Development of Avocado Derived Polyhydroxylated Fatty Alcohols As Metabolic Modulators

Development of avocado derived polyhydroxylated fatty alcohols as metabolic modulators

Nawaz Ahmed

A Thesis presented to The University of Guelph

In partial fulfilment of requirements for the degree of Doctor of Philosophy in Food Science

Guelph, Ontario, Canada

ABSTRACT

DEVELOPMENT OF AVOCADO DERIVED POLYHYDROXYLATED FATTY ALCOHOLS AS METABOLIC MODULATORS

Nawaz Ahmed Advisor: University of Guelph, 2019 Professor P.A. Spagnuolo

Obesity and associated disorders like cardiovascular disease and Type 2 Diabetes share a common pathophysiological process in which adipose tissue dysfunction causes fatty acid spillover and accumulation in non-adipose tissues like skeletal muscle, heart, pancreas and liver.

Recent evidence strongly points towards the inhibition of fatty acid oxidation (FAO) in metabolically active tissue as an effective treatment strategy for obesity and associated disorders, however, safe and efficacious FAO inhibitors are currently not clinically available. Avocado derived polyhydroxylated fatty alcohols (PFAs), avocadene and avocadyne, have been previously shown to selectively induce cell death in leukemia stem cells by inhibiting FAO. This thesis investigates the potential of avocado PFAs to be developed as FAO inhibitors or metabolic modulators for the treatment of obesity and insulin resistance.

First, a validated analytical method for the quantitation of avocadene and avocadyne in an array of biological matrices was developed. Second, delivery systems of avocadene and avocadyne were tested in two in vitro digestion models (i.e. static and dynamic (i.e. TIM-1)) and a pilot pharmacokinetic in vivo study. The effects of a 1:1 mixture of pure avocadene and avocadyne

(AVO) were then evaluated in treatment and prevention mouse models of diet-induced obesity

(DIO). In the treatment DIO study, five-week oral administration of AVO (twice weekly) slowed

weight gain, improved whole body glucose tolerance, reversed insulin resistance, and increased postprandial skeletal muscle glucose utilization. AVO’s mechanism of action was further explored in cell culture models of lipotoxicity where AVO inhibited excessive FAO and restored glucose oxidation which restored insulin action in C2C12 myotubes (mouse skeletal muscle cells) and increased glucose stimulated insulin secretion (GSIS) in INS-1 (832/13) cells (rat pancreatic β-cell line), respectively. Finally, a nutritional supplement containing 50 or 200 mg AVO was manufactured for evaluation in a Phase I, double-blind, placebo-controlled human clinical trial.

AVO consumption (once daily for 60 days) showed no dose limiting toxicities in healthy human participants.

In summary, the use of a safe and potent FAO inhibitor like AVO was shown to ameliorate obesity-associated pathologies. This thesis lays the groundwork for future large scale preclinical and clinical studies for AVO.

ACKNOWLEDGMENTS

First and foremost, I would like to thank my supervisor, Dr. Paul A. Spagnuolo, for round- the-clock guidance and support throughout my PhD. The success of this project over the past three years is a testament of your vision and exemplary mentorship. I will never forget the serendipity and good fortune with which I was chosen to work in your lab. Your patience, enthusiasm, motivation, humor, and immense knowledge never failed to raise my spirits through all ups and downs of my PhD. It has been a privilege to have had you as a mentor and it is with great joy I look forward to a life-long friendship. I would like to thank my advisory committee members, Dr. Doug Goff and Dr. Amanda Wright, for their feedback and input throughout my PhD. Thank you for letting me barge into your labs and offices unannounced to get your feedback or use your equipment. I am very thankful to all our collaborators at the University of Waterloo without whom the analytical method development presented in this thesis would not have been possible. I’m very grateful for the mentorship of Dr. Richard Smith; thank you for sharing your invaluable knowledge of mass spectrometry and for keeping me accountable through countless hours of LC-MS. I also want to thank Dr. Ken Stark and Dr. Juan Aristizabal-Henao for their vital contributions towards this project. Juan, I feel very fortunate to have worked with you and will always cherish our friendship. I am obliged to Dr. Michael Rogers and Dr. Alejandro Marangoni for always asking the hard questions which incented me to add depth to my overall scientific approach. I will always be in awe of your accomplishments and am inspired by the discipline and rigor with which you address challenges. I would also like to thank our collaborators Dr. Tariq Akhtar and Kevin Rea for their contributions towards key cell culture experiments. I am very grateful to the team at Advanced Orthomolecular Research for their time and assistance with the avocado clinical trial. My special thanks to Dr. Traj Nibber, Dr. Pamela Ovadje, and Mohammad Islam for allowing me to spend time at AOR and supporting me with all aspects of the clinical trial. I also gratefully acknowledge the funding sources and scholarships that made my PhD work possible. I sincerely thank my labmates and fellow graduate students for ensuring my time at UoG was a memorable one. Alessia and Matt, thank you for your invaluable contributions towards my

v research but most of all thank you for your friendship. I will remember with great fondness how we started and just how much we achieved together; my appreciation for you is endless. I thank Dr. Preethi Jayanth for her contributions to my research, being a fantastic lab manager, and most importantly for enduring my quirks and pranks. I also thank our undergraduate students Michael Buraczynski and Sarah Walker for directly assisting with various aspects of my PhD research. Saving the best for last, I want to acknowledge the significant contributions of my parents. Thank you for your unconditional love, support and patience. I am forever indebted to you for all the hardships you have faced and sacrifices you have made to support my career choices. I dedicate this thesis to you.

TABLE OF CONTENTS ABSTRACT ...... ii ACKNOWLEDGMENTS ...... iv TABLE OF CONTENTS ...... vi LIST OF TABLES ...... ix LIST OF FIGURES ...... x ABBREVIATIONS ...... xii Chapter 1. Nutraceuticals as metabolic modulators for the treatment of obesity and associated diseases ...... 1 1.1. Introduction ...... 2 1.2. Metabolic flexibility in health ...... 3 ...... 6 1.3. Obesity, lipotoxicity and metabolic inflexibility ...... 7 1.5. Avocado derived lipids as metabolic modulators ...... 15 1.6. Author Contributions...... 16 Thesis objectives and hypothesis ...... 17 Chapter 2: Analytical method to detect and quantify Avocatin B in Hass avocado seed and pulp matter ...... 18 2.1. Abstract ...... 19 2.2. Introduction ...... 19 2.3. Experimental section ...... 22 2.3.1. Chemicals and reagents...... 22 2.3.2. MS tuning experiments ...... 22 2.3.3. Ultra-High performance liquid chromatography and mass spectrometry ...... 23 2.3.4. Method validation—calibration standards and quality control (QC) samples ...... 24 2.3.5. Method Validation— Precision and Accuracy ...... 25 2.3.6. Method Validation— Specificity and Selectivity ...... 25 2.3.7. Plant Material Preparation ...... 26 2.3.8. Total Lipid Extraction from Plant Material ...... 26 2.3.9. Methanolic saponification of avocado pulp and seed ...... 27 2.3.10. Sample preparation of extracts for LC-MS analysis ...... 28 2.3.11. Statistics ...... 28 2.4. Results and discussion ...... 29 2.4.1. MS tuning...... 29 2.4.2. Method validation ...... 30

vii

2.4.3. Avocadene and avocadyne in total lipid extracts of Hass avocado pulp and seed . 33 2.4.4. Methanolic Saponification of Avocado Pulp and Seed ...... 34 2.5. Conclusions ...... 36 2.6. Acknowledgments ...... 38 2.7. Author contributions ...... 38 2.8. Supporting information (Chapter 2) ...... 39 Chapter 3: Bioaccessability and bioavailability of avocado polyhydroxylated fatty alcohols ...... 44 3.1. Abstract ...... 45 3.2. Introduction ...... 45 3.3. Experimental Section ...... 48 3.3.1. Materials ...... 48 3.3.2. Determining avocadene and avocadyne concentration in lipid extracts of lyophilized avocado pulp powder ...... 49 3.3.3. Static in vitro digestion ...... 49 3.3.4. Avocatin B emulsion preparation and characterization ...... 50 3.3.5. TIM-1 Studies ...... 51 3.3.6. In vivo pilot pharmacokinetic study ...... 54 3.3.7. Statistical data analysis ...... 55 3.4. Results ...... 56 3.4.1. Avocadene and avocadyne are hydrolyzed from their bound forms after static and dynamic in vitro digestion of lyophilized avocado pulp powder ...... 56 3.4.2. Avocatin B self-assembles in an O/W microemulsion ...... 61 3.4.3. Avocadene and avocadyne delivered via AVO emulsion have moderately greater bioaccessibility compared to avocado pulp powder delivery ...... 64 3.4.4. AVO emulsion shows bioavailability and biodistribution in a pilot in vivo pharmacokinetic study ...... 67 3.5. Discussion ...... 70 3.6. Acknowledgments ...... 73 3.7. Author contributions ...... 73 3.8. Supplementary information (chapter 3) ...... 74 3.8.1. Lyophilized avocado powder proximate analysis ...... 74 3.8.2. AVO bio-analytical method ...... 75 Chapter 4: Avocatin B protects against lipotoxicity and improves insulin sensitivity in diet- induced obesity ...... 81 4.1. Abstract ...... 82

viii

4.2. Introduction ...... 82 4.3. Experimental Section ...... 84 4.3.1. Pre-clinical mouse studies...... 84 4.3.2. In vitro studies...... 87 4.3.3. Immunoblotting analysis ...... 94 4.3.4. Statistical Analysis ...... 95 4.3.5. Avocatin B oral consumption pilot clinical study...... 95 4.4. Results ...... 99 4.4.1. AVO inhibits FAO and improves glucose tolerance and insulin sensitivity after DIO is established in mice ...... 99 4.4.2. AVO inhibits FAO during lipotoxicity in pancreatic β-islet cells which improves glucose stimulated insulin secretion ...... 105 4.4.3. AVO inhibits FAO during lipotoxicity in C2C12 myotubes which improves insulin signaling ...... 110 4.4.4. AVO was well-tolerated in a pilot clinical study ...... 114 4.5. Discussion ...... 117 4.6. Acknowledgments ...... 120 4.7. Author contributions ...... 120 4.8. Supporting Information (Chapter 4) ...... 121 Chapter 5: Integrated discussion, conclusions and future work ...... 130 5.1. Preamble ...... 130 5.2. Summary and discussion of major findings ...... 130 5.3. Strengths, limitations and future work ...... 133 5.4. Conclusions ...... 137 REFERENCES ...... 138 APPENDICES ...... 161 Appendix A. University of Guelph research ethics board (REB) certificate of approval for AVO clinical study ...... 161 Appendix B. Health Canada NHPD notice of authorization for AVO clinical study...... 162 Appendix C. AVO clinical study recruitment poster ...... 163 Appendix D. AVO clinical study consent form ...... 164 Appendix E. AVO clinical study case report form ...... 170 Appendix F. AVO clinical study safety ...... 181 Appendix G. AVO clinical study Weekly telephone log ...... 183

LIST OF TABLES

Table 2.1. Summary of key molecular ion species detected when avocatin B standard is directly infused into a QE-MS in the +ESI mode ...... 29

Table 2.2. Projected Amounts of Avocadene and Avocadyne in Total Lipid Extract of Seed and Pulp of a Hass Avocadoa...... 34

Table 2.3. Projected amounts of avocadene and avocadyne in saponified extract of seed and pulp of a hass avocadoa ...... 36

Supplementary Table 2.1. Accuracy and precision for the analysis of avocadene and avocadyne in solvent ...... 43

Supplementary Table 2.2. Recovery of avocadene and avocadyne from hass avocado pulp and seed total lipid extracts...... 43

Table 3.1. Amount avocadene and avocadyne quantified before and after saponification of one- gram lyophilized avocado pulp powder total lipid content ...... 57

Table 3.2. Area under curve analysis on non-cumulative concentration—time graphs for jejunum, ileum, and jejunum+ileum for avocado milk beverage administered to TIM-1 ...... 61

Table 3.3. Area under curve analysis on non-cumulative concentration—time graphs for jejunum, ileum, and jejunum+ileum for AVO emulsion administered to TIM-1 ...... 66

Table 3.4. Fitted parameter logistical model applied on cumulative data showing PFA bioaccessibility, induction time and rate of bioaccessibility in jejunum, ileum, and jejunum+ileum for avocado powder milk beverage and AVO emulsion administered to TIM-1 67

Table 3.5. Avocatin B pilot non-compartmental pharmacokinetic analysis ...... 69

Supplementary Table 3.1. Manufacturer provided certificate of analysis for lyophilized avocado pulp powder ...... 74

Supplementary Table 3.2. AVO bioanalytical method validation parameters for pilot PK study...... 79 Supplementary Table 3.1. AVO analytical method validation parameters for TIM-1 studies. . 80

Supplementary Table 4.1. Macro-and-micro nutrient composition of avocatin B clinical trial formulations ...... 126

LIST OF FIGURES Chapter 1 Figure 1.1. The glucose—fatty acid cycle...... 6 Figure 1.2. Lipotoxicity in an obese and insulin resistant adult...... 8 Figure 1.3. Molecular mechanism of lipotoxicity and metabolic inflexibility in skeletal muscle.11

Chapter 2 Figure 2.1. Chemical structure of avocadene and avocadyne (A) and their acetate forms (acetogenins) (B)...... 22 Figure 2.2. Representative AVO LC-MS chromatographs...... 32 Figure 2.3. Hass avocado pulp and seed total lipid extraction...... 34 Figure 2.4. Methanolic potassium hydroxide saponification of Hass avocado pulp and seed. .... 36 Supplementary Figure 2.1. Representative MS/MS spectra for avocadene ...... 39 Supplementary Figure 2.2. Representative MS/MS spectra for avocadyne...... 40 Supplementary Figure 2.3. MS spectrum for AVO from LC-MS method...... 41 Supplementary Figure 2.4. Representative LC-SIM-MS chromatographs of AVO...... 42

Chapter 3 Figure 3.1. Avocadyne and avocadene are hydrolyzed from their bound forms after static in vitro digestion of lyophilized avocado powder...... 58

Figure 3.2. Cumulative bioaccessibility of avocadene and avocadyne during 360-min dynamic in vitro TIM-1 digestion of lyophilized avocado powder...... 60

Figure 3.3. AVO reduces droplet size of a NeoBee®M5— polysorbate 80 based O/W microemulsion...... 63

Figure 3.4. Bioaccessibility of avocadene and avocadyne during 360-min dynamic in vitro TIM- 1 digestion of AVO emulsion...... 65

Figure 3.5. AVO emulsion shows bioavailability and biodistribution in a pilot in vivo pharmacokinetic study...... 69

Chapter 4 Figure 4.1. Effects of AVO in a treatment model of DIO...... 102 Figure 4.2. Effects of AVO in a prevention model of DIO...... 104 Figure 4.3. AVO inhibits FAO in INS-1 (832/13) cells and improves glucose oxidation...... 106 Figure 4.4. Inhibition of FAO by AVO under lipotoxic conditions in INS-1 (832/13) cells improves glucose stimulated insulin secretion...... 109

Figure 4.5. AVO inhibits FAO in C2C12 myotubes and improves glucose oxidation...... 111 Figure 4.6. Inhibition of FAO by AVO under lipotoxic conditions improves insulin signaling in C2C12 myotubes...... 113

Figure 4.7. AVO does not impart toxicity after human oral consumption...... 116 Supplementary Figure 4.1. Complete blood count with differential (CBC w/diff) analysis on whole blood collected from 3-5 animals per group at treatment-DIO study endpoint...... 121

Supplementary Figure 4.2. Complete blood count with differential (CBC w/diff) analysis on whole blood collected from 3-5 animals per group at prevention-DIO study endpoint...... 122

Supplementary Figure 4.3. Cell viability assessments...... 123 Supplementary Figure 4.4. High resolution respirometry (HRR) example oxygraphs...... 124 Supplementary Figure 4.5. Mitochondrial mass determination...... 125 Supplementary Figure 4.6. Method of preparation for avocatin B clinical trial formulations. . 125 Supplementary Figure 4.8. Complete blood count with differential (CBC w/diff) analysis on whole blood collected from all clinical trial participants on day 0, day 30, and day 60 (endpoint)...... 128

xii

ABBREVIATIONS

7-AAD 7-aminoactinomycin D ANOVA Analysis of variance ASM Acid soluble metabolites AVO Avocatin B BMI Body mass index BSA Bovine serum albumin CoA Coenzyme A CONSORT Consolidated Standards of Reporting Trials CPT-1 Carnitine palmitoyltransferase I CVD Cardiovascular disease DAG Diacylglycerol DHA Docosahexaenoic acid DIO Diet-induced obesity DMEM Dulbecco’s modified eagle medium EPA Eicosapentaenoic acid FA Formic acid FAO Fatty acid oxidation FBS Fetal bovine serum FFA Free fatty acid GLUT 4 Glucose transporter type 4 GSIS Glucose stimulated insulin secretion HFD High fat diet HOMA-IR homeostatic model assessment of insulin resistance HRR High resolution respirometry LC-MS Liquid-chromatography—mass spectrometry MAG Monoacylglycerol MMP Mitochondrial membrane potential NAO Nonyl-acridine orange PA-BSA Palmitate-BSA

xiii

PDH Pyruvate dehydrogenase PDI Polydispersity index PDK Pyruvate dehydrogenase kinase PDP Pyruvate dehydrogenase phosphatase PFA Polyhydroxylated fatty alcohol PI Propidium iodide PSD Particle size distribution RCF Relative centrifugal force REB Research ethics board Rho123 Rhodamine 123 RIPA Radioimmunoprecipitation assay ROS Reactive oxygen species SD Standard deviation SEDDS Self-emulsifying drug delivery system SEM Standard error of mean SIM Single ion monitoring TAG triacylglycerol TCA tricarboxylic acid TIM-1 TNO gastrointestinal model (upper tract)

Chapter 1. Nutraceuticals as metabolic modulators for the treatment of obesity and associated diseases

Nawaz Ahmed†, Sarah Walker† and Paul A. Spagnuolo†*

†University of Guelph, Department of Food Science, 50 Stone Road East, Guelph, ON, N1G 2W1, Canada *To whom correspondence should be addressed. Email: [email protected]

Contents of this chapter have been minimally modified from a soon-to-be published book chapter.

1.1. Introduction

Obesity is a global pandemic spreading at an alarming rate (WHO, 2017) and is strongly related to the soaring rates of other metabolic diseases such as Type 2 Diabetes (T2D) and heart disease. Insulin resistance is a clinical hallmark of T2D defined by the inability of cells in metabolically active tissues to respond to insulin leading to glucose intolerance and lethal vascular complications. In diet-induced obesity (DIO), insulin resistance can result from excess amounts of circulating plasma free fatty acids (FFAs) accumulating into non-adipose tissues such as skeletal muscle, liver, and pancreas (Unger et al., 2010).

The optimal metabolic health of a cell is governed by the ability of mitochondria to switch freely between oxidative substrates in response to physiological and nutritional cues.

Mitochondrial dysfunction in key tissues is therefore believed to be at the centre of obesity and associated disorders, which are now denoted by a new clinical hallmark: metabolic inflexibility.

Metabolic inflexibility is defined as the inability of an organism or a single cell to adapt fuel (i.e., substrate) oxidation to fuel availability (Goodpaster & Sparks, 2018) that results in lower glucose oxidation during insulin-stimulated conditions. The importance of metabolic inflexibility is highlighted by numerous studies showing that oversupply of FFAs in metabolically active tissues overloads the mitochondria leading to incomplete fatty acid oxidation (FAO) and increased oxidative stress that worsens and/or perpetuates insulin resistance (Adams et al., 2009; Anderson et al., 2009; Gavin et al., 2018; Koves et al., 2008; Mihalik, 2010; Muoio & Neufer, 2012; Muoio

& Newgard, 2006, 2008b). In light of this emerging evidence, strategies that target mitochondrial fuel selection by modulating lipid and glucose metabolism have been proposed as therapies to improve insulin resistance (Foley J. E., 1992; Keung et al., 2013; Lopaschuk, 2016; Muoio &

Newgard, 2008a). While lifestyle interventions (e.g., diet, exercise and stress management) are

undoubtedly the most economical treatment and prevention options, they have failed to make a lasting impact at the population level. This highlights the need for pharmacotherapy as an adjunct to lifestyle interventions; however, available drugs vary in efficacy, side effect profiles, and are often inaccessible due to inadequate health insurance coverage (Bessesen & Van Gaal, 2018). This chapter will i) provide a brief review of metabolic flexibility and how metabolic modulators have the potential to restore impairments in mitochondrial lipid and glucose metabolism in obesity and associated disorders, and ii) outline objectives and hypothesis related to this thesis.

1.2. Metabolic flexibility in health

The metabolism of a cell or tissue must be able to adapt to fluctuations in energy demand or substrate availability to survive. Metabolic flexibility is evident in how healthy tissues respond to changing conditions such as going from a fasting to fed state, rest to exercise, and even to changes in circadian rhythm (Galgani et al., 2008; Goodpaster & Sparks, 2018). The concept of metabolic flexibility was first coined in 1963 when the Randle Cycle was proposed as an explanation for the relationship between fatty acid oxidation (FAO; or β-oxidation) and glucose oxidation (Hue & Taegtmeyer, 2009). Randle and colleagues discovered that while the insulin- glucagon ratio from the fasting to fed states controls anabolic vs. catabolic metabolism, there is a finer level of hormone-independent metabolic control (Hue & Taegtmeyer, 2009). This is the relationship between fatty acid and glucose oxidation where the oxidation of one substrate directly inhibits the oxidation of the other (Hue & Taegtmeyer, 2009). For example, going from a fasting to fed state involves a shift in substrate utilization from fatty acids to glucose in skeletal muscle tissue (Goodpaster & Sparks, 2018). The increased oxidation of glucose is, of course, driven by insulin but the subsequent down regulation of FAO is a result of the glucose-fatty acid cycle

(Goodpaster & Sparks, 2018). In the fasted state, fat stores are broken down to provide FFAs as

metabolic fuel substrates (Hue & Taegtmeyer, 2009). This inhibits glucose oxidation and, importantly, preserves glucose for tissues that depend on it (i.e., brain) until the next meal is consumed (Hue & Taegtmeyer, 2009). During exercise, skeletal muscle tissues need metabolic flexibility to adapt to increased energy demands where intensity and duration of exercise dictates the substrates that are utilized (Goodpaster & Sparks, 2018).

On a systemic level, the biochemical transitions described above rely on endocrine signalling between multiple organs which coordinate control of available fuel. Figure 1.1 highlights cellular components of the glucose-fatty acid cycle where allosteric inhibition of key regulatory enzymes in opposing metabolic pathways enables a rapid switch in substrate oxidation.

In the post-prandial period of a glucose rich meal, blood glucose and insulin levels spike causing glucose uptake, glycolysis, and subsequent pyruvate oxidation to acetyl Co-A in the mitochondria via the pyruvate dehydrogenase (PDH) complex (i.e., the first key regulatory enzyme complex).

Additionally, the increase in pyruvate from glycolysis inhibits pyruvate dehydrogenase kinases

(PDK) and increases the activity of pyruvate dehydrogenase phosphatases (PDP), which collectively activates the PDH complex and increases glucose oxidation (Patel et al., 2014; Randle,

1998). These glucose dependent processes increase the concentration of the cytosolic substrate malonyl-CoA, which allosterically inhibits carnitine palmitoyltransferase-1 (CPT-1). CPT-1 is a key regulatory mitochondrial enzyme that controls entry of long chain fatty acids into the mitochondria for FAO. Inhibition of CPT-1 via malonyl-CoA generally re-routes fatty acids towards triglyceride synthesis and storage. Conversely, during fasting, glucose-mediated inhibition of FAO is released by the action of the energy stress sensor adenosine monophosphate- activated protein kinase (AMPK), which inhibits acetyl-CoA carboxylase (ACC) (Randle, 1998).

Inhibition of ACC lowers malonyl-CoA concentrations and results in increases in CPT-1 activity

and the rate of mitochondrial β-oxidation. FAO increases mitochondrial acetyl Co-A and lowers nicotinamide adenine dinucleotide (NADH) levels, which collectively inhibit PDH activity via both allosteric inhibition and PDK activation (Randle, 1998). Additionally, an increase in fat availability elevates cytosolic citrate levels that inhibit glucose transporters (GLUT) and the glycolytic enzyme phosphofructokinase-1 (PFK-1), which collectively impede glucose uptake and utilization (Randle, 1998). As mentioned above, this allosteric inhibition of glucose oxidation via

FAO processes helps to conserve glucose for anabolic pathways and brain metabolism.

Importantly, when FAO inhibits glucose oxidation, a decrease in pyruvate oxidation enables its use as a substrate for pyruvate carboxylase (PC) in energetically demanding tissues, which produces oxaloacetate that feeds into the tricarboxylic acid (TCA) cycle and acts as an anaplerotic substrate. In the liver, however, lower pyruvate oxidation (due to elevated FAO) leads to the use of pyruvate as a precursor for gluconeogenesis (production of glucose from non-carbohydrate sources) to avoid hypoglycemia (Randle, 1998); this pathway becomes extremely problematic in obesity and T2D as will be explained in the next section.

Figure 1.1. The glucose—fatty acid cycle. Yellow lines showing glucose dependent processes suppressing FAO and purple lines showing FAO processes inhibiting glucose oxidation. Abbreviations: LCFA, long chain fatty acid; CD36, cluster of differentiation 36 (membrane fatty acid transporter protein); LCFA-CoA, long chain fatty acyl-coenzyme A; CPT-1, carnitine palmitoyltransferase-1; CACT, carnitine acylcarnitine translocase; NADH; nicotinamide adenine dinucleotide reduced form; FADH; flavin adenine dinucleotide reduced form; CTP, citrate transport protein; ACL, ATP-citrate lyase; ACC, acetyl- CoA carboxylase; FAS, fatty acid synthase; AMPK, adenosine monophosphate-activated protein kinase; GLUT, glucose transporter; HK, hexokinase; Glu-6-P; glucose 6-phosphate; Fru-6-P; fructose 6-phosphate; PFK, phosphofructokinase; OMM, outer mitochondrial membrane; IMS, intermembrane space; IMM, inner mitochondrial membrane, MPC, mitochondrial pyruvate carrier; PC, pyruvate carboxylase; PDH, pyruvate dehydrogenase complex; PDK, pyruvate dehydrogenase kinase; PDP, pyruvate dehydrogenase phosphatase; TCA, tricarboxylic acid cycle.

1.3. Obesity, lipotoxicity and metabolic inflexibility

The imbalance between energy intake and expenditure in obese individuals leads to expansion of adipose tissue deposits. Adipose tissue expansion is a physiological and protective process (Ruge et al., 2009) that is heavily impaired in obesity. The inability of adipose tissue to store FFAs from systemic circulation causes overflow and accumulation into non-adipose tissues

(termed ectopic fat) such as muscle, liver and pancreas (Cusi, 2010; Iozzo, 2009). Ectopic fat can lead to lipotoxicity. In lipotoxicity, tissue mitochondria are overloaded, and FAO is chronically ramped up without co-ordinated increases in TCA cycle or electron transport chain fluxes. This ultimately leads to incomplete FAO, which is characterized by a build up of intra-mitochondrial toxic lipid intermediates and free radicals that cause insulin resistance and impairment in organ function (Koves et al., 2008). Figure 1.2 highlights key metabolic impairments in an obese and insulin resistant individual.

Figure 1.2. Lipotoxicity in an obese and insulin resistant adult. Caloric excess and a sedentary lifestyle reduce expandability of adipose tissue which causes spillover of FFA to non-adipose tissue (brown arrows) leading to insulin resistance in liver, muscle, heart and pancreas. Insulin resistance in non-adipose tissue leads to key metabolic impediments (as highlighted in the figure) that causes whole body metabolic inflexibility and enables the onset of further complications like T2D and cardiovascular disease.

Clinical observations in obese and insulin resistant individuals show that insulin is unable to efficiently supress FAO or sufficiently stimulate glucose oxidation (Figure 1.2). Thus, there is an impairment in switching between fat and glucose oxidation in fasting and insulin-stimulated conditions, respectively (D. Kelley, 2005). The molecular mechanisms of obesity and T2D related metabolic inflexibility in different tissue types is a topic of great debate and continued research. It is unclear which is the first organ to become metabolically inflexible or whether metabolic inflexibility precedes or proceeds insulin resistance. There is, however, strong consensus on the molecular mechanisms of lipotoxicity and metabolic inflexibility in skeletal muscle (Koves et al.,

2008). This mechanism is highlighted in Figure 1.3 and shows how insulin resistance in skeletal

muscle is characterized by excessive β-oxidation in the post-prandial state. Excessive β-oxidation gives rise to incomplete FAO which produces lipotoxic lipid intermediates that impair glucose utilization and insulin signalling, deplete organic intermediates of the tricarboxylic acid (TCA) cycle, cause reactive oxygen species (ROS), and alters mitochondrial protein acetylation states.

This cascade of events chronically impedes mitochondrial performance and oxidative capacity

(Koves et al., 2008). Additionally, it is important to note that excessive FAO enhances PDK activity and supresses PDP activity which imposes a double setback on skeletal muscle glucose oxidation (LeBlanc et al., 2008).

Several elements of the mechanism of lipotoxicity and metabolic inflexibility in skeletal muscle have also been observed in other non-adipose tissues. In the pancreas, the site of insulin secretion and synthesis, lipotoxicity or increased FAO contributes to the pathogenesis of T2D by causing dysfunction and apoptosis of insulin-producing β-cells (Lupi et al., 2002). In pre-diabetes, where β-cell metabolic dysfunction instead of apoptosis is more prominent, FFA-induced lipotoxicity disrupts the glucose-fatty acid cycle (Zhou & Grill, 1994, 1995) and impairs glucose- stimulated insulin secretion (GSIS) resulting in hyperinsulinemia (Biden et al., 2002; Erion et al.,

2015; Ježek et al., 2018; Zhou & Grill, 1994). During ischemic heart disease, blood supply is restricted to cardiac tissue resulting in a lack of oxygen for oxidative phosphorylation and an increased reliance on FAO (Goodpaster & Sparks, 2018; Lopaschuk, 2017). In this scenario, insufficient energy production is not due to lack of available substrate but rather to metabolic inflexibility resulting from a lack of oxygen—a key distinction from the mechanism in skeletal muscle. Here, metabolic flexibility is severely impaired in the ischemic and post-ischemic heart

(as well as in angina and heart failure), as cardiac muscle cells are exposed to excess circulating

FFAs in amounts far greater than their oxidative capacity. As such, cardiac muscle cells become

locked in a fatty acid dependent state where 60-90% of energy is derived from FAO (Lopaschuk et al., 1994)—a far less efficient metabolic pathway yielding less ATP per molecule of oxygen compared to glucose oxidation. Excess FAO during and post-ischemia perpetuates ischemic damage by i) reducing glucose oxidation, ii) causing mitochondrial dysfunction and oxidative stress, and iii) exacerbating ischemia induced ionic imbalance (Lopaschuk et al., 2010). In the liver, insulin resistance perpetuates lipid accumulation, which is clinically observed as a hallmark of non-alcoholic fatty liver disease (NAFLD). In NAFLD, hepatic uptake of lipids and de novo lipogenesis are increased causing a compensatory enhancement of FAO (Ipsen et al., 2018).

Increased FAO in fatty liver is insufficient in normalizing lipid levels and instead promotes oxidative stress, gluconeogenesis and metabolic inflexibility (Gastaldelli, 2017).

Figure 1.3. Molecular mechanism of lipotoxicity and metabolic inflexibility in skeletal muscle. Caloric excess and inactivity results in a surplus of fatty acids in mitochondria that promotes β- oxidation without a coordinated increase in TCA cycle flux. Chronic lipid overload of the β- oxidative pathway leads to an increase in redox potential and depletion of free CoA, NAD, and carnitine. Together this exerts negative pressure on the TCA cycle and electron transport chain (ETC) while also inhibiting glucose disposal through PDH by increasing PDK activity and inhibiting PDP activity. The increased oxidative and reductive stress disrupts mitochondrial performance and alters the acetylation state of mitochondrial proteins (red stars). As metabolic by- products of incomplete FAO accumulate (e.g., acylcarnitines, NADH, Reactive Oxygen Species), LCFA-CoAs are rerouted towards TAG synthesis, however, the mitochondrial overload causes cytosolic buildup of DAGs and ceramides. Collectively, mitochondrial and lipid-derived stresses activate serine kinases that block insulin receptor (IR) signaling, GLUT4 translocation (through IRS-1/PI3K pathway) and glucose metabolism. Abbreviations: DAG, diacylglycerol; IR, insulin receptor; IRS-1, insulin receptor substrate-1; PI3K, phosphoinositide 3-kinase. Adopted and modified from Koves et al., 2008.

1.4. Metabolic modulators and modes of modulation

Obesity and associated metabolic disorders share a common underlying pathological event characterized by alterations in fatty acid metabolism that impairs glucose metabolism. Habitual exercise has been shown to combat this mitochondrial lipid overload by increasing TCA cycle flux and mitochondrial biogenesis thereby restoring mitochondrial function and enhancing muscle insulin sensitivity (Houmard et al., 2002). However, if exercise cannot be easily implemented, a potential treatment strategy is to target mitochondrial dysfunction by directly or indirectly modulating fatty acid and/or glucose metabolism. Pharmacological agents as metabolic modulators have been researched primarily in the field of heart disease and a number of these agents are now in clinical trials or clinically available. Based on current preclinical and clinical evidence, Figure 1.4 highlights the main modes by which metabolic modulation can be achieved in a variety of tissues that are affected by lipotoxicity in obesity and associated metabolic diseases.

One of the biggest successes for metabolic modulators is trimetazidine, which is currently approved as a treatment for angina pectoris. It acts by inhibiting the final enzyme of β-oxidation thereby decreasing FAO and leading to a subsequent increase in glucose oxidation in myocardial tissue (Chrusciel et al., 2014). Trimetazidine upregulates PDH activity in cardiomyocytes and helps re-couple glucose oxidation to glycolysis (Chrusciel et al., 2014). It is also involved in restoring function in mitochondria damaged during ischemia and directly inhibits cardiac fibrosis

(Chrusciel et al., 2014). There are few larger scale clinical trials testing trimetazidine; however, smaller trials and meta-analyses demonstrate its safety and, since it does not decrease resting blood pressure, it can be used with standard treatments like beta blockers (Steggall et al., 2017).

Ranolazine is another metabolic modulator approved for use in angina pectoris that is showing promising results in patients with heart failure (Maier et al., 2013). Although initially proposed as

a partial FAO inhibitor (i.e., causes modest (~30-40%) FAO inhibition (Luiken et al., 2009)), it imparts activity mainly by restoring ionic homeostasis during heart failure (Ardehali et al., 2012).

Etomoxir is a direct inhibitor of CPT-1 and is typically used to experimentally study FAO, as it exhibited severe dose limiting liver toxicity in clinical trials (Holubarsch et al., 2007). Perhexiline is also a CPT-1 inhibitor with a narrow therapeutic window (i.e., potential hepato- and neuro- toxicity) that has shown clinical promise for heart failure (Jaswal et al., 2011; Killalea & Krum,

2001; Steggall et al., 2017). There are no additional classes of FAO inhibitors currently being evaluated in clinical trials for ischemic heart disease; however, compounds that inhibit malonyl

CoA decarboxylase, an enzyme that catalyzes malonyl CoA to acetyl CoA thereby reducing CPT-

1 activity, are showing preclinical promise in mouse models of heart failure (Ahmadi et al., 2017).

Unlike for heart disease, pharmacological modulators of metabolism have yielded limited clinical success when evaluated as treatments for obesity and T2D. While the beneficial effects of trimetazidine as a partial FAO inhibitor in heart muscle are well established, its efficacy in the treatment of obesity and diabetes is hampered due to its limited activity in skeletal muscle (J. R.

Ussher et al., 2014). Ranolazine has recently been approved for use in combination therapies for the treatment and management of T2D; however; its beneficial activity is not due to FAO inhibition in skeletal muscle (Pettus et al., 2016). Recently, ranolazine was demonstrated to increase liver

PDH activity, which reduced fat accumulation, body weight gain and glucose intolerance in a mouse model of obesity (Batran et al., 2019). Future clinical studies are likely to elucidate its mechanism of action in obese and diabetic humans. The unfavourable safety profile of etomoxir has limited the clinical development of CPT-1 inhibitors for the treatment of diabetes and obesity

(Chiodi, 2017); nonetheless, recent efforts in pilot clinical studies have reported favourable effects of etomoxir with dose adjustments (Timmers et al., 2012). A reversible and selective inhibitor of

liver specific CPT-1, Teglicar, demonstrated impressive preclinical results where it significantly lowered fasting blood glucose, insulin resistance and postprandial gluconeogenesis in obese and diabetic mice (Conti et al., 2011); Phase I clinical trial results of Teglicar are not yet published. In the 1990s, thiazolidines (TZDs) were introduced as a new class of drug for the treatment of T2D.

TZDs (e.g., rosiglitazone) are agonists for the transcription factor peroxisome proliferator- activated receptor gamma (PPAR-γ). The activation of PPAR-γ in adipose tissue leads to adipogenesis that reduces ectopic fat accumulation in other tissues and can restore metabolic flexibility. TZDs currently in the market have known adverse side effects like cardiac toxicity, hepatotoxicity and gastric discomfort, which hinders patient compliance impacting overall effectiveness (Tyagi et al., 2011). Lastly, malonyl CoA decarboxylase inhibitors (J R. Ussher et al., 2016) and a novel class of peptide based CPT-1 inhibitors (Gao et al., 2015) have shown preclinical promise in mouse models of obesity but are far from clinical translation.

Figure 1.4. Modes of metabolic modulation. Modes of metabolic modulation that improve glucose oxidation in metabolically active tissues affected by obesity-associated lipotoxicity.

1.5. Avocado derived lipids as metabolic modulators

The search for alternative treatments for obesity and associated diseases has led to the discovery of thousands of natural compounds derived from plants or food (Rochlani et al., 2017).

