The Role of Lipid Physical State in Determining In Vitro Digestibility and

In Vivo Postprandial Metabolism

by

Surangi H. Thilakarathna

A Thesis

presented to

The University of Guelph

In partial fulfilment of requirements

for the degree of

Doctor of Philosophy

in

Human Health and Nutritional Sciences

Guelph, Ontario, Canada

© Surangi H. Thilakarathna, April 2019

i

ABSTRACT

THE ROLE OF LIPID PHYSICAL STATE IN DETERMINING IN VITRO

DIGESTIBILITY AND IN VIVO POSTPRANDIAL METABOLISM

Surangi H. Thilakarathna Advisor:

University of Guelph Dr. Amanda J. Wright

Lipid digestibility and postprandial lipemia (PPL) have implications for health and disease,

although the possible contributions of triacylglycerol (TAG) physical properties to these

processes remain ill-defined. This thesis investigated the role of TAG physical state, specifically

on in vitro digestibility and PPL using compositionally similar interesterified (IE) lipids with

different solid fat contents (SFCs) and compositionally identical undercooled liquid (LE) and

partially crystalline solid (SE) emulsion droplets. A secondary aim was to compare the in vitro

and in vivo results obtained.

The first study with stearic-rich IE lipids showed lower free fatty acid (FFA) bioaccessibility and

higher levels of “excreted” stearic acid in the undigested TAG for non-IE vs IE blends. Those

results obtained using the TIM-1 dynamic digestion simulator indicated lower digestibility for

solid fat versus liquid oil and correlated well with previous postprandial TAG human data for the

same lipids. Tempering was used to formulate and compare compositionally identical LE and SE

emulsion droplets based on palm stearin. The droplets were similarly sized, charged, and shaped,

eliminating factors that might have otherwise influenced TAG digestibility. Using a static in

vitro digestion method, SE lipolysis was slower and reduced compared with LE. The presence of

crystalline fat was also associated with extensive partial coalescence with exposure to simulated

ii

gastric conditions and differences in in vitro lipolysis between LE and SE were exaggerated

dependent on in vitro digestion conditions, i.e. gastric pH and shear. In a randomized double-

blinded crossover acute meal study with healthy male participants, SE showed a significantly

delayed postprandial plasma TAG increase from baseline and a decrease in plasma TAG

incremental area under the curve. Consistently, the postprandial plasma TAG parameters

indicated attenuated response to SE versus LE, validating the in vitro observations. The observed

differences point specifically to isolated effects of SFC on PPL. Overall, this thesis evidences

that the presence of solid fat decreases digestibility compared to lipids in the liquid state. It

highlights that TAG melting temperature impacts digestibility, with implications for lipemia,

shown both for bulk and emulsified lipids (i.e. stearic-rich IE lipids and tempered partially

crystallized and undercooled palm stearin emulsions, respectively).

iii

DEDICATION

I want to dedicate this thesis to my parents…. for their unconditional love, kindness, and care.

I am who I am today because of you!

iv

ACKNOWLEDGEMENTS

First and foremost, I want to thank my advisor, Dr. Amanda Wright for giving me this invaluable

opportunity to do a PhD project under her supervision. Amanda, you always had faith in me that

I would be successful. Your encouraging and reassuring words always lifted me up especially,

when I was facing challenging situations, in the lab and at home. You always want the best for

your students and made sure that they reach their full potential. I learnt so much from you during

these past few years. As an expert in the research field having a successful career, I truly admire

your devotion to work, dedication to your students, and at the same time, balancing work and

life. Amanda, you are my role model! Thank you for your patience, guidance, and kindness.

I want to thank Dr. Alejandro G. Marangoni and Dr. Lindsay Robinson for serving in my

advisory committee, my qualifying examination committee, and their contribution to this thesis.

It was my privilege to work with you. I learnt so much from these experts in the field. Given the

interdisciplinary nature of the project between food science and nutrition, it was such a learning

experience for me. Your guidance and intriguing questions helped me understand the strong and

fascinating connection between food science and nutrition. I also want to thank Dr. Bill Bettger

and Dr. Michael Rogers for serving in my qualifying examination committee. You both made the

stressful examination process a fun learning experience and I was able to expand my knowledge

beyond my project which helped me understand the network connecting many aspects of food

science and nutrition.

