 2001). In addition to steric hindrance of G protein-coupling, arrestins also target GPCRs for endocytosis by recruitment to clathrin-coated pits, allowing the now internalised GPCRs to be dephosphorylated by endosomal phosphatases and recycled to the surface, or labelled with ubiquitin and flagged for lysosomal degradation (Bohm et al., 1997). Of relevance to this thesis, not all receptors can be recycled, with protease-activated receptors being partially degraded by virtue of their activation (Drake et al.,

2006). It should also be noted that some GPCRs have been shown to sustain signalling once

-arrestin-dependent pathways (Drake et al., 2006; Thomsen et al., 2016). internalised, through β

Figure 1.2: Schematic of heterotrimeric G protein-mediated signalling in response to

GPCR activation

Classically, a GPCR is activated following binding by its cognate ligand, enabling the receptor to serve as a guanine nucleotide exchange factor, facilitating the exchange of guanosine diphosphate

(GDP) for guanosine triphosphate (GTP) on the α subunit of the heterotrimeric G protein (α, β, γ). 22

The activated

Gα may modulate a variety of signalling pathways. Different GPCRs are coupled with biologicaldifferent Gα responses. species, mediatingAdapted from the diverseMarinissen intracellular et al. (2001) effects and ofDorsam GPCR activationet al. (2007) and. Ca subsequent2+, calcium; cAMP, cyclic adenosine monophosphate; DAG, diacylglycerol, GEF, guanine nucleotide exchange factor; PLC- -

β, phospholipase C β; PKC, protein kinase C.

1.3. The GPCR N-terminus In order to place a key finding arising from this thesis into context, one aspect of the GPCR structure, the extracellular N-terminus, must be explored in more depth. The significance of the

GPCR N-terminus is usually defined in the context of receptor trafficking, from the endoplasmic reticulum (ER) and Golgi apparatus, or as facilitating ligand interactions with secretin-like (class B)

(Neumann et al., 2008) and glutamate-like (class C) GPCRs (Kniazeff et al., 2011). With regards to cellular trafficking, the N-terminus can be subject to N-linked glycosylation that occurs at canonical

NxS/T motifs; an important post-translational modification for correct integration of a number of

GPCRs into the cell membrane (Dong et al., 2007). The importance of this N-linked glycosylation has been demonstrated for certain GPCRs that display reduced surface expression if their N-terminal N- linked glycosylation motifs are removed by site- 2- adrenoceptors (Rands et al., 1990), vasoactive intestinaldirected peptide mutagenesis, 1 receptor including(Couvineau the et βal.,

1996), angiotensin II type 1 receptor (Deslauriers et al., 1999), the melanin-concentrating hormone receptor 1 (Saito et al., 2003) and the melanocortin 2 receptor (Roy et al., 2010). For completeness, it should be acknowledged that GPCRs may be subject to N-linked glycosylation on domains other than their N-termini, as reported to be necessary for expression of the rat AT1A receptor at the plasma membrane (Deslauriers et al., 1999). Moreover, a limited number of GPCRs have been shown to be subject to O-linked glycosylation at serine or threonine residues, of yet unclear significance, but shown to affect signalli 1-adrenergic receptor

(Goth et al., 2017) and expression of ngthe and bradykinin proteolytic B2 processing receptor (ofMichineau the β et al., 2006), for example. In addition to being a substrate for glycosylation, for a small subset of GPCRs (5-10%), the

N-terminus provides further influence on surface expression and trafficking through possession of a hydrophobic signal peptide sequence, which is typically necessary for correct protein translation and export from the ER for such receptors (Schulein et al., 2012). GPCRs that lack an N-terminal signal peptide are believed to require cryptic or uncleaved signal anchor sequences, commonly

24 found within their first transmembrane domains. Both the N-terminal signal peptide or transmembrane signal anchor are similarly recognised by the cytosolic signal recognition particle

(SRP), leading to a temporary arrest of protein translation (Coleman et al., 2017). This facilitates transfer of the ribosome and nascent protein to the ER translocon complex, where translation can resume. For receptors lacking an N-terminal signal peptide, the N-terminus is synthesised before the signal anchor, meaning that by the time the signal anchor is recognised and the growing protein transferred to the ER translocon complex, the N-terminus is not yet on the correct side of the ER membrane, and must be post-translationally translocated across it (Schulein et al., 2012). This is a challenge for GPCRs with longer N-termini or with N-termini abundant in positively charged amino acids (Lys and Arg), and accordingly, GPCRs with these properties more commonly possess a signal peptide within their N-termini, to facilitate the translocation of the N-terminus across the ER membrane as it is synthesised (Wallin et al., 1995). Similarly, some GPCR N-termini contain stably folded domains that may inhibit post-translational ER translocation, necessitating the presence of signal peptides, as postulated to be the case for the endothelin B (ETB) receptor (Kochl et al., 2002).

Following translation, signal peptides are typically removed by ER signal peptidases, but not universally. The corticotropin-releasing factor receptor 2 is one such example, harbouring an uncleaved signal peptide that both reduces plasma membrane expression and receptor dimerisation (as the signal peptide contains bulky N-linked glycans) (Rutz et al., 2015) 2C- adrenoceptor signal peptide remains similarly uncleaved, but with a contrasting result,. allowing The α more binding-competent receptor to reach the cell membrane (Jahnsen et al., 2012). Whilst this thesis explores the pharmacology and physiology of a member of the rhodopsin-like GPCRs, it warrants mentioning that for receptors outside of this family, the N-terminus is also appreciated for roles in ligand binding. Peptide ligands of the secretin-like GPCRs bind first to the receptor’s N- terminus before interacting with its transmembrane domain, as exemplified by parathyroid hormone and its cognate receptor (Vilardaga et al., 2011). Glutamate-like GPCRs, such as the

25 metabotropic glutamate receptors, are bound by their ligand solely within a so-called ‘venus fly trap’ domain, that is formed by the extensive N-termini of two dimerised receptors (Hermans et al.,

2001). More recently however, even rhodopsin-like GPCRs have been shown to involve their N- termini in ligand binding, with structural biology studies revealing the N-terminus to play a role in ligand binding for such rhodopsin-like GPCRs as the human CB1 receptor (Hua et al., 2016) and the

CXCR4 receptor (Qin et al., 2015).

Despite these well-established roles of GPCR N-termini in receptor trafficking and expression, the ability of this domain to be a direct contributor to GPCR signalling is becoming increasingly apparent (Coleman et al., 2017). This paradigm was perhaps established by the discovery of the first protease-activated receptor (PAR) PAR1 in 1991 (Rasmussen et al., 1991; Vu et al., 1991).

Members of this family of four receptors (PAR1-4) become activated following protease-mediated removal of part of their N-termini, revealing a ‘tethered ligand’ sequence on the remaining N- terminus that binds the receptor to which it is anchored, resulting in receptor activation

(Ossovskaya et al., 2004; Sidhu et al., 2014). N-terminal tethered agonists have since been discovered for other GPCRs that require no proteolysis to cause activation of their cognate receptors, including the melanocortin-4 receptor; the only receptor known, at time of writing, to have an endogenous inverse agonist (agouti-related peptide) which antagonises the activity caused by the receptor’s own tethered ligand (Srinivasan et al., 2004; Ersoy et al., 2012). These are but a few of the examples summarised in Table 1.1, that when taken together, help recast the unassuming N-terminus as a true participant in the exquisite workings of GPCR activation.

Table 1.1: Summary of the diverse modalities by which the GPCR N-terminus can directly and indirectly contribute to receptor signalling

1. Signalling modified by N-terminal proteolysis

Protease-activated Protease-inactivated

Receptor Mechanism Receptor Mechanism

Regulated cleavage of N-terminus Ligand-enhanced metalloprotease-mediated N- Protease-activated exposes new N-terminal sequence – a Endothelin B terminal proteolysis may perturb biphasic ERK receptors 1-4 tethered ligand that auto-activates the Receptor response to agonists (Grantcharova et al., 2006a). receptor (Ossovskaya et al., 2004).

Proteolytic removal of N-terminal domain may liberate the receptor from Thyroid-stimulating constraints on constitutive signalling hormone receptor (Zhang et al., 2000; Vlaeminck-Guillem et al., 2002).

2. Expression modified by N-terminal proteolysis

Protease-augmented Protease-diminished

Receptor Mechanism Receptor Mechanism

Partial removal of N-terminus necessary Removal of N-terminus plays a speculative role in Angiotensin II α1D-adrenoceptor for proper trafficking to cell surface receptor turn over and degradation (Cook et al., receptor Type I (Kountz et al., 2016) 2007).

Ligand-enhanced metalloprotease-mediated N- terminal proteolysis hypothesised to attenuate β1-adrenoceptor signalling and initiate receptor turnover (Hakalahti et al., 2010).

Ligand-enhanced metalloprotease activity Endothelin B decreases receptor surface expression Receptor (Grantcharova et al., 2002).

3. Signalling modified by N-terminus (protease-independent)

N-terminus-enhanced N-terminus-constrained

Receptor Mechanism Receptor Mechanism

Discrete residues in the N-terminus reduce Potential tethered agonist domains GPR61 GPR83 constitutive signalling, particularly through Gαq/11 within N-terminus (Toyooka et al., 2009). (Muller et al., 2016).

Single residue polymorphisms in N-terminus found Melanocortin-4 N-terminus contains a tethered agonist Serotonin receptor to enhance agonist-induced signalling (Belmer et Receptor (Srinivasan et al., 2004; Ersoy et al., 2012). 2B al., 2014). yeast pheromone α- N-terminal LDLa module essential for Region of the N-terminus constrains RXFP1 & RXFP2 factor receptor agonist stimulation (Scott et al., 2006). responsiveness to agonist (Uddin et al., 2016). Ste2p

1.4. Orphan GPCRs When considering the sheer scale of the GPCR family, it is perhaps not surprising that over 15% of GPCRs have yet to be paired with an endogenous ligand, and these receptors are termed

‘orphans’ (Ngo et al., 2016). The aforementioned utility of GPCRs as drug targets means that much research interest surrounds these orphan receptors, but in absence of high-affinity ligands and, in many cases, without knowledge of the receptor’s downstream signalling pathways, understanding their function remains a challenging undertaking (Ngo et al., 2016). At time of writing, there remains 87 rhodopsin family orphans (Alexander et al., 2017) and one of these will be the focus of this thesis - the GPR37-like 1 receptor, or GPR37L1.

As mentioned above, orphan receptors are a considerable challenge to study, as by their very definition, not only is their stimulus for activation unknown, but so too are the cellular and physiological consequences of that activation (Ngo et al., 2015). Deorphanisation has traditionally occurred through reverse pharmacology, where ligand libraries are screened for activity against an orphan of interest in a recombinant expression assay system that provides the GPCR with necessary G protein signalling and trafficking machinery (Wise et al., 2004). However, one hurdle to this process is that assays that depend on measuring G protein activity may require prior knowledge of which G protein-dependent pathway to actually monitor; an obstacle that is overcome within this thesis by use of an unbiased yeast G protein identification assay (Dowell et al.,

2009; Ngo et al., 2015). Other developments in the field involve measurement of ligand-stimulated

-arrestin activity, which does not require prior knowledge of G protein-coupling and is of great utilityβ to high-throughput screening efforts (Southern et al., 2013; Kroeze et al., 2015). Regardless of the assay chosen, candidate ligands must then be selected for screening against the orphan of interest, and may be done so in an unbiased manner (i.e. by testing thousands of compounds from large chemical libraries), or through more targeted approaches (i.e. ligands selected for known interaction with receptors having phylogenetic similarity to the orphan of interest, or flagged

28 through bioinformatic identification)(Wise et al., 2004; Ngo et al., 2016). Ultimately, any hits identified through these approaches require chemical optimisation to develop compounds that bind with high affinity to the orphan of interest, and even then, this approach may not lead to identification of the true endogenous ligand, hence waning rates of successful deorphanisation (Ngo et al., 2016). Indeed, whilst approximately 7-8 GPCRs were deorphanised per year between 1999 and 2004 (Civelli, 2005), only 15 rhodopsin family GPCRs have been deorphanised since 2005 (Ngo et al., 2016).

1.5. GPR37 and GPR37L1 First described in 1997 by Zeng et al. (1997), the orphan receptor GPR37 was discovered as a

614 amino acid transcript in a human hippocampal cDNA library and subsequently named as the human endothelin B receptor-like protein, based on its 52% similarity and 27% identity to the ETB receptor. Owing to this similarity, GPR37 has been screened for binding of endothelins, but none could be demonstrated (Marazziti et al., 1997; Zeng et al., 1997; Valdenaire et al., 1998). The first endogenous ligand proposed for GPR37 was head activator (Rezgaoui et al., 2006), an undecapeptide discovered in Hydra. Yet this pairing was wrought with inconsistencies, the most glaring being the lack of a human head activator homologue (Smith, 2015). Recently, a study has reported that prosaptide, a synthetic 14-amino acid, neuroprotective peptide derived from the saposin C domain of prosaposin, stimulates GPR37 (and the yet to be introduced GPR37L1) to

i (Meyer et al., 2013), but without convincing independent validation, GPR37 remains signalan orphan via Gα (Smith, 2015). In terms of physiology, GPR37 is most notable for its putative role in

Parkinson’s disease pathogenesis, and was identified in a yeast-2-hybrid screen as a novel interacting partner for parkin (Imai et al., 2001); a protein implicated in autosomal recessive juvenile parkinsonism (Smith, 2015). Indeed, a role in central nervous system (CNS) pathophysiology agrees with the expression profile of GPR37, with Northern blot analysis revealing this receptor to be present in the brain, particularly in the corpus callosum and substantia nigra

(Marazziti et al., 1997; Zeng et al., 1997; Donohue et al., 1998). Additionally, Northern blotting by multiple groups revealed expression in the spinal cord, placenta, liver, stomach and testis (Smith,

2015). Reports of GPR37 in other peripheral tissues are variable; Marazziti et al. (1997) could not detect transcript in human kidney, skeletal muscle, heart, lung or pancreas, Leng et al. (1999) reported low levels in rat heart, liver, kidney and lung, and Zeng et al. (1997) reported faint transcription in human kidney and pancreas.

Soon after the discovery of GPR37 came that of its closest phylogenetic neighbour: GPR37-like

1, or GPR37L1 (Figure 1.3). The discovery of GPR37L1 was reported in 1998 by two independent research groups (Valdenaire et al., 1998; Leng et al., 1999), and termed the endothelin B receptor- like protein 2 by the former and GPCR/CNS2 by the latter. Whilst both groups reported receptor transcript to be abundant in the CNS, only Leng et al. (1999) detected GPR37L1 in peripheral organs such as the heart and kidney, albeit at low levels therein. Whilst GPR37L1, like GPR37, shows amino acid similarity to the ETB receptor (56% similarity and 32% identity) (Valdenaire et al., 1998; Leng et al., 1999), endothelin or related peptides were not found to bind the receptor

(Valdenaire et al., 1998). As discussed above, GPR37 is also an orphan and, therefore, cannot guide ligand identification for GPR37L1, and the proposal of prosaptide as a ligand for both receptors

(Meyer et al., 2013) remains to be validated. Fascinatingly, our laboratory observed that three peptides previously identified in human cerebrospinal fluid (CSF) (Stark et al., 2001; Schutzer et al.,

2010; Zhao et al., 2010) are identical to three different regions of the GPR37L1 N-terminus, suggesting this receptor undergoes regulated N-terminal proteolysis. Recalling examples of protease-modulated receptors in Table 1.1, one may speculate a similar mechanism may be occurring for GPR37L1, and studies testing this hypothesis are reported in this thesis.

Whilst the pharmacology of GPR37L1 remains poorly defined, genetic approaches have begun to shed light on this receptor’s physiology. The work of Marazziti et al. (2013) described a role for

GPR37L1 as a regulator of the development of the cerebellum, with mice lacking GPR37L1

(GPR37L1 knockout) exhibiting transient differences in cerebellar cytoarchitecture, yet with improved motor skills that are retained into adulthood. Also implicating GPR37L1 in CNS physiology is a recent study by Giddens et al. (2017), in which a human variant of GPR37L1 was identified and shown to be responsible for fatal progressive myoclonus epilepsies. Further,

GPR37L1 knockout mice were identified as having enhanced susceptibility to experimentally- induced seizures (Giddens et al., 2017). Yet, the first report of GPR37L1 physiology implicated the receptor as a cardiovascular regulator, with a specific role in blood pressure control. Whilst the primary data was not shown, Min et al. (2010) reported the receptor to be downregulated in heart failure and subsequently investigated the blood pressure difference between GPR37L1 knockout mice and GPR37L1 cardiac-specific overexpressor mice, unfortunately without reference to wild- type controls. Nevertheless, the GPR37L1 knockout mice were found to have an astounding ~60 mmHg elevation in blood pressure and concomitant evidence of left ventricular hypertrophy. The magnitude of this effect, considered in the context of the absence of vital experimental controls, has made GPR37L1 a pressing candidate for further investigation and independent validation, the results of which are described herein. As GPCRs are overwhelmingly successful as pharmaceutical targets, and with GPR37L1 identified as a potential blood pressure regulator, I postulate that pharmacological modulation of GPR37L1 could represent a new strategy in the management of high blood pressure, or hypertension, which is the single biggest contributor to the global burden of cardiovascular disease (Lim et al., 2012).

Figure 1.3: Predicted secondary structure of human GPR37L1 and annotated features

Diagram adapted from Protter Visualisation (Omasits et al., 2014) and based on UniProt entry

O60883 (The UniProt, 2017). The approximate locations of the DRY and NPxxY motifs

(characteristic features of rhodopsin-like GPCRs) and a putative disulphide bond are shown (Audet et al., 2012). The locations of phosphoserine and phosphothreonine residues are inferred by similarity to mouse GPR37L1 (Huttlin et al., 2010).

1.6. The physiology of hypertension Hypertension is usually defined as a resting systolic or diastolic blood pressure above 140 mmHg or 90 mmHg, respectively (James et al., 2014). In most instances, hypertension occurs without known aetiology and is termed primary or essential hypertension (Coffman, 2011). In fewer than 10% of cases, hypertension occurs secondary to a clear identifiable cause (Rimoldi et al.,

2014). Of these cases, the most common are consequences of conditions including obstructive sleep apnoea, renal disease, primary aldosteronism and thyroid disease (Coffman, 2011; Rimoldi et al.,

2014). Given the lack of understanding of the pathophysiology of primary hypertension, the disease can be difficult to treat and in 30% of cases, patients’ blood pressure cannot be adequately controlled with existing pharmacotherapies (Rimoldi et al., 2015).

Blood pressure is governed by both cardiac output and peripheral resistance, and thus, any physiological factors that influence either parameter could contribute to the dysregulation and elevation of blood pressure. Unfortunately for the hypertension researcher, these factors are many, with an early example of a systems biology study identifying well over 300 subsystems that contribute to circulatory regulation (Guyton et al., 1972b). Yet, in addition to describing the sheer complexity of circulatory regulation, Guyton and colleagues were pioneers of a unifying and renal- centric hypothesis of hypertension, suggesting the disease is overwhelmingly driven by the body’s need to regulate its fluid volume (Guyton, 1989). Guyton’s theory of hypertension (Guyton et al.,

1970; Guyton et al., 1972a; Guyton, 1981; Guyton, 1989) is as follows and whilst its criticisms are valid and will be mentioned in part, it does serve as an appropriate framework by which the main physiological mediators of blood pressure, and their interconnectedness, can be introduced.

Guyton proposed the following: an initial increase in cardiac output (possibly driven by renal or vascular dysregulation, minute yet sustained increases in extracellular fluid volume, effectors of the

CNS, or some other trigger), causes compensatory vasoconstriction to increase total peripheral

33 resistance, and thus cardiac afterload, in order to achieve a concomitant reduction in, and normalisation of, cardiac output. This comes at the expense of increased peripheral resistance.

However, the induced vasoconstriction affects the kidney, causing decreased perfusion of the organ and preventing normal salt excretion, thus increasing extracellular fluid volume and blood pressure. Further, the kidney may raise arterial blood pressure in an attempt to increase pressure diuresis, via components of the renin-angiotensin system, triggered by the aforementioned renal vasoconstriction causing a decrease in renal perfusion as sensed by juxtaglomerular cells. Thus, the circulatory system is maintained in a long-term state of elevated pressure arising from both an increase in total peripheral resistance, compensatory to increased cardiac output, and the consequences of that increased resistance for renal physiology.

The first determinant of blood pressure introduced in Guyton’s hypothesis above is cardiac output, which itself is a function of heart rate and stroke volume. Heart rate, in turn, is chiefly controlled by the CNS (Anrep et al., 1926), and stroke volume controlled by the central nervous, vascular and renal systems, through influencing factors including cardiac contractility, afterload, preload and blood volume (Hamilton et al., 1948; Bugge-Asperheim et al., 1973; Greenway et al.,

1986). Elevated cardiac output is not commonly observed in established hypertension (Lund-

Johansen, 1980), but hypothesised to contribute to the onset of the disease (Guyton et al., 1970), supported by observations of increased stroke volume and heart rate in subsets of humans with early hypertension (Julius et al., 1971). The next contributor to Guyton’s hypothesis of hypertension is peripheral resistance, which comprises a sum of the resistances to flow through all organs and tissues and is principally determined by blood vessel diameter (Conway, 1984), which is primarily controlled by vascular smooth muscle as regulated by sympathetic nervous system (SNS) tone

(Oparil et al., 2003) and circulating factors including angiotensin II (AngII) (Coffman, 2011), estrogen (Crews et al., 1999) and endothelin-1 (Schiffrin, 2001). Again, this implicates the CNS, vasculature and kidneys in progression of the disease. Furthermore, the vasculature of

34 hypertensive patients is often hypersensitive to sympathetic stimulation, and with chronic stimulation, may undergo restrictive vascular remodelling (Oparil et al., 2003). Blood viscosity also influences total peripheral resistance and may contribute to hypertension pathogenesis, with increased viscosity observed in human patients (Letcher et al., 1981; Linde et al., 1993). Next in

Guyton’s hypothesis is that this state of increased total peripheral resistance causes a decrease in renal salt excretion, such that (i.) extracellular fluid volume is increased and (ii.) the kidneys begin to drive hypertension, to achieve adequate perfusion and excretion of salt (Guyton et al., 1972a).

Whilst changes in extracellular fluid volume are often immeasurable in hypertension, Guyton

(1989) proposed that the change required to cause hypertension may be smaller than can be accurately quantified. Regardless, excess renal vasoconstriction and inadequate blood flow can cause the kidney to increase blood pressure by endocrine, paracrine and autocrine factors (Luke,

1993). Primarily, the kidney responds to inadequate perfusion by increasing activity of the renin- angiotensin system, potentiating SNS activity, vasoconstriction, salt and water retention, and thus, increasing blood pressure (Peach, 1977). In agreement with renal mechanisms of essential hypertension is the strong association between high blood pressure and high salt intake (Meneton et al., 2005). Differences in individual’s sensitivity to salt-mediated changes in blood pressure are thought to stem from traits that favour salt retention, and may have once granted an evolutionary advantage during times of nutritional scarcity, but now promote hypertension in the setting of the relatively high-salt intake in the diets of people in the developed world (Coffman, 2011; Choi et al.,

2015). Crucially, not only can the kidney potentiate SNS activity, with circulating AngII able to bind its receptors on the brain’s circumventricular organs, but the CNS can potentiate the actions of the kidney. This can occur by sympathetic innervation of the organ (DiBona, 2002), influencing the kidney’s renin release, glomerular filtration rate and renal tubular reabsorption of sodium (Parati et al., 2012), or by the renal effects of antidiuretic hormone (ADH), as released from the pituitary gland, on water resorption. (Share, 1988).

Whilst Guyton’s renal-centric view of hypertension has not been discredited by modern research (Coffman, 2014), evidence supports more substantial contributions of both neural and vascular systems to disease pathogenesis (Michael et al., 2008; Osborn et al., 2011; Coffman, 2014).

Critics of Guyton’s theory of hypertension question whether blood volume changes would lead to dysregulation of homeostatic mechanisms and accordantly elevated total peripheral resistance, whether the hypothesis’ linear mathematical assumptions are valid, and the degree to which primary hypertension is driven by renal factors (Korner, 2007). Indeed, an alternative, yet equally unifying, view of hypertension was championed by Korner, and dictated that all vascular, renal and hormonal contributions to hypertension were secondary to the disease being driven by the CNS

(key regions of CNS cardiovascular control shown in Figure 1.4).

Figure 1.4: Schematic of key central nervous system pathways of cardiovascular control

Information from the sensory nervous system, baroreceptors (via the 9th and 10th cranial nerves) and spinal afferents are integrated at a number of levels within the CNS. Corresponding effectors include the vagus nerve, the hypothalamic-

pituitary axis and sympathetic ganglia. Σ

37 denotes area of integration. NTS, nucleus tractus solitarius; RVLM/CVLM, rostral and caudal ventrolateral medulla; pre-GM, preganglionic sympathetic motor neurons; SG, sympathetic ganglia.

Adapted from Korner (2007).

Korner hypothesised that genetically-susceptible individuals progressively become hypersensitive to mental stressors, exacerbated by environmental and lifestyle factors, resulting in a SNS response to stimuli that ordinarily would not elicit such a response in a non-genetically predisposed normotensive subject. SNS-mediated vasoconstriction increases total peripheral resistance and reduces extracellular space, increasing extracellular sodium concentration, and making blood pressure more sensitive to sodium. Increased stress susceptibility also enhances activity of the hypothalamic pituitary axis, decreasing nitric oxide (a potent vasodilator) through elevated release of cortisol, and enhancing the permeability of the blood-brain barrier to sodium, leading to increased brain AngII and further blood pressure increases. Sustained blood pressure increases lead to vascular remodelling to compensate for increased wall stress, maintaining the elevation of total peripheral resistance and predisposing to inadequate organ perfusion and ischaemia. As discussed above, inadequate renal perfusion is met with the kidney’s attempts to increase blood pressure further to restore its perfusion. As will be discussed in the next section, ischaemic damage to the kidney impairs renal function and exacerbates the disease.

Whilst the involvement of the CNS in hypertension aetiology is well demonstrated, it is likely the disease is driven by both the CNS, as championed by Korner, and by the kidney as proposed by

Guyton. The true pathophysiology of primary hypertension may involve both hypotheses, one or the other in different patient subgroups, or neither. In partial refutation of both hypotheses, the failure of renal denervation to significantly reduce systolic blood pressure compared to sham surgery, in resistant hypertensive patients in the SYMPLICITY HTN-3 trial, serves to demonstrate that sympathetic innervation of the kidney is not as contributory to the maintenance of hypertension as once thought (Bhatt et al., 2014). Whilst both Guyton’s and Korner’s hypotheses are distinct, they have in common a concept of seemingly innocuous stimuli driving progressive and self-sustaining dysregulation of blood pressure control mechanisms in genetically-susceptible individuals. Importantly, although not without its critics, Guyton’s landmark unifying theory of

39 hypertension represents a pioneering integrative approach to the physiological modulators of blood pressure, and remains a useful model for understanding the nexus between them.

1.7. Physiological consequences of hypertension In contrast to our incomplete understanding of the aetiology of primary hypertension, disease outcomes are well characterised. These complications have long been known to be cardiovascular, cerebral, retinal and renal (Paullin, 1926). A meta-analysis of 30 echocardiographic studies revealed that approximately one third of hypertensive patients show evidence of left ventricular hypertrophy (LVH) (Cuspidi et al., 2012), an abnormal increase in the heart’s left ventricle (LV) muscle mass in response to increased cardiac workload (Dorn, 2007). Yet, the reported prevalence of LVH in hypertension is highly variable, and increases with severity and duration of hypertension

(Ruilope et al., 2008). LVH is initially compensatory, serving to normalise wall stress through the stimulation of cardiomyocyte sarcomeres to multiply in parallel in response to increased afterload

(Prisant, 2005). Yet this eventually leads to worsened performance of the heart. The thicker ventricle is less elastic, so its filling during diastole becomes more dependent on left-atrial systole, increasing LV end-diastolic pressure and eventually causing pulmonary congestion (Prisant, 2005).

Cardiac oxygen demand is also increased by hypertrophy, which may not be sufficiently met during exercise or if LVH is advanced (Prisant, 2005). Further complications of LVH include: (i.) loss of contractility due to cardiomyocyte apoptosis, (ii.) dysfunction and arrhythmia due to fibroblast- mediated collagen deposition, and (iii.) remodelling of intramyocardial vasculature predisposing to inadequate myocardial perfusion (Dorn, 2007; Nadruz, 2015). Classically, these conditions predispose to LV dilatation and/or myocardial infarction, diminishing LV contractility and cardiac output, and eventually, proceeding to systolic heart failure (Prisant, 2005).

Hypertension also causes cerebrovascular complications, and is the strongest risk factor for stroke (Iadecola et al., 2008). The relationship between the two is linear, with one study

40 demonstrating that every 1 mmHg increase in treated systolic or diastolic pressures, raises risk of death by stroke by 2% and 3% respectively (Palmer et al., 1992). A multitude of vascular effects underlie the increased risk of stroke. Vessel occlusion can occur by atherosclerotic plaques in the cerebral circulation, leading to ischaemic injury, and causing necrosis of vessels supplying white matter (Iadecola et al., 2008). Further, inadequate perfusion may also be induced by adaptive vascular remodelling, aimed at reducing wall stress which would otherwise compromise the blood- brain barrier and lead to cerebral oedema. This remodelling is driven by trophic effects of sympathetic innervation, and by mechanical stress through growth factor-, oxidative stress- and nitric oxide-mediated responses (Iadecola et al., 2008). Independent of gross stroke, the compromised cerebrovascular circulation may lead to cognitive decline involving multiple faculties, including those of learning and memory, and psychomotor and visuospatial skills, as well as being associated with cerebral atrophy and Alzheimer’s disease (Manolio et al., 2003).

The kidney is also at risk of damage in primary hypertension, which is the second leading cause of end-stage renal disease, after diabetes (Mennuni et al., 2014). Classically, hypertension-induced renal damage has been classified on the basis of nephrosclerosis. Benign nephrosclerosis involves vascular hyaline arteriosclerotic lesions with nephron loss occurring progressively, but without overt proteinuria or compromise of renal function (Bidani et al., 2004) Malignant nephrosclerosis occurs in severe hypertension, and is characterised by vascular and glomerular injury with fibrinoid thrombosis and necrosis. However, hypertension also accelerates the progression, and increases the severity, of most chronic kidney diseases (Bidani et al., 2004). The damage that hypertension inflicts on the kidney is thought to primarily be the result of barotrauma to the microvasculature, as compensatory mechanisms such as glomerular afferent arteriole constriction, which limits blood flow to the kidney, become overwhelmed (Bidani et al., 2004; Mennuni et al.,

2014). As in the brain, structural changes occur in the kidney to allow the vessels to withstand increased wall stress (Mennuni et al., 2014). Such compensatory responses increase medial wall

41 thickness, which narrows the lumen and increases the distance across which oxygen must diffuse, promoting ischaemic injury of glomerular and tubulo-interstitial structures (Mennuni et al., 2014).

Moreover, this compensatory response limits renal perfusion, which when sensed by the kidney, stimulates its attempts to further increase mean arterial pressure to restore its perfusion, thus further increasing blood pressure and damage to the kidney (Mennuni et al., 2014).

1.8. The renin-angiotensin system Although briefly introduced above for its role in hypertension pathophysiology, the renin- angiotensin system (RAS) also regulates blood pressure under normal physiological circumstances

(Kaschina et al., 2003) (Figure 1.5), and must be explored in more detail, as manipulation of this system is used as an experimental model of mouse hypertension within this thesis. Classically, the

RAS is an endocrine system, that mediates its effects through its final effector, AngII (Fyhrquist et al., 2008) constitutively, which released is synthesised into the circulation as follows. from Angiotensinogen the liver (Griendling (AGT) et al. is , an1993 α2) , and macroglobulin is cleaved into the angiotensin I (AngI) decapeptide by activity of the aspartyl protease renin (REN) released from the kidney (Griendling et al., 1993; Lavoie et al., 2003). AGT synthesis is typically thought to be hormonally regulated, modulated by glucocorticoids and estrogens, for example (Corvol et al.,

1997), and REN release, acutely stimulated by decreases in renal perfusion pressure or increases in sympathetic outflow to the kidneys (Kurtz, 2011). The resulting AngI is further cleaved by the membrane-bound carboxypeptidase, angiotensin converting enzyme (ACE), to yield the octopeptide AngII and C-terminal dipeptide, His-Leu (Brown et al., 1998; Lavoie et al., 2003). ACE is located in the vascular endothelium of organs including, in particular, the lung, but also liver, adrenal cortex, kidney, and spleen (Caldwell et al., 1976) and gastrointestinal tract (Bruneval et al.,

1986), as well as in epithelial cells of the gastrointestinal tract and kidney (Bruneval et al., 1986).

ACE is also present in a soluble form in the circulation (Das et al., 1977), as well as having a

42 homologue, ACE 2 (Donoghue et al., 2000; Tipnis et al., 2000), which contrasts ACE in that AngII is its preferred substrate, over AngI (Tipnis et al., 2000; Vickers et al., 2002).

The biological activity of AngII is primarily mediated through its activation of the AT1 and AT2

GPCRs. AT1 q-mediated coupling (Kaschina et al.,

2003), and istypically expressed activates in brain, phospholipase heart, lung, Cliver, through adrenal, Gα and vascular tissues (Lavoie et al.,

2003). The generalised effects of AT1 activation increase blood pressure through systemic vasoconstriction, increased SNS activity, enhanced proximal tubular sodium resorption and stimulation of the release of aldosterone from the adrenal cortex and antidiuretic hormone from the pituitary gland (Severs et al., 1970; Kaschina et al., 2003). Study of the RAS is further complicated in rodents by the existence of two isoforms of the AT1 receptor; the AT1A and the AT1B.

The latter has a more restricted expression profile and contributes less to the pressor response

(Kaschina et al., 2003) q-coupling of AT1, the signalling activity of AT2, which shares only ~34%. In contrastamino acid to the identity known with Gα AT1, has not been clearly defined, and whilst

s i coupled, it has not been conclusively demonstrated to be either (Porrello etreported al., 2009 to) ,be and Gα a orrecently Gα reported crystal structure of AT2 in an ‘active’ state, appeared to show the receptor’s helix 8 adopting a non-canonical structure whereby G protein interactions are obstructed (Zhang et al., 2017). Although predominating in abundance over the AT1 in fetal tissue, the AT2 is not as abundant in the adult, but is still widely expressed in vascular, brain, lung, adrenal, kidney and heart tissue, and its expression is these regions is modulated in disease states

(Steckelings et al., 2005). Physiologically, the AT2 has traditionally been thought to oppose AT1 activation, and hypothesised to prevent overstimulation of AT1 signalling (Steckelings et al., 2005), but with different AT2-deletion mouse models exhibiting opposing effects on blood pressure, cardiac fibrosis and hypertrophy, its physiological role remains to be clearly defined (Porrello et al.,

2009).

In addition to the systemic and classical perspective of RAS activity described above, largely self-contained RAS systems can be found within specific organs, including the vasculature, heart, kidney and brain (Bader, 2010). This is well demonstrated for the latter, with the brain being largely impermeable to circulating AngII, but having high capacity for AngII generation, as AGT produced by astroglia is found in high abundance in CSF (Bader, 2010). Yet further complexity is provided by the existence of non-canonical but biologically active angiotensin peptides including

Ang 1-9 (produced from AngI by activity of ACE 2), Ang 1-7 (produced from AngII by ACE 2 or from

Ang 1-9 by ACE), AngIII (produced from AngII by aminopeptidase A) and AngIV (produced from

AngIII by aminopeptidase M) (Fyhrquist et al., 2008). These peptides have varying specificities for

AT1 and AT2, but Ang 1-7 may oppose the physiological consequences of AT1 activation, by interacting with an atypical GPCR, the Mas receptor [(reviewed in detail by Fyhrquist et al. (2008)].

Figure 1.5: Schematic of the classical renin-angiotensin system

Renin (released from the kidney in response to decreased renal perfusion) catalyses the conversion of angiotensinogen (released from the liver), into angiotensin I. Angiotensin converting enzyme, located the endothelium of organs including the lung and kidney, catalyses the conversion of angiotensin I into the classical biological effector of the renin-angiotensin system (RAS),

45 angiotensin II. Dotted arrow denotes tissue release of RAS-effector, curved arrow denotes substrate conversion, blunted line denotes inhibition of renin release by restoration of renal perfusion. ACE, angiotensin converting enzyme; ADH, anti-diuretic hormone; CNS, central nervous system; SNS, sympathetic nervous system.

1.9. Treating hypertension As previously introduced, the multifactorial nature of the pathophysiology of hypertension makes the disease resistant to treatment. The adoption of healthy lifestyle factors is not only preventative in development of hypertension, but an indispensable part of disease management

(Chobanian et al., 2003). Such lifestyle factors that are recommended for their contribution to the lowering of blood pressure include: (i.) the maintenance of a body mass index of 18.5-24.9 kg/m2,

(ii.) a diet high in fruits and vegetables and low in saturated fats and sodium, (iii.) regular aerobic exercise and (iv.) moderation of alcohol consumption (Chobanian et al., 2003). However, where lifestyle interventions are insufficient to achieve a patient’s goal blood pressure (commonly a systolic pressure below 140 mmHg and a diastolic pressure below 90 mmHg) pharmacological interventions are necessary (Chobanian et al., 2003). Yet two thirds of patients receiving pharmacological treatment for their blood pressure require more than one type of medication

(Cushman et al., 2002).

Whilst there are many guidelines for the pharmacological management of hypertension, such as those set by the European Societies of Hypertension and Cardiology (Mancia et al., 2013), I will use those set by the American Medical Association (James et al., 2014) as a linear framework by which the different classes of antihypertensive pharmaceuticals can be introduced. The most recent

American Medical Association guidelines for high blood pressure management recommend that most patients (those without chronic kidney disease) are treated first with a thiazide-type diuretic

(eg. bendroflumethazide, hydrochlorothiazide) (James et al., 2014). By inhibiting the Na+/Cl- symporter in the renal distal convoluted tubule, Na+/Cl- and water are lost from the body, resulting in a reduction in fluid volume, and therefore, cardiac output and blood pressure (Duarte et al.,

2010). Yet, decreased baroreceptor stimulation and renal perfusion promotes RAS and SNS activity, quickly counter-acting the medication’s efficacy (Duarte et al., 2010). In the face of this, the actual mechanism by which thiazide-type diuretics remain effective is not defined, but theorised to be

47 from vasodilatory effects mediated by the endothelium and vascular smooth muscle (Duarte et al.,

2010). These drugs are also associated with various metabolic side effects including hyperlipidemia, hyperglycemia, hypokalemia, hyperuricemia and RAS stimulation (Duarte et al.,

2010).