However, only a handful of natural compounds have the potential to act as classic metabolic modulators, i.e., agents that exert direct action in mitochondria to modulate glucose and lipid metabolism thereby improving metabolic flexibility. These compounds include L-carnitine (Bene et al., 2018; Koves et al., 2012), R-lipoic acid (Gomes et al., 2012; Korotchkina et al., 2004), thiamine (González-Ortiz et al., 2011; Rabbani et al., 2009), and nicotinamide riboside (Dollerup et al., 2018; Jing et al., 2013) to highlight a few.

Lipids are seldomly found to act as mitochondrial metabolic modulators. The effects of omega-3 fatty acids in the mitochondria are well reported. However, their modulatory effect on mitochondrial metabolism is secondary to their activation of nuclear transcriptional factors that stimulate mitochondrial biogenesis and capacity which indirectly improves metabolic flexibility

(Serrano et al., 2016). Surprisingly, Spagnuolo and colleagues showed that lipids derived from avocados can impart direct modulatory activity in the mitochondria. Seventeen to twenty-one carbon-chain polyhydroxylated fatty alcohols (PFAs) were first discovered in avocado seeds

(Persea americana Mill.; Lauraceae) in 1969 (Kashman et al., 1969a) and have since been found in avocado pulp (the edible portion of the avocado)(Ahmed et al., 2018; Patent No.

20160249613A1, 2013). Spagnuolo and colleagues previously determined that a mixture of

PFAs called avocatin B (1:1 mixture of avocadene and avocadyne, herein referred to as AVO), possess novel anticancer activity by accumulating in mitochondria and selectively inducing apoptosis of leukemia and leukemia stem cells through inhibition of FAO (E. A. Lee et al.,

2015). Leukemia stem cells exhibit an altered metabolic phenotype characterized by an excessive

reliance on FAO for ATP production and survival (Samudio et al., 2010). In contrast, normal cells from metabolic active tissues like skeletal muscle exhibit greater metabolic flexibility where FAO, glucose oxidation and amino acid utilization is dictated by substrate availability and energy demand (Muoio, 2014). Hence, AVO induces cell death in leukemia stem cells in a selective manner where normal cells are unaffected.

1.6. Author Contributions

NA conceptualized chapter content, performed literature search, composed figures and drafted the review. SW assisted NA with introductory section on metabolic flexibility and edited the review. PAS edited the review.

Thesis objectives and hypothesis

The overall hypothesis of the present work is that AVO inhibits FAO in in vitro and in vivo models of lipotoxicity to restore insulin signaling and glucose tolerance. To test this hypothesis, this thesis addressed the following objectives:

1. Develop and validate an analytical method for the detection and quantification of AVO in

avocado pulp and seed— Study 1

2. Develop optimal AVO formulations for delivery in mice and healthy human participants

and evaluate their bioaccessibility and bioavailability— Study 2

3. Determine if oral supplementation with AVO can delay the onset of obesity and insulin

resistance in C57BL/6J mice on high fat diet (HFD) for 12 weeks (prevention study) —

Study 3

4. Determine if oral supplementation with AVO can reverse or reduce markers of obesity

and insulin resistance in C57BL/6J mice on HFD for 12 weeks (treatment study) —

Study 3

5. Determine if AVO can safely inhibit FAO and increase glucose utilization under

lipotoxic conditions in a pancreatic β-islet cell line (INS-1 (832/13) and in a skeletal

muscle cell line (differentiated C2C12 myotubes) — Study 3

6. Determine safety of avocado dietary supplement, standardized to amounts of AVO, in

healthy human participants — Study 3

Chapter 2: Analytical method to detect and quantify Avocatin B in Hass avocado seed and pulp matter

Nawaz Ahmed,† Richard W. Smith,‡ Juan J. Aristizabal Henao,§ Ken D. Stark,§ and Paul A. Spagnuolo*,†

†University of Guelph, Department of Food Science, 50 Stone Road East, Guelph, ON, N1G 2W1, Canada ‡University of Waterloo Mass Spectrometry Facility, Department of Chemistry, 200 University Avenue West, Waterloo, ON, N2L 3G1, Canada §University of Waterloo, Department of Kinesiology, 200 University Avenue West, Waterloo, ON, N2L 3G1, Canada

*To whom correspondence should be addressed. Email: [email protected]

Contents of this chapter are published and reprinted with permission from the Journal of Natural Products.

Ahmed, N., Smith, R. W., Henao, J. J. A., Stark, K. D., & Spagnuolo, P. A. (2018). Analytical Method To Detect and Quantify Avocatin B in Hass Avocado Seed and Pulp Matter. Journal of Natural Products, 81(4), 818–824. DOI:10.1021/acs.jnatprod.7b00914.

2.1. Abstract

Avocatin B, an avocado-derived compound mixture, was demonstrated recently to possess potent anticancer activity by selectively targeting and eliminating leukemia stem cells. Avocatin

B is a mixture of avocadene and avocadyne, two 17-carbon polyhydroxylated fatty alcohols

(PFAs), first discovered in avocado seeds; their quantities in avocado pulp are unknown.

Analytical methods to detect avocado seed PFAs have utilized NMR spectroscopy and GC-MS; both of these lack in quantitative capacity and accuracy. Herein, we report a sensitive LC-MS method for the quantitation of avocadene and avocadyne in avocado seed and pulp. The method has a reliable and linear response range of 0.1-50 µM (0.03-17.2 ng/µL) for both avocadene and avocadyne (r2>0.990) with a lower limit of quantitation (LLOQ) of 0.1 µM. The intra-and inter- assay accuracy and precision of the quality control (QC) samples at LLOQ showed ≤18.2% percentage error and ≤14.4% coefficient of variation (CV). The intra-and inter-assay accuracy and precision for QC samples at low and high concentrations were well below 10% percentage error and CV. This method was successfully applied to quantify avocadene and avocadyne in total lipid extracts of Hass avocado pulp and seed matter.

2.2. Introduction

Kashman et al. in 1969 were the first to discover C17 to C21 carbon-chain polyhydroxylated fatty alcohols (PFAs) in avocado seeds (Persea americana Mill.; Lauraceae) (Kashman et al.,

1969a, 1969b). Avocado seed-derived PFAs are a group of lipids with a long aliphatic chain, having a terminal unsaturation of either an olefinic (alkene) or an acetylenic (alkyne) nature, and multiple hydroxylations on the opposing end. Since their discovery, avocado PFAs have been used in topical cosmetic formulations for skin care products (Patent No. 9,416,333, 2014; Patent No.

20150175933, 2015), and as food additives due to their insecticidal, antimicrobial, and spore-

inhibiting properties (Adikaram et al., 1992; Patent No. 20160249613A1, 2013; Oberlies et al.,

1998). Our laboratory previously determined that avocatin B, a mixture of avocadene and avocadyne (Figure 2.1A), possesses novel anticancer activity by accumulating in mitochondria and selectively inducing apoptosis of leukemia and leukemia stem cells (Lee et al., 2015).

Mechanistically, avocatin B inhibited fatty acid oxidation (FAO), a cellular process utilizing fat for energy, which was determined recently to be critical in acute myeloid leukemia (AML)

(Samudio et al., 2010). Our interest in avocatin B is highlighted by its potency, exhibiting bioactivity at concentrations 10 times that of etomoxir (Lee et al., 2015), a standard compound used to explore the FAO pathway. The potential clinical utility of avocatin B is further highlighted by its ability to enhance the anti-cancer activity of cytarabine synergistically (Tcheng et al., 2017).

Studies to date have not clearly identified efficient extraction and analytical methods to specifically detect and quantify avocadene and avocadyne. Currently, gas chromatography-mass spectrometry (GC-MS) is the only analytical technique reported in patent literature for quantification of PFAs like avocadene and avocadyne in avocado seed and pulp extracts,

(Beyazova et al., 2010; Patent No. 6,582,688 B1, 2003; Patent No. 2011/0217251, 2011; Patent

No. 9,416,333, 2014; Patent No. 20150175933, 2015) where specific details of analytical procedures have not been outlined to allow comparisons with existing GC methods (Brown, 1973).

Simple HPLC-based methods for PFAs have also been reported in the literature that rely on

UV/Vis (Hashimura et al., 2001; Oberlies et al., 1998), or photodiode (PDA) detectors (Rodríguez-

López et al., 2015; Rodriguez-Sanchez et al., 2013). Quantifying PFAs like avocadene and avocadyne using existing GC- and HPLC-based methods requires extensive extraction and sample preparation procedures where identified eluted peaks need to be validated using surrogate methods like proton nuclear magnetic resonance (1H NMR) and Fourier-transform infrared (IR)

spectroscopy. Kashman and colleagues also discovered acetogenins in avocado seeds, specifically avocadene acetate and avocadyne acetate (Figure 2.1B), for which quantitative HPLC-PDA

(Rodríguez-López et al., 2015) and HPLC-MS (Rodríguez-López et al., 2015; Rosenblat et al.,

2011) methods do exist. While structurally similar, the bioactivity of acetogenins and PFAs are quite distinct,(D’Ambrosio et al., 2011; Ding et al., 2007; Patent No. 20170304251A1, 2016) thus the objective of this study was to specifically develop an analytical method to quantify the amounts of avocadene and avocadyne in Hass avocado seed and pulp matter.

To develop a simple and effective analytical method, a reversed-phase chromatographic separation coupled to a high-resolution quadrupole-orbitrap mass spectrometer (high resolution

LC-MS) was utilized for sensitive and complete quantitation in which no extensive extraction steps, sample preparation or additional structure validation would be required. A commercially available standard of avocatin B was used to ascertain the exact masses of the avocadene and avocadyne molecular ions (MS) and their subsequent fragmentations (MS/MS) in the positive electrospray ionization mode (ESI) by high resolution quadrupole-orbitrap mass spectrometry. The

+ fragmentation peaks [M+H-H2O] of avocadene and avocadyne were the most intense and enabled the development of a sensitive high mass resolution UHPLC-MS method for quantification. For efficient extraction of avocadene and avocadyne from avocado pulp and seeds, the total lipid extraction method of Folch and colleagues (Folch et al., 1957) was utilized. Saponification of both the total lipid extracts as well as direct seed and pulp matter was also carried out to measure the release of bound PFAs from complex ester linkages. This work also demonstrated that avocadene and avocadyne are found in appreciable quantities in the pulp and seeds of Hass avocados, the most cultivated and widely consumed variety of avocados in the world (Dreher & Davenport,

2013).

Figure 2.1. Chemical structure of avocadene and avocadyne (A) and their acetate forms (acetogenins) (B).

2.3. Experimental section

2.3.1. Chemicals and reagents

HPLC grade solvents acetonitrile (ACN), chloroform, methanol (MeOH), and hexane were purchased from Thermo-Fisher Scientific (Napean, ON, Canada). Reagents including monobasic sodium phosphate, formic acid (Optima grade), 12 M sodium hydroxide were purchased from

Sigma (St. Louis, MO, USA). Avocatin B standard was purchased from Microsource Discovery

Systems Inc. (CT, USA) where NMR spectroscopy was used to determine that the ratio of avocadyne:avocadene was 1:1.

2.3.2. MS tuning experiments

Direct infusion studies of avocatin B standard (1:1 mixture of avocadene and avocadyne) into the Thermo Scientific Q-Exactive (QE) Quadrupole-Orbitrap Mass Spectrometer (Thermo-

Fisher Scientific, Waltham, MA) were performed to record all abundant molecular ion species,

adducts and fragments. For all direct infusion studies, the Q-Exactive mass spectrometer (QE) was operated using the following ESI parameters: heater temperature, 0 °C; sheath gas flow rate, 2; auxiliary gas flow rate, 0; sweep gas, 0; spray voltage, 3.2 kV; capillary temperature, 275 °C, and

S-lens 60. Scan parameters were set to: full-scan range, m/z 100 to 1000; polarity, positive; mass resolution, 70000; lock mass, m/z 371.10123 (polysiloxane); automatic gain control (AGC) target,

1e6; and maximum injection time (maxIT) was 50 ms. Avocatin B standard stock was made up to

5 mmol/L (5 mM or 2.85 mg/mL) in 1:1 CH3CN-MeOH and was diluted to 10 µM in 1:1 methanol- water and 0.1% formic acid (FA). Ten-µM standards were directly infused into the QE using a 500

µL Hamilton syringe (Hamilton Company, NV, USA) connected directly to the ESI source via fused-silica tubing at a flow rate of 10 µL/min. By comparing blank runs against PFA standards, the most abundant molecular ions were determined. All ion elemental compositions were confirmed by high resolution, accurate mass determinations with lock mass correction. Targeted

MS/MS experiments on precursor ions ([M+H] +) of avocadene and avocadyne were also performed using the same 10 µM avocatin B standard injections as described above where mass resolution was the only ESI parameter changed to 35000. MS/MS data were collected at normalized collisional energy (NCE) values that ranged from 10 to 40.

2.3.3. Ultra-High performance liquid chromatography and mass spectrometry

For LC-MS method development, the QE was coupled to a Dionex UltiMate 3000 UPLC

System (Dionex Corporation, Bannockburn, IL) using Chromeleon Xpress (Version 7.2; Thermo-

Fisher Scientific, Waltham, MA, USA). For LC-MS, only three MS parameters were altered compared to the direct infusion studies: (i) capillary temperature (300 °C), (ii) spray voltage

(3.5kV), (iii) lock mass (di-isooctyl phthalate, m/z ratio 391.28429 as was present in mobile

phase). A C18 Zorbax Eclipse Plus UHPLC reversed-phase column was used (2.1 x 5.1 mm, 1.8 micron particle size) (Agilent, Santa Clara, CA, USA) for UHPLC. The mobile phases used for the chromatographic separation of avocatin B was water+0.1% formic acid (FA) (solvent A) and

CH3CN+0.1% FA (solvent B). The multi-step gradient was 40% B for 0-2 min, 100% B for 2-15 min, 100% B hold for 15-19 min, 40% B for 19-20 min and allowed to equilibrate at 40% B for

20-30 min. The flow rate was set at 0.2 mL/min and the column compartment was kept at room temperature. The Thermo Xcalibur QualBrowser (Version 2.1; Thermo-Fisher Scientific) was used for extracting ion chromatograms, exporting MS spectra, and integrating peak areas for quantification.

In an attempt to increase the sensitivity of the LC-MS method described above, the quadrupole mass filter capabilities of the Q-Orbitrap were used by specifying the m/z ratios of avocadene and avocadyne fragmentation peaks [M+H-H2O]+ in a targeted inclusion list. Running the LC-MS method described above in this high-resolution selected-ion monitoring (SIM) mode enabled avocadene and avocadyne fragmentation ions to be detected by the QE and achieved a much lower detection limit compared to the LC-MS method.

2.3.4. Method validation—calibration standards and quality control (QC) samples

On LC-MS analysis days, a stock solution of avocatin B standard was prepared in 1:1

CH3CN-MeOH at 5 mM (2.85 mg/mL). The stock solution was then diluted in 1:1 CH3CN-MeOH to produce a working solution containing 100 µM avocatin B. Eight calibration standards were prepared from the working solution by making serial dilutions at the final concentrations of 0.1, 5,

10, 20, 30, 40, 50 and 60 µM (a standard curve with a range of 0.03-17.2 ng/µL of avocadene or avocadyne injected using the avocatin B standard) in the LC starting gradient (60% water-40%

CH3CN with 0.1% FA). QC samples containing 0.1, 5 and 50 µM avocatin B were prepared the same way using a different 5 mM avocatin B stock. Ten µL of all prepared calibration standards and QC samples were injected into the LC-MS method. Standard curves were constructed using least-square linear regression of peak area under curve versus nominal standard concentrations.

Linearity was assessed by evaluating the slope, intercept and coefficient of determination (r2) of three different calibration curves produced on separate analytical days.

2.3.5. Method Validation— Precision and Accuracy

Intra- and inter-day accuracy and precision of the developed method were determined by assaying three concentrations of QCs (LLOQ = 0.1 µM, low QC = 5 µM, and high QC = 50 µM) in triplicates on three different analytical days. Precision was reported as percent coefficient of variation (%CV) of replicates within one sample run (intra-assay) or between sample runs (inter- assay). Intra- and inter-assay accuracy was reported as percent relative error (% RE) or the percent deviation of QC replicates from nominal concentration. The acceptance limit for accuracy and precision, at low and high QC concentration levels, were set to 15% RE and 15% CV, respectively.

For LLOQ, accuracy and precision acceptance limits below 20% RE and 20% CV, respectively were acceptable in keeping with United States Food and Drug Administration guidelines (Zimmer,

2014).

2.3.6. Method Validation— Specificity and Selectivity

The total lipid profiles of all avocado seed and pulp samples were expected to contain avocadene and avocadyne (Kashman et al., 1969a, 1969b). To test if unknown constituents of the extracted samples interfere with the analysis of the target analytes, 5 µM avocatin B post-extraction

spike control samples were prepared for every unspiked sample (Chambers et al., 2007). Both samples were injected into the method and percentage recovery calculations were performed to assess selectivity and specificity. For instance, if the concentration of an unspiked sample was reported from the method to be at 2.5 µM avocadene, its respective post-extraction spike control sample was expected to contain 7.5 µM ±15% avocadene.

2.3.7. Plant Material Preparation

Three ripe Hass avocados were purchased from Sobey's Canada (Guelph, ON, Canada) on three separate occasions in July 2016 and prepared separately. The seed and pulp of each avocado was separated from the peel and 3-5 mm thick slices (4 inches in length) were placed evenly inside a ventilated oven for 2 h at 65 °C. Dried pulp slices were further cut into smaller particles whereas dried seed slices were pulverized in a high-speed food processor. Dried pulp and seed matter from each of the three avocados were frozen at -80 °C until extraction day. One gram of dried pulp and seed samples from each avocado were used for all solvent extraction methods.

2.3.8. Total Lipid Extraction from Plant Material

For total lipid extraction of the prepared dried avocado seed and pulp material, a modified

Folch protocol (Folch et al., 1957) was adopted. Briefly, 1 g of sample was macerated in 20 mL of 2:1 chloroform-methanol (v/v) for different periods of time (immediate, 24 h, 48 h, 72 h) at room temperature, away from light. After maceration, 3 mL of 0.2 M sodium-phosphate (NaHPO4) buffer in ddH2O (pH 4.4) were added to induce layer separation. After inversion, samples were centrifuged for 5 min at 1500 rcf. The total lipid containing organic layer was collected and an additional 9 mL chloroform was added to the aqueous layer as a wash step and as an additional

round of extraction. The second organic layer was combined with the first and the buffer layer was discarded. Chloroform extracts were rotary evaporated and stored at 4 °C until sample preparation was required for LC-MS analysis. All total lipid extraction experiments were performed in duplicate for each of the three avocados.

2.3.9. Methanolic saponification of avocado pulp and seed

Literature evidence from environmental and bacterial samples show that several classes of fatty alcohols are bound to complex ester linkages (Cook et al., 2016; Mudge, 2005). To release free avocadene and avocadyne from all potential bound forms (e.g., wax esters, acetogenins

(Figure 2.1B), or triacylglycerols (Takenaga et al., 2008)) in avocado seed and pulp material, three different saponification methods were tested. First, total lipids were extracted from seed and pulp samples, as outlined above, after which 50 mL of methanol containing 100 µM sodium hydroxide was added and allowed to shake at a moderate speed for 16 h at 37 °C. After saponification, a liquid-liquid extraction was performed with 40 mL hexane to extract unsaponifiable substances

(note that avocadene and avocadyne are not soluble in hexane). The methanol (lower phase) was collected and re-acidified with 100 µM HCl to protonate fatty acid salts and any free fatty alcohols that may have been deprotonated in the saponification. The methanol phase was rotary evaporated and stored at 4 °C until sample preparation was required for LC-MS analysis. This method is referred to as "overnight saponification of total lipids—low KOH." A variation of this method was also tested where 200 mM KOH was used instead of 100 µM, this method is referred to as

"overnight saponification of total lipids—high KOH." The third saponification method, termed

"direct methanolic saponification," was applied to seed and pulp material directly to release bound fatty alcohols. Briefly, 1 g of seed or pulp material was saponified under reflux in 50 mL of 1 M

KOH for 1 h after which the solution was left to cool. Unsaponifiable constituents were collected in 40 mL hexane and discarded as outlined above and the same acidification step was applied to the methanol phase which was then evaporated and stored at 4 °C until sample preparation was required. All saponification experiments were performed in duplicate for each of the three avocados.

2.3.10. Sample preparation of extracts for LC-MS analysis

All dried extracts were dissolved in 10 mL 1:1 CH3CN-MeOH and vortexed for several min. Ten µL of dissolved extracts were added to 990 µL of LC starting gradient (100-fold dilution) to prepare the unspiked final samples. For each unspiked sample, a post-extraction spiked sample was also prepared which contained 5 µM avocatin B standard. All samples were prepared in duplicates and 10 µL of each was injected. To minimize degradation of target analyte, samples were injected within 24 h of reconstitution.

2.3.11. Statistics

All data were expressed as means ± standard deviations and were analyzed using two-way analyses of variance (ANOVA) and Bonferroni's post hoc test using GraphPad Prism 6.0

(GraphPad Software, San Diego, CA, USA). For histograms, * = p <0.05, ** = p <0.01, **** = p

<0.0001.

2.4. Results and discussion

2.4.1. MS tuning

Direct infusion studies with an avocatin B standard in full-scan positive ESI mode revealed

+ + + + three key molecular ions ([M+H] , [M+Na] and [M+K] ) and three fragment ions ([M+H-H2O] ,

+ + + [M+H-2H2O] , [M+H-3H2O] ), as outlined in Table 2.1. Avocadene and avocadyne [M+H] ion

+ intensities were relatively low in abundance in comparison to their fragment ions [M+H-H2O]

+ and [M+H-2H2O] , which showed high intensities. The fragment ions were confirmed by MS/MS to be from the avocadene and avocadyne molecular ions [M+H]+ (Supplementary Figures 2.1 and

2.2, Supporting Information Chapter 2). Sodium adduct peaks, [M+Na]+, were the most abundant for both avocadene and avocadyne in comparison to their potassium adduct peaks and were completely resistant to fragmentation with a range of normalized collision energies (NCE = 10-

40) (data not shown). Given that the original source of the molecular adduct ion peaks was

+ unknown, all direct infusion studies revealed that [M+H-H2O] ions for both avocadene and avocadyne could potentially be suitable for quantification in a LC-MS method.

Table 2.1. Summary of key molecular ion species detected when avocatin B standard is directly infused into a QE-MS in the +ESI mode

abundant theoretical mass from elemental composition uncertainty (delta) molecular ion m/z value (ppm) species detected

avocadene avocadyne avocadene avocadyne avocadene avocadyne

+ [M+H] 287.25807 285.24241 C17H35O3 C17H33O3 -0.1 -0.06

+ [M+H-H2O] 269.24749 267.23183 C17H33O2 C17H31O2 -0.01 -0.18

+ [M+H-2H2O] 251.23703 249.22137 C17H31O C17H29O 1.21 0.09

+ [M+H-3H2O] 233.22737 231.21084 C17H29 C17H27 1.14 0.12

+ + + [M+Na] 309.24004 307.22438 C17H34O3Na C17H32O3Na -1.14 -0.21

+ + + [M+K] 325.21404 323.19838 C17H34O3K C17H32O3K 1.43 0.16

2.4.2. Method validation

A typical UHPLC-MS chromatograph for 100 µM avocatin B standard (28.4 and 28.6 ng/µL avocadyne and avocadene, respectively) is presented in Figure 2.2A. Retention times for avocadyne and avocadene were 5.65 and 7.94 min, respectively. As predicted from direct infusion

+ studies, the intensities of [M+H-H2O] ions for both avocadyne and avocadene were consistently

10-fold higher compared to their [M+H] + molecular ion peaks (Figure 2.2A) and were thus chosen for quantification. The sodium and potassium adduct peaks for both avocadene and avocadyne did not have any significant intensity in the LC-MS method (Supplementary Figure 2.3) compared to what was observed in direct infusion studies. Linearity was assessed based on the average of three standard curves from three separate LC-MS validation days where acceptable linearity was achieved in the range of 0.1 µM-60 µM avocatin B standard (or 0.03-17.2 ng/µL of avocadyne or avocadene), with correlation coefficients (r2)>0.990 for all validation batches (Figure 2.2B). For avocadyne, a typical regression equation for calibration curves was y = 9.036e7x + 1.442e8, where y represents the peak area under the curve and x represents the concentration of avocatin B. For

8 8 avocadene, a typical regression equation for calibration curves was y = 1.232e x+2.214e .

The assay sensitivity of 0.1 µM avocatin B (0.03 ng/µL avocadene or avocadyne) (Figure

2.2C) was determined by the analysis of LLOQ triplicate samples on different analytical days (n

= 6). The inter-assay comparison for LLOQ samples yielded acceptable accuracies of 18.2% percentage error for avocadene and 13.3% percentage error for avocadyne (Supplementary Table

2.1). The inter-assay precision for LLOQ were also acceptable at 13.8% CV for avocadene and

14.4% CV for avocadyne (Supplementary Table 2.1). The intra- and inter-assay precisions for avocadene and avocadyne low and high QC samples were ≤ 7.5% CV and accuracies (percentage error) ranged from -0.5-9.3% (Supplementary Table 2.1). Both precision and accuracy for all QC

samples were found to be within the pre-defined acceptance criteria suggesting the developed LC-

MS method was suitable for the quantification of avocadene and avocadyne. The assay sensitivity increased to 5 nM avocatin B (0.001 ng/µL avocadene or avocadyne) when the LC-MS method described was run in a high- resolution selected-ion monitoring (SIM) mode, which only enabled

+ the avocadene and avocadyne fragment ions, [M+H-H2O] , to be detected (Supplementary Figure

2.4). Given the relatively high abundance of avocadene and avocadyne in seed and pulp matter, the added sensitivity provided by the LC-SIM-MS method was not required for this study.

Since there are no commercially available internal standards of avocatin B, the specificity and selectivity of the LC-MS method were assessed using the post-extraction spike method

(Chambers et al., 2007). The percentage recovery of spiked avocadene and avocadyne (from 5 µM avocatin B standard) was assessed using two total lipid extracts of Hass avocado seed and pulp, respectively. Supplementary Table 2.2 demonstrates that the percent recovery of avocadene and avocadyne from the total lipid extracts of seed and pulp were well within acceptance limits, further suggesting that endogenous compounds in the pulp or seed did not interfere with the ionization of the target analytes (106.5± 10.0% and 104.4± 2.8% from the pulp and seed for avocadene, respectively; 100.8± 3.1% and 101.5± 3.3% from the pulp and seed for avocadyne, respectively).

+ A v o c a d e n e [M + H -H O ] B. 2 A. 8 .0  1 0 9 + A v o c a d y n e [M + H -H 2 O ] Total ion current 7 .0  1 0 9

e 6 .0  1 0

v r

u Y = 1.232e+008*X + 2.214e+008

C 9 2

5 .0  1 0 R = 0 .9 9 0 9

e d

n + 9

U 4 .0  1 0

Avocadyne [M+H] a

m/z range: e r 9 A 3 .0 1 0 285.24000-285.25000 

k Y = 9.036e+007*X + 1.442e+008 a

e 2 R = 0 .9 9 2 3 P 2 .0  1 0 9

1 .0  1 0 9 Avocadene [M+H]+ m/z range: 0 0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0 4 5 5 0 5 5 6 0 287.25000-287.26000 [A v o c a tin B ]  M C. + Avocadyne [M+H-H2O] m/z range: + Avocadyne [M+H-H2O] 267.23000-267.24000 m/z range: 267.23000-267.24000

+ Avocadene [M+H-H2O] + Avocadene [M+H-H2O] m/z range: m/z range: 269.24000-269.25000 269.24000-269.25000

Figure 2.2. Representative AVO LC-MS chromatographs. (A) Representative LC-MS chromatogram of 100 µM avocatin B standard injected in the described analytical method. (B) Typical standard curve for serial dilutions of avocatin B standard injected into LC-MS method. (C) Avocatin B extracted ion chromatograms at LLOQ (0.1 µM—0.03 ng/µL of avocadene or avocadyne) employing an extraction window of 10 mDa.

2.4.3. Avocadene and avocadyne in total lipid extracts of Hass avocado pulp and seed

Total lipid extracts from 1 g of Hass avocado dried seed and pulp samples were prepared and analyzed to determine the total amounts of avocadene and avocadyne. To date, this is the first report in the literature that shows the extraction of avocadene and avocadyne from both the pulp and seed using the Folch method (Folch et al., 1957). The amount of avocadene and avocadyne extracted from both the pulp and seed increased in proportion to the solvent maceration time

(Figure 2.3A and B). In total lipid extractions after maceration for 72 h, avocadyne was found in the pulp and seed at 0.18 ±0.04 mg/g dry weight (DW) and 0.41 ± 0.02 mg/g DW, respectively.

Similarly, avocadene was extracted in the pulp and seed at 0.22 ± 0.04 mg/g DW and 0.43 ± 0.04 mg/g DW, respectively (Table 2.2). The concentration of both avocadene and avocadyne were close to two-fold higher in the seed when compared to the pulp (p < 0.0001). However, there were no differences between avocadene and avocadyne concentrations within the pulp or seed matter

(Table 2.2). These findings were also reported by Kashman and colleagues (Kashman et al., 1969a,

1969b) where a soxhlet extraction method was used for the avocado seeds and pulp. Total lipid extractions of pulp and seed were reflective of free avocadyne and avocadene in dried pulp and seed thus levels below 0.5 mg/g DW were expected. Assuming the dried weight of a typical Hass avocado is 30 g for the pulp and 10 g for seed (Haas, 1951; S. K. Lee et al., 1983), Table 2.2 also shows extrapolated amounts of avocadyne and avocadene in the pulp and seeds of one Hass avocado. As expected, higher PFA levels were determined in the pulp (5.25 ±1.15 mg avocadyne and 6.65 ±1.25 mg avocadene) compared to the seed given the larger expected average dry mass of the pulp.

A . B . T o ta l L ip id E x tra c tio n  A v o c a d y n e T o ta l lip id e x tr a c tio n  A v o c a d e n e **** 0 .5 0 **** 0 .5 0 * 0 .4 5 **** 0 .4 5 0 .4 0 0 .4 0 2 4 h 2 4 h 0 .3 5 0 .3 5 4 8 h 4 8 h

W 0 .3 0 W 0 .3 0

7 2 h 7 2 h

D g

g 0 .2 5 0 .2 5

/ g g

0 .2 0 0 .2 0

m m 0 .1 5 0 .1 5 0 .1 0 0 .1 0 0 .0 5 0 .0 5 0 .0 0 0 .0 0

p d lp d l e e u u e e P P S S Figure 2.3. Hass avocado pulp and seed total lipid extraction. Avocadyne (A) and avocadene (B) content expressed in mg per g dried weight (DW) of Hass avocado pulp or seed material subject to total lipid extraction with varying exposures to extraction solvent (24 h-72 h). Each box plot represents means± standard deviation of two biological replicates * = p <0.05, **** = p <0.0001.

Table 2.2. Projected Amounts of Avocadene and Avocadyne in Total Lipid Extract of Seed and Pulp of a Hass Avocadoa

avocadyne avocadene mean amount projected amount in mean amount projected amount in per gram dry average dry weight of one per gram dry average dry weight of one weight Hass avocado weight Hass avocado (mg/g ± S.D). (mg ± S.D) (mg/g ± S.D). (mg ± S.D) pulp 0.18 ± 0.04b 5.25 ± 1.15 0.22 ± 0.04b 6.65 ± 1.25 seed 0.41 ± 0.02c 4.10 ± 0.23 0.43 ± 0.04c 4.25 ± 0.44 aMean amount per gram dry weight ± standard deviation of two biological replicates representative of Figure 3 from the 72 h solvent maceration condition. Means with different letters b, c within same column are significantly different (p < 0.0001).

2.4.4. Methanolic Saponification of Avocado Pulp and Seed

Methanolic saponification of the total lipid content of the seed and pulp matter or directly on solid seed and pulp matter was carried out to test if avocadene and avocadyne could be released from all bound forms (complex ester linkages). Figure 2.4 shows that saponification of total lipid

extracts or direct saponification of solid pulp and seed material increases the yield of total avocadyne and avocadene by greater than ten-fold compared to amounts quantified in total lipid extracts only (Table 2.2). No significant differences were seen in the total amounts of PFAs recovered with any of the three methods of saponification. Table 2.3 summarizes results only from overnight saponification of total lipids with low KOH extractions. Avocadyne was detected in the pulp and seeds at 1.97 ± 0.55 mg/g DW and 6.04 ± 3.02 mg/g DW (p < 0.01), respectively.

Similarly, 3.92 ± 1.75 mg/g DW and 7.09 ± 3.17 mg/g DW of avocadene was recovered in the pulp and seeds. Overall, Table 2.3 shows that far greater amounts of avocadene and avocadyne can be released if complex ester linkages in avocado pulp and seed can be saponified to release the free fatty alcohols present. With solvent-based saponification, 59.05 ± 16.40 mg and 60.35 ±

30.23 mg of avocadyne can be expected in the dry weight of one Hass avocado pulp and seed, respectively. Similarly, with solvent saponification, 117.50 ± 52.53 mg and 70.92 ± 31.69 mg of avocadene can be anticipated in the dry weight of one Hass avocado pulp and seed, respectively.

It is important to note that patent and literature reports have previously outlined methods of alkaline saponification of cold-pressed avocado pulp oil that not only yield polyhydroxylated fatty alcohols but mainly avocado alkyl furans and other avocado unsaponifiables (Patent No. 6,582,688

B1, 2003; Patent No. 9,416,333, 2014; Farines et al., 1995). Saponification methods utilized in this study are specific to recovering avocadyne and avocadene only and did not allow for total extraction and characterization of avocado unsaponifiable substances.

A . A v o c a d y n e B . A v o c a d e n e

* 1 2 1 2 ** ** 1 0 1 0

O v e rn ig h t s a p o n ific a tio n o f to ta l O v e rn ig h t s a p o n ific a tio n o f to ta l 8 8

lip id s  lo w K O H lip id s  lo w K O H

W W

D O v e rn ig h t s a p o n ific a tio n o f to ta l

D O v e rn ig h t s a p o n ific a tio n o f g

g 6 6 /

/ to ta l lip id s  h ig h K O H lip id s  h ig h K O H

g m m D ire c t s a p o n ific a tio n D ire c t s a p o n ific a tio n 4 4

2 2

0 0

p d p d l e l e u e u e P S P S

Figure 2.4. Methanolic potassium hydroxide saponification of Hass avocado pulp and seed. Avocadyne (A) and avocadene (B) content expressed in mg per g dried weight (DW) of Hass avocado pulp or seed material subject to three outlined saponification methods. Each box plot represents means± standard deviation of two biological replicates * = p < 0.05, ** = p < 0.01, **** = p < 0.0001.

Table 2.3. Projected amounts of avocadene and avocadyne in saponified extract of seed and pulp of a hass avocadoa

avocadyne avocadene

mean amount projected amount in mean amount projected amount in per gram dry average dry weight of one per gram dry average dry weight of one weight Hass avocado weight Hass avocado (mg/g ± S.D.) (mg ± S.D.) (mg/g ± S.D.) (mg ± S.D.)

pulp 1.97 ± 0.55b 59.05 ± 16.40 3.92 ± 1.75 117.50 ± 52.53

seed 6.04 ± 3.02c 60.35 ± 30.23 7.09 ± 3.17 70.92 ± 31.69

aMean amount per gram dry weight ± standard deviation of two biological replicates representative of Figure 4 from the overnight saponification of total lipids—low KOH condition. Means with different letters b, c within same column are significantly different (p < 0.05).

2.5. Conclusions

Avocadene and avocadyne are bioactive compounds with physiological benefits and were

first identified in avocado seed. Although their quantity in avocado pulp has not been reported

consistently, their known presence in the pits provided an ideal source to test the proposed analytical method. As such, the developed LC-MS method was successfully applied to lipid extracts of Hass avocado seed and pulp matter. An established total lipid extraction method, adopted from Folch and colleagues (Folch et al., 1957) was used in tandem with methanolic saponification to extract and quantify the target PFAs. The amount of avocadene and avocadyne

(in mg/g DW) quantified in total lipid extracts were two-fold higher in the seed compared to the pulp, a finding well supported in literature (Patent No. 20160249613A1, 2013).

The more than ten-fold increased recovery of avocadyne and avocadene after methanolic saponification of either total lipid extracts or original starting material further supports previous evidence that fatty alcohols largely exist naturally as compounds bound to complex ester linkages

(Mudge, 2005; Mudge et al., 2012). While more avocatin B was found in the seeds, the pulp still contained appreciable amounts suggesting that dietary avocado consumption results in avocatin B intake (Hargrove et al., 2004; R. F. Lee et al., 1970; Tande et al., 2016) however, future studies will be needed to directly quantify the bioavailability of avocatin B from the ingestion of avocados.

In summary, this study presents a simple, sensitive and reproducible LC-MS method of quantifying avocatin B and can be used to evaluate other methods of extracting polyhydroxylated fatty alcohols available in the literature (Rodríguez-López et al., 2015; Rosenblat et al., 2011;

Patent No. 20150175933, 2015; Silva-Platas et al., 2012). Future studies can also utilize the described analytical method to further explore differences in the amounts of avocadene and avocadyne present in different avocado cultivars.

2.6. Acknowledgments

The authors acknowledge the support of Dr. Tarek Mohamed and Dr. Praveen N. Rao

(University Waterloo School of Pharmacy) for mentoring and planning initial extraction and method development protocols.

2.7. Author contributions

N.A. and P.A.S. conceptualized the experimental design, analyzed data and wrote the manuscript. R.W.S. supervised N.A. with LC-MS method development and validation. J.A.H and K.D.S. assisted with the experimental design of saponification experiments and edited the manuscript.

2.8. Supporting information (Chapter 2) A. Supplementary Figure 2.1. precursor ion MS/MS fragments elemental relative selected composition intensity (%) Representative MS/MS + spectra for avocadene avocadene [M+H] 233.22577 C17H29 100.00 251.23623 C17H31O 33.31 (A) Mass spectra of major 269.24674 C17H33O2 66.85 MS/MS fragments generated 287.25723 C17H35O3 1.79 from Avocadene precursor ion [M+H]+ at normalized collisional energy (NCE) value of 10 after 10 µM avocatin B standard is directly infused into QE-MS in +ESI mode. Table above MS/MS spectra summarizes m/z values of all relevant molecular ions. (B) MS/MS fragment ion intensity traces are displayed for all relevant Avocadene fragments as NCE value is increased from 10- 40 after 10 µM avocatin B standard is directly infused into QE-MS in +ESI mode.