I want to extend my sincere thanks to Dr. Amy Tucker, the manager of the Human Nutraceutical

Research Unit (HNRU), our exceptional team of phlebotomists Premila Sathasivam, James

v

Turgeon and Nina Andrejic for all their assistance and contribution towards a successful human

clinical trial. I also want to thank the “Lipid Digestion Study” participants as without them, this

human study would not have been possible. I also want to thank the department of Food science,

University of Guelph for all the great collaborations. Especially Dr. Michael Rogers who was a

collaborator through the New Jersey Institute of Food, Nutrition, and Health, Department of

Food Science, Rutgers State University, New Jersey, USA but later through Department of Food

Science, University of Guelph. I also want to thank Dr. Lisa Duizer from the Department of

Food Science, University of Guelph for providing access to the food formulation kitchen to

formulate our lipid emulsions for the human study. I want to thank Dr. Alejandro Marangoni

from the same department for providing access to use numerous equipment use in his laboratory.

I also want to thank Dr. David Ma and Ms Lyn Hillyer from the Department of Human Health

and Nutritional Sciences (HHNS), University of Guelph for all their assistance for the gas

chromatographic analysis. I want to extend my gratitude to the past and present graduate

program coordinators, Drs. David Wright and Graham Holloway for taking all their efforts

towards a successful graduate program and best graduate student experience. I cannot forget the

amazing “A-team” Ms. Andra Williams, Ms. Ann Stride, and Ms. Anne Lovett-Hutchinson for

their excellent administration work as well as great conversations. Our “A-team” most definitely

deserves an A+.

I want to thank the past and present Wright lab team members including Sally Huynh, Amanda

Cuncins, Hannah Neizer, Samar Hamad, Melissa Brown, Dr. Nilourfar Rafiee Tari, Samantha

Hart, Kevin (Kim) Min, Don Lochana Ekanayake, and Darrah Condino with a very special

thanks to my good friend and colleague Xinjie (Lois) Lin. Lois, I was very fortunate to do the

PhD side by side with you and I had such a wonderful experience which would not have been the

vi

same if it wasn’t for you. Your friendship and kindness helped me get through some of the

hardest and challenging times. At the same time, I learnt so much from you. Your critical

thinking skills, thirst for adventure, bright and bubbly personality lightened up the lab. Your

perspective of life taught me to think more positively and accept challenges. I also want to thank

my fellow HHNS colleagues for all their support and encouragement and friendship that helped

me keep my sanity as well as lift my spirits.

I want to extend my very special thanks to Mrs. Thakshila De Zoysa for her kind words and

support during hardest times of my life trying to balance work and life, with a very young infant

in my hands. I almost gave up but you encouraged me to keep going, to hang in there and I will

be proud of myself one day. And here I am! I’m forever grateful to have such good friends in my

life. I want to thank my dear loving parents, for their unconditional love and care that brought me

all the way here. You taught me the importance and value of education, you taught me the value

of becoming an empathetic human being who cares for others. Even though you are so far away

from me, you are always close to my heart. Last but not the least, I want to thank my husband,

my significant other, Dr. Malinda Thilakarathna, for everything done from being my personal

mentor, motivational speaker, to my loving and caring partner. Without your support at home, all

this would only be a dream. We had very tough times and when looking back, I’m really proud

of ourselves. We survived two PhDs! I want to thank my two little girls, Hasini and Dinithi, my

very own person cheerleaders. Thank you for keeping your mom sane, thank you for all the

much-needed distractions and your hugs and kisses. Life was tough but if I’m to do this all over

again, I would not change a thing. You brought purpose and meaning to my life!