Where thiazide-type diuretics alone are insufficient to achieve goal blood pressure, the

American Medical Association recommends introducing a second class of medication: either an ACE inhibitor (eg. captopril), an angiotensin receptor blocker (ARB; eg. candesartan) or a calcium channel blocker (CCB; eg. amlodipine) (James et al., 2014). ACE inhibitors reduce circulating AngII, and therefore its pressor effects, by inhibiting the ability of ACE to cleave AngI into AngII (Brown et al., 1998). Side effects of ACE inhibitors include cough and, more rarely and seriously, angioneurotic oedema, which can be fatal through respiratory stress and obstruction (Israili et al., 1992). ARBs are AT1 receptor antagonists and reduce the pressor response that AngII ordinarily mediates through these receptors, whilst preserving the potentially beneficial effects that AngII may mediate through AT2 (Ruilope et al., 2005). Whilst plasma AngII increases in response to AngII receptor blockade, rebound hypertension has not been found to occur on withdrawal from treatment, and the side effect profile of ARBs does not differ from that of placebo (Burnier, 2001). Separate from these modifiers of the RAS, CCBs, either specifically targeting L-type calcium channels in vascular smooth muscle (dihydropyridines; DHP), or broadly targeting L-type calcium channels in the vasculature, myocardium, and sinoatrial and atrioventricular nodes (non-dihydropyridines; non-

DHP) (Basile, 2004), may also be used as second-line antihypertensives. The former causes more potent vasodilation, and the latter also causes negative chronotropy and decreased SNS activity

(Basile, 2004). Either way, both types of CCB prevent the excitatory effects of calcium entering the cell in response to voltage-mediated opening of the L-type channel (Flynn et al., 2000). DHPs more commonly cause headaches, dizziness, flushing and peripheral oedema than non-DHPs, whilst the

48 ability of non-DHPs to diminish cardiac conduction and contractility contraindicates their use in patients with severe cardiac dysfunction (Basile, 2004).

If a patient’s blood pressure remains uncontrolled by a combination of thiazide-type diuretics,

ACE inhibitors, ARBs and CCBs, the American Medical Association recommends the introduction of

-blockers’ (BB; eg. atenolol) (James et al., 2014) -adrenoceptor antagonists (Baker,

2005‘β ) 1- 2-adrenoceptors,, which decreasing are β the cardiac adrenergic response in the heart. BBs by target antagonising both the βtheand former, β but may cause bronchospasm by antagonising the latter

(Baker, 2005) and therefore are contraindicated for use in asthmatic patients (Mancia et al., 2013).

Alt 1-andrenoceptor exist, those used clinically have rather poorhough in BBsvivo with higher affinity for the β 2-adrenoceptor (Baker, 2005). Once recommended asselectivity first-line antihypertensives, for this receptor BBs over have the been β shown to be less efficacious than the previously discussed treatments, and hence their positioning for use only in combination therapy

(Lindholm et al., 2005). It should also be noted that the pharmacological interventions discussed in this section should be used in conjunction with the healthy lifestyle modifications introduced above, which further serve to enhance the efficacy of antihypertensive medications (Chobanian et al., 2003).

1.10. Mouse models of hypertension and left ventricular hypertrophy The laboratory mouse has proven to be a vital contributor to scientific research since the 15th century and, owing to the animal’s small size and resource demands, short generation time, and relative ease of breeding and genetic manipulation, continues to aid researchers in the elucidation of mammalian biology and pathology (Coleman et al., 2015). The study of complex diseases involving multiple physiological systems has benefitted immensely from this model organism, and cardiovascular disease is, of course, no exception. A selection of common pharmacological and

49 surgical cardiovascular disease models will now be briefly introduced, with an emphasis on models of hypertension and LVH.

As introduced previously, blood pressure is regulated at many levels, although experimental hypertension in mice is usually introduced by modification of renal or vascular mechanisms. Whilst most of these models require surgery, the least invasive of these is likely the pharmacological induction of hypertension by AngII infusion, delivered subcutaneously by implantation of an osmotic pump, which absorbs extracellular fluid through its outer membrane, displacing AngII- containing solution from an internal reservoir, at a constant rate mediated by a flow regulator. The released AngII is able to stimulate all the pressor systems regulated by endogenous AngII (i.e enhancing SNS activity, vasoconstriction, renal sodium resorption, etc). Depending on the AngII dose and strain of mouse, this model can reliably increase conscious systolic blood pressure by approximately 30-40 mmHg in male mice, reaching a maximum pressor response by approximately one week (Xue et al., 2005; Kvakan et al., 2009; Haggerty et al., 2015) and may then be maintained for a further three, depending on pump reservoir volume (Haggerty et al., 2015). Importantly for the findings of this thesis, the blood pressure increase in response to AngII infusion is weaker for females, whilst for animals of both sexes, AngII-induced blood pressure increases are less marked when measured in anaesthetised animals (Haggerty et al., 2015). AngII infusion leads to LVH, which occurs in response to the increased cardiac afterload, and potentially through consequences of

AngII-induced paracrine signalling or direct action on cardiomyocytes (Kim et al., 2000). This hypertension model was used in the experiments described herein, owing to its reproducibility and minimal invasiveness compared to the following hypertension models.

An alternative pharmacological model of hypertension involves administration of mineralocorticoid deoxycorticosterone acetate (DOCA) and a high salt diet, with (Wang et al., 2009) or without (Nath et al., 2007) uninephrectomy. As a point of difference, blood pressure is slower to

50 rise in the DOCA-salt model than with AngII infusion, with peak blood pressure increases (~20-30 mmHg) (Obst et al., 2004; Jia et al., 2010; Ji et al., 2012) taking at least two weeks to develop (Obst et al., 2004; Yemane et al., 2010). Intriguingly, the actual means by which blood pressure is increased through this model are not completely known, with a combination of factors including

SNS activity (through renal-dependent and -independent means), endothelin-1 and ADH thought to contribute (Yemane et al., 2010). LVH in this model occurs by both blood pressure-dependent and – independent (i.e. endothelin-1, adrenergic signalling) mechanisms (Karatas et al., 2008).

Unlike the previously described models, which are primarily pharmacologically driven, surgical models may also be used to induce hypertension in mice. Whereas the AngII infusion model requires an exogenous source of AngII, the two kidney one clip (2K1C) model is an alternative that increases endogenous circulating AngII. This achieves blood pressure increases (20-30 mmHg) comparable to the AngII infusion model (Wiesel et al., 1997; Tyagi et al., 2005; Sun et al., 2009), but can be sustained for longer (months), without relying on an external AngII source. In the 2K1C model, whose origins trace back as early as 1934 (Goldblatt et al., 1934), one renal artery is constricted, chronically reducing the kidney’s perfusion, and stimulating increased release of renin

(Wiesel et al., 1997). This perpetual activation of the RAS maintains hypertension, but the preservation of one normally functioning kidney prevents hypervolemia (Wiesel et al., 1997). An alternative model, with a greater pressor response, is the one kidney one clip (1K1C) model, which differs in that whilst one kidney undergoes renal artery restriction, the other is surgically removed

(Wiesel et al., 1997). In this case, hypertension is maintained not solely by the RAS, but also by hypervolemia (Wiesel et al., 1997). Whilst subtly different in their mechanism of hypertension, both models induce LVH (de Simone et al., 1993; Corbier et al., 1994).

Whilst arguably less representative of the systemic nature of human hypertension than the previously discussed models (Doggrell et al., 1998), surgical obstruction of cardiac outflow through

51 aortic constriction, typically induced by ligation between the brachiocephalic and left common carotid arteries (Rockman et al., 1991), instantaneously increases total peripheral resistance, and induces rapid and marked effects on blood pressure and LVH in mice. Our laboratory has shown that such partial thoracic aortic constriction (TAC) can increase systolic pressure by ~50 mmHg and cause LVH within seven days (Soetanto et al., 2016) and progress to heart failure within 12 weeks (Du et al., 2004). Whilst TAC is typically considered a RAS-independent model, a RAS- dependent aortic constriction model also exists whereby the aorta is ligated immediately above the suprarenal arteries, which decreases renal perfusion, thus stimulating the RAS (Byrne et al., 2003).

Regardless of the model used to induce LVH, the measurable endpoints include molecular, cellular and morphometric changes. The latter is often assessed by expressing the weight of the whole heart or its individual chambers as a ratio of the animal’s body weight or tibia length, with this ratio increasing as the heart becomes heavier from hypertrophic cell growth and fibrotic remodelling (Sakata et al., 1998). LVH can also be measured at the cellular level, with hypertrophied cells exhibiting increased cross-sectional area and a corresponding decrease in cardiomyocyte density, that is, the number of cells per area of tissue (Li et al., 2008). Remodelling of the heart as a result of cardiomyocyte hypertrophy is accompanied by a measurable molecular signature whereby the heart returns to a transcriptional state reminiscent of that seen during its fetal development (Frey et al., 2003), which includes enhanced transcription of atrial natriuretic pept - -skeletal actin genes; a response hypothesised to represent the heart’side, attempt β myosin to sustainheavy chain energy and metabolism α in an increasingly hypoxic environment (Taegtmeyer et al., 2010).

1.11. Rationale and aims of PhD project

1.11.1. Rationale As an orphan GPCR, the pharmacology and physiology of GPR37L1 remain largely undefined.

Whilst a prior study suggested a role for GPR37L1 in blood pressure control, a more thorough, independent investigation of the cardiovascular effects of GPR37L1 is necessary to better determine the potential cardiovascular role of GPR37L1 and to elucidate whether this orphan receptor may indeed be a potential target for pharmacological management of human hypertension.

Yet, evaluation of the pharmacological targeting of GPR37L1 requires an understanding of the molecular basis of its activation and regulation. The presence of peptides identical to regions of

GPR37L1 N-terminal peptides, as identified in human CSF, hints at an intriguing proteolytic regulatory mechanism, but without understanding of the receptor’s signalling pathways, the consequences of this phenomenon remain unclear. Further, lack of knowledge of the intracellular signalling pathway(s) of GPR37L1 also hinders the search for the receptor’s endogenous ligand.

Accordingly, the experiments in this thesis were undertaken to elucidate both the pharmacological and physiological properties of GPR37L1.

1.11.2. Hypotheses and aims Firstly, I hypothesise that the presence of peptides in human CSF with identity to the GPR37L1

N-terminus indicates that the receptor undergoes regulated proteolysis. Accordingly, my aims were to:

• Investigate GPR37L1 N-terminal proteolysis in vitro,

• Define the cognate signalling pathway of GPR37L1 and identify ligands or mechanisms

that may modulate its activity in vitro.

Secondly, I hypothesise that GPR37L1 is a cardiovascular regulator with a specific role in blood pressure control. To address this hypothesis, my aims were to:

• Characterise the expression of GPR37L1 in the mouse brain, heart and kidney,

• Determine if GPR37L1 is a blood pressure-regulating gene in the mouse,

• Ascertain whether GPR37L1 conveys protection in a mouse model of hypertension.

Chapter Two: Methods

2.1. Ethics statement All animal work was performed in accordance with the Australia Code for the Care and Use of

Animals for Scientific Purposes, 8th Edition (2013) and was approved by the relevant Animal Ethics

Committees (Garvan Institute for Medical Research/St Vincent’s Hospital, project numbers AEC

11/25, 13/30, 17/11; University of Melbourne, project number 1212573; University of New South

Wales, project number 15/12B).

2.2. Materials and reagents

2.2.1. General reagents All reagents were sourced from Sigma-Aldrich (USA) or Thermo Fisher Scientific (including Life

Technologies and Invitrogen, USA), unless otherwise indicated. Unless otherwise stated, water refers to deionised water. All experiments performed at room temperature (RT) unless other stated.

2.2.2. DNA constructs Constructs used in in vitro GPR37L1 studies were sourced or generated as follows: A previously reported human GPR37L1 construct (Meyer et al., 2013) was sourced from Multispan (USA). This construct had an N-terminal Flag tag and lacked the endogenous GPR37L1 signal peptide, so the construct was restored to the original human GPR37L1 sequence with a Hind III restriction site and

Kozak sequence by polymerase chain reaction (performed by Dr Nicola Smith). Further, the

GPR37L1 construct obtained from Multispan contained a rare variant (K91R), which was restored to the wild-type sequence by QuickChange (Agilent, USA) site-directed mutagenesis (performed by

Ms Nadine Mrad). N-terminal - - -GPR37L1-eYFP) were generated by polymerase chain reactiontruncation (performed mutants by (Δ25Dr Nicola, Δ80 Smith, Δ122 and Ms Nadine Mrad). N- terminal alanine mutants (106Ala108, 109Ala111, 112Ala114, 115Ala117, 118Ala120, 121Ala123, 124Ala126; all tagged with e-YFP) were generated by polymerase chain reaction (performed by Dr Chu Kong

Liew). For creation of stable cell lines, human GPR37L1 (or N-terminal truncation mutants) was cloned into pcDNA5/FRT/TO with a C-terminal eYFP tag (performed by Dr Nicola Smith). For S. cerevisiae experiments, human GPR37L1 (or N-terminal truncation mutant) was cloned into p426GPD with a C-terminal eYFP tag (performed by Ms Nadine Mrad). Constructs were verified by

DNA sequencing performed by the Garvan Molecular Genetics facility (Sydney, Australia). N- terminal “CSF” and “tethered” peptides were synthesised by Genscript (China).

2.3. In vitro experiments and cell maintenance

2.3.1. Stable HEK293 cells

2.3.1.1. Generation and culture of HEK293 cells Prior to transfection, Flp-In™ T-REx™ cells (Flp-In) (Invitrogen) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 4 mM L-glutamine, 10% v/v fetal bovine serum (FBS),

10 µg/ml blasticidin and 50 units/ml penicillin/streptomycin. Cells were maintained in a

o humidified incubator at 37 C in air/CO2 (19:1) unless otherwise stated. Inducible Flp-In T-REx

HEK293 cells expressing human GPR37L1- - - -GPR37L1-eYFP N-terminal truncation mutants, or triple-alanine GPR37L1eYFP - (oreYFP Δ25 N-,terminal Δ80 , Δ122 mutants) under control of a doxycycline-inducible promoter were generated as follows. Briefly, cells were transfected with pcDNA5/FRT/TO (containing GPR37L1-eYFP) and pOG44 (1:9) using LipofectAMINE (Invitrogen) as per manufacturer’s protocol (Invitrogen). After 48 h, the media was changed to that described above, supplemented with 200 µg/ml hygromycin B (to select for cells that successfully integrated the gene of interest). Stably transfected cells were subsequently maintained in this media. Stable cell lines were generated by myself, Dr Tony Ngo, and Dr Nicola J Smith. Expression of GPR37L1 was induced in stably transfected cells by addition of 500 ng/ml doxycycline to cell media for 24 h.

2.3.1.2. Treatments and harvest of HEK293 cells

2.3.1.2.1. Induction of GPR37L1 expression in HEK293 cells and inhibition of metalloproteases For use in experiments, GPR37L1-eYFP-transfected Flp-In cells were cultured on 0.1 mg/ml

Poly-D-lysine coated 6-well plates at a density of 6 x 105 cells/well. 24 h later, media was aspirated and replaced and GPR37L1-eYFP expression was induced by addition of 500 ng/ml doxycycline.

Where inhibitors including BB94 (batimastat, 5 mM stock), TAPI-1 (5 mM stock) and TAPI-2 (10 mM stock) (Santa Cruz Biotechnology, USA), and TIMP1 (4.8 mM stock), TIMP2 (4.7 mM stock),

TIMP3 (4.6 mM stock) and TIMP4 (8.9 mM stock) (R&D systems, USA), were used, they were added at time of induction (at final concentrations indicated in text) of wild-type or mutant GPR37L1- eYFP expression. All inhibitors were stored in aliquots at -20oC.

2.3.1.2.2. Harvest of HEK293 cells After 24 h, plates were placed on ice and media was aspirated from cells. Cells were then washed with ice-cold phosphate-buffered saline (PBS) before harvesting with ice-cold radio- immunoprecipitation (RIPA) buffer [50 mM HEPES, 150 mM NaCl, 1% Triton X100, 0.5% deoxycholate, 0.1% SDS, 10 mM NaF, 5 mM ethylenediaminetetraacetic acid (EDTA), 10 mM

NaH2PO4, 5% ethylene glycol, Complete EDTA-free protease inhibitor cocktail (Roche Diagnostics,

Australia)]. Cell lysates were cleared of insoluble debris by centrifugation at 16,000 r.c.f. for 15 min, and the supernatant mixed with Laemmli buffer (final concentrations: 63 mM Tris, 50 mM dithiothreitol, 80 mM SDS, 10% glycerol, 0.004% bromophenol blue, pH 6.8), followed by 15 min incubation in a 37oC water bath.

2.3.1.2.3. HEK293 cell surface biotinylation Cell surface biotinylation experiments were performed by Dr Nicola Smith as previously described (Canals et al., 2009). Wild-type or mutant GPR37L1-eYFP-transfected Flp-In cells were cultured on 0.1 mg/ml Poly-D-lysine coated 6-well plates at a density of 6 x 105 cells/well and

58 induced to express receptor constructs by addition of 500 ng/ml doxycycline for 24 h. Media was removed and cells were washed with ice-cold borate buffer (10 mM boric acid, 154 mM NaCl,

7.2 mM KCl and 1.8 mM CaCl2, pH 9.0) and incubated on ice with 1 ml of 0.8 mM EZ-Link Sulfo-NHS-

SS-Biotin (Thermo Fisher Scientific) in borate buffer for 15 min. Cells were then rinsed in 0.192 M glycine and 25 mM Tris-HCl (pH 8.3) solution to quench biotin excess and then harvested in RIPA buffer. Cells were cleared of insoluble debris by centrifugation for 30 min (14,000 r.c.f., 4oC). The supernatant was then incubated with 100 µl immunopure immobilised streptavidin (Thermo

Fisher Scientific) for 1 h at 4oC with constant rotation. Mixtures were then centrifuged for 1 min

(16,000 r.c.f., 4oC) and streptavidin beads were washed 3 times in RIPA buffer. Biotinylated proteins were then eluted with 100 µl Laemmli buffer.

2.3.1.2.4. Deglycosylation of HEK293 cell lysates For removal of N-linked glycosylation from proteins in Flp-In cell lysates (as prepared above), lysates were treated with 2 U of peptide N-glycosidase F (Roche) for 2 h in a 37oC waterbath. For removal of high mannose N-glycans, cleared lysates were treated with 100 mU of endoglycosidase

H (Roche) for 2 h in a 37oC waterbath. For removal of both forms of glycosylation, peptide N- glycosidase F and endoglycosidase H were added simultaneously at the amounts indicated above.

Enzyme treatment was performed on cleared lysates prior to addition of Laemmli buffer.

2.3.2. Primary astrocyte cultures

2.3.2.1. Generation and culture of primary astrocytes Primary cortical astrocyte cultures were established from mixed sex C57BL/6J mouse pups

(day 2-4) as previously described (Schildge et al., 2013). Briefly, the brain was removed and placed into a dish of ice-cold Hank’s balanced salt solution (HBSS) (Life Technologies), before the cortices were removed and the meninges dissected off the cortices. Dissected cortices were then cut into approximately eight pieces, and digested in HBSS with 0.25% Trypsin (Life Technologies) in a 37oC

59 water bath for 30 min. Cortical tissue was harvested with centrifugation (5 min, 300 g) and resuspended in astrocyte plating media [High glucose DMEM (Life Technologies), 10% fetal bovine serum, 1% Penicillin/Streptomycin] and plated in a T75 flask coated with 50 µg poly-D-lysine. Cells

o were maintained in a humidified incubator at 37 C in air/CO2 (19:1). After seven days, T75 flask was shaken for 30 min at 180 rpm at RT to detach microglia from the astrocyte layer, before media was discarded and replaced. A further 6 h of shaking at 240 rpm at RT was used to detach oligodendrocyte precursor cells, before media was discarded and replaced. The remaining astrocyte-enriched cell population was passaged into new flasks 7-14 days before harvest or plating

o for experiments, and maintained in a humidified incubator at 37 C in air/CO2 (19:1).

2.3.2.2. Treatment and harvest of primary astrocytes

2.3.2.2.1. Immunocytochemistry of primary astrocytes To confirm the purity of primary astrocyte cultures, as assessed by glial fibrillary acidic protein

(GFAP) immunofluorescence, 60,000 or 120,000 cells were plated per well on a two-chambered slide (previously coated with 500 µl of 0.1 mg/ml poly-D-lysine) in 2 ml astrocyte plating media

(described above). 2 slides were prepared simultaneously to allow for an isotype control during

o subsequent experiments. Slides were incubated overnight (12-24 hours) at 37 C in air/CO2 (19:1).

The following day, media was removed and the cells rinsed with PBS (RT) and then fixed with 4% paraformaldehyde (PFA) in PBS for 10 min at RT. PFA/PBS was then removed. Cells were then made permeable, and non-specific binding was blocked, by application of blocking buffer [PBS, 3% skim milk powder (Coles, Australia), 0.15% triton X-100] for 10 min at RT. Blocking buffer was then removed and primary antibody [1:500 rabbit anti-GFAP (Z0334, Agilent Technologies, USA)] or isotype control [rabbit IgG (Santa Cruz, USA), added to protein concentration equivalent to that of the primary antibody], diluted in blocking buffer, was then added to the cells for 1 h at RT. Primary antibody solution was then removed and cells were rinsed 3 times for 5 min in ice-cold PBS.

Se -rabbit Alexa488; 1:5000 4’,6-

Diamidinocondary- antibody2-phenylindole solution dihydrochloride [blocking solution (DAPI); containing D9542, 1:1000 Sigma α Aldrich, USA)] was then incubated for 1 h at RT whilst shielded from light (slides are now kept shielded from light henceforth). Secondary antibody solution was then removed and cells were rinsed 3 times for 5 min in ice-cold PBS. Plastic chambers were then removed from slides, and glass cover slips mounted using 2.5% w/v DABCO (1,4-diazabicyclo-[2,2,2]-octane; D2522, Sigma Aldrich), 10% w/v polyvinyl alcohol (P8136, Sigma Aldrich), 25% w/v glycerol (G5516, Sigma Aldrich), and 100 mmol/L Tris buffer (pH 8.7), as adapted from a previous method (Ono et al., 2001). Slides were imaged on a LSM 7 Duo confocal microscope (Zeiss, Germany).

2.3.2.2.2. Harvest of primary astrocytes To attempt to detect GPR37L1 protein in primary astrocyte cultures, a pellet of approximately 2 x

106 cells (approximately half the population of a T75 flask after following the protocol in section

2.3.2.1) was suspended in 100 µl RIPA buffer and allowed to dissolve for 30 min on rotating mixer.

Lysates were cleared of insoluble debris by centrifugation for 15 min (4oC, 16,000 r.c.f.). The concentration of protein in the resulting supernatant was measured by Direct Detect (Merck, USA).

Supernatant was then mixed with Laemmli buffer (final concentrations: 63 mM Tris, 50 mM dithiothreitol, 80 mM SDS, 10% glycerol, 0.004% bromophenol blue, pH 6.8), and incubated for 15 min in a 37oC water bath.

2.3.3. Mouse cerebellar slice cultures

2.3.3.1. Generation and culture of cerebellar slice cultures Organotypic cerebellar slice cultures were prepared essentially as described for the hippocampus (Stoppini et al., 1991; Cimarosti et al., 2006). Briefly, 15 day old mixed-sex C57BL/6J or GPR37L1null/null mice (Min et al., 2010) were anaesthetised in a chamber filled with 5% isofluorane in air, before decapitation. Brains were removed and placed in a petri dish with ice-cold

HBSS (Life Technologies) supplemented with 30 mM D-glucose and penicillin/streptomycin (100

U/ml), and saturated with carbogen (95% oxygen, 5% carbon dioxide). The cerebella were removed and sliced into 400 µm sagittal sections using a McIlwain tissue chopper (Stoelting Co.,

USA). Sections from a single cerebellum were plated evenly across two Millicell cell culture inserts

(Merck), in six-well plates (approximately six to eight slices per well) containing 1 ml of 50% minimum essential media (Life Technologies), 25% HBSS (Life Technologies), 25% heat-inactivated horse serum, 36 mM D-glycose, 25 mM Hepes, 4 mM NaHCO3, 2mM GlutaMAX, 1% Fungizone, and penicillin/streptomycin (100 U/ml). All cultures were maintained at 37oC in an atmosphere of 5%

CO2.

2.3.3.2. Treatment and harvest of cerebellar slice cultures

2.3.3.2.1. Inhibition of metalloproteases in cerebellar slice cultures For experiments with BB94, slices of a cerebellum were plated across two wells as described, with one of the two wells containing BB94 at the indicated concentration. 24 h later, slices from each well were harvested into a microfuge tube, without media, and snap frozen in liquid nitrogen

2.3.3.2.2. cAMP measurement in cerebellar slice cultures For measurement of cAMP accumulation, slices of a cerebellum were plated across two wells as previously described, with one of the two wells containing the phosphodiesterase inhibitor 3- isobutyl-1-methylxanthine (IBMX, 1 mM). Exactly 1 h later, slices were harvested into a microfuge tube and snap frozen in liquid nitrogen. To ensure sufficient tissue for the assay, slices from two to four separate pups were pooled and considered to represent n=1. Tissue was homogenized by hand using a conical pestle, in microfuge tubes partially submerged in a bath of ethanol and dry ice to ensure tissue did not thaw. cAMP in cerebellar slice homogenates was then determined using the

DirectX Direct cAMP enzyme-linked immunosorbent assay (ELISA) kit (Arbor Assays, USA) as per manufacturer’s instructions.

2.3.4. Saccharomyces cerevisiae

2.3.4.1. Maintenance of Saccharomyces cerevisiae 11 individual FUS1- -galactosidase Saccharomyces cerevisiae yeast reporter strains were obtained from GlaxoSmithKlineregulated β (UK). As described by Dowell et al. (2009), each strain was

x modifiedfusion proteins, to express where the x endogenousis the last five yeast amino Gpa1 acids Gα of subunit, each of orthe chimeras h consisting of Gpa1/Gα host yeast strains were stored at -80oC, and streaked onto YPD agar uman(70 g premixedGα G proteins. powder Stocks in 1 of L water, ForMedium, UK) plates to establish working stocks. Plates were incubated at 30oC for approximately 2-4 days, or until colonies of 3-5 mm diameter had grown. Colonies were then picked for use in the assay (as described below), or plates were sealed with plastic paraffin film and stored facedown at 4oC for later use (for up to 4 weeks). Unless otherwise stated, all substances used in yeast culture were sterilised by autoclave (15 min, 121oC) and stored at RT, and culture was performed with aseptic technique, using presterilised consumables, under flame.

2.3.4.2. Transformation of Saccharomyces cerevisiae A single colony was picked from each working stock plate and transferred into 10 ml of YPD broth (50 g premixed powder in 1 L water), and then incubated for 12-24 h at 30oC in a shaking incubator set to 200-250 oscillations per min. These cultures were then used to establish 50 ml pre- transformation cultures in YPD broth, diluted to equal cell density [OD600 = 0.2, measured with an

Ultraspec 10 cell density meter (Amersham Biosciences, UK)]. Pre-transformation cultures were

o incubated at 30 C (200-250 oscillations per min in a shaking incubator) until an OD600 of approximately 1.0 was reached (4-6 h). Cells were then harvested by centrifugation (RT, 2 min,

~1000 r.c.f.) and the supernatant was discarded. Cell pellet was then resuspended by gentle swirling in 10 ml water before pelleting the cells again by centrifugation (RT, 2 min, ~1000 r.c.f.) and discarding the supernatant. 250 µl of LiAc/TE solution [5 ml of 1 M Lithium Acetate, 500 µl 1 M

63 sterile Tris-HCl (pH 7.5), 100 µl 500 mM EDTA] was added to each cell pellet and gently swirled to resuspend cells. The 250 µl cell suspension was used to create 50 µl aliquots for transformation.

Four 50 µl aliquots per strain underwent the transformation protocol, with one aliquot receiving no DNA (negative control), one receiving empty p426GPD (high-copy vector), one receiving p426GPD-GPR37L1-eYFP and one receiving p426GPD- -GPR37L1-eYFP. With the exception of the negative control, a total of 1 µg of DNA (in water) wasΔ122 added to each 50 µl aliquot.

300 µl of LiAc/TE solution (with 40% PEG3350) was then added to each tube and gently mixed by inverting tubes 3-4 times. All transformation mixtures were then incubated at 30oC (30 min), subjected to heat shock at 42oC (15 min), and removed to RT. 200 µl of each transformation solution was then spread onto selection plates [SD agar, -ura, -met; 20 g agar, 20 g glucose, 6.8 g yeast nitrogen base without amino acids (Sigma-Aldrich), 1.85 g synthetic complete amino acid drop-out mix (-ura, -met) (ForMedium) in 1 L water] which were then grown at 30oC for 2-4 days, or until colonies of 3-5 mm diameter had grown. 8 colonies were then picked from each plate and streaked onto new selection plates (divided into 8 segments, one segment per streaked colony) and grown as per the previous step.

2.3.4.3. Saccharomyces cerevisiae G protein-coupling assay preparation Liquid cultures of the transformants isolated above were established by spiking individual transformants into 1 ml of SD broth [SD broth, -ura, -met, 20 g glucose, 6.8 g yeast nitrogen base without amino acids (Sigma-Aldrich), 1.85 g synthetic complete amino acid drop-out mix (-ura, - met) (ForMedium)] in wells of a 24-well plate. Typically, 6 of the 8 isolated transformants were cultured per condition. Cultures were grown overnight (12-24 h) in a shaking incubator set at 30oC and 200-250 oscillations per min. Next 1 µl of each mini-culture established in the previous step was spiked into 100 µl of ‘pre-assay media’ [37.4 ml SD broth (-ura, -his)(described in next sentence), 5 ml 10x YNB solution (6.8 g yeast nitrogen base without amino acids, 100 ml water,

o filter sterilised, stored at 4 C), 5 ml 10x BU salt solution (37 g Na2HPO4 anhydrous, 30 g NaH2PO4, 1

L water, pH 7), 2.5 ml 40% glucose (40 g glucose, 100 ml water, filter sterilised), 100 µl 1 M 3-AT solution (1 M 3-amino-1,2,4-triazole dissolved in water, filtered sterilised, and stored at 4oC whilst protected from light. Obtained from ForMedium)] in wells of a 96-well plate. SD broth (-ura, -his) was prepared as follows: 1.85 g synthetic complete amino acid drop-out mix (-ura, -his)

(ForMedium) was added to 850 ml water, pH adjusted to 7.0, and autoclaved. After sterilisation,

100 ml of 10 x YNB solution and 50 ml 40% glucose solution were added. After establishing pre- assay cultures, plates were then briefly shaken (5 s) on an orbital shaker. 10 µl of each well was then transferred into 90 µl of ‘FDG assay media’ (as per ‘pre-assay media’ with Fluorescein di- - galactopyranoside added to a final concentration of 11.11 µM) in duplicate in the wells of a blackβ

96-well plate. Final culture plates were then grown for 24 h in a shaking incubator at 30oC and 200-

250 oscillations per min, before fluorescence was detected with a PHERAstar FS (BMG Labtech,

Germany) (excitation 485 nm/emission 520 nm). Where compounds were screened for ability to modulate reporter gene expression, compounds are diluted in 10 µl pre-assay media and added to black 96 well plates along with 80 µl of ‘FDG assay media’ (instead of 90 µl) and 10 µl of pre-assay culture. Duplicates were averaged before further data analysis.

2.4. In vivo experiments

2.4.1. Mouse line generation The European Conditional Mouse Mutagenesis (EUCOMM) program, a member of the

International Knockout Mouse Consortium [IKMC, (Bradley et al., 2012)], developed a powerful gene targeting strategy for conditional gene inactivation using Cre/loxP and Flp1/FRT gene- trapping (Skarnes et al., 2011). This resource was used to generate C57BL/6J GPR37L1flx/flx,

GPR37L1wt/wt, GPR37L1wt/KO, GPR37L1KO/KO and GPR37L1lacZ/wt mice as follows: C57BL/6N embryonic steam (ES) cells containing the GPR37L1 tm1a allele (to be subsequently described),

65 were obtained from EUCOMM (clone number EPD0486_2_B09), and subsequent creation of heterozygote tm1a mice was outsourced to the Monash University Embryonic Stem Cell-to-Mouse

(ES2M) initiative (Cotton et al., 2015). The founder GPR37L1 tm1a allele, which allows generation of all subsequent experimental mice, was designed by EUCOMM as follows: An IRES:lacZ trapping cassette (internal ribosome entry site upstream of lacZ) was placed 5’ of a loxP-flanked, promoter- driven, neomycin-resistance selection cassette, which lies immediately upstream of exon 2 of

GPR37L1 (which has only two exons). The neomycin-resistance cassette permitted in vitro selection of targeted ES cells, while the lacZ cassette was used in generation of tm1b -galactosidase reporter

-galactosidase is expressed instead of GPR37L1. A thirdβ loxP site was inserted immediatelyanimals, where after β exon 2 of GPR37L1 to facilitate its removal and therefore generation of experimental GPR37L1 knockout animals. To facilitate removal of both the lacZ and neomycin- resistance cassettes, thus allowing restoration of the GPR37L1 tm1a allele to that of a pseudo-wild- type allele (tm1c, also referred to as ‘floxed’), two FRT (flippase recognition target) sites were inserted into the allele: one upstream of the lacZ cassette, and one between the neomycin- resistance cassette and the loxP -galactosidase reporter mice

(tm1b) were generated by crossingsite tm1a immediately mice with before global exon“Cre deleter” 2. β mice [a transgenic B6.C-

Tg(CMV-Cre)1Cgn/J (The Jackson Laboratory, USA, Stock number 006054) mouse expressing Cre recombinase ubiquitously under the control of a human cytomegalovirus (CMV) promoter], resulting in recombination at loxP sites and deletion of the neomycin selection cassette and exon 2.

Tm1b mice were then bred to remove the potentially confounding Cre transgene [as outlined in

Coleman et al. (2015)] and heterozygotes (GPR37L1wt/lacZ) were used for experiments herein.

‘Floxed’ mice (tm1c, GPR37L1flx/flx), were generated as the founder line for creation of true knockout mice, by the crossing of heterozygous tm1a mice with transgenic C57BL/6J mice expressing an enhanced variant of Saccharomyces cerevisiae FLP1 recombinase (FlpE) in all tissues,

-actin promoter (transgenic B6.Cg-Tg(ACTFlpE)9205Dym/J, available from The under the human β 66

Jackson Laboratory, USA, Stock number 005703). Where offspring inherited both the tm1a allele and the FlpE allele, the FRT-flanked region of tm1a was excised and recombined, to remove both the lacZ and neomycin-resistance cassettes and restore GPR37L1 expression (recombination was confirmed by PCR). GPR37L1 ‘knockout’ mice (tm1d, GPR37L1KO/KO) were subsequently generated by crossing GPR37L1flx/flx mice (tm1c) with global Cre deleter mice (as used above), as exon 2 of

GPR37L1 was still loxP-flanked following removal of lacZ and neomycin-resistance cassettes. FlpE and Cre recombinase transgenes were bred out of the line, as outlined in Coleman et al. (2015).

Although all mice described above were generated on a C57BL/6J background, the initial ES cells were on a C57BL/6N background, and thus, inbred wild-type C57BL/6J mice were not the most suitable control. Therefore, experimental and control mice were bred by a two-tiered stem and expansion colony management strategy as per Coleman et al. (2015). Briefly, stem breeding was performed by the mating of heterozygous tm1d mice. Resulting offspring were inbred for a single generation to expand colonies of GPR37L1wt/wt, GPR37L1wt/KO and GPR37L1KO/KO mice for experimental use.

GPR37L1 null and cardiac-specific over-expressors, as reported by Min et al. (2010) were obtained via material transfer agreement (Daiichi Sankyo, Japan) and reanimated from frozen sperm by

Charles River Laboratories (Japan).

2.4.2. Mouse line maintenance Mice were group-housed in a temperature-controlled environment with a 12 h light-dark cycle

(lights on at 0700). Standard chow and water were available ad libitum.

2.4.3. Blood pressure measurement

2.4.3.1. Hemodynamic measurement and analysis At age 10-12 weeks, mice were anaesthetised with 1-2% isofluorane (Zoetis, USA) in oxygen

(flow rate 0.5 L/min), delivered via nose cone regulated using a mechanical ventilator (150 strokes/min, 200 µl stroke volume). Hemodynamics in the aorta and left ventricle were recorded with a high-fidelity pressure transducer (Millar Instruments, USA) following its insertion into the right carotid artery as described previously (Li et al., 2008; Pacher et al., 2008) (performed by Dr

Jianxin Wu). Heart rate was maintained at approximately 500 bpm by manual adjustment of the isofluorane concentration. Data were recorded and analysed using AcqKnowledge software

(version 3.9.0 for Windows, BIOPAC Systems, USA) in a blinded fashion. Hemodynamics on osmotic mini pump animals were recorded using a Transonic Scisense (FTH-1211B-001, USA) high-fidelity pressure transducer.

2.4.3.2. Radiotelemetry blood pressure measurement At 12-14 weeks of age, mice were anaesthetised with 2% isofluorane (Abbott Laboratories,

USA), with the rate adjusted to provide a deep surgical plane of anaesthesia. PA-C10 telemeters

(Data Sciences International, USA) were implanted as previously described (Butz et al., 2001), with the pressure-sensing catheter inserted into the left carotid artery. Analgesia was administered before and after surgery (Carprofen, 0.5 mg/100g, IP; Norbrook, UK). Hydration was maintained by administration of 0.5 ml 0.9% NaCl (IP), given at time of operation and again 24 h later.

Approximately 2 weeks after telemeter implantation, blood pressure monitoring was performed twice-daily in the home cage, between 09:00-12:00 (light phase) and between 21:00-24:00 (dark phase) for nine days (performed and analysed by Dr Nikola Jancovski). The average of the nine days of measurements in each phase was taken to give a single value per animal per phase.

2.4.4. Angiotensin II infusion Osmotic mini pumps (product 1007D, Alzet) were filled with either vehicle (0.15 mol/L NaCl, 1 mmol/L acetic acid) or AngII (A9525, Sigma Aldrich) at a concentration adjusted to allow 2 mg/kg/day release of AngII for 7 days. After filling, pumps were primed by incubation in 0.9% NaCl for 16 h at 37oC. Mice (10-12 week-old) were anaesthetised [inhaled 3-4% isofluorane (Zoetis) in air (flow rate: 0.8-1.0 L/min) delivered by nose cone] and pumps implanted subcutaneously, as per the manufacturer instructions. Following pump implantation, topical analgesia was administered in the form of bupivacaine (AstraZeneca, UK), and systemic analgesia was administered (SC) in the form of buprenorphine (0.075 mg/kg; Reckitt Benckiser, UK), before completion of surgery and recovery from anaesthesia. Animals were allowed to recover from anaesthesia on a heated pad, before being single-housed for the remainder of the study.