Avocadene [M+H]+

Avocadene [M+H-H O]+ 2

Avocadene [M+H-2H O]+ 2

+ Avocadene [M+H-3H2O]

↑ ↑ ↑ ↑ ↑ ↑ NCE: 10 15 20 25 30 40

Supplementary Figure 2.2. Representative MS/MS spectra for avocadyne. A. precursor ion MS/MS fragments elemental relative selected composition intensity (%)

+ (A) Mass spectra of major MS/MS avocadyne [M+H] 231.21067 C17H27 9.36 fragments generated from 249.22117 C17H29O 13.37 Avocadyne precursor ion [M+H]+ at 267.23181 C17H31O2 8.55 285.24228 C17H33O3 1.39 normalized collisional energy (NCE) value of 10 after 10 µM avocatin B standard is directly infused into QE- MS in +ESI mode. Table above MS/MS spectra summarizes m/z values of all relevant molecular ions. (B) MS/MS fragment ion intensity traces are displayed for all relevant Avocadyne fragments as NCE value is increased from 10-40 after 10 µM avocatin B standard is directly infused into QE-MS in +ESI mode.

Avocadyne [M+H]+

+ Avocadyne [M+H-H2O]

+ Avocadyne [M+H-2H2O]

+ Avocadyne [M+H-3H2O]

↑ ↑ ↑ ↑ ↑ ↑ NCE: 10 15 20 25 30 40

Supplementary Figure 2.3. MS spectrum for AVO from LC-MS method. (A) Avocadyne [M+H]+ and (B) Avocadene [M+H]+extracted ion chromatograms at retention time 5.65 min and 7.94 min, respectively for 100 µM Avocatin B standard injected into described LC-MS method.

Supplementary Figure 2.4. Representative LC-SIM-MS chromatographs of AVO. 5 nM avocatin B standard (0.001 ng/µL of avocadene or avocadyne) injected in described LC-SIM-MS analytical method, employing an extraction window of 10 mDa.

Supplementary Table 2.1. Accuracy and precision for the analysis of avocadene and avocadyne in solvent

nominal accuracy (% error) precision (%, CV) concentration (µM) intra-assay inter-assay intra-assay inter-assay

avocadene avocadyne avocadene avocadyne avocadene avocadyne avocadene avocadyne

0.1 (LLOQ) 16.3 16.7 18.2 13.3 13.6 13.1 13.8 14.4

5 7.3 9.3 4.5 4.7 5.3 7.5 4.6 7.0

50 1.8 -0.5 1.2 0.1 2.4 1.2 1.8 1.2

Supplementary Table 2.2. Recovery of avocadene and avocadyne from hass avocado pulp and seed total lipid extracts

avocadene avocadyne

concentration in concentration % concentration in concentration % total lipid extract unspiked in spiked recovery unspiked in spiked recovery sample (µM) sample (µM) sample (µM) sample (µM)

pulp extract 1 0.895 6.575 113.6 0.542 5.692 103 pulp extract 2 1.031 6.001 99.4 0.786 5.716 98.6 seed extract 1 2.925 8.245 106.4 1.321 6.281 99.2 seed extract 2 3.012 8.132 102.4 1.578 6.768 103.8

Chapter 3: Bioaccessability and bioavailability of avocado polyhydroxylated fatty alcohols

1 2 1 1 1 Nawaz Ahmed , Richard W. Smith , Alessia Roma , Peter X. Chen , Michael A. Rogers , Paul A. Spagnuolo1* 1 Department of Food Science, University of Guelph, Guelph, Ontario, N1G 2WI, Canada 2 University of Waterloo Mass Spectrometry Facility, Department of Chemistry, 200 University Avenue West, Waterloo, ON, N2L 3G1, Canada

*To whom correspondence should be addressed. Email: [email protected]

Manuscript in preparation

3.1. Abstract

Avocado derived polyhydroxylated fatty alcohols, avocadene and avocadyne, have recently been identified as potent mitochondrial metabolic modulators that selectively induce leukemia cell death and reverse pathologies associated with diet-induced obesity. The absorption of avocadene and avocadyne from avocado pulp consumption in humans has not been reported.

The purpose of this study was to investigate if Hass avocado pulp digestion in vitro can lead to relevant bioaccessibility of avocadene and avocadyne. Static and dynamic (TNO dynamic intestinal model-1 (TIM-1)) in vitro digestion was performed on lyophilized Hass avocado pulp powder that contained avocadene and avocadyne in bound forms sensitive to sodium methoxide saponification. Results from static and dynamic digestion on avocado pulp powder revealed that the action of lipolytic gastrointestinal enzymes led to appreciable bioaccessibility of avocadene and avocadyne. Furthermore, TIM-1 digestion of a 1:1 ratio of pure avocadene and avocadyne

(avocatin B or AVO) formulated in an oil-in-water microemulsion resulted in moderately higher bioaccessibility compared to the avocado pulp powder highlighting avocado pulp to be a relevant dietary source of these bioactive molecules. AVO emulsion delivered orally to C57BL/6J mice showed noticeable absorption in blood and key target tissues in a pilot in vivo study. The results of this study provide impetus for further research on the nutritional significance of dietary long chain fatty alcohols often considered to be biologically inert.

3.2. Introduction

Avocados are recognized as a therapeutic dietary component for humans suffering from chronic metabolic diseases like obesity, type 2 diabetes (T2D), dyslipidemia, hypertension and cardiovascular disease (Dreher & Davenport, 2013; Tabeshpour et al., 2017). Oil constitutes up to

60% of the dry weight of an avocado pulp thus the majority of pre-clinical and clinical studies to date have highlighted mono- and polyunsaturated fatty acids, phytosterols, carotenoids and fat- soluble vitamins as the major bioactive constituents in avocado that exert beneficial health effects

(Khan et al., 2019; Park et al., 2018; Tabeshpour et al., 2017; Zhu et al., 2019). However, the biological activity and nutritional significance of avocado pulp unsaponifiable constituents

(fraction of the oil that remains insoluble in water after prolonged action of alkaline base) like polyhydroxylated fatty alcohols (PFAs) and furan lipids are less well studied (Farines et al., 1995;

Kashman et al., 1969b).

Avocado PFAs like avocadene and avocadyne have been shown to inhibit mitochondrial fatty acid oxidation (FAO) which selectively induces cell death in leukemia stem cells while sparing healthy cells (E. A. Lee et al., 2015). Similarly, avocadene and avocadyne have been reported to inhibit excess FAO in skeletal muscle and pancreatic β-cells which reverses obesity associated pathologies (article in press). Given the significant implications of these PFAs for human nutrition and health, studies are required to investigate their bioaccessability and bioavailability from avocado pulp and oil consumption. Previously, we reported the combined recovery of close to 200 mg of avocadene and avocadyne (up to 0.7% of the dry weight) after saponification of one dried Hass avocado pulp with methanolic sodium hydroxide (Ahmed et al.,

2018). Patent literature has similarly outlined the use of saponification on crude, cold pressed avocado pulp oil to yield fractions enriched in PFAs and furan lipids for application in cosmetic products (Patent No. 6,582,688 B1, 2003; Patent No. 7,371,420 B2, 2008). Traditionally however, commercially available avocado oils are void of PFAs and other unsaponifiable constituents as cold crystallization and winterization is performed on the crude oil to meet quality standards of refined oils (Flores et al., 2019; Patent No. 7,81,6547 B2, 2010). To date, PFAs like avocadene

and avocadyne have only been reported to exist as free fatty alcohols or as monoacetates in the unsaponifiable portion of avocado pulp oil and seeds (Farines et al., 1995). The possibility that avocadene and avocadyne may exist in more complex bound forms likes wax esters has never been proposed.

Classically, waxes are defined as esters formed between long chain fatty alcohols and long chain fatty acids which are abundant in typical human diets that comprise of cereal grains, germs, bran, leaves, seeds, nuts and unrefined oils (Hargrove et al., 2004). The bioaccessability (quantity of an ingested compound available for systemic absorption after digestion) of free fatty alcohols and fatty acids from dietary wax consumption in humans is generally assumed to be low due to the lower capacity for wax ester hydrolysis in humans compared to seabirds and chickens

(Hargrove et al., 2004; Place, 1992). Wax hydrolysis in seabirds, chickens and pigs has been shown to be dependent on bile salt-activated lipase (or carboxyl ester lipase) and colipase (Place, 1992).

Additionally, presence of dietary fats has been postulated to inhibit wax hydrolysis in the mammalian intestinal milieu as triacylglycerol hydrolysis occurs at 10-50 times the rate of wax hydrolysis (Hargrove et al., 2004; Savary, 1971). More recent in vitro evidence suggests that wax esters characterized by long chain polyunsaturated fatty acid moieties are more efficiently hydrolyzed by porcine pancreatic lipase (equivalent to human pancreatic lipase) compared to waxes with highly saturated moieties (Capozzoli et al., 2018). This finding is well supported by studies reporting the bioaccessability, bioavailability and bioactivity of wax ester bound n-3 (ω-3) polyunsaturated fatty acids (PUFA), eicosapentaenoic acid (EPA) and dacosahexaenoic acid

(DHA), in oil extracted from marine copepod Calanus finmarchicus (Cook et al., 2016; Eilertsen et al., 2012). Furthermore, the nutritional significance of ingesting long chain free fatty alcohols like octacosanol (C28H58OH), most abundant in sugarcane wax, has resurfaced (Sharma et al.,

2019) following initial contradictory reports of its cholesterol-lowering effects in humans (Gouni-

Berthold & Berthold, 2002; Lin et al., 2004).

Here, we utilize static and dynamic in vitro digestion models to report for the first time that avocado pulp PFAs, avocadene and avocadyne, exist in bound forms that are bioaccessible after gastric and pancreatic digestion. We also report the enhanced in vitro bioaccessability of an oil-in- water (O/W) microemulsion comprising of avocatin B (AVO; 1:1 ratio of avocadene—avocadyne) and confirm its in vivo bioavailability (fraction of an ingested compound that reaches systemic circulation) and biodistribution in a pilot pharmacokinetic study in mice. The results of this study provide strong rationale for further research into the full analytical characterization of avocado pulp wax components which are proposed to contain important bioactive molecules like avocadene and avocadyne.

3.3. Experimental Section

3.3.1. Materials

Chemicals and reagents purchased from Sigma Chemical Co. (MO, USA) were: pepsin

(from porcine pancreas, P7012), pancreatin (from porcine pancreas, P1750), pancreatic lipase

(from porcine pancreas, L3126), colipase (from porcine pancreas, C3028), amylase (A6380), trypsin (T1426), deoxycholic acid (D2510), hydroxypropylmethylcellulose (HPMC; H7509), polysorbate 80 (59924). Cell culture grade phosphate buffered saline (PBS) was purchased from

Fisher Scientific (ON, Canada; SH3025601LR). Neobee®M-5 was provided as a gift from Stepan

Company (IL, USA). Lyophilized avocado pulp powder was purchased from New Zealand

Avocado Oil Ltd (bath number 0115). 1% milk was purchased from Natrel, Canada. Fresh porcine bile was collected from Conestoga meat packers (Kitchener, Ontario, Canada). Avocatin B was purchased and used as received from Microsource Discovery Systems Inc. (CT, USA).

3.3.2. Determining avocadene and avocadyne concentration in lipid extracts of lyophilized avocado pulp powder

The amount of free avocadene and avocadyne in lyophilized avocado pulp powder was determined as previously described (Ahmed et al., 2018). Briefly, 1g of the lyophilized avocado powder was macerated in 20 mL of 2:1 choloroform—methanol for 72 h in ambient conditions. After maceration, 3 mL of 0.2 M sodium-phosphate (NaHPO4) buffer in ddH2O (pH 4.4) were added to induce layer separation. After inversion, samples were centrifuged for 5 min at 1500 rcf. The total lipid containing organic layer was collected and an additional 9 mL chloroform was added to the aqueous layer as an additional round of extraction. The second organic layer was combined with the first and the buffer layer was discarded. Chloroform extracts were rotary evaporated and stored at 4 °C until sample preparation was required for LC-MS analysis.

To determine the amount of bound avocadene and avocadyne in the avocado powder, methanolic sodium hydroxide saponification was carried out as previously described (Ahmed et al., 2018). Briefly, total lipids were extracted from avocado powder, as outlined above, after which

50 mL of methanol containing 0.8% w/v potassium hydroxide was added and allowed to shake at a moderate speed for 16-24 h at 37 °C. After saponification, the methanolic potassium hydroxide mixture was neutralized with 6 M HCl solution to protonate fatty acid salts and any free fatty alcohols that may have been deprotonated in the saponification. Methanol was rotary evaporated and dry samples were stored at 4 °C until sample preparation was required for LC-MS analysis.

3.3.3. Static in vitro digestion

Lyophilized avocado powder was subject to static in vitro digestion experiments simulating gastric and intestinal conditions using a standard protocol (Minekus et al., 2014). Simulated

digestion fluids were prepared as described (Minekus et al., 2014) with the exceptions that phospholipids were not included in the simulated gastric fluid (SGF) and simulated intestinal fluids

(SIF) comprised of 1000 U/mL lipase, 1 mg/mL colipase and 17.5 mmol/L deoxycholic acid as a substitute for bile. The digestions were performed in glass jars wrapped in aluminum foil and placed in a 37 °C incubator set to shake at 200 rpm (MaxQ 4450, Barnstead/Lab-Line, IL, USA).

1g avocado powder was mixed thoroughly with 5 mL SGF which contained 2,000 U/mL pepsin.

Gastric digestion was performed for 2 h at pH 3.0. Ten mL of SIF were then added to start the intestinal phase of digestion. The pH of the reaction mixture was adjusted to 7.0 by addition of 1 mmol/L sodium hydroxide and intestinal digestion was carried out for 4 h. After completion of digestion, total lipids were extracted from digestion mixture by the addition of 100 mL 2:1 chloroform—methanol in a separatory funnel which was inverted 5-10 times. Bottom organic layer was collected after which 65 mL of chloroform was added to the aqueous digestate layer and a double extraction was performed. Organic layer was rotary evaporated and dry samples were placed at 4 ºC until LC-MS analysis. Static digestion of avocado powder was performed in both gastric and intestinal conditions (G+I), gastric conditions only (G), intestinal conditions only (I), gastric pH control (control:G), intestinal pH control (control:I), and gastric and intestinal pH control (control: G+I). Experiments were performed in triplicate.

3.3.4. Avocatin B emulsion preparation and characterization

Self-emulsifying drug delivery systems (SEDDS) (Buyukozturk et al., 2010; Callender et al., 2017; Nardin & Köllner, 2018) that spontaneously form oil-in-water (O/W) emulsions into which AVO could be easily incorporated were tested. Formulations were chosen based on their self-emulsification properties, small droplet size (below 300 nm), low polydispersity index (PDI;

below 0.4), and no requirement for co-surfactants. A 1:1 mixture of NeoBee®M5 (a commercially available medium chain triglyceride oil) and polysorbate 80 (Tween 80) was the chosen oil- surfactant combination. 1-20 mg of avocatin B was added to 100 µL of a 1:1 mixture of

NeoBee®M5—Tween 80 and heated to 75 °C for 2 hr in an incubator (MaxQ 4450,

Barnstead/Lab-Line, IL, USA). After dissolution of avocatin B in the oil phase, 900 µL water phase (PBS) was added to the oil phase and vortexed for 30 s. Visual appearance of avocatin B emulsions were noted followed by droplet size and PDI measurements on day zero (immediate measurement on fresh emulsion), day 3 and weekly for 4 weeks. The mean droplet size (Z-average in nm) and polydispersity index (PDI) for each formulation was determined using dynamic light scattering (Zetasizer Nano ZS, Malvern Instruments, Malvern, U.K.). A refractive index of 1.33 was used for the aqueous phase, and refractive indices for each oil/surfactant mix was determined using a table top refractometer (Zeiss Abbe, NY, USA). Droplet size was measured on three independently prepared formulations and averaged from three readings per formulation. All measurements were carried out at 25 °C.

3.3.5. TIM-1 Studies

3.3.5.1. TIM-1 protocol

TIM-1 simulates digestion in the upper gastrointestinal tract of an adult human (Minekus et al., 1995). TIM-1 includes four compartments: stomach, duodenum, jejunum, and ileum. Prior to introducing the meal into the TIM-1, a small intestinal electrolyte solution (SIES: 132 mM NaCl,

5.28 mM KCl, 0.75 mM CaCl2), a 7% (w/v) pancreatin solution, a duodenal start residue (15 g

SIES, 30 g fresh 20% porcine bile solution, 2 mg trypsin, 15 g of 7% pancreatin solution), a jejunal start residue (35 g SIES, 70 g fresh 20% porcine bile solution, 35 g of 7% pancreatin solution),

and an ileal start residue (140 g SIES) were injected into their respective compartments and the system was heated to 37 °C. The initial amount of gastric juice was simulated by loading the gastric compartment with 5 g (fed protocol) or 15 g (fasted protocol) of gastric enzyme solution (4800

U/mL pepsin, 20 U/mL lipase, and 47 U/mL amylase in gastric electrolyte solution: 150 mM NaCl,

20 mM KCl, 1 mM CaCl2, 10 mM sodium acetate buffer) and 5 g (fed protocol) or 15 g (fasted protocol) of 0.4% HPMC/0.04% bile solution. The rate of secretion of digestive juices, peristaltic movements, nutrient and water absorption, gastric emptying, pH, and transit times in each compartment was set in accordance to manufacturer’s protocol (Triskelion/TIM021 and 040) for fed state-lipid digestion or a fasted state-lipid digestion. Simulated digestive fluids were prepared fresh on the day of the experiment, and enzyme solutions were stored between 0-5 ℃ before use.

For fed state digestions the pH of the stomach compartment was set up to evolve with time: 6.5 at

0 min, 4.2 at 30 min, 2.9 at 60 min, 2.0 at 120 min, 1.7 between 210-360 min. For fasted state digestion the pH of the stomach was pre-set as follows: 3.0 at 0 min, 2.2 at 10 min, 1.8 at 30 min, and 1.7 between 60-360 min. For both fed and fasted state digestions the pH in all other compartments was kept constant (5.9, 6.5, and 7.4 for the duodenum, jejunum and ileum, respectively). Meals were administered to TIM-1 and samples were collected from the jejunal filtrates, ileal filtrates and ileal efflux (effluent) at 60, 120, 180, 240, 300 and 360 min after which the TIM runs were terminated, and samples were solvent extracted the same day. A water control was also subjected to TIM-1 digestion (fed protocol) where dialysates were sampled at the same time points and used to generate matrix matched calibration curves and perform analyte recovery experiments for LC-MS method validation and analysis. See supplementary information section for sample extraction and analytical method development and validation information. All TIM-1 digestion experiments were completed in triplicate.

3.3.5.2. Avocado powder milk beverage preparation and administration

Lyophilized avocado powder was blended in 1% milk to create a beverage that would be administered to the TNO gastrointestinal model (TIM-1, Zeist, The Netherlands) using the fed state-lipid digestion protocol. Briefly, 22.44 g of lyophilized avocado powder (determined to contain up to 200 mg of a 3:2 mixture of avocadene—avocadyne) was mixed (using a Magic

Bullet® blender, MBR-1701) in 355 mL of 1% milk until a consistent suspension formed. 209 g of the beverage was mixed with 81 g of start solution (75 g gastric electrolyte solution, 5.7 g HPMC and 11 mg amylase) and added to the stomach compartment through a funnel.

3.3.5.3. AVO O/W emulsion preparation and administration

An emulsion containing 2% (w/w) AVO was prepared in deionized water as described above and administered to TIM-1 for digestion using the fasted state-lipid digestion protocol.

Avocatin B emulsion was added to 270 g of start solution (251 g gastric electrolyte solution, 18.9 g HPMC, 11 mg amylase) such that a total of 100 mg of avocatin B was added to the stomach compartment. The particle size of the emulsion in the start solution was measured, as described above, prior to administration to TIM-1.

3.3.5.4. Determination of bioaccessibility

Cumulative bioaccesibility was the accumulation of avocadene or avocadyne quantified during each time interval in the jejunum, ileum and the sum of the two filtrates, expressed as a percent of the input amount (absolute fatty alcohol content). The amounts quantified in the effluent were also expressed as a percent of the input amount but excluded from the bioaccessibility

analysis. Non-cumulative, absolute concentrations of avocadene and avocadyne was plotted against time to perform area under curve (AUC) analysis on GraphPad Prism 6.0 software (CA,

USA).

Mathematical modelling was also applied on all TIM-1 cumulative data which displayed a sigmoidal pattern (initial lag phase followed by a steep increase in fatty alcohol bioaccessibility that approaches a plateau). A three-parameter shifted logistic model commonly used to characterize free fatty acids generated as a result of lipolytic activity in the gastrointestinal tract as a function of time (t) (Fondaco et al., 2015; Speranza et al., 2013; Troncoso et al., 2012):

퐶푎푠푦푚푝 퐶푎푠푦푚푝 퐶(푡) = − 1 + 푒[푘(푡퐶 − 푡)] 1 + 푒[푘푡퐶]

Here, Casymp is the total amount of fatty alcohol released (asymptotic level at the plateau region), k is the rate of fatty alcohol released per unit time, and tC is the critical time (induction time) at which half the total amount of fatty alcohol released is achieved. Nonlinear analysis was performed in

Graphpad Prism 6.0 software for each compartment on the three bioaccessibility parameters from the equation above (Casymp, k, and tC).

3.3.6. In vivo pilot pharmacokinetic study

Ten-week-old, female, C57BL/6J mice were purchased (Jackson Laboratory, Bar Harbor,

ME) and allowed to acclimatize for 1 week. After acclimatization, mice were randomly assigned

(n=3 per treatment group) to receive an oral bolus dose of 100 mg/kg body weight (b.w.) avocatin

B (formulated as 1% w/w SEDDS) or vehicle control (control SEDDS). The gavage volume was

5 mL/kg b.w. After 2 hr and 6 hr post gavage, up to 100 µL of whole blood was drawn per animal via tail bleed and collected in K2EDTA coated tubes (Sarstedt, Canada). At endpoint (24 hr post

gavage), animals were euthanized via CO2 followed by exsanguination from which 500-800 µL of whole blood was collected and stored as described above. Tissue (inguinal fat pad, gonadal fat pad, liver, pancreas, heart, femural bone marrow, and brain) was harvested and flash frozen in liquid N2. All blood and tissue samples were then maintained at -80 °C until extraction was performed for the analytical determination of avocadene and avocadyne in whole blood and tissues. The bioanalytical, quantitative LC-MS method and its validation is detailed in supplementary information section of this chapter. Results for avocadene or avocadyne quantitation are presented as mean ± S.D., in µg/ml for whole blood or in ng/g wet tissue for tissues, respectively.

Pharmacokinetic parameters were calculated using non-compartmental analysis with the

PK Functions (Joel I. Usansky, PhD, Atul Desai, MS and Diane Tang-Liu, PhD, Department of

Pharmacokinetics and Drug Metabolism, Allergan, Irvine, CA 92606, USA) add-in for Microsoft®

Excel. The total area under the curve (AUC0−t) was determined with the linear trapezoidal rule from the time of dosing while AUC0−inf was extrapolated to infinity. The elimination rate constant (kel) and plasma concentration half life (t½) were determined by regression of the two terminal data points on the semi-logarithmic concentration versus time plot. The maximal concentration (Cmax) and time at maximal (Tmax) were obtained directly from the concentration versus time plot.

3.3.7. Statistical data analysis

Unless otherwise stated, all in vitro digestion data is presented as mean ± SEM, droplet size and PDI measurements for AVO emulsions are presented as mean ± SEM, and results from in vivo pharmacokinetics study are presented as mean ± SD. Data were analyzed with GraphPad

Prism 6.0 using one or two-way ANOVA with Sidak’s post hoc analysis for between group comparisons. P < 0.05 was accepted as being statistically significant. Student’s t-tests were also used where appropriate and as indicated in figure legends.

3.4. Results

3.4.1. Avocadene and avocadyne are hydrolyzed from their bound forms after static and dynamic in vitro digestion of lyophilized avocado pulp powder

Lyophilized Hass avocado pulp powder was chosen as the ideal material to effectively study the bioaccessibility of avocadene and avocadyne (see Supplementary Table 3.1 for manufacturer provided nutritional information of this avocado pulp powder). The lyophilization process is carried out under low temperature and pressure which removes up to 95% of water content in the pulp, while retaining all macro-and- micronutrients as well as volatile materials responsible for the aroma, flavor, and the rigid structure of avocado pulp (Patent No. 7,678,396

B2, 2010; Patent No. 2011 009 1616A1, 2011). Most importantly, the homogenous nature of the lyophilized powder allowed for precise quantification of absolute amounts of avocadene and avocadyne via solvent extraction. Table 3.1 illustrates that close to 70% of avocadene and avocadyne content of this avocado powder were bound and released as free fatty alcohols only after methanolic sodium hydroxide saponification. In our previous report, where Hass avocado pulp were cut into 2-5 mm thick slices and oven dried, up to 90% of avocadene and avocadyne were quantified as bound (Ahmed et al., 2018) which also highlights the greater solvent extraction efficiency achievable with lyophilized avocado pulp powder.

Table 3.1. Amount avocadene and avocadyne quantified before and after saponification of one- gram lyophilized avocado pulp powder total lipid content

Non-saponified: free fatty alcohol content Saponified: absolute fatty alcohol content (mg) (mg) Avocadene Avocadyne Avocadene Avocadyne 2.0 ± 0.06 1.1 ± 0.01 5.55 ± 0.03 3.35 ± 0.05 Values represent mean (± S.D.), n=3

Static in vitro digestion was carried out on this lyophilized avocado powder to determine whether primary intestinal lipid digestion enzymes like lipase and colipase can hydrolyze bound forms of avocadene and avocadyne similar to methanolic sodium hydroxide saponification. This experiment revealed that the action of pancreatic lipase and colipase alone (I) for 4 h released almost 80% of the absolute amounts of avocadene and avocadyne in the avocado pulp powder

(Figure 3.1A). This value decreased by 20-30% when the powder was subject to both gastric and intestinal digestion (G+I) likely due to the highly acidic nature of gastric digestion which may have deprotonated the free fatty alcohols (unbound PFAs) that could not be quantified in the LC-

MS method. Blank digestions (no enzymes) of the avocado powder in gastric, intestinal or gastric and intestinal pH conditions alone did not release significant amounts of avocadene and avocadyne from the avocado powder (Figure 3.1B) further confirming that the hydrolysis observed was a result of lipase and colipase action.

o t

1 1 0

e A b s o lu te A m o u n t (S a p o n ifie d ) v 1 0 0

t c

a 9 0 c F re e F a tty a lc o h o l l

e n

r 8 0

u d G

e o l 7 0

b d m

i I a

s 6 0

s e

t 5 0 G + I

u l

c a a a

o 4 0

o i

b 3 0

b a b b

A 2 0 F 1 0

0 % e e n n y e d d a a c c o o v v A A

B o

e 1 0 0 A b s o lu te A m o u n t (S a p o n ifie d )

v i t 8 0 a p H C o n tro l ( G )

e n

r 6 0

u p H C o n tro l (I )

e o

l 4 0

b m

i p H C o n tro l ( G + I ) a

s 3 0

e e

t 2 5

u l

c a o 2 0

s a

o a i

b 1 5 b a a a a 1 0 A a F 5

0 % e e n n y e d d a a c c o o v v A A Figure 3.1. Avocadyne and avocadene are hydrolyzed from their bound forms after static in vitro digestion of lyophilized avocado powder. (A) 1 g lyophilized avocado powder was subject to gastric (G), intestinal (I), or gastric and intestinal (G+I) digestion after which digestates were subject to lipid extraction. Data represents percent avocadene or avocadyne quantified relative to absolute amounts determined in avocado powder via methanolic sodium hydroxide saponification. (B) Avocadene and avocadyne were quantified after 1 g lyophilized avocado powder was mixed and incubated at appropriate gastric, intestinal or gastric and intestinal pH controls. For (A) and (B) data represents three biological replicates where different letter alphabets indicate significant differences between conditions, p<0.0001, two-way ANOVA, Sidak’s post hoc test.

Further to static digestion experiments, dynamic TIM-1 digestion experiments were performed on the avocado powder to determine the bioaccessibility of avocadene and avocadyne.

Bioaccessibility in this context was defined as the portion of the hydrolyzed PFA that become solubilized in the aqueous phase thus are expected to be available for absorption through the gut wall (Minekus et al., 1995). Given the poor water solubility of the lyophilized avocado powder,

1% milk was used to suspend the avocado powder which was administered to TIM-1 as a beverage in the fed-lipid digestion protocol run for 6 h. Amounts of avocadene and avocadyne extracted from the jejunum dialysate, ileum dialysate and ileum effluent were used to determine cumulative and non-cumulative bioaccessibility. Cumulative bioaccessibility of avocadene and avocadyne in the jejunum compartment was determined to be 33 ± 0.5% and 36 ± 1.0%, respectively (Figure

3.2A), in comparison to which only up to 20% of avocadene and avocadyne were bioaccessible in the ileum compartment (Figure 3.2B). From the sum of jejunal and ileal compartments, the cumulative bioaccessibility of avocadyne (50.1 ± 1.4%) was slightly greater than avocadene (55.2

± 2.1%) at 360 min (Figure 3.2C). Close to 10% avocadene and avocadyne was quantified in the effluent (data not shown). Overall, cumulative bioaccessibility determinations from TIM-1 were in good agreement with results of static in vitro digestion experiments.

Non-cumulative bioaccessibility analysis performed by plotting absolute concentrations versus time for all TIM-1 compartments provided mechanistic insight on bioaccessibility kinetics of avocadene and avocadyne in the avocado pulp powder. For both avocadene and avocadyne, a gradual increase in bioaccessibility in the first two hours, a slight plateau between two and three hours, followed by a rapid drop in levels was observed (Figure 3.2 D-F). Area under the curve analysis on concentration—time curves revealed almost 1.5 times more avocadene was bioaccessible compared to avocadyne (Table 3.2), a finding in agreement with the naturally

occurring ratio of both analytes in avocado pulp powder. Overall, the non-cumulative analysis here

also confirms that intestinal digestion is essential for avocadene and avocadyne hydrolysis and

bioaccessibility from bound forms when lyophilized avocado powder is consumed.

A B C 1 0 0 1 0 0 1 0 0 A v o c a d e n e 9 0 9 0 9 0

A v o c a d y n e y

i i

) 8 0 i 8 0 8 0

(

s 7 0 )

s 7 0

)

( %

c 6 0

c 6 0 * e

6 0 (

u i i 5 0 5 0

5 0 b

e i

v 4 0

v 4 0 e

v 4 0

* i

a l

i 3 0

j l

3 0 u

i m

m 2 0 2 0 u

2 0 u

C C C 1 0 1 0 1 0 0 0 0 0 6 0 1 2 0 1 8 0 2 4 0 3 0 0 3 6 0 0 6 0 1 2 0 1 8 0 2 4 0 3 0 0 3 6 0 0 6 0 1 2 0 1 8 0 2 4 0 3 0 0 3 6 0 T im e (m in ) T im e (m in ) T im e (m in )

D E F J e ju n u m Ile u m J e ju n u m + Ile u m 1 5 3 0 A v o c a d e n e 5 0 ** A v o c a d y n e 2 5 ** *** 4 0 *** 1 0 2 0 ***

3 0

/ / 1 5 /

g *



  2 0 1 0 5

1 0 5

0 0 0 0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6 T im e (h ) T im e (h ) T im e (h )

Figure 3.2. Cumulative bioaccessibility of avocadene and avocadyne during 360-min dynamic in vitro TIM-1 digestion of lyophilized avocado powder. Twelve gram lyophilized avocado powder (equivalent to one third dry weight of an Hass avocado pulp) was blended in 200 mL of 1% milk and administered to TIM-1. Cumulative bioaccessibility of avocadene and avocadyne in (A) jejunum, (B) ileum, and (C) the sum of jejunum and ileum. Non-cumulative, absolute concentrations of avocadene and avocadyne over time in (D) jejunum, (E) ileum, (F) the sum of jejunum and ileum. For (A-C) data represents mean (±SEM) % cumulative bioaccessibility relative to input amount. For (D-F) data represents mean (±SEM) concentration. For (A-F) *p<0.05, **p<0.01, ***p<0.001 for avocadene vs. avocadyne, two-way ANOVA, Sidak’s post hoc test, n=3.

Table 3.2. Area under curve analysis on non-cumulative concentration—time graphs for jejunum, ileum, and jejunum+ileum for avocado milk beverage administered to TIM-1

Jejunum Ileum Jejunum+Ileum a a a Avocadene 79.3 ± 1.0 38.2 ± 2.2 117.5 ± 3.0 (µg*h/mL) b b b Avocadyne 51.9 ± 1.5 25.9 ± 1.8 77.8 ± 3.0 (µg*h/mL) Different superscript letters within each column indicates significant differences (P < 0.05) between avocadene and avocadyne, two-way ANOVA with Sidak’s post hoc test (n=3).

3.4.2. Avocatin B self-assembles in an O/W microemulsion

A delivery system consisting of pure avocadene and avocadyne (in the form of AVO) was developed to compare its bioaccessibility with what was determined for lyophilized avocado pulp powder. Oil-in-water self-emulsifying systems (also known as SEDDS) are commonly used in food and pharmaceutical formulations (Buyukozturk et al., 2010; Callender et al., 2017; Nardin &

Köllner, 2018). Self-emulsifying systems are complex, lipid based drug delivery systems comprised of a mixture of oils, co-solvents, surfactants and co-surfactants that spontaneously self- assemble into oil-in-water (O/W) or water-in-oil (W/O) nano (~100-400 nm in diameter) or micro

(<100 nm in diameter) emulsions (Mcclements, 2015). We tested a variety of formulations for self-emulsification, droplet size, polydispersity index (PDI), and long-term thermodynamic stability among which a 1:1 combination of a medium chain triglycerides oil (Neobee®M5) and a non-ionic surfactant (Tween 80) was chosen.

A 1:1 Neobee®M5—Tween 80 ratio self-emulsifies into a turbid emulsion with a mean droplet diameter close to 200 nm. AVO self-assembled into O/W emulsions when incorporated into SEDDS. AVO (1-20 mg) was pre-dissolved in 100 µL of the oil/surfactant phase (1:1

Neobee®M5—Tween 80) by heating to 75 °C for 2 hr, after which 900 µL water phase (PBS) was added to the pre-heated oil phase and vortexed for 30 sec (Figure 3.3A). This method of preparation

was necessary to produce consistent emulsions droplet sizes. Other methods, such as directly adding hot oil phase to the water phase under constant stirring, did not produce consistent results.

When 10-20 mg/mL AVO (or 1-2% (w/w)) was incorporated into SEDDS, a significant reduction in turbidity (as characterized by an increase in transparency) and mean droplet size was observed.

Incorporation of 2% (w/w) AVO reduced emulsion droplet diameter to below 25 nm, almost a

90% reduction of the control (blank emulsion) droplet size (Figure 3.3B). Ambient temperature polydispersity (Fig. 3.3C) and droplet size measurements over a course of four weeks revealed no stark differences except at 1.5-2% (w/w) AVO containing emulsions, which showed an increase in droplet size and polydispersity over time compared to the control emulsion. The incorporation of small amounts of AVO into the self-emulsifying system resulted in drastic reductions in droplet diameter without additional energy input or surfactant. This is strongly indicative of AVO self- assembling at the oil-water interface which alters the interface curvature and film flexibility

(Gradzielski, 2002).

Given the stability of AVO O/W emulsions, a 2% (w/w) AVO emulsion was prepared in deionized water and spiked into 270 g of TIM-1 starting solution (a mixture of gastric electrolyte solution, HPMC and amylase) such that 100 mg of AVO would be administered to the TIM.

Particle size distribution analysis of the TIM-1 starting solution revealed a mean droplet diameter of 410.8 ± 0.1 nm (PDI: 0.4 ± 0.1) (Figure 3.3D), whereas the mean droplet diameter of the TIM-

1 starting solution with the spiked AVO emulsion was 25.3 ± 0.01 nm (PDI: 0.1 ± 0.01) (Figure

3.3E). Spiking AVO emulsion into TIM-1 starting solution resulted in a transparent mixture void of any signs of creaming or coalescence. Collectively, a unique self-emulsifying O/W emulsion- based delivery system for AVO was created and subject to TIM-1 digestion for direct comparison with lyophilized avocado pulp powder.

Shake and Heat (2h) Water phase Vortex 30s Emulsion droplet 1:1 oil : surfactant added directly size and + to oil phase polydispersity 1-20 mg avocado polyol (1:10 dilution) measurements

Step 1 Step 2 Step 3 Step 4

® ® A v o c a tin B e m u ls io n (1 :1 N e o B e e M -5 -T w e e n 8 0 ) B A v o c a tin B e m u ls io n (1 :1 N e o B e e M -5  T w e e n 8 0 ) C **** 0 .5 0 * 2 5 0 **** **** **** **** **** **** *** 0 .4 5

) **** **** **** **** **** ) 2 2 5 *** *** ** *** ***

m 0 .4 0 n

. 2 0 0 d

( 0 .3 5 C o n tro l

e 1 7 5 g

a 1 m g /m L 0 .3 0

r 1 5 0

I e

v 5 m g /m L

D 0 .2 5

a P - 1 2 5

1 0 m g /m L

Z (

0 .2 0 1 0 0

e 1 5 m g /m L z i 0 .1 5

S 7 5 2 0 m g /m L

e l

c 0 .1 0

i 5 0

t r

a 0 .0 5

P 2 5

0 0 .0 0

0 3 7 4 1 8 0 3 7 4 1 8 1 2 2 1 2 2 D a y D a y

D E

3 0 2 0

)

t t

n 1 5

c c

2 0 r r

( (

1 0

i s

s n

n 1 0

e t

t 5

n I I 0 0 1 1 0 1 0 0 1 0 0 0 1 0 0 0 0 1 1 0 1 0 0 1 0 0 0 1 0 0 0 0 S iz e (d .n m ) S iz e (d .n m )

Figure 3.3. AVO reduces droplet size of a NeoBee®M5— polysorbate 80 based O/W microemulsion. (A) Emulsion method of preparation. (B) Effect of AVO concentration on average hydrodynamic diameter (Z-average) of NeoBee®—Tween 80 emulsion over time. Inset: visual appearance of control (blank) emulsion and avocatin B containing emulsion on day 0. (C) Polydispersity index of emulsions described in (B). For (B-C) values are means ± SEM of three independent experiments; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 compared to control, two-way ANOVA, Dunnett’s post hoc test. (D) Mean particle size distribution (PSD) of TIM-1 start solution and (E) start solution spiked with AVO emulsion (20 mg/mL) to administer 100 mg to TIM-1. For (D-E) data represents mean PSD (±SEM) from three independent experiments.