vii

TABLE OF CONTENTS

ABSTRACT i

DEDICATION iii

ACKNOWLEDGEMENTS iv

TABLE OF CONTENTS vii

LIST OF TABLES x

LIST OF FIGURES xi

LIST OF ABBREVIATIONS xiii

LIST OF APPENDICES xv

Chapter 1 1

1.1 Introduction 2

1.2 Brief overview of human lipid digestion 5

1.3 Saturated fat and impact on health 8

1.4 Techniques to compare lipid digestibility 9

1.5 Investigating digestibility differences between solid vs liquid lipids 11

1.5.1 Interesterified (IE) lipids 11

1.5.2 Dispersed lipids 14

1.6 Summary 18

Chapter 2 19

2.1 Rationale 20

2.2 Objectives and Hypotheses 21

2.2.1 Specific objective 1 and hypothesis (for Chapter 3) 21

2.2.2 Specific objective 2 and hypothesis (for Chapter 4) 21

2.2.3 Specific objective 3 and hypothesis (For Chapter 5) 22

Chapter 3 23

3.1 Abstract 24

3.2 Introduction 25

3.3 Methodology 28

viii

3.3.1 Materials 28

3.3.2 The test fats 29

3.3.3 LC/MS regiospecific analysis of the test fats 29

3.3.4 Meal preparation 30

3.3.5 TIM-1 in vitro digestions 30

3.3.6 Sample collection 32

3.3.7 Total FFA measurements 32

3.3.8 FA composition of bioaccessible lipids 32

3.3.9 FA composition of ileal efflux lipids 33

3.3.10 Melting behavior by differential scanning calorimetry (DSC) 33

3.3.11 Statistical analysis 34

3.4 Results and Discussion 34

3.4.1 TIM-1 digestion: lipid bioaccessibility 34

3.4.2 TIM-1 digestion: unabsorbed lipids in the ileal efflux 42

3.4.3 Correlations between TIM-1 and human study results 45

3.5 Conclusions 46

3.6 Acknowledgements 47

3.7 Bridge to Chapter 4 47

Chapter 4 49

4.1 Abstract 50

4.2 Introduction 51

4.3 Materials & Methods 53

4.3.1 Materials 53

4.3.2 Triacylglycerol composition analysis 53

4.3.3 Preparation and analysis of palm stearin emulsions 54

4.3.4 Particle size and -potential 55

4.3.5 Melting and re-crystallization behavior 55

4.3.6 Lipid polymorphism 56

4.3.7 Solid fat content (SFC) 56

4.3.8 Emulsion droplet morphology 56

4.3.9 In vitro digestion during exposure to gastric and duodenal conditions 56

4.3.10 Data and statistical analysis 58

4.4 Results and Discussion 58

4.4. 1 Compositional, physical property and morphological analyses 58

4.4.1 In vitro duodenal lipolysis without exposure to a gastric phase 63

ix

4.4.2 In vitro duodenal lipolysis following exposure to a gastric phase 65

4.4.3 Role of gastric parameters in determining duodenal lipolysis 69

4.5 Bridge to Chapter 5 72

Chapter 5 73

5.1 Abstract 74

5.2 Introduction 76

5.3 Materials and Methods 79

5.3.1 Study design 79

5.3.2 Participants 80

5.3.3 Study visit protocols 81

5.3.4 Emulsion preparation 82

5.3.5 Emulsion droplet characterization 83

5.3.6 Blood sample collection and analysis 84

5.3.7 Statistical analysis 85

5.4 Results 87

5.4.1 Participant characteristics 87

5.4.2 Test emulsion characteristics 89

5.4.3 Postprandial plasma TAG responses 90

5.4.4 Postprandial plasma NEFA response 93

5.4.5 Changes in postprandial CMRF 94

5.4.6 Post-study visit gastrointestinal and bowel habits 96

5.4.7 Correlation between previous in vitro and human study results 98

5.5 Discussion 99

Chapter 6 107

6.1 Preamble 108

6.2 Summary of major findings, strengths and limitations 109

6.3 Limitations in the different digestion models used 112

6.4 Future directions 114

6.5 Conclusions 115

REFERENCE 116

APPENDICES 126

x

LIST OF TABLES

CHAPTER 3

Table 3.1 Fatty acid and triacylglycerol compositions of NIE, CIE, and EIE. 35

Table 3.2 Fatty acid (palmitic, stearic, oleic, and linoleic acids) concentration

(cumulative, wt%) of bioaccessible lipids following 6 h TIM-1

digestion of NIE, CIE, and EIE1 41

CHAPTER 5

Table 5.1 Characteristics of the 15 healthy male participants at baseline1 89

xi

LIST OF FIGURES

CHAPTER 1

Figure 1.1 A schematic summarizing the main steps involved in triacylglycerol

digestion in the small intestine 6

CHAPTER 3

Figure 3.1 A schematic of the TNO intestinal model (TIM)-1 28

Figure 3.2 Bioaccessibility (%) of NIE, CIE, and EIE during 6 h in vitro

TIM-1 digestion (n=2) 36

Figure 3.3 DSC thermograms of NIE, CIE, and EIE fats and extracted

lipids from the NIE, CIE, and EIE ileal efflux from the TIM-1

digestion simulator (n=3) 39

Figure 3.4 Palmitic, stearic, oleic, and linoleic acid composition (wt %) of

triacylglycerol and free fatty acid fractions of lipids remaining

in the 6 h pooled TIM-1 efflux from NIE, CIE and EIE (mean ± SEM, n=2) 43

Figure 3.5 Correlation analyses between TIM-1 6 h bioaccessibility (%, error bars