2.4.5. Tissue morphometry Tissues were retrieved after sacrifice, briefly rinsed of blood in 0.9% NaCl and then blotted dry.

Tissue weights were measured immediately and expressed as a ratio of each animal’s own tibia length, measured with caliMAX fine calipers (WIHA, Germany).

2.4.6. Microscopy and imaging

2.4.6.1. Tissue fixation for β-galactosidase imaging Tissue fixation was performed as adapted from a previous method (Chen et al., 2010) weeks of age, GPR37L1wt/wt and GPR37L1lacZ/wt mice of either sex were anaesthetised with. At ≥125% isofluorane in oxygen (flow rate 1 L/min) delivered via nose cone. The thoracic cavity was opened and the left ventricle was punctured with a needle that was then used to perfuse the heart with phosphate-buffered saline (PBS) (1 min; flow rate: 20 ml/min), achieving sacrifice by exsanguination. Immediately after, tissues were fixed by further 2 min perfusion with 4% PFA in

PBS (flow rate 20 ml/min). Following tissue fixation (confirmed by observing blanching of tissue

69 and the onset of tail rigidity), brain, heart and kidney were collected and further fixed in 4%

PFA/PBS for 48 h at RT (PFA/PBS volume was at least 5 times that of tissue volume). After post- fixation, tissue was stored in sucrose cryoprotectant buffer (5 mmol/L Trizma base, 77 mmol/L

o NaCl, 4 mmol/L Na2HPO4, 1.3 mmol/L NaH2PO4, 20% w/v sucrose) at 4 cryoprotected tissue was embedded in Tissue Freezing Medium (TFM-5, TriangleC for ≥48Biomedical, h. Fixed USA) and and frozen. Frozen tissue was cut into 12 µm sections using a CM1950 Cryostat (Leica Biosystems,

Germany) and mounted on Superfrost Plus Slides (Thermo Fisher Scientific).

2.4.6.2. β-galactosidase immunofluorescent staining The protocol was performed at RT unless otherwise specified. Slides containing mounted tissue sections from the above procedure were washed for 10 min, 3 times, in immunobuffer [TPBS (10 mmol/L Tris, 8 mmol/L Na2HPO4, 2.6 mmol/L NaH2PO4, 0.9% NaCl, 0.05% thimerosal) with 0.3%

Triton X-100]. Non-specific antibody binding was blocked by incubating each section in immunobuffer containing 10% v/v normal donkey serum (NDS) (017-000-121, Jackson

ImmunoResearch Laboratories, USA) for 30 min, in a humidifying chamber. Primary antibodies

[1:250 chicken anti- -galactosidase (AB9361, Abcam, Cambridge, UK) for all tissue, 1:500 rabbit anti-GFAP (Z0334, Agilentβ Technologies, USA) for brain only] were diluted in immunobuffer with

10% v/v NDS and applied to each section for 16 h, in a humidifying chamber at 4oC. Slides were then washed for 10 min, 3 times, in TPBS. For the remainder of the protocol, slides were shielded from ambient light. Secondary antibodies [1:500 donkey anti-chicken Cy3 (703-165-155, Jackson

ImmunoResearch Laboratories) for all tissue, 1:500 donkey anti-rabbit Alexa Fluor 647 (711-605-

152, Jackson ImmunoResearch) for brain only], DAPI (1:5000 4’,6-Diamidino-2-phenylindole dihydrochloride; D9542, Sigma Aldrich, USA) nuclear stain and wheat germ agglutinin (WGA) membrane stain (for heart and kidney only, 1:500 Alexa Fluor 488 WGA conjugate; W11261,

Invitrogen) were diluted in immunobuffer containing 10% v/v NDS and applied to each section for

1 h, in a humidifying chamber. Slides were then washed for 10 min in TPBS. To reduce

70 autofluorescence, each section was then treated with 0.1% w/v Sudan black B (S2380, Sigma

Aldrich) in 70% v/v ethanol for 1 h, in a humidifying chamber. Excess dye was then removed by washing for 10 min, 3 times, in TPBS. Slides were then dried for 16 h in a humidifying chamber before being sealed with glass cover slips mounted using 2.5% w/v DABCO (1,4-diazabicyclo-

[2,2,2]-octane; D2522, Sigma Aldrich), 10% w/v polyvinyl alcohol (P8136, Sigma Aldrich), 25% w/v glycerol (G5516, Sigma Aldrich), and 100 mmol/L Tris buffer (pH 8.7), as adapted from a previous method (Ono et al., 2001). Slides were imaged on a LSM 7 Duo confocal microscope (Zeiss,

Germany).

2.4.6.3. Cardiomyocyte density Cardiomyocyte density was determined using an adaptation of a previous method (Li et al.,

2008). Briefly, left ventricle tissue was fixed in 4% PFA/PBS for 24 h at RT before post-fixation, sectioning and processing as described above for immunofluorescent staining and confocal microscopy. Cell membranes were visualised using 1:500 Alexa Fluor 488 WGA conjugate. Using

ImageJ, an observer (Ms Margaret Mouat), blinded to sex, genotype and treatment, counted cardiomyocytes from two representative areas of each left ventricle and averaged them to give one value per animal. Cardiomyocyte counts were expressed as cells/mm2. Cardiomyocyte density was used instead of cardiomyocyte cross-sectional area as the former was subject to less inter-observer variability (data not shown).

2.4.6.4. Fibrosis staining Left ventricle tissue was fixed as described for cardiomyocyte density. Cytoplasm (red) and collagen (blue) were stained using Masson’s Trichrome stain as per the manufacturer’s protocol

(HT15, Sigma Aldrich) and performed by Ms Margaret Mouat. Coverslips were mounted onto slides using DEPEX mounting media (VWR International, USA). Sections were visualised with bright field microscopy on a DM6000 power mosaic microscope (Leica Biosystems) and images were retained

71 in RGB format. The degree of fibrosis was quantified using ImageJ and expressed as a percentage of area occupied by blue-hue pixels relative to total cross-sectional tissue area, as adapted from previous methods (Tejada et al., 2016).

2.5. Molecular biology

2.5.1. Western blotting and in-gel imaging

2.5.1.1. Sample preparation

2.5.1.1.1. Cell lysates After cell harvest of HEK293 cells or primary astrocytes, as described above, lysates (mixed with loading buffer as described) were loaded onto gel for SDS-PAGE.

2.5.1.1.2. Tissue homogenates All immunoblots of homogenized brain (including cerebellum), heart and kidney tissue were performed identically. Snap frozen tissues were homogenized in RIPA buffer using a PRO200 homogenizer (Bio-Gen, USA). Tissue lysates were cleared of insoluble debris by centrifugation at

16,000 r.c.f. for 15 min. Equal amounts of protein (as determined by Merck Direct Detect) were mixed with 4x LDS sample buffer (Invitrogen), and 50 mM dithiothreitol before gel loading for SDS-

PAGE.

2.5.1.1.3. Crude membrane preparations Crude membrane preparations from tissues and cerebellar slice cultures were performed as follows: Tissue was homogenized in a sucrose buffer (5 mM HEPES, 250 mM sucrose, 1 mM EGTA,

Complete EDTA-free protease inhibitor cocktail, pH 7.4) using a PRO200 homogenizer (Bio-Gen).

Homogenates were centrifuged at 600 r.c.f. for 10 min at 4oC, and the subsequent supernatant was collected and further centrifuged at 16,000 r.c.f. for 45 min at 4oC. The supernatant was discarded and the pellet resuspended in 100 µl HEMD buffer (20 mM HEPES, 12.5 mM MgCl2, 1mM EGTA, 500

µM dithiothrietol, Complete EDTA-free protease inhibitor cocktail) and passed ~10 times through a

29 gauge needle. Equal amounts of protein (as determined by Merck Direct Detect) were mixed with 4x LDS sample buffer (Invitrogen), and 50 mM dithiothreitol before gel loading for SDS-PAGE.

2.5.1.2. SDS-PAGE Cell lysates and tissue homogenates were separated with 4-12% Bis-Tris SDS-PAGE in NuPAGE

MOPS SDS running buffer (Invitrogen). For SDS-PAGE prior to in-gel fluorescence imaging, running buffer was kept ice-cold and SDS-PAGE was performed at 4oC (to maintain eYFP excitability).

2.5.1.3. In-gel fluorescence imaging Following SDS-PAGE, the gel was imaged using the FLA-5100 imager (Fujifilm, Japan).

Densitometry was performed using ImageJ software (Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA, https://imagej.nih.gov/ij/, 1997-2016).

2.5.1.4. Western blotting After SDS-PAGE, proteins were then transferred to 0.45 µm pore-size polyvinylidene difluoride

(PVDF) transfer membranes (Merck, USA) and treated with antibodies, as detailed below. In all cases, three 10 min washes with Tris-buffered saline containing 0.1% Tween 20 (TBS-T) were performed after primary and secondary antibody application.

2.5.1.4.1. Western blotting of HEK293 lysates Non-specific binding was blocked with Tris-buffer saline containing TBS-T and 5% skim milk powder (Coles, Australia) for 1 h at RT. C-terminal eYFP epitope was detected with an antibody directed against GFP (Abcam, UK) diluted 1:10,000 in TBS-T with 5% skim milk powder at 4oC overnight (12-24 h). Primary antibody was detected with a horseradish peroxidase (HRP)- conjugated IgG anti-rabbit secondary antibody (GE healthcare, UK) diluted 1:10,000 in TBS-T with

5% skim milk powder for 1 h at RT. HRP was detected with Western Lightning ECL reagent

(PerkinElmer, USA).

2.5.1.4.2. Western blotting of mouse tissue, slice culture and primary cell homogenates Non-specific binding was blocked with TBS-T and 5% skim milk powder (Coles, Australia) for 1 h at RT. Endogenous GPR37L1 was detected by goat anti-GPR37L1 (C-12) antibody (1:1000 in TBS-

T and 5% skim milk powder; sc-164532, Santa Cruz Biotechnology, USA) overnight (12-24 h) at 4oC and then HRP-conjugated rabbit anti-goat antibody (1:5000-1:7500 TBS-T and 5% skim milk powder; 61-1620, Invitro - - galactosidase antibody (1:5000gen) for in 1TBS h at-T RT.and β5%galactosidase skim milk waspowder; detected A-11132; by rabbit Invitrogen) anti β overnight (12-24 h) at 4oC and then donkey anti-rabbit (1:15000, in TBS-T and 5% skim milk powder; NA934; GE Healthcare, Australia) for 1 h at RT. For detection of GAPDH after endogenous

GPR37L1, membranes were rinsed in TBS-T before further non-specific binding was blocked with

5% bovine serum albumin (BSA) (Bovogen Biologicals, Australia) in TBS-T for 1 h at RT. GAPDH was detected by rabbit anti-GAPDH antibody (1:10000 in TBS-T and 2.5% BSA; 14C10, Cell

Signaling, USA) for 1 h at RT and then donkey anti-rabbit (1:20000 in TBS-T and 2.5% BSA, NA934,

GE Healthcare) for 1 h at RT. SuperSignal™ West Pico (34080, Invitrogen) or Clarity™ (1705060,

Bio-rad, USA) were used as chemiluminescent substrates.

2.5.1.5. Quantification of immunoblot and fluorescence images Relative pixel density of bands of interest was measured using Image J software (Rasband, W.S.,

ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA, https://imagej.nih.gov/ij/,

1997-2016). The background pixel density was also recorded for each image and subtracted from density scores of all measured bands. Where indicated, GAPDH pixel density was also measured

(and background levels subtracted) and used to adjust pixel density scores of other bands of interest to account for differences in protein loading.

2.5.2. Verification of TIMP activity Verification of TIMP-1, -2, -3 and -4 activity (as used in Flp-In HEK293 experiments) was performed according to the manufacturer’s protocol (R&D systems) as follows: 100 µg/ml recombinant human MMP-2 (R&D Systems) in Assay Buffer [50 mM Tris, 10 mM CaCl2, 150 mM

NaCl, 0.05% Brij-35 (w/v), pH 7.5] was activated by incubation at 37oC for 1 h with 1 mM p- aminophenylmercuric acid (APMA). A curve of MMP-2 inhibition points was established by serial dilutions of TIMPs in Assay Buffer (2 µM to 2 nM for TIMP-1 and -2, 5 µM to 2 nM for TIMP-3 and -

4). After activation, MMP-2 was diluted to 12.5 µg/ml in Assay Buffer. Reaction mixtures were established by addition of 25 µl of activated MMP-2 (12.5 µg/ml) to 16 µl of each TIMP serial dilution and 119 µl of Assay Buffer. Reaction mixtures were incubated at 37oC for 2 h (an enzyme- only control without TIMPs was included). After incubation, reaction mixtures were diluted 5-fold in Assay Buffer. 50 µl of each diluted reaction mixture was added to wells of a black 96-well plate, before injection of 50 µl fluorogenic MMP-2 substrate (10 µM, MCA-Pro-Leu-Gly-Leu-DPA-Ala-Arg-

NH2, R&D systems) and measurement of fluorescence kinetics for 5 min by FLUOStar

(BMG)(excitation 320 nm, emission 405 nm).

2.5.3. Quantitative PCR Tissue was homogenized in 1000 µl TRIzol (15596026, Thermo Fisher Scientific, USA) using a

PRO200 homogenizer (Bio-Gen, USA) and RNA was extracted as follows: 200 µl chloroform was added to tissue homogenate and shaken vigorously for 15 s. After 5 min incubation at RT, and 15 min centrifugation (16,000 r.c.f. at 4oC), approximately 500 µl of the resulting aqueous layer was removed and combined with 5 µl of glycogen (5 mg/ml stock, Thermo Fisher Scientific) and 500 µl of isopropanol. Tubes were shaken vigorously for 15 s before 10 min incubation at RT. RNA was pelleted by 15 min centrifugation (16,000 r.c.f. at 4oC), before supernatant was removed and the pellet gently washed by addition of 1 ml ethanol (75%). Centrifugation was repeated for 10 min

(16,000 r.c.f. at 4oC). After removal of supernatant, RNA pellets were air-dried under light for 30-60

75 min, and then thoroughly resuspended in 100 µl of DEPC-treated water. RNA was further purified as follows: after resuspension in water, 10 µl sodium acetate (3 M) and 275 µl ethanol (100%) was added. Tubes were mixed by vortex and stored overnight (12-24 h) at -20oC. Next, tubes were centrifuged for 15 min (16,000 r.c.f. at 4oC), before supernatant was removed and 1 ml of ethanol

(75 %) added. Tubes were centrifuged again for 10 min (16,000 r.c.f. at 4oC) before supernatant was removed and RNA pellet air-dried as before. RNA pellet was thoroughly resuspended in 100 µl

DEPC-treated water, and then stored in aliquots at -80oC until required for cDNA synthesis.

DNA contamination was removed from extracted RNA with TURBO DNAse kit treatment (AM1907,

Invitrogen, USA) before cDNA synthesis was performed on 1 µg of RNA using SuperScript® III First-

Strand Synthesis Supermix as per manufacturer’s protocol (11752050, Invitrogen). cDNA was stored at -20oC until use, and diluted by a factor of 10 before use in quantitative PCR (qPCR). qPCR was performed on cDNA using TaqMan Gene Expression Master Mix (4369016, Applied Biosystems,

USA) and Taqman probes (Applied Biosystems). All qPCR was performed using a LightCycler 480

(Roche, Switzerland) (cycling conditions: preincubation at 95oC for 10 min; 55 amplification cycles of 95oC for 10 s, 58oC for 30 s, 72oC for 30 s; final cooling at 50oC for 10 s). Probe Mm00661872_m1 was used to detect GPR37L1 in mouse tissue. The following probes were used to detect components of the fetal gene program in mouse tissue (performed by Ms Margaret Mouat): ANP

- -SkA

(Mm01255747_g1), BNP (Mm01255770_g1), β MHC (Mm00600555_m1) and α quantification(Mm00808218_g1). of GPR37L1 ΔCt val andues fetal were gene calculated program in transcripts reference in to EUCOMM β2M (Mm00437762_m1) mice. Relative mRNA for abundance was determined using the 2- method (Livak et al., 2001). ΔΔCt

2.5.4. RNA sequencing Human post-mortem brain RNA sequencing data was obtained from the Genotype-Tissue

Expression Consortium (GTEx) (Consortium, 2015). GTex raw sequence counts were examined for

76 differentially expressed genes using the edgeR statistical package (Robinson et al., 2010), where samples were grouped by sex for each brain tissue sample (performed by Dr David T. Humphreys).

2.6. Statistics and data analysis

2.6.1. In vitro All data are presented as mean ± standard error of the mean (s.e.m.). In Saccharomyces cerevisiae experiments, n represents data derived from a single transformant. Data were analysed with a one-way ANOVA with Bonferroni’s post hoc analysis. In mouse cerebellar slice culture assays, n represents a data point derived from pooled tissue from 2-4 mice. Data were analysed with one- to be statisticallyway ANOVA significant. with Dunnett’s multiple comparison’s test. In all tests, p≤0.05 was considered

2.6.2. In vivo All data are presented as mean ± s.e.m.. The number of independent observations (n) in animal/human data represents a single animal/donor. Where multiple measurements were taken per animal, measurements were averaged to give a single observation (n) per animal. Where two untreated genotypes (per sex) were analysed, a two-tailed Student’s t-test was used, or a one-tailed

Mann-Whitney test, as indicated. Where >2 untreated genotypes (per sex) were analysed, a one- way ANOVA was used with Dunnett’s multiple comparisons tests, or Kruskal-Wallis test with

Dunn’s multiple comparisons test where ANOVA assumptions were violated. Where two genotypes

(per sex), with and without treatment, were analysed, two-way ANOVA was used with Tukey’s multiple comparisons test. For all tests, p ed to be statistically significant.

≤0.05 was consider

Chapter Three: GPR37L1 is a constitutively active orphan G protein-coupled receptor subject to regulation by metalloprotease-mediated N-terminal cleavage

3.1. Introduction With over 800 unique members, GPCRs are the largest family of integral membrane proteins in the human genome (Katritch et al., 2013), and are paramount to the eukaryotic cell’s ability to mount an intracellular signal in response to an extracellular stimuli (Coleman et al., 2017). Owing to the high degree of specificity between ligand and GPCR, the extracellular location of their ligand binding site(s), and the discrete profiles of GPCR expression throughout tissues and organs, these receptors are a major class of pharmacological target (Vassilatis et al., 2003; Tang et al., 2012). By virtue of evolutionary diversification, each GPCR has remarkable specificity for its cognate stimuli, which may include photons, ions, odorants, nucleotides, amino acids, peptides or proteins

(Bockaert et al., 1999), interactions with which stabilise a GPCR in an active conformation, resulting in heterotrimeric G protein activation, subsequent activation of downstream effector molecules, and signal amplification (Rosenbaum et al., 2009). Notably, some GPCRs have constitutive activity, signalling in the absence of a ligand, and this may be observed both in vitro and in vivo (Milligan,

2003).

Of interest to our laboratory is GPR37L1; a GPCR implicated in the physiology of blood pressure regulation (Min et al., 2010), suggesting it to be a promising target for the development of novel anti-hypertensive treatments. Despite being identified as a peptide receptor (and belonging to the family A, or rhodopsin-like, GPCRs) (Valdenaire et al., 1998), GPR37L1 has yet to be paired with an endogenous ligand and remains a so-called orphan receptor. It was originally identified as endothelin B receptor-like protein 2 or GPCR/CNS2 by two independent groups (Valdenaire et al.,

1998; Leng et al., 1999), and has high central nervous system expression, with lower levels of transcript found in cardiovascular organs such as the heart and kidney (Leng et al., 1999). Whilst

GPR37L1 shares 32% identity and 56% similarity with the endothelin B receptor (Valdenaire et al.,

1998; Leng et al., 1999), GPR37L1 does not bind endothelin or related peptides (Valdenaire et al.,

1998). The closest relative of GPR37L1 is GPR37, which is also an orphan receptor, and thus cannot

79 guide the ligand identification process for GPR37L1. However, a study has reported that prosaptide, a synthetic 14-amino acid, neuroprotective peptide derived from the saposin C domain of

i signalling (Meyer et al., 2013), prosaposin,but without stimulates independent both validation,GPR37 and bothGPR37L1 receptors to signal remain via Gα as orphans. Fascinatingly, our laboratory identified three peptides in previous analyses of human CSF (Stark et al., 2001; Schutzer et al., 2010; Zhao et al., 2010) with identical amino acid sequences to three regions within the N- terminus of GPR37L1. Accordingly, I hypothesise that the receptor’s N-terminus is subject to proteolysis and perhaps is involved in activation of the receptor, in a mechanism comparable to that of the PARs.

Here I show that GPR37L1 is subject to N-terminal proteolysis, occurring both in cultured cells and in the mouse cerebellum. In a mechanism of receptor regulation distinct from that of the PARs,

GPR37L1 displays robust constitutive activity, and is actually inactivated by proteolysis. The contents of this chapter have been adapted from a previously published manuscript (Coleman et al.,

2016), for which I share first authorship with Dr Tony Ngo, and to which I contributed experimental data and co-wrote the manuscript.

3.2. Results

3.2.1. GPR37L1 is subject to N-terminal processing Hypothesising that the presence of GPR37L1 N-terminal fragments in human CSF (Stark et al.,

2001; Schutzer et al., 2010; Zhao et al., 2010) indicates that the receptor undergoes regulated N- terminal processing, doxycycline-inducible, Flp-In HEK293 cells were stably transfected with human GPR37L1, fused to enhanced yellow fluorescent (eYFP) protein at its C-terminus (GPR37L1- eYFP), to investigate the receptor’s expression. In addition to full length GPR37L1-eYFP, several N- terminal truncation mutants were also created. These included the receptor lacking (i.) its putative signal peptide (Leng et al., 1999) -GPR37L1-eYFP), (ii.) the signal peptide and the region

(Δ25 -GPR37L1-eYFP) and (iii.) the whole of the N- containing the CSF-GPR37L1 GPR37L1-eYFP) peptide (Figure fragments 3.1A). (Δ80 terminus (Δ122 In-gel eYFP fluorescence imaging of doxycycline-induced HEK293 lysates revealed GPR37L1- eYFP resolves at two molecular weight species: a larger band of Mr ~65-70 kD (Figure 3.1B, open circle) and a smaller band of Mr~40-45 kD (Figure 3.1B, closed circle). The smaller species was

-GPR37L1- eYFanalogousP construct in size was to induced the only (Figure species 3 .that1B). wasCell surfacedetected biotinylation when expression revealed of boththe Δ122 large and small

GPR37L1-eYFP species expression at the surface, and of note, the apparent large:small band ratio was consistent with that seen in whole-cell lysate fluorescent imaging (compare Figure 3.1B & C).

-GPR37L1-eYFP was also detected at the cell surface, suggesting the smaller band seen for

GPR37L1Δ122 -eYFP may be an N-terminally processed product of the full-length receptor. Concordantly, whilst treatment with deglycosylating enzyme PNGaseF reduced the size of the full-length

GPR37L1-eYFP [presumably via removing predicted glycosylation at residue Asn105 (Leng et al.,

1999)], the deglycosylated product was still larger than the 40 kD smaller band (Figure 3.1D), indicating the smaller GPR37L1-eYFP species was unlikely to represent an immature precursor of

81 the full-length receptor. Because deglycosylation did not change the apparent Mr of the smaller

GPR37L1- -GPR37L1-eYFP species, the full-length GPR37L1-eYFP was presumedeYFP tospecies, undergo or ofN- terminalthe only Δ122processing C-terminal to the Asn105 residue [NB: C-terminal processing was excluded as a potential explanation for the presence of the smaller GPR37L1-eYFP species as the in-gel fluorescence imaging required the C-terminal eYFP moiety be intact].

Treatment with EndoH did not change the apparent molecular weight of either the full-length or truncated GPR37L1-eYFP, demonstrating that the receptor does not have high mannose N-glycans

(Figure 3.1D).

-GPR37L1-eYFP was detected both in whole-cell lysates and at the cell surface

[presumablyWhilst Δ122 made possible by a cryptic signal sequence within its first transmembrane domain

(Rutz et al., 2015) -GPR37L1- -GPR37L1-eYFP could not be detected in either whole cell lysates ],or Δ25 at the cell surfaceeYFP by and biotinylation. Δ80 Further experiments (performed by Dr

Nicola J. Smith) demonstrated that these receptor species could not be detected by confocal microscopy, and that inhibition of proteosomal degradation with inhibitor MG132 (Z-Leu-Leu-Leu-

-GPR37L1- -GPR37L1-eYFP, suggesting

H)these did constructs not result are in expression improperly of synthesised either Δ25 or are misfoldedeYFP and [as Δ80 published in Coleman et al.

(2016)].

Figure 3.1: GPR37L1 is expressed in HEK293 cells as both a full-length and N-terminally truncated isoform

A. Representative schematic of the GPR37L1 N-terminus and derived cerebrospinal-fluid (CSF) peptides within. Asn105 is a predicted site of N-linked glycosylation (WT, wild-type; TM1, beginning of transmembrane domain 1). B. Immunoblot of cell lysates from each stable Flp-In HEK293 cell line with an antibody targeted to green fluorescent protein (GFP). Expression of receptor constructs was induced with 500 ng/ml doxycycline for 24 hours antibody detecting GFP and eYFP). C. Immunoprecipitationbefore cellof biotinylated harvest (WB, lysates Western shown blot; inαGFP, (B)

(open circle, full-length receptor; closed circle, truncated receptor; IP, immunoprecipitation). D.

Lysates of doxycycline-induced (24 hours, 500 ng/ml) GPR37L1-eYFP-expressing stable HEK293 cell lines were treated for 2 hours with endoglycosidase H (EndoH; 100 mU) or peptide N- glycosidase F (PNGaseF; 2 U), or both, before immunoblotting with antibody targeted to GFP (Star symbol; deglycosylated receptor). E. Cell lysates from doxycycline-induced (24 hours, 500 ng/ml)

GPR37L1-eYFP-expressing HEK293 cells were visualised by in-gel fluorescence (488 nm) for evidence of C-terminal proteolysis. All images are representative of three independent experiments.

3.2.2. The N-terminus of GPR37L1 is subject to metalloprotease-mediated cleavage I hypothesised that activity of MMPs (matrix metalloproteases) or ADAMs (a disintegrin and metalloproteases) might be responsible for creation of the N-terminally processed GPR37L1 species, as multiple GPCRs are reported to undergo proteolysis in this manner (Coleman et al.,

2017). When GPR37L1-eYFP expression was induced in the presence of broad-spectrum

MMP/ADAM inhibitor BB94 (batimastat), the appearance of the small GPR37L1-eYFP species was almost completely prevented (Figure 3.2A). The effect of BB94 was found to be concentration- dependent, concomitantly increasing the full-length GPR37L1-eYFP whilst decreasing the smaller species (Figure 3.2B -GPR37L1-eYFP, consistent with this receptor species). BB94 lacking was itsnot N found-terminus. to alter abundance of Δ122

Next, I tested the Tissue Inhibitors of Metalloproteases (TIMPs) for their ability to inhibit the processing of full-length GPR37L1-eYFP into the cleaved form, as the greater selectivity of these inhibitors (Brew et al., 2010) may provide insight into the identity of the proteases responsible for the N-terminal processing of the receptor. TIMP-1 inhibits ADAM10 and shows weak inhibition towards MMP-14, MMP-16, MMP-19 and MMP-24. TIMP-2 inhibits all MMPs and ADAM 12. TIMP-3 also inhibits all MMPs and ADAM10, ADAM12, ADAM17, ADAM28, ADAM33. TIMP-4 has a wide

MMP inhibition profile and also inhibits ADAM17, ADAM28 and ADAM 33. At no concentration tested did any of the four TIMPs alter the relative abundances of the full-length and cleaved

GPR37L1-eYFP species, despite all TIMPs of the same batch proving capable of inhibiting recombinant MMP-2 (Figure 3.2C-F). These results suggest that GPR37L1 N-terminal proteolysis is not mediated by a MMP, but by an ADAM; and that ADAM10, ADAM 12, ADAM17, ADAM28 and

ADAM33 are unlikely to be responsible for GPR37L1 N-terminal proteolysis.

Whilst less selective than the TIMPs [yet more selective than BB94, (Arribas et al., 1996)], the tumour necrosis factor- -1) and 2 (TAPI-2) prevented cleavage of

GPR37L1-eYFP, in a concentrationα protease inhibitors-dependent 1 (TAPImanner (Figure 3.2G & H). Reviewing and comparing the specific inhibitory profiles of the TIMPs and TAPI-1 and -2, revealed that likely mediators of GPR37L1 N-terminal proteolysis may include ADAM8, ADAM9, ADAM15, ADAM19,

ADAM20, ADAM21 and ADAM30.

Figure 3.2: GPR37L1 is subject to N-terminal proteolysis by an ADAM

A. Flp-In HEK293 cells were induced by doxycycline (500 ng/ml) to express either wild-type or

-GPR37L1-eYFP, in the presence of BB94 (20 µM, 24 h). Cells were lysed and visualised by in-

Δ122 87 gel fluorescence. B. Wild-type GPR37L1-eYFP expression was induced in Flp-In HEK293 cells by doxycycline and cultured in increasing concentrations of BB94 as indicated (24 h). Cells were lysed and visualised by in-gel fluorescence. Quantification in right panel. C-F. Wild-type GPR37L1-eYFP expression was induced in Flp-In HEK293 cells by doxycycline and cultured in increasing concentrations of recombinant human TIMP-1, TIMP-2, TIMP-3 and TIMP-4 as indicated (24 h).

Cells were lysed and immunoblotted with anti-GFP antibody. Biological activity of recombinant

TIMPs was validated using activated MMP-2 and a fluorogenic substrate (shown on right of each immunoblot). AFU, arbitrary fluorescence units; MMP-2 act., MMP-2 activity. G & H. Wild-type

GPR37L1-eYFP expression was induced in Flp-In HEK293 cells by doxycycline and cultured in increasing concentrations of TAPI-1 and TAPI-2 as indicated (24 h). Cells were lysed and visualised by in-gel fluorescence. Quantification in right panels. Open circle represents full-length GPR37L1- eYFP (~65-70 kD), closed circle represents truncated species (~40-50 kD). Each image represents at least four individual experiments.

3.2.3. Mutation of the GPR37L1 N-terminus does not abrogate proteolysis Given that the cleaved GPR37L1-eYFP species was found to not be glycosylated and was

-GPR37L1-eYFP species, the experiments described above defined the analogoussite of GPR37L1 in size N to-terminal the Δ122 proteolysis to be C-terminal to residue 105 and N-terminal to residue

122. In an attempt to further define the cleavage site, site-directed mutagenesis was performed within this region and further into the start of transmembrane domain 1. Residues within this region were mutated to alanine, in groups of three, with the aim of interrupting any metalloprotease cleavage consensus sequences that may be harboured within (mutants illustrated in Figure 3.3A). These constructs (all C-terminally tagged with eYFP) were stably expressed in Flp-

In HEK293 cells, with expression induced by doxycycline as per the previous experiments. Cells were lysed, and their expression was assessed by immunoblotting with an antibody directed against the C-terminal eYFP epitope. With the exception of 109Ala111, which could not be detected, all mutants were expressed and all existed as two distinct species, as per the wild-type GPR37L1-eYFP construct (Figure 3.3B). Interestingly, the full-length 106Ala108 GPR37L1-eYFP resolved at a smaller molecular weight to that of the full-length wild-type GPR37L1 (Figure 3.3B), and given that residues 105, 106 and 107 form part of the receptor’s NxS/T glycosylation motif (shown to be glycosylated on the wild-type receptor in Figure 3.1D), our laboratory hypothesised this species to have a glycosylation defect arising from the mutation. Indeed, deglycosylation of lysates from doxycycline-induced cells revealed that the full-length GPR37L1 species in all alanine mutants, shifted to a lower molecular weight following removal of N-linked glycosylation, with the exception of 106Ala108, consistent with this receptor lacking its NxS/T glycosylation motif (Figure 3.3C). To confirm that the presence of two distinct protein species, as seen for all GPR37L1 alanine mutants, was due to metalloprotease-dependent proteolysis (as per the wild-type receptor, as shown in

Figure 3.2), expression of alanine mutants was induced in the presence of metalloprotease- inhibitor BB94. In all cases, (with the exception of the non-expressing 109Ala111 mutant), BB94

89 inhibited receptor cleavage, as evident by the increased amount of full-length receptor, confirming that all mutants are still subject to metalloprotease-mediated N-terminal cleavage (Figure 3.3D).

The inability of mutagenesis to inhibit GPR37L1 N-terminal cleavage was surprising, but consistent with the notion that metalloproteases lack clearly defined consensus sequences (Tocchi et al.,

2013). Of interest, Mattila et al. (2016) recently reported the related orphan GPCR GPR37 to also be subject to metalloprotease-mediated N-terminal proteolysis and similarly observed that mutation of the suspected cleavage site was not sufficient to prevent cleavage.

Figure 3.3: N-terminal proteolysis of GPR37L1 is not interrupted by alanine mutagenesis

A. Schematic illustrating the GPR37L1-eYFP alanine mutants (106Ala108, 109Ala111, 112Ala114,

115Ala117, 118Ala120, 121Ala123, 124Ala126; all tagged with e-YFP) and the locations of their respective mutations. B. GPR37L1-eYFP (and alanine mutants) expression was induced in stably transfected

HEK293 cells by addition of doxycline. 24 hours later, cells were lysed and immunoblotted with anti-GFP antibody. C. Performed as per panel B, but cell lysates were treated with PNGaseF to remove N-linked glycosylation prior to immunoblot. D. Performed as per panel B, but BB94 (10

µM) was added at time of induction of receptor expression. Open circle represents full length

GPR37L1-eYFP, closed circle represents cleaved GPR37L1-eYFP, star symbol represents GPR37L1- eYFP without N-terminal glycosylation. Images representative of three independent experiments.

3.2.4. GPR37L1 constitutively couples to chimeric Gαs in Saccharomyces cerevisiae To investigate if loss of the GPR37L1 N-terminus has functional consequences for the receptor’s signalling ability, it was necessary to first define the signal transduction pathways activated by

GPR37L1. GPR37L1–eYFP was transformed into multiple reporter yeast strains that feature a modified pheromone-response pathway where, crucially, each strain expresses one of a panel of

(Dowell et al., 2009;

Ngochimeric et al. Gpa1/Gα, 2015) proteins that are capable of activation by mammalian GPCRs dissociation, which. Intransduces this assay, a signal activation through of athe mitogen G protein-activated by receptor protein kinasecoupling cascade, leads causingto Gβγ

-galactosidase reporter gene regulated by the FUS1 promoter. transcription of a β GPR37L1- s 16 G proteins (Figure 3.4).

eYFP exhibited robust coupling to chimeric Gα andq/11 Gα 12 13 14 proteins

Coupling(Figure 3 was.4). Contrary not detected to a previous between report GPR37L1 (Meyer and et chimeric al., 2013 Gα) , Gα , Gα i/o norwas Gαnot observed

(Figure 3.4 16 G protein is known to couple promiscuously, activity to GPCRs via Gα that ordinarily signal

s). Thei/o Gα q/11 12/13 (Kostenis et al., 2005;

Giannonethrough Gα et, Gαal., or2010 Gα), suggestingG proteins, thebut notcoupling with GPCRsprofile that observed couple Gα for GPR37L1 in yeast is

s- -GPR37L1-eYFP demonstrated representativeno statistically ofsignificant it being aability Gα coupled to couple receptor. with any Intriguingly, of the G proteinΔ122 chimeras tested, suggesting loss of the N-terminus may result in a functional silencing of the receptor.

This finding was then validated in a mammalian cell system, using reporter gene assays

(performed by Dr Tony Ngo), in which transfection with the full-length and untagged GPR37L1 caused constitutive cAMP generation, as measured via accumulation of luciferase, transcription of which was driven by a cAMP-response element promoter [as published in Coleman et al. (2016)].

-GPR37L1 was unable to constitutively generate cAMP and,

Using this assay, we found that Δ122

92 furthermore, treatment of the wild-type GPR37L1 with metalloprotease inhibition caused increases in cAMP accumulation (Coleman et al., 2016). The present data support GPR37L1 being a

s-coupling receptor, capable of driving cAMP production until its N-terminus is cleavedconstitutively by a metalloprotease Gα (likely an ADAM), causing silencing of receptor signalling.

Figure 3.4: The GPR37L1 N-terminus is necessary for constitutive coupling to chimeric

Gαs G proteins

Wild-type GPR37L1-eYFP (WT) -GPR37L1-eYFP , or empty vector (p426GPD),

and Δ122 (Δ122) - galactosidasewas transformed activity into was yeast measured strains in arbitrary expressing fluorescence individual units G (AFU) protein 24 h chimeras,ours later. before(One-way β

ANOVA with Bonferroni’s post hoc analysis, n = 4-6 individual transformants per strain, *** =

p≤0.001).

3.2.5. Prosaptide does not activate GPR37L1 A previous study suggested that TX14A (also known as prosaptide), a synthetic 14-amino acid peptide derivative of neuroprotective peptide prosaposin, is an agonist for both GPR37 and

i signal transduction assays (Meyer et al.,

2013GPR37L1,) as assessed in receptor internalisation and Gα i in yeast (. FigureDespite 3 having.4) or i nbeen a mammalian unable to observereporter that gene GPR37L1 assay [performed is capable by of Drcoupling Tony Ngo chimeric as reported Gα in Coleman et al. (2016)], I screened TX14A against chimeric yeast G proteins for evidence of stimulation. Consistent with results in Figure 3.4, GPR37L1- -

s-expressing yeast, but not in theeYFP Gpa1 stimulated control strain, expression nor in of the β galactosidasei1 i3 strains in the (chimericFigure 3 .Gα5). Application of 10 µM TX14A did not cause any difference in the G

Gαproteinand-coupling Gα activity of GPR37L1-eYFP (Figure 3.5). Further experiments by Dr Tony Ngo

i family G proteins in a mammalianfailed to show reporter any significant gene assay coupling and further, between TX14A GPR37L1 did not(untagged) alter ands Gα i activity in the same assay (Coleman et al., 2016). In additional efforts to reproduce the Gαfindingsor Gα of Meyer et al. (2013),

Dr Nicola Smith incubated membranes from HEK293 cells expressing GPR37L1, with biotinylated

TX14A, and used streptavidin-coated beads in an attempt to immunoprecipate TX14A bound to

GPR37L1. However, no GPR37L1 could be detected by immunoblotting of samples following immunoprecipitation (Coleman et al., 2016).