3.4.3. Avocadene and avocadyne delivered via AVO emulsion have moderately greater bioaccessibility compared to avocado pulp powder delivery

Cumulative TIM-1 bioaccessibility analysis of AVO emulsion surprisingly revealed that avocadene had significantly greater bioaccessibility than avocadyne in both ileum and jejunum as early as two hours into digestion (Figure 3.4A-C). In the jejunum, the cumulative bioaccessibility for avocadene and avocadyne was 60.3 ± 0.5% and 47.1 ± 0.7%, respectively. In the ileum, cumulative bioaccessibility for avocadene and avocadyne was 12.0 ± 0.4% and 15.2 ± 0.4%, respectively. In the effluent, less than 8% of both avocadene and avocadyne was quantified (data not shown). Collectively, the cumulative bioaccessibility (sum of jejunum and ileum compartments) for avocadene and avocadyne delivered via emulsion was determined to be 72.2 ±

0.9% and 62.3 ± 1.0%, respectively. Interestingly, these cumulative values were only moderately higher than what was observed for the avocado pulp powder digestion, suggesting some level of equivalency between the two delivery systems of avocadene and avocadyne.

Non-cumulative, absolute concentration—time analysis (Figure 3.4D-F) showed a slight lag in avocadene and avocadyne bioaccessibility in the first hour followed by a rapid increase until two hours. Between one and four hours, avocadene had significantly higher bioaccessibility than avocadyne in the jejunum, however the opposite trend (statistically insignificant) was observed in the ileum where avocadyne bioaccessibility was elevated above that of avocadene until 6 hours. A plateau was seen between two and three hours after which amounts of avocadene and avocadyne rapidly declined towards baseline. AUC analysis on the concentration—time curves revealed almost 1.2 folds higher bioaccessibility of avocadene compared to avocadyne (Table 3.3). This finding was surprising as the initial ratio of the two analytes at input was 1:1, suggesting physicochemical differences as well stoichiometric ratios of avocadene and avocadyne play a

significant role in bioaccessibility.

A B C 1 0 0 A v o c a d e n e 1 0 0 1 0 0

9 0 A v o c a d y n e 9 0 9 0

i i

8 0 8 0 i 8 0

)

i i

i * * * *

% s

s * * * *

(

)

s s 7 0 7 0

s 7 0 * * * *

)

c c ( * * * *

* * * * %

* * * * c c c 6 0 6 0

6 0 ( e

* * * * *

+ i

i 5 0 * * * * 5 0 5 0

v v

e 4 0

n 4 0

v 4 0

u i

j * * *

a l

* * * * l

e l

3 0 3 0 j 3 0

m m

2 0 u 2 0 2 0 u

u * * * * * * * *

C C

C * * * * 1 0 1 0 * * 1 0

0 0 0 0 6 0 1 2 0 1 8 0 2 4 0 3 0 0 3 6 0 0 6 0 1 2 0 1 8 0 2 4 0 3 0 0 3 6 0 0 6 0 1 2 0 1 8 0 2 4 0 3 0 0 3 6 0 T im e (m in ) T im e (m in ) T im e (m in )

D E F J e ju n u m Ile u m J e ju n u m + Ile u m 1 0 5 0 A v o c a d e n e 6 0 A v o c a d y n e 5 0 **** **** 8 **** 4 0 ****

4 0

3 0 6

/ /

/ 3 0 **

g 

****   2 0 4 2 0

1 0 2 1 0

0 0 0 0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6 T im e (h ) T im e (h ) T im e (h )

Figure 3.4. Bioaccessibility of avocadene and avocadyne during 360-min dynamic in vitro TIM-1 digestion of AVO emulsion. 100 mg avocatin B powder was emulsified in an O/W emulsion and administered to TIM-1. Cumulative bioaccessibility of avocadene and avocadyne in (A) jejunum, (B) ileum, and (C) the sum of jejunum and ileum. Non-cumulative bioaccessibility of avocadene and avocadyne in (D) jejunum, (E) ileum, (F) the sum of jejunum and ileum. Non-cumulative bioaccessibility rate of avocadene and avocadyne in (G) jejunum, (H) ileum, (I) the sum of jejunum and ileum, at each time interval. For (A-C) data represents mean (±SEM) % cumulative bioaccessibility relative to input amount. For (D-F) data represents mean (±SEM) non-cumulative bioaccessibility relative to input amount. For (G-I) data represents mean (±SEM) % PFA bioaccessible (relative to input) per minute for each time interval. For (A-I) **p<0.01, *** p<0.001, **** p<0.0001 for avocadene vs. avocadyne, two-way ANOVA, Sidak’s post hoc test, n=3.

Table 3.3. Area under curve analysis on non-cumulative concentration—time graphs for jejunum, ileum, and jejunum+ileum for AVO emulsion administered to TIM-1

Jejunum Ileum Jejunum+Ileum a a a Avocadene 116.4 ± 1.8 24.4 ± 2.0 140.8 ± 3.8 (µg*h/mL) b a b Avocadyne 85.9 ± 3.1 28.8 ± 3.0 114.7 ± 6.1 (µg*h/mL)

Different superscript letters within each column indicates significant differences (P < 0.0001) between avocadene and avocadyne, two-way ANOVA with Sidak’s post hoc test (n=3).

A three-parameter shifted logistical model for lipolytic release of free fatty acids was adopted and used to characterize free fatty alcohol release in each digestive compartment as a function of time. Fitting the cumulative bioaccessibility data to this model also allowed for direct comparison of additional bioaccessibility parameters between the avocado pulp powder and the

AVO emulsion. Table 3.4 highlights the significantly greater jejunal bioaccessibility (Casymp) of both avocadene and avocadyne for the emulsion compared to the avocado powder milk beverage.

This was expected given the emulsion administered was free fatty alcohols whereas the powder contained the target analytes in bound forms. Overall (jejunum and ileum) however, only avocadene had significantly greater bioaccessibility in the emulsion compared to the avocado powder milk beverage. This contrasted with what was expected as the avocado powder contained a higher absolute ratio of avocadene—avocadyne (3:2) whereas the emulsion delivered a 1:1 ratio, confirming the importance of this ratio on overall bioaccessibility parameters. No significant differences in induction time (tc) between or within groups were found for avocadene and avocadyne, however, overall (jejunum and ileum) analysis shows an expected trend towards moderately higher induction times in the avocado powder milk beverage. Lastly, the rate constant

(k) for both avocadene and avocadyne was significantly higher in the jejunum and overall for the

AVO emulsion compared to the avocado powder beverage. Collectively, the mathematical model

applied here confirmed that i) avocadene and avocadyne bioaccessibility from avocado pulp

consumption is dependent on lipolytic activity in the gastrointestinal tract, and ii) bioaccesibility

of pure, formulated AVO is only moderately higher than consuming avocado pulp.

Avocado powder milk beverage AVO Emulsion Avocadene Avocadyne Avocadene Avocadyne Jejunum a,# a, ѱ a, # b, ѱ PFA Bioaccessibility (Casymp) (%) 36.3 ± 1.3 39.1 ± 1.4 61.5 ± 0.6 48.8 ± 0.6 a a a a Induction time (tc) (min) 134.1 ± 4.0 139.6 ± 4.1 132.2 ± 1.4 129.5 ± 1.8 Rate constant (k) (%/min) 0.0181 ± 0.0015a,# 0.0181 ± 0.0016a, ѱ 0.0270 ± 0.0011a,# 0.0247 ± 0.0011a, ѱ Ileum a, # a a, # a PFA Bioaccessibility (Casymp) (%) 18.3 ± 1.3 21.1 ± 1.9 12.3 ± 0.3 15.7 ± 0.4 a a a a Induction time (tc) (min) 177.9 ± 8.7 180.4 ± 10.4 176.1 ± 3.7 182.4 ± 3.7 Rate constant (k) (%/min) 0.0167 ± 0.0028a 0.0155 ± 0.0031a 0.0218 ± 0.0018a 0.0206 ± 0.0016a Jejunum+Ileum a, # a a, # b PFA Bioaccessibility (Casymp) (%) 54.5 ± 2.2 60.2 ± 2.8 73.8 ± 0.9 64.5 ± 1.1 a a a a Induction time (tc) (min) 147.7 ± 4.6 151.8 ± 5.2 138.2 ± 1.8 140.3 ± 2.2 Rate constant (k) (%/min) 0.0172 ± 0.0017a,# 0.0169 ± 0.0019a, ѱ 0.0254 ± 0.0012a,# 0.0227 ± 0.0013a, ѱ

Different superscript letters within each row (within meal type) indicates significant difference (P < 0.0001) between avocadene and avocadyne. # denotes significant difference between avocadene in milk beverage and avocadene in emulsion (P < 0.05), ѱ denotes significant difference between avocadyne in milk beverage and avocadyne in emulsion (P < 0.05). Two-way ANOVA with Tukey’s post hoc test (n=3).

3.4.4. AVO emulsion shows bioavailability and biodistribution in a pilot in vivo

pharmacokinetic study

The nutritional significance along with bioavailability and biodistribution of long chain

fatty alcohols in vivo is not well studied (Ahmed et al., 2018; Hargrove et al., 2004; Place, 1992).

Through an in vivo pilot study in mice, we assessed the pharmacokinetics of a 100 mg/kg (body

weight) oral bolus dose delivered from a 20 mg/mL AVO emulsion. Non-compartmental pharmacokinetic analysis showed that the maximum concentration (Cmax) of avocadyne and avocadene in whole blood was 1687.90 and 1544.83 ng/mL, respectively (Table 3.5). Despite similar maximal blood concentrations (Figure 3.5A), exposure to avocadyne was greater than avocadene due to its higher half-life (t½) of 4.37 hours compared to 3.55 hours for avocadene.

However, this finding cannot be correlated with TIM-1 data due to the pilot nature of this study which had insufficient blood collection time points. Furthermore, we found that avocadyne and avocadene were detectable in various metabolically active tissues with highest accumulation in the liver (Figure 3.5B). A future full-scale pharmacokinetic study will assess the effects of both an oral and intravenous bolus dose of the AVO emulsion to obtain a more descriptive plasma and blood concentration curve and to determine absolute bioavailability of avocadene and avocadyne when delivered in an emulsion. Despite the limitations, this pilot study allowed for the development and validation of a complete bioanalytical method that showed good selectivity, linearity, extraction recovery, accuracy and precision for mouse whole blood and tissue matrices

(Supplementary Table 3.2).

A. B. 2 .0 A v o c a d y n e 2 0 0 A v o c a d y n e A v o c a d e n e 1 8 0 A v o c a d e n e 1 6 0 1 .5

1 4 0

u L

s 1 2 0

m t

/ 1 .0 1 0 0

g /

 8 0 g

n 6 0 0 .5 4 0 2 0 0 .0 0

0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 w rt s r d d n o a a e a a i r e e iv p p ra r r L t t T im e (h ) a H c a a B M n F F a l l e P a a n n o d i a u B n g o n G I

Figure 3.5. AVO emulsion shows bioavailability and biodistribution in a pilot in vivo pharmacokinetic study. AVO emulsion (2% w/w) was delivered via gavage (100 mg/kg body weight (b.w.)) to 6-8 week old female C57BL/6J mice. (A) 50-100 µL blood was collected via tail-snips at 2h and 6h and at endpoint (24h) and avocadene and avocadyne were quantified in blood using a validated LC-MS bio-analytical method. (B) Tissues (bone marrow, heart, pancreas, liver, gonadal fat pad, inguinal fat pad, and brain) were collected at endpoint for avocadene and avocadyne quantification. Data are shown as mean ± S.D., N=3 in each group.

Table 3.5. Avocatin B pilot non-compartmental pharmacokinetic analysis

Avocadyne Avocadene

Mean ± S.D. Mean ± S.D.

Cmax (µg/mL) 1.69 ± 0.08a 1.54 ± 0.22a Tmax (h) 2.00 ± 0.00 2.00 ± 0.00

Kel (h-1) 0.18 ± 0.06a 0.20 ± 0.03a

a a t1/2 (h) 4.37 ± 1.88 3.55 ± 0.50 AUC0_t (µg*h/ml) 9.50 ± 1.24a 13.03 ± 0.84b

AUC0_inf (µg*h/mL) 9.07 ± 3.48a 9.87 ± 2.25a

Cmax denotes maximum concentration of bioactive in blood; Tmax denotes time at which Cmax occurs; Kel denotes elimination rate constant; t1/2 denotes half-life of bioactive in blood; AUC0_t denotes area- under-the-curve (AUC) from time 0-24h; AUC0_inf denotes AUC from time 0 to infinity. Different superscript letters within each row indicates significant differences (P < 0.05) between avocadene and avocadyne, two-tailed Student’s t-test (n=3).

3.5. Discussion

This study demonstrates for the first time that avocado derived PFAs like avocadene and avocadyne are indeed bioaccessible in vitro and bioavailable in vivo. By using static and dynamic in vitro digestion models on lyophilized avocado pulp, it was found that the bioaccessibility of avocadene and avocadyne was largely dependent on lipolytic processes in the gastrointestinal tract.

This finding alone confirms the nutritional significance and relevance of these bioactive molecules given the global rise in avocado pulp consumption in humans. We have previously reported that avocadene and avocadyne in Hass avocado pulp and seed largely exist in bound forms that are liberated as free fatty alcohols in the presence of a strong alkali like sodium methoxide (Ahmed et al., 2018). All potential bound forms of PFAs however have still not been identified. Mechanical pressing or extraction of crude avocado oil from specialized oil containing cells (parenchyma cells and ideoblasts) of avocado pulp mesocarp tissue have been shown to contain saponifiable and nonsaponifiable matter (Farines et al., 1995; Platt-Aloia & Thomson, 1981). The nonsaponifiable matter of crude avocado oil has been characterized to contain sterols (2-4%), squalene (0.5-5%), furan lipids (50-70%), tocopherols (trace amounts), hydrocarbons (5-20%), free fatty acids, ketones and pigments (5-20%) and finally, PFAs like avocadene and avocadyne (5-25%) (Patent

No. 6,582,688 B1, 2003; Patent No. 9,416,333, 2014; Kashman et al., 1969a). Conversely, the saponifiable fraction of avocado pulp oil has been shown to include tri-, di- and mono-glycerides, phospho-, glyco-, sulfo- and galacto-lipids, where triglycerides (TAGs) comprise approximately

85% of avocado pulp lipids (Patent No. 7,81,6547 B2, 2010; Platt-Aloia & Thomson, 1981). From these findings, it is plausible that avocadene and avocadyne that were bioaccessibile after in vitro digestion of avocado pulp powder were from a free fatty alcohol pool, a monoacetate pool

(avocadene and avocadyne acetates) and potentially a complex wax ester pool. Unlike plant waxes

like rice bran wax and sunflower wax (Blake et al., 2014), avocado pulp wax esters have never been characterized. The LC-MS based analytical method used in this study was also not conducive for the detection of waxes or monoacetates (Chen et al., 2015) in the lyophilized avocado pulp powder utilized for in vitro digestion studies.

Assuming avocadene and avocadyne do exist as part of waxes in avocado pulp, their bioaccessibility should have been a lot lower than what was found in this study as wax esters have been shown to be poor substrates for lipolytic enzymes like pancreatic lipase (Place, 1992).

Pancreatic lipase catalyzes the hydrolysis of primary ester linkages in TAG molecules to produce diacylglycerols (namely sn-1,2(2,3)-DAG), monoacylglycerols (namely sn-2-MAG), free fatty acids and glycerol. During TAG hydrolysis, the flow of water molecules into the active site of lipase allows for lipolysis to occur at a maximal rate where the products of hydrolysis (MAGs,

DAGs and protonated fatty acids) form a crystalline phase porous to this water flow (Patton et al.,

1985). In contrast, the products of lipase mediated wax ester hydrolysis (free fatty alcohols and fatty acids) are thought to impede this water flow as they form an insoluble oil or solid phase in water (Place, 1992). In opposition to this theory, avocadene and avocadyne were found to self assemble at the interface of an O/W emulsion characterized in this study, suggesting that their surface-active properties are more likely to enable and not impede lipase mediated hydrolysis that was observed in vitro. Furthermore, the significantly higher assimilation of wax esters in birds has been attributed to the presence of the “enterogastric reflux” which allows gastric and intestinal digestates to be pumped into the gizzard for further emulsification and processing with biliary

TAGs and products of luminal hydrolysis (MAGs and fatty acids) (Place, 1992). In line with this observation, it is possible the combination of bile, milk lipids and avocado powder saponifiable as well as protein content aided the hydrolysis of bound forms of avocadene and avocadyne. Further

studies are needed to elucidate the exact bound forms of avocadene and avocadyne in avocado pulp and seed subsequent to which the mechanism of lipase mediated hydrolysis reported in this study should be confirmed.

Interestingly, we also report the amphiphilic behavior of avocadene and avocadyne in an

O/W emulsion which likely effected two critical interfacial parameters. First, the spontaneous self- assembly of avocadene and avocadyne at the oil-water interface modified film curvature by absorbing at the interface and aligning the polar portion of the fatty alcohol towards the aqueous phase while the apolar aliphatic chain aligned toward the oil phase. Second, the self-assembly of the PFAs at the interface likely increased the elasticity of the interfacial film (referred to as the bending moduli) (Garti et al., 2001; Gradzielski, 2002; Morales et al., 2003). Despite the ideal stability and extremely small particle size of the chosen delivery system for AVO, TIM-1 digestion of the AVO emulsion showed only a modest enhancement in bioaccessibility compared to what was observed for the avocado powder. It is plausible that not all avocadene and avocadyne delivered via the emulsion remained solubilized in the aqueous phase of the bioaccessible portion of digestates. The potential displacement of avocadene and avocadyne from the interface by bile and the inhibitory effect of non-ionic surfactants like Tween 80 on bioaccessibility of O/W emulsions (Speranza et al., 2013) should be further explored along with the precise mechanism by which AVO exerts surface activity. The AVO emulsion did however assimilate in vivo where appreciable quantities were observed in mouse whole blood and key metabolically active tissues.

A full-scale pharmacokinetic evaluation that compares oral and intravenous doses of AVO will determine absolute bioavailability and allow some level of comparison with cumulative and non- cumulative TIM-1 bioaccessibility data reported in this study. Biodistribution analysis from this pilot study did however point to accumulation in mouse liver which is highly suggestive of

avocadene and avocadyne biotransformation into secondary metabolites like fatty aldehydes and fatty acids (Hargrove et al., 2004) that could not be accounted for with the developed bio-analytical method. The synthesis of radiolabeled stable-isotopes of avocadene and avocadyne is essential to develop analytical methods that can quantify the parent compounds as well as their metabolites in whole blood, plasma and tissues.

3.6. Acknowledgments

We would like to acknowledge Natalie Ng and Chloe Chotard for their time and contributions towards all TIM-1 experiments.

3.7. Author contributions

N.A. and P.A.S. conceptualized the experimental design, analyzed data and wrote the manuscript. P.X.C. operated TIM-1 and assisted N.A. in completing all TIM-1 experiments.

M.A.R. supervised N.A. with all TIM-1 experiments and edited the manuscript. R.W.S. supervised

N.A. with bio-analytical method development for all TIM-1 and animal studies. A.R. assisted N.A. with the animal study and PK data analysis.

3.8. Supplementary information (chapter 3)

3.8.1. Lyophilized avocado powder proximate analysis

Supplementary Table 3.1. Manufacturer provided certificate of analysis for lyophilized avocado pulp powder

3.8.2. AVO bio-analytical method

Liquid Chromatography-Mass Spectrometry

A previously developed LC-MS method for the quantitation of avocadene and avocadyne in avocado seed and pulp (Ahmed et al., 2018) was further validated for mouse whole blood and tissue as well as TIM-1 jejunum, ileum and effluent dialysates. All chromatography and mass spectrometer parameters utilized were the same as previously described (Ahmed et al., 2018) except all quantitative analysis was performed in high-resolution selected-ion monitoring (SIM)

+ mode which enabled avocadene and avocadyne fragmentation ions ([M+H-H2O] ) to be detected at low detection limits compared to the LC-MS method.

Whole blood and tissue extraction and sample preparation

For whole blood extraction, 100 µL of whole blood was extracted using a modified Folch protocol

(Folch et al., 1957). Briefly, 100 µL of whole blood was macerated and mixed in 3 mL of 2:1 chloroform-methanol (v/v) at room temperature after which 0.5 mL of 0.2 M sodium-phosphate

(NaHPO4) buffer in ddH2O (pH 4.4) was added to induce layer separation. After inversion, samples were centrifuged for 5 min at 1500 rcf. The total lipid containing organic layer was collected and an additional 2 mL chloroform was added to the aqueous layer as a wash step and as an additional round of extraction. The second organic layer was combined with the first and the buffer layer was discarded. Chloroform extracts were dried under a gentle stream of nitrogen and stored at 4 °C until sample preparation was required for LC-MS analysis. Blank whole blood (blood from non- treated or control mice) was also extracted the same way for recovery experiments and the generation of matrix-matched AVO standard curves. Modified Folch protocol was also utilized for tissue extraction where 75 mg of flash frozen and pulverized tissue was macerated in 3 mL of 2:1 chloroform-methanol (v/v) and homogenized 40 times with a tissue grinder. 0.5 mL of NaHPO4

was then added to induce layer separation and double extraction of organic layer was completed as described before. Tissue chloroform extracts were dried under a gentle stream of nitrogen and stored at 4 °C until sample preparation was required for LC-MS analysis. Tissue from non-treated or control mice were also extracted the same way for recovery experiments and the generation of matrix-matched AVO standard curves. On analysis day, all dried blood and tissue samples were reconstituted in 250 µL of LC starting gradient (60% water-40% acetonitrile + 0.1% FA) and 10

µL of each was injected into the LC-SIM-MS method.

TIM-1 samples extraction and sample preparation

100 µL of TIM-1 dialysates from jejunum, ileum, and effluent were extracted and samples prepared with the same procedures described for mouse blood. Blank TIM-1 dialysates (water administration to TIM-1 in fed state protocol) from 1 and 6 h samples were also extracted the same way for recovery experiments and the generation of matrix-matched AVO standard curves.

Method validation

Selectivity and Linearity

The effect of endogenous matrix constituents interfering with retention times of avocadyne and avocadene was assessed by extracting and analyzing individual batches (in duplicates) of blank whole blood, tissue or TIM-1 dialysates using the extraction procedure and chromatographic/mass spectroscopic conditions described above. Responses of the analyte at the lower limit of quantitation (LLOQ) concentration were compared with the responses in the blank samples.

Selectivity was found to be acceptable as no interference from endogenous matrix constituents was found for whole blood, tissue and TIM-1 blank samples. Linearity was assessed by analyzing six- point matrix-matched calibration curves of AVO in blank mouse whole blood, tissue or TIM-1 dialysates extracted as described above. Standard curves were constructed using least-square linear

regression of peak area under curve versus nominal standard concentrations. Linearity was assessed by evaluating the slope, intercept and coefficient of determination (r2) of two different calibration curves produced on separate analytical days. Supplementary Table 3.2 and 3.3 highlight linear range and correlation coefficients (r2) for whole blood, tissues and TIM-1 dialysates.

Accuracy and Precision

Intra- and inter-day accuracy and precision of the developed method were determined by assaying three concentrations of quality control (QC) samples (for blood matrix: LLOQ = 2 ng/mL, low = 100 ng/mL, and high QC = 1800 ng/mL; for tissue matrix: see Supplementary Table 3) in duplicates on two different analytical days. Precision was reported as percent coefficient of variation (%CV) of replicates within one sample run (intra-assay) or between sample runs (inter- assay). Intra- and inter-assay accuracy was reported as percent relative error (% RE) or the percent deviation of QC replicates from nominal concentration. The acceptance limit for accuracy and precision, at low and high QC concentration levels, were set to 15% RE and 15% CV, respectively.

For LLOQ, accuracy and precision acceptance limits were set to below 20% RE and 20% CV, respectively (Zimmer, 2014). See Supplementary Tables 3.2 and 3.3 for accuracy and precision parameters for whole blood, tissue and TIM-1 dialysates.

Extraction Recovery and Matrix Effect

Recovery of avocadene and avocadyne from mouse whole blood, tissue and TIM-1 dialysates was determined by comparing the peak areas of extracted QC samples (at LLOQ, low and high concentrations as highlighted in Supplementary Tables 3.2 and 3.3) with the peak areas of post-extraction blood, tissue or TIM-1 dialysate blanks spiked at corresponding concentrations.

The matrix effect of the different samples on the ionization of avocadene and avocadyne was evaluated by comparing the peak areas of post-extraction blank biological or TIM-1 samples spiked at concentrations of QC samples with the areas obtained by QC samples prepared in solvent

(LC starting gradient). This analysis was performed for biological replicates. See Supplementary

Tables 3.2 and 3.3 for extraction recovery and matrix effect for whole blood, tissue and TIM-1 dialysates.

Supplementary Table 3.2. AVO bioanalytical method validation parameters for pilot PK study.

Sample Linear Range, r2 QC Accuracy (% error) Precision (% CV) Recovery (%) Matrix Effect (%) concentrations

Intra-day Inter-day Intra-day Inter-day

Avocadyne Avocadene Avocadyne Avocadene Avocadyne Avocadene Avocadyne Avocadene Avocadyne Avocadene

Whole blood 2-1800 ng/mL, 2-1800 ng/mL, ng/mL 0.996 0.999 2 10.3 11.4 4.5 5.9 82.2 ± 2.3 79.2 ± 0.3 80.2 ± 1.9 80.0 ± 1.8 100 3.3 3.9 2.9 2.2 89.9 ± 1.1 89.3 ± 1.6 83.3 ± 0.9 84.5 ± 0.3 1800 1.1 1.9 1.3 1.8 92.8 ± 0.8 95.8 ± 2.8 88.7 ± 1.5 88.9 ± 0.7

Bone marrow 0.5-10 ng/g, 0.5-10 ng/g, ng/g 0.995 0.998 0.5 9.6 11.6 3.3 2.8 78.3 ± 3.3 77.1 ± 1.9 77.2 ± 1.8 76.6 ± 2.8 2 4.6 3.8 3.9 2.4 83.9 ± 0.6 89.1 ± 1.1 80.0 ± 0.5 79.9 ± 0.6 10 3.9 3.8 1.9 2.9 89.1 ± 1.8 90.3 ± 0.2 87.9 ± 0.8 86.6 ± 1.5 Heart 1-200 ng/g, 1-200 ng/g, ng/g 0.998 0.996 0.5 8.9 12.3 2.8 3.9 76.1 ± 1.1 73.1 ± 0.9 73.9 ± 1.9 72.8 ± 1.0 2 6.9 7.6 3.6 4.2 79.9 ± 1.4 74.4 ± 1.3 79.9 ± 2.5 78.6 ± 0.9 10 4.3 5.9 1.3 1.1 83.4 ± 2.1 85.6 ± 0.9 82.1 ± 1.9 83.1 ± 0.7 Pancreas 1-200 ng/g, 1-200 ng/g, ng/g 0.998 0.999 0.5 13.9 11.8 3.9 7.5 79.8 ± 0.9 77.9 ± 0.2 79.6 ± 0.8 78.5 ± 0.2 2 10.8 10.7 4.5 3.1 80.3 ± 1.9 79.9 ± 1.8 86.4 ± 1.3 84.6 ± 0.9 10 7.2 5.6 4.0 2.8 85.6 ± 2.9 83.2 ± 1.5 88.1 ± 0.9 88.6 ± 1.1 Liver 1-200 ng/g, 1-200 ng/g, ng/g 0.995 0.999 1 10.8 14.6 5.9 7.9 73.0 ± 2.9 74.1 ± 1.9 77.3 ± 0.3 77.1 ± 0.9 10 5.3 11.2 3.6 7.1 79.5 ± 3.3 79.6 ± 1.8 82.4 ± 1.2 83.5 ± 1.3 200 1.9 5.6 2.8 5.5 81.2 ± 1.8 82.2 ± 0.5 89.6 ± 1.4 88.7 ± 0.6 Gonadal Fat 1-200 ng/g, 1-200 ng/g, ng/g Pad 0.995 0.996 1 11.2 15.9 4.6 5.9 70.5 ± 1.2 69.9 ± 0.8 80.6 ± 0.1 78.4 ±1.5 10 10.8 11.9 3.1 5.1 75.2 ± 2.2 73.5 ± 1.2 84.6 ± 0.9 83.5 ±1.8 200 9.7 10.8 2.9 4.1 80.1 ± 1.8 76.6 ± 0.8 88.1 ± 1.1 87.4 ± 0.8 Inguinal Fat 1-200 ng/g, 1-200 ng/g, ng/g Pad 0.998 0.997 1 14.8 14.9 4.6 6.9 72.3 ± 0.6 70.5 ± 0.9 81.2 ± 5.5 79.6 ± 1.3 10 11.6 12.1 3.3 4.6 76.9 ± 1.2 74.6 ± 1.3 84.9 ± 0.2 81.2 ± 0.6 200 5.9 4.7 1.8 3.3 81.2 ± 0.8 80.0 ± 1.5 89.6 ± 0.5 87.6 ± 1.2 Brain 1-200 ng/g, 1-200 ng/g, ng/g 0.998 0.995 1 11.8 10.7 3.9 3.5 76.5 ± 0.9 75.9 ± 0.4 80.2 ± 0.3 82.8 ± 0.5 10 5.9 3.5 1.8 1.8 80.6 ± 0.7 79.8 ± 1.1 87.5 ± 1.5 86.6 ± 1.3 200 4.3 1.2 1.3 1.9 89.9 ± 1.3 85.5 ± 0.2 89.1 ± 0.2 89.6 ± 0.2

Supplementary Table 3.1. AVO analytical method validation parameters for TIM-1 studies.

Sample Linear Range, r2 QC Accuracy (% error) Precision (% CV) Recovery (%) Matrix Effect (%) concentrations

Intra-day Inter-day Intra-day Inter-day

Avocadyne Avocadene Avocadyne Avocadene Avocadyne Avocadene Avocadyne Avocadene Avocadyne Avocadene

Jejunum 1h 0.1-10 µg/mL, 0.1-10 µg/mL, µg/mL 0.995 0.999 0.1 13.3 14.1 5.5 6.3 73.2 ± 2.3 71.3 ± 0.6 76.2 ± 0.7 74.0 ± 2.8 2 9.3 8.9 3.9 4.7 79.9 ± 3.1 79.3 ± 1.4 82.3 ± 0.3 81.5 ± 0.9 10 2.1 4.9 2.3 3.6 85.8 ± 0.8 84.8 ± 0.8 89.7 ± 1.9 89.9 ± 0.7

Jejunum 6h 0.1-10 µg/mL, 0.1-10 µg/mL, µg/mL 0.997 0.999 0.1 12.6 14.6 4.3 5.8 70.3 ± 2.3 70.1 ± 0.9 74.2 ± 1.8 73.6 ± 0.8 2 8.6 9.8 3.9 4.4 80.9 ± 07 83.1 ± 1.8 79.0 ± 3.5 83.1 ± 1.6 10 2.9 3.8 2.9 3.9 88.1 ± 1.3 88.3 ± 0.6 89.9 ± 5.8 88.6 ± 3.5 Ileum 1h 0.1-5 µg/mL, 0.1-5 µg/mL, µg/mL 0.995 0.997 0.1 18.9 15.3 5.8 6.9 70.1 ± 0.1 70.1 ± 0.8 69.9 ± 0.3 70.8 ± 2.0 1 12.9 9.6 3.6 3.2 75.9 ± 2.4 73.4 ± 1.4 72.9 ± 1.8 75.6 ± 3.9 5 6.3 5.9 4.3 3.1 80.6 ± 5.1 80.0 ± 3.9 80.1 ± 3.9 80.1 ± 1.7 Ileum 6h 0.1-5 µg/mL, 0.1-5 µg/mL, µg/mL 0.997 0.996 0.1 17.9 11.8 3.8 4.5 71.8 ± 1.2 73.0 ± 0.8 68.5 ± 0.3 69.5 ± 0.3 1 10.8 9.9 4.5 4.6 78.3 ± 2.5 77.1 ± 1.1 76.4 ± 1.6 74.7 ± 2.3 5 7.0 4.5 3.0 3.7 81.2 ± 0.9 85.2 ± 1.4 80.1 ± 0.4 83.4 ± 2.1 Effluent 1h 0.1-5 µg/mL, 0.1-5 µg/mL, µg/mL 0.996 0.998 0.1 14.0 14.3 3.0 3.3 72.0 ± 1.9 73.2 ± 1.9 75.2 ± 2.3 77.3 ± 1.9 1 8.3 10.0 2.8 3.9 74.3 ± 2.3 77.4 ± 1.7 80.1 ± 3.4 80.4 ± 2.2 5 3.0 2.6 1.8 2.6 83.4 ± 3.3 83.6 ± 0.8 84.4 ± 2.6 86.3 ± 0.9 Effluent 6h 0.1-5 µg/mL, 0.1-5 µg/mL, µg/mL 0.995 0.998 0.1 13.1 11.0 5.5 6.8 73.4 ± 0.6 70.9 ± 1.2 79.5 ± 0.9 71.4 ±1.9 1 9.4 10.8 3.2 4.2 76.0 ± 4.1 74.3 ± 2.4 80.4 ± 2.3 80.9 ±1.2 5 4.4 3.9 1.9 2.2 89.6 ± 7.0 79.9 ± 1.3 86.2 ± 2.5 84.4 ± 0.4

Chapter 4: Avocatin B protects against lipotoxicity and improves insulin sensitivity in diet-induced obesity

Nawaz Ahmed1, Matthew Tcheng1, Alessia Roma1, Michael Buraczynski1, Preethi Jayanth1, Kevin Rea2, Tariq A. Akhtar2, Paul A. Spagnuolo1*

1Department of Food Science, University of Guelph, Guelph, Ontario, Canada 2Department of Molecular and Cellular Biology, University of Guelph, Guelph, Ontario, Canada

*To whom correspondence should be addressed. Email: [email protected]

Modified version of this manuscript is in press at the Journal of Molecular Nutrition and Food Research

4.1. Abstract

Scope: The effects of an avocado-derived fatty acid oxidation (FAO) inhibitor, avocatin B (AVO), on glucose and lipid metabolism in models of diet-induced obesity (DIO) and in vitro models of lipotoxicity was evaluated. The safety of its oral consumption in humans was also evaluated.

Methods and Results: Mice were given high fat diets (HFD) for 8 weeks. Thereafter, AVO (100 mg/kg body weight) or vehicle was administered orally twice weekly for 5 weeks. AVO inhibited

FAO which led to improved glucose tolerance, glucose utilization and insulin sensitivity. AVO’s effects on metabolism under lipotoxic conditions were evaluated in vitro in pancreatic β-islet cells

(INS-1 832/13) and C2C12 myotubes. AVO inhibited FAO and increased glucose oxidation resulting in lowering of mitochondrial reactive oxygen species that improved insulin responsiveness in C2C12 myotubes and insulin secretion in INS-1 (832/13) cells, respectively. A randomized, double-blind, placebo-controlled clinical trial in healthy human participants was conducted to assess the safety of AvoB consumption (50 mg or 200 mg/day for 60 days). AVO was well-tolerated and not associated with any dose-limiting toxicity.

Conclusion: Therapeutic agents that are safe and effectively inhibit FAO and improve DIO- associated pathologies are currently not available. AVO’s mechanism of action and favorable safety profile highlight its nutritional and clinical importance.

4.2. Introduction

Diet-induced obesity (DIO) is a central risk factor for the onset of metabolic complications like insulin resistance, type 2 diabetes (T2D) and cardiovascular diseases (Anderson et al., 2009;

Després & Lemieux, 2006; Muoio & Neufer, 2012). The mechanism(s) by which lipotoxicity (i.e.,

elevated free fatty acids, FFAs) cause metabolic dysfunction continue to be widely debated; however, Randle and colleagues (Randle et al., 1963) were the first to show that metabolic flexibility, which is the capacity of tissues to adapt fuel oxidation to substrate availability

(Goodpaster & Sparks, 2018), is heavily impaired in obese and diabetic patients. This led to a prominent theory that insulin resistance is driven by elevated rates of skeletal muscle fatty acid oxidation (FAO; or β-oxidation) (Randle, 1998). Randle’s theory has been challenged (Cline et al., 1994; Cline et al., 1997; Kelley et al., 1999; Roden et al., 1996; Wolfe, 1998); nonetheless, there is a particular interest, given that FAO is modifiable through drug-targeting, in exploiting fatty acid metabolism to improve metabolic complications associated with DIO and lipotoxicity.

Incomplete mitochondrial FAO has emerged as a key factor in skeletal muscle insulin resistance, as excess β-oxidation in the post-prandial state gives rise to incomplete FAO which produces lipotoxic acylcarnitines that impair carbohydrate utilization, deplete organic intermediates of the tricarboxylic acid (TCA) cycle and generates reactive oxygen species (ROS)

(Adams et al., 2009; Gavin et al., 2018; Koves et al., 2008; Mihalik, 2010). In the pancreas, the site of insulin synthesis and secretion, lipotoxicity disrupts the glucose-fatty acid cycle in β-cells

(Zhou & Grill, 1994, 1995), which impairs glucose-stimulated insulin secretion (GSIS) resulting in hyperinsulinemia in vitro and in vivo (Biden et al., 2002; Erion et al., 2015; Ježek et al., 2018;

Zhou & Grill, 1994). Therefore, in periods of lipotoxicity, diverting fuel source selection towards glucose by inhibiting FAO in skeletal muscle and pancreatic β-cells may improve insulin sensitivity and reduce hyperinsulinemia, respectively.