indicate the SEM, n=2) and human study 6 h postprandial serum TAG

concentration of non-obese subjects (mmol/L, n=10) fed the same test fats 46

CHAPTER 4

Figure 4.1 Representative DSC thermograms of melting and crystallization of

liquid (LE) and solid (SE) droplets, bulk palm stearin (PS) tempered

as per SE with PS SFC curve inset, X-ray diffractograms for bulk PS and SE 59

Figure 4.2 Representative particle size distributions of LE and SE 60

Figure 4.3 Microstructure of LE, SE, and 10% canola oil-0.4%

Span 60 emulsion droplets under bright field and polarized light conditions 62

Figure 4.4 Free fatty acid release during duodenal digestion without the

gastric phase, duodenal digestion with the gastric phase,

and gastro-duodenal digestion with added glass beads (d=10mm) 65

Figure 4.5 Representative graphs showing particle size distributions of duodenal

digestion of LE and SE 66

xii

Figure 4.6 Gastric phase impact on LE and SE particle size distribution 70

CHAPTER 5

Figure 5.1 Participant flow through the study 88

Figure 5.2 Characteristics of the undercooled liquid (LE) and crystalline solid (SE)

emulsions. Representative DSC thermograms of LE and SE showing

the melting behavior and particle size distributions and the solid fat

content of bulk palm stearin 90

Figure 5.3 Mean change in plasma TAG concentration from baseline and TAG

iAUC0-6h and iAUC1-6h after participants consumed 50 g of 10 % palm stearin

and 0.4% span 60 in-water emulsion either as undercooled liquid (LE) or

crystalline (SE) emulsion droplets 92

Figure 5.4 Mean change in plasma NEFA concentration change from baseline after 15

participants consumed 50 g of 10 % palm stearin and 0.4% span 60 in-water

emulsion with either undercooled liquid (LE) or crystalline (SE) emulsion

droplets 93

Figure 5.5 CMRF mean concentration of saturated and unsaturated fatty acids,

proportional FA composition in the initial emulsion, CMRF-TAG

fraction at fasting state and 4 h postprandial consuming LE and SE, and

CMRF particle size change through the 6 h postprandial period 95

Figure 5.6 Self-reported gastrointestinal symptoms and bowel habits for the 24 h

period after participants’ each study visit 97

Figure 5.7 Correlation analysis between in vitro lipolysis (%, n=3) and baseline

adjusted plasma TAG concentrations (mmol/L, n=15) for both LE and

SE as a group and LE and SE separately, during 1-4 h in vitro and in vivo

digestions 99

xiii

LIST OF ABBREVIATIONS

ANOVA Analysis of variance

AUC Area under the curve

BMI Body mass index

CIE Chemically interesterified

CMRF Chylomicron rich fraction

CVD Cardiovascular disease

DAG Diacylglycerol

DSC Differential scanning calorimetry

EDTA Ethylenediaminetetraacetic acid

EIE Enzymatically interesterified

FA Fatty acid

FFA Free fatty acid

GC Gas chromatography

GLM General linear model

HDL High density lipoprotein

HNRU Human Nutraceutical Research Unit

iAUC Incremental area under the curve

IE Interesterified

LC/MS Liquid chromatography/mass spectroscopy

LDL Low density lipoprotein

LE Undercooled liquid emulsion

MAG Monoacylglycerol

NEFA Non-esterified fatty acid

NIE Non-interesterified

OOO Triolein

PPL Postprandial lipemia

PPP Tripalmitin

PS Palm stearin

REB Research ethics board

SD Standard deviation

SDS Sodium dodecyl sulphate

SE Crystalline solid emulsion

SEM Standard error of mean

SFC Solid fat content

Span 60 Sorbitan monostearate

SSL Sodium stearoyl lactylate

SSS Tristearin

xiv

TAG Triacylglycerol

TIM-1 TNO intestinal model-1

TIM-2 TNO intestinal model-2

TLC Thin layer chromatography

Wt% Weight percentage

XRD X-ray diffraction

xv

LIST OF APPENDICES

Appendix 1 University of Guelph Research Ethics Board Certificate of Approval 126

Appendix 2 The Lipid Digestion Study Informed Consent Form 127

Appendix 3 The Lipid Digestion Study Participant Recruitment Poster 146

Appendix 4 The Lipid Digestion Study Phone Screening Questionnaire 147

Appendix 5 The Lipid Digestion Study In-Person Screening Eligibility Questionnaire 156

1

CHAPTER 1

INTRODUCTION AND OVERVIEW

2

1.1 Introduction

Dietary lipids, including triacylglycerols (TAGs), with their constituent fatty acids (FAs), are

important to human health. However, their excessive consumption and certain species of FA

(trans species, in particular) have been linked to various metabolic diseases and conditions,

including obesity, metabolic syndrome, cardiovascular disease (CVD). Moreover, lipid

consumption has implications for satiety and appetite control, energy regulation, as well as

bioactive delivery and release. Despite this, there remains much to learn about the specific

contributions of dietary fat to digestive processes that impact subsequent metabolism and disease

risk. Contradictory findings and discrepant opinions highlight the need for further study and new

lines of investigation. They also rationalize the use of different approaches to study lipid

digestion topics. For example, opinions on the relationships between