Figure 3.5: Prosaptide (TX14A) does not stimulate GPR37L1 signalling

- s i1 i3), transformed to expressβ galactosidase either p426GPD activity vector in chimeric or wild- type yeast GPR37L1 strains - (Gpa1 or Gα , Gα , or Gα

24 hours. (One-way ANOVA with Bonferroni’s post hoceYFP analysis and stimulated, n = 12 indi withvidual 10 transformantsμM TX14A for

per strain, *p≤0.05).

3.2.6. Cerebrospinal fluid peptides do not activate GPR37L1 As TX14A was unable to alter GPR37L1 signalling in yeast (Figure 3.5) or mammalian cells

(Coleman et al., 2016), I screened regions of the GPR37L1 N-terminus against the GPR37L1-

s-chimera yeast strain, hypothesising that they may function as putative ‘tethered ligands’.expressing Peptides Gα screened included the three GPR37L1 N-terminal peptides found in human CSF

(Stark et al., 2001; Schutzer et al., 2010; Zhao et al., 2010), designated CSF 1-3 from most N- terminal to most C-terminal, as well as two peptides designated ‘tethered 1’ and ‘tethered 2’.

Tethered 1 is a region immediately after CSF 1, and tethered 2 is a region immediately after CSF 3

(there is only one residue separating CSF 2 and CSF3) (Figure 3.6A). At no concentration tested,

s activity of vector-, GPR37L1- -GPR37L1- transformeddid any of the yeast five (Figure peptides 3.6 modifyB-F). This the result Gα was mirrored in the CRE-luciferase, or Δ122 reporter assay in mammalian cells performed by Dr Tony Ngo (Coleman et al., 2016).

Figure 3.6: N-terminal peptides to not modify GPR37L1 Gαs signalling in yeast

A. Schematic showing sequences of N-terminal GPR37L1 peptides hypothesised to be potential agonists for GPR37L1. B-F. s were transformed with empty vector

(p426GPD), wild-type (WT) YeastGPR37L1 expressing- chimeric-GP GαR37L1-eYFP and treated with tethered 1

(B), tethered 2 (C), CSF 1 (D), CSF 2 (EeYFP) or orCSF Δ122 3 (F) for 24 hours (concentrations as indicated)

(Data expressed as fold of peptide- -galactosidase activity over signal from media alone, one-way ANOVA with Bonferonni’sstimulated post hoc analysis, β n=9 individual transformants per strain).

3.2.7. GPR37L1 predominates in vivo as a proteolytically processed species To investigate whether the two distinct species of GPR37L1-eYFP observed in vitro

(heterologous expression in Flp-In HEK293 cells) was representative of an in vivo phenomenon, endogenous GPR37L1 was immunoblotted from crude membrane preparations of cerebella from

12 week old, male C57BL/6J and GPR37L1null/null mice [the latter mice were generated by Min et al.

(2010)]. Endogenously expressed GPR37L1 was detected with high specificity (as inferred by lack of immunoreactivity in cerebellar membranes from GPR37L1null/null mice) and found to exist as two molecular weight species (Figure 3.7), as per the receptor’s in vitro expression profile (Figure 3.2).

Intriguingly, the GPR37L1 species of lower molecular weight, equivalent to the N-terminally truncated and functionally-silent species in vitro, was the predominant form, with more film exposure time required to detect the full-length receptor, indicating its lower relative abundance.

The antibody used herein for immunoblotting endogenous GPR37L1 targets a C-terminal epitope, consistent with the smaller receptor being the product of N-terminal proteolysis.

Figure 3.7: GPR37L1 exists as two species in the mouse cerebellum

Immunoblotting was performed using an anti-GPR37L1 antibody on crude plasma membranes prepared from cerebella of GPR37L1null/null and C57BL/6J mice (10-12 weeks age, male). Images show n = 3 at two exposure times. Open circle, full-length receptor; closed circle, truncated receptor.

100

3.2.8. GPR37L1 is not expressed in primary cortical astrocyte cultures Despite evidence that GPR37L1 is subject to N-terminal proteolysis in vivo, it was necessary to confirm that a metalloprotease was responsible for this phenomenon, as per in vitro findings. I initially prepared primary astrocyte cultures, because Meyer et al. (2013) used this primary system to investigate the potential of prosaptide to convey protection from oxidative stress to astrocytes, via GPR37L1 and GPR37. Despite preparing viable astrocyte cultures, as confirmed by the enriched presence of glial fibrillary acid protein-expressing cells (Figure 3.8A), GPR37L1 was not expressed in these cells, as determined by the immunoblotting of cell lysates from multiple independent culture preparations (Figure 3.8B). This was perhaps not unexpected, as primary astrocyte cultures are prepared from mice of two to four days of age, whilst Marazziti et al. (2013) demonstrated that cerebellar GPR37L1 expression increases gradually between five and 15 days of age, with expression on day five being very low.

101

Figure 3.8: Primary cortical astrocyte cultures do not express GPR37L1

A. Primary cortical astrocytes were prepared as previously described (Schildge et al., 2013), plated onto chambered slides, fixed and processed for immunohistochemistry. Cells were treated with rabbit IgG isotype control, or rabbit antibody directed against glial fibrillary acidic protein

(GFAP, green), before labelling with secondary antibody. Cell nuclei marked with DAPI (blue).

Images representative of n=1. Scale bar represents 50 µm. B. Immunoblot of cerebellar homogenates from C57BL/6J (WT) and GPR37L1null/null mice (null) (10-12 weeks of age, mixed-sex), and lysates of four independent primary astrocyte preparations. Image representative of n=4. Equal protein was loaded as determined by Merck Direct Detect.

102

3.2.9. GPR37L1 is processed by a metalloprotease in organotypic cerebellar cultures Owing to the lack of GPR37L1 expression in primary astrocyte cultures, I prepared organotypic cerebellar slice cultures from cerebella of 15 day old, mixed-sex C57BL/6J and GPR37L1null/null (Min et al., 2010) mice, and cultured them for 24 hours in 0, 20 or 40 µM BB94, before harvest for membrane preparation and immunoblotting. In contrast to the primary cortical astrocyte cultures,

GPR37L1 expression was readily observed in these organotypic cultures. As per the investigation of

GPR37L1 in whole mouse cerebellum (Figure 3.7), the cleaved species of the receptor also predominated in the ex vivo cerebellar cultures, and after 24 hours in culture without BB94, full- length GPR37L1 was almost undetectable (Figure 3.9A). However, with BB94 treatment, expression of the full-length GPR37L1 was restored, thus demonstrating the involvement of metalloproteases in the cerebellar processing of GPR37L1 (Figure 3.9A).

103

3.2.10. GPR37L1 increases cAMP in organotypic mouse cerebellar cultures I further used the organotypic cerebellar slice culture model to investigate the signalling of

s- coupling,GPR37L1, cAMP which-generating yeast and receptor. mammalian To investigate reporter assay this datain the showed ex vivo to cerebellum, be a constitutive organotypic Gα cerebellar slice cultures were prepared from cerebella of 15 day old, mixed-sex C57BL/6J and

GPR37L1null/null (Min et al., 2010) mice, and cultured for 1 hour in the presence of phosphodiesterase inhibitor IBMX to prevent cAMP degradation. Over the course of 1 hour, significantly more cAMP accumulated in the C57BL/6J cerebellar cultures than GPR37L1null/null

s-coupled, cAMP-generating receptor

(cultures,Figure 3 consistent.9B). I postulate with GPR37L1 that despite being the an predominance endogenously of Gα the truncated and presumably inactive

GPR37L1 species in the cerebellum, the remaining full-length receptor is responsible for a potent effect on cAMP generation.

104

Figure 3.9: GPR37L1 stimulates cAMP synthesis and is cleaved by a metalloprotease in the ex vivo cerebellum

A. Mouse organotypic cerebellar slice cultures were prepared from GPR37L1null/null (null) or

C57BL/6J (WT) mice (15 days of age, mixed-sex) and cultured for 24 hours in the presence or absence of BB94 (concentrations as indicated) before preparation of crude cell membranes and immunoblot for endogenous GPR37L1. Equal protein was loaded accross lanes as determined by

Merck Direct Detect. (Open circle, full-length receptor; closed circle, truncated receptor). Image shows two biological replicates of n=4. B. Mouse organotypic cerebellar slice cultures were prepared from GPR37L1null/null (null) or C57BL/6J mice (WT) mice and cultured for 1 hour in the presence or absence of phosphodiesterase inhibitor IBMX (1 mM). ELISA was used to determine cAMP concentration (n=6, where each n consists of tissue pooled from 2-4 mice. One-way ANOVA with Dunn

ett’s multiple comparison’s test, * = p≤0.05).

105

3.3. Discussion The findings in -galactosidase

this chapter show that when expresseds-coupled in a heterologous GPCR, and that yeast if it βis expressed in treporterhe same system, system GPR37L1 without itsis a N constitutively-terminus, it active,is inactive. Gα We have also confirmed the constitutive activity of full-length GPR37L1 in a mammalian luciferase reporter assay, and confirmed the inactivity of the truncated, lower molecular weight species in this system (Coleman et al., 2016).

Further, we have found metalloprotease-mediated N-terminal truncation of the receptor occurs both in cultured cells and the mouse cerebellum. Owing to the discovery of three peptides in human

CSF (Stark et al., 2001; Schutzer et al., 2010; Zhao et al., 2010) that are homologous to regions of the

GPR37L1 N-terminus, we hypothesised GPR37L1 activation to resemble that of the PARs, whereby a tethered agonist is revealed on the receptor’s own N-terminus following proteolytic processing of the receptor (Coughlin, 2000; Coleman et al., 2017). Instead, the present study supports the opposite – that GPR37L1 is constitutively active and that N-terminal processing by a metalloprotease silences this activity. This mechanism of signalling has not previously been discovered for GPR37L1 and may prove essential to understanding the receptor’s role in cerebellar development (Marazziti et al., 2013) and cardiovascular disease (Min et al., 2010).

Using yeast containing a chimeric variant of th

eir native Gpa1 Gα protein (in which the last five amino acids of the Gpa1 protein are substituted for the slast-coupled five amino receptor. acids Perhaps of a mammalian this provides Gα protein),an explanati I haveon providedfor the deleteriousevidence that effects GPR37L1 of GPR37L1is a Gα -overexpression in cardiomyocytes as reported by Min et al. (2010) s/adenylyl cyclase signalling by

-adrenoceptors decreases cardiomyocyte, given that excessive viability, stimulation potentially of Gαthrough cAMP-mediated calcium overloadβ (Mann et al., 1992). My studies with GPR37L1 in yeast also show the receptor’s propensity

16 15), s i/o q/11, but to couple12/13 to Gα (now recognised as Gα reflecting that15/16 GPCRs (Kostenis linked et toal. ,Gα 2005, Gα; Giannoneor Gα et al., not Gα , are subject to promiscuous coupling by Gα 106

2010). Thus, the observed GPR37L1- 15/16 s-coupling propensity as, to my knowledge, the Gαliteraturecoupling contains likely no supportsexamples the of receptor’sGPCRs that Gα exclusively

15/16 15/16, the formercouple tobeing Gα abundant. In addition, in the GPR37L1CNS (Leng expression et al., 1999 does) and not the correlate latter, in with stem that cells of and Gα thymic epithelial cells (Giannone et al., 2010), with only low expression in the brain (Wilkie et al., 1991).

s, wi i, contradicts a

The finding that GPR37L1 couples i- tocoupled Gα (Meyerth no observedet al., 2013 coupling). The same to Gα report proposed previousthat prosaposin, report thatand theits receptorneuroprotective is Gα derivative prosaptide (TX14A), are agonists for both

GPR37L1 and its closest-relative, GPR37. Meyer et al. (2013) transiently transfected HEK293T cells

35 with Flag- i/o-mediated [ S]guanosine 5’-O-(3’- thio)triphosphatetagged GPR37L1 accumulation and reportedin response a 5% to increaseTX14A; ain modest Gα signal compared to that seen in our laboratory’s previous stu i/o-coupled GPCRs, which typically result in a 100% increase in activity following ligand stimulationdies of Gα (Stoddart et al., 2008; Schmidt et al., 2011; Smith et al.,

2011). In this chapter, I failed to substantiate i-coupled GPCR, as it

- evidence for GPR37L1 being a Gα i/o- did not cause a significant increase in β galactosidase transcription in yeasti/o-coupled containing GPCRs any (Brown Gα etfamily al., 2003 chimeras,; Jenkins despite et al. , this2010 assay’s) successful application to known Gα i/o, receptors that

15/16 . In furtheri/o-selective support (Kostenis of GPR37L1 et al. ,not 2005 coupling), yet we to observedGα significant fail to couple to Gα tend15/16 toin be yeast. Gα Further, our studies of GPR37L1 in a mammalian reporter coupling to chimeric Gα i/o-reporting SRE-luciferase element,gene assay compared showed to only t statisticallys-reporting insignificant CRE -luciferase activity via activity the Gα (Coleman et al., 2016).

he robust Gα In this chapter, I also demo s i/o family chimeric G proteins, was not modifiednstrated by additionthat GPR37L1 of TX14A activity - the throughneuroprotective either Gα peptideor Gα proposed by Meyer et al. (2013) to be the ligand for GPR37L1. Our group has also been unable to observe the

107

s i/o reporter assays (despite using multipleactivity of batches TX14A of in the the pepti aforementionedde), nor substantiate mammalian a high Gα-affinityand Gα interaction between GPR37L1 and

TX14A as assessed by the lack of GPR37L1 immunoreactivity in cell membranes immunoprecipitated with biotinylated TX14A and streptavidin-coated beads (Coleman et al., 2016).

Adsorption of endogenous prosaposin from cell media during the CRE-luciferase and SRE-luciferase assays reduced reporter gene activity in both vector and receptor-transfected cells, demonstrating a non-specific effect, which crucially, was not rescued by addition of exogenous TX14A (Coleman et al., 2016). A previously reported large- -arrestin recruitment screen also failed to demonstrate the activity of TX14A on eitherscale GPR37L1 β or GPR37 (Southern et al., 2013). Very recently, the group responsible for the initial report of the TX14A:GPR37L1 pairing has replicated

s and acknowledged that evidence for TX14A being ourthe endogenousdata showing ligand that ofGPR37L1 GPR37L1 couples is tenu toous Gα (Giddens et al., 2017).

An additional recent study by another independent laboratory reported prosaptide signalling via GPR37L1 to inhibit astrocyte glutamate transporters and reduce neuronal N-methyl-D- aspartate (NMDA) receptor (NMDAR) activity, as shown in studies of mouse hippocampal slice cultures (Jolly et al., 2017). Although the resting electrical properties of astrocytes and neurons in hippocampal slices were found to not differ between slices prepared from wild-type and knockout mice at baseline, Jolly et al. (2017) observed that repeated applications of NMDA caused increasing

NMDAR responses in GPR37L1 knockout hippocampal slices, but only when the second NMDA application was performed in the presence of a high concentration (10 µM) of prosaptide. However, whilst not discussed by the authors, the electric currents evoked by application of D-aspartate and

NMDA to GPR37L1 knockout hippocampal slice cultures appear be different in magnitude and kinetics from those evoked in the wild-type slices (Jolly et al., 2017). For example, although maximal D-aspartate-evoked inward current in astrocytes was quantified to be the same between wild-type and GPR37L1 knockout astrocytes, the example current trace included in the manuscript

108 showed this maximal to be reached twice as quickly in the knockout cells. Further, although maximal inhibition of this current by TFB-TBOA was not different between the genotypes, the example current trace in the paper shows the GPR37L1 knockout astrocyte to respond more slowly to the inhibition. This makes it difficult to conclude whether any effects of prosaptide stimulation are the result of true agonism at GPR37L1 or secondary to some cellular consequence of loss of the receptor. A further inconsistency was revealed in that prosaptide conveyed protection to hippocampal slices from chemically-induced ischaemic cell death, even in GPR37L1 knockout tissue, demonstrating the protective effects of this peptide cannot be solely, if at all, mediated by

GPR37L1 (Jolly et al., 2017). Whilst the search for endogenous ligands of GPR37L1 must continue, recent work from our laboratory has shown that the predicted binding-pocket of GPR37L1 shows similarity to that of the phylogenetically distant orexin, bombesin and neuropeptide S receptors, and this information will provide much-needed direction in guiding future GPR37L1 deorphanisation attempts (Ngo et al., 2017).

The N-terminal processing of GPR37L1 by a metalloprotease as demonstrated in this chapter is not without precedent, although the consequences for receptor activity are novel. The ETB receptor

(Grantcharova et al., 2002) 1-adrenoceptors (Hakalahti et al., 2010) are two GPCRs for which metalloprotease-mediated, andN-terminal β cleavage influences receptor expression/turnover, but not signalling activity directly [although Grantcharova et al. (2006a) report that cleavage of the endothelin-B receptor by a metalloprotease may perturb biphasic ERK signalling]. This contrasts with the cleavage of the PAR N-terminus, which reveals a ‘tethered ligand’ sequence that is only accessible once the N-terminus proximal to this sequence has been cleaved off (Ossovskaya et al.,

2004; Coleman et al., 2017). Of relevance to GPR37L1, metalloproteases have been shown to cause atypical activation of the PARs. For example, the cleavage of PAR1 is typically mediated by thrombin, but aberrant cleavage by MMP-1 promotes tumorigenesis in breast carcinoma cells

(Boire et al., 2005). Although distinct from the rhodopsin-like GPCR family, the adhesion GPCRs too

109 are subject to proteolysis (in this case autoproteolysis) that sees a tethered agonist sequence

[designated the stachel sequence by Liebscher et al. (2014)] revealed following removal of the N- terminal extracellular domain. Yet this proceeds without involvement of a separate protease, as exemplified by GPR126 and GPR133 (Liebscher et al., 2014). Accordingly, we hypothesised that

GPR37L1 could be activated via a tethered ligand, but found no evidence that the three human CSF

s activity, in either yeast

(Chapterpeptides, 3) nor or two HEK293 regions cell between,s (Coleman served et al. , to 2016 modulate). Rather, receptor it is possible Gα that the human CSF peptides are products of proteolytic degradation of the greater N-terminus, after its metalloprotease-mediated liberation from the receptor. This is consistent with the activity of prolyl carboxypeptidase on the peptide herein designated “CSF 3”, which requires that a peptide one amino acid longer must have existed (Zhao et al., 2010). Whilst lacking efficacy as agonists of

GPR37L1, the physiological role that these peptides play, if any, remains to be elucidated; speculatively they may have biological activity via other peptide receptors, or simply represent products of degradation.

The present study has shown proteolysis of GPR37L1 to contrast that of the PARs – GPR37L1 is a constitutively active receptor until it is inactivated by metalloprotease activity (Figure 3.10).

Whilst not subject to regulation by a metalloprotease, the constitutive activity of the melanocortin-

4 receptor is maintained by auto-agonism by its own N-terminus, facilitating the evolution of agouti-related peptide as an endogenous inverse agonist acting via antagonism of the N-terminus

(Lu et al., 1994; Haskell-Luevano et al., 2001; Ersoy et al., 2012). As an orphan receptor, perhaps

GPR37L1 has an endogenous agonist, endogenous inverse agonist, or both. More speculatively, it is possible that the receptor is entirely regulated by metalloprotease-activity and is without ligand.

Whilst few specific inhibitors of MMPs and ADAMs exist, I was able to identify a number of likely mediators of GPR37L1 N-terminal proteolysis (ADAM8, ADAM9, ADAM15, ADAM19,

110

ADAM20, ADAM21 and ADAM30) by using metalloprotease inhibitors with varied and overlapping, but not identical, inhibitory profiles. Of these, ADAM19 is an attractive candidate, as like GPR37L1, it is reportedly expressed in the heart (Wei et al., 2001; Edwards et al., 2008) and following genetic deletion, is responsible for cardiac defects (Kurohara et al., 2004; Horiuchi et al., 2005), albeit defects that are far more severe than those reported for GPR37L1 ablation (Min et al., 2010). Like

GPR37L1, ADAM19 mRNA abundance in the human brain is highest in the cerebellum (Wei et al.,

2001), where murine studies show that GPR37L1 plays a regulatory role in development of the organ and, subsequently, motor skills (Marazziti et al., 2013). Owing to the lack of highly specific inhibitors of the MMPs and ADAMs, experiments involving genetic inhibition of ADAM19 (for example, through siRNA knockdown) will be required to confirm the involvement of this protease, although it is not unlikely that multiple members of the MMPs and ADAMs could be responsible, owing to redundancy within the family. Conversely, identification of the site of N-terminal proteolysis is complicated by the lack of clear consensus sequences for MMP and ADAM cleavage

(Tocchi et al., 2013). As a result, I was unable to define an area of the N-terminus that was sensitive to mutation in a manner that interrupted proteolysis, consistent with these proteases having their substrate target specificity regulated at a number of levels; often tolerating amino acid substitutions within, or in close proximity to, their substrate’s cleavage site (Ratnikov et al., 2014).

Moreover, specificity of cleavage location is often mediated by the existence of substrate recognition sites outside of the protease’s catalytic domain termed exosites (Hadler-Olsen et al.,

2011). Whilst, I was not able to observe expression of the 109Ala111 mutant at any molecular weight, this is unlikely to be due to interrupted proteolysis, as this phenomenon would presumably enhance the amount of full-length receptor. The reasons for the lack of expression of this mutant remain unclear, but this observation suggests the mutated residues are essential for transcription, or that their mutation causes a translation defect that interrupts protein synthesis before the C- terminal eYFP epitope can be synthesised. Ultimately, the present study has at least localised the

111 site of proteolysis as being after the glycosylated Asn105 residue and before the start of transmembrane domain 1 (residue Pro122).

The present study has revealed the intriguing finding that GPR37L1 N-terminal proteolysis is observable in vivo, and that metalloproteases have been shown to be responsible for this in the ex vivo cerebellum, mirroring the results of the heterologous in vitro experiments. Fascinatingly, the truncated GPR37L1 predominates in the cerebellum, whereas it was approximately equal in abundance to the full length receptor in vitro. Given that the present study has shown that the truncated form of GPR37L1 does not exhibit the constitutive activity of its full-length counterpart, it is possible that in vivo GPR37L1 is largely maintained in an inactive state. Yet, I have demonstrated that the ex vivo GPR37L1null/null cerebellum produces less cAMP than that of wild-type mice, suggesting that the low abundance full-length GPR37L1 is still signal-competent. Presumably this means that GPR37L1 is sufficiently sensitive to respond to slight changes in metalloprotease activity, although the physiological consequences of this interaction remain to be determined. One may speculate that a transgenic mouse expressing only the N-terminally truncated GPR37L1 may mirror the precocious cerebellar development phenotype seen in mice lacking the receptor in its entirety (Marazziti et al., 2013).

s-coupled receptor and thatThe its signallingexperiments activity reported is silenced here show by proteolysisGPR37L1 to of be its a constitutivelyN-terminus. A Gα previous report that the receptor is activated by prosaptide (Meyer et al., 2013) could not be confirmed, and the present study suggests that, in contrast to the PARs, GPR37L1 does not have a tethered-ligand. Further, I have shown that N-terminal proteolysis of GPR37L1 occurs in the mouse cerebellum, a setting where the cleaved receptor predominates, and this raises the possibility that proteolysis is an important mechanism in the regulation of GPR37L1, with intriguing implications for receptor physiology.

112

Figure 3.10: Schematic contrasting protease-activated and inactivated receptors

A. Protease-activated receptor 1 (PAR1) contains a tethered ligand sequence within its N- terminus that is revealed following N-terminal cleavage by thrombin, causing receptor activation.

B. Orphan receptor GPR37L1 has constitutive activity until proteolysis by a metalloprotease removes the entire N-terminus and silences receptor signalling (MMP, matrix metalloprotease;

ADAM, a disintegrin and metalloprotease). Figure adapted from Coleman et al. (2017).

113

Chapter Four: Characterising GPR37L1 tissue expression

114

4.1. Introduction From the first published reports of its discovery, the orphan GPCR GPR37L1 has been known to be most abundantly expressed in the CNS (Valdenaire et al., 1998; Leng et al., 1999). Valdenaire et al. (1998) cloned GPR37L1 from a human caudate nucleus cDNA library, identified it as a putative

GPCR, and subsequently screened various human tissues for the presence of the receptor’s mRNA. A strong hybridisation signal was observed in human brain, and in discrete regions such as the cerebellum, cortex, medulla and spinal cord. Valdenaire et al. (1998) also used a 35S-labelled oligonucleotide to perform in situ hybridisation on the brain, reporting high levels of GPR37L1 transcript in cerebellar Bergmann glia of the Purkinje cell layer, but with no hybridisation to

Purkinje and granule cells. Concurrently, Leng et al. (1999) cloned GPR37L1 from a rat hippocampal cRNA library and performed a similar Northern blot experiment, detecting the receptor in regions including the cerebellum, cortex, hypothalamus and hippocampus.

More recently, reports of GPR37L1 protein expression in the CNS have appeared in the literature. Marazziti et al. (2013) described the role of GPR37L1 in murine cerebellar development and motor learning regulation, and using immunofluorescent staining, observed the receptor colocalised with Bergmann glia cells in the cerebellum. As described in the previous chapter, I have also successfully detected GPR37L1 protein in both the in vivo and ex vivo mouse cerebellum, as assessed by immunoblotting of crude membrane preparations. Similarly, Meyer et al. (2013) detected GPR37L1 in lysates of whole mouse brain using immunoblot analysis.

Whilst GPR37L1’s abundance in the CNS is increasingly well established, its expression and role in peripheral tissues are not. Despite effective detection of GPR37L1 mRNA in human brain,

Valdenaire et al. (1998) reported no hybridisation signal for the receptor in heart, placenta, lung, liver, skeletal muscle, kidney or pancreas tissue. This is at odds with the work of Leng et al. (1999), who detected GPR37L1 mRNA in certain rat peripheral tissues (such as heart and kidney), although

115 even four days of exposure did not produce a hybridisation signal as strong as a 12 hour exposure of brain tissue. Despite the low levels of GPR37L1 mRNA reported in peripheral tissue, determining the expression profile of the GPR37L1 protein is vital to understanding the potential role of

GPR37L1 in the cardiovascular system (Min et al., 2010), and to addressing the hypothesis that this orphan receptor is a previously undefined regulator of cardiovascular homeostasis.

Here I present the expression profile of GPR37L1 in the brain and peripheral cardiovascular organs (heart, kidney), as determined by immunofluorescence microscopy and immunoblotting experiments using - -galactosidase is expressed in place ofboth GPR37L1) β galactosidase and wild-type reporter mouse mousetissues. tissues (in which β

116

4.2. Results

4.2.1. Generation of GPR37L1 knockout and β-galactosidase reporter mice Although interested in the cardiovascular role of GPR37L1 as reported by Min et al. (2010), I first sought to investigate the expression profile of the receptor. Yet, the characterisation of the tissue expression of GPCRs is often hindered by lack of specific antibodies (Beermann et al., 2012;

Hafko et al., 2013; Herrera et al., 2013), and as such, I preferred to instead assess the expression of

-galactosidase reporter gene. To this end, C57BL/6J mice containing a GPR37L1 allele conditionallya β -targeted by a Cre/loxP and Flp1/FRT gene-trapping strategy (designated the tm1a allele, Figure 4.1) were generated from ES cells created and supplied by the European Conditional

Mouse Mutagenesis (EUCOMM) program (Skarnes et al., 2011), as described in Chapter 2 and previously by our laboratory (Coleman et al., 2015). This allele design allows creation of both reporter and knockout mice for a given gene of interest, as illustrated in Figure 4.1. By crossing tm1a mice with ‘Cre deleter’ mice (C57BL/6J mice globally and constitutively expressing Cre recombinase), the tm1a allele was converted into the tm1b -galactosidase was expressed instead of GPR37L1 (Figure 4.1). For use in theallele, next whereby chapter βof my thesis, mice expressing the tm1a allele were also bred with ‘Flp deleter’ mice (C57BL/6J mice globally and constitutively expressing FlpE recombinase) to create a GPR37L1 ‘floxed’ or ‘knockout-ready’ allele

(designated tm1c), which were then bred with ‘Cre deleter’ mice to generate a GPR37L1 knockout allele (designated tm1d), -galactosidase was present and the second of the two exons of

GPR37L1 was deleted (Figurein which 4. 1β). No unexpected phenotypes were observed at any stage of modification of the GPR37L1 allele, and as shown in Table 4.1, the expected Mendelian inheritance was observed, indicating that modifications to GPR37L1 had no effect on animal viability.

Potentially confounding recombinase genes were bred out of each mouse line after desired alleles had been generated, and experimental and control animals were generated by stem and expansion breeding strategies as outlined in Chapter 2 and by Coleman et al. (2015).

117

Figure 4.1: EUCOMM targeting vector design and alleles resulting from crosses with Cre or Flp deleter mice

Schematic of the EUCOMM vector design illustrating the targeting and recombination process for GPR37L1. Mice with the tm1a allele were bred with mice expressing Cre recombinase to

-galactosidase reporter mice (tm1b). Alternatively, ‘floxed’ or ‘knockout-ready’ mice generate(tm1c) were β generated by breeding of tm1a mice with those expressing Flp recombinase. Floxed mice were then bred with mice expressing Cre recombinase to generate GPR37L1 knockout mice

(tm1d). Black boxes represent exons. Figure adapted from Skarnes et al. (2011) and Coleman et al.

(2015).

118

Table 4.1: Breeding statistics for the first generation of each EUCOMM line

Line Strategy # # # % Avg. Pre-wean Genotype % Genotype by (F x M) BPs litters total weaned litter mortality (expected %) sex (% desired per BP pups males size genotype) Desired Other M F tm1a tm1a/wt x 10 2 139 61.2 5.8 5 54.1 45.9 wt 58.8 46.5 wt, or wt x tm1a (50) tm1a/wt (50) tm1b tm1a/wt x 8 1-2 88 0*(52.3) 7.3 1.1 36.1 63.9 n.a 36.6 Cre/Y tm1b tm1a or (25) Cre only (75) tm1c Flp/wt x 10 1-2 120 57.9 6.6 16.8 26.5 73.5 29.5 23.1 tm1a/wt tm1c tm1a or (25) Flp only (75) tm1d Cre/Cre x 7 1-2 80 38.7 7.3 0 34.2 65.8 wt 32.2 35.4 tm1c/wt tm1d (50) (50)

* indicates that only one sex was used (males indicated in parentheses). Avg, average; BP,

breeding pair; F, female; M, male; wt, wild-type. Table adapted from Coleman et al. (2015).

119

4.2.2. GPR37L1 is abundant in the cerebellum as assessed by β- galactosidase reporter gene expression As described above, we generated GPR37L1lacZ/wt - galactosidase reporter protein in place of one allele of GPR37L1),mice (heterozygote in order to characterise mice expressing the tissue a β expression profile of GPR37L1. Immunofluorescence microscopy was used to visualis -

-galactoside thease β wasgalacatosidase expressed reporterabundantly protein in astrocytes in brain, heartand cerebellar and kidney Bergmann tissue. Within glia cells, the brain, as determined β by its expression in GFAP-positive cells (Figure 4.2 -galactosidase within the brain was observed in the cerebellum,). Indeed, where the immunostaining highest abundance localised of β the protein to

GFAP-positive fibres of the Bergmann glia that populate the molecular layer of the cerebellum, and to GFAP-negative cell bodies in the Purkinje cell layer (Figure 4.2). The GPR37L1-expressing, but

GFAP-negative cells, in the Purkinje cell layer are likely also Bergmann glia; weak GFAP- immunoreactivity of the Bergmann glia cell body relative to fibres has been reported previously and thought to be due to the cytoarchitecture of this cell type (Yamada et al., 2000). Moreover,

Valdenaire et al. (1998) also reported GPR37L1 mRNA presence in Bergmann glia cell bodies of the

Purkinje cell layer, and absence in neighbouring Purkinje cells, as assessed by Northern Blot analysis.

120

Figure 4.2: Expression profile of β-galactosidase in EUCOMM-derived reporter mouse cerebellum

Paraformaldehyde-fixed brain tissue from GPR37L1lacZ/wt and GPR37L1wt/wt mice (mixed-

12 weeks of age) was embedded in tissue-freezing medium, frozen and cut into 12 µm sectionssex, for ≥

- - immunofluorescence microscopy. Nuclei (blue; DAPI); GPR37L1 β galactosidase reporter (green; β

121 galactosidase antibody); astrocytes and Bergmann glia (red, glial fibrillary acidic protein antibody,

GFAP). Scale bars: 500 µm (10x images) and 100 µm (63x images). Representative images of n=3.

122

4.2.3. GPR37L1 is widely expressed in the central nervous system as assessed by β-galactosidase reporter gene expression -galactosidase expression in the cerebellum, I next investigated its Havingexpression demonstrated in other strongbrain regionsβ including the hippocampus, isocortex and brainstem.

Throughout these regions, astrocytes (as determined by GFAP immunostaining) were found to

-galactosidase in GPR37L1lacZ/wt mice, most evidently in the astrocyte reproduciblypopulation of expressthe hippocampus β (Figure 4.3 -galactosidase, but not GFAP, were found in the isocortex and brainstem (Figure). Cells 4 that.3); a expressed finding consistent β with recent reports of

GPR37L1 expression in cells of oligodendrocyte lineage (Jolly et al., 2017; Reddy et al., 2017). Of relevance to the hypothesis -galactosidase staining was observed closeed to, role but of not GPR37L1 within, intyrosine cardiovascular hydroxylase control,-positive diffuse catecholaminergic β neurons in several key cardiovascular centres of the CNS including the caudal and rostral ventrolateral medulla (CVLM, RVLM), the nucleus tractus solitarius (NTS) and the A5 nucleus

(Figure 4.4).

123

Figure 4.3: Expression profile of β-galactosidase in EUCOMM-derived reporter mouse brain

124

Paraformaldehyde-fixed brain tissue from GPR37L1lacZ/wt and GPR37L1wt/wt mice (mixed-

12 weeks of age) was embedded in tissue-freezing medium, frozen and cut into 12 µm sectionssex, for ≥

- - galactosidaseimmunofluorescence antibody); microscopy. astrocytes Nuclei and Bergmann (blue; DAPI); glia GPR37L1 (red, glial β fibrillarygalactosidase acidic reporter protein (green;antibody β,

GFAP). Scale bars: 200 µm (20x images) and 100 µm (63x images). Representative images of n=3.

125

Figure 4.4: Expression of β-galactosidase in central nervous system cardiovascular control regions in EUCOMM-derived reporter mouse

Paraformaldehyde-fixed brain tissue from a GPR37L1lacZ/wt n=1) was embedded in tissue-freezing medium, frozen andmouse cut into (male, 20 ≥ 12µm weekssections of age,for immunofluorescence microscopy. Catecholaminergic neurons were stained with an anti-tyrosine hydroxylase antibody (TH; red) and GPR37L1 was stained with an anti- -galactosidase antibody

(green). Scale bar indicates 500 µm. A5, A5 nucleus; CVLM, caudal ventrolβ ateral medulla; RVLM, rostral ventrolateral medulla; NTS, nucleus tractus solitarius.

126

4.2.4. GPR37L1 protein is undetectable in the mouse heart Despite reports of GPR37L1 mRNA being present in the heart (Leng et al., 1999; Min et al.,

2010), there are no reports of GPR37L1 expression at the protein level in this organ. I first sought to

-galactosidase reporter gene in the LV of GPR37L1lacZ/wt mice.

Despiteexamine successful the expression detection of theof the β reporter gene in brain tissue from GPR37L1lacZ/wt mice, I could

-galactosidase signal above that of background fluorescence in the LV (Figure 4.5A). notTo bedetect confident a β that a low level of background tissue auto-fluorescence was not obscuring

-galactosidase reporter protein in this tissue, I performed immunoblot analysesdetection on of lowhomogenized abundance whole β cerebellum and heart, from male and female GPR37L1lacZ/wt mice, using a different anti- -galactosidase antibody to that used in the immunohistochemistry

β-galactosidase was detected with this antibody upon immunoblotting of experiments.the cerebellar Although homogenate, β it was not evident in whole heart (Figure 4.5B). I next attempted to detect endogenous GPR37L1 in the LV of GPR37L1flx/flx and GPR37L1KO/KO mice by immunoblot analysis, using an antibody validated against both the endogenous full-length and truncated

GPR37L1 species [as described in the previous chapter]. No specific bands could be distinguished in the LV homogenates of GPR37L1flx/flx mice, as compared to LV homogenates of GPR37L1KO/KO mice

(Figure 4.5C). In an attempt to reduce non-specific targets of GPR37L1 antibody binding that may be obscuring the detection of endogenous GPR37L1, crude membrane preparations were performed on another set of LV tissue before Western blot. No differences in antibody binding could be observed across genotypes (Figure 4.5D). Finally, Dr Nicola Smith performed an immunoblot on the brain and LV of GPR37L1 cardiac-specific overexpressing (cTg) and GPR37L1 null mice, both obtained from Min et al. (2010), revealing again the endogenous expression of

GPR37L1 protein in the brain, but not the heart (Figure 4.5E).

127

Figure 4.5: GPR37L1 protein was not detected in the heart

128

A. Paraformaldehyde-fixed LV tissue from GPR37L1lacZ/wt and GPR37L1wt/wt mice (mixed-

12 weeks of age) was embedded in tissue-freezing medium, frozen and cut into 12 µm sectionssex, for ≥

-galactosidase was immunofluorescencedetected via antibody microscopy. (green) and Nucleicell membra were detectednes were with detected DAPI with stain WGA (blue), stain β (red). Images are representative of n=3. Scale bar indicates 200 µm. B. Immunoblot of cerebellar and left-ventricular homogenates from GPR37L1lacZ/wt and GPR37L1wt/wt mice (mixed- -

lacZ/wt galactosidase was detected in GPR37L1 cerebellar lysates (asterisk,sex, ≥M r 12~ 100 weeks kD ) of but age). not βin

LV. Image shows n=2 from 2 different animals. C. Immunoblot of LV homogenates prepared from

GPR37L1flx/flx and GPR37L1KO/KO mice (male and female as indicated, 10-12 weeks of age).

Cerebellar lysates shown as antibody specificity controls. Image shows n=2 from 2 different animals. D. Immunoblot of LV cell membranes prepared from GPR37L1flx/flx and GPR37L1KO/KO mice

(male and female as indicated, 10-12 weeks of age). Cerebellar membrane preparations are shown as controls for antibody specificity. Image shows n=2 from 2 different amimals. E. Immunoblot of brain and LV homogenate prepared from GPR37L1KO/KO and GPR37L1-cTg [cardiomyocyte-specific

GPR37L1-overexpressor (Min et al., 2010)] (Mixed sex, 7-8 weeks). Open circle, Mr of full-length

GPR37L1, ~50 kD; closed circle, Mr of cleaved GPR37L1, ~30 kD. Br, brain; KO, GPR37L1 knockout;

WT, wild-type.