Inhibiting FAO has indeed emerged as a potential therapeutic strategy in DIO (Collier et al., 1993; Foley J. E., 1992; Gao et al., 2015; Keung et al., 2013; Lopaschuk, 2016; Muoio &

Newgard, 2008a; Pettus et al., 2016); however, safe and well-tolerated small molecules that

effectively target this metabolic pathway are currently unavailable. Our lab recently discovered that the avocado-derived lipid, avocatin B (AVO), is an FAO inhibitor that accumulates in mitochondria to induce leukemia cell apoptosis without imparting toxicity toward normal cells (E.

A. Lee et al., 2015). However, AVO’s mechanism of FAO inhibition and/or metabolic modulation in other tissue types has not been studied; therefore, we sought to examine if AVO can ameliorate the effects of lipotoxicity in both skeletal muscle and pancreatic tissue in vitro and in vivo and also test its safety in a Phase I human study.

4.3. Experimental Section

4.3.1. Pre-clinical mouse studies

As a treatment study, twelve-week-old male C57BL/6J mice were purchased from Jackson

Laboratory (Bar Harbor, ME) and allowed to acclimatize for 1 week. After acclimatization, mice received either a high-fat diet (HFD) (60% kcal from lard; Research Diet D12492; Research Diets,

USA) or a standard diet (10-13% kcal from fat; Teklad 2014; Envigo, USA) for eight weeks. At the end of week eight, animals continued on their respective diets but were administered AVO

(100mg/kg body weight (b.w.)) or vehicle via gavage twice weekly for 5 weeks (see Fig. 1A).

AVO was formulated in a self-emulsifying drug delivery system (SEDDS) as previously described

(Buyukozturk et al., 2010) where the oil phase for both vehicle and AVO containing SEDDS comprised of Tween 80 (5% (v/v); Sigma, MO, USA) and a medium chain triglyceride oil:

NeoBee®M5 (5% (v/v); Gattefosse, St-Priest, France). Animal weights were monitored twice weekly. At study completion, ad libitum animals were euthanized (10-12 h into their dark cycle) via CO2 followed by exsanguination, after which tissue and blood was collected.

For the prevention study, eight-week-old male C57BL/6J mice were also purchased from

Jackson Laboratory and allowed to acclimatize for 1 week. After acclimatization, mice received the same diets as outlined above for thirteen weeks. AVO or vehicle supplementation was initiated as described above at the onset of standard or HFDs. Animal weights were monitored twice weekly. At study completion, 6-8h fasted animals were euthanized via CO2 followed by exsanguination, after which tissue and blood was collected.

All animal studies were carried in accordance to the regulations of the Canadian Council on Animal Care (CCAC) and with the approval of the Animal Care Committee at the University of Guelph.

4.3.1.1. Glucose and insulin tolerance tests

Intraperitoneal glucose and insulin tolerance tests (GTT and ITT) were performed 8 hours after food withdrawal. For GTT, a glucose dose of 1.5 g/kg was used for mice on both HFD and standard diet. For ITT, an insulin (Humulin, Eli Lilly, USA) dose of 1 U/kg and 0.25 U/kg was used for mice on HFD and standard diet, respectively. Blood glucose levels were determined at 0,

10, 20, 30, 60, and 90 minutes after glucose/insulin administration via tail bleed using the

OneTouch Blood Ultra 2 glucose meters (LifeScan Europe, Switzerland).

4.3.1.2. Measurement of complete blood counts, plasma and tissue biochemical markers

Complete blood count with differential (CBC w/diff) analysis was done on 500 µL of whole blood by the animal health laboratory at the University of Guelph. Plasma insulin, FFAs, and triacylglycerols (TAGs) were measured using a rat/mouse insulin ELISA kit (Millipore,

Rat/Mouse Insulin Detection kit), FFA fluorometric assay kit (Cayman, MI, USA), and the TAG colorimetric assay kit (Cayman, MI, USA), respectively. Manufacturer’s protocols were followed

for all kits.

4.3.1.3. Insulin resistance index (HOMA-IR) calculations

The homeostasis model assessment was used to calculate the insulin resistance (HOMA-

IR) index (Matthews et al., 1985) using the values of fasting plasma glucose (FPG) and fasting plasma insulin (FPI) as follows: HOMA-IR=FPG×FPI/22.5, with FPG expressed as mmol/L and

FPI as mIU/L.

4.3.1.4. Pyruvate dehydrogenase (PDH) activity assay

After mice were euthanized, muscle tissues were freeze-clamped and flash-frozen in liquid nitrogen. The PDH activity from flash-frozen whole gastrocnemius muscle lysates was measured using a colorimetric microplate assay kit (ab109902, Abcam, Cambridge, MA, USA) following the manufacturer’s protocol and as previously described (Gao et al., 2015). Briefly, PDH proteins from whole gastrocnemius muscle lysates (lysed in the presence of 10 mM sodium fluoride (NaF)

(a phosphatase inhibitor) to preserve the phosphorylated status of PDH or to determine native PDH activity as a result of HFD feeding) were immunocaptured on a microplate by loading equal amounts of protein for each sample. While kinase inhibitors (e.g., dichloroacetate) were not present in the lysis buffer, the presence of phosphatase inhibitors increased the likelihood that native PDH activity was measured. PDH activity was determined spectrophotometrically by monitoring the reduction of NAD+ to NADH, coupled to the reduction of a reporter dye at an absorbance of 450 nm.

4.3.2. In vitro studies

4.3.2.1. Cell culture

Cells were cultured in a humidified atmosphere containing 5% CO2 at 37°C. INS-1

(832/13) rat pancreatic β-cell line was cultured in RPMI 1640 medium containing 11.1 mM glucose and supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, 2mM

L-glutamine, 1 mM sodium pyruvate, and 50 μM β-mercaptoethanol. INS-1(832/13) cells are a robust β-islet cell culture model routinely used given their glucose responsiveness and endogenous insulin production. C2C12 mouse skeletal myoblast were cultured in growth media consisting of low-glucose Dulbecco’s Modified Eagles Medium (DMEM; Hyclone, ThermoFisher) supplemented with 10% FBS and 1% penicillin/streptomycin. Differentiation of C2C12 myoblasts into myotubes was induced by switching 90% confluent cells to differentiation media consisting of low-glucose DMEM supplemented with 2% horse-serum and 1% penicillin/streptomycin.

Differentiation media was changed every 24h for up to 5 days prior to all experimental treatments.

At day 5 of the differentiation process, myoblasts are fully converted into myotubes, as determined by morphological assessment and immunoblotting for myogenin, a skeletal muscle-specific protein (Supplementary Figure 4.3C).

4.3.2.2. Cell viability

Quantitative analysis of cell death was evaluated by flow cytometry using propidium iodide staining (Biovision, Mountainview, CA), where specified, according to the manufacturer’s protocol and as previously described (E. A. Lee et al., 2015). Viable cells were identified as PI negative (PI-).

4.3.2.3. Conjugation of palmitic acid to fatty acid-free bovine serum albumin

Briefly, fatty acid and endotoxin free BSA (Sigma) was dissolved in distilled water at 37 oC to produce a 7% BSA solution. Sodium palmitate (Sigma) was dissolved in distilled water at

80 oC to a final concentration of 5 mM. The sodium palmitate solution was added dropwise to the

BSA solution and stirred for 1 hour. PA-BSA conjugation was quantified with the NEFA quantification kit (Wako Pure Chemicals) and stored at -20 oC until further use. PA-BSA was used to assess FAO-supported intact cell high resolution respirometry and liquid scintillation counting studies.

4.3.2.4. High resolution respirometry

Cells treated with test compounds for 24h were trypsinized, counted, and resuspended in phosphate buffer saline (PBS). High-resolution O2 consumption measurements were conducted in

2 ml of respiration medium (PBS, pH 7.5, stir speed 750 rpm) using the Oroboros Oxygraph-2k

(Oroboros Instruments, Corp., Innsbruck, Austria) set to 37 oC with a gain of 2 in both chambers.

Briefly, after the injection of 5 million treated INS-1 (832/13) cells or C2C12 myotubes per chamber, basal respiration was measured once steady-state respiratory flux was obtained; this represents coupled respiration fueled by available substrates from the incubation with test compounds. Uncoupled respiration was then measured following addition of the ATP synthase inhibitor oligomycin (0.25µM). This was followed by measuring maximal respiratory rate by stepwise titration of carbonyl cyanide p-trifluoromethoxy phenyl hydrazone (FCCP; within 0.1-

0.25 µM) in the presence of 1mmol/L pyruvate. Finally, respiration was inhibited by addition of rotenone (0.5 µM; complex I inhibitor) to obtain extra-mitochondrial residual oxygen consumption. Data was recorded with DatLab software 7.4 (Oroboros Instruments, Innsbruck,

Austria), oxygen consumption at each step of the protocol was normalized to vehicle control and illustrated in histograms.

C2C12 myoblasts or INS-1 (832/13) cells were treated in low glucose media with AVO

(25µM) or etomoxir (100µM) in the presence of 0.25 mM PA/1 mM carnitine for 24h before being collected and injected into the respirometer to assess PA supported respiration. Additionally, to indirectly assess metabolic flexibility, cells were treated the same way as described, collected and subjected to a 15 min treatment of 10 mM D-glucose before being washed and injected into the respirometer; these experiments were done to assess if cells incubated in low glucose/high fat conditions for 24h can easily switch to glucose utilization upon acute incubation with D-glucose.

For all HRR protocols, cell viability was determined after trypsinization and also after end of HRR runs via trypan blue cell staining and counting using a hemocytometer. No significant differences in cell viability between control and treated cells were observed.

4.3.2.5. Fatty acid oxidation

FAO in cell lines was determined using established methods (Molecular & Carolina, 2014).

Briefly, cells were pretreated with AVO (25µM) or etomoxir (100µM) in the presence of 0.5 mM

PA/1mM L-carnitine for 24h after which they were washed with PBS and subject to radiolabeled

[1-14C]-palmitate for 3h. [1-14C]-palmitate was prepared and applied to cells as follows: purchased stock supplied in ethanol (PerkinElmer, MA, USA) was dried in vacuo and then re-suspended in a solution containing unlabeled carrier in BSA to yield a mixture of 7% BSA [w/v], 2.5 mM palmitate with [1-14C]-palmitate at a final concentration of 10 μCi/mL. Following a 16 hour incubation at 37°C, the solution containing radiolabeled palmitate was diluted with respective serum free cell culture media (low glucose RPMI media for INS-1 (832/13) cells and low glucose

DMEM for C2C12 myotubes) containing 1 mM L-carnitine to a final concentration of 0.3% BSA,

100 μM palmitate, and 0.4 μCi/mL [1-14C]-palmitate. Following incubation of treated cells with

[1-14C]-palmitate for 3h, cells were gently scraped and 400 µL of well media was transferred to acidification vials containing perchloric acid and capped with CO2 filter traps impregnated with

1M NaOH. The vials were then incubated at room temperature for one hour to allow for CO2 capture and precipitation of undigested palmitate. The CO2 trap was transferred to scintillation vials containing 4 mL of scintillation fluid and complete FAO was quantified using a Tri-Carb

2910 TR liquid scintillation analyzer (PerkinElmer). The amount of radioactivity in the media was also determined after pelleting cells and debris (14,000 x g, 10 min); this quantified acid soluble metabolites (ASMs), an indirect measure of incomplete FAO.

4.3.2.6. Glucose oxidation and uptake

C2C12 myotubes and INS-1 (832/13) cells treated as in the FAO assays described above and then incubated with D-[14C(U)]-glucose (0.58µC/mL) and unlabeled glucose (final glucose

14 concentration 200 µM) for 4h subsequent to which [ C]-CO2 (representing complete glucose oxidation) were captured and quantified as described in the FAO method. For C2C12 myotubes, glucose oxidation and uptake was measured with and without a 30 min pre-incubation with 100 nM insulin to determine basal and insulin-mediated glucose oxidation or uptake, respectively.

To measure glucose uptake in C2C12 myotubes, a fluorescent D-glucose analog 2-(N-[7- nitrobenz-2-oxa-1,3-diazol-4-yl]amino)-2-deoxy-D-glucose (2-NBDG) was utilized. Briefly, myotubes were treated for 24h with test compound after which they were washed twice with PBS and starved for 1h in serum and glucose free DMEM. Following starvation, 100 nM insulin was added for 30 min after which 60 µM 2-NBDG was added to each well and allowed to incubate for

30 min. The 2-NBDG uptake reaction was then stopped by removing the incubation medium and washing the cells twice with PBS after which the cells were trypsinized and collected for flow cytometry analysis (Guava 8HT; EMD Millipore, Billerica, MA).

4.3.2.7. Glucose stimulated insulin secretion (GSIS) in INS-1 (832/13) cells

To measure glucose stimulated insulin secretion (GSIS) in INS-1(832/13) cells a standard protocol was adopted (Patterson et al., 2014). Briefly, cells were grown to 80% confluency in 6 well plates and treated with AVO or etomoxir with or without palmitate as described for the FAO assay. After treatments were complete, cell media was replaced with KRB (Krebs-Ringer

Bicarbonate) buffer. Cells were then challenged with 3 mM glucose for 2h in KRB buffer after which 500 µL KRB buffer was collected and frozen in -80°C until analysis. For another 2h, cells were challenged with 16 mM glucose after which 500 µL KRB was collected and frozen in -80°C until analysis. The amount of insulin released into the KRB buffer was determined by an ELISA kit (Millipore, Rat/Mouse Insulin Detection kit). Data was normalized for cellular protein content, as determined by the Micro-BCA Protein Assay kit. The glucose-dependent insulin secretory index

(GSIS16/3), defined as the ratio between the insulin secretion at 16 mM (stimulatory, surrogate to postprandial glucose levels) and 3 mM (basal, surrogate to fasting glucose levels) glucose was also calculated.

4.3.2.8. Reactive oxygen species, mitochondrial membrane potential, and mitochondrial mass determination

For all ROS studies, INS-1 (832/13) cells and C2C12 myotubes were treated with or without palmitate and FAO inhibitors for 24h in low glucose (5.5 mM) cell culture media as

described for the FAO assay. After treatment, cells were trypsinized and re-suspended in PBS containing 5µM of the fluorescent dye MitoSOX® (Molecular Probes, Invitrogen) and allowed to incubate for 5-15 minutes in a humidified atmosphere containing 5% CO2 at 37 °C. Cells were then washed in PBS, placed in a 96 well plate and fluorescence was measured by flow cytometry. ROS was quantified by normalizing the mean red fluorescence values of live gated cells to vehicle control.

Mitochondrial membrane potential (MMP) in INS-1 (832/13) and C2C12 myotubes was measured using the cationic fluorescent dye JC-1 (Molecular Probes). JC-1 accumulates in the mitochondrial matrix in an MMP dependent manner. JC-1 fluoresces green in the cytosol where it exists as a monomer, after entering the mitochondria JC-1 monomers form J aggregates and fluoresce red, thus the red/green fluorescence ratio is used to determine MMP. After all treatments, cells were washed, trypsinized and re-suspended in PBS containing 1 µM JC-1 and allowed to incubate for 5-10 minutes in a humidified atmosphere containing 5% CO2 at 37 °C. Cells were then washed in PBS, placed in a 96 well plate and fluorescence was measured by flow cytometry.

MMP was quantified by normalizing the mean fluorescence values of live gated cells to vehicle control.

Mitochondrial mass was measured using 10-N-nonyl acridine orange (NAO; Enzo Life

Sciences, ENZ-52306), which accumulates in the mitochondria in an MMP-independent manner where it binds to cardiolipin in the inner mitochondrial membrane. Cells were treated, washed and collected as described and incubated with 0.35 μM NAO and green fluorescence was measured using flow cytometry.

In all analyses (ROS, MMP and mitochondrial mass), forward vs side scatter plots were gated to exclude debris (e.g., dead cells), as cell death is associated with increased ROS and

altered membrane potentials.

4.3.2.9. Cellular ATP determination

Total cellular ATP content was determined using the ATP bioluminescent assay kit

(Calbiochem, 119107). Briefly, cells were seeded in white walled 96-well plates

(1.2x104cells/well for INS-1 and 2x104 cells/well for C2C12) and treated for 24h (C2C12 cells were differentiated for 5 days prior to treatments whereas INS-1 cells were treated 24h after seeding). After treatment, culture media was removed, and cells were lysed with nucleotide releasing buffer with gentle shaking for 5 min. ATP monitoring enzyme was then added to cell lysates and luminescence was read in a Biotek Synergy HT spectrophotometer (Biotek;

Winooski, VT).

4.3.2.10. Cytosolic calcium determination

INS-1 (832/13) cells were treated as described for the FAO assays but also with metabolic modulators dichloroacetate (DCA) (1 mM) and trimetazidine (25 µM) as well as 10 µM endoplasmic reticulum calcium blocker 8-(N,N-Diethylamino)-octyl-3,4,5-trimethoxybenzoate

HCl (TMB-8) (Sigma Aldrich) or 25 nM cyclosporin (Sigma Aldrich) in the presence of 0.5 mM palmitate/1 mM L-carnitine. Cytosolic calcium was measured in INS-1 (832/13) cells using fluo-

3AM (Invitrogen, F1241) charged fluorescent dye that accumulates in the cytosol and emits green fluorescence after binding calcium ions. A stock solution of 10 mM fluo-3AM was diluted to 5 µM in fluo-3AM loading buffer (10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 137 mM NaCl, 5nM KCl, 1 mM Na2HPO4, 5 mM glucose, and 0.5 mM MgCl2

(pH 7.4)]. After treatment, cells were incubated in fluo-3AM loading buffer for 30 min at 37 oC

with gentle agitation after which cells were washed twice with PBS, trypsinized and collected for flow cytometry analysis. Cytosolic calcium was quantified by taking the mean green fluorescence values of live gated cells and normalizing to vehicle control.

4.3.3. Immunoblotting analysis

For all cell culture experiments, after drug treatments cells were washed with chilled phosphate-buffered saline (PBS), and lysed in chilled lysis buffer with 1x protease inhibitor cocktail (Sigma). RIPA lysis buffer (Sigma-Aldrich) was used for INS-1 (832/13) cells and muscle lysis buffer for C2C12 myotubes (20mM HEPES, 10mM NaCl, 1.5mM MgCl2, 1 mM DTT, 20% glycerol, and 0.1% Triton-X100). Cells were scraped, mechanically homogenized using a syringe and centrifuged at 14,000 x g for 20 min at 4 °C and the supernatants were collected. For preparation of all gastrocnemius muscle lysates, 10 mg frozen tissue (wet weight) was washed in chilled PBS, placed in a dounce homogenizer, lysed in muscle lysis buffer and centrifuged at

14,000 x g for 20 min at 4 ºC and the supernatant was collected.

Protein content for cell and tissue lysates was measured using the BCA protein assay kit

(ThermoFisher) according to manufacturer’s protocol. Thirty µg protein lysates were prepared in loading buffer and immunoblotting was performed by heating lysates for 5 minutes at 95 °C and subjecting them to gel electrophoresis on 10% SDS-polyacrylamide gels at 150 V for 75 minutes.

Using a semi-dry transfer apparatus (Bio-Rad) the gels were then transferred at 25V for 45 minutes to a PVDF membrane and blocked with 5% BSA (Sigma) in tris-buffered saline-tween (TBS-T) for 1 hour. The membrane was incubated overnight with the target primary antibody and a loading control (GAPDH or α-tubulin) primary antibody (1:15000 or 1:1000; ThermoFisher) at 4°C. The following primary antibodies were also used at the specified dilutions: pAKTSer473 (9271, 1:500,

60 kDa), AKT (9272, 1:1000, 60 kDa), pAMPKα-Thr172 (2531, 1:500, 62 kDa), AMPKα (2532,

1:1000, 62 kDa), pERK1/2-Thr202/Tyr204 (4370, 1:500, 42 and 44 kDa), Erk1/2 (4695, 1:1000,

42 and 44 kDa), pP38-Thr180/Tyr182 (4511, 1:500, 43 kDa), and P38 (8690, 1:1000, 40 kDa) was purchased from Cell Signaling Technology, myogenin (F5D, 1:200, 34 kDa) was purchased from

Developmental Studies Hybridoma Bank, PGC-1α (ab54481, 1:500, 105 kDa) and CPT1A

(ab128568, 1:500, 88 kDa) were purchased from Abcam. PVDF Membranes were then washed and incubated with the appropriate horseradish peroxidase- (HRP) conjugated secondary antibody

(1:3000) for 1 hour at room temperature. Bands were visualized using clarity enhanced chemiluminescence (ECL) substrates (Bio-Rad) and the ChemiGenius 2 Bio-Imaging System

(Syngene). The approximate molecular weight for each protein was estimated using Precision Plus

Protein WesternC ladder (Bio-Rad) that was electrophoresed and transferred onto the membrane.

4.3.4. Statistical Analysis

Unless otherwise stated, in vitro results are presented as mean ± SEM whereas in vivo results are presented as mean ± SD. Data were analyzed with GraphPad Prism 6.0 (GraphPad

Software, USA) using one or two-way ANOVA with Bonferroni’s or Dunnett’s post hoc analysis for between group comparisons. Standard student’s t-tests were also used where appropriate. P

<0.05 was accepted as being statistically significant. Normality for all data sets was verified by using the Shapiro-Wilk normality test on GraphPad Prism 6.0 and non-parametric tests were used to analyze some data as specified in their figure legends.

4.3.5. Avocatin B oral consumption pilot clinical study

Our previous work has demonstrated that up to 200 mg of AVO is present in the dry weight 95

of one Hass avocado pulp (Ahmed et al., 2018). As such, a single center, double-blind, placebo- controlled, randomized, pilot clinical study was executed to determine the safety of consuming 50

(equivalent to consuming a quarter Hass avocado pulp) or 200 mg (equivalent to consuming one

Hass avocado pulp) AVO per day for 60 days in healthy human participants (NCT03898505). All study protocols were approved by the research ethics board (REB) at the University of Guelph and

Health Canada Natural and Non-prescription Health Products Directorate (NNHPD) (see appendices A-G for regulatory documents related to this clinical trial).

4.3.5.1. Avocatin B and placebo supplement preparation

The investigational product was approved by Health Canada (natural product number

(NPN) 80074296). To create the product, freeze-dried avocado pulp powder was sourced from

Avocado Oil New Zealand Ltd., (Tauranga, New Zealand) and tested for absolute quantities of

AVO per gram of powder using a validated analytical method (Ahmed et al., 2018). Clinical trial material was then standardized to contain low dose (50 mg) or high dose (200 mg) AVO

(Supplementary Figure 4.6). The placebo product was formulated similarly to the investigational product except it only contained the non-medicinal food-grade ingredients used in the test product and other excipients that were used to simulate appearance, smell, texture and taste of the investigational product. Participants in every group consumed 40 g of material per day for 60 days by dissolving/blending the product in 12-16 ounces of a smoothie like diluent (e.g., milk (with or without lactose), soy milk, coconut milk, or fruit juice of the participant's choice). All three products were approximately matched for total calories per serving (see Supplementary Table 4.1 for nutritional composition of each product).

4.3.5.2. Study Participants and study objectives

Healthy volunteers (13 males and 17 females; age range 20-54 years, BMI range 19.1-29.9) were enrolled in the study. Exclusion criteria were: presence of active clinical disease and history of diabetes, hypertension, dyslipidemia, major depressive disorders, chronic liver disorders, kidney disorders, or blood disorders. Volunteers that were pregnant, had history of previous bariatric surgery, or were actively using prescription or non-prescription medications that impact weight gain or loss were also excluded. Baseline characteristics of participants are highlighted in

Table 4.1. All participants provided written informed consent for enrollment in the study and for taking part in all study related protocols. Primary outcome was the analysis of adverse experiences

(AEs) measured throughout the study. Secondary outcomes were to ascertain safety and tolerability which were assessed through clinically relevant changes in standard laboratory evaluations (blood chemistry and hematology) measured on days 0, 30 and 60. Body mass index

(BMI) and glycated hemoglobin (HbA1c) were also measured as part of secondary outcomes.

Laboratory evaluations included serum alanine aminotransferase (ALT) to assess liver function; serum creatinine to assess kidney function; serum creatine phosphokinase to assess muscle injury; total serum bilirubin to assess blood cell lysis, complete blood count to assess overall health, and

HbA1c as part of secondary outcomes. All AEs were rated by the study investigators for intensity and relationship to study drug as outlined in the Common Terminology Criteria for Adverse Events

(CTCAE) version 5.0. All blood collection and analysis was completed at a private diagnostic laboratory (LifeLabs, Canada).

4.3.5.3. Sample size and statistical analysis

The sample size for this protocol was determined as 30 by performing a power analysis on the primary endpoint (adverse event rates). Briefly, an a priori chi-square goodness-of-fit test was used in the software G*Power to calculate sample size needed to detect statistically significant differences in adverse event rates, for a range of effect sizes, between the three trial arms. A plot of effect size vs total sample size was generated which was used to extrapolate sample sizes that would be needed for effect sizes that were calculated for likely adverse event rates. This analysis revealed that if the incidence rate of any given adverse event is 10% in the placebo group and over 20% in the treatment groups, the trial needs a total sample size of 30 to detect statistically significant differences across treatment arms. Buffering for an estimated drop- out rate of 20%, a sample size of 30 was determined to be sufficient.

Mean change from baseline at day 30 and 60 for all laboratory evaluations and body weight were analysed by the non-parametric Kruskal–Wallis – One-Way Analysis of Variance on Ranks. using GraphPad Prism 6.0 (GraphPad Software, USA) where differences with p <0.05 were considered to be statistically significant. AEs were tabulated (for all randomized patients who received at least one dose of study supplements) and analyzed using Fisher’s exact test of independence in Statistical Analysis System (SAS) university edition where differences with p

<0.05 were considered to be statistically significant.

4.4. Results

4.4.1. AVO inhibits FAO and improves glucose tolerance and insulin sensitivity after DIO is established in mice

AVO was assessed in vivo in a conventional treatment model of DIO. C57BL/6J mice were given a standard (STD; 10-13% kcal from fat) or high fat diet (HFD; 60% kcal from fat) for a total of 12 weeks. At week 8, AVO (100mg/kg b.w.) or vehicle control was administered by oral gavage twice weekly for a total of 5 weeks (Fig. 4.1A). At study completion, HFD mice treated with AVO weighed significantly less than HFD-control mice (Fig. 4.1B; t(16)=3.84, p<0.01) with lower gonadal and mesenteric fat pad (GFP and MFP) weights (Fig. 4.1C: GFP: t(16) =3.34; p<0.01;

MFP: t(16) =2.50; p<0.05). AVO had no effect on blood markers of toxicity compared to HFD- control mice (Supplementary Figure 4.1). HFD mice treated with AVO had significantly improved glucose tolerance compared to HFD-control mice (Fig. 4.1D and E; F(2,25) = 51.27; p<0.0001) and also had greater insulin sensitivity (Fig. 4.1F and G; F(2,23)= 6.37; p<0.01). The homeostasis model assessment index for insulin resistance (HOMA-IR) was increased in HFD-control mice

(Fig. 4.1H) whereas AVO-treated-HFD mice had HOMA-IR index values comparable to lean,

STD mice (Fig. 4.1H; F(2,25)= 17.22; p<0.0001). AVO treatment also increased plasma levels of

FFA (Fig. 4.1I; U= 0, p<0.001) and triacylglyercols (TAG) (Fig. 4.1J; U= 0, p<0.001) in ad libitum mice compared to control-HFD ad libitum mice which together are indirect indicators of decreased whole body FAO and increased glucose utilization as has been previously reported (Gao et al.,

2015; Keung et al., 2013; J. R. Ussher et al., 2014). The plasma FFA and triacylglycerol levels of ad libitum lean mice on standard diets were observed to be higher than HFD-mice due to significant differences in carbohydrate and fat content of the two diets.

The activity of muscle pyruvate dehydrogenase (PDH; a mitochondrial enzyme that

facilitates pyruvate oxidation to acetyl-CoA) is suppressed in DIO due to excessive FAO which impairs the balance between fatty acid and glucose oxidation (Guo, 2015). Given its role in the measure of metabolic flexibility in obese and diabetic patients (Blaak et al., 2006; D. Kelley, 2005) and in mouse models of DIO (Turner et al., 2007), we quantified PDH activity in skeletal muscle excised from ad libitum mice at endpoint. HFD-control mice had decreased gastrocnemius muscle native PDH activity consistent with elevated FAO. In contrast, PDH activity in AVO-treated-HFD mice was restored to that of lean, STD mice (Fig. 4.1K; F(2,25)= 9.235, p<0.001) suggesting higher glucose oxidation in AVO-treated-HFD mice. In addition to higher PDH activity, AVO- treated-HFD mice also had increased phosphorylation of AKTSer473, a key marker of insulin signaling, in gastrocnemius muscle compared to HFD-control mice (Fig. 4.1L). Furthermore,

HFD-control mice had higher protein levels of CPT-1A in gastrocnemius muscle compared to

AVO-treated-HFD mice, which eludes to higher rates of FAO in an ad libitum state (Fig. 4.1L).

Collectively, these results suggest AVO inhibited whole body FAO and altered skeletal muscle substrate preference from fatty acids to glucose thereby reversing insulin resistance in a treatment model of DIO.

We also assessed whether AVO, when provided (100mg/kg (b.w.), twice weekly) at the onset of high-fat feeding (Fig. 4.2A), could prevent HFD-induced pathologies. At study completion, AVO had no effect on body weight between STD or HFD groups (Fig. 4.2B) and there were also no changes in blood markers of toxicity (Supplementary Figure 4.2). Interestingly, AVO did not impact glucose tolerance (Fig. 4.2C and D: t(18)=1.055, p<0.306); however, AVO-HFD mice showed trends towards being more insulin sensitive than HFD-control mice (Fig. 4.2E and

F: t(18)=1.782, p<0.092), which was supported by a slight, but statistically significant, decrease in HOMA-IR (Fig. 4.2G: t(18)=2.269, p<0.05). AVO did not impart striking effects in the

100

prevention study as it did in the treatment study, suggesting FAO inhibition is likely most effective only after the onset of DIO-induced mitochondrial perturbations in key metabolically active tissues. The effects of FAO inhibitors in prevention models of DIO have previously not been reported.

101

Figure 4.1. Effects of AVO in a treatment model of DIO. (A) Treatment study design schematic. Mice (N=28) received either a high-fat diet (HFD) or a low-fat standard diet (STD) for 8 weeks. At the end of week 8, mice were randomly divided into three groups: (i) vehicle-STD (STD); n=10, (ii) vehicle-HFD (HFD); n=9, and (iii) HFD mice treated with 100mg/kg b.w., AVO twice weekly (HFD+AVO; n=9). STD and HFD mice were treated with vehicle-control twice weekly. (B) Mouse weights at treatment study termination (week 13). (C) Gonadal (left) and mesenteric (right) fat pads (GFP and MFP) were excised and weighed upon study termination. (D) Effect of AVO on glucose tolerance. For GTT, an i.p. glucose injection was administered, after which blood glucose levels were monitored at regular intervals for 90 min. (E) Area under the curve (AUC) calculated for the glucose excursion curve of the GTT. (F) Effect of AVO on whole body insulin sensitivity. For ITT, an i.p. insulin injection was administered, after which blood glucose levels were monitored at regular intervals for 60 min. (G) AUC calculated for the glucose excursion curve of the ITT. (H) HOMA-IR index, a measure of insulin resistance, was calculated as described in the methods section. (I) Plasma free fatty acids and (J) plasma 102

triacylglycerols (TAG) was measured at endpoint from ad libitum animals. (K) Native pyruvate dehydrogenase (PDH) activity in whole gastrocnemius muscle lysates in the presence of phosphatase inhibitors was measured. (L) Phosphorylation of Akt-Ser473 (a marker of insulin signaling) and CPT-1A enzyme protein levels were assessed via western blot on gastrocnemius muscle lysates from ad libitum animals. Figure shows representative western blots for pAkt-Ser473, total AKT, CPT-1A, and GAPDH (loading control); histograms (right) represent fold change in pAkt-Ser473 and CPT-1A levels relative to STD mice. For (B-C) data presented as mean weight ± S.D., *p<0.05, **p<0.01, unpaired, two-tailed, student’s t-test. For (E, G, H, K) data represents mean ± S.D., N=8-10/group. *p<0.05, **p<0.01; *** p<0.001, one-way ANOVA, Bonferroni’s post hoc test. For (I-J) data represents mean ± S.D., N=9- 10/group for I and n=5 for J, *p<0.05, **p<0.01; *** p<0.001, Mann-Whitney U test. For (L) data represents mean ± S.D. N=5-6/group, **p<0.01; *** p<0.001, one-way ANOVA, Bonferroni’s post hoc test.

103

5 0 A B n .s . Vehicle—STD (n=10)

) 4 0

(

AVO—STD (n=9) t

h 3 0

g i

Vehicle—HFD (n=10) e W

2 0 y

AVO—HFD (n=10) d o B 1 0

D D D D T F T S F S H H + + O O V V A A C D 3 5 n .s . S T D 3 0 0 0 3 0

) S T D + A V O

/ l

H F D ) o

2 5 n

i m

H F D + A V O e

m m

s 2 0 0 0

(

0 o

e 2 0

L l

1 5 /

o a

U 1 0 0 0

m m 1 0 A

( P 5 0 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 D O D O T V F V T im e (m in ) S A H A + + D D T F S H E F n .s . 1 1 4 0 0 1 0

) 3 5 0

L 9

/ H F D

l o 8 A V O + H F D ) 3 0 0

s m m 7

( 2 5 0

6 u

s x

2 0 0

5 /

u o l

U 1 5 0

g 4

m ( o 3 1 0 0

o l 2 B 5 0 1 0 0 O 0 1 0 2 0 3 0 4 0 5 0 6 0 D V F H A + T im e (m in ) D F H

* G 2 4

2 0

1 6

I -

1 2 M

O H 8 4 0

D O D O T V F V S A H A + + D D T F S H Figure 4.2. Effects of AVO in a prevention model of DIO. (A) Prevention study design schematic. Mice (N=39) were randomly divided into four groups: (i) vehicle- STD; n=10, (ii) AVO-STD; n=9, (iii) vehicle-HFD; n=10, and (iv) AVO-HFD; n=10. Supplementation with vehicle or AVO was started at the onset of HFD or STD feeding. (B) Mouse weights at prevention study termination (week 13). (C) Prevention study GTT and (D) calculated AUC. (E) Prevention study ITT and (F) calculated AUC. For (E-F) STD mice data was omitted as mice became hypoglycemic within 30 min of insulin injection. (G) HOMA-IR index for prevention study. For (B, D, F, G) values are mean ± 104

S.D., N=10/group, unpaired, two-tailed, student’s t-test.

4.4.2. AVO inhibits FAO during lipotoxicity in pancreatic β-islet cells which improves glucose stimulated insulin secretion

To determine the mechanism by which AVO ameliorates mitochondrial lipid overload (i.e., lipotoxicity), we first confirmed its ability to inhibit FAO in pancreatic β-cells (INS-1 (832/13) in the presence of excess lipids that cause mitochondrial dysfunction but not cell death (i.e., >250μM palmitate-BSA (PA-BSA) with 1mM L-carnitine for 24h; herein referred to as lipotoxic conditions; Supplementary Figure 4.3). In β-cells, AVO inhibited complete FAO, as determined

14 by measuring [1- C]-palmitate oxidation to CO2 (Fig. 4.3A; F(3,16)= 46.76; p<0.0001), and did not increase levels of acid soluble metabolites (ASM), an indicator of incomplete FAO (Fig. 4.3A;

F(1,16)= 55.59; p<0.0001). FAO inhibition under lipotoxic conditions was also tested using high resolution respirometry (HRR) where AVO inhibited palmitate supported basal and maximal- uncoupled respiration in INS-1 (832/13) cells (Fig. 4.3B; F(6,24)= 14.44; p<0.0001; see

Supplementary Figure 4.4A for example HRR oxygraphs). AVO imparted this activity at concentrations four-fold lower than etomoxir (ETO), a conventional FAO inhibitor used as a positive control, and had no effect on cell viability (Supplementary Figure 4.3).

To determine whether AVO-induced inhibition of FAO could increase glucose utilization

14 in INS-1 cells, we measured oxidation of D-[ C(U)]-glucose to CO2. Under lipotoxic conditions,

AVO caused greater glucose oxidation compared to cells treated with palmitate only (Fig. 4.3C;

F(3,8)= 77.25, p<0.0001). We confirmed this effect using HRR, where β-cells were incubated under high fat/low glucose conditions in the presence or absence of AVO for 24h and then challenged with 10 mM glucose for 15 min prior to injection into the respirometer. Addition of glucose, in the presence of AVO, increased basal and maximal respiration compared to the 105

palmitate-only control (Fig. 4.3D; F(2,24)= 10.81;p<0.0004; see Supplementary Figure 4.4C for

HRR oxygraphs). These results demonstrate that AVO inhibits FAO and increases glucose utilization in pancreatic tissue.