129

4.2.5. GPR37L1 protein is undetectable in the mouse kidney Owing to the purported role of GPR37L1 in the cardiovascular system (Min et al., 2010) but apparent absence of the receptor in the heart, I next sought to detect GPR37L1 in the kidney; an organ vital for blood pressure control and where transcript has also been reported (Leng et al.,

1999). Mirroring the results of my heart expression studies (Figure 4.5), I could not detect a

-galactosidase signal in the kidney of GPR37L1lacZ/wt mice, either by specificimmunohistochemistry β (Figure 4.6A) or by immunoblot analysis (Figure 4.6B). I next performed immunoblotting on crude membranes prepared from the kidneys of male and female GPR37L1wt/wt and GPR37L1KO/KO mice, and again, could not detect bands that were specific to the GPR37L1wt/wt samples (Figure 4.6C).

130

Figure 4.6: GPR37L1 protein was not detected in the kidney

A. Paraformaldehyde-fixed kidney tissue from GPR37L1lacZ/wt and GPR37L1wt/wt mice (mixed-

-freezing medium, frozen and cut into 12 µm sex, ≥ 12 weeks of age) was embedded in tissue - galactosidasesections for immunofluorescencewas detected via antibody microscopy. (green) and Nuclei cell were membra detectednes were with detected DAPI stainwith WGA (blue), stain β

131

(red). Images are representative of n=3. Scale bar indicates 200 µm. B. Immunoblot of cerebellar and kidney homogenates from GPR37L1lacZ/wt and GPR37L1wt/wt mice (mixed-

lacZ/wt -galactosidase was detected in GPR37L1 cerebellar lysates (asterisk,sex, Mr ~ ≥ 100 12 weeks kD ) but of age).not in β kidney. Image shows n=2 from 2 different animals. C. Immunoblot of kidney cell membranes prepared from GPR37L1wt/wt and GPR37L1ko/ko mice (male and female as indicated, 10-12 weeks of age). Cerebellar membrane preparations shown as antibody specificity controls. Image shows n=2 from 2 different animals. Immunoblot cropped from same exposure as that used in Figure 4.5B, and thus the cerebellar control tissue is duplicated. Open circle, Mr of full-length GPR37L1, ~50 kD; closed circle, Mr of cleaved GPR37L1, ~30 kD.

132

4.2.6. GPR37L1 transcription is efficiently abrogated in EUCOMM-derived knockout mice To independently investigate the role of GPR37L1 in vivo, our laboratory generated our own

GPR37L1 knockout animals using the EUCOMM system (Skarnes et al., 2011; Coleman et al., 2015) as described above. Before investigating the physiology of these animals I sought to investigate the expression of endogenous GPR37L1 in the resulting GPR37L1wt/wt and GPR37L1flx/flx mice and to confirm the reduction and loss of GPR37L1 in the GPR37L1wt/KO and GPR37L1KO/KO mice, respectively. Quantitative PCR was used to assess the extent of gene deletion in GPR37L1 knockout heterozygote and homozygote animals, and to assess any perturbations of GPR37L1 expression potentially arising in the GPR37L1flx/flx mice. The synthesis of complementary DNA was performed on RNA extracted from the cerebellum of male and female mice (10-12 weeks of age), owing to the abundance of GPR37L1 in the brain and its reported role in cerebellar development (Valdenaire et al., 1998; Leng et al., 1999; Marazziti et al., 2013). Male and female GPR37L1KO/KO mice showed significant loss of cerebellar GPR37L1 transcript, as compared to the transcript levels in sex- matched GPR37L1wt/wt cerebella (Figure 4.7A). Further, neither male nor female GPR37L1flx/flx mice significantly differed in cerebellar GPR37L1 transcript levels, as compared to levels in sex-matched

GPR37L1wt/wt cerebella (Figure 4.7A).

To verify the deletion of GPR37L1 protein, immunoblotting was used to detect the presence of the receptor in crude cerebellar lysates (Figure 4.7B & C). Densitometry was performed on the cleaved GPR37L1 species, as shown in the previous chapter to be the predominant form of the receptor in vivo (Coleman et al., 2016). Male and female GPR37L1wt/KO and GPR37L1KO/KO mice had significantly reduced anti-GPR37L1 immunoreactivity, as compared to that observed in cerebella from sex-matched GPR37L1flx/flx mice (Figure 4.7C). [NB: Neither male (Figure 4.7D) nor female

(Figure 4.7E) GPR37L1flx/flx mice differed from GPR37L1wt/wt mice in abundance of cerebellar

GPR37L1 protein and these mice were not used for further experiments].

133

Figure 4.7: Cerebellar GPR37L1 transcript and protein is efficiently ablated in knockout mice

A. qPCR was performed on cerebella from GPR37L1wt/wt, GPR37L1flx/flx, GPR37L1wt/KO and

GPR37L1KO/KO mice, revealing efficient abrogation of GPR37L1 transcript in male and female

GPR37L1KO/KO mice (10-12 weeks age, n=6, one-way ANOVA with Dunnett’s multiple comparisons

134

B. Immunoblot and (C.) densitometry (normalised to GAPDH abundance) of test,cerebellar *p≤0.05). homogenate revealed marked reduction of GPR37L1 protein in male and female

GPR37L1wt/KO and GPR37L1KO/KO mice, relative to the GPR37L1flx/flx genotype (10-12 weeks age, n=3, one-way ANOVA with Dunnet

GPR37L1 in the GPR37L1flx/flx cerebellart’s multiple homogenate comparisons was unchanged test, *p≤0.05). from ProteinGPR37L1 abundancewt/wt in both of male (D.) and female (E.) samples (10-12 weeks age, n=3, two-tailed student’s T-test). Closed circle, Mr of predominant and cleaved GPR37L1 species, ~30 kD.

135

4.3. Discussion Prior to the present study, reports of GPR37L1 being located outside of the brain existed only at the resolution of mRNA transcript. The initial reports of the cloning of GPR37L1 revealed that the receptor is expressed predominantly in the CNS, with minimal transcription in peripheral tissues

(Valdenaire et al., 1998; Leng et al., 1999). Indeed, my experiments show GPR37L1 to be abundant in the brain, where its highest expression is in cerebellar glial cells, as well as in astrocytes and presumed oligodendrocytes throughout the brain and brainstem. This is consistent with what is known of the physiological role of GPR37L1. Whilst a hypertensive phenotype has been reported following deletion of GPR37L1 in mice (Min et al., 2010), and is to be furthered explored in the next chapter of my thesis, prior studies of the physiology of GPR37L1 largely implicate the receptor in

CNS functions.

The first description of the role of GPR37L1 in the CNS investigated the receptor’s involvement in cerebellar development (Marazziti et al., 2013). Using immunohistochemistry, Marazziti et al.

(2013) observed the receptor in Bergmann glia cells in the molecular and external granular layers of the mouse cerebellum, and more specifically, in association with the basal membrane of the

Bergmann glia primary cilia. Global GPR37L1 knockout mice were found to exhibit premature decreases in proliferation of neurons within the external granular layer and precocious development of the Bergmann glia and Purkinje neurons in the molecular layer. This corresponded to a decrease in thickness of the external granular layer and an increase in thickness of the molecular layer; evident at postnatal day 10 and partially at day 15, but resolved by adulthood

(Marazziti et al., 2013). The authors also implicated GPR37L1 in Sonic hedgehog (Shh) signalling and thus, discovered a potential mechanistic explanation for the altered developmental timecourse of the GPR37L1 knockout mouse cerebellum. Colocalisation was observed between the expression of GPR37L1 and Patched 1 (Ptch1) protein in Bergmann glia, and GPR37L1 knockout mice had increased Ptch1 protein at day 5 and 10, increased Smoothened (Smo) protein at day 5, decreased

136

Smo protein at days 10 and 15, and increased Shh protein levels at day 5. All three proteins are components of the Shh pathway - a known modulator of differentiation and maturation of cerebellar glia (Dahmane et al., 1999). Marazziti et al. (2013) also discovered physiological consequences of the observed alterations in cerebellar development and interestingly, some of the differences between GPR37L1 knockout and wild-type mice remained long after differences in cytoarchitecture have resolved. At three, six and 12 months of age, GPR37L1 knockout mice outperformed their wild-type counterparts on the rotarod test, indicating superior motor skills.

GPR37L1 knockout mice also had improved climbing reflexes, but only at postnatal day 10.

However, if the loss of GPR37L1 is indeed beneficial for motor skills, then the question of why it has been retained by natural selection remains. Two recent studies perhaps shed light on why this may be.

Giddens et al. (2017) identified a new role for GPR37L1 in the CNS. Whole exome sequencing was performed on the family of a consanguineous couple, after five of their children died in adolescence from progressive myoclonus epilepsies – a group of disorders characterised by myoclonic and generalised seizures and progressive neurological degeneration. Affected individuals in the study were found to be homozygous carriers of a GPR37L1 variant with an amino acid substitution in the distal end of transmembrane 6, whereby the 349th residue (Lysine) is mutated to Asparagine (residue 6.25 according to Ballesteros-Weinstein numbering). Whilst the authors’ in vitro experiments demonstrated this mutant to have unperturbed total and surface expression relative to the wild-type receptor, experiments were performed with GPR37L1 constructs carrying an N-terminal FLAG tag, and the receptor was identified using antibodies directed against the FLAG epitope. Our laboratory’s finding that GPR37L1 is subject to N-terminal proteolysis (Chapter 3) precludes quantification of GPR37L1 protein abundance by N-terminal tags, and thus, a difference in GPR37L1 expression may yet be found for the Lys349Asp mutant if it were to be investigated with a C-terminal antibody. Nevertheless, Lys349Asp was not different from the

137 wild-type receptor in terms of constitutive cAMP generation, using the same cAMP luciferase reporter assay applied to GPR37L1 by our laboratory (Coleman et al., 2016). Using GPR37L1 knockout mice, Giddens et al. (2017) also demonstrated GPR37L1 knockout mice to have increased susceptibility to experimentally-induced models of seizures.

In addition to its reported role in suppression of seizures, GPR37L1 also appears to inhibit potentiation of NMDA receptor responses in CA1 pyramidal neurons in mouse hippocampal slice cultures, presumably by some mechanism dependent on neuronal and glial cross talk (Jolly et al.,

2017). Further, Jolly et al. (2017) demonstrated that loss of GPR37L1 resulted in more hippocampal cell death in an ex vivo mouse model of hypoxia. Whilst our understanding of the physiology of

GPR37L1 in the brain is still in its infancy, these three studies reveal potential roles of GPR37L1 in the CNS, which may include such diverse functions as brain and motor skill development, inhibition of seizures, regulation of glial and neuronal conductance, and protection against ischaemic cell death.

As research begins to elucidate the role of GPR37L1 in the CNS, the question of what role it may play in the periphery remains unaddressed. However, my experiments herein suggest GPR37L1 may play little direct role in the periphery, as I could not detect the protein in my peripheral organs of interest (heart, kidney), despite the presence of receptor mRNA transcript in these tissues. This is perhaps unsurprising, as GPR37L1 mRNA in peripheral tissues is so sparing that Valdenaire et al.

(1998) were unable to detect it in any human peripheral tissue tested when they first cloned the receptor. Further, whilst Leng et al. (1999) did detect GPR37L1 in rat peripheral tissues including the heart and kidney, even four days of exposure did not produce a GPR37L1 mRNA hybridisation signal as strong as a 12 hour exposure of brain tissue. Additionally, in studying the role of GPR37L1 in mouse brain development, Marazziti et al. (2013) demonstrated wide GPR37L1 mRNA transcript by in situ hybridisation in the day five mouse brain, but that abundance of the protein was

138 comparatively low at this age compared to other developmental time points. Perhaps this suggests that GPR37L1 is one of the estimated 30% of mammalian protein-coding genes subject to post- transcriptional regulation by micro RNAs (Filipowicz et al., 2008). More likely however, the discrepancy between GPR37L1 mRNA and protein is perhaps explained by the relatively higher sensitivity of assays to detect the former compared to the latter, suggesting that if GPR37L1 protein resides in peripheral tissues it is in such low amounts as to be below the limits of detection, and therefore, of negligible physiological relevance to these tissues. Whilst in the present study,

GPR37L1 expression was not specifically investigated in the vasculature (a key contributor to haemodynamic control), this receptor has never been reported to express in vascular tissue.

-galactosidase staining in any tissue examined. Moreover, a large scaleHerein, transcriptomic I observed no profiling vascular of β male and female human post-mortem tissue, as performed by the

Genome-Tissue Expression consortium (GTEx), shows GPR37L1 mRNA abundance in human blood vessels to be as comparatively low as that of the heart and kidney (Consortium, 2015); a level insufficient to result in detectable protein in these organs. Taken together, the findings of the present study agree with that of the Human Protein Atlas (Uhlen et al., 2015), which reports that

GPR37L1 protein is not detectable outside of the brain.

In summary, the present chapter has shown that GPR37L1 is a CNS receptor without detectable peripheral expression. Whilst most predominant in cerebellar glial cells, GPR37L1 is also found in astrocytes and other non-neuronal cells throughout the brain. Despite using multiple methods of detection, I could not detect GPR37L1 in the heart or kidney, suggesting that, if confirmed as a cardiovascular regulator, GPR37L1 must exert its control over cardiovascular function as a result of its expression in the brain.

139

Chapter Five: The cardiovascular physiology of GPR37L1

140

5.1. Introduction High blood pressure is responsible for >45% of heart disease deaths worldwide and is the single biggest contributing risk factor to global disease burden (WHO, 2013). The human cost of hypertension is in part explained by the multifactorial pathogenesis of the disease, which makes the condition difficult to treat (Coffman, 2011). Approximately a third of patients receiving treatment for hypertension do not have their blood pressure adequately controlled by current pharmacotherapy (Rimoldi et al., 2015). Clearly, a better understanding of the mechanisms that regulate blood pressure is required to more effectively manage hypertension, and more broadly, cardiovascular disease.

One such blood pressure regulator may be the GPCR GPR37L1. This orphan receptor is predominantly expressed in the brain (Zeng et al., 1997; Valdenaire et al., 1998), and contributes to cerebellar development and motor skills (Marazziti et al., 2013). Gpr37l1 also lies within eight quantitative trait loci linked to blood pressure regulation in the rat (Dubay et al., 1993; Samani et al., 1996; Kovacs et al., 1997; Zhang et al., 1997; Stoll et al., 2001; Kato et al., 2003; Bilusic et al.,

2004). Furthermore, a previous study (Min et al., 2010) reported that GPR37L1 knockout mice have

LVH and ~60 mmHg higher systolic blood pressure than that of cardiac-specific, GPR37L1- overexpressing transgenic mice. However, the mouse strain, sex, age, heart rate and depth of anaesthesia used for blood pressure measurement in this study were not indicated, and blood pressure was not compared to appropriate wild-type controls. Despite the lack of GPR37L1 protein in the heart and kidney, as reported in the previous chapter, the magnitude of the hypertensive phenotype reported for deletion of GPR37L1 (Min et al., 2010) is greater than that observed following inactivation of any other GPCR to date, and, thus, warrants independent validation and further investigation. In this chapter, I sought to definitively establish whether GPR37L1 regulates blood pressure using GPR37L1 knockout mice that our laboratory generated and maintained on a pure C57BL/6J background (Skarnes et al., 2011; Coleman et al., 2015). I report that GPR37L1 is

141 indeed involved in blood pressure regulation, and also in protection against cardiovascular challenge, yet in a sexually dimorphic manner.

142

5.2. Results

5.2.1. GPR37L1 regulates baseline blood pressure in female but not male mice Baseline haemodynamic measures were first assessed in anaesthetised female mice (10-12 weeks of age) by catheterisation using a high fidelity pressure-transducing catheter. Depth of anaesthesia was controlled to ensure that a heart rate of approximately 500 beats per minute was maintained, allowing the involvement of GPR37L1 in blood pressure regulation to be assessed independently of heart rate (Figure 5.1, Table 5.1). Thus, heart rate did not significantly differ between genotypes (GPR37L1wt/wt, GPR37L1wt/KO and GPR37L1KO/KO) during acquisition of either aortic or ventricular recordings (Figure 5.1, Table 5.1). Under these conditions, female

GPR37L1wt/KO and GPR37L1KO/KO mice displayed significantly elevated aortic systolic pressure compared to GPR37L1wt/wt controls (GPR37L1wt/KO, +10.3 mmHg; GPR37L1KO/KO, +11.6 mmHg;

Figure 5.1, Table 5.1). Furthermore, GPR37L1KO/KO p≤0.05)increased ( aortic diastolic, mean arterial and pulse pressures females(Figure showed5.1, Table significantly 5.1). No genotype(p≤0.05) differences in LV systolic pressure, end-diastolic pressure, dP/dT max or dP/dT min were observed.

To confirm that this phenotype was preserved in conscious animals, blood pressure was monitored for three hours daily, for nine days, by radiotelemetry on a separate cohort of conscious

GPR37L1wt/wt and GPR37L1KO/KO female mice (14-16 weeks of age). Conscious heart rate was not different between genotypes (Figure 5.1). In agreement with the anaesthetised cohort,

GPR37L1KO/KO mmHg) pressure,females during had both significantly light and (p≤0.05) dark phases elevated of the systolic diurnal (+ cycle,~8 mmHg) relative and to diastolic GPR37L1 (+wt/wt ~6 controls (Figure 5.1).

Baseline haemodynamic measures were also assessed in male mice. GPR37L1wt/wt,

GPR37L1wt/KO and GPR37L1KO/KO males showed no differences in any hemodynamic parameter

143 measured by pressure-transducing catheter under anaesthesia (Figure 5.1, Table 5.1). Conscious blood pressure monitoring by radiotelemetry was also performed on a separate cohort of conscious male GPR37L1wt/wt and GPR37L1KO/KO animals (14-16 weeks of age) and in agreement with the anaesthetised cohort, no blood pressure differences were found between genotypes (Figure 5.1).

Conscious heart rate also did not differ between genotypes (Figure 5.1). Consistent with previous reports (Barsha et al., 2016), GPR37L1wt/wt males had higher systolic and diastolic blood pressures

(Figure 5.1C) and lower heart rates (Figure 5.1D) than their female counterparts. The expected circadian dip was seen in both sexes (Figure 5.1C).

144

Figure 5.1: Female GPR37L1 knockout mice have elevated blood pressure

A. Micromanometry under anaesthesia was performed on male and female GPR37L1wt/wt,

GPR37L1wt/KO and GPR37L1KO/KO mice (10-12 week- -way ANOVA with Dunnett’s post- hoc test, or Kruskal-Wallis with Dunn’s multiple comparisonsold, n≥13, one where ANOVA assumptions were

B. Heart rate during period of anaesthetised blood pressure recordings. C.

Conviolated.scious *p≤0.05). radiotelemetry in male and female GPR37L1wt/wt and GPR37L1KO/KO mice. Data represents averaged recordings during light and dark phases of diurnal cycle over nine days (14-16

145 week- -tailed Student’s t-test between genotypes within each sex). D. Heart rate duringold, n=5, period *p≤0.05, of conscious two blood pressure recordings.

146

Table 5.1: Blood pressure by micromanometry in GPR37L1 knockout mice

Male Female

wt/wt wt/KO KO/KO wt/wt wt/KO KO/KO HR (bpm) 498.7±2.1 499±3.3 497.3±1.4 497.3±3.0 499.2±1.2 491.5±4.1

SP (mmHg) 113.1±2.3 115.5±2.3 109.7±2.1 107.7±2.3 118±3.5* 119.3±2.7* DP (mmHg) 80.4±1.6 79.8±1.8 77.9±1.4 78.0±2.3 85.1±2.4 85.9±2.0* Aorta MAP (mmHg) 91.3±1.8 91.6±1.9 88.5±1.6 88.0±2.2 96.0±2.7 97.0±2.2*

PP (mmHg) 32.7±1.0 35.7±1.2 31.8±1.1 29.7±1.1 32.9±1.2 33.4±1.1*

HR (bpm) 500.4±1.0 493.4±2.3 500.6±1.1 501.7±1.9 496.7±2.2 489.3±4.6 SP (mmHg) 111.3±2.2 113.1±2.0 113.2±3.0 107.7±2.4 111.1±2.8 116.1±2.5

LV EDP (mmHg) 8.4±0.8 8.6±0.8 8.1±0.6 7.46±0.4 8.1±0.6 8.1±0.4 dP/dT max 10957± 10574± 10577± 10406± 10265± 11262± (mmHg/s) 696.9 447.9 429.5 523.7 495.5 813.6 dP/dT min -9879± -10731± -10326± -9088± -9243± -9575± (mmHg/s) 501 522.4 566.5 497.3 473.4 540.7

Underline and * wt/wt within

indicates a significant (p≤0.05)-way ANOVA difference with asDunnett’s compared post to- GPR37L1hoc test or Kruskal-

Walliseach sex with (n≥13 Dunn’s per multiple genotype, comparisons per sex. One test where ANOVA assumptions were violated). HR, heart rate; SP, systolic pressure; DP, diastolic pressure; LV, left ventricle; MAP, mean arterial pressure;

PP, pulse pressure; EDP, end-diastolic pressure.

147

5.2.2. GPR37L1 protects against AngII-induced cardiovascular stress in male mice

I hypothesised that the increased blood pressure in female GPR37L1KO/KO mice may lead to more deleterious outcomes in the setting of hypertensive stress. To test this, female GPR37L1wt/wt and GPR37L1KO/KO mice (10-12 weeks of age) were infused for seven days with AngII (2 mg/kg/day), or vehicle solution as a control, by subcutaneous osmotic mini pumps, to induce hypertension. AngII infusion caused greater increases in systolic pressure in the GPR37L1wt/wt females (GPR37L1wt/wt, +26.2 mmHg; GPR37L1KO/KO, +11.2 mmHg) (Table 5.2), presumably because GPR37L1KO/KO -tailed Mann-

Whitney test) (Table 5.females2), consistent have with higher our baseline previous, systolic independent pressure cohort (p≤0.05, (Figure one 5.1).

However, systolic blood pressure of female mice after seven days of AngII did not significantly differ between genotypes (GPR37L1wt/wt, 135.5 mmHg; GPR37L1KO/KO, 130.9 mmHg, p>0.05) (Table

5.2) and both developed equivalent LVH in response to AngII, as determined by LV to tibia length ratios (GPR37L1wt/wt, 6.0 mg/mm; GPR37L1KO/KO, 6.2 mg/mm, p>0.05) (Table 5.2). The LVH response was confirmed to be due to enlargement of cardiomyocytes, as indicated by a significant decrease in the number of cardiomyocytes per cross sectional area of LV (Figure 5.2).

In contrast to the females, hypertension was not sustained in male mice after seven days of

AngII infusion (Table 5.3), although it was clear from the various physiological endpoints (to be discussed) that the mice had received pressure overload. While AngII treatment induced gross LVH

(assessed by the LV weight to tibia length ratio) for males of both genotypes, the extent of this hypertrophy was significantly greater in GPR37L1KO/KO mice (GPR37L1wt/wt, 6.9 mg/mm;

GPR37L1KO/KO Table 5.3). Cardiomyocyte density reflected this difference, with only GPR37L1, 8.2 mg/mm,KO/KO LV sections p≤0.05) indicating ( significant cardiomyocyte hypertrophy in response to AngII (Figure 5.2). In concert with exacerbated hypertrophy, we observed a significantly blunted

148 contractile response (dP/dT max.) in AngII-treated GPR37L1KO/KO males (9413 mmHg/s) compared to GPR37L1wt/wt counterparts (12844 mmHg/s), suggestive of systolic heart failure (Table 5.3).

Consistent with our previous cohort (Figure 5.1), the baseline aortic systolic pressure of male

GPR37L1KO/KO did not differ from that of GPR37L1wt/wt controls (one-tailed Mann-Whitney test, p>0.05) (Table 5.3).

The induction of transcription of components of the fetal gene program also indicated that genes classically associated with pathological hypertrophy were transcribed in response to AngII in both male and female GPR37L1wt/wt and GPR37L1KO/KO mice (Figure 5.3).

149

Table 5.2: Hemodynamic assessment of AngII-treated (2 mg/kg/day) female

GPR37L1wt/wt and GPR37L1KO/KO mice by micromanometry under anaesthesia

Female

Genotype wt/wt KO/KO

Treatment Vehicle AngII Vehicle AngII

HR (bpm) 515.6± 8.0 512.3± 7.1 -3.3Δ 504.4± 6.8 503.6± 3.6 -0.8Δ

SP (mmHg) 109.3± 3.3 135.5± 4.1* 26.2 119.7± 6.3 130.9± 3.7 11.2

Aorta DP (mmHg) 78.1± 2.3 83.7± 2.5 5.6 83.1± 2.8 86.8± 2.4 3.7

MAP (mmHg) 88.5±2.6 101±2.8* 12.5 95.3±3.9 101.5±2.7 6.2

PP (mmHg) 31.2±1.2 51.8±3.0* 20.6 36.6±3.6 44.0±2.2 7.4

HR (bpm) 500.9±8.2 500.4±2.2 -0.5 514.0±9.6 508.9±4.8 -5.1

SP (mmHg) 106.5±3.1 128.8±1.5* 22.3 112.9±4.7 126.4±2.6* 13.5

LV EDP (mmHg) 6.4±0.5 7.0±1.1 0.6 5.3±0.3 8.7±1.1* 3.4 dP/dT max 9799±860.0 10998±426.2 1199 11343±745 12455±383.9 1112 (mmHg/s) dP/dT min -9019± 853.7 -9253± 584.0 -234 -9262± 470.2 -9718± 330.8 -456 (mmHg/s) Body weight (g) 21.2± 0.3 22.3± 0.3 1.1 23.3± 0.5 22.3± 0.3 -1.0

Tibia length (mm) 16.1± 0.1 16.1± 0.1 0 16.4± 0.1 16.4± 0.1 0 HW:TL 6.5±0.1 8.1±0.2* 1.6 6.8±0.1 8.2±0.2* 1.4 (mg/mm) LVW:TL 4.6±0.1 6.0±0.2* 1.4 4.9±0.1 6.2±0.1* 1.3 (mg/mm) Morphometry RVW:TL 1.4±0.0 1.4±0.0 0 1.4±0.0 1.4±0.0 0 (mg/mm) Lung:TL 9.9±0.4 9.5±0.3 -0.4 9.3±0.5 9.7±0.4 0.4 (mg/mm)

Two-way ANOVA results displayed in Table S1 difference between AngII and vehicle-treated cohorts,. Underline within and each * indicates genotype a significant (Tukey’s (p≤0.05)multiple comparison’s test). Underline and † indicates a - treated animals of different genotypes (Tukey’s multiplesignificant comparison’s (p≤0.05) differencetest). Hemodynamic between AngII data,

meann≥8; morphometry arterial pressure; data, PP, n≥18. pulse HR, pressure; heart rate; EDP, SP, end systolic-diastolic pressure; pressure; DP, HW,diastolic heart pressure; weight; LVW,MAP, left ventricle weight; TL, tibia length; LV, left ventricle; RVW, right ventricle weight.

150

Table 5.3: Hemodynamic assessment of AngII-treated (2 mg/kg/day) male GPR37L1wt/wt and GPR37L1KO/KO mice by micromanometry under anaesthesia

Male

wt/wt KO/KO Genotype

Treatment Vehicle AngII Vehicle AngII

Δ Δ HR (bpm) 487.2± 11 498.5± 13.3 11.3 486.± 7.2 506.3± 19.4 20.3

SP (mmHg) 114± 2.8 127.1± 6.5 13.1 107.9± 4.4 118.5± 6.3 10.6

Aorta DP (mmHg) 78.4± 1.3 82.8± 4.6 4.4 75.4± 3.4 77.3± 6.0 1.9

MAP (mmHg) 90.3±1.6 97.6±5.2 7.3 86.3±3.6 91.0±6.0 4.7

PP (mmHg) 35.6±2.2 44.3±2.7 8.7 32.5±2.3 41.2±2.6 8.7

HR (bpm) 490.5±21.7 513.6±9.4 23.1 493.5±8.1 493.5±12.1 0

SP (mmHg) 113.8±4.0 127.6±6.3 13.8 109.5±2.2 114.3±5.2 4.8

LV EDP (mmHg) 7.6±1.0 7.3±0.5 -0.3 5.8±0.5 9.9±3.0 4.1 dP/dT max 10123±953.3 12844±605.2 2721 8772±438.1 9413±733.1† 641 (mmHg/s) dP/dT min -9753± 1024 -8684± 650.6 1069 -8517± 736.7 -7653± 594.6 864 (mmHg/s) Body weight (g) 28.5± 0.5 25.7± 0.4* -2.8 29.4± 0.5 26.5± 0.3* -2.9

Tibia length (mm) 16.6± 0.1 16.6± 0.1 0 16.7± 0.1 16.7± 0.1 0 HW:TL 8.3±0.2 9.3±0.2* 1.0 8.6±0.2 10.9±0.4*† 2.3 (mg/mm) LVW:TL 5.9±0.2 6.9±0.5* 1.0 6.2±0.1 8.2±0.2*† 2.0 (mg/mm) Morphometry RVW:TL 1.7±0.1 1.6±0.1 -0.1 1.8±0.0 1.7±0.1 -0.1 (mg/mm) Lung:TL 10.5± 0.7 10.3± 0.3 -0.2 10.0± 0.3 11.7± 0.5* 1.7 (mg/mm)

Two-way ANOVA results displayed in Table S1. difference between AngII and vehicle-treated cohorts,Underline within and each * indicates genotype a significant (Tukey’s (p≤0.05)multiple

- treatedcomparison’s animals test). of different Underline genotypes and † indicates (Tukey’s amultiple significant comparison’s (p≤0.05) differencetest). Hemodynamic between AngII data,

meann≥7; morphometry arterial pressure; data, PP, n≥15. pulse HR, pressure; heart rate; EDP, SP, end systolic-diastolic pressure; pressure; DP, HW,diastolic heart pressure; weight; LVW,MAP, left ventricle weight; TL, tibia length; LV, left ventricle; RVW, right ventricle weight.

151

Figure 5.2: AngII infusion decreased cardiomyocyte density

At 10-12 weeks of age, male and female GPR37L1wt/wt and GPR37L1KO/KO mice were infused for

7 days with vehicle or AngII (2 mg/kg/day). LV tissue was cut into 12 µm sections and stained with

WGA to highlight cell membranes. After imaging, cardiomyocytes were counted and expressed per mm2 as shown for (A) males and (B) females. Representative images of (C) male and (D) female cardiomyocyte density displayed in gre Two-way ANOVA results displayed in Table S1. yscale. Scale bar=100 µm. n≥6.

152

Figure 5.3: Activation of the cardiac fetal gene program in vehicle- and AngII-treated mice as determined by quantitative PCR analysis

153

At 10-12 weeks of age, male and female GPR37L1wt/wt and GPR37L1KO/KO mice were infused for

7 days with vehicle or AngII (2 mg/kg/day). LVs were harvested and cDNA was prepared for analysis of (A) ANP, (B) BNP, (C -MHC and (D -SkA transcription. Transcript abundance

) β ) α le-infused GPR37L1wt/wt mice within each normalisedFor to A β2M-D, two and-way expressed ANOVA as results proportion displayed of vehic in Table S2. multiplesex. n≥9. comparison’s test. ANP, atrial natriuretic peptide; BNP, brain*p≤0.05 according to Tukey’s-MHC,

-myosin heavy chain; - - -2 microglobulin.natriuretic peptide; β

β α SkA, α skeletal actin; β2M, β

154

In concert with exacerbated LVH, and decreased contractile function, male AngII-treated

GPR37L1KO/KO mice had increased relative lung weight (normalised to tibia length) (Figure 5.4), indicative of pulmonary edema resulting from congestive heart failure (Adachi et al., 2003; Liao et al., 2005; Kuroda et al., 2010) and suggesting that GPR37L1KO/KO males are susceptible to AngII- induced heart failure. Lung weight did not increase in female mice of either genotype in response to

AngII (Figure 5.4).

155

Figure 5.4: Male GPR37L1KO/KO mice develop pulmonary edema in response to AngII infusion

At 10-12 weeks of age, male and female GPR37L1wt/wt and GPR37L1KO/KO mice were infused for

7 days with vehicle or AngII (2 mg/kg/day), before lung tissue was harvested. Lung weight normalised to each animal’s own tibia length Two-way ANOVA results displayed in Table S1,

. n≥15.arison’s test. and *p≤0.05 according to Tukey’s multiple comp

156

5.2.3. GPR37L1 deletion protects against AngII-induced cardiac fibrosis

In contrast to the features of worsened cardiovascular response observed in the AngII-treated

GPR37L1KO/KO males was the apparent protection from AngII-induced cardiac fibrosis seen in both male and female GPR37L1KO/KO mice. Although GPR37L1wt/wt females developed cardiac fibrosis in response to AngII treatment, this was significantly reduced in GPR37L1KO/KO females (Figure 5.5).

Similarly to the females, GPR37L1KO/KO males did not develop significant fibrosis (Figure 5.5).

However, the more marked hypertrophy and therefore increased tissue area in the GPR37L1KO/KO males may have obscured the extent of absolute fibrosis, which was not grossly different between genotypes (Figure 5.5).

157

Figure 5.5: GPR37L1KO/KO mice are protected from AngII-induced cardiac fibrosis

At 10-12 weeks of age, male and female GPR37L1wt/wt and GPR37L1KO/KO mice were infused for

7 days with vehicle or AngII (2 mg/kg/day). LV tissue was cut into 12 µm sections, and cardiac

158 fibrosis was quantified by Masson’s Trichrome stain, and (A) expressed as percent area of total cross-sectional LV area. Representative images of cardiac fibrosis in (B) male and (C) female tissue.

Two-way ANOVA results displayed in Table S1 comparison’sn≥6. test. Scale bar as indicated. , and *p≤0.05 according to Tukey’s multiple

159

5.2.4. GPR37L1 abundance does not differ between sexes To investigate if the sexually dimorphic cardiovascular effects of GPR37L1 were due to differences in receptor abundance between the sexes, GPR37L1 mRNA transcript and protein in

GPR37L1wt/wt mouse cerebellum, where GPR37L1 is most enriched, were compared between males and females at 10-12 weeks of age. No significant differences in GPR37L1 expression were observed between the sexes (Figure 5.6). Additionally, our analysis of human post-mortem brain

RNA-sequencing [collected by the Genotype-Tissue Expression Consortium (Consortium, 2015)] shows the cortex to be the only sub-region with significant sex differences in GPR37L1 mRNA abundance, wherein GPR37L1 mRNA is higher in females (Figure 5.6).

160

Figure 5.6: Comparison of male and female GPR37L1 expression in mouse and human

161

A. qPCR was performed on cerebellum from male and female GPR37L1wt/wt mice (10-12 weeks age, n=6, two-tailed Student’s t-test). B. GPR37L1 immunoblot and (C) densitometry of male and female GPR37L1wt/wt cerebellar homogenate (10-12 weeks age, n=3, two-tailed Student’s t-test).

Densitometry was performed on the predominating, cleaved GPR37L1 species (indicated by closed circle, Mr~30 kD). D. log2 GPR37L1 mRNA counts per million (CPM) in a heterogeneous cohort of human post-mortem brain tissue (female, red, n=17-27; male, blue, n=31-68), as sequenced by the

Genotype-Tissue Expression Consortium (GTEx)(Consortium, 2015). GTEx raw sequence counts were examined for differentially expressed genes using the edgeR statistical package(Robinson et al., 2010), where samples were grouped by sex for each brain tissue sample. edgeR statistical

brainoutput region. for GPR37L1 Age distribution was analysed. of samples * p≤0.05 shown between in panel female E. and(i. Amygdala, male samples ii. Anterior within thecingulate same cortex, iii. Caudate, iv. Cerebellar hemisphere, v. Cortex, vi. Frontal cortex, vii. Hippocampus, viii.

Hypothalamus, ix. Nucleus accumbens, x. Putamen, xi. Spinal cord, xii. Substantia nigra).

162

5.3. Discussion In this chapter, I have evaluated the cardiovascular role of GPR37L1 using a gene-inactivation mouse model. I report that blood pressure was increased only in female GPR37L1KO/KO mice, compared to their GPR37L1wt/wt counterparts. In contrast, AngII-induced LVH was exacerbated only in male mice lacking GPR37L1, and was associated with depressed contractility (reduced dP/dT max) and heart failure (increased relative lung weight). In light of the receptor’s CNS-restricted expression, as reported in the previous chapter, these findings implicate GPR37L1 as an important central regulator of blood pressure homeostasis, but with differing cardiovascular effects in males and females.

Female GPR37L1KO/KO mice were found to have ~12 mmHg greater systolic blood pressure than

GPR37L1wt/wt counterparts. Such an increase is not trivial, as it would translate to a doubling in CVD risk if it were observed in humans (Jones et al., 2004). Whilst basal blood pressure was unaltered in

GPR37L1KO/KO males, these mice failed to mount compensatory responses to AngII-induced hypertensive stress, and exhibited greater LVH, impaired LV contractile function, and pulmonary congestion. The most parsimonious interpretation of this finding is that GPR37L1KO/KO males had an exaggerated pressor response to AngII. However, I did not observe significant blood pressure increases in males of either genotype in response to AngII, despite each having obvious LVH. Given that males are more susceptible to AngII-induced hypertension than females (Xue et al., 2005), this result was unexpected. However, the dose of AngII used in the present study was higher than that typically employed to model hypertension in mice, and I hypothesise that the severity of the AngII challenge induced pressure overload in males of both genotypes, causing systolic dysfunction, as seen in previous rodent pressure overload studies (Norton et al., 2002). Indeed, only males lost body weight in response to AngII, supportive of AngII being more detrimental to their health

(Table 5.3). This broadly agrees with human heart failure epidemiology - low systolic pressure and

LV function at hospital admission is more common in men (Gheorghiade et al., 2006). Further,

163 systolic function is better preserved in women with aortic stenosis than men, despite equivalent hypertensive stress (Villari et al., 1995). Interestingly, GPR37L1KO/KO mice appeared to be protected from cardiac fibrosis in response to AngII; a phenomenon observed in both sexes, but more pronounced in females despite both genotypes having equivalent LVH and blood pressure following the infusion. In light of the absence of detectable GPR37L1 protein in the heart, it is possible that, unlike the direct, pro-fibrotic effects of cardiac- 2 adrenoceptor, AT1 receptor and ETA (Travers et al., 2016)], GPR37L1expressed may indirectly GPCRs [e.g. influence the β fibrosis by central modulation of circulatory factors that regulate the extracellular matrix, though this remains to be investigated. Alternatively, it may be that the reduced cardiac fibrosis is indirectly linked to effects of the LV GPR37L1 transcript, rather than protein, through regulation of miRNA, which are known to play a role in cardiac fibrosis (Small et al., 2011). Ultimately, detailed studies using radiotelemetry in concert with echocardiography are required to fully evaluate temporal changes in hemodynamic response to AngII, and the mechanism of decompensation in the GPR37L1KO/KO male mice, and whether GPR37L1 conveys cardiovascular protection in other CVD models (i.e. thoracic aortic constriction, myocardial infarction).

The observation of sexual dimorphism in the cardiovascular effects of GPR37L1 is intriguing, although sex differences in the cardiovascular control systems are known to contribute to differential disease aetiology and outcomes (Fermin et al., 2008). While GPR37L1 has previously been linked to hypertension (Min et al., 2010), the sex of the animals studied was not reported.

Moreover, previous studies of GPR37L1 in the brain investigated only male mice (Marazziti et al.,

2013; Meyer et al., 2013), or did not disclose the sex of the animals studied (Jolly et al., 2017).

Notably, sexual dimorphism has been reported for GPR37, the closest phylogenetic neighbour to

GPR37L1 (Smith, 2015), and shown to be more abundant in human females in a variety of brain sub-regions at different developmental stages (Shi et al., 2016). Further, GPR37 deletion in mice leads to sex-dependent effects on neurotransmitter levels and anxiety (Mandillo et al., 2013). In our

164 study, we did not see significant sex differences in cerebellar GPR37L1 mRNA or protein levels in the mouse. Additionally, our analysis of human post-mortem brain RNA-sequencing [collected by the Genotype-Tissue Expression Consortium (Consortium, 2015)] shows the cortex to be the only sub-region with sex differences in GPR37L1 mRNA abundance, which is increased in females. Taken together this suggests that GPR37L1 may exert its sex-specific effects on blood pressure via a mechanism that is independent of its expression level. Such effects might involve hormonal influences; estrogen, for example, is well known to sensitize central baroreflex and modulate sympathetic nerve activity (Saleh et al., 2000).

Sex differences alone cannot reconcile the blood pressure phenotype of GPR37L1 deletion in the present study (+12 mmHg, female-specific), with that reported previously (+60 mmHg, sex not disclosed) (Min et al., 2010). Although both studies determined blood pressure in anaesthetised animals, our additional blood pressure measures by radiotelemetry in conscious, free-moving mice show that the ~12 mmHg increase in GPR37L1KO/KO females is independent of heart rate and anaesthesia. The magnitude of the previously reported GPR37L1 hypertensive phenotype (Min et al., 2010) is also difficult to reconcile with the effects of deletion of other major cardiovascular regulators. For example, deletion of regulator of G protein signaling 2 (Heximer et al., 2003), endothelial nitric oxide synthase (Godecke et al., 1998), adenosine A2A receptor (Ledent et al.,

1997), and bradykinin B2 receptor (Madeddu et al., 1998) reportedly increased blood pressure by

45, 28, 25 and 20 mmHg, respectively. Therefore, it is likely that the effect of GPR37L1 on blood pressure was overestimated, due to the lack of comparisons to wild-type controls. Moreover, we obtained the same reported GPR37L1 knockout mouse line (Min et al., 2010) (reanimated from frozen sperm by Charles River Laboratories, Japan) and failed to detect a change in blood pressure by tail cuff plethysmography in a small cohort, suggesting that any blood pressure difference in these mice is also likely to be modest (data not shown). We additionally found these mice to be of mixed genetic background. Such confounding genetic issues were avoided in the present study by

165 the use of EUCOMM conditional-ready mice (Skarnes et al., 2011; Coleman et al., 2015), enabling comparison of wild-type and null mice on the same C57BL/6J background.

How then might GPR37L1 regulate cardiovascular function? Whilst in this thesis, I have shown

s G protein (Coleman et al., 2016), and despite our laboratory’sGPR37L1 to discovery constitutively that coupleit shares to theantagonists Gα with orexin and bombesin GPCRs (Ngo et al.,

2017), the lack of a high-affinity ligand means pharmacological probing of its physiology is not possible at present. While vascular function remains to be explored in GPR37L1KO/KO mice, we did

-galactosidase immunoreactivity in the heart and kidney of

GPR37L1not observelacZ/wt anymice, vascular and the β mRNA abundance of GPR37L1 in human artery and aorta is comparable to that of the heart and kidney (Consortium, 2015), where we could not detect receptor protein. This raises the fascinating possibility that GPR37L1 is a previously undiscovered component of the CNS-cardiovascular axis. Widely expressed in astrocytes throughout the brain and brainstem, GPR37L1 may exert cardiovascular control through its proximity to autonomic control regions such as the CVLM, RVLM and NTS (Guyenet, 2006), where astrocytes contribute to the brain RAS (Stornetta et al., 1988; O'Callaghan et al., 2011). Alternatively, GPR37L1 could be involved in cerebellar circuits known to influence the cardiovascular system (Bradley et al., 1987a;

Bradley et al., 1987b). Indeed, the cerebellum has a long-suspected role in the modulation of baroreflexes and the development of essential hypertension (Korner, 1971; Korner, 2010). More speculatively, cerebellar control of respiration (Xu et al., 2002) may exert secondary influences on blood pressure, with a recent study implicating excessive respiratory modulation by brainstem neurons in the development of rat and human hypertension (Menuet et al., 2017).

The studies presented in this chapter indicate that the deletion of the orphan GPCR, GPR37L1, causes increased systolic and diastolic blood pressure in female mice. In males, however, deletion of

GPR37L1 blunted a compensatory response to AngII challenge. However, as described in the

166 preceding chapter, I could not detect GPR37L1 protein in peripheral tissue of cardiovascular relevance, including the heart and kidney. Thus, while the mechanism by which GPR37L1 regulates blood pressure in a sexually dimorphic manner has yet to be determined, it is likely that its role in cardiovascular homeostasis is confined to its CNS actions, where the protein is abundant. Thus, the present study suggests that GPR37L1 is a contributor to CNS cardiovascular control, and sex differences therein, and may contribute to the role of the CNS in hypertension and heart failure.

167

Chapter Six: General Discussion

168

At the commencement of my PhD studies, both the pharmacology and physiology of the orphan

GPCR GPR37L1 were underexplored in existing literature. The receptor had been cloned by multiple groups, based on its similarity to the endothelin B receptor (Valdenaire et al., 1998; Leng et al., 1999), and its mRNA had been identified in tissues including the brain, heart and kidney

(Leng et al., 1999). Yet this offered few insights into the receptor’s physiology. Over a decade after

GPR37L1 was first cloned, Min et al. (2010) reported a potential role for the receptor in blood pressure control, (Min et al., 2010) and subsequently, Marazziti et al. (2013) demonstrated a role for GPR37L1 in cerebellar development and motor skills (Marazziti et al., 2013). Owing to its high similarity to the endothelin B receptor, Valdenaire et al. (1998) screened endothelin and related peptides for activity at GPR37L1 and its nearest phylogenetic neighbour, the orphan GPCR GPR37, but neither receptor was activated by any peptide tested. Meyer et al. (2013) subsequently reported that both GPR37L1 and GPR37 could be activated by the neuroprotective peptides prosaposin, and its active portion, prosaptide; a finding that remained to be validated when I commenced my thesis. Of particular interest to myself and my laboratory was the apparent role of

GPR37L1 in blood pressure control, as reported by Min et al. (2010). That mice lacking GPR37L1 appeared severely hypertensive (Min et al., 2010) suggested to our laboratory that pharmacological activation of GPR37L1 could be a novel strategy for the treatment of hypertension. However, certain methodological shortcomings of the study made GPR37L1 an attractive candidate for a more thorough and careful investigation – an undertaking that I embarked upon as part of this PhD thesis. Moreover, I investigated the expression of the receptor protein as a means to help define its physiological role. Ultimately, experiments I performed, as described in this thesis, have contributed to the identification of s as the cognate G protein for GPR37L1 and shown a novel mechanism for modulation of signallingGα activity by metalloprotease-mediated N-terminal proteolysis. In contrast to previous reports, I could not substantiate any interaction of the receptor with t i protein, nor could I demonstrate that prosaptide affected receptor activity. In addition

he Gα 169 to these pharmacological findings, I have provided more detail about both the expression of

GPR37L1, which I report to be CNS-specific, and about the physiology of GPR37L1 and its role in cardiovascular homeostasis. The findings of my studies have confirmed that GPR37L1 is indeed a cardiovascular regulator in the mouse, but that its effects on baseline blood pressure are observed only in females. In contrast, in male mice I have shown GPR37L1 to be necessary for compensatory responses to AngII-induced hypertensive stress. In this chapter, I will discuss these findings and their broader pharmacological and physiological implications.

170

6.1. Major findings of this PhD project

6.1.1. Gαs is the cognate G protein for GPR37L1 signalling As described in Chapter 3, using heterologous expression of GPR37L1 in a yeast reporter gene

Gassay protein. system, Further I was experimentsable to demonstrate from our constitutive laboratory activity verified of that GPR37L1 GPR37L1 via itsindeed coupling constitutively to the Gα

s signalling pathway, as determined by cAMP reporter gene assays in HEK293 cells,couples transfected to the Gα to express GPR37L1 (Coleman et al., 2016). However, these approaches relied on the heterologous expression of GPR37L1 and thus I desired to confirm the finding in a model with endogenous expression of GPR37L1. Using cultures of the ex vivo mouse cerebellum, herein

(Chapter 3) I was able to demonstrate that cerebella which natively express GPR37L1 generate more cAMP than cerebella that lack the receptor. Ultimately, the identification of the cognate G protein mediating GPR37L1 signalling is an important finding that precedes a better understanding of the receptor’s cellular physiology and pharmacology. As discussed earlier (Chapter 1) traditional means of GPCR deorphanisation rely on prior knowledge of receptor and G protein pairings (Wise et al., 2004) s will aid future attempts to discover the receptor’s endogenous, and thus, ligand. the pairing Indeed, of our GPR37L1 laboratory with has Gα subsequently identified several surrogate

s activity of GPR37L1 in vitro, and hope to useligands this that informatio we haven todemonstrated guide computational to modulate modelling the Gα of the receptor and in silico screening for high-affinity GPR37L1 ligands (Ngo et al., 2017).

s, as reported within this thesis (Chapter 3), is at odds with a

Yet the coupling of GPR37L1 to Gαi (Meyer et al., 2013). I believe the modest (~5%) activity of previous report of it coupling toi G Gαprotein, as reported by Meyer et al. (2013), could be explained by

GPR37L1 -actingindependent through effects the Gα of the proposed ligands prosaptide and prosaposin (Meyer et al.,

2013). Indeed, prosaposin-induced phosphorylation of ERK in murine primary astrocytes is not

171 affected by siRNA knockdown of GPR37L1 (Meyer et al., 2013). This is perhaps unsurprising, as I could not demonstrate the expression of GPR37L1 in murine primary astrocytes (Chapter 3), and

Marazziti et al. (2013) show GPR37L1 to be only lowly expressed in whole mouse cerebellum before post-natal day 10. Similarly, Jolly et al. (2017) demonstrated that the protective effects of prosaptide against ischaemic cell death in mouse hippocampal slices were not lost in GPR37L1 knockout tissue, suggesting that, at the very least, the protective effects of prosaptide are GPR37L1- in i coupling of GPR37L1 has

dependent. Moreover, the laboratory which initially s reported to GPR37L1 the Gαand did not assess GPR37L1 since replicated ouri in data their reporting most recent the publication coupling of ( GαGiddens et al., 2017). activity through Gα Whilst the lack of independent evidence supporting the pairing of prosaptide to GPR37L1 is unfortunate, it is just one of many examples of orphan GPCRs ‘paired’ with ligands in single studies or in those that are not supported by the work of independent laboratories. Indeed, in the most recent list of recommendations for new receptor ligand pairings, as published by the International

Union of Basic and Clinical Pharmacology (IUPHAR) (Davenport et al., 2013), thirty human rhodopsin-like GPCRs remained paired with a ligand in only a single publication, or were paired in publications with discordant results to those of other independent publications. It is for this reason that the IUPHAR sets strict criteria on the classification of a ligand as being the ‘endogenous’ ligand for a GPCR. The most vital being that the pairing of the proposed ligand and receptor must be demonstrated by two or more peer-reviewed papers from independent research groups, and that the potency of the proposed ligand is consistent with a physiological function (Davenport et al.,

2013). Further evidence that is considered may include data from radioligand binding assays, functional assays performed in native tissues, experiments whereby selective antagonism blocks the action of the proposed ligand, and whether the tissue expression profile of the ligand is compatible with that of the GPCR (Davenport et al., 2013). The reasons for the low level of reproducibility in orphan GPCR research are varied, but some key considerations include whether

172 the assays used measured the appropriate G protein(s) or other downstream effectors (Ngo et al.,

2016), whether the identities and relative amounts of interacting proteins and signalling molecules in the chosen assay are representative of those of the endogenous environment of the orphan GPCR of interest, and whether artefactual activity is induced by heterologous expression of the GPCR

(Thomsen et al., 2005).

6.1.2. GPR37L1 is subject to metalloprotease-mediated N-terminal cleavage Whilst I was unable to demonstrate that prosaptide affected receptor activity in yeast reporter gene assays, I was able to generate evidence that contributed to our laboratory’s discovery of a novel, alternative mechanism of GPR37L1 signal regulation. The presence of three discrete fragments of the GPR37L1 N-terminus in human CSF suggested that GPR37L1 undergoes post- translational modification involving N-terminal proteolysis. Indeed, in Chapter 3, I show that

GPR37L1 undergoes N-terminal proteolysis as mediated by a metalloprotease, both in vitro (in

HEK293 cells) and in the ex vivo mouse cerebellum, resulting in a receptor species that lacks its entire N-terminus.

Of particular relevance to my thesis, GPR37 has also recently been shown to undergo metalloprotease-mediated N-terminal cleavage (Mattila et al., 2016), suggesting it could be regulated in a manner similar to that shown for GPR37L1. By performing deglycosylation and

Western blotting experiments, using lysates and conditioned media from HEK293 cells transfected to express an N- and C-terminally tagged GPR37, Mattila et al. (2016) demonstrated that the GPR37

N-terminus is cleaved distal to its first putative N-linked glycosylation at residue 36, but proximal to its second and third N-linked glycosylations at residues 222 and 293. In experiments similar to those I performed on GPR37L1 in this thesis, broad spectrum metalloprotease inhibitors including

GM6001 and Marimastat were found to attenuate GPR37 cleavage in vitro, demonstrating the

173 involvement of metalloprotease enzymes in the N-terminal proteolysis of GPR37 (Mattila et al.,

2016). Whilst in my thesis I have provided evidence for GPR37L1 N-terminal proteolysis being an in vivo phenomenon, this remains to be demonstrated for GPR37. However, given the close phylogenetic relationship between the two receptors, and that they are both largely CNS proteins, I predict N-terminal cleavage occurs in vivo for both GPR37 and GPR37L1.

However, an outstanding question remains with regards to where within the GPR37L1 N- terminus cleavage occurs. Herein, I was able to define the region to be after residue 105 and before

122, by comparing the molecular weights of wild-type GPR37L1 with (i.) a mutant GPR37L1 lacking its N-terminus and (ii.) wild-type GPR37L1 after deglycosylation. Unfortunately, I was unable to further define the cleavage site as proteolysis was resistant to mutagenesis within this region of the

GPR37L1 N-terminus, presumably owing to the tolerance of metalloproteases to amino acid substitutions at their target sites (Ratnikov et al., 2014). Interestingly, whilst Mattila et al. (2016) were able to determine that the location of GPR37 N-terminal cleavage was between residues 167 and 168 using N-terminal sequencing by Edman degradation, mutation of this region was insufficient to abolish proteolysis, echoing the findings of my similar experiment with GPR37L1.

The inability of mutagenesis to inhibit N-terminal cleavage of GPR37L1 was disappointing, since development of a transgenic mouse expressing a cleavage-resistant GPR37L1 would have provided crucial insight into the physiological role of N-terminal proteolysis in vivo. Whilst our laboratory still wishes to further define the site of GPR37L1 N-terminal cleavage, we have not been able to immunoprecipitate either the full-length or cleaved receptor species in high enough quantities, nor with enough purity, to yield conclusive mass spectrometry results.

174

6.1.3. Constitutive activity of GPR37L1 is regulated by the N-terminus Intriguingly, my yeast reporter gene assays, as described in Chapter 3, show that the full-length

GPR37L1 constitutively signals through G s whereas the cleaved species does not. Our laboratory has subsequently confirmed this finding inα HEK293 cell reporter gene assays in vitro (Coleman et al., 2016). I further showed that in the ex vivo mouse cerebellum, metalloprotease activity is responsible for creating an N-terminally truncated species of GPR37L1, as per the in vitro experiments (Chapter 3). However, the means by which the GPR37L1 N-terminus causes constitutive signalling remain to be elucidated.

Classically, the best described examples of N-terminal proteolysis altering GPCR signalling are the PARs, which become activated following proteolytic removal of a portion of their N-terminus, revealing a peptide ligand, tethered to the remaining N-terminus, which activates the receptor

(Ossovskaya et al., 2004). Could GPR37L1 contrast the PARs in that a tethered ligand exists on the full-length GPR37L1 N-terminus, prior to cleavage, and that proteolysis removes the tethered ligand? Certainly the melanocortin-4 receptor has constitutive activity maintained by an N-terminal tethered ligand, requiring no prior proteolysis, binding in the same location as the receptor’s endogenous ligand -melanocyte-stimulating hormone -MSH) (Ersoy et al., 2012). In this thesis I tested the hypothesisα that the GPR37L1 N-terminal fragments(α found in human CSF (and two N- terminal peptides derived from regions separating them) are peptide agonists for the receptor, by screening them against GPR37L1 in the yeast reporter gene assay system. However, I found no evidence that any of the peptides could alter GPR37L1 signalling, suggesting GPR37L1 constitutive signalling may not be due to a tethered ligand mechanism. In contrast, our laboratory’s identification of several surrogate GPR37L1 inverse agonists suggests that there may be an orthosteric interaction between the GPR37L1 N-terminus and the receptor’s transmembrane domains, and this will be discussed in more detail in section 6.2.4. Ultimately, future experiments

175 should screen a more comprehensive library of synthetic GPR37L1 N-terminal mimetic peptides, before the tethered ligand hypothesis can be excluded.

Alternatively, perhaps the GPR37L1 N-terminus constrains the receptor’s structure in a conformation permissive of G protein-coupling. Indeed, the N-terminal domain of the thyroid- stimulating hormone receptor is thought to maintain the receptor in a conformation that inhibits constitutive signalling (Zhang et al., 2000; Vlaeminck-Guillem et al., 2002), and the converse may be true for GPR37L1. Although the mechanism by which the GPR37L1 N-terminus activates the receptor remains unknown, the data in this thesis support the hypothesis that GPR37L1 exists in contrast to the PARs; receptor activity is silenced, rather than enhanced, following N-terminal proteolysis. To my knowledge, this is the only published example of a GPCR with constitutive activity that is negatively regulated by N-terminal proteolysis, and although the neighbouring

GPR37 is subject to metalloprotease-mediated N-terminal cleavage too (Mattila et al., 2016), this has yet to be studied in terms of receptor activity and function.

Although the GPR37L1 N-terminus is necessary for constitutive signalling, whether it is required for ligand-mediated receptor activation, or if the receptor has a ligand at all, remains unclear. Certainly, N-terminal proteolysis does not preclude the possibility that GPR37L1 has an endogenous ligand. It may be that proteolysis influences interactions with an endogenous ligand, in a manner comparable to that of the ETB receptor, for which N-terminal proteolysis reportedly perturbs bi-phasic responsiveness to agonists (Grantcharova et al., 2006b) -factor receptor Ste2p, for which the N-terminus constrains agonist-induced signalling, or to (Uddin the yeast et al. α, 2016).

However, it could be that GPR37L1 is without an endogenous ligand, and that its constitutive activity is its primary means of signal transduction. Indeed, there is precedent for a rhodopsin-like

GPCR having evolved constitutive activity before it evolved to respond to its endogenous ligand, again, as suggested by the work of Ersoy et al. (2012) on the melanocortin-4 receptor. The

176 constitutive activity maintained by the melanocortin-4 receptor N-terminus involves an aromatic cluster (a so called ‘toggle switch’) that is highly conserved throughout most rhodopsin-like GPCRs, whereas activation by MSH does not (Ersoy et al., 2012). Whilst it will be difficult to prove that

GPR37L1 lacks an endogenousα ligand, as absence of evidence is not evidence as absence, the precedent of the melanocortin-4 receptor (Ersoy et al., 2012) certainly raises a fascinating possibility that N-terminus-mediated constitutive active could be a primary mechanism of activation for at least some GPCRs.

6.1.4. Cleaved GPR37L1 is the predominant species in mouse cerebellum A particularly intriguing finding of this thesis is that in mouse cerebellum, the cleaved, and presumably inactive, GPR37L1 is the predominant species (Chapter 3). This suggests the physiological effects of GPR37L1 are mediated by a small reserve of active receptor, and raises the possibility that GPR37L1 signalling may be sensitive to temporal changes in metalloprotease activity. I postulate that in certain diseases where brain metalloprotease activity is increased, the amount of active GPR37L1 may be reduced by means of increased N-terminal processing. Such diseases that feature increased metalloprotease expression or activity include Alzheimer’s disease, multiple sclerosis and malignant glioma (Yong et al., 1998), cerebral ischaemia (Romanic et al.,

1998; Yang et al., 2010), and infection with human immunodeficiency virus (Sporer et al., 1998;

Zhang et al., 2003). Given that the physiological activity of GPR37L1 appears broadly beneficial in the CNS [as seen in its inhibitory effects on seizures (Giddens et al., 2017) and ischaemic cell death

(Jolly et al., 2017)], perhaps increased silencing of functional GPR37L1 could contribute to pathogenesis of the aforementioned diseases. Conversely, perhaps GPR37L1 activity, and its presumed CNS-protective effects, are enhanced during states of increased deposition of other metalloprotease substrates [such as during fibrotic scar formation following traumatic brain injury

(Kawano et al., 2012)], effectively antagonising the N-terminal proteolysis of GPR37L1 and preserving more active receptor.

177

Whether GPR37L1 processing, or even expression, in the brain is modified in disease states remains to be established. Whilst insight could be gained from interrogation of ever-expanding, and publicly available, RNA sequencing resources (Rung et al., 2013), the importance of post- translational processing for GPR37L1 means its transcription cannot be studied in isolation, and the receptor protein must also be studied to interrogate the dynamics of receptor activity. Accordingly, immunoblot analysis of healthy and diseased brain tissue could be used to investigate if there are differences in the ratios of cleaved to full-length GPR37L1 in different CNS diseases, although careful consideration should be given to the choice of brain regions to be investigated, given that

GPR37L1 is expressed in varying abundance, and in multiple cell types, throughout the brain.

Moreover, differential analysis of the proteome of CSF from healthy and diseased subjects may reveal differences in abundance of the three known GPR37L1 N-terminal fragments identified in human CSF (Stark et al., 2001; Schutzer et al., 2010; Zhao et al., 2010). One also wonders if the

GPR37L1 N-terminal fragments in human CSF are degradation products generated following removal of the N-terminus, or if they may represent peptides with biological activity, as reported for PAR1 (Furman et al., 1998). Perhaps screening these peptides for activity against remaining

CNS-expressed orphan GPCRs may reveal ‘cross-talk’ between GPR37L1 and other receptors, or provide weight to the idea that they are biologically inert.

The predominance of the cleaved GPR37L1 species in vivo will surely prove an important consideration in future attempts to pair the receptor with an endogenous ligand, or in understanding how the N-terminus is responsible for maintaining constitutive signalling.

Prospective ligands should be screened at both GPR37L1 species, as perhaps the GPR37L1 N- terminus, whilst providing constitutive activity, antagonises an orthosteric site on the receptor’s transmembrane domain [as demonstrated for the melanocortin-4 receptor (Ersoy et al., 2012)].

Our laboratory’s successful identification of several surrogate ligands for GPR37L1 (Ngo et al.,

2017) indirectly suggests the N-terminus has an orthosteric interaction with the GPR37L1

178 transmembrane domains. A computational tool (GPCR-CoINPocket) developed by our laboratory was used to predict GPCRs with similar binding pockets to GPR37L1, based on amino acid similarity at key residues of ligand contact, as derived from all available GPCR-ligand crystallographic structures at the time of study (Ngo et al., 2017). Subsequently, bombesin, orexin and neuropeptide

S receptors were predicted to have similar binding pockets to that of GPR37L1, and ligands for these receptors were screened in vitro at full-length GPR37L1 for the ability to modify cAMP production. Whilst this study led to the identification of the first surrogate ligands for GPR37L1

(two bombesin receptor antagonists and one neuropeptide S receptor antagonists), a curious observation is that they all behaved as inverse agonists at GPR37L1 (Ngo et al., 2017), suggesting that they may compete with the GPR37L1 N-terminus for access to a currently unidentified orthosteric site.

From a physiological perspective, it remains difficult to assess the relevance of the predominating

-GPR37L1 mouse cleavedline, which GPR37L1 expresses species. the IN propose-terminally that truncated it could be version assessed of bythe creation receptor of in a Δ122place of the wild-type version. It would be interesting to use these mice to measure cAMP levels in the ex vivo cerebellum model, to see whether loss of the N-terminus is sufficient to abrogate cAMP signalling as per experiments in yeast (Chapter 3) and HEK293 cells (Coleman et al., 2016), or if the receptor remains active, perhaps indicating the presence of an endogenous ligand in the cerebellum. Such an animal model would also be invaluable in studying the role of the cleaved GPR37L1 species in vivo, and if consistent with our in vitro experiments, I wou -GPR37L1 mouse line to recapitulate the phenotypes observed in GPR37L1KO/KOld mice. predict However, the Δ122 if as I hypothesised earlier,

-GPR37L1 mice may have an intermediatethat GPR37L1 phenotype, could still wh haveereby an N endogenous-terminus-mediated ligand, perhaps signalling Δ122 is lost, but sensitivity to ligand is not.

179

6.1.5. The GPR37L1 protein is abundant in the CNS but not detectable in the heart and kidney In Chapter 4 of this thesis I used a range of approaches to investigate the tissue expression of the GPR37L1 protein. A key finding of these studies was that, in contrast to reports of GPR37L1 mRNA transcript in the mouse heart and kidney, I could not detect the protein itself in either organ despite exhaustive efforts. However, the receptor was readily detected in the mouse brain, where it was localised to a variety of glial cells including the cerebellar Bergmann glia and cerebellar astrocytes, cortical and hippocampal astrocytes, as well as presumed cells of oligodendrocyte lineage, in many regions including the brainstem. Taken together, this expression pattern is consistent with that reported by other laboratories during the time of this thesis (Marazziti et al.,

2013; Jolly et al., 2017) and suggests that the cardiovascular effects of GPR37L1 reported in

Chapter 5 are solely due to the receptor’s expression in the CNS.

What then is the role of GPR37L1 mRNA in the heart? One speculative possibility is that

GPR37L1 mRNA itself has direct functions in the heart, since many protein-coding genes contain micro RNAs (miRNAs) within their introns, and such miRNAs are known to be involved in heart development and disease (Small et al., 2011). An unexpected finding of this thesis was that AngII- induced cardiac fibrosis was mitigated in GPR37L1KO/KO mice, and given that miRNAs are known to be involved with pathological remodelling of the heart, and in particular, fibrosis (Small et al.,

2011), perhaps miRNAs arising from GPR37L1 are involved in regulation of the extracellular matrix. However, the hypothesis I most favour is that cardiac GPR37L1 mRNA transcription simply represents ‘leaky transcription’, whereby the gene’s transcription may not be tightly controlled enough to allow complete silencing in a particular cell type (Yanai et al., 2006). Such ‘leaky’ or ectopic expression is thought to occur where a specific gene is targeted for transcription but regional chromatin rearrangements allow transcriptional machinery to access neighbouring genes

(Yanai et al., 2006). Yanai et al. (2006) suggest that evolutionary selection against unnecessary

180 transcription is not particularly strong, given that repression of transcription can itself be energetically costly, and that more stringent control is often performed at the post-transcriptional level, with approximately 30% of human genes subject to regulation by miRNAs (Lewis et al., 2005).

6.1.6. Deletion of GPR37L1 increases blood pressure in female mice In Chapter 5 of this thesis I describe the determination of the basal blood pressure phenotype of mice lacking GPR37L1. As key points of difference to a previous study (Min et al., 2010), the mice developed for my studies share a genetic background, and, also, a suitable wild-type control strain was included in the study. These experiments revealed that female, but not male, mice lacking

GPR37L1 have elevated systolic and diastolic blood pressure, and that this phenotype was preserved in both anaesthetised and conscious, freely-moving mice.

In light of the CNS-restricted expression of GPR37L1, the discovery of an elevated blood pressure phenotype in GPR37L1KO/KO female mice suggests that the receptor is both a component of central cardiovascular control and one with sexually dimorphic effects. Even momentarily overlooking the complexity of the sex-dependent nature of the GPR37L1KO/KO phenotype, many questions still remain as to how GPR37L1 controls blood pressure. Firstly, which of the many brain regions expressing GPR37L1 might explain the cardiovascular phenotype? The abundance of

GPR37L1 in the cerebellum suggests the receptor has important functions in this region. Indeed,

Marazziti et al. (2013) demonstrated GPR37L1 to regulate cerebellar development and to influence motor learning. Given the well described role of the cerebellum in motor learning (Halsband et al.,

2006), it is likely the effects of GPR37L1 on motor learning are attributable to its expression in the cerebellum. As inferred from the study of patients with cerebellar lesions, the organ is also involved in executive function, spatial cognition, personality and linguistics (Schmahmann et al., 1998), yet

GPR37L1 remains to be associated with these traits. But is there a precedent for a cardiovascular role of the cerebellum? Stimulation of many cerebellar regions including the fastigial nucleus,

181 anterior vermis, posterior vermis, uvula and nodulus has been shown to have a wide variety of modulatory effects on haemodynamic control, many of which were dependent on sympathetic nerve activity (Nisimaru, 2004). Moreover, regions of the cerebellum have either been shown, or hypothesised, to be capable of communicating with important cardiovascular control nuclei elsewhere in the brain, such as the parabrachial nucleus (PBN), NTS and RVLM (Nisimaru, 2004).

Further, Korner (2007) reasoned that the close linkage between skeletal muscle- and autonomic nervous system-responses to exercise represents the cerebellum’s ability to coordinate both functions. However, despite links between the cerebellum and cardiovascular control processes, I and others (Marazziti et al., 2013) observe cerebellar GPR37L1 to be most concentrated in the

Bergmann glia of the cerebellar molecular layer – a region and cell-type that, to my knowledge, are yet to be linked to cardiovascular function.

If not the cerebellum, which other brain regions may mediate the cardiovascular effects of

GPR37L1? It is noteworthy that I have observed expression of GPR37L1 in the brainstem, and within astrocytes in proximity to catecholaminergic neurons within brainstem nuclei (including the

CVLM, RVLM and NTS), given the well described role of these nuclei and neurons in sympathetic control of blood pressure (Guyenet, 2006). It is possible that GPR37L1 may influence sympathetic control of the kidney or vasculature, via its activity within the brainstem. Although the expression of GPR37L1 in different brain cell types has yet to be compared between the sexes, I did not find any sex differences in expression of GPR37L1 in the mouse cerebellum or human brain (Chapter 5).

Perhaps the influence of sex on the GPR37L1 phenotype could be explained by the targeting of catecholaminergic neurons in the brainstem by both male and female hormones (Heritage et al.,

1980), in a manner sensitive to regulation by GPR37L1-expressing cells. Physiologically, estrogen has broadly hypotensive effects via its actions in the brain and its sub regions, including the RVLM and NTS (Hay et al., 2014), and androgens can sensitise pressor responses and the baroreflex (El-

Mas et al., 2001). Combining the expression profile of GPR37L1 with the female-specific nature of

182 the phenotype after its deletion, perhaps GPR37L1 is a component of a sympathetic cardiovascular control pathway that is sensitised by, or responsive to, estrogen.

However, a contrasting hypothesis to those presented above is that the haemodynamic effects of GPR37L1 are secondary to it being a mediator of stress. I observed the mouse hippocampus to contain many GPR37L1-expressing astrocytes, and this is of interest owing both to the ability of the hippocampus to provide negative feedback regulation of the hypothalamic-pituitary axis (HPA)

(Jacobson et al., 1991) and the cardiovascular effects of the HPA’s chief glucocorticoid effectors, including cortisol (Whitworth et al., 2005) and corticosterone (Whitworth et al., 2001).

Glucocorticoids, released from the adrenal cortex following stimulation by adrenocorticotropic hormone (ACTH), which itself is released from the pituitary gland in response to stress, increase blood pressure by a variety of mechanisms including increased renal vascular resistance and decreased nitric oxide-mediated vasodilation (Whitworth et al., 2005). As a point of difference between mouse and human glucocorticoid-induced hypertension, corticosterone is the predominant glucocorticoid in the former, and cortisol in the latter (Whitworth et al., 2001). Yet, of relevance to the sexual dimorphism observed in the present study, sex and species differences are found within this system: men appear to have greater HPA activity than women, yet female mice appear to have greater HPA activity than males (Kudielka et al., 2005). Speculatively, loss of

GPR37L1 in the mouse, perhaps specifically within the hippocampus, may disinhibit the HPA, increasing glucocorticoid levels in the GPR37L1KO/KO mice and further potentiating the cardiovascular response to stress in female mice. Given that men have greater HPA activity than women (Kudielka et al., 2005), perhaps the opposite could be true in the context of human health, and that loss of GPR37L1 could disinhibit the stress response in both sexes, but be more deleterious in males than in females.

183

The present study has generated many hypotheses about the mechanism of cardiovascular control by GPR37L1, and future experiments from our laboratory will be aimed at testing these. A study commencing soon in our laboratory aims to determine the respective contributions of stress and the SNS to the GPR37L1 blood pressure phenotype. These experiments will involve radiotelemetry blood pressure measurement in conscious mice during a variety of treatments, including exposure to stressful environmental stimuli and pharmacological ganglionic blockade

[similar to experiments described by Davern et al. (2009)]. Cross-spectral analysis of the resulting radiotelemetry blood pressure and heart rate data will also allow assessment of baroreflex sensitivity in GPR37L1KO/KO mice. Further, it would be of interest to investigate if there are increases in tyrosine hydroxylase immunoreactivity in different brain regions in the GPR37L1KO/KO mice, as an indicator of enhanced SNS activity therein (Burgi et al., 2011). In tandem, we will be characterising the function of ex vivo aortae and resistance vessels from GPR37L1KO/KO mice using wire myography, to determine if they differ from wild-type vessels in responsiveness to stimuli including noradrenaline, nitric oxide, endothelin-1 or AngII [similar to experiments described by Stanley et al.

(2014)]. Moreover, as I hypothesised that GPR37L1KO/KO mice may have enhanced HPA activity leading to heightened sympathetic sensitivity, a future study could investigate if the transcription, translation, or surface expression of adrenergic receptors in GPR37L1KO/KO tissue (vasculature, kidney, heart and brain) is increased relative to wild- 1- adrenoceptor for example (Davies et al., 1984). This idea type,could as be is further known explored to occur by formeasuring the β genotype differences in levels of components of the HPA, such as corticosterone or ACTH [similar to experiments described by Timpl et al. (1998)], in GPR37L1KO/KO mice.

184

6.1.7. Male mice lacking GPR37L1 have an exacerbated response to AngII- induced hypertension To determine if GPR37L1 is protective in the setting of cardiovascular stress, mice lacking

GPR37L1 were infused with AngII for seven days to model hypertension. Despite their baseline blood pressure elevation, female GPR37L1KO/KO mice responded to AngII similarly to their wild-type counterparts, and did not differ in terms of post-treatment blood pressure or LVH, and neither genotype showed signs of heart failure (Chapter 5). In contrast, male mice lacking GPR37L1 fared worse than their wild-type counterparts in response to AngII, displaying heightened LVH and signs of heart failure (pulmonary oedema and depressed LV contractile function) (Chapter 5).

The fact that male GPR37L1KO/KO mice showed no baseline phenotype, yet responded more poorly to AngII than their female counterparts (who did have a baseline phenotype), was an unexpected finding of this study. However, this could be explained by sex differences in sensitivity to AngII infusion. Females are less susceptible to AngII-induced hypertension than males, with a previous study reporting that a dose (1.152 mg/kg/day) sufficient to increase conscious mean arterial pressure by ~7 mmHg in females, increased it by ~35 mmHg in males (Xue et al., 2005).

Considering that in our study, that female wild-type mice responded to AngII with a ~12 mmHg increase in mean arterial pressure, yet male wild-type mice did not exhibit significant AngII hypertension by seven days, I concluded that the higher dose (2 mg/kg/day) used in my studies caused a blood pressure increase in males sufficient to induce pressure overload, resulting in decompensation and reduced blood pressure, as known to occur in rodent pressure overload

(Norton et al., 2002).

Perhaps the deleterious effects of the GPR37L1KO/KO genotype were only observed in males because they were subject to a greater AngII-mediated increase in blood pressure than the females.

One wonders if the male GPR37L1KO/KO phenotype of exacerbated LVH, decreased contractile

185 function and pulmonary congestion in response to AngII may be seen in female GPR37L1KO/KO mice, had they been subject to equivalent hypertensive stress.

However, the reasons for the genotype-dependent response to AngII, as observed in the male mice, remain unclear. I propose three subtly distinct hypotheses that could each explain this phenomenon. The first is that both genotypes were equally sensitive to the molecular consequences of AngII, yet wild-type mice were better able to evoke compensatory mechanisms to an equivalent haemodynamic challenge than the GPR37L1KO/KO mice. This infers that irrespective of the choice of model used to induce hypertension, a broad haemodynamic control mechanism has been lost in the

GPR37L1KO/KO mice. For instance, perhaps GPR37L1KO/KO mice could not effectively mount baroreflex-mediated reductions in SNS activity. This would result in greater haemodynamic stress and would lead to a greater degree of LVH [known to increase proportionally with blood pressure in both mouse and human hypertension (Rowlands et al., 1981; Crowley et al., 2006)], which could be sufficient to induce LV dysfunction and heart-failure (Prisant, 2005), as observed in this thesis

(Chapter 5). The second hypothesis is that loss of GPR37L1 specifically increases tissue sensitivity to circulating AngII, causing an unequal hypertensive challenge between genotypes. For example,

KO/KO GPR37L1 mice might have greater expression of AT1 receptors in the brain’s AngII-responsive circumventricular organs (Guyenet, 2006), which could cause heightened sympathetic outflow following stimulation of the brain by infused AngII (Guyenet, 2006). Again, in this hypothesis, the heart failure phenotype of GPR37L1KO/KO males is explained directly by effects of a greater hypertensive load. Whilst it could be entertained that the greater LVH seen in GPR37L1KO/KO males is a direct effect of the myocardium being more sensitive to AngII, the absence of detectable

GPR37L1 protein in this tissue (Chapter 4) makes it unlikely that GPR37L1 ablation has directly affected the heart, and more likely that differences in LVH are consequences of differences in pressure. It could also be that the genotypes differed in sensitivity to the non-haemodynamic effects of AngII. For example, perhaps GPR37L1KO/KO males had enhanced myocardial NADPH oxidase

186 activity, which contributes to LVH, and is known to be stimulated in mice by doses of AngII as low as 0.3 mg/kg/day (Bendall et al., 2002). Alternatively, perhaps there were differences in inflammatory responses to AngII, as more extreme doses in mice are shown to invoke IL-6 signalling [5.184 mg/kg/day (Lee et al., 2006)] and increases in monocyte chemoattractant protein-

1 in the vasculature [3.6 mg/kg/day (Jin et al., 2012)]. The third possibility relates to a hypothesis I invoked to explain how GPR37L1 may regulate baseline blood pressure (in female mice), and is that

GPR37L1KO/KO mice may have elevated serum glucocorticoids (i.e corticosterone) or HPA activity. It has long been known that adrenal steroid hormones can sensitise tissues to adrenergic stimulation

(Davies et al., 1984), and GPR37L1KO/KO mice may be subject to adrenergic hypersensitivity, mediating enhanced vasoconstriction and LVH. Perhaps, AngII induced an equivalent hypertensive challenge between genotypes, yet for the GPR37L1KO/KO males, the stimulatory effects of AngII on the SNS were enhanced by greater tissue sensitivity to catecholamines.