A n 1 4 0 B 1 5 0 o i C O C T L

t # 2

L a

1 2 0 T 1 2 5

d A V O i

A S M C ****

O 1 0 0

t E TO

1 0 0

t v

*** i

a 8 0 # ****

t t

i *** ****

a 7 5 ****

m l **** e

6 0 r

R 5 0 ]

4 0 C

4 O

- 2 5

2 0

[

% 0 0 l l n e e o A O a i t tr O s c a n P V T y o n E a v n o A B m ru e + + o t C A y o A ig P le P P l R c + i O P h e C V C F

C D n

o 1 4 0 2 5 0 i

t C T L

d L

T ** * i 1 2 0

T A V O

x **** 2 0 0

o ** *

1 0 0 t E TO

e o

e 1 5 0 v

** *** i

c v

8 0 t

- r

e 6 0 ]

1 0 0

)

( C

O 4 0

C O

4 5 0

[ %

t 2 0

- D 0 0

% l n e L a i R A O O s c Y n T P T a y o C V P n A E B m + e + + o P t A g o A li C P P R O C F

Figure 4.3. AVO inhibits FAO in INS-1 (832/13) cells and improves glucose oxidation. (A) INS-1 (832/13) cells were treated for 24h with 0.5% BSA/1mM L-carnitine (Control) or 0.5 mM palmitate-BSA/1 mM L-carnitine (PA) or PA+ 25 µM AVO (AVO+PA) or PA+ 100 µM etomoxir (ETO+PA) in low glucose (5.5 mM) media. After 24h, fatty acid oxidation (FAO) was determined by 14 measuring [ C] incorporation into CO2 and acid soluble metabolites (ASMs), representing complete and incomplete oxidation, respectively. (B) Palmitate supported oxygen consumption rate (OCR) was measured using high resolution respirometry (HRR) in INS-1 (832/13) cells treated as described in (A) except with 0.25 mM palmitate-BSA/1mM L-carnitine; basal, uncoupled (oligomycin addition), and maximal uncoupled (FCCP + pyruvate addition) respiration was assessed. (C) INS-1 (832/13) cells were treated for 14 24h as described in (A) and glucose oxidation was determined by measuring [ C] incorporation into CO2 representing complete glucose oxidation. (D) To indirectly assess metabolic flexibility towards glucose utilization, oxygen consumption rate (OCR) was measured using high resolution respirometry (HRR) in INS-1 (832/13) cells treated as described in (A) except with 0.25mM palmitate-BSA/1mM L-carnitine for 24 h. After treatment, cells were trypsinized, collected and subjected to D-glucose addition (10 mM) for 15 min prior to HRR measurements; basal, uncoupled, and maximal uncoupled respiration was assessed as 106

described in (C). For (A) data represent means ± SEM from three independent experiments performed in triplicate. ***p<0.001; ****p<0.0001 vs vehicle control for CO2 group or ASM group, two-way ANOVA, Sidak’s post hoc test. For (B) data represent means ± SEM from three independent experiments, ****p<0.0001 vs vehicle control, two-way ANOVA, Sidak’s post hoc test. For (C) data represent means ± SEM from three independent experiments performed in triplicate, **p<0.01; ***p<0.001; ****p<0.0001 vs PA group, one-way ANOVA, Sidak’s post hoc test. For (D) data represents means ± SEM from three independent experiments, *p<0.05; **p<0.01; vs control group, Sidak’s post hoc test.

Lipotoxicity desensitizes β-islet cells to glucose resulting in reduced GSIS and hyperinsulinemia; hallmarks of obesity and insulin resistance (Biden et al., 2002; Erion et al.,

2015; Ježek et al., 2018; Zhou & Grill, 1994). To further understand the mechanism by which

AVO-induced FAO inhibition improves insulin sensitivity in vivo, we determined its effect on

GSIS, generation of mitochondrial superoxide, cellular ATP levels along with cytosolic calcium levels in INS-1 cells. Lipotoxicity blunted GSIS (Fig. 4.4A; p<0.0001); a phenotype that was rescued in the presence of AVO (Fig. 4.4A; p<0.001; Fig. 4.4B; p<0.001). AVO also lowers palmitate-induced increases in mitochondrial superoxide levels (Fig. 4.4C; F(2,12)= 11.69, p=0.0015), and mitochondrial membrane potential (MMP) (Fig. 4.4D; F(1,12)=109.9, p=0.001).

AVO restores total cellular ATP content in palmitate challenged cells to control levels suggesting metabolic modulation does not lead to an ATP crisis (Fig. 4.4E; F(9,20)= 8.52, p<0.0001).

Together, these results are indicative of AVO’s ability to restore oxidative metabolism and improve efficiency of the electron transport system.

Chronic exposure to excess palmitate can alter calcium homeostasis and cause calcium leak from the endoplasmic reticulum into the cytosol which perturbs GSIS (Dai Ly et al., 2018).

Consistent with these findings, palmitate treated cells had increased cytosolic Ca2+ levels under non-stimulatory glucose conditions where AVO, ETO and the endoplasmic reticulum calcium blocker, TMB-8 were able to reduce cytosolic Ca2+ levels (Fig. 4.4F; F(7,16)= 23.08, p<0.0001).

These results provide direct support to the observation that in lipotoxicity, GSIS was higher in the

107

presence of the FAO inhibitors (i.e., AVO and ETO). Lastly, the effects of AVO in INS-1 cells were determined to be independent of AMP-activated protein kinase (AMPK) activation (Fig.

4.4G) and mitochondrial biogenesis which was assessed via protein levels of PGC-1α (Fig. 4.4G) and the fluorescent dye NAO (Supplementary Figure 4.5A). Collectively, these results demonstrate that AVO acts as a metabolic modulator to restore GSIS in pancreatic β-islet cells under lipotoxic conditions by inhibiting FAO that improves glucose utilization, mitochondrial oxidative stress, and calcium homeostasis.

108

A 8 0 0 3 m M G lu c o se B 4 C

e 2 .5

1 6 m M G lu c o se c

n C T L

)

c x

6 0 0 3 A e

s A V O

t 2 .0

e o

* * * * (

d r ***

o E TO

n L

****

g l

T 1 .5

/ m

* * * * C

/ 4 0 0

(

I * 1 .0

a s 2 0 0 G

1 l

n e

I 0 .5

l o

0 0 F 0 .0 - le A O O O + c O e i P V T V T l A O O O O A A h A E A E c P T V P P e + + i V T V A A h A E A E P P e + + V A A P P

D E F e 2 .0 e 2 5 0 0

2 .0 c c

C T L n

)

l c

- **

c s

A V O o t **

A 2 0 0 0 *** r

s **

e t

n 1 .5

e n

1 .5 ** o

( r

E TO c

c o

o ****

L l

u * *

l T

1 5 0 0

****

1 .0

o 1 .0

v i

e 1 0 0 0

a a

. 0 .5

h l

0 .5 R

h u

5 0 0

(

l o

o 0 .0 0 F

F 0 .0 l e A o o A A A A A - + l A A v O r ic P P P T t P P P P P A + A n + + + A h + E 8 + P P e O o o O O - P t s V V E C V T B A A E M C T

l o

Control PA AVO+PA ETO+PA RGZ r t 1 .5 0 1 .7 5 n **

o *** 

C 1 .5 0

1 .2 5 1 *

o t

pAMPK r

t C

e 1 .2 5

v g

1 .0 0 o

l ** 1 .0 0

t g

AMPK 0 .7 5 **

R e

0 .7 5

o P

0 .5 0 a

PGC-1α F 0 .5 0

R /

0 .2 5 o 0 .2 5

F P

GAPDH M 0 .0 0 0 .0 0 A p l A Z l A A A Z A A o P o P P P G r P P G tr + t + + R + R n O n O O o O o V V T C T C A E A E Figure 4.4. Inhibition of FAO by AVO under lipotoxic conditions in INS-1 (832/13) cells improves glucose stimulated insulin secretion. (A) INS-1 (832/13) cells were treated for 24h with 0.5% BSA/1mM L-carnitine (Control) or 0.5 mM palmitate-BSA/1 mM- L-carnitine (PA) or PA+ 25 µM AVO (AVO+PA) or PA+ 100 µM etomoxir (ETO+PA) in low glucose (5.5 mM) media. After treatment, GSIS assay was performed where insulin secreted into media (KRB) was quantified using an ELISA kit. Insulin in KRB was normalized to total protein content as described in the Methods section. (B) Data from (A) presented as the glucose- dependent insulin secretory index (GSIS16/3), defined as the ratio between the insulin secretion at 16 mM (stimulatory, surrogate to postprandial glucose levels) and 3 mM (basal, surrogate to fasting glucose levels) glucose. (C) Mitochondrial ROS was measured in INS-1 (832/13) cells treated as described in (A) using MitoSOX® Red. (D) Mitochondrial membrane potential was measured in INS-1 (832/13) cells treated as described in (A) using JC-1 dye. (E) Total cellular ATP content was measured in INS-1 (832/13) cells treated as described in (A). (F) Cytoplasmic Ca2+ was measured in INS-1 (832/13) using Fluo-3AM dye. Cells were treated as described in (A) and with 10 µM TMB-8 (endoplasmic reticulum calcium blocker) or 25 nM cyclosporin (mitochondrial calcium blocker) in the presence of PA. (G) Phosphorylation of AMPKα (Thr-172), and PGC-1α protein levels were assessed via western blot on INS-1 (832/13) cells treated as described in (A) and also with 2.5 µM rosiglitazone (RGZ). Figure shows representative western blots for pAMPKα-Thr172, total AMPKα, PGC-1α, and GAPDH (loading control); histograms (right) represent fold change in pAMPK and PGC-1α levels relative to vehicle 109

control. For (A) data presented as mean ± SEM (N=4), **p<0.01; ****p<0.0001 vs vehicle- control for 16 mM glucose group, two-way ANOVA, Dunnett’s post hoc test. For (B) data presented as mean ± SEM (N=4) *p<0.05 vs vehicle, One-way ANOVA, Dunnett’s post hoc test. For (C-D) data are presented as mean ± SEM (N=3), *p<0.05; **p<0.01;***p<0.001; ****p<0.0001 vs CTL (PA- or PA+) group, two- way ANOVA, Dunnett’s post hoc test. For (E) data represents mean ± SEM (N=3), **p<0.01;****p<0.0001 vs Control, One-way ANOVA, Dunnett’s post hoc test; for (F) data represents mean ± SEM (N=3), **p<0.01; ***p<0.001; ****p<0.0001 vs PA group, one-way ANOVA, Dunnett’s post hoc test. For (G) data represents mean ± SEM (N=3), *p<0.05; **p<0.01; ***p<0.001 vs Control group, one-way ANOVA, Dunnett’s post hoc test.

4.4.3. AVO inhibits FAO during lipotoxicity in C2C12 myotubes which improves insulin signaling

Skeletal muscle accounts for approximately 70% of whole body glucose disposal

(DeFronzo et al., 1981); thus, the interplay between pancreatic insulin secretion and skeletal muscle glucose utilization is key to whole-body insulin sensitivity. Similar to INS-1 cells, AVO inhibited complete FAO in C2C12 myotubes (Fig. 4.5A; F(3,20)= 24.64; p<0.0001) and did not increase incomplete FAO as assessed by ASM levels (Fig. 4.5A; F(1,20)= 61.69; p<0.0001).

Inhibition of FAO in C2C12 myotubes was further confirmed with HRR where AVO inhibited palmitate supported basal and maximal-uncoupled respiration (Fig. 4.5B; F(6,24)= 6.074; p<0.001; see Supplementary Figure 4.4B for HRR oxygraphs). Consistent with the known effects of FAO inhibition, AVO increased basal and insulin stimulated glucose oxidation in C2C12 myotubes in the presence of excess palmitate as measured by oxidation of D-[14C(U)]-glucose to

CO2 (Fig. 4.5C; F(3,16)= 5.284, p<0.01) and HRR (Fig. 4.5D; F(2,24)= 6.978, p<0.004; see

Supplementary Figure 4.4D for HRR oxygraphs).

110

B 3 0 0

A n 1 4 0 #

o C T L

i C O t

2 L a

1 2 0 T 2 5 0 A V O

d i

A S M C

o E TO O

1 0 0 t

# 2 0 0

v i

a 8 0 t t ** i ****

a 1 5 0

l m

e **** l

6 0 r a **** ****

P 1 0 0

R ]

4 0 C ** **

4 O

1 5 0 * -

2 0 %

[

0 % 0 l l a in te e ro A O O s c a n t P T a y v o n V n A E B m ru e o + o t C + y o A A ig P e P l + R l P O ic P h C e C V F

C D n

o 1 4 0 2 5 0 i

L # t

T **** C T L

a B a sa l

C i

1 2 0

T l

x In s u lin 2 0 0 ** A V O a

* * * * C

a 1 0 0

e E TO

o 1 5 0

c t

8 0 i

# t

l e

# a

e -

t 6 0 ]

r 1 0 0

)

a * *

R U

e * (

4 0 C

2 O

4 5 0

1 [

2 0 %

- C

t 0 0

% l a in te e L A O O s c a n T y o P V T a v n C A E B m ru e + + o y t A g o A li P P P + R O P C C F Figure 4.5. AVO inhibits FAO in C2C12 myotubes and improves glucose oxidation. (A) C2C12 myotubes were treated for 24h with 0.5% BSA/1 mM L-carnitine (Control) or 0.5 mM palmitate-BSA/1mM L-carnitine (PA) or PA+ 25 µM AVO (AVO+PA) or PA+ 100 µM etomoxir (ETO+PA) in low glucose media. After 24h, FAO was determined by measuring [14C] incorporation into CO2 and acid soluble metabolites (ASMs), representing complete and incomplete oxidation, respectively. (B) Palmitate supported oxygen consumption rate (OCR) was measured using high resolution respirometry (HRR) in C2C12 myotubes treated as described in (A) except with 0.25 mM palmitate-BSA/1mM L- carnitine; basal, uncoupled (oligomycin addition), and maximal uncoupled (FCCP + pyruvate addition) respiration was assessed. (C) C2C12 myotubes were treated as described in (A). After 24h, glucose oxidation was determined in insulin stimulated and non-insulin stimulated (basal) conditions by measuring 14 [ C] incorporation into CO2 representing complete glucose oxidation. (D) Metabolic flexibility towards glucose utilization was measured using HRR in C2C12 myotubes treated as described in (A) except with 0.25 mM palmitate-BSA/1mM L-carnitine for 24 h. After treatment, cells were trypsinized, collected and subjected to D-glucose addition (10 mM) for 15 min prior to HRR measurements; basal, uncoupled, and maximal capacity for electron flux (by FCCP-uncoupling) were then assessed. For (A) data represent means ± SEM from three independent experiments performed in triplicate. **p<0.01; ****p<0.0001 vs vehicle control for CO2 group or ASM group, two-way ANOVA, Sidak’s post hoc test. For (B) data represent means ± SEM from three independent experiments, **p<0.01; ****p<0.0001 vs vehicle control, two-way ANOVA, Sidak’s post hoc test. For (C) data represent means ± SEM from three independent experiments performed in triplicate. *p<0.05; **p<0.01; ****p<0.0001 vs PA basal group, and #p<0.0001 vs PA insulin stimulated group, two-way ANOVA, Sidak’s post hoc test. For (D) data represents means ± SEM from three independent experiments, **p<0.01; ****p<0.0001; vs control group, Sidak’s post hoc test. 111

Products of incomplete FAO have been shown to directly inhibit insulin signaling in skeletal muscle (Boucher et al., 2014; Koves et al., 2008), thus, we evaluated if AVO induced FAO inhibition under lipotoxic conditions in C2C12 myotubes would restore or improve insulin signaling. AVO was able to restore the suppressive effects of lipotoxicity on insulin stimulated phosphorylation of AKTSer473 (Fig. 4.6A; F(3,16)= 11.58; p<0.0003), and also reduce palmitate induced phosphorylation of mitogen-activated protein kinases (MAPK): ERK1/2 and P38 (Fig.

4.6A). Consistent with improving glucose oxidation and reducing the attenuation of insulin signaling under lipotoxic conditions, AVO also increased glucose uptake compared to palmitate only treated cells as measured using the fluorescent glucose analogue, 2-NDBG, which accumulates via the skeletal muscle dominant glucose transporter, GLUT 4 (Zou et al., 2005) (Fig.

4.6B; F(2, 17)= 16.03; p<0.001).

Similar to observations in INS-1 cells, AVO lowered palmitate induced increases in mitochondrial superoxide levels (Fig. 4.6C; F(2,12)= 4.644, p<0.05) and MMP (Fig. 4.6D;

F(1.12)= 83.56, p<0.0001) in C2C12 myotubes. AVO also restored cellular ATP content to that of control cells (Fig. 4.6E; F(9, 20)= 2.951, p= 0.021) suggesting restoration of mitochondrial function under lipotoxic conditions. Consistent with previous literature on metabolic modulation in DIO, the pyruvate dehydrogenase kinase inhibitor, dichloroacetate (DCA)(I. K. Lee, 2014) also restored cellular ATP content (Fig. 4.6E) and insulin signaling (Fig. 4.6F) whereas the partial FAO inhibitor trimetazidine (TRI) did not exert such effects (J. R. Ussher et al., 2014). Finally, the effects of AVO in C2C12 myotubes were determined to be independent of AMPK activation (Fig.

4.6G) or mitochondrial biogenesis (Fig. 4.6G; Supplementary Figure 4.5B).

Collectively, AVO reduces the negative impact of lipotoxicity on insulin signaling in

C2C12 myotubes by improving mitochondrial function and increasing glucose uptake and

112

oxidation.

A Control PA AVO+PA ETO+PA Insulin - + - + - + - +

pAKTSer473 **** *

1 .4 * **** 6

T )

8 ) ****

l **

)

l K

**** a

B a sa l a

A a

AKT s

s R / 7 **

s **** 5

3 a

7 /

In s u lin b

(

4 b

( 1 .2

r K

** (

6 l

o 8

R 4

t r pErk T 5

n K

***

8 e

o 3 A 1 .0

4 C

** o

e t

t 3

g 2

e e

Erk n

a v

0 .8

t t

i 2

l l

l 1

e 1

R l

pP38 R o 0 0 0 .6 F l l l o A A o A A A o A A r A tr P r A t P P P P P t P P P n + + n + + n + + o o O o O O O P38 O O C V T C V T C V T A E A E A E

GAPDH

B 4 0 C D E e 1 .6 e 3 0 0 0

c 1 .8

) n

n C T L

3 0 C T L l

)

- c

G 2 5 0 0

c t

A A V O A

s A V O s D 1 .6

n * *

e P

2 0 e

e )

B 1 .4

(

E TO

N o

o E TO n

o 2 0 0 0

t e

u *

o 2

l *

1 0 t n

1 .4 c

F **

c * 1 5 0 0

n v

o 1 .2

0 i

o e

1 .2 l

p v

% 1 0 0 0

(

-1 0 L

c l

1 .0 l

n e

e .

I 1 .0 5 0 0

-2 0 (

l o o 0

F 0 .8 -3 0 F 0 .8 I le A A A A A o O A - + - + c v R L A O O i P P P P P A T C T T P T A A A A h + + + + E C V P o I D A E P P P e O t A R + + V V C E T A A P A D P

F T 2 5 G

)

o r

A *

a **** t / 2 .0 1 .2

s pAMPK

n a 7

2 0 *** o

 (

*** C

l 1

l **** 1 .0

o o

AMPK t

T t 1 .5 C

t 1 5

G v

g 0 .8

PGC-1α C

e o

1 0 t g

t 0 .6

g 1 .0

n GAPDH

t o

P 0 .4 a

5 C

C F

e d

e 0 .5

/ o

R 0 .2

o K

0 F

l l

A A

o o

P r

r 0 .0 t

t 0 .0

I n n

A l O

p T o o A A O l A T A A A R A A A A o

o r P P P P P E C V E

T r P P P P t C C P T t + + + + + M n + + I+ M

A n I D E O A o O O A R o O R V C C V T C T C T T A E D A E D Insulin - + + + + + + pAKTSer473 AKT GAPDH Figure 4.6. Inhibition of FAO by AVO under lipotoxic conditions improves insulin signaling in C2C12 myotubes. (A) C2C12 myotubes treated with 0.5% BSA/1 mM L-carnitine (Control) or 0.5 mM palmitate-BSA/1mM L-carnitine (PA) or PA+ 25 µM AVO (AVO+PA) or PA+ 100 µM etomoxir (ETO+PA) in low glucose media for 24 h after which myotubes were starved and stimulated with 100 nM insulin for 30 min prior to lysis for Western Blot. Insulin signaling was measured by assessing phosphorylation of Akt-Ser473. Inflammatory stress was assessed by phosphorylation of ERK1/2 and phosphorylation of P38. Figure shows representative western blots for pAkt-Ser473, total AKT, pERK1/2, total ERK, pP38, total p38, and GAPDH (loading control); histograms (right) represent fold change in pAkt-Ser473, pERK1/2, and pP38 relative to vehicle control (basal). (B) Insulin stimulated glucose uptake was measured using the fluorescent glucose analogue 2-NDBG via flow cytometry in C2C12 myotubes treated as described in (A). Data presented as % incorporated 2-NDBG relative to control. (C) Mitochondrial ROS was measured using MitoSOX® Red in C2C12 myotubes treated as described in (A). (D) Mitochondrial membrane potential was measured in C2C12 myotubes treated as described in (A) using JC-1 dye. (E) Total cellular ATP content was measured in C2C12 myotubes treated as described in (A) as well as with two other metabolic modulators: 1 mM 113

Dichloroacetate (DCA), and 25 µM trimetazidine (TRI) in the presence or absence of PA. (F) The effect of DCA and TRI on insulin signaling (pAkt-Ser473) in C2C12 myotubes was also assessed via western blot as described for (A). (G) Phosphorylation of AMPKα (Thr-172), and PGC-1α protein levels were- assessed via Western Blot on C2C12 myotubes treated as described in (A) and also with 1 mM metformin (MET). Figure shows representative western blots for pAMPKα-Thr172, total AMPKα, PGC-1α, and GAPDH (loading control); histograms (right) represent fold change in pAMPK and PGC-1α levels relative to vehicle control. For (A) data are expressed as mean ± SEM (N=4), *p<0.05; **p<0.01; **** p<0.0001 vs basal or insulin stimulated PA group (as indicated), two-way ANOVA, Sidak’s post hoc test. For (B) data presented as mean ± SEM from 3 independent experiments. *p<0.05, **p<0.01; *** p<0.001 vs control, one-way ANOVA, Dunnett’s post hoc test. For (C-D) data are presented as mean ± SEM (N=3), *p<0.05; **p<0.01; vs CTL (PA- or PA+) group, two-way ANOVA, Dunnett’s post hoc test. For (E) data represents mean ± SEM (N=3), **p<0.01;****p<0.0001 vs Control, One-way ANOVA, Dunnett’s post hoc test; for (F) data represents mean ± SEM (N=3), ***p<0.001; ****p<0.0001 vs Control (basal), one-way ANOVA, Dunnett’s post hoc test. For (G) data represents mean ± SEM (N=3), *p<0.05 vs Control group, one-way ANOVA, Dunnett’s post hoc test.

4.4.4. AVO was well-tolerated in a pilot clinical study

A combination of dose-limiting toxicities and lack of specificity have prevented the clinical use of currently available FAO inhibitors for the treatment of obesity to date (Steggall et al., 2017).

Thus, as a first step in assessing AVO’s potential clinical utility, a Phase I pilot study

(NCT03898505) was undertaken to determine the safety of its general use. Here, 50 or 200 mg of

AVO (roughly equivalent to consuming a quarter or one Hass avocado pulp matter, respectively) or placebo per day was consumed for 60 days and serological and physiological assessments were measured at baseline (day 0), day 30, and 60 (see Fig. 4.7A for study schematic, Table 4.1 for participant baseline characteristics and Supplementary Figure 4.6 for supplement description).

Analysis of adverse events (AE) revealed that AVO was generally well-tolerated with minor gastrointestinal issues noted across all groups (Table 4.2). Three participants in the high dose group developed a skin rash two weeks into supplementation. While this resolved after one-week of supplement discontinuation, one participant resumed supplementation to day 30 whereas two withdrew from the study prior to day 30 (see Supplementary Figure 4.7 for trial CONSORT diagram of participant flow). Supplementation with AVO had no effect on several blood markers of kidney, liver, and muscle toxicity (Fig. 4.7B-E) or complete blood counts (Supplementary 114

Figure 4.8) compared to placebo. Levels of all safety markers were within normal reference ranges as reported by the third-party diagnostic laboratory (LifeLabs Canada) that performed the blood analysis. Interestingly, we observed a trend where decreases in body weight were noted in both

AVO supplemented groups between baseline and day 30 relative to placebo (Fig. 4.7F). Even though the trend in weight change was not statistically significant, due to the small sample size

(p=0.10), it is likely that future larger Phase II efficacy studies will be able to link human supplementation with the noted pre-clinical activity of AVO on body weight. No clinically relevant changes from baseline in body mass index or glycated hemoglobin (HbA1c) were observed in either AVO treated groups compared to placebo (Fig. 4.7G-H). Overall, the results of this pilot clinical study highlight the translational relevance, potential bioactivity and favorable safety profile of AVO.

115

A Placebo N=10 N=10 N=9

Avocatin B N=10 N=10 N=10 Low Dose

Avocatin B N=10 N=8 N=6 High Dose

Time (day): 0 30 60 * * * *denotes ascertainment of primary and secondary endpoints

2 0 B 1 0 C 6 0 D P la c e b o 5 0 1 5

L o w D o se

)

L L

/ 4 0

H ig h D o s e / 1 0 s

0 s

e l

3 0 l

o o

) 5

 /

2 0 

(

n e

0 i

1 0 n T

b -1 0

A l

-5

i 0



r l

a -1 0

C t

-1 0 o

-2 0 

T -2 0

 -1 5 -3 0 -2 0 -3 0 -4 0 0 0 0 0 y y 0 0 a a y y y y a a a a -D -D D D 0 0 -D -D - - 0 0 3 6 0 0 3 6 y y 3 6 a a y y y y a a a a D D D D D D E 4 0 0 F 1 0 G 2

) 8

/ U

( 2 0 0

6 1

e s

a 4

)

g k

0 k

o m

( 2 0

s (

g 0

e h

-2 0 0 M P -2

-1



n i

t -4 a

e -4 0 0

r -6 -2

 -8 -6 0 0 -1 0 -3 0 0 y y 0 0 0 0 a a y y y y -D -D a a a a D D 0 0 -D -D - - 3 6 0 0 0 0 y y 3 6 3 6 a a y y y y D D a a a a D D D D

H 0 .4 0 .3

0 .2 )

0 .1

(

1 0 .0

b H

-0 .1  -0 .2 -0 .3

-0 .4

0 0 y y a a -D -D 0 0 3 6 y y a a D D

Figure 4.7. AVO does not impart toxicity after human oral consumption. (A) Schematic of clinical study design. Healthy human participants consumed placebo, low dose AVO or high dose AVO supplement once per day for 60 days. Participants dropped out of the study at different times as denoted by the “N” value at day 0, 30 or 60 of the clinical trial. Mean changes between day 30- day 0 and day 60-day 0 are presented for (B) total bilirubin, (C) alanine aminotransferase (ALT), (D) creatinine, (E) creatine phosphokinase, (F) body weight, (G) body mass index (BMI), (H) glycated hemoglobin (HbA1c). Data expressed as mean ± S.D., Kruskal–Wallis – One-Way Analysis of Variance on Ranks. 116

4.5. Discussion

High dietary fat intake has long been associated with insulin resistance and studies directly point to mitochondria of metabolically active tissues as a central therapeutic target. While our findings are unable to reconcile the current controversy on whether HFD-induced insulin resistance precedes or is followed by mitochondrial dysfunction, our results clearly point towards a shift in metabolic substrate preference as being an important mediator of insulin sensitivity.

These findings are in agreement with a growing body of literature supporting FAO inhibition as a strategy in the management of DIO-induced insulin resistance (Lopaschuk, 2016; Zhang et al.,

2010). The most compelling evidence in support of excess FAO contributing to the pathophysiology of HFD-induced insulin resistance is from transgenic studies where mice engineered to increase flux through FAO develop insulin resistance. In contrast, transgenic mice with limited FAO capacity challenged with HFDs had normal insulin sensitivity (Finck et al.,

2005; Guerre-Millo et al., 2001; Koves et al., 2008; Wicks et al., 2015).

Accelerating FAO as a means to process excess FFAs has been proposed as a therapeutic strategy to overcome nutrient overload (Abu-Elheiga et al., 2003; Bruce et al., 2008; Dobbins et al., 2001; Henique et al., 2010; Iglesias et al., 2002; Sebastián et al., 2007). However, these studies question the traditional view of the Randle Cycle in skeletal muscle and argue that impaired glucose oxidation is a direct result of defective insulin stimulated glucose uptake, neither of which is related to increased FAO (Roden et al., 1996; Wolfe, 1998). More recent studies, however, point to nutrient overload causing pathologies as a result of high rates of incomplete FAO (that produces lipotoxic acyl carnitines) and/or increases in oxidative stress (i.e., excessive ROS production)

(Anderson et al., 2009; Koves et al., 2008). Moreover, HFDs cause insulin resistance in rodents while increasing skeletal muscle mitochondrial biogenesis and β-oxidation (Hancock et al., 2008;

117

Turner et al., 2007) where induced alterations in skeletal muscle mitochondrial structure and function occur only after the onset of insulin resistance (Bonnard et al., 2008). In further support of this are key observations that skeletal muscle oxidative capacity is far in excess of the energy demands of resting muscle (Andersen & Saltin, 1985) and that despite reductions in mitochondrial content observed in obese and diabetic patients, insulin-resistant skeletal muscle have normal mitochondrial function (Boushel et al., 2007; Holloway et al., 2007). Our study further adds to this literature in that the striking effects of restoring insulin sensitivity in vivo were only evident when AVO was provided as a treatment to DIO and not as a preventative agent. This supports the notion that DIO causes elevated FAO and that inhibition of FAO is only necessary when it is dysregulated (i.e., aberrantly elevated) (Collier et al., 1993; Dobbins et al., 2001; Gao et al., 2015;

Keung et al., 2013; Koves et al., 2008; J. R. Ussher et al., 2014).

The ability of AVO to inhibit FAO, lower ROS and improve glucose oxidation under lipotoxic conditions point toward the Randle Cycle as an important mediator of insulin resistance.

The interplay between pancreatic β-cell insulin secretion and skeletal muscle glucose uptake is often underplayed in scientific literature, likely owing to the larger role of skeletal muscle in absorbing systemic glucose. To our knowledge, no study to date has examined the effects of small molecule-induced FAO inhibition on glucose oxidation or insulin secretion in pancreatic tissue.

Importantly, a pilot in vivo study to assess pharmacokinetics showed accumulation of AVO in both skeletal muscle (soleus and gastrocnemius) and pancreas following oral administration (data not shown). Our study shows that AVO inhibits excessive FAO in β-cells which improved mitochondrial glucose oxidation and restored GSIS (through improved mitochondrial function and reduced ER calcium leak). These improvements in pancreatic β-cell function along with the improved insulin sensitivity in skeletal muscle likely combine to contribute to the beneficial

118

whole-body insulin responsiveness and glucose tolerance observed in vivo. AVO also blocked palmitate-induced mitochondria-derived ROS and did not impact incomplete FAO in both INS-1 and C2C12 myotubes. This is critical to AVO’s favorable mechanism of action given previous work suggesting that lowering ROS or incomplete FAO is key to reversing insulin resistance in

DIO (Anderson et al., 2009; Gavin et al., 2018; Koves et al., 2008). Collectively, and in keeping with Randle’s original hypothesis of the mutually exclusive fuel selection process between glucose and fatty acid substrates, inhibiting and not accelerating FAO by the addition of molecules like

AVO is key to increasing glucose utilization and reversing insulin resistance.

The preclinical findings reported in this work also warrant further investigation on AVO’s mechanism of action. First, determining the inhibitory effects of AVO and its individual constituents (avocadene and avocadyne) on all enzymatic steps of the mitochondrial β-oxidation pathway in muscle and pancreatic cells in vitro and in vivo is important to define a structure- function relationship. Our initial report that AVO accumulates in mitochondria to induce cell death in leukemia cell lines independent of CPT-1 (E. A. Lee et al., 2015) suggests AVO targets a specific enzymatic step of the β-oxidation pathway. Second, AVO’s effects on peroxisomal FAO

(a key organelle interconnected with mitochondria (Fransen et al., 2017)) need to be determined as peroxisomal FAO and lipotoxicity are strongly correlated in muscle (Wicks et al., 2015) and β- islet cells (Elsner et al., 2011).

AVO was well tolerated in mouse and human studies which makes it superior to conventional FAO inhibitors that impart dose-limiting toxicities such as ETO (Holubarsch et al.,

2007; Timmers et al., 2012). Other pharmacological agents that directly or indirectly inhibit FAO have shown a wide-range of protective benefits in DIO (Gao et al., 2015; Keung et al., 2013; J. R.

Ussher et al., 2014; J R. Ussher et al., 2016); however, these are limited by their variable pre-

119

clinical effectiveness, availability and/or lack of clinical testing. Indeed, ranolazine, a partial FAO inhibitor used to treat angina, was recently approved for use in combination therapies for obesity and diabetes; however; its beneficial activity is primarily in the liver via a mechanism not related to FAO (Batran et al., 2019; Pettus et al., 2016; J. R. Ussher et al., 2014). Moreover, these pharmacological agents have no reported activity in pancreatic tissue, the site of insulin secretion.

In this study, we directly link FAO inhibition with improved insulin sensitivity through increased glucose utilization and modulation of ROS using a novel small molecule that imparts activity in both pancreatic and skeletal muscle tissue. In a Phase I study, AVO exerted no toxicity. Taken together, through metabolic modulation, AVO has potential clinical utility and warrants further development as an agent for the treatment of obesity and associated metabolic disorders.

4.6. Acknowledgments

We are grateful to Advanced Orthomolecular Research (AOR, Canada) for their assistance in manufacturing the clinical trial material and obtaining the natural product number (NPN). We are also very grateful to Drs. Kim Bretz and Mary Michele Worndl for assisting us with all clinical trial related activities. We thank Drs. Jaime Joseph (University of Waterloo) and Joe Quadrilatero

(University of Waterloo) for the generous gift of the INS-1 (832/13) and C2C12 cells, respectively.

We thank all the members of the central animal facility at the University of Guelph for their support with all animal studies.

4.7. Author contributions

PAS and NA conceptualized the experimental design, performed experiments, analyzed data and wrote the manuscript. MT assisted NA in performing HRR experiments. AR assisted NA

120

with all animal experiments. MB assisted NA in completing INS-1 cell line experiments. PJ assisted NA with Western Blot analysis. TAA and KR performed radiolabeled experiments and analyzed data. All authors read and revised the manuscript.

4.8. Supporting Information (Chapter 4)

A B C 1 8 0 0 .6 1 1 1 6 0 1 0

) 0 .5 9

) L

/ 1 4 0

)

g /

L 8

(

1 2 0 0 .4

(

2 n

b r

1 0 0 0 6

c 1

l 0 .3

g t

8 0 ( 5

o a

C 4 m

6 0 m 0 .2

B e

e 3

a R 4 0 H H 0 .1 2 2 0 1 0 0 .0 0 S T D H F D H F D + A V O S T D H F D H F D + A V O S T D H F D H F D + A V O

D E F 8 1 .2 7

7 ) )

1 .0 L 6

9 )

6 9

L 0

/ 5

1 0 .8

x 9

5 (

(

1 4

l e

x 4 i

0 .6 t

(

y c

p 3

C 3 o

B r

0 .4 h

p u

W 2 2

m N

1 0 .2 y 1 L

0 0 .0 0 S T D H F D H F D + A V O S T D H F D H F D + A V O S T D H F D H F D + A V O

G H I 0 .8 0 .0 8 1 5 0 0

0 .0 7 )

) 1 2 5 0

L )

/ 0 .6

L 0 .0 6

/ 9

9 0

1 0 .0 5 1 0 0 0

1 (

0 .4 x

( 0 .0 4

(

l s

i 7 5 0

i t

h 0 .0 3

e p

0 .2 l

e o 0 .0 2 t

n 5 0 0

a o 0 .0 1 P

0 .0 B

E 2 5 0 0 .0 0 -0 .2 -0 .0 1 0 S T D H F D H F D + A V O S T D H F D H F D + A V O S T D H F D H F D + A V O Supplementary Figure 4.1. Complete blood count with differential (CBC w/diff) analysis on whole blood collected from 3-5 animals per group at treatment-DIO study endpoint. (A) Hemoglobin. (B) Hematocrit. (C) Red blood cells (RBC). (D) White blood cells. (E) Neutrophils. (F) Lymphocyte. (G) Eosinophils. (H) Basophils. (I) Platelets. N=3-5/group, one-way ANOVA with Tukey’s post hoc test.

121

Supplementary Figure 4.2. Complete blood count with differential (CBC w/diff) analysis on whole blood collected from 3-5 animals per group at prevention-DIO study endpoint. (A) Hemoglobin. (B) Hematocrit. (C) Red blood cells (RBC). (D) White blood cells. (E) Neutrophils. (F) Lymphocyte. (G) Eosinophils. (H) Basophils. (I) Platelets. N=3-5/group, one-way ANOVA with Tukey’s post hoc test.

122

A. B.

1 1 0 1 1 0 1 0 0 IN S -1 (8 3 2 /1 3 ) 1 0 0 - P A -B S A 9 0 C 2 C 1 2 M y o tu b e s 9 0 + P A -B S A (2 5 0  M ) 8 0 8 0 + P A -B S A (5 0 0  M )

7 0 7 0

l b

b 6 0 6 0

i i

V 5 0

V 5 0 % % 4 0 4 0 3 0 3 0 2 0 2 0 1 0 1 0 0 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 l ro M M t   n 5 0 o [P A -B S A ,  M ] 2 5 C o o le v v ic A A h e V C. D. 1 1 0 Day 1 0 0 - P A -B S A 0 5 6 9 0 + P A -B S A (2 5 0  M ) 8 0 Myogenin + P A -B S A (5 0 0  M )

7 0

e l

(33-40 kD) b 6 0

a i

α-Tubulin V 5 0 (50 kD) % 4 0 3 0 2 0 1 0 0 l ro M M t   n 5 0 o 2 5 C o o le v v ic A A h e V

Supplementary Figure 4.3. Cell viability assessments. (A) INS-1 (832/13) cells or day five C2C12 myotubes were treated for 24h with 0.5% BSA/1mM L- carnitine (Control) or 0.1-0.5 mM palmitate-BSA/1mM L-carnitine in low glucose media (5.5 mM) after which cell viability was measured via flow cytometry using propidium iodide. (B) INS-1 (832/13) cells were treated for 24h with 25 or 50 µM avocatin B in the absence or presence of 0.25 or 0.5 mM palmitate-BSA/1mM L-carnitine after which cell viability was measured via flow cytometry using propidium iodide. (C) C2C12 cells were differentiated in differentiation media for 5-6 days, lysed and collected for Western Blot analysis of myogenin (skeletal muscle specific protein). (D) Day 5 C2C12 myotubes were treated for 24h with 25 or 50 µM avocatin B in the absence or presence of 0.25 or 0.5 mM palmitate-BSA/1mM L-carnitine after which cell viability was measured via flow cytometry using propidium iodide. All data represents mean ± S.D. N=2.