As for the study of GPR37L1KO/KO mice at baseline, the aforementioned AngII infusion experiments have raised many questions about the cardiovascular role of GPR37L1. Whilst I predict that radiotelemetry and vessel myography experiments will enable better interpretation of the present findings, much could still be learned from studying the response of GPR37L1KO/KO mice to a different cardiovascular stressor. Thoracic aortic constriction (TAC), for example, would allow a better matching of the extent of cardiovascular challenge between the sexes, and help avoid the confounding influence of sex differences in sensitivity to AngII-infusion hypertension. This experiment might help reveal both (i.) whether the deleterious response of male GPR37L1KO/KO mice to cardiovascular challenge is truly a sex-specific phenomenon and (ii.) whether the apparent importance of GPR37L1 for compensation to cardiovascular challenge is model-specific or more broadly applicable.

187

Further, there are two more unexpected findings in the present study that cannot be completely resolved with the data herein. Firstly, why was AngII hypertension not sustained in male mice of either genotype? Whilst echocardiography was not performed in the present study due to resource demands, it would have been invaluable for following the temporal changes in LV structure and function that preceded the apparent pressure-overload in males of both genotypes. Similarly, if it had been possible to perform radiotelemetry blood pressure measurement, instead of measurement via catheterisation, then the time course of blood pressure could have been tracked across the seven days of infusion. If the study were to be repeated, or done using a different model of cardiovascular disease, it should ideally be performed using both radiotelemetry and echocardiography. The second curiosity that cannot be addressed by the present data is why

GPR37L1KO/KO mice were protected from AngII-induced cardiovascular fibrosis. Although, I formerly suggested a role of GPR37L1 transcript in miRNA-regulation of the extracellular matrix, this is highly speculative. However, this could be explored by comparing the transcriptome of AngII- treated wild-type and GPR37L1KO/KO hearts, using RNA sequencing. This would be performed using female tissues, as I have shown that both genotypes achieve a similar degree of LVH in response to

AngII, whereas the males are confounded by their genotype-dependent LVH response. This data could be used to interrogate changes in genes known to play a role in remodelling, which may indirectly regulated by GPR37L1 transcript.

188

6.2. Conclusion The experiments described within this thesis have generated new understanding of GPR37L1, both in terms of pharmacology and physiology. Importantly, these findings have demonstrated that much can be learned about orphan receptors, despite the difficulties intrinsic to studying them without ligands. However, for every hypothesis tested herein, many more have been raised.

Exciting questions remain in regards to the mechanisms that link the CNS expression of GPR37L1 with its sexually-dimorphic effects on the cardiovascular system, and in regards to the significance of N-terminal proteolysis for the broader physiological roles of GPR37L1.

189

Chapter Seven: Appendices

190

Males Females

Parameter Effect F (DFn, DFd) P value F (DFn, DFd) P value

Aortic HR Interaction F (1, 32) = 0.08 P=0.782 F (1, 33) = 0.04 P=0.852

Genotype F (1, 32) = 0.06 P=0.814 F (1, 33) = 2.38 P=0.139

AngII F (1, 32) = 1.08 P=0.307 F (1, 33) = 0.10 P=0.760

P=0.828 Aortic SP Interaction F (1, 32) = 0.05 F (1, 33) = 2.81 P=0.103

Genotype F (1, 32) = 1.65 P=0.208 F (1, 33) = 0.42 P=0.522

AngII F (1, 32) = 4.27 P=0.047 F (1, 33) = 17.24 P<0.001

Aortic DP Interaction F (1, 32) = 0.07 P=0.791 F (1, 33) = 0.15 P=0.670

Genotype F (1, 32) = 0.82 P=0.373 F (1, 33) = 2.62 P=0.115

AngII F (1, 32) = 0.43 P=0.518 F (1, 33) = 3.45 P=0.072

Aortic MAP Interaction F (1, 32) = 0.07 P=0.801 F (1, 33) = 1.08 P=0.307

Genotype F (1, 32) = 1.15 P=0.292 F (1, 33) = 1.46 P=0.235

AngII F (1, 32) = 1.47 P=0.235 F (1, 33) = 9.39 P<0.001

Aortic PP Interaction F (1, 32) < 0.01 P=0.994 F (1, 33) = 5.70 P=0.023

Genotype F (1, 32) = 1.53 P=0.225 F (1, 33) = 0.19 P=0.672

AngII F (1, 32) = 12.45 P=0.001 F (1, 33) = 25.99 P<0.001

LV HR Interaction F (1, 31) = 0.80 P=0.378 F (1, 31) = 0.12 P=0.733

Genotype F (1, 31) = 0.44 P=0.511 F (1, 31) = 2.58 P=0.118

AngII F (1, 31) = 0.81 P=0.375 F (1, 31) = 0.18 P=0.677

LV SP Interaction F (1, 31) = 0.87 P=0.359 F (1, 31) = 1.92 P=0.176

Genotype F (1, 31) = 3.29 P=0.080 F (1, 31) = 0.38 P=0.540

AngII F (1, 31) = 3.71 P=0.064 F (1, 31) = 31.92 P<0.001

LV EDP Interaction F (1, 31) = 1.50 P=0.231 F (1, 31) = 2.64 P=0.114

Genotype F (1, 31) = 0.06 P=0.811 F (1, 31) = 0.10 P=0.754

AngII F (1, 31) = 1.09 P=0.305 F (1, 31) = 5.47 P=0.026

LV dP/dT max Interaction F (1, 31) = 2.28 P=0.142 F (1, 31) < 0.01 P=0.945

Genotype F (1, 31) = 12.03 P=0.002 F (1, 31) = 5.71 P=0.023

191

AngII F (1, 31) = 5.95 P=0.021 F (1, 31) = 3.39 P=0.075

LV dP/dT min Interaction F (1, 31) = 0.02 P=0.891 F (1, 31) = 0.04 P=0.851

Genotype F (1, 31) = 2.36 P=0.135 F (1, 31) = 0.36 P=0.553

AngII F (1, 31) = 1.71 P=0.200 F (1, 31) = 0.34 P=0.563

HW:TL Interaction F (1, 86) = 5.25 P=0.024 F (1, 69) = 0.35 P=0.557

Genotype F (1, 86) = 13.44 P<0.001 F (1, 69) = 2.13 P=0.149

AngII F (1, 86) = 40.15 P<0.001 F (1, 69) = 86 P<0.001

LVW:TL Interaction F (1, 86) = 6.27 P=0.014 F (1, 69) = 0.33 P=0.569

Genotype F (1, 86) = 19.17 P<0.001 F (1, 69) = 5.59 P=0.021

AngII F (1, 86) = 67.31 P<0.001 F (1, 69) = 129.7 P<0.001

RVW:TL Interaction F (1, 86) = 0.49 P=0.485 F (1, 69) = 0.48 P=0.489

Genotype F (1, 86) = 1.86 P=0.176 F (1, 69) = 0.45 P=0.507

AngII F (1, 86) = 2.22 P=0.140 F (1, 69) = 0.51 P=0.477

Lung:TL Interaction F (1, 81) = 3.79 P=0.055 F (1, 69) = 0.91 P=0.344

Genotype F (1, 81) = 0.93 P=0.337 F (1, 69) = 0.13 P=0.717

AngII F (1, 81) = 2.36 P=0.128 F (1, 69) < 0.01 P=0.986

CM density Interaction F (1, 22) = 3.06 P=0.094 F (1, 22) = 0.16 P=0.694

Genotype F (1, 22) = 0.53 P=0.473 F (1, 22) = 2.16 P=0.156

AngII F (1, 22) = 12.02 P=0.002 F (1, 22) = 25.15 P<0.001

Fibrosis Interaction F (1, 21) = 3.77 P=0.066 F (1, 24) = 10.3 P=0.009 F (1, 24) = 2.02 Genotype F (1, 21) = 0.71 P=0.408 P=0.168

AngII F (1, 21) = 17.64 P<0.001 F (1, 24) = 0.53 P=0.474

pressure; MAP, meanUnderline arterial highlightspressure; PP,p ≤0.05. pulse HR, pressure; heart rate; EDP, SP, end systolic-diastolic pressure; pressure; DP, HW,diastolic heart weight; LVW, left ventricle weight; TL, tibia length; LV, left ventricle; RVW, right ventricle weight; CM, cardiomyocyte.

192

Males Females

Parameter Effect F (DFn, DFd) P value F (DFn, DFd) P value

F (1, 35) = 7.74 P<0.001 ANP Interaction F (1, 37) = 0.12 P=0.732

F (1, 35) = 3.14 P=0.085 Genotype F (1, 37) = 0.35 P=0.560

F (1, 35) = 23.53 P<0.001 AngII F (1, 37) = 19.47 P<0.001

BNP Interaction F (1, 37) = 0.11 P=0.738 F (1, 35) = 2.65 P=0.113

Genotype F (1, 37) = 0.19 P=0.663 F (1, 35) = 0.69 P=0.414

AngII F (1, 37) = 2.87 P=0.099 F (1, 35) = 24.24 P<0.001

β-MHC Interaction F (1, 37) = 1.35 P=0.258 F (1, 35) = 0.32 P=0.577

Genotype F (1, 37) = 1.80 P=0.188 F (1, 35) = 1.99 P=0.167

AngII F (1, 37) = 7.69 P=0.009 F (1, 35) = 2.60 P=0.116

α-SkA Interaction F (1, 37) = 1.20 P=0.274 F (1, 35) = 0.03 P=0.860

Genotype F (1, 37) = 0.92 P=0.344 F (1, 35) = 0.18 P=0.677

AngII F (1, 37) = 19.59 P<0.001 F (1, 35) = 13.53 P<0.001

ANP, atrial natriuretic peptide; BNP, bra -MHC, - -

SkA, - in natriuretic peptide; β β myosin heavy chain; α

α skeletal actin. Underline highlights p ≤0.05.

193

Chapter Eight: Bibliography

194

Adachi Y, Saito Y, Kishimoto I, Harada M, Kuwahara K, Takahashi N, et al. (2003). Angiotensin II type 2 receptor deficiency exacerbates heart failure and reduces survival after acute myocardial infarction in mice. Circulation 107(19): 2406-2408.

Alexander SP, Christopoulos A, Davenport AP, Kelly E, Marrion NV, Peters JA, et al. (2017). THE CONCISE GUIDE TO PHARMACOLOGY 2017/18: G protein-coupled receptors. British journal of pharmacology 174 Suppl 1: S17-S129.

Anrep GV, Segall HN (1926). The central and reflex regulation of the heart rate. The Journal of physiology 61(2): 215-231.

Arribas J, Coodly L, Vollmer P, Kishimoto TK, Rose-John S, Massague J (1996). Diverse cell surface protein ectodomains are shed by a system sensitive to metalloprotease inhibitors. The Journal of biological chemistry 271(19): 11376-11382.

Audet M, Bouvier M (2012). Restructuring G-protein- coupled receptor activation. Cell 151(1): 14- 23.

Bader M (2010). Tissue renin-angiotensin-aldosterone systems: Targets for pharmacological therapy. Annual review of pharmacology and toxicology 50: 439-465.

Baker JG (2005). The selectivity of beta-adrenoceptor antagonists at the human beta1, beta2 and beta3 adrenoceptors. British journal of pharmacology 144(3): 317-322.

Barsha G, Denton KM, Mirabito Colafella KM (2016). Sex- and age-related differences in arterial pressure and albuminuria in mice. Biology of sex differences 7: 57.

Basile J (2004). The role of existing and newer calcium channel blockers in the treatment of hypertension. Journal of clinical hypertension 6(11): 621-629; quiz 630-621.

Beermann S, Seifert R, Neumann D (2012). Commercially available antibodies against human and murine histamine H(4)-receptor lack specificity. Naunyn-Schmiedeberg's archives of pharmacology 385(2): 125-135.

Belmer A, Doly S, Setola V, Banas SM, Moutkine I, Boutourlinsky K, et al. (2014). Role of the N- terminal region in G protein-coupled receptor functions: negative modulation revealed by 5-HT2B receptor polymorphisms. Molecular pharmacology 85(1): 127-138.

195

Bendall JK, Cave AC, Heymes C, Gall N, Shah AM (2002). Pivotal role of a gp91(phox)-containing NADPH oxidase in angiotensin II-induced cardiac hypertrophy in mice. Circulation 105(3): 293- 296.

Bhatt DL, Kandzari DE, O'Neill WW, D'Agostino R, Flack JM, Katzen BT, et al. (2014). A controlled trial of renal denervation for resistant hypertension. The New England journal of medicine 370(15): 1393-1401.

Bidani AK, Griffin KA (2004). Pathophysiology of hypertensive renal damage: implications for therapy. Hypertension 44(5): 595-601.

Bilusic M, Bataillard A, Tschannen MR, Gao L, Barreto NE, Vincent M, et al. (2004). Mapping the genetic determinants of hypertension, metabolic diseases, and related phenotypes in the lyon hypertensive rat. Hypertension 44(5): 695-701.

Bockaert J, Pin JP (1999). Molecular tinkering of G protein-coupled receptors: an evolutionary success. The EMBO journal 18(7): 1723-1729.

Bohm SK, Grady EF, Bunnett NW (1997). Regulatory mechanisms that modulate signalling by G- protein-coupled receptors. The Biochemical journal 322 ( Pt 1): 1-18.

Boire A, Covic L, Agarwal A, Jacques S, Sherifi S, Kuliopulos A (2005). PAR1 is a matrix metalloprotease-1 receptor that promotes invasion and tumorigenesis of breast cancer cells. Cell 120(3): 303-313.

Bradley A, Anastassiadis K, Ayadi A, Battey JF, Bell C, Birling MC, et al. (2012). The mammalian gene function resource: the International Knockout Mouse Consortium. Mammalian genome : official journal of the International Mammalian Genome Society 23(9-10): 580-586.

Bradley DJ, Ghelarducci B, Paton JF, Spyer KM (1987a). The cardiovascular responses elicited from the posterior cerebellar cortex in the anaesthetized and decerebrate rabbit. The Journal of physiology 383: 537-550.

Bradley DJ, Pascoe JP, Paton JF, Spyer KM (1987b). Cardiovascular and respiratory responses evoked from the posterior cerebellar cortex and fastigial nucleus in the cat. The Journal of physiology 393: 107-121.

Brew K, Nagase H (2010). The tissue inhibitors of metalloproteinases (TIMPs): an ancient family with structural and functional diversity. Biochimica et biophysica acta 1803(1): 55-71.

196

Brown AJ, Goldsworthy SM, Barnes AA, Eilert MM, Tcheang L, Daniels D, et al. (2003). The Orphan G protein-coupled receptors GPR41 and GPR43 are activated by propionate and other short chain carboxylic acids. The Journal of biological chemistry 278(13): 11312-11319.

Brown NJ, Vaughan DE (1998). Angiotensin-converting enzyme inhibitors. Circulation 97(14): 1411-1420.

Bruneval P, Hinglais N, Alhenc-Gelas F, Tricottet V, Corvol P, Menard J, et al. (1986). Angiotensin I converting enzyme in human intestine and kidney. Ultrastructural immunohistochemical localization. Histochemistry 85(1): 73-80.

Bugge-Asperheim B, Kiil F (1973). Preload, contractility, and afterload as determinants of stroke volume during elevation of aortic blood pressure in dogs. Cardiovascular research 7(4): 328-341.

Burgi K, Cavalleri MT, Alves AS, Britto LR, Antunes VR, Michelini LC (2011). Tyrosine hydroxylase immunoreactivity as indicator of sympathetic activity: simultaneous evaluation in different tissues of hypertensive rats. American journal of physiology. Regulatory, integrative and comparative physiology 300(2): R264-271.

Burnier M (2001). Angiotensin II type 1 receptor blockers. Circulation 103(6): 904-912.

Butz GM, Davisson RL (2001). Long-term telemetric measurement of cardiovascular parameters in awake mice: a physiological genomics tool. Physiological genomics 5(2): 89-97.

Byrne JA, Grieve DJ, Bendall JK, Li JM, Gove C, Lambeth JD, et al. (2003). Contrasting roles of NADPH oxidase isoforms in pressure-overload versus angiotensin II-induced cardiac hypertrophy. Circulation research 93(9): 802-805.

Caldwell PR, Seegal BC, Hsu KC, Das M, Soffer RL (1976). Angiotensin-converting enzyme: vascular endothelial localization. Science 191(4231): 1050-1051.

Canals M, Lopez-Gimenez JF, Milligan G (2009). Cell surface delivery and structural re-organization by pharmacological chaperones of an oligomerization-defective alpha(1b)-adrenoceptor mutant demonstrates membrane targeting of GPCR oligomers. The Biochemical journal 417(1): 161-172.

Chen D, Bassi JK, Walther T, Thomas WG, Allen AM (2010). Expression of angiotensin type 1A receptors in C1 neurons restores the sympathoexcitation to angiotensin in the rostral ventrolateral medulla of angiotensin type 1A knockout mice. Hypertension 56(1): 143-150.

197

Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jr., et al. (2003). Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension 42(6): 1206-1252.

Choi HY, Park HC, Ha SK (2015). Salt Sensitivity and Hypertension: A Paradigm Shift from Kidney Malfunction to Vascular Endothelial Dysfunction. Electrolyte & blood pressure : E & BP 13(1): 7-16.

Christopoulos A, Kenakin T (2002). G protein-coupled receptor allosterism and complexing. Pharmacological reviews 54(2): 323-374.

Cimarosti H, O'Shea RD, Jones NM, Horn AP, Simao F, Zamin LL, et al. (2006). The effects of estradiol on estrogen receptor and glutamate transporter expression in organotypic hippocampal cultures exposed to oxygen--glucose deprivation. Neurochemical research 31(4): 483-490.

Civelli O (2005). GPCR deorphanizations: the novel, the known and the unexpected transmitters. Trends in pharmacological sciences 26(1): 15-19.

Coffman TM (2011). Under pressure: the search for the essential mechanisms of hypertension. Nature medicine 17(11): 1402-1409.

Coffman TM (2014). The inextricable role of the kidney in hypertension. The Journal of clinical investigation 124(6): 2341-2347.

Coleman JL, Brennan K, Ngo T, Balaji P, Graham RM, Smith NJ (2015). Rapid Knockout and Reporter Mouse Line Generation and Breeding Colony Establishment Using EUCOMM Conditional-Ready Embryonic Stem Cells: A Case Study. Frontiers in endocrinology 6: 105.

Coleman JL, Ngo T, Schmidt J, Mrad N, Liew CK, Jones NM, et al. (2016). Metalloprotease cleavage of the N terminus of the orphan G protein-coupled receptor GPR37L1 reduces its constitutive activity. Science signaling 9(423): ra36.

Coleman JL, Ngo T, Smith NJ (2017). The G protein-coupled receptor N-terminus and receptor signalling: N-tering a new era. Cellular signalling 33: 1-9.

Consortium GT (2015). Human genomics. The Genotype-Tissue Expression (GTEx) pilot analysis: multitissue gene regulation in humans. Science 348(6235): 648-660.

Conway J (1984). Hemodynamic aspects of essential hypertension in humans. Physiological reviews 64(2): 617-660.

198

Cook JL, Mills SJ, Naquin RT, Alam J, Re RN (2007). Cleavage of the angiotensin II type 1 receptor and nuclear accumulation of the cytoplasmic carboxy-terminal fragment. American journal of physiology. Cell physiology 292(4): C1313-1322.

Corbier A, Lecaque D, Secchi J, Depouez B, Hamon G (1994). Effects of 4 weeks of treatment with trandolapril on renal hypertension and cardiac and vascular hypertrophy in the rat. Journal of cardiovascular pharmacology 23 Suppl 4: S26-29.

Corvol P, Jeunemaitre X (1997). Molecular genetics of human hypertension: role of angiotensinogen. Endocrine reviews 18(5): 662-677.

Cotton LM, Meilak ML, Templeton T, Gonzales JG, Nenci A, Cooney M, et al. (2015). Utilising the resources of the International Knockout Mouse Consortium: the Australian experience. Mammalian genome : official journal of the International Mammalian Genome Society 26(3-4): 142-153.

Coughlin SR (2000). Thrombin signalling and protease-activated receptors. Nature 407(6801): 258-264.

Couvineau A, Fabre C, Gaudin P, Maoret JJ, Laburthe M (1996). Mutagenesis of N-glycosylation sites in the human vasoactive intestinal peptide 1 receptor. Evidence that asparagine 58 or 69 is crucial for correct delivery of the receptor to plasma membrane. Biochemistry 35(6): 1745-1752.

Crews JK, Khalil RA (1999). Gender-specific inhibition of Ca2+ entry mechanisms of arterial vasoconstriction by sex hormones. Clinical and experimental pharmacology & physiology 26(9): 707-715.

Crowley SD, Gurley SB, Herrera MJ, Ruiz P, Griffiths R, Kumar AP, et al. (2006). Angiotensin II causes hypertension and cardiac hypertrophy through its receptors in the kidney. Proceedings of the National Academy of Sciences of the United States of America 103(47): 17985-17990.

Cushman WC, Ford CE, Cutler JA, Margolis KL, Davis BR, Grimm RH, et al. (2002). Success and predictors of blood pressure control in diverse North American settings: the antihypertensive and lipid-lowering treatment to prevent heart attack trial (ALLHAT). Journal of clinical hypertension 4(6): 393-404.

Cuspidi C, Sala C, Negri F, Mancia G, Morganti A, Italian Society of H (2012). Prevalence of left- ventricular hypertrophy in hypertension: an updated review of echocardiographic studies. Journal of human hypertension 26(6): 343-349.

199

Dahmane N, Ruiz i Altaba A (1999). Sonic hedgehog regulates the growth and patterning of the cerebellum. Development 126(14): 3089-3100.

Das M, Hartley JL, Soffers RL (1977). Serum angiotensin-converting enzyme. Isolation and relationship to the pulmonary enzyme. The Journal of biological chemistry 252(4): 1316-1319.

Davenport AP, Alexander SP, Sharman JL, Pawson AJ, Benson HE, Monaghan AE, et al. (2013). International Union of Basic and Clinical Pharmacology. LXXXVIII. G protein-coupled receptor list: recommendations for new pairings with cognate ligands. Pharmacological reviews 65(3): 967-986.

Davern PJ, Nguyen-Huu TP, La Greca L, Abdelkader A, Head GA (2009). Role of the sympathetic nervous system in Schlager genetically hypertensive mice. Hypertension 54(4): 852-859.

Davies AO, Lefkowitz RJ (1984). Regulation of beta-adrenergic receptors by steroid hormones. Annual review of physiology 46: 119-130.

de Mendoza A, Sebe-Pedros A, Ruiz-Trillo I (2014). The evolution of the GPCR signaling system in eukaryotes: modularity, conservation, and the transition to metazoan multicellularity. Genome biology and evolution 6(3): 606-619.

de Simone G, Devereux RB, Camargo MJ, Wallerson DC, Laragh JH (1993). Influence of sodium intake on in vivo left ventricular anatomy in experimental renovascular hypertension. The American journal of physiology 264(6 Pt 2): H2103-2110.

Deslauriers B, Ponce C, Lombard C, Larguier R, Bonnafous JC, Marie J (1999). N-glycosylation requirements for the AT1a angiotensin II receptor delivery to the plasma membrane. The Biochemical journal 339 ( Pt 2): 397-405.

DiBona GF (2002). Sympathetic nervous system and the kidney in hypertension. Current opinion in nephrology and hypertension 11(2): 197-200.

Doggrell SA, Brown L (1998). Rat models of hypertension, cardiac hypertrophy and failure. Cardiovascular research 39(1): 89-105.

Dong C, Filipeanu CM, Duvernay MT, Wu G (2007). Regulation of G protein-coupled receptor export trafficking. Biochimica et biophysica acta 1768(4): 853-870.

Donoghue M, Hsieh F, Baronas E, Godbout K, Gosselin M, Stagliano N, et al. (2000). A novel angiotensin-converting enzyme-related carboxypeptidase (ACE2) converts angiotensin I to angiotensin 1-9. Circulation research 87(5): E1-9.

200

Donohue PJ, Shapira H, Mantey SA, Hampton LL, Jensen RT, Battey JF (1998). A human gene encodes a putative G protein-coupled receptor highly expressed in the central nervous system. Brain research. Molecular brain research 54(1): 152-160.

Dorn GW, 2nd (2007). The fuzzy logic of physiological cardiac hypertrophy. Hypertension 49(5): 962-970.

Dorsam RT, Gutkind JS (2007). G-protein-coupled receptors and cancer. Nature reviews. Cancer 7(2): 79-94.

Dowell SJ, Brown AJ (2009). Yeast assays for G protein-coupled receptors. Methods in molecular biology 552: 213-229.

Drake MT, Shenoy SK, Lefkowitz RJ (2006). Trafficking of G protein-coupled receptors. Circulation research 99(6): 570-582.

Du XJ, Fang L, Gao XM, Kiriazis H, Feng X, Hotchkin E, et al. (2004). Genetic enhancement of ventricular contractility protects against pressure-overload-induced cardiac dysfunction. Journal of molecular and cellular cardiology 37(5): 979-987.

Duarte JD, Cooper-DeHoff RM (2010). Mechanisms for blood pressure lowering and metabolic effects of thiazide and thiazide-like diuretics. Expert review of cardiovascular therapy 8(6): 793-802.

Dubay C, Vincent M, Samani NJ, Hilbert P, Kaiser MA, Beressi JP, et al. (1993). Genetic determinants of diastolic and pulse pressure map to different loci in Lyon hypertensive rats. Nature genetics 3(4): 354-357.

Edwards DR, Handsley MM, Pennington CJ (2008). The ADAM metalloproteinases. Molecular aspects of medicine 29(5): 258-289.

El-Mas MM, Afify EA, Mohy El-Din MM, Omar AG, Sharabi FM (2001). Testosterone facilitates the baroreceptor control of reflex bradycardia: role of cardiac sympathetic and parasympathetic components. Journal of cardiovascular pharmacology 38(5): 754-763.

Ersoy BA, Pardo L, Zhang S, Thompson DA, Millhauser G, Govaerts C, et al. (2012). Mechanism of N- terminal modulation of activity at the melanocortin-4 receptor GPCR. Nature chemical biology 8(8): 725-730.

201

Ferguson SS (2001). Evolving concepts in G protein-coupled receptor endocytosis: the role in receptor desensitization and signaling. Pharmacological reviews 53(1): 1-24.

Fermin DR, Barac A, Lee S, Polster SP, Hannenhalli S, Bergemann TL, et al. (2008). Sex and age dimorphism of myocardial gene expression in nonischemic human heart failure. Circulation. Cardiovascular genetics 1(2): 117-125.

Filipowicz W, Bhattacharyya SN, Sonenberg N (2008). Mechanisms of post-transcriptional regulation by microRNAs: are the answers in sight? Nature reviews. Genetics 9(2): 102-114.

Flynn JT, Pasko DA (2000). Calcium channel blockers: pharmacology and place in therapy of pediatric hypertension. Pediatric nephrology 15(3-4): 302-316.

Foster SR, Roura E, Molenaar P, Thomas WG (2015). G protein-coupled receptors in cardiac biology: old and new receptors. Biophysical reviews 7(1): 77-89.

Fredriksson R, Lagerstrom MC, Lundin LG, Schioth HB (2003). The G-protein-coupled receptors in the human genome form five main families. Phylogenetic analysis, paralogon groups, and fingerprints. Molecular pharmacology 63(6): 1256-1272.

Frey N, Olson EN (2003). Cardiac hypertrophy: the good, the bad, and the ugly. Annual review of physiology 65: 45-79.

Furman MI, Liu L, Benoit SE, Becker RC, Barnard MR, Michelson AD (1998). The cleaved peptide of the thrombin receptor is a strong platelet agonist. Proceedings of the National Academy of Sciences of the United States of America 95(6): 3082-3087.

Fyhrquist F, Saijonmaa O (2008). Renin-angiotensin system revisited. Journal of internal medicine 264(3): 224-236.

Gheorghiade M, Abraham WT, Albert NM, Greenberg BH, O'Connor CM, She L, et al. (2006). Systolic blood pressure at admission, clinical characteristics, and outcomes in patients hospitalized with acute heart failure. Jama 296(18): 2217-2226.

Giannone F, Malpeli G, Lisi V, Grasso S, Shukla P, Ramarli D, et al. (2010). The puzzling uniqueness of the heterotrimeric G15 protein and its potential beyond hematopoiesis. Journal of molecular endocrinology 44(5): 259-269.

202

Giddens MM, Wong JC, Schroeder JP, Farrow EG, Smith BM, Owino S, et al. (2017). GPR37L1 modulates seizure susceptibility: Evidence from mouse studies and analyses of a human GPR37L1 variant. Neurobiology of disease 106: 181-190.

Godecke A, Decking UK, Ding Z, Hirchenhain J, Bidmon HJ, Godecke S, et al. (1998). Coronary hemodynamics in endothelial NO synthase knockout mice. Circulation research 82(2): 186-194.

Goldblatt H, Lynch J, Hanzal RF, Summerville WW (1934). Studies on Experimental Hypertension : I. The Production of Persistent Elevation of Systolic Blood Pressure by Means of Renal Ischemia. The Journal of experimental medicine 59(3): 347-379.

Goth CK, Tuhkanen HE, Khan H, Lackman JJ, Wang S, Narimatsu Y, et al. (2017). Site-specific O- Glycosylation by Polypeptide N-Acetylgalactosaminyltransferase 2 (GalNAc-transferase T2) Co- regulates beta1-Adrenergic Receptor N-terminal Cleavage. The Journal of biological chemistry 292(11): 4714-4726.

Grantcharova E, Furkert J, Reusch HP, Krell HW, Papsdorf G, Beyermann M, et al. (2002). The extracellular N terminus of the endothelin B (ETB) receptor is cleaved by a metalloprotease in an agonist-dependent process. The Journal of biological chemistry 277(46): 43933-43941.

Grantcharova E, Reusch HP, Beyermann M, Rosenthal W, Oksche A (2006a). Endothelin A and endothelin B receptors differ in their ability to stimulate ERK1/2 activation. Experimental biology and medicine 231(6): 757-760.

Grantcharova E, Reusch HP, Grossmann S, Eichhorst J, Krell HW, Beyermann M, et al. (2006b). N- terminal proteolysis of the endothelin B receptor abolishes its ability to induce EGF receptor transactivation and contractile protein expression in vascular smooth muscle cells. Arteriosclerosis, thrombosis, and vascular biology 26(6): 1288-1296.

Greenway CV, Lautt WW (1986). Blood volume, the venous system, preload, and cardiac output. Canadian journal of physiology and pharmacology 64(4): 383-387.

Griendling KK, Murphy TJ, Alexander RW (1993). Molecular biology of the renin-angiotensin system. Circulation 87(6): 1816-1828.

Guyenet PG (2006). The sympathetic control of blood pressure. Nature reviews. Neuroscience 7(5): 335-346.

Guyton AC (1981). The relationship of cardiac output and arterial pressure control. Circulation 64(6): 1079-1088.

203

Guyton AC (1989). Dominant role of the kidneys and accessory role of whole-body autoregulation in the pathogenesis of hypertension. American journal of hypertension 2(7): 575-585.

Guyton AC, Coleman TG, Bower JD, Granger HJ (1970). Circulatory control in hypertension. Circulation research 27: Suppl 2:135+.

Guyton AC, Coleman TG, Cowley AV, Jr., Scheel KW, Manning RD, Jr., Norman RA, Jr. (1972a). Arterial pressure regulation. Overriding dominance of the kidneys in long-term regulation and in hypertension. The American journal of medicine 52(5): 584-594.

Guyton AC, Coleman TG, Granger HJ (1972b). Circulation: overall regulation. Annual review of physiology 34: 13-46.

Hadler-Olsen E, Fadnes B, Sylte I, Uhlin-Hansen L, Winberg JO (2011). Regulation of matrix metalloproteinase activity in health and disease. The FEBS journal 278(1): 28-45.

Hafko R, Villapol S, Nostramo R, Symes A, Sabban EL, Inagami T, et al. (2013). Commercially available angiotensin II At(2) receptor antibodies are nonspecific. PloS one 8(7): e69234.

Haggerty CM, Mattingly AC, Gong MC, Su W, Daugherty A, Fornwalt BK (2015). Telemetric Blood Pressure Assessment in Angiotensin II-Infused ApoE-/- Mice: 28 Day Natural History and Comparison to Tail-Cuff Measurements. PloS one 10(6): e0130723.

Hakalahti AE, Vierimaa MM, Lilja MK, Kumpula EP, Tuusa JT, Petaja-Repo UE (2010). Human beta1- adrenergic receptor is subject to constitutive and regulated N-terminal cleavage. The Journal of biological chemistry 285(37): 28850-28861.

Halsband U, Lange RK (2006). Motor learning in man: a review of functional and clinical studies. Journal of physiology, Paris 99(4-6): 414-424.

Hamilton WF, Remington JW (1948). Some factors in the regulation of the stroke volume. The American journal of physiology 153(2): 287-297.

Haskell-Luevano C, Monck EK (2001). Agouti-related protein functions as an inverse agonist at a constitutively active brain melanocortin-4 receptor. Regulatory peptides 99(1): 1-7.

Hay M, Xue B, Johnson AK (2014). Yes! Sex matters: sex, the brain and blood pressure. Current hypertension reports 16(8): 458.

204

Heritage AS, Stumpf WE, Sar M, Grant LD (1980). Brainstem catecholamine neurons are target sites for sex steroid hormones. Science 207(4437): 1377-1379.

Hermans E, Challiss RA (2001). Structural, signalling and regulatory properties of the group I metabotropic glutamate receptors: prototypic family C G-protein-coupled receptors. The Biochemical journal 359(Pt 3): 465-484.

Herrera M, Sparks MA, Alfonso-Pecchio AR, Harrison-Bernard LM, Coffman TM (2013). Lack of specificity of commercial antibodies leads to misidentification of angiotensin type 1 receptor protein. Hypertension 61(1): 253-258.

Heximer SP, Knutsen RH, Sun X, Kaltenbronn KM, Rhee MH, Peng N, et al. (2003). Hypertension and prolonged vasoconstrictor signaling in RGS2-deficient mice. The Journal of clinical investigation 111(8): 1259.

Horiuchi K, Zhou HM, Kelly K, Manova K, Blobel CP (2005). Evaluation of the contributions of ADAMs 9, 12, 15, 17, and 19 to heart development and ectodomain shedding of neuregulins beta1 and beta2. Developmental biology 283(2): 459-471.

Hua T, Vemuri K, Pu M, Qu L, Han GW, Wu Y, et al. (2016). Crystal Structure of the Human Cannabinoid Receptor CB1. Cell 167(3): 750-762 e714.

Huttlin EL, Jedrychowski MP, Elias JE, Goswami T, Rad R, Beausoleil SA, et al. (2010). A tissue- specific atlas of mouse protein phosphorylation and expression. Cell 143(7): 1174-1189.

Iadecola C, Davisson RL (2008). Hypertension and cerebrovascular dysfunction. Cell metabolism 7(6): 476-484.

Imai Y, Soda M, Inoue H, Hattori N, Mizuno Y, Takahashi R (2001). An unfolded putative transmembrane polypeptide, which can lead to endoplasmic reticulum stress, is a substrate of Parkin. Cell 105(7): 891-902.

Israili ZH, Hall WD (1992). Cough and angioneurotic edema associated with angiotensin-converting enzyme inhibitor therapy. A review of the literature and pathophysiology. Annals of internal medicine 117(3): 234-242.

Jacobson L, Sapolsky R (1991). The role of the hippocampus in feedback regulation of the hypothalamic-pituitary-adrenocortical axis. Endocrine reviews 12(2): 118-134.

205

Jahnsen JA, Uhlen S (2012). The predicted N-terminal signal sequence of the human alpha(2)C- adrenoceptor does not act as a functional cleavable signal peptide. European journal of pharmacology 684(1-3): 51-58.

James PA, Oparil S, Carter BL, Cushman WC, Dennison-Himmelfarb C, Handler J, et al. (2014). 2014 evidence-based guideline for the management of high blood pressure in adults: report from the panel members appointed to the Eighth Joint National Committee (JNC 8). Jama 311(5): 507-520.

Jean-Baptiste G, Yang Z, Khoury C, Gaudio S, Greenwood MT (2005). Peptide and non-peptide G- protein coupled receptors (GPCRs) in skeletal muscle. Peptides 26(8): 1528-1536.

Jenkins L, Brea J, Smith NJ, Hudson BD, Reilly G, Bryant NJ, et al. (2010). Identification of novel species-selective agonists of the G-protein-coupled receptor GPR35 that promote recruitment of beta-arrestin-2 and activate Galpha13. The Biochemical journal 432(3): 451-459.

Ji X, Naito Y, Weng H, Endo K, Ma X, Iwai N (2012). P2X7 deficiency attenuates hypertension and renal injury in deoxycorticosterone acetate-salt hypertension. American journal of physiology. Renal physiology 303(8): F1207-1215.

Jia Z, Aoyagi T, Yang T (2010). mPGES-1 protects against DOCA-salt hypertension via inhibition of oxidative stress or stimulation of NO/cGMP. Hypertension 55(2): 539-546.

Jin W, Reddy MA, Chen Z, Putta S, Lanting L, Kato M, et al. (2012). Small RNA sequencing reveals microRNAs that modulate angiotensin II effects in vascular smooth muscle cells. The Journal of biological chemistry 287(19): 15672-15683.

Jolly S, Bazargani N, Quiroga AC, Pringle NP, Attwell D, Richardson WD, et al. (2017). G protein- coupled receptor 37-like 1 modulates astrocyte glutamate transporters and neuronal NMDA receptors and is neuroprotective in ischemia. Glia.