123

A. CTL B.

oli FCCP+Pyr Rot AVO

oli FCCP+Pyr Rot

C. D. CTL CTL

AVO oli FCCP+Pyr Rot oli FCCP+Pyr Rot AVO

oli FCCP+Pyr Rot oli FCCP+Pyr Rot

Supplementary Figure 4.4. High resolution respirometry (HRR) example oxygraphs. (A-B) PA-BSA supported respiration was measured in (A) INS-1 (832/13) cells or (B) day 5 C2C12 myotubes by treating with FAO inhibitors (avocatin B or etomoxir) in the presence of 0.25 mM PA- BSA/1 mM L-carnitine for 24h after which cells were trypsinized, collected and injected into the respirometer. (C-D) To assess metabolic flexibility, (C) INS-1 (832/13) cells or (D) day 5 C2C12 myotubes were treated with FAO inhibitors and PA-BSA as previously described, after which cells were trypsinized, collected and subject to a 15 min incubation with 10 mM D-glucose before being washed and injected into the respirometer.

124

A. B.

1 .5 C T L C T L 1 .5 A V O

e A V O e

c * *

)

- n

) E TO

e -

E TO e

s s

( 1 .0 1 .0

(

t e

0 .5 0 .5

a o e 0 .0

F 0 .0 M - - + + A A A A P P P P

Supplementary Figure 4.5. Mitochondrial mass determination. INS-1 (832/13) (A) or C2C12 myotubes (B) were treated for 24h with 0.5% BSA/1 mM L-carnitine (Control), or 0.5 mM palmitate-BSA/1 mM L-carnitine (PA), or PA+ 25 µM AVO (AVO+PA), or PA+ 100 µM etomoxir (ETO+PA), or AVO or ETO alone in low glucose media (5.5 mM). Cells were then collected and stained with NAO dye for mitochondrial mass determination. Data represents means ± SEM from three independent experiments, *p<0.05; vs PA- or PA+ control group, Dunnett’s post hoc test.

Sourced freeze-dried Hass avocado pulp Clinical Trial Formulations powder (New Zealand) Analytical testing Low dose (1 serving= 5.61 g avocado powder= 50 mg avocatin B)

High dose (1 serving= 22.22 g avocado powder= 200 mg avocatin B)

Placebo

Supplementary Figure 4.6. Method of preparation for avocatin B clinical trial formulations.

125

Supplementary Table 4.1. Macro-and-micro nutrient composition of avocatin B clinical trial formulations

126

Supplementary Figure 4.7. CONSORT participant flow diagram for AVO clinical trial 127

A B 1 8 0 0 .5 1 6 0

) 0 .4

1 4 0 )

(

1 2 0 D a y 0 (

t i i 0 .3

D a y 3 0 r b

1 0 0 c

o l

D a y 6 0 o

g a

o 8 0

m 0 .2 m

e 6 0

H H 4 0 0 .1 2 0 0 0 .0

e e o e o e s s s s b b o o o o e D e D c D c D la w h la w h P o g P o g L i L i H H

C 6 D 1 0 0

5 8 0

t n

4 )

) L

f 6 0

( /

1 V

0 3

x B

(

M 4 0 R 2 2 0 1

0 0

e e o s e o s e b s b s o o o o e D e D c D c D la w h la w h P o g P o g L i L i H H E F 3 6 4 0 0 3 2

2 8

3 0 0

) )

2 4 L /

(

2 0 C

H 2 0 0 H

C 1 6

C M

1 2 M

1 0 0 8 4 0 0

e e o e o e s s s s b o b o o o e D e D c D c D la w h la w h P o g P o g L i L i H H

G 1 5

) 1 0

(

D R 5

e o s e b s o o e D c D la w h P o g L i H Supplementary Figure 4.8. Complete blood count with differential (CBC w/diff) analysis on whole blood collected from all clinical trial participants on day 0, day 30, and day 60 (endpoint). (A) Hemoglobin. (B) Hematocrit. (C) Red blood cells (RBC). (D) Mean corpuscular volume. (E) Mean corpuscular hemoglobin. (F) Mean corpuscular hemoglobin concentration. (G) Red cell distribution width. N=3-10/group, Kruskal–Wallis – One-Way Analysis of Variance on Ranks.

128

H I 1 2 1 0

1 0 8

D a y 0 )

L /

8 D a y 3 0 9

n 1

) 6 u

D a y 6 0 x

(

o /

C l

p B

(

o 4

t W

4 u

e N 2 2

0 0

o e e o e e b s s b s s e o o e o o c D c D a D a D l w h l w h P o g P o g L i L i H H J K 4 1 .0

0 .8

) )

L 3

9 0

1 1

x 0 .6

(

s s

t 2

c c o 0 .4

m M y 1 L 0 .2

0 0 .0

e o e e o s e b s s b o s e o o e o c D c D D a D la l w h w h P o g P o ig L i L H H

L 1 .0 M 0 .1 5

0 .8

) 0 .1 0

)

/ 9

0 .6 9

0 0

(

l s i 0 .4 0 .0 5

n i

s 0 .2

o B E 0 .0 0 0 .0 -0 .2 -0 .0 5

o e e o e e b s s b s s e o o e o o c D c D a D a D l w h l w h P o g P o g L i L i H H

N 4 0 0

3 0 0

)

L /

x (

2 0 0

l e

l P 1 0 0

o e e b s s e o o c D a D l w h P o g L i H Supplementary Figure 4.8 continued. (H) White blood cells. (I) Neutrophils. (J) Lymphocytes. (K) Monocytes. (L) Eosinophils. (M) Basophils. (N) Platelets. 129

Chapter 5: Integrated discussion, conclusions and future work

5.1. Preamble

This thesis addresses a controversial topic in obesity, diabetes and metabolism research.

The concept of metabolic flexibility is used to reconcile opposing views on whether treatment strategies should aim to increase or decrease FAO in key tissues affected by obesity and insulin resistance. Here, the use of a metabolic modulator like AVO is proposed as an effective treatment strategy to spare glucose oxidation in the context of excessive FAO and low energy expenditure.

A validated and sensitive bioanalytical method was used to develop optimal formulations of AVO that showed appreciable bioaccessibility and bioavailability for preclinical and clinical studies.

More importantly, cell culture, preclinical, and clinical data presented here provides mechanistic details of how AVO safely and effectively modulates lipid and glucose metabolism in the mitochondria of key tissues affected by lipotoxicity and insulin resistance. This chapter will highlight limitations of the current work in order to propose specific objectives for future work that is required for further development of avocado PFAs as metabolic modulators.

5.2. Summary and discussion of major findings

Chapter 2 outlines the use of a high-resolution mass spectrometer to develop a sensitive

LC-MS method for the detection and quantitation of avocadene and avocadyne in Hass avocado seed and pulp. This method was fully validated for selectivity, linearity, accuracy, and precision and requires no extensive sample preparation or additional structural validation using surrogate analytical techniques like NMR. This method was further modified to operate in single-ion- monitoring (SIM) mode which allowed for extremely low limits of detection for both avocadene and avocadyne. The LC-SIM-MS method was then validated for the quantitation of avocadene and

130

avocadyne in complex biological matrices like TIM-1 compartments, mouse whole blood and tissues (Chapter 3). This chapter also reports the suitability of the Folch method (Folch et al., 1957) for the extraction and acceptable recovery of avocadene and avocadyne from several matrices.

Methanolic potassium hydroxide saponification of total lipids extracted from seed and pulp via

Folch method resulted in the discovery that a significant portion of PFAs like avocadene and avocadyne are bound to ester linkages.

Utilizing the analytical method developed in Chapter 2, the next set of studies evaluated the bioaccessibility and bioavailability of avocadene and avocadyne from lyophilized avocado pulp powder or an O/W microemulsion comprised of pure AVO. Here, lyophilized avocado pulp powder with known absolute amounts of avocadene and avocadyne was subjected to two in vitro simulated digestion models (i.e. static and dynamic (TIM-1)) to determine if bound forms of avocadene and avocadyne could be hydrolyzed by lipolytic enzymes like lipase and colipase. It was hypothesized that a very small percentage of avocadene and avocadyne would be bioaccessible in vitro as mammalian lipase has been shown to have low affinity for wax esters compared to complex lipids like TAGs, DAGs, MAGs, phospholipids and cholesterol esters.

Contrary to what was expected, the bioaccessibility of avocadene and avocadyne was determined to be upwards of 50% which was suggestive of the possibility that avocadene and avocadyne could be bound to TAGs and complex lipids other than wax esters. Pure AVO was formulated in a microemulsion with a mean droplet size of 25 nm that was subject to TIM-1 digestion in order to compare bioaccessibility with the lyophilized avocado powder. The pure AVO emulsion formulation was determined to possess moderately greater bioaccessibility for avocadene compared to the avocado pulp powder where stark differences in bioaccessibility kinetics were noted. Furthermore, the pure AVO emulsion formulation showed appreciable accumulation in

131

mouse whole blood and tissues after a 100 mg/kg-b.w., oral dose was administered in a pilot pharmacokinetic study. The results of Chapter 3 are seminal in nature as the bioaccessibility and bioavailability of very long chain fatty alcohols like avocadene and avocadyne has never been reported.

Utilizing the delivery systems for AVO developed in Chapter 3, preclinical and clinical studies were carried out to investigate the main hypothesis of this thesis: AVO acts as a metabolic modulator to delay or reverse pathologies associated with obesity and lipotoxicity in key metabolic tissues. Here, a 13-week treatment study was carried out in mice on HFD where AVO or vehicle

(as part of an O/W emulsion) was administered orally twice weekly for 5 weeks prior to which obesity and insulin resistance was established in mice due to 8 weeks of high fat feeding. AVO improved glucose tolerance, reversed insulin resistance, slowed weight gain, and increased skeletal muscle glucose utilization in the post-prandial state compared to control mice on HFD. In contrast,

AVO had no such effects when it was administered at the onset of HFD feeding in a prevention study. The results of these preclinical studies were in good agreement with current literature which specifies the use of metabolic modulators solely in the context of dysregulated mitochondrial substrate utilization and efficiency (Keung et al., 2013; Koves et al., 2008; Lopaschuk, 2016;

Muoio, 2014; Muoio & Newgard, 2008a). Furthermore, in vitro models of lipotoxicity in C2C12 myotubes and INS-1 pancreatic β-cells were utilized for mechanistic studies. In this model, AVO directly shifted mitochondrial substrate utilization from FAO to glucose oxidation independent of inducing mitochondrial biogenesis in cells exposed to high concentrations of palmitate-BSA. This metabolic modulation that AVO exerts in vitro was subsequently shown to reduce insulin resistance in C2C12 myotubes and enhance glucose stimulated insulin secretion (GSIS) in INS-1 cells, respectively. Lastly, in collaboration with an industry partner (Advanced Orthomolecular

132

Research), a natural health product containing lyophilized avocado pulp powder was manufactured to conduct a Health Canada approved randomized, double-blind, placebo-controlled clinical study in healthy human participants to determine the safety of AVO consumption. The results of this clinical trial suggested that the consumption of 50 or 200 mg AVO per day for 60 days had no dose limiting toxicity as determined by hematological markers of kidney, liver, and muscle damage. The occurrence of skin rash in three 200 mg AVO/day participants was noted and allowed for the determination of amounts of lyophilized avocado pulp powder that can be safely consumed in healthy humans. More interestingly, a trend towards reduction in total body weight from baseline was observed at day 30 in both 50 and 200 mg AVO/day participants compared to placebo, a finding that was statistically inconclusive but in line with the observed effects of AVO in preclinical models.

In summary, AVO is a bioaccessible avocado-derived FAO inhibitor which shows significant clinical potential. FAO inhibitors for use in obesity and associated disorders are generally limited by their lack of safety (Holubarsch et al., 2007; Vickers, 2009) and lack of efficacy in target tissues (Batran et al., 2019; J. R. Ussher et al., 2014). This work reports for the first time the safety and bioactivity of AVO towards improving various obesity associated pathologies. Moreover, given the bioaccessibility and bioavailability of the formulations presented in this thesis (avocado pulp powder or O/W emulsion delivery system) it is plausible AVO’s actions can be achieved through oral supplementation.

5.3. Strengths, limitations and future work

The work presented in this thesis comprehensively highlights avocado polyhydroxylated fatty alcohols to be potent bioactive molecules that target mitochondria and influence substrate

133

utilization in the context of lipotoxicity or excessive FAO. The overall strength of this work lies in the systematic presentation of in vitro, in vivo, and clinical evidence in support of a hypothesis that stems from the initial work performed with AVO in models of acute myeloid leukemia (E. A.

Lee et al., 2015). The development of a validated, sensitive and reliable analytical method is a significant contribution to the literature which is currently lacking high resolution—mass spectrometry based methods. The analytical method developed in this work requires no derivatization of extracted samples or complex sample preparation and can be employed for a variety of biological matrices. Furthermore, this work uses gold standard in vitro digestion models to present new data showcasing the appreciable bioaccessibility of avocadene and avocadyne from avocado pulp digestion. Chapter 3 also reports the development of a novel self-emulsifying in vivo delivery system for AVO which elegantly highlights the cosurfactant-like properties of avocadene and avocadyne in O/W emulsions. The bioavailability of AVO delivered via emulsion in vivo also addresses a gap in literature where long chain fatty alcohols are generally assumed to bypass absorption and accumulation in tissues. Lastly, chapter 4 presents definitive evidence of AVO’s ability to inhibit FAO and restore glucose oxidation in vitro and in vivo which aligns directly with a recent cluster of high impact studies (cited in chapter 4) which propose FAO inhibition as a viable therapeutic strategy for obesity and associated disorders. The results of the Phase I clinical study related to this work presents strong evidence for AVO’s safety in healthy human participants and provides impetus for future efficacy clinical studies.

The limitations of the current work provide clear objectives for future work. First, the analytical method outlined in chapter 2 still requires the incorporation of synthesized, deuterated internal standards for both avocadene and avocadyne. Such internal standards can greatly enhance reproducibility and reliability and significantly reduce analysis time (Zimmer, 2014). Avocadene

134

and avocadyne have previously been synthesized into four respective stereoisomers (Becker et al.,

1990; Sugiyama et al., 1982) where carbons 2 and 4 were determined to be chiral centers. Based on these studies, the standard for AVO that was utilized in the development of our analytical method is thought to be a 1:1 mixture of (2R,4R)-avocadene and (2R,4R)-avocadyne, however optical activity assays are warranted to further confirm this. Future optimization of the analytical method outlined here should include i) the evaluation of deuterated internal standards for the biologically relevant stereoisomers of avocadene and avocadyne in both LC-MS and LC-SIM-MS modes, ii) optimization of the multi-step gradient of the mobile phase to shorten run time from 30 min to 10-15 min, and iii) determine if the current method can be validated for the detection and quantitation of 19 and 21 carbon PFA analogs as well as relevant monoacetates.

The major limitation of findings reported in chapter 3 was the lack of characterization of the saponifiable and unsaponifiable fractions of total lipid extracts of the lyophilized avocado pulp powder. Several experiments are needed to further develop the work presented in this chapter pertinent to the bioaccessibility of the bound forms of avocadene and avocadyne. First, total quantitation of mono acetates and alkyl furans present in total lipid extracts of lyophilized avocado pulp powder is warranted to rule out these compounds as significant contributors to the bioaccessibility of avocadene and avocadyne. Second, free fatty acid distribution and complex lipid profiles in saponified and unsaponified total lipid extracts of the lyophilized avocado pulp powder needs to be determined using a variety of analytical methods. Third, differential scanning calorimetry should be performed on cold crystallized wax fractions of lyophilized avocado pulp powder lipid extract for subsequent analytical characterization. Lastly, chapter 3 presents results from a pilot in vivo study where AVO emulsion was administered via oral gavage, however a larger scale study is required to report the absolute bioavailability of avocadene and avocadyne

135

and enable a comparisons with TIM-1 non-cumulative bioaccessibility data.

Chapter 4 is comprised of substantial amounts of in vitro, in vivo, and clinical data that corroborates the main hypothesis of this thesis. However, many experiments are required to gain an even more comprehensive understanding of AVO’s mechanism of action in vitro and in vivo.

First, the activity of avocadene and avocadyne should be evaluated separately in vitro and results should be compared to that of AVO as presented in this thesis. Once the potency of avocadene and avocadyne has been elucidated, their individual and combined effects on the FAO pathway should be studied systematically. The catabolic β-oxidation enzymatic spiral for long-chain fatty acid substrates is comprised of very long-chain-CoA dehydrogenase (VLCAD) and the mitochondrial trifunctional protein (TFP) (Wang et al., 2019) . TFP is a tetramer protein containing the activities enoyl-CoA hydratase, 3-hydroxyacyl-CoA dehydrogenase (LCHAD, both a part of the HADHA subunit), and 3-ketoacyl-CoA thiolase (HADHB) (Wang et al., 2019). To date, the HADHB inhibitor, trimetazidine, has been tested for efficacy in the context of DIO and lipotoxicity where no effects on glucose tolerance and insulin sensitivity were reported (J. R. Ussher et al., 2014).

Given each possible mode of FAO inhibition does not result in equivalent biological outcomes, the effects of avocadene and avocadyne on all four steps of β-oxidation must be evaluated in vitro.

These experiments can be performed using previously established techniques that profile short, medium and long chain acylcarnitine and acyl-CoA species in vitro and in vivo (Koves et al., 2008;

John R. Ussher et al., 2010) and through optimization of currently available two-dimensional protein electrophoresis methods that can illustrate physical interactions between proteins from mitochondrial lysates (Wang et al., 2019). Along with these experiments, avocadyne and avocadene induced FAO inhibition should be directly measured in vivo (mice and humans) using indirect calorimetry or other surrogate techniques (Even & Nadkarni, 2012; San-Millán & Brooks,

136

2018). These experiments are essential to establish a definitive structure-activity relationship for avocadene and avocadyne which show great potential and warrant further clinical evaluation.

5.4. Conclusions

This work highlights i) the nutritional significance of avocadene and avocadyne in avocado pulp where results from in vitro digestion models reveal significant bioaccessibility that can lead to physiological effects; ii) ideal delivery systems of AVO for preclinical and clinical studies; iii) the mechanism by which AVO acts as a novel FAO inhibitor which modulates fat and glucose metabolism in cell culture and animal models of lipotoxicity; and iv) the safety of AVO’s long term consumption in healthy human adults. Further studies are needed to provide more comprehensive insight into the structure-activity relationships of avocadene and avocadyne by investigating all steps of the FAO pathway these PFAs target in vitro and in vivo.

137

REFERENCES

Abu-Elheiga, L., Oh, W., Kordari, P., & Wakil, S. J. (2003). Acetyl-CoA carboxylase 2 mutant

mice are protected against obesity and diabetes induced by high-fat/high-carbohydrate diets.

Proceedings of the National Academy of Sciences, 100(18), 10207–10212.

Adams, S. H., Hoppel, C. L., Lok, K. H., Zhao, L., Wong, S. W., Minkler, P. E., … Garvey, W.

T. (2009). Plasma Acylcarnitine Profiles Suggest Incomplete Long-Chain Fatty Acid -

Oxidation and Altered Tricarboxylic Acid Cycle Activity in Type 2 Diabetic African-

American Women. Journal of Nutrition, 139(6), 1073–1081.

Adikaram, N., Ewing, D., Karunaratne, A., & Wijeratne, E. (1992). Antifungal compounds from

immature avocado fruit peel. Phytochemistry, 31(1), 93–96.

Ahmadi, A., Renaud, J., Petryk, J., Hadizad, T., Ruddy, T., Dyck, J., … Mielniczuk, L. (2017).

Inhibition of fatty acid oxidation is associated with improved right ventricular function and

metabolism in two rat models of pulmonary arterial hypertension. Canadian Journal of

Cardiology, 33(10), S167–S168.

Ahmed, N., Smith, R. W., Henao, J. J. A., Stark, K. D., & Spagnuolo, P. A. (2018). Analytical

Method To Detect and Quantify Avocatin B in Hass Avocado Seed and Pulp Matter.

Journal of Natural Products, 81(4), 818–824.

Andersen, P., & Saltin, B. (1985). Maximal perfusion of skeletal-muscle in man. J Physiol,

366(366), 233–249.

Anderson, E. J., Lustig, M. E., Boyle, K. E., Woodlief, T. L., Kane, D. A., Lin, C. T., … Neufer,

P. D. (2009). Mitochondrial H2O2 emission and celedox state link excess fat intake to

insulin resistance in both rodents and humans. J Clin Invest, 119(3), 573–581.

Ardehali, H., Sabbah, H. N., Burke, M. A., Sarma, S., Liu, P. P., Cleland, J. G., … Gheorghiade,

138

M. (2012). Targeting myocardial substrate metabolism in heart failure: potential for new

therapies. Eur J Heart Fail, 14(2), 120–129.

Batran, R. Al, Gopal, K., Aburasayn, H., Eshreif, A., Almutairi, M., Greenwell, A. A., …

Ussher, J. R. (2019). The antianginal ranolazine mitigates obesity-induced nonalcoholic

fatty liver disease and increases hepatic pyruvate dehydrogenase activity. JCI Insight, 4(1).

Becker, D., Sahali, Y., & Raviv, M. (1990). The absolute configuration effect on the activity of

the avocado rooting promoter. Phytochemistry, 29(7), 2065–2067.

Bene, J., Hadzsiev, K., & Melegh, B. (2018). Role of carnitine and its derivatives in the

development and management of type 2 diabetes. Nutrition & Diabetes, 8(1), 8.

Bessesen, D. H., & Van Gaal, L. F. (2018). Progress and challenges in anti-obesity

pharmacotherapy. The Lancet Diabetes & Endocrinology, 6(3), 237–248.

Beyazova, M. L., Dewis, M. L., Gravina, S. A., Kraut, K., Trinnaman, L., Yang, Y., & Bardsley,

K. (2010). Use of Avocado Derivatives in Flavor Applications. Retrieved from

http://www.freepatentsonline.com/y2010/0034944.html

Biden, T. J., Robinson, D., Cordery, D., Hughes, W. E., & Busch, A. K. (2002). Chronic Effects

of Fatty Acids on Pancreatic beta-Cell Function: New Insights From Functional Genomics.

Diabetes, 53(suppl 1), 159–165.

Blaak, E., Hul, G., Verdich, C., Stich, V., Martinez, A., Petersen, M., … Saris, W. (2006). Fat

oxidation before and after a high fat load in the obese insulin resistant state. Journal of

Clinical Endocrinology and Metabolism, 91(4), 1462–1469.

Blake, A. I., Co, E. D., & Marangoni, A. G. (2014). Structure and physical properties of plant

wax crystal networks and their relationship to oil binding capacity. Journal of the American

Oil Chemists’ Society, 91(6), 885–903.

139

Bonnard, C., Durand, D., Peyrol, S., Chanseaume, E., Chauvin, M. a, Morio, B., … Rieusset, J.

(2008). Mitochondrial dysfunction results from oxidative stress in the skeletal muscle of

diet-induced insulin resistant mice. J Clin Invest, 118(2), 789–800.

Boucher, J., Kleinridders, A., & Kahn, C. R. (2014). Insulin receptor signaling in normal and

insulin-resistant states. Cold Spring Harbor Perspectives in Biology, 6(1), a009191.

Boushel, R., Gnaiger, E., Schjerling, P., Skovbro, M., Kraunsøe, R., & Dela, F. (2007). Patients

with type 2 diabetes have normal mitochondrial function in skeletal muscle. Diabetologia,

50(4), 790–796.

Broutin, N., Legrnd, J., & Pieeirilli, A. (2003). Patent No. 6,582,688 B1. U.S.: U.S. Patent and

Trademark Office.

Brown, B. (1973). Qualitative and quantitative estimation by thin-layer and gas chromatography

of a series of C17 oxygenated aliphatic compounds in the avocado (Persea americana).

Journal of Chromatography, 86, 239–245.

Bruce, C. R., Hoy, A. J., Turner, N., Watt, M. J., Allen, T. L., Carpenter, K., … Kraegen, E. W.

(2008). Overexpression of carnitine palmitoyltransferase-1 in skeletal muscle is sufficient to

enhance fatty acid oxidation and improve high fat diet-induced insulin resistance. Diabetes,

58(3), 550–558.

Buyukozturk, F., Benneyan, J. C., & Carrier, R. L. (2010). Impact of emulsion-based drug

delivery systems on intestinal permeability and drug release kinetics. Journal of Controlled

Release, 142(1), 22–30.

Callender, S. P., Mathews, J. A., Kobernyk, K., & Wettig, S. D. (2017). Microemulsion utility in

pharmaceuticals: Implications for multi-drug delivery. International Journal of

Pharmaceutics, 526(1–2), 425–442.

140

Capozzoli, C., Galliani, G., Ricot, F., & Terraneo, A. (2018). Competitive Kinetics of Lipase

Catalyzed Hydrolysis explain Differences in Digestibility of Wax Esters. Journal of

Biochemistry and Analytical Studies, 3(1).

Chambers, E., Wagrowski-Diehl, D. M., Lu, Z., & Mazzeo, J. R. (2007). Systematic and

comprehensive strategy for reducing matrix effects in LC/MS/MS analyses. J Chromatogr

B Analyt Technol Biomed Life Sci, 852(1–2), 22–34.

Chen, J., Green, K. B., & Nichols, K. K. (2015). Characterization of wax esters by electrospray

ionization tandem mass spectrometry: double bond effect and unusual product ions. Lipids,

50(8), 821–836.

Chiodi, P. (2017). Diabetes Mellitus: could the Inhibition of a Single Enzyme (CPT-1) Involved

in the Beta-Oxidation Process Improve this Complex Disease? Current Research in

Diabestes & Obesity Journal, 1(3).

Chrusciel, P., Rysz, J., & Banach, M. (2014). Defining the role of trimetazidine in the treatment

of cardiovascular disorders: some insights on its role in heart failure and peripheral artery

disease. Drugs, 74(9), 971–980.

Cline, G. W., Magnusson, I., Rothman, D. L., Petersen, K. F., Laurent, D., & Shulman, G. I.

(1997). Muscle Glycogen Synthesis in Insulin-dependent Diabetes Mellitus Mechanism of

Impaired Insulin-stimulated Muscle Glucose Metabolism in Subjects with Insulin-

dependent Diabetes Mellitus. J. Clin. Invest, 99(9), 2219–2224.

Cline, G. W., Rothman, D. L., Magnusson, I., Katz, L. D., & Shulman, G. I. (1994). 13C-nuclear

magnetic resonance spectroscopy studies of hepatic glucose metabolism in normal subjects

and subjects with insulin-dependent diabetes mellitus. The Journal of Clinical Investigation,

94(6), 2369–2376.

141

Collier, G. R., Traianedes, K., Macaulay, S. L., & O’Dea, K. (1993). Effect of fatty acid

oxidation inhibition on glucose metabolism in diabetic rats. Hormone and Metabolic

Research, 25(01), 9–12.

Conti, R., Mannucci, E., Pessotto, P., Tassoni, E., Carminati, P., Giannessi, F., & Arduini, A.

(2011). Selective reversible inhibition of liver carnitine palmitoyl-transferase 1 by teglicar

reduces gluconeogenesis and improves glucose homeostasis. Diabetes, 60(2), 644–651.

Cook, C. M., Larsen, T. S., Derrig, L. D., Kelly, K. M., & Tande, K. S. (2016). Wax Ester Rich

Oil From The Marine Crustacean, Calanus finmarchicus, is a Bioavailable Source of EPA

and DHA for Human Consumption. Lipids, 51(10), 1137–1144.

Cusi, K. (2010). The role of adipose tissue and lipotoxicity in the pathogenesis of type 2

diabetes. Current Diabetes Reports, 10(4), 306–315.

D’Ambrosio, S. M., Han, C., Pan, L., Kinghorn, A. D., & Ding, H. (2011). Aliphatic acetogenin

constituents of avocado fruits inhibit human oral cancer cell proliferation by targeting the

EGFR/RAS/RAF/MEK/ERK1/2 pathway. Biochem Biophys Res Commun, 409(3), 465–

469.

Dai Ly, L., Xu, S., Choi, S.-K., Ha, C.-M., Thoudam, T., Cha, S.-K., … Park, K.-S. (2018).

Oxidative stress and calcium dysregulation by palmitate in type 2 diabetes. Experimental &

Molecular Medicine, 49(2), 291.

DeFronzo, R. A., Jacot, E., Jequier, E., Maeder, E., Wahren, J., & Felber, J. P. (1981). The effect

of insulin on the disposal of intravenous glucose. Results from indirect calorimetry and

hepatic and femoral venous catheterization. Diabetes, 30(12), 1000–1007.

Després, J. P., & Lemieux, I. (2006). Abdominal obesity and metabolic syndrome. Nature,

444(7121), 881–887.

142

Ding, H., Chin, Y. W., Kinghorn, A. D., & D’Ambrosio, S. M. (2007). Chemopreventive

characteristics of avocado fruit. Seminars in Cancer Biology, Vol. 17, pp. 386–394.

Dobbins, R. L., Szczepaniak, L. S., Bentley, B., Esser, V., Myhill, J., & McGarry, J. D. (2001).

Prolonged inhibition of muscle carnitine palmitoyltransferase-1 promotes intramyocellular

lipid accumulation and insulin resistance in rats. Diabetes, 50(1), 123–130.

Dollerup, O. L., Christensen, B., Svart, M., Schmidt, M. S., Sulek, K., Ringgaard, S., …

Treebak, J. T. (2018). A randomized placebo-controlled clinical trial of nicotinamide

riboside in obese men: safety, insulin-sensitivity, and lipid-mobilizing effects. The

American Journal of Clinical Nutrition, 108(2), 343–353.

Dreher, M. L., & Davenport, A. J. (2013). Hass Avocado Composition and Potential Health

Effects. Critical Reviews in Food Science and Nutrition, 53(7), 738–750.

Eilertsen, K. E., Maehre, H. K., Jensen, I. J., Devold, H., Olsen, J. O., Lie, R. K., … Osterud, B.

(2012). A wax ester and astaxanthin-rich extract from the marine copepod Calanus

finmarchicus attenuates atherogenesis in female apolipoprotein E-deficient mice. J Nutr,

142(3), 508–512.

Elsner, M., Gehrmann, W., & Lenzen, S. (2011). Peroxisome-generated hydrogen peroxide as

important mediator of lipotoxicity in insulin-producing cells. Diabetes, 60(1), 200–208.

Erion, K. A., Berdan, C. A., Burritt, N. E., Corkey, B. E., & Deeney, J. T. (2015). Chronic

exposure to excess nutrients left-shifts the concentration dependence of glucose-stimulated

insulin secretion in pancreatic β-cells. Journal of Biological Chemistry, 290(26), 16191–

16201.

Even, P. C., & Nadkarni, N. A. (2012). Indirect calorimetry in laboratory mice and rats:

principles, practical considerations, interpretation and perspectives. American Journal of

143

Physiology-Regulatory, Integrative and Comparative Physiology, 303(5), R459–R476.

Farines, M., Soulier, J., Rancurel, A., Montaudoin, M. G., & Leborgne, L. (1995). Influence of

Avocado Oil Processing on the Nature of Some Unsaponifiable Constituents. Journal of the

American Oil Chemists Society, 72(4), 473–476.

Finck, B. N., Bernal-Mizrachi, C., Han, D. H., Coleman, T., Sambandam, N., LaRiviere, L. L.,

… Kelly, D. P. (2005). A potential link between muscle peroxisome proliferator-activated

receptor-α signaling and obesity-related diabetes. Cell Metabolism, 1(2), 133–144.

Flores, M., Saravia, C., Vergara, C. E., Avila, F., Valdés, H., & Ortiz-Viedma, J. (2019).

Avocado Oil: Characteristics, Properties, and Applications. Molecules, 24(11), 2172.

Folch, J., Lees, M., & Sloane Stanley, G. H. (1957). A simple method for the isolation and

purification of total lipides from animal tissues. J Biol Chem, 226(1), 497–509.

Foley J. E. (1992). Rationale and Application Of Fatty Acid Oxidation Inhibitors in Treatment of

Diabetes Mellitus. Diabetes Care, 15(6), 773–784.

Fondaco, D., AlHasawi, F., Lan, Y., Ben-Elazar, S., Connolly, K., & Rogers, M. A. (2015).

Biophysical Aspects of Lipid Digestion in Human Breast Milk and SimilacTM Infant

Formulas. Food Biophysics, 10(3), 282–291.

Fransen, M., Lismont, C., & Walton, P. (2017). The peroxisome-mitochondria connection: how

and why? International Journal of Molecular Sciences, 18(6), 1126.

Galgani, J. E., Moro, C., & Ravussin, E. (2008). Metabolic flexibility and insulin resistance.

American Journal of Physiology-Endocrinology and Metabolism, 295(5), E1009–E1017.

Gao, S., McMillan, R. P., Zhu, Q., Lopaschuk, G. D., Hulver, M. W., & Butler, A. A. (2015).

Therapeutic effects of adropin on glucose tolerance and substrate utilization in diet-induced

obese mice with insulin resistance. Mol Metab, 4(4), 310–324.

144

Garcia, E. D., Antonio, J., Hernandez, Z., Vargas, R. D., & Cervantes, S. C. (2010). Patent No.

7,678,396 B2. U.S.: U.S. Patent and Trademark Office.

Garti, N., Yaghmur, A., Leser, M. E., Clement, V., & Watzke, H. J. (2001). Improved oil

solubilization in oil/water food grade microemulsions in the presence of polyols and

ethanol. Journal of Agricultural and Food Chemistry, 49(5), 2552–2562.

Gastaldelli, A. (2017). Insulin resistance and reduced metabolic flexibility: cause or

consequence of NAFLD? Portland Press Limited.

Gavin, T. P., Ernst, J. M., Kwak, H. B., Caudill, S. E., Reed, M. A., Garner, R. T., … Dohm, G.

L. (2018). High Incomplete Skeletal Muscle Fatty Acid Oxidation Explains Low Muscle

Insulin Sensitivity in Poorly Controlled T2D. The Journal of Clinical Endocrinology and

Metabolism, 103(3), 882–889.

Gomes, M. B., Negrato, C. A., Haritoglou, C., Gerss, J., Hammes, H. P., Kampik, A., …

Muderrisoglu, H. (2012). Glycemic and oxidative status of patients with type 2 diabetes

mellitus following oral administration of alphalipoic acid: A randomized double-blinded

placebocontrolled study. Asia Pacific Journal of Clinical Nutrition, 21(1), 12–21.

González-Ortiz, M., Martínez-Abundis, E., Robles-Cervantes, J. A., Ramírez-Ramírez, V., &

Ramos-Zavala, M. G. (2011). Effect of thiamine administration on metabolic profile,

cytokines and inflammatory markers in drug-naïve patients with type 2 diabetes. European

Journal of Nutrition, 50(2), 145–149.

Goodpaster, B. H., & Sparks, L. M. (2018). Metabolic Flexibility in Health and Disease. Cell

Metabolism, 25(5), 1027–1036.

Gouni-Berthold, I., & Berthold, H. K. (2002). Policosanol: clinical pharmacology and

therapeutic significance of a new lipid-lowering agent. American Heart Journal, 143(2),

145

356–365.

Gradzielski, M. (2002). Effect of the Cosurfactant Structure on the Bending Elasticity in

Nonionic Oil-in-Water Microemulsions. Langmuir, 14(21), 6037–6044.

Guerre-Millo, M., Rouault, C., Poulain, P., André, J., Poitout, V., Peters, J. M., … Staels, B.

(2001). PPAR-α–Null Mice Are Protected From High-Fat Diet–Induced Insulin Resistance

— Diabetes. Diabetes, 50(12), 2809–2814.

Guo, Z. (2015). Pyruvate dehydrogenase, Randle cycle, and skeletal muscle insulin resistance.

Proceedings of the National Academy of Sciences, 112(22), 2854.

Haas, R. C. (1951). Variation in the composition of avocado seed. California Avocado Society

1951 Yearbook, (35), 139–152.

Hancock, C. R., Han, D.-H., Chen, M., Terada, S., Yasuda, T., Wright, D. C., & Holloszy, J. O.

(2008). High-fat diets cause insulin resistance despite an increase in muscle mitochondria.

Proceedings of the National Academy of Sciences, 105(22), 7815–7820.

Hargrove, J. L., Greenspan, P., & Hartle, D. K. (2004). Nutritional significance and metabolism

of very long chain fatty alcohols and acids from dietary waxes. Exp Biol Med (Maywood),

229(3), 215–226.

Hashimura, H., Ueda, C., Kawabata, J., & Kasai, T. (2001). Acetyl-CoA carboxylase inhibitors

from avocado (Persea americana Mill) fruits. Biosci Biotechnol Biochem, 65(7), 1656–1658.

Henique, C., Mansouri, A., Fumey, G., Lenoir, V., Girard, J., Bouillaud, F., … Cohen, I. (2010).

Increased mitochondrial fatty acid oxidation is sufficient to protect skeletal muscle cells

from palmitate-induced apoptosis. Journal of Biological Chemistry, 285(47), 36818–36827.

Hernandez-brenes, C., Garcia-cruz, M. I., Gutierrez-uribe, J. A., Benavides-lozano, J. A., &

Rodriguez-sanchez, D. G. (2013). Patent No. 20160249613A1. U.S.: U.S. Patent and

146

Trademark Office.

Holloway, G. P., Thrush, A. B., Heigenhauser, G. J. F., Tandon, N. N., Dyck, D. J., Bonen, A., &

Spriet, L. L. (2007). Skeletal muscle mitochondrial FAT/CD36 content and palmitate

oxidation are not decreased in obese women. AJP: Endocrinology and Metabolism, 292(6),

E1782–E1789.

Holubarsch, C. J. F., Rohrbach, M., Karrasch, M., Boehm, E., Polonski, L., Ponikowski, P., &

Rhein, S. (2007). A double-blind randomized multicentre clinical trial to evaluate the

efficacy and safety of two doses of etomoxir in comparison with placebo in patients with

moderate congestive heart failure: the ERGO (etomoxir for the recovery of glucose

oxidation) stud. Clinical Science, 113(4), 205–212.

Houmard, J. A., Tanner, C. J., Yu, C., Cunningham, P. G., Pories, W. J., MacDonald, K. G., &

Shulman, G. I. (2002). Effect of weight loss on insulin sensitivity and intramuscular long-

chain fatty acyl-CoAs in morbidly obese subjects. Diabetes, 51(10), 2959–2963.

Hue, L., & Taegtmeyer, H. (2009). The Randle cycle revisited: a new head for an old hat.

American Journal of Physiology-Endocrinology and Metabolism, 297(3), E578–E591.

Iglesias, M. A., Ye, J., Frangioudakis, G., Saha, A. K., Tomas, E., Ruderman, N. B., … Kraegen,

E. W. (2002). AICAR Administration Causes an Apparent Insulin-Resistant High-Fat – Fed

Rats. Diabetes, 51, 2886–2894.

Iozzo, P. (2009). Viewpoints on the way to the consensus session: Where does insulin resistance

start? The adiposse tissue. Diabetes Care, 32(suppl 2), 168–173.

Ipsen, D. H., Lykkesfeldt, J., & Tveden-Nyborg, P. (2018). Molecular mechanisms of hepatic

lipid accumulation in non-alcoholic fatty liver disease. Cellular and Molecular Life

Sciences, 75(18), 3313–3327.

147

Jaswal, J. S., Keung, W., Wang, W., Ussher, J. R., & Lopaschuk, G. D. (2011). Targeting fatty

acid and carbohydrate oxidation—a novel therapeutic intervention in the ischemic and

failing heart. Biochimica et Biophysica Acta (BBA)-Molecular Cell Research, 1813(7),

1333–1350.

Ježek, P., Jabůrek, M., Holendová, B., & Plecitá-Hlavatá, L. (2018). Fatty Acid-Stimulated

Insulin Secretion vs. Lipotoxicity. Molecules, 23(6), 1483.

Jing, E., O’Neill, B. T., Rardin, M. J., Kleinridders, A., Ilkeyeva, O. R., Ussar, S., … Kahn, C. R.

(2013). Sirt3 regulates metabolic flexibility of skeletal muscle through reversible enzymatic

deacetylation. Diabetes, 62(10), 3404–3417.

Kashman, Y., Neeman, I., & Lifshitz, A. (1969a). 6 New C17-Olefinic and Acetylenic

Oxygenated Compounds from Avocado Pear. Israel Journal of Chemistry, 7(1), 173-.

Kashman, Y., Neeman, I., & Lifshitz, A. (1969b). New Compounds from Avocado Pear.

Tetrahedron, 25(18), 4617-.

Kelley, D. (2005). Skeletal muscle fat oxidation: timing and flexibility are everything. Journal of

Clinical Investigation, 115(7), 1699–1702.

Kelley, D. E., Goodpaster, B., Wing, R. R., & Simoneau, J.-A. (1999). Skeletal muscle fatty acid

metabolism in association with insulin resistance, obesity, and weight loss. American

Journal of Physiology-Endocrinology And Metabolism, 277(6), E1130–E1141.

Keung, W., Ussher, J. R., Jaswal, J. S., Raubenheimer, M., Lam, V. H., Wagg, C. S., &

Lopaschuk, G. D. (2013). Inhibition of carnitine palmitoyltransferase-1 activity alleviates

insulin resistance in diet-induced obese mice. Diabetes, 62(3), 711–720.

Khan, N., Edwards, C., Thompson, S., Burke, S., Walk, A., Reeser, G., … Holscher, H. (2019).

Effects of Avocado Consumption on Abdominal Adiposity and Glucose Tolerance: Findings

148

from the Persea Americana for Total Health (PATH) Randomized Controlled Trial (P21-

005-19). Oxford University Press.

Killalea, S. M., & Krum, H. (2001). Systematic review of the efficacy and safety of perhexiline

in the treatment of ischemic heart disease. American Journal of Cardiovascular Drugs,

1(3), 193–204.

Korotchkina, L. G., Sidhu, S., & Patel, M. S. (2004). R-lipoic acid inhibits mammalian pyruvate

dehydrogenase kinase. Free Radical Research, 38(10), 1083–1092.

Koves, T. R., Kheterpal, I., Davies, M. N., DeBalsi, K. L., Stevens, R. D., Ilkayeva, O. R., …

Mynatt, R. L. (2012). Muscle-Specific Deletion of Carnitine Acetyltransferase

Compromises Glucose Tolerance and Metabolic Flexibility. Cell Metabolism, 15(5), 764–

777.

Koves, T. R., Ussher, J. R., Noland, R. C., Slentz, D., Mosedale, M., Ilkayeva, O., … Muoio, D.

M. (2008). Mitochondrial Overload and Incomplete Fatty Acid Oxidation Contribute to

Skeletal Muscle Insulin Resistance. Cell Metabolism, 7(1), 45–56.

LeBlanc, P. J., Mulligan, M., Antolic, A., MacPherson, L., Inglis, J. G., Martin, D., … Peters, S.

J. (2008). Skeletal muscle type comparison of pyruvate dehydrogenase phosphatase activity

and isoform expression: effects of obesity and endurance training. American Journal of

Physiology-Regulatory, Integrative and Comparative Physiology, 295(4), R1224–R1230.

Lee, E. A., Angka, L., Rota, S. G., Hanlon, T., Mitchell, A., Hurren, R., … Spagnuolo, P. A.

(2015). Targeting mitochondria with avocatin B induces selective leukemia cell death.

Cancer Research, 75(12), 2478–2488.

Lee, I. K. (2014). The role of pyruvate dehydrogenase kinase in diabetes and obesity. Diabetes &

Metabolism Journal, 38(3), 181–186.

149

Lee, R. F., Paffenhofer, G. A., Nevenzel, J. C., & Benson, A. A. (1970). The metabolism of wax

ters and other lipids by the marine copepod, Calanus helgolandicus. J Lipid Res, 11(3), 237–

240.

Lee, S. K., Young, R. E., Schiffman, P. M., & Coggins, C. W. (1983). Maturity Studies of

Avocado Fruit Based on Picking Dates and Dry-Weight. Journal of the American Society

for Horticultural Science, 108(3), 390–394.

Lin, Y., Rudrum, M., van der Wielen, R. P. J., Trautwein, E. A., McNeill, G., Sierksma, A., &

Meijer, G. W. (2004). Wheat germ policosanol failed to lower plasma cholesterol in

subjects with normal to mildly elevated cholesterol concentrations. Metabolism, 53(10),

1309–1314.

Lopaschuk, G. D. (2016). Fatty Acid Oxidation and Its Relation with Insulin Resistance and

Associated Disorders. Annals of Nutrition and Metabolism, 68(3), 15–20.

Lopaschuk, G. D. (2017). Metabolic modulators in heart disease: past, present, and future.

Canadian Journal of Cardiology, 33(7), 838–849.

Lopaschuk, G. D., Collins-Nakai, R., Olley, P. M., Montague, T. J., McNeil, G., Gayle, M., …

Finegan, B. A. (1994). Plasma fatty acid levels in infants and adults after myocardial

ischemia. Am Heart J, 128(1), 61–67.

Lopaschuk, G. D., Ussher, J. R., Folmes, C. D., Jaswal, J. S., & Stanley, W. C. (2010).

Myocardial fatty acid metabolism in health and disease. Physiol Rev, 90(1), 207–258.

Luiken, J. J. F. P., Niessen, H. E. C., Coort, S. L. M., Hoebers, N., Coumans, W. A., Schwenk,

R. W., … Glatz, J. F. C. (2009). Etomoxir-induced partial carnitine palmitoyltransferase-I

(CPT-I) inhibition in vivo does not alter cardiac long-chain fatty acid uptake and oxidation

rates . Biochemical Journal, 419(2), 447–455.

150

Lupi, R., Dotta, F., Marselli, L., Del Guerra, S., Masini, M., Santangelo, C., … Marchetti, P.

(2002). Prolonged Exposure to Free Fatty Acids Has Cytostatic and Pro-Apoptotic Effects

on Human Pancreatic Islets. Diabetes, 51(5), 1437–1442.

Maier, L. S., Layug, B., Karwatowska-Prokopczuk, E., Belardinelli, L., Lee, S., Sander, J., …

Jacobshagen, C. (2013). RAnoLazIne for the treatment of diastolic heart failure in patients

with preserved ejection fraction. The RALI-DHF proof-of-concept study. JACC: Heart

Failure, 1(2), 115–122.

Matthews, D. R., Hosker, J. P., Rudenski, A. S., Naylor, B. A., Treacher, D. F., & Turner, R. C.

(1985). Homeostasis model assessment: insulin resistance and β-cell function from fasting

plasma glucose and insulin concentrations in man. Diabetologia, 28(7), 412–419.

Mcclements, D. J. (2015). Nanoemulsions versus microemulsions : Terminology , differences ,

and similarities Nanoemulsions versus microemulsions : terminology , differences , and

similarities. The Royal Society of Chemistry, 8, 1719–1729.

Mendoza, D. J. (2011). Patent No. 2011 009 1616A1. U.S. Patent and Trademark Office.

Meretzki, S., Rosenblat, G., & Segal, J. (2011). Patent No. 2011/0217251. United States, U.S.:

U.S. Patent and Trademark Office.

Mihalik, S. J. (2010). Increased Levels of Plasma Acylcarnitines in Obesity and Type 2 Diabetes

and Identification of a Marker of Glucolipotoxicity. Obesity, 18(9), 1695–1700.

Minekus, M., Alminger, M., Alvito, P., Ballance, S., Bohn, T., Bourlieu, C., … Dupont, D.

(2014). A standardised static in vitro digestion method suitable for food–an international

consensus. Food & Function, 5(6), 1113–1124.

Minekus, M., Marteau, P., Havenaar, R., & Huis in ’t Veld, J. H. J. (1995). A

multicompartmental dynamic computer-controlled model simulating the stomach and small

151

intestine. Alternatives to Laboratory Animals, 23(August 2015), 197–209.

Molecular, D., & Carolina, N. (2014). Measurement of Fatty Acid Oxidation Rates in Animal

Tissues and Cell Lines. In Conceptual background and bioenergetic/mitochondrial aspects

of oncometabolism (Vol. 542, pp. 391–405).

Morales, D., Gutiérrez, J. M., García-Celma, M. J., & Solans, Y. C. (2003). A study of the

relation between bicontinuous microemulsions and oil/water nano-emulsion formation.

Langmuir, 19(18), 7196–7200.

Msika, P, Legrand, J., & Garnier, S. (2014). Patent No. 9,416,333. United States, U.S.: U.S.

Patent and Trademark Office.

Msika, Philippe, & Legrand, J. (2010). Patent No. 7,81,6547 B2. U.S.: U.S. Patent and

Trademark Office.

Mudge, S. M. (2005). Fatty alcohols—a review of their natural synthesis and environmental

distribution. Exec Summ Soap Deterg Assoc, (132), 1–141.

Mudge, S. M., DeLeo, P. C., & Dyer, S. D. (2012). Quantifying the anthropogenic fraction of

fatty alcohols in a terrestrial environment. Environmental Toxicology and Chemistry, 31(6),

1209–1222.

Muoio, D. M. (2014). Metabolic inflexibility: When mitochondrial indecision leads to metabolic

gridlock. Cell, 159(6), 1253–1262.

Muoio, D. M., & Neufer, D. P. (2012). Lipid-induced mitochondrial stress and insulin action in

muscle. Cell Metabolism, 15(5), 595–605.

Muoio, D. M., & Newgard, C. B. (2006). Obesity-related derangements in metabolic regulation.

Annu. Rev. Biochem., 75, 367–401.

Muoio, D. M., & Newgard, C. B. (2008a). Fatty acid oxidation and insulin action: when less is

152

more. Diabetes, 57(6), 1455–1456.

Muoio, D. M., & Newgard, C. B. (2008b). Molecular and metabolic mechanisms of insulin

resistance and β-cell failure in type 2 diabetes. Nature Reviews Molecular Cell Biology,

9(3), 193.

Nardin, I., & Köllner, S. (2018). Successful development of oral SEDDS: Screening of

excipients from the industrial point of view. Advanced Drug Delivery Reviews.

Oberlies, N. H., Rogers, L. L., Martin, J. M., & McLaughlin, J. L. (1998). Cytotoxic and

insecticidal constituents of the unripe fruit of Persea americana. J Nat Prod, 61(6), 781–

785.

Park, E., Edirisinghe, I., & Burton-Freeman, B. (2018). Avocado Fruit on Postprandial Markers

of Cardio-Metabolic Risk: A Randomized Controlled Dose Response Trial in Overweight

and Obese Men and Women. Nutrients, 10(9), 1287.

Patel, M. S., Nemeria, N. S., Furey, W., & Jordan, F. (2014). The pyruvate dehydrogenase

complexes: Structure-based function and regulation. Journal of Biological Chemistry, Vol.

289, pp. 16615–16623.

Patterson, J. N., Cousteils, K., Lou, J. W., Fox, J. E. M., MacDonald, P. E., & Joseph, J. W.

(2014). Mitochondrial metabolism of pyruvate is essential for regulating glucose-stimulated

insulin secretion. Journal of Biological Chemistry, 289(19), 13335–13346.

Patton, J. S., Vetter, R. D., Hamosh, M., Borgstrom, B., Lindstrom, M., & Carey, M. C. (1985).

The light microscopy of triglyceride digestion. Food Structure, 4(1), 5.

Pettus, J., McNabb, B., Eckel, R. H., Skyler, J. S., Dhalla, A., Guan, S., … Henry, R. H. (2016).

Effect of ranolazine on glycaemic control in patients with type 2 diabetes treated with either

glimepiride or metformin. Diabetes, Obesity and Metabolism, 18(5), 463–474.

153

Piccirilli, A., & Legrand, J. (2008). Patent No. 7,371,420 B2. U.S.: U.S. Patent and Trademark

Office.

Place, A. R. (1992). Comparative aspects of lipid digestion and absorption: physiological

correlates of wax ester digestion. American Journal of Physiology-Regulatory, Integrative

and Comparative Physiology, 263(3), R464–R471.

Platt-Aloia, K. A., & Thomson, W. W. (1981). Ultrastructure of the mesocarp of mature avocado

fruit and changes associated with ripening. Annals of Botany, 48(4), 451–466.

Rabbani, N., Alam, S. S., Riaz, S., Larkin, J. R., Akhtar, M. W., Shafi, T., & Thornalley, P. J.

(2009). High-dose thiamine therapy for patients with type 2 diabetes and microalbuminuria:

a randomised, double-blind placebo-controlled pilot study. Diabetologia, 52(2), 208–212.

Randle, P. J. (1998). Regulatory interactions between lipids and carbohydrates: The glucose fatty

acid cycle after 35 years. Diabetes/Metabolism Reviews, 14(4), 263–283.

Randle, P. J., Garland, P. B., Hales, C. N., & Newsholme, E. A. (1963). The glucose fatty-acid

cycle. Its role in insulin sensitivity and the metabolic disturbances of diabetes mellitus.

Lancet, 1(7285), 785–789.

Rochlani, Y., Pothineni, N. V., Kovelamudi, S., & Mehta, J. L. (2017). Metabolic syndrome:

pathophysiology, management, and modulation by natural compounds. Therapeutic

Advances in Cardiovascular Disease, 11(8), 215–225.

Roden, M., Price, T. B., Perseghin, G., Petersen, K. F., Rothman, D. L., Cline, G. W., &

Shulman, G. I. (1996). Mechanism of free fatty acid-induced insulin resistance in humans.

Journal of Clinical Investigation, 97(12), 2859–2865.

Rodríguez-López, C., Hernández-Brenes, C., & Díaz de la Garza, R. (2015). A Targeted

Metabolomics Approach To Characterize Acetogenin Profiles In Avocado Fruit (Persea

154

Americana Mill.). RSC Adv., 5, 106019–106029.

Rodriguez-Sanchez, D. G., Pacheco, A., Garcia-Cruz, M. I., Gutierrez-Uribe, J. A., Benavides-

Lozano, J. A., & Hernandez-Brenes, C. (2013). Isolation and Structure Elucidation of

Avocado Seed (Persea americana) Lipid Derivatives That Inhibit Clostridium sporogenes

Endospore Germination. Journal of Agricultural and Food Chemistry, 61(30), 7403–7411.

Rosenblat, G., Meretski, S., Segal, J., Tarshis, M., Schroeder, A., Zanin-Zhorov, A., …

Hochberg, M. (2011). Polyhydroxylated fatty alcohols derived from avocado suppress

inflammatory response and provide non-sunscreen protection against UV-induced damage

in skin cells. Archives of Dermatological Research, 303(4), 239–246.

Ruge, T., Hodson, L., Cheeseman, J., Dennis, A. L., Fielding, B. A., Humphreys, S. M., …

Karpe, F. (2009). Fasted to Fed Trafficking of Fatty Acids in Human Enhanced Fat Storage.

The Journal of Clinical Endocrinology & Metabolism, 94(May), 1781–1788.

Samudio, I., Harmancey, R., Fiegl, M., Kantarjian, H., Konopleva, M., Korchin, B., … Andreeff,

M. (2010). Pharmacologic inhibition of fatty acid oxidation sensitizes human leukemia cells

to apoptosis induction. J Clin Invest, 120(1), 142–156.

San-Millán, I., & Brooks, G. A. (2018). Assessment of metabolic flexibility by means of

measuring blood lactate, fat, and carbohydrate oxidation responses to exercise in

professional endurance athletes and less-fit individuals. Sports Medicine, 48(2), 467–479.

Savary, P. (1971). The action of pure pig pancreatic lipase upon esters of long-chain fatty acids

and short-chain primary alcohols. Biochimica et Biophysica Acta (BBA)-Lipids and Lipid

Metabolism, 248(2), 149–155.

Sebastián, D., Herrero, L., Serra, D., Asins, G., Hegardt, F. G., Sebastian, D., … Hegardt, F. G.

(2007). CPT I overexpression protects L6E9 muscle cells from fatty acid-induced insulin

155

resistance. American Journal of Physiology. Endocrinology and Metabolism, 292(3), E677-

86.

Segal, J., & Rosenblat, G. (2015). Patent No. 20150175933. United States: U.S. Patent and

Trademark Office.

Serrano, J., Cassanye, A., Martín-Gari, M., Granado-Serrano, A., & Portero-Otín, M. (2016).

Effect of Dietary Bioactive Compounds on Mitochondrial and Metabolic Flexibility.

Diseases, 4(1), 14.

Sharma, R., Matsuzaka, T., Kaushik, M. K., Sugasawa, T., Ohno, H., Wang, Y., … Mizunoe, Y.

(2019). Octacosanol and policosanol prevent high-fat diet-induced obesity and metabolic

disorders by activating brown adipose tissue and improving liver metabolism. Scientific

Reports, 9(1), 5169.

Silva-Platas, C., Garcia, N., Fernandez-Sada, E., Davila, D., Hernandez-Brenes, C., Rodriguez,

D., & Garcia-Rivas, G. (2012). Cardiotoxicity of acetogenins from Persea americana occurs

through the mitochondrial permeability transition pore and caspase-dependent apoptosis

pathways. J Bioenerg Biomembr, 44(4), 461–471.

Spagnuolo, P. A., Schimmer, A. D., & Lee, E. A. (2016). Patent No. 20170304251A1. U.S.: U.S.

Patent and Trademark Office.

Speranza, A., Corradini, M. G., Hartman, T. G., Ribnicky, D., Oren, A., & Rogers, M. A. (2013).

Influence of emulsifier structure on lipid bioaccessibility in oil-water nanoemulsions.

Journal of Agricultural and Food Chemistry, 61(26), 6505–6515.

Steggall, A., Mordi, I., & Lang, C. (2017). Targeting Metabolic Modulation and Mitochondrial

Dysfunction in the Treatment of Heart Failure. Diseases, 5(2), 14.

Sugiyama, T., Sato, A., & Yamashita, K. (1982). Synthesis of all four stereoisomers of

156

antibacterial component of avocado. Agricultural and Biological Chemistry, 46(2), 481–

485.

Tabeshpour, J., Razavi, B. M., & Hosseinzadeh, H. (2017). Effects of Avocado ( Persea

americana ) on Metabolic Syndrome : A Comprehensive Systematic Review. Phytotherapy

Research, 837(April), 819–837.

Takenaga, F., Matsuyama, K., Abe, S., Torii, Y., & Itoh, S. (2008). Lipid and fatty acid

composition of mesocarp and seed of avocado fruits harvested at northern range in Japan. J

Oleo Sci, 57(11), 591–597.

Tande, K. S., Vo, T. D., & Lynch, B. S. (2016). Clinical safety evaluation of marine oil derived

from Calanus finmarchicus. Regul Toxicol Pharmacol, 80, 25–31.

Tcheng, M., Samudio, I., Lee, E. A., Minden, M. D., & Spagnuolo, P. A. (2017). The

mitochondria target drug avocatin B synergizes with induction chemotherapeutics to induce

leukemia cell death. Leuk Lymphoma, 58(4), 986–988.

Timmers, S., Nabben, M., Bosma, M., van Bree, B., Lenaers, E., van Beurden, D., …

Schrauwen, P. (2012). Augmenting muscle diacylglycerol and triacylglycerol content by

blocking fatty acid oxidation does not impede insulin sensitivity. Proceedings of the

National Academy of Sciences, 109(29), 11711–11716.

Troncoso, E., Aguilera, J. M., & McClements, D. J. (2012). Fabrication, characterization and

lipase digestibility of food-grade nanoemulsions. Food Hydrocolloids, 27(2), 355–363.

Turner, N., Bruce, C. R., Beale, S. M., Hoehn, K. L., So, T., Rolph, M. S., & Cooney, G. J.

(2007). Excess Lipid Availability Increases Mitochondrial Fatty Acid Oxidative Capacity in

Muscle: Evidence Against a Role for Reduced Fatty Acid Axidation in Lipid-induced

Insulin Resistance in Rodents. Diabetes, 56(8), 2085–2092.

157

Tyagi, S., Gupta, P., Saini, A. S., Kaushal, C., & Sharma, S. (2011). The peroxisome

proliferator-activated receptor: a family of nuclear receptors role in various diseases.

Journal of Advanced Pharmaceutical Technology & Research, 2(4), 236.

Unger, R. H., Clark, G. O., Scherer, P. E., & Orci, L. (2010). Lipid homeostasis, lipotoxicity and

the metabolic syndrome. Biochimica et Biophysica Acta (BBA)-Molecular and Cell Biology

of Lipids, 1801(3), 209–214.

Ussher, J. R., Keung, W., Fillmore, N., Koves, T. R., Mori, J., Zhang, L., … Lopaschuk, G. D.

(2014). Treatment with the 3-Ketoacyl-CoA Thiolase Inhibitor Trimetazidine Does Not

Exacerbate Whole-Body Insulin Resistance in Obese Mice. Journal of Pharmacology and

Experimental Therapeutics, 349(3), 487–496.

Ussher, J R., Fillmore, N., Keung, W., Zhang, L., Mori, J., Sidhu, V. K., … Lopaschuk, G. D.

(2016). Genetic and pharmacological inhibition of malonyl CoA decarboxylase does not

exacerbate age-related insulin resistance in mice. Diabetes, 65(7), 1883–1891.

Ussher, John R., Cadete, V. J. J., Zhang, L., Jaswal, J. S., Lopaschuk, D. G., Keung, W., …

Swyrd, S. J. (2010). Inhibition of de novo ceramide synthesis reverses diet-induced insulin

resistance and enhances whole-body oxygen consumption. Diabetes, 59(10), 2453–2464.

Vickers, A. E. M. (2009). Characterization of hepatic mitochondrial injury induced by fatty acid

oxidation inhibitors. Toxicologic Pathology, 37(1), 78–88.

Wang, Y., Palmfeldt, J., Gregersen, N., Makhov, A. M., Conway, J. F., Wang, M., … Vockley, J.

(2019). Mitochondrial fatty acid oxidation and the electron transport chain comprise a

multifunctional mitochondrial protein complex. Journal of Biological Chemistry,

jbc.RA119.008680.

WHO. (2017). WHO Fact Sheet: Obesity and overweight. Retrieved July 9, 2018, from

158

http://www.who.int/en/news-room/fact-sheets/detail/obesity-and-overweight

Wicks, S. E., Vandanmagsar, B., Haynie, K. R., Fuller, S. E., Warfel, J. D., Stephens, J. M., …

Mynatt, R. L. (2015). Impaired mitochondrial fat oxidation induces adaptive remodeling of

muscle metabolism. Proceedings of the National Academy of Sciences of the United States

of America, 112(25), E3300–E3309.

Wolfe, R. R. (1998). Metabolic interactions between glucose and fatty acids in humans.

American Journal of Clinical Nutrition, 67(3), 519S-526S.

Zhang, L., Keung, W., Samokhvalov, V., Wang, W., & Lopaschuk, G. D. (2010). Role of fatty

acid uptake and fatty acid β-oxidation in mediating insulin resistance in heart and skeletal

muscle. Biochimica et Biophysica Acta - Molecular and Cell Biology of Lipids, 1801(1), 1–

22.

Zhou, Y. P., & Grill, V. E. (1994). Long-term exposure of rat pancreatic islets to fatty acids

inhibits glucose-induced insulin secretion and biosynthesis through a glucose fatty acid

cycle. Journal of Clinical Investigation, 93(2), 870–876.

Zhou, Y. P., & Grill, V. E. (1995). Palmitate-induced beta-cell insensitivity to glucose is coupled

to decreased pyruvate dehydrogenase activity and enhanced kinase activity in rat pancreatic

islets. Diabetes, 44(4), 394–399.

Zhu, L., Huang, Y., Edirisinghe, I., Park, E., & Burton-Freeman, B. (2019). Using the Avocado

to Test the Satiety Effects of a Fat-Fiber Combination in Place of Carbohydrate Energy in a

Breakfast Meal in Overweight and Obese Men and Women: A Randomized Clinical Trial.

Nutrients, 11(5), 952.

Zimmer, D. (2014). New US FDA draft guidance on bioanalytical method validation versus

current FDA and EMA guidelines: chromatographic methods and ISR. Bioanalysis, 6(1),

159

13–19.

Zou, C., Wang, Y., & Shen, Z. (2005). 2-NBDG as a fluorescent indicator for direct glucose

uptake measurement. Journal of Biochemical and Biophysical Methods, 64(3), 207–215.

160

APPENDICES Appendix A. University of Guelph research ethics board (REB) certificate of approval for AVO clinical study

161

Appendix B. Health Canada NHPD notice of authorization for AVO clinical study

162

Appendix C. AVO clinical study recruitment poster

163

Appendix D. AVO clinical study consent form

ONTARIO AGRICULTURAL COLLEGE Department of Food Science CONSENT TO PARTICIPATE IN RESEARCH

RESEARCH PROJECT TITLE Clinical investigation on the safety of avocado pulp lipids You are asked to participate in a research study conducted by Dr. Paul Spagnuolo and his research team from the Department of Food Science at the University of Guelph. This research is sponsored by Dr. Spagnuolo's laboratory and Advanced Orthomolecular Research. The results of this research will contribute to graduate research.

If you have any questions or concerns about the research, please feel free to contact any of the three contacts below: Dr. Paul Spagnuolo Dr. Kim Bretz Nawaz Ahmed, PhD [email protected] [email protected] Candidate 519-824-4120 ext. 53732 519-585-0010 [email protected] 226-978-3164

We request that you first contact the graduate student (Mr. Nawaz Ahmed) associated with this trial for all emergencies and trial related questions.

PURPOSE OF THE STUDY Obesity and diabetes are a significant global burden and there is an immediate need for novel treatments and management strategies. Our lab has shown that an avocado based supplement inhibits the process responsible for the cellular breakdown of fat for energy. Preliminary in vivo experiments have shown that this avocado based supplement is an ideal compound for the management of metabolic diseases, as it imparts physiological activity without toxicity. The purpose of this study is to assess the safety of an avocado based supplement in healthy human subjects after oral consumption. This is a single center (one trial clinical site), double-blind (you and the research team will not know which treatment group you are part of), placebo-controlled (one group of participants will not be consuming product containing the medicinal test ingredient), randomized (you will be randomly allocated to one of the study treatment groups) clinical trial.

164

PROCEDURES If you volunteer to participate in this study, we would ask you to do the following things:

1. Recruitment Interview (first clinic visit) At the start of the study, you will visit the trial clinic site (Fundamental of Health Naturopathic Medicine Clinic, 420 Erb St., West, Waterloo, ON) for a recruitment interview with a trained naturopath (Dr. Kim Bretz ND). During this recruitment interview your weight and height measurements will be taken to obtain your BMI value. The Naturopath will also need to acquire your detailed medical history to ensure you are healthy and do not have any medical conditions that may prevent you from being enrolled in the trial. The inclusion criteria for this trial is as follows: ▪ Adults (male or female) 18 to 60 years of age ➢ includes non-pregnant, non-breastfeeding women on adequate birth control* * Please note that a urine pregnancy test will be administered to all females in this interview where they will also be inquired about the type of contraception they are currently using. Acceptable effective contraceptive methods for females with child-bearing potential include: barrier methods (condoms), total abstinence, hormonal birth control methods (oral, injectable, transdermal, or intra-vaginal), intrauterine devices, and confirmed successful vasectomy of partner. ▪ Stable body weight (BMI: 18.5-29.9) ▪ Written informed consent obtained from subject and ability for subject to comply with the requirements of the study.

The exclusion criteria for this trial is as follows and the Naturopath will ask you several detailed questions: ▪ Pregnant or breastfeeding females ▪ History or presence of diabetes; ▪ History or presence of hypertension; ▪ History or presence of dyslipidemia; ▪ History or presence of major depressive disorders; ▪ History or presence of chronic liver disorders; ▪ History or presence of kidney disorders; ▪ History or presence of blood disorders; ▪ Previous bariatric surgery (or any major surgeries or medical procedures to be scheduled within the time frame of the study); ▪ Use of medication that causes significant weight gain or loss • medications that cause significant weight gain: selective serotonin re-uptake inhibitors (SSRIs) and other neuropsychiatric medications • medications that cause significant weight loss: diabetes medications, or medications clinically prescribed specifically for weight loss in obese patients

165

• You are also expected to provide information regarding any natural health products you currently consume ▪ Allergies to or inhibitions consuming all three choices of 2% lactose free milk, soy milk, or coconut milk ▪ Allergies to any ingredients in the placebo/investigational product ▪ Presence of a condition or abnormality that in the opinion of the Investigator would compromise the safety of the patient or the quality of the data.

At the end of this recruitment interview the Naturopath will determine if you are eligible to participate in the clinical trial and inform you of their decision. The Naturopath will give you a detailed explanation in case you are ineligible for the study. If you are determined to be eligible for the trial, the Naturopath will inform you that you have been recruited in the study and give you an appointment for the next clinical site visit. Please note that the investigators do not receive any monetary compensations or fees for enrolling study participants.

2. Second clinic visit (participant debriefing, provision of study materials and baseline blood collection) When you arrive for your second study visit, your height and weight will be measured again to record your BMI. At this stage you will already have been randomized into one of three study groups: i) placebo, ii) low dose avocado powder, iii) high dose avocado powder. There are 30 participants planned for this study (10 participants to be randomized to each treatment group) therefore you have a 33% chance of being randomized into any one of the three study groups. The research team will be debriefing you on all study related details and will answer all your questions as concerns. In this visit you will be provide with clinical trial related materials which include: i. Two 1kg plastic jars containing the clinical trial material. ➢ this is a 60-day supply in total, therefore you must consume 2 scoops (~40g) of powder per day for 60 days. You can prepare the test product using any of the following diluents: 1-2% milk (with or without lactose), soy milk, coconut milk, or fruit juice. ii. Trial diary (you must use this diary to log daily consumption of the study test products, record any adverse reactions, and bring with you to every trial clinic visit) iii. Blood collection kits to have your blood drawn at the local LifeLabs (65 University Avenue East, Waterloo) iv. Pre-paid gift cards in the amount of $60 (this is your compensation for taking part in the clinical trial and we expect you to use this amount for all study related expenses as well) After you are clear on all study procedures, you will be asked to head over to the local LifeLabs to have your blood drawn (7-10mL of blood (about 2 teaspoons) for baseline measurements. Note that you will have your blood drawn 3 times during the entire course of this study as it allows us to measure and record the following safety markers:

• serum alanine aminotransferase (ALT), • serum creatinine, • serum creatine phosphokinase, 166

• total serum bilirubin • glycated hemogobin (HbA1c) • complete blood count (CBC)

3. Third clinic visit (day 30 of trial) You will come in for a third clinic visit to once again have your weight and height measured (to record BMI). You will be required to fill out a safety questionnaire and be asked questions to verify if your medication history has changed or if you are pregnant. You will be given another blood collection kit so you can have your blood drawn at LifeLabs for a second time.

4. Fourth clinic visit (day 60 of trial) This will be your final clinic visit to have all the same measurements taken as visits 2, 3 and 4 and have blood drawn at LifeLabs.

. Notes on Procedures and trial Information ➢ The non-medicinal ingredients in the test products contain small quantities of the following food-grade ingredients: sodium bicarbonate, rosemary extract, rice bran extract, rice hull powder, silicon dioxide, crystalline cellulose, strawberry flavouring. ➢ If you become pregnant during the course of this trial, you must inform us to be withdrawn from the study. ➢ You must inform the research team of any new medications or natural health products you may start taking during the course of the trial. ➢ You will receive a phone call once per week to check on your progress throughout the study and to discuss any concerns you may have

POTENTIAL RISKS AND DISCOMFORTS Our priority for every participant is their well-being. This research project involves minimal risk as our avocado based supplement is a natural product formulated from a commercially-available food-grade product which should not pose any health risk or cause major discomfort. The procedures you will be subject to during the course of this study are low risk and pose little to no discomfort. Our safety questionnaires or weekly telephone follow-ups will not ask invasive or personal questions and all your responses will be kept anonymous. The study will involve monthly trips (3 in total) to the clinic and the local LifeLabs facility which will be a time commitment you will be reimbursed for. There may be some discomfort involved during the blood draw procedure but this procedure will be conducted by a certified phlebotomist who is trained to cause minimal discomfort. Please note there may be unknown risks associated with taking the investigational products. In case of a severe adverse event, please contact Nawaz Ahmed directly who will direct your call to the study Monitor (Medical Doctor).

POTENTIAL BENEFIT TO PARTICIPANT/SOCIETY Obesity and diabetes affect many people worldwide and new therapies to manage these illnesses are urgently required. Previous studies have shown that this avocado based supplement may aid in the metabolism of fat and in the management of metabolic diseases. This study is necessary to create a clinical product that may help improve the lives of millions currently suffering from debilitating metabolic disorders like obesity and diabetes. Please note that there is currently no proven clinical benefit associated with our test product. 167

TIME COMMITMENT Participants will be required to spend 5-10 minutes a day for 60 days to prepare and consume the avocado based supplement. Additionally, on day 0, 30 and 60 of the trial, participants will be required to visit the naturopath clinic for follow-up (10 min appointment) and then have their blood drawn at LifeLabs (15 minute appointment). The estimated total time commitment for this study will not exceed 15 hours. The total duration of this study should not exceed 61 days (please allow additional 2 days in case of scheduling related issues)

CONFIDENTIALITY Every effort will be made to ensure the confidentiality of any identifying information that is obtained in this study.

Who will have access to information about you in this research? All our research records are stored securely in locked cabinets and password protected computers. Only Dr. Spagnuolo, Dr. Bretz and Nawaz Ahmed will be able to view your information as they are conducting the trial and are closely concerned with the research. Please also note that monitor (Medical Doctor associated with this trial), the auditor(s), the Research Ethics Board, Health Canada, and other regulatory authority(ies) will be granted direct access to your original medical records for verification of clinical trial procedures and/or data.

What happens to your blood samples and information collected from you? Individual names are removed from all samples and case report forms (where we collect your information) and replaced by codes to ensure samples and case report forms can only be linked to the participant by people closely concerned with the research. 80% of the blood collected from you will be used by LifeLabs to produce reports on the safety of our avocado supplement. 20% of the blood collected from you will be transported to the University of Guelph to analytically detect levels of the active medicinal ingredients in your blood. Please note that confidentiality cannot be guaranteed while data are in transit over the internet. All records identifying you will be kept confidential for 25 years. If the results of the clinical trial are published, your identity will not be revealed.

PARTICIPATION AND WITHDRAWAL All participation in research is voluntary, i.e., it is your choice to participate in this study. If you volunteer to participate in this study, you may withdraw without any consequences. You may also choose to remove your data from the study. You may also choose to refrain from answering any questions in the study and remain in the study. We request you only withdraw or request to have your data removed between day zero and day 60 of the study, i.e., the duration for which you have to consume the avocado supplement or placebo. Without special circumstances or justification, requests for voluntary withdrawal after study completion (day 60) will not be possible.

The investigator may also choose to withdraw you from the study if certain circumstances warrant doing so. These circumstances would include failing to take the avocado based supplement daily or missing blood draw and clinic visit appointments.

168

RESEARCH ETHICS APPROVAL FOR THIS STUDY This project has been reviewed by the Research Ethics Board for compliance with federal guidelines for research involving human participants.

RIGHTS OF PARTICIPANTS You do not waive any legal rights by agreeing to take part in this study. Please see the guideline on consent at http://www.uoguelph.ca/research/document/reb-application-guideline-obtaining- consent. If you have questions regarding your rights and welfare as a research participant in this study (REB# 1612885), please contact: Director, Research Ethics University of Guelph [email protected] 519-824--4120 (ext. 56606) SIGNATURES

PARTICIPANTS I have read the information provided for the study Clinical investigation on the safety of avocado pulp lipids as described herein. My questions have been answered to my satisfaction and I agree to participate in this study. I have been provided a copy of this consent form.

Name of Participant (please print)

Signature of Participant Date

DESIGNEE or INVESTIGATOR

I certify that I have followed the study SOP to obtain consent from the participant. The participant understood the nature and the purpose of the study and consents to participate in this study. The participant has been given opportunity to ask questions which have been answered satisfactorily.

Name of Designee/Investigator (please print)

Signature of Designee/Investigator Date

169

Appendix E. AVO clinical study case report form

170

171

172

173

174

175

176

177

178

179

180

Appendix F. AVO clinical study safety questionnaire

181

182

Appendix G. AVO clinical study Weekly telephone log

183