Jones DW, Hall JE (2004). Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure and evidence from new hypertension trials. Hypertension 43(1): 1-3.

Julius S, Schork MA (1971). Borderline hypertension--a critical review. Journal of chronic diseases 23(10): 723-754.

Karatas A, Hegner B, de Windt LJ, Luft FC, Schubert C, Gross V, et al. (2008). Deoxycorticosterone acetate-salt mice exhibit blood pressure-independent sexual dimorphism. Hypertension 51(4): 1177-1183.

206

Kaschina E, Unger T (2003). Angiotensin AT1/AT2 receptors: regulation, signalling and function. Blood pressure 12(2): 70-88.

Kato N, Mashimo T, Nabika T, Cui ZH, Ikeda K, Yamori Y (2003). Genome-wide searches for blood pressure quantitative trait loci in the stroke-prone spontaneously hypertensive rat of a Japanese colony. Journal of hypertension 21(2): 295-303.

Katritch V, Cherezov V, Stevens RC (2013). Structure-function of the G protein-coupled receptor superfamily. Annual review of pharmacology and toxicology 53: 531-556.

Kawano H, Kimura-Kuroda J, Komuta Y, Yoshioka N, Li HP, Kawamura K, et al. (2012). Role of the lesion scar in the response to damage and repair of the central nervous system. Cell and tissue research 349(1): 169-180.

Kenakin T (1995). Agonist-receptor efficacy. I: Mechanisms of efficacy and receptor promiscuity. Trends in pharmacological sciences 16(6): 188-192.

Kim S, Iwao H (2000). Molecular and cellular mechanisms of angiotensin II-mediated cardiovascular and renal diseases. Pharmacological reviews 52(1): 11-34.

Kniazeff J, Prezeau L, Rondard P, Pin JP, Goudet C (2011). Dimers and beyond: The functional puzzles of class C GPCRs. Pharmacology & therapeutics 130(1): 9-25.

Kochl R, Alken M, Rutz C, Krause G, Oksche A, Rosenthal W, et al. (2002). The signal peptide of the G protein-coupled human endothelin B receptor is necessary for translocation of the N-terminal tail across the endoplasmic reticulum membrane. The Journal of biological chemistry 277(18): 16131- 16138.

Korner PI (1971). The central nervous system and physiological mechanisms of "optimal" cardiovascular control. The Australian journal of experimental biology and medical science 49(4): 319-343.

Korner PI (2007). Essential Hypertension and Its Causes. edn. Oxford University Press: New York.

Korner PI (2010). The phenotypic patterns of essential hypertension are the key to identifying "high blood pressure" genes. Physiological research 59(6): 841-857.

Kostenis E, Waelbroeck M, Milligan G (2005). Techniques: promiscuous Galpha proteins in basic research and drug discovery. Trends in pharmacological sciences 26(11): 595-602.

207

Kountz TS, Lee KS, Aggarwal-Howarth S, Curran E, Park JM, Harris DA, et al. (2016). Endogenous N- terminal Domain Cleavage Modulates alpha1D-Adrenergic Receptor Pharmacodynamics. The Journal of biological chemistry 291(35): 18210-18221.

Kovacs P, Voigt B, Kloting I (1997). Alleles of the spontaneously hypertensive rat decrease blood pressure at loci on chromosomes 4 and 13. Biochemical and biophysical research communications 238(2): 586-589.

Krishnan A, Schioth HB (2015). The role of G protein-coupled receptors in the early evolution of neurotransmission and the nervous system. The Journal of experimental biology 218(Pt 4): 562- 571.

Kroeze WK, Sassano MF, Huang XP, Lansu K, McCorvy JD, Giguere PM, et al. (2015). PRESTO-Tango as an open-source resource for interrogation of the druggable human GPCRome. Nature structural & molecular biology 22(5): 362-369.

Kudielka BM, Kirschbaum C (2005). Sex differences in HPA axis responses to stress: a review. Biological psychology 69(1): 113-132.

Kuroda J, Ago T, Matsushima S, Zhai P, Schneider MD, Sadoshima J (2010). NADPH oxidase 4 (Nox4) is a major source of oxidative stress in the failing heart. Proceedings of the National Academy of Sciences of the United States of America 107(35): 15565-15570.

Kurohara K, Komatsu K, Kurisaki T, Masuda A, Irie N, Asano M, et al. (2004). Essential roles of Meltrin beta (ADAM19) in heart development. Developmental biology 267(1): 14-28.

Kurtz A (2011). Renin release: sites, mechanisms, and control. Annual review of physiology 73: 377- 399.

Kvakan H, Kleinewietfeld M, Qadri F, Park JK, Fischer R, Schwarz I, et al. (2009). Regulatory T cells ameliorate angiotensin II-induced cardiac damage. Circulation 119(22): 2904-2912.

Lavoie JL, Sigmund CD (2003). Minireview: overview of the renin-angiotensin system--an endocrine and paracrine system. Endocrinology 144(6): 2179-2183.

Ledent C, Vaugeois JM, Schiffmann SN, Pedrazzini T, El Yacoubi M, Vanderhaeghen JJ, et al. (1997). Aggressiveness, hypoalgesia and high blood pressure in mice lacking the adenosine A2a receptor. Nature 388(6643): 674-678.

208

Lee DL, Sturgis LC, Labazi H, Osborne JB, Jr., Fleming C, Pollock JS, et al. (2006). Angiotensin II hypertension is attenuated in interleukin-6 knockout mice. American journal of physiology. Heart and circulatory physiology 290(3): H935-940.

Lee SM, Booe JM, Pioszak AA (2015). Structural insights into ligand recognition and selectivity for classes A, B, and C GPCRs. European journal of pharmacology 763(Pt B): 196-205.

Lefkowitz RJ, Cotecchia S, Samama P, Costa T (1993). Constitutive activity of receptors coupled to guanine nucleotide regulatory proteins. Trends in pharmacological sciences 14(8): 303-307.

Leng N, Gu G, Simerly RB, Spindel ER (1999). Molecular cloning and characterization of two putative G protein-coupled receptors which are highly expressed in the central nervous system. Brain research. Molecular brain research 69(1): 73-83.

Letcher RL, Chien S, Pickering TG, Sealey JE, Laragh JH (1981). Direct relationship between blood pressure and blood viscosity in normal and hypertensive subjects. Role of fibrinogen and concentration. The American journal of medicine 70(6): 1195-1202.

Lewis BP, Burge CB, Bartel DP (2005). Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell 120(1): 15-20.

Li M, Naqvi N, Yahiro E, Liu K, Powell PC, Bradley WE, et al. (2008). c-kit is required for cardiomyocyte terminal differentiation. Circulation research 102(6): 677-685.

Liao Y, Takashima S, Maeda N, Ouchi N, Komamura K, Shimomura I, et al. (2005). Exacerbation of heart failure in adiponectin-deficient mice due to impaired regulation of AMPK and glucose metabolism. Cardiovascular research 67(4): 705-713.

Liebscher I, Schon J, Petersen SC, Fischer L, Auerbach N, Demberg LM, et al. (2014). A tethered agonist within the ectodomain activates the adhesion G protein-coupled receptors GPR126 and GPR133. Cell reports 9(6): 2018-2026.

Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adair-Rohani H, et al. (2012). A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 380(9859): 2224-2260.

Linde T, Sandhagen B, Hagg A, Morlin C, Wikstrom B, Danielson BG (1993). Blood viscosity and peripheral vascular resistance in patients with untreated essential hypertension. Journal of hypertension 11(7): 731-736.

209

Lindholm LH, Carlberg B, Samuelsson O (2005). Should beta blockers remain first choice in the treatment of primary hypertension? A meta-analysis. Lancet 366(9496): 1545-1553.

Livak KJ, Schmittgen TD (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25(4): 402-408.

Lu D, Willard D, Patel IR, Kadwell S, Overton L, Kost T, et al. (1994). Agouti protein is an antagonist of the melanocyte-stimulating-hormone receptor. Nature 371(6500): 799-802.

Luke RG (1993). Essential hypertension: a renal disease? A review and update of the evidence. Hypertension 21(3): 380-390.

Lund-Johansen P (1980). Haemodynamics in essential hypertension. Clinical science 59 Suppl 6: 343s-354s.

Lv X, Liu J, Shi Q, Tan Q, Wu D, Skinner JJ, et al. (2016). In vitro expression and analysis of the 826 human G protein-coupled receptors. Protein & cell 7(5): 325-337.

Madeddu P, Milia AF, Salis MB, Gaspa L, Gross W, Lippoldt A, et al. (1998). Renovascular hypertension in bradykinin B2-receptor knockout mice. Hypertension 32(3): 503-509.

Mancia G, Fagard R, Narkiewicz K, Redon J, Zanchetti A, Bohm M, et al. (2013). 2013 ESH/ESC guidelines for the management of arterial hypertension: the Task Force for the Management of Arterial Hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). European heart journal 34(28): 2159-2219.

Mandillo S, Golini E, Marazziti D, Di Pietro C, Matteoni R, Tocchini-Valentini GP (2013). Mice lacking the Parkinson's related GPR37/PAEL receptor show non-motor behavioral phenotypes: age and gender effect. Genes, brain, and behavior 12(4): 465-477.

Mann DL, Kent RL, Parsons B, Cooper Gt (1992). Adrenergic effects on the biology of the adult mammalian cardiocyte. Circulation 85(2): 790-804.

Manolio TA, Olson J, Longstreth WT (2003). Hypertension and cognitive function: pathophysiologic effects of hypertension on the brain. Current hypertension reports 5(3): 255-261.

Marazziti D, Di Pietro C, Golini E, Mandillo S, La Sala G, Matteoni R, et al. (2013). Precocious cerebellum development and improved motor functions in mice lacking the astrocyte cilium-,

210 patched 1-associated Gpr37l1 receptor. Proceedings of the National Academy of Sciences of the United States of America 110(41): 16486-16491.

Marazziti D, Golini E, Gallo A, Lombardi MS, Matteoni R, Tocchini-Valentini GP (1997). Cloning of GPR37, a gene located on chromosome 7 encoding a putative G-protein-coupled peptide receptor, from a human frontal brain EST library. Genomics 45(1): 68-77.

Marinissen MJ, Gutkind JS (2001). G-protein-coupled receptors and signaling networks: emerging paradigms. Trends in pharmacological sciences 22(7): 368-376.

Mattila SO, Tuusa JT, Petaja-Repo UE (2016). The Parkinson's-disease-associated receptor GPR37 undergoes metalloproteinase-mediated N-terminal cleavage and ectodomain shedding. Journal of cell science 129(7): 1366-1377.

May LT, Leach K, Sexton PM, Christopoulos A (2007). Allosteric modulation of G protein-coupled receptors. Annual review of pharmacology and toxicology 47: 1-51.

Meneton P, Jeunemaitre X, de Wardener HE, MacGregor GA (2005). Links between dietary salt intake, renal salt handling, blood pressure, and cardiovascular diseases. Physiological reviews 85(2): 679-715.

Mennuni S, Rubattu S, Pierelli G, Tocci G, Fofi C, Volpe M (2014). Hypertension and kidneys: unraveling complex molecular mechanisms underlying hypertensive renal damage. Journal of human hypertension 28(2): 74-79.

Menuet C, Le S, Dempsey B, Connelly AA, Kamar JL, Jancovski N, et al. (2017). Excessive Respiratory Modulation of Blood Pressure Triggers Hypertension. Cell metabolism 25(3): 739-748.

Meyer RC, Giddens MM, Schaefer SA, Hall RA (2013). GPR37 and GPR37L1 are receptors for the neuroprotective and glioprotective factors prosaptide and prosaposin. Proceedings of the National Academy of Sciences of the United States of America 110(23): 9529-9534.

Michael SK, Surks HK, Wang Y, Zhu Y, Blanton R, Jamnongjit M, et al. (2008). High blood pressure arising from a defect in vascular function. Proceedings of the National Academy of Sciences of the United States of America 105(18): 6702-6707.

Michineau S, Alhenc-Gelas F, Rajerison RM (2006). Human bradykinin B2 receptor sialylation and N-glycosylation participate with disulfide bonding in surface receptor dimerization. Biochemistry 45(8): 2699-2707.

211

Milligan G (2003). Constitutive activity and inverse agonists of G protein-coupled receptors: a current perspective. Molecular pharmacology 64(6): 1271-1276.

Milligan G, Kostenis E (2006). Heterotrimeric G-proteins: a short history. British journal of pharmacology 147 Suppl 1: S46-55.

Min KD, Asakura M, Liao Y, Nakamaru K, Okazaki H, Takahashi T, et al. (2010). Identification of genes related to heart failure using global gene expression profiling of human failing myocardium. Biochemical and biophysical research communications 393(1): 55-60.

Muller A, Berkmann JC, Scheerer P, Biebermann H, Kleinau G (2016). Insights into Basal Signaling Regulation, Oligomerization, and Structural Organization of the Human G-Protein Coupled Receptor 83. PloS one 11(12): e0168260.

Nadruz W (2015). Myocardial remodeling in hypertension. Journal of human hypertension 29(1): 1- 6.

Nath KA, d'Uscio LV, Juncos JP, Croatt AJ, Manriquez MC, Pittock ST, et al. (2007). An analysis of the DOCA-salt model of hypertension in HO-1-/- mice and the Gunn rat. American journal of physiology. Heart and circulatory physiology 293(1): H333-342.

Neumann JM, Couvineau A, Murail S, Lacapere JJ, Jamin N, Laburthe M (2008). Class-B GPCR activation: is ligand helix-capping the key? Trends in biochemical sciences 33(7): 314-319.

Neves SR, Ram PT, Iyengar R (2002). G protein pathways. Science 296(5573): 1636-1639.

Ngo T, Coleman JL, Smith NJ (2015). Using constitutive activity to define appropriate high- throughput screening assays for orphan g protein-coupled receptors. Methods in molecular biology 1272: 91-106.

Ngo T, Ilatovskiy AV, Stewart AG, Coleman JL, McRobb FM, Riek RP, et al. (2017). Orphan receptor ligand discovery by pickpocketing pharmacological neighbors. Nature chemical biology 13(2): 235- 242.

Ngo T, Kufareva I, Coleman J, Graham RM, Abagyan R, Smith NJ (2016). Identifying ligands at orphan GPCRs: current status using structure-based approaches. British journal of pharmacology 173(20): 2934-2951.

Nisimaru N (2004). Cardiovascular modules in the cerebellum. The Japanese journal of physiology 54(5): 431-448.

212

Norton GR, Woodiwiss AJ, Gaasch WH, Mela T, Chung ES, Aurigemma GP, et al. (2002). Heart failure in pressure overload hypertrophy. The relative roles of ventricular remodeling and myocardial dysfunction. Journal of the American College of Cardiology 39(4): 664-671.

O'Callaghan EL, Bassi JK, Porrello ER, Delbridge LM, Thomas WG, Allen AM (2011). Regulation of angiotensinogen by angiotensin II in mouse primary astrocyte cultures. Journal of neurochemistry 119(1): 18-26.

Obst M, Gross V, Luft FC (2004). Systemic hemodynamics in non-anesthetized L-NAME- and DOCA- salt-treated mice. Journal of hypertension 22(10): 1889-1894.

Oldham WM, Hamm HE (2008). Heterotrimeric G protein activation by G-protein-coupled receptors. Nature reviews. Molecular cell biology 9(1): 60-71.

Omasits U, Ahrens CH, Muller S, Wollscheid B (2014). Protter: interactive protein feature visualization and integration with experimental proteomic data. Bioinformatics 30(6): 884-886.

Ono M, Murakami T, Kudo A, Isshiki M, Sawada H, Segawa A (2001). Quantitative comparison of anti-fading mounting media for confocal laser scanning microscopy. The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society 49(3): 305-312.

Oparil S, Zaman MA, Calhoun DA (2003). Pathogenesis of hypertension. Annals of internal medicine 139(9): 761-776.

Osborn JW, Fink GD, Kuroki MT (2011). Neural mechanisms of angiotensin II-salt hypertension: implications for therapies targeting neural control of the splanchnic circulation. Current hypertension reports 13(3): 221-228.

Ossovskaya VS, Bunnett NW (2004). Protease-activated receptors: contribution to physiology and disease. Physiological reviews 84(2): 579-621.

Pacher P, Nagayama T, Mukhopadhyay P, Batkai S, Kass DA (2008). Measurement of cardiac function using pressure-volume conductance catheter technique in mice and rats. Nature protocols 3(9): 1422-1434.

Palmer AJ, Bulpitt CJ, Fletcher AE, Beevers DG, Coles EC, Ledingham JG, et al. (1992). Relation between blood pressure and stroke mortality. Hypertension 20(5): 601-605.

213

Parati G, Esler M (2012). The human sympathetic nervous system: its relevance in hypertension and heart failure. European heart journal 33(9): 1058-1066.

Paullin J (1926). Ultimate Results Of Essential Hypertension. Journal of the American Medical Association 87(12): 925-928.

Peach MJ (1977). Renin-angiotensin system: biochemistry and mechanisms of action. Physiological reviews 57(2): 313-370.

Porrello ER, Delbridge LM, Thomas WG (2009). The angiotensin II type 2 (AT2) receptor: an enigmatic seven transmembrane receptor. Frontiers in bioscience 14: 958-972.

Prisant LM (2005). Hypertensive heart disease. Journal of clinical hypertension 7(4): 231-238.

Qin L, Kufareva I, Holden LG, Wang C, Zheng Y, Zhao C, et al. (2015). Structural biology. Crystal structure of the chemokine receptor CXCR4 in complex with a viral chemokine. Science 347(6226): 1117-1122.

Rands E, Candelore MR, Cheung AH, Hill WS, Strader CD, Dixon RA (1990). Mutational analysis of beta-adrenergic receptor glycosylation. The Journal of biological chemistry 265(18): 10759-10764.

Rasmussen UB, Vouret-Craviari V, Jallat S, Schlesinger Y, Pages G, Pavirani A, et al. (1991). cDNA cloning and expression of a hamster alpha-thrombin receptor coupled to Ca2+ mobilization. FEBS letters 288(1-2): 123-128.

Ratnikov BI, Cieplak P, Gramatikoff K, Pierce J, Eroshkin A, Igarashi Y, et al. (2014). Basis for substrate recognition and distinction by matrix metalloproteinases. Proceedings of the National Academy of Sciences of the United States of America 111(40): E4148-4155.

Reddy AS, O'Brien D, Pisat N, Weichselbaum CT, Sakers K, Lisci M, et al. (2017). A Comprehensive Analysis of Cell Type-Specific Nuclear RNA From Neurons and Glia of the Brain. Biological psychiatry 81(3): 252-264.

Reimann F, Tolhurst G, Gribble FM (2012). G-protein-coupled receptors in intestinal chemosensation. Cell metabolism 15(4): 421-431.

Rezgaoui M, Susens U, Ignatov A, Gelderblom M, Glassmeier G, Franke I, et al. (2006). The neuropeptide head activator is a high-affinity ligand for the orphan G-protein-coupled receptor GPR37. Journal of cell science 119(Pt 3): 542-549.

214

Rimoldi SF, Messerli FH, Bangalore S, Scherrer U (2015). Resistant hypertension: what the cardiologist needs to know. European heart journal 36(40): 2686-2695.

Rimoldi SF, Scherrer U, Messerli FH (2014). Secondary arterial hypertension: when, who, and how to screen? European heart journal 35(19): 1245-1254.

Robinson MD, McCarthy DJ, Smyth GK (2010). edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26(1): 139-140.

Rockman HA, Ross RS, Harris AN, Knowlton KU, Steinhelper ME, Field LJ, et al. (1991). Segregation of atrial-specific and inducible expression of an atrial natriuretic factor transgene in an in vivo murine model of cardiac hypertrophy. Proceedings of the National Academy of Sciences of the United States of America 88(18): 8277-8281.

Romanic AM, White RF, Arleth AJ, Ohlstein EH, Barone FC (1998). Matrix metalloproteinase expression increases after cerebral focal ischemia in rats: inhibition of matrix metalloproteinase-9 reduces infarct size. Stroke 29(5): 1020-1030.

Rosenbaum DM, Rasmussen SG, Kobilka BK (2009). The structure and function of G-protein- coupled receptors. Nature 459(7245): 356-363.

Rowlands DB, Ireland MA, Glover DR, McLeay RA, Stallard TJ, Littler WA (1981). The relationship between ambulatory blood pressure and echocardiographically assessed left ventricular hypertrophy. Clinical science 61 Suppl 7: 101s-103s.

Roy S, Perron B, Gallo-Payet N (2010). Role of asparagine-linked glycosylation in cell surface expression and function of the human adrenocorticotropin receptor (melanocortin 2 receptor) in 293/FRT cells. Endocrinology 151(2): 660-670.

Ruilope LM, Rosei EA, Bakris GL, Mancia G, Poulter NR, Taddei S, et al. (2005). Angiotensin receptor blockers: therapeutic targets and cardiovascular protection. Blood pressure 14(4): 196-209.

Ruilope LM, Schmieder RE (2008). Left ventricular hypertrophy and clinical outcomes in hypertensive patients. American journal of hypertension 21(5): 500-508.

Rung J, Brazma A (2013). Reuse of public genome-wide gene expression data. Nature reviews. Genetics 14(2): 89-99.

215

Rutz C, Klein W, Schulein R (2015). N-Terminal Signal Peptides of G Protein-Coupled Receptors: Significance for Receptor Biosynthesis, Trafficking, and Signal Transduction. Progress in molecular biology and translational science 132: 267-287.

Saito Y, Tetsuka M, Yue L, Kawamura Y, Maruyama K (2003). Functional role of N-linked glycosylation on the rat melanin-concentrating hormone receptor 1. FEBS letters 533(1-3): 29-34.

Sakata Y, Hoit BD, Liggett SB, Walsh RA, Dorn GW, 2nd (1998). Decompensation of pressure- overload hypertrophy in G alpha q-overexpressing mice. Circulation 97(15): 1488-1495.

Saleh MC, Connell BJ, Saleh TM (2000). Autonomic and cardiovascular reflex responses to central estrogen injection in ovariectomized female rats. Brain research 879(1-2): 105-114.

Samama P, Cotecchia S, Costa T, Lefkowitz RJ (1993). A mutation-induced activated state of the beta 2-adrenergic receptor. Extending the ternary complex model. The Journal of biological chemistry 268(7): 4625-4636.

Samani NJ, Gauguier D, Vincent M, Kaiser MA, Bihoreau MT, Lodwick D, et al. (1996). Analysis of quantitative trait loci for blood pressure on rat chromosomes 2 and 13. Age-related differences in effect. Hypertension 28(6): 1118-1122.

Santos R, Ursu O, Gaulton A, Bento AP, Donadi RS, Bologa CG, et al. (2017). A comprehensive map of molecular drug targets. Nature reviews. Drug discovery 16(1): 19-34.

Schiffrin EL (2001). Role of endothelin-1 in hypertension and vascular disease. American journal of hypertension 14(6 Pt 2): 83S-89S.

Schildge S, Bohrer C, Beck K, Schachtrup C (2013). Isolation and culture of mouse cortical astrocytes. Journal of visualized experiments : JoVE(71).

Schioth HB, Fredriksson R (2005). The GRAFS classification system of G-protein coupled receptors in comparative perspective. General and comparative endocrinology 142(1-2): 94-101.

Schmahmann JD, Sherman JC (1998). The cerebellar cognitive affective syndrome. Brain : a journal of neurology 121 ( Pt 4): 561-579.

Schmidt J, Smith NJ, Christiansen E, Tikhonova IG, Grundmann M, Hudson BD, et al. (2011). Selective orthosteric free fatty acid receptor 2 (FFA2) agonists: identification of the structural and chemical requirements for selective activation of FFA2 versus FFA3. The Journal of biological chemistry 286(12): 10628-10640.

216

Schulein R, Westendorf C, Krause G, Rosenthal W (2012). Functional significance of cleavable signal peptides of G protein-coupled receptors. European journal of cell biology 91(4): 294-299.

Schutzer SE, Liu T, Natelson BH, Angel TE, Schepmoes AA, Purvine SO, et al. (2010). Establishing the proteome of normal human cerebrospinal fluid. PloS one 5(6): e10980.

Scott DJ, Layfield S, Yan Y, Sudo S, Hsueh AJ, Tregear GW, et al. (2006). Characterization of novel splice variants of LGR7 and LGR8 reveals that receptor signaling is mediated by their unique low density lipoprotein class A modules. The Journal of biological chemistry 281(46): 34942-34954.

Severs WB, Summy-Long J, Taylor JS, Connor JD (1970). A central effect of angiotensin: release of pituitary pressor material. The Journal of pharmacology and experimental therapeutics 174(1): 27- 34.

Share L (1988). Role of vasopressin in cardiovascular regulation. Physiological reviews 68(4): 1248- 1284.

Shi L, Zhang Z, Su B (2016). Sex Biased Gene Expression Profiling of Human Brains at Major Developmental Stages. Scientific reports 6: 21181.

Sidhu TS, French SL, Hamilton JR (2014). Differential signaling by protease-activated receptors: implications for therapeutic targeting. International journal of molecular sciences 15(4): 6169-6183.

Skarnes WC, Rosen B, West AP, Koutsourakis M, Bushell W, Iyer V, et al. (2011). A conditional knockout resource for the genome-wide study of mouse gene function. Nature 474(7351): 337-342.

Small EM, Olson EN (2011). Pervasive roles of microRNAs in cardiovascular biology. Nature 469(7330): 336-342.

Smith NJ (2015). Drug Discovery Opportunities at the Endothelin B Receptor-Related Orphan G Protein-Coupled Receptors, GPR37 and GPR37L1. Frontiers in pharmacology 6: 275.

Smith NJ, Ward RJ, Stoddart LA, Hudson BD, Kostenis E, Ulven T, et al. (2011). Extracellular loop 2 of the free fatty acid receptor 2 mediates allosterism of a phenylacetamide ago-allosteric modulator. Molecular pharmacology 80(1): 163-173.

Soetanto R, Hynes CJ, Patel HR, Humphreys DT, Evers M, Duan G, et al. (2016). Role of miRNAs and alternative mRNA 3'-end cleavage and polyadenylation of their mRNA targets in cardiomyocyte hypertrophy. Biochimica et biophysica acta 1859(5): 744-756.

217

Southern C, Cook JM, Neetoo-Isseljee Z, Taylor DL, Kettleborough CA, Merritt A, et al. (2013). Screening beta-arrestin recruitment for the identification of natural ligands for orphan G-protein- coupled receptors. Journal of biomolecular screening 18(5): 599-609.

Sporer B, Paul R, Koedel U, Grimm R, Wick M, Goebel FD, et al. (1998). Presence of matrix metalloproteinase-9 activity in the cerebrospinal fluid of human immunodeficiency virus-infected patients. The Journal of infectious diseases 178(3): 854-857.

Srinivasan S, Lubrano-Berthelier C, Govaerts C, Picard F, Santiago P, Conklin BR, et al. (2004). Constitutive activity of the melanocortin-4 receptor is maintained by its N-terminal domain and plays a role in energy homeostasis in humans. The Journal of clinical investigation 114(8): 1158- 1164.

Sriram K, Insel PA (2018). GPCRs as targets for approved drugs: How many targets and how many drugs? Molecular pharmacology.

Stanley CP, O'Sullivan SE (2014). Cyclooxygenase metabolism mediates vasorelaxation to 2- arachidonoylglycerol (2-AG) in human mesenteric arteries. Pharmacological research 81: 74-82.

Stark M, Danielsson O, Griffiths WJ, Jornvall H, Johansson J (2001). Peptide repertoire of human cerebrospinal fluid: novel proteolytic fragments of neuroendocrine proteins. Journal of chromatography. B, Biomedical sciences and applications 754(2): 357-367.

Steckelings UM, Kaschina E, Unger T (2005). The AT2 receptor--a matter of love and hate. Peptides 26(8): 1401-1409.

Stoddart LA, Smith NJ, Jenkins L, Brown AJ, Milligan G (2008). Conserved polar residues in transmembrane domains V, VI, and VII of free fatty acid receptor 2 and free fatty acid receptor 3 are required for the binding and function of short chain fatty acids. The Journal of biological chemistry 283(47): 32913-32924.

Stoll M, Cowley AW, Jr., Tonellato PJ, Greene AS, Kaldunski ML, Roman RJ, et al. (2001). A genomic- systems biology map for cardiovascular function. Science 294(5547): 1723-1726.

Stoppini L, Buchs PA, Muller D (1991). A simple method for organotypic cultures of nervous tissue. Journal of neuroscience methods 37(2): 173-182.

Stornetta RL, Hawelu-Johnson CL, Guyenet PG, Lynch KR (1988). Astrocytes synthesize angiotensinogen in brain. Science 242(4884): 1444-1446.

218

Sun Y, Carretero OA, Xu J, Rhaleb NE, Yang JJ, Pagano PJ, et al. (2009). Deletion of inducible nitric oxide synthase provides cardioprotection in mice with 2-kidney, 1-clip hypertension. Hypertension 53(1): 49-56.

Taegtmeyer H, Sen S, Vela D (2010). Return to the fetal gene program: a suggested metabolic link to gene expression in the heart. Annals of the New York Academy of Sciences 1188: 191-198.

Tang XL, Wang Y, Li DL, Luo J, Liu MY (2012). Orphan G protein-coupled receptors (GPCRs): biological functions and potential drug targets. Acta pharmacologica Sinica 33(3): 363-371.

Tejada T, Tan L, Torres RA, Calvert JW, Lambert JP, Zaidi M, et al. (2016). IGF-1 degradation by mouse mast cell protease 4 promotes cell death and adverse cardiac remodeling days after a myocardial infarction. Proceedings of the National Academy of Sciences of the United States of America 113(25): 6949-6954.

The UniProt C (2017). UniProt: the universal protein knowledgebase. Nucleic acids research 45(D1): D158-D169.

Thomsen ARB, Plouffe B, Cahill TJ, 3rd, Shukla AK, Tarrasch JT, Dosey AM, et al. (2016). GPCR-G Protein-beta-Arrestin Super-Complex Mediates Sustained G Protein Signaling. Cell 166(4): 907-919.

Thomsen W, Frazer J, Unett D (2005). Functional assays for screening GPCR targets. Current opinion in biotechnology 16(6): 655-665.

Timpl P, Spanagel R, Sillaber I, Kresse A, Reul JM, Stalla GK, et al. (1998). Impaired stress response and reduced anxiety in mice lacking a functional corticotropin-releasing hormone receptor 1. Nature genetics 19(2): 162-166.

Tipnis SR, Hooper NM, Hyde R, Karran E, Christie G, Turner AJ (2000). A human homolog of angiotensin-converting enzyme. Cloning and functional expression as a captopril-insensitive carboxypeptidase. The Journal of biological chemistry 275(43): 33238-33243.

Tocchi A, Parks WC (2013). Functional interactions between matrix metalloproteinases and glycosaminoglycans. The FEBS journal 280(10): 2332-2341.

Toyooka M, Tujii T, Takeda S (2009). The N-terminal domain of GPR61, an orphan G-protein- coupled receptor, is essential for its constitutive activity. Journal of neuroscience research 87(6): 1329-1333.

219

Travers JG, Kamal FA, Robbins J, Yutzey KE, Blaxall BC (2016). Cardiac Fibrosis: The Fibroblast Awakens. Circulation research 118(6): 1021-1040.

Tyagi N, Moshal KS, Lominadze D, Ovechkin AV, Tyagi SC (2005). Homocysteine-dependent cardiac remodeling and endothelial-myocyte coupling in a 2 kidney, 1 clip Goldblatt hypertension mouse model. Canadian journal of physiology and pharmacology 83(7): 583-594.

Uddin MS, Hauser M, Naider F, Becker JM (2016). The N-terminus of the yeast G protein-coupled receptor Ste2p plays critical roles in surface expression, signaling, and negative regulation. Biochimica et biophysica acta 1858(4): 715-724.

Uhlen M, Fagerberg L, Hallstrom BM, Lindskog C, Oksvold P, Mardinoglu A, et al. (2015). Proteomics. Tissue-based map of the human proteome. Science 347(6220): 1260419.

Valdenaire O, Giller T, Breu V, Ardati A, Schweizer A, Richards JG (1998). A new family of orphan G protein-coupled receptors predominantly expressed in the brain. FEBS letters 424(3): 193-196.

Vassilatis DK, Hohmann JG, Zeng H, Li F, Ranchalis JE, Mortrud MT, et al. (2003). The G protein- coupled receptor repertoires of human and mouse. Proceedings of the National Academy of Sciences of the United States of America 100(8): 4903-4908.

Vickers C, Hales P, Kaushik V, Dick L, Gavin J, Tang J, et al. (2002). Hydrolysis of biological peptides by human angiotensin-converting enzyme-related carboxypeptidase. The Journal of biological chemistry 277(17): 14838-14843.

Vilardaga JP, Romero G, Friedman PA, Gardella TJ (2011). Molecular basis of parathyroid hormone receptor signaling and trafficking: a family B GPCR paradigm. Cellular and molecular life sciences : CMLS 68(1): 1-13.

Villari B, Campbell SE, Schneider J, Vassalli G, Chiariello M, Hess OM (1995). Sex-dependent differences in left ventricular function and structure in chronic pressure overload. European heart journal 16(10): 1410-1419.

Vlaeminck-Guillem V, Ho SC, Rodien P, Vassart G, Costagliola S (2002). Activation of the cAMP pathway by the TSH receptor involves switching of the ectodomain from a tethered inverse agonist to an agonist. Molecular endocrinology 16(4): 736-746.

Vu TK, Hung DT, Wheaton VI, Coughlin SR (1991). Molecular cloning of a functional thrombin receptor reveals a novel proteolytic mechanism of receptor activation. Cell 64(6): 1057-1068.

220

Wallin E, von Heijne G (1995). Properties of N-terminal tails in G-protein coupled receptors: a statistical study. Protein engineering 8(7): 693-698.

Wang Y, Wang DH (2009). Aggravated renal inflammatory responses in TRPV1 gene knockout mice subjected to DOCA-salt hypertension. American journal of physiology. Renal physiology 297(6): F1550-1559.

Wei P, Zhao YG, Zhuang L, Ruben S, Sang QX (2001). Expression and enzymatic activity of human disintegrin and metalloproteinase ADAM19/meltrin beta. Biochemical and biophysical research communications 280(3): 744-755.

Whitworth JA, Schyvens CG, Zhang Y, Mangos GJ, Kelly JJ (2001). Glucocorticoid-induced hypertension: from mouse to man. Clinical and experimental pharmacology & physiology 28(12): 993-996.

Whitworth JA, Williamson PM, Mangos G, Kelly JJ (2005). Cardiovascular consequences of cortisol excess. Vascular health and risk management 1(4): 291-299.

WHO (2013). A global brief on hypertension: silent killer, global public health crisis.

Wiesel P, Mazzolai L, Nussberger J, Pedrazzini T (1997). Two-kidney, one clip and one-kidney, one clip hypertension in mice. Hypertension 29(4): 1025-1030.

Wilkie TM, Scherle PA, Strathmann MP, Slepak VZ, Simon MI (1991). Characterization of G-protein alpha subunits in the Gq class: expression in murine tissues and in stromal and hematopoietic cell lines. Proceedings of the National Academy of Sciences of the United States of America 88(22): 10049-10053.

Wise A, Jupe SC, Rees S (2004). The identification of ligands at orphan G-protein coupled receptors. Annual review of pharmacology and toxicology 44: 43-66.

Wootten D, Christopoulos A, Sexton PM (2013). Emerging paradigms in GPCR allostery: implications for drug discovery. Nature reviews. Drug discovery 12(8): 630-644.

Xu F, Frazier DT (2002). Role of the cerebellar deep nuclei in respiratory modulation. Cerebellum 1(1): 35-40.

Xue B, Pamidimukkala J, Hay M (2005). Sex differences in the development of angiotensin II- induced hypertension in conscious mice. American journal of physiology. Heart and circulatory physiology 288(5): H2177-2184.

221

Yamada K, Fukaya M, Shibata T, Kurihara H, Tanaka K, Inoue Y, et al. (2000). Dynamic transformation of Bergmann glial fibers proceeds in correlation with dendritic outgrowth and synapse formation of cerebellar Purkinje cells. The Journal of comparative neurology 418(1): 106- 120.

Yanai I, Korbel JO, Boue S, McWeeney SK, Bork P, Lercher MJ (2006). Similar gene expression profiles do not imply similar tissue functions. Trends in genetics : TIG 22(3): 132-138.

Yang Y, Candelario-Jalil E, Thompson JF, Cuadrado E, Estrada EY, Rosell A, et al. (2010). Increased intranuclear matrix metalloproteinase activity in neurons interferes with oxidative DNA repair in focal cerebral ischemia. Journal of neurochemistry 112(1): 134-149.

Yemane H, Busauskas M, Burris SK, Knuepfer MM (2010). Neurohumoral mechanisms in deoxycorticosterone acetate (DOCA)-salt hypertension in rats. Experimental physiology 95(1): 51- 55.

Yong VW, Krekoski CA, Forsyth PA, Bell R, Edwards DR (1998). Matrix metalloproteinases and diseases of the CNS. Trends in neurosciences 21(2): 75-80.

Zeng Z, Su K, Kyaw H, Li Y (1997). A novel endothelin receptor type-B-like gene enriched in the brain. Biochemical and biophysical research communications 233(2): 559-567.

Zhang H, Han GW, Batyuk A, Ishchenko A, White KL, Patel N, et al. (2017). Structural basis for selectivity and diversity in angiotensin II receptors. Nature 544(7650): 327-332.

Zhang K, McQuibban GA, Silva C, Butler GS, Johnston JB, Holden J, et al. (2003). HIV-induced metalloproteinase processing of the chemokine stromal cell derived factor-1 causes neurodegeneration. Nature neuroscience 6(10): 1064-1071.

Zhang M, Tong KP, Fremont V, Chen J, Narayan P, Puett D, et al. (2000). The extracellular domain suppresses constitutive activity of the transmembrane domain of the human TSH receptor: implications for hormone-receptor interaction and antagonist design. Endocrinology 141(9): 3514- 3517.

Zhang QY, Dene H, Deng AY, Garrett MR, Jacob HJ, Rapp JP (1997). Interval mapping and congenic strains for a blood pressure QTL on rat chromosome 13. Mammalian genome : official journal of the International Mammalian Genome Society 8(9): 636-641.

222

Zhao X, Southwick K, Cardasis HL, Du Y, Lassman ME, Xie D, et al. (2010). Peptidomic profiling of human cerebrospinal fluid identifies YPRPIHPA as a novel substrate for prolylcarboxypeptidase. Proteomics 10(15): 2882-2886.

Zlotnik A, Yoshie O (2000). Chemokines: a new classification system and their role in immunity. Immunity 12(2): 121-127.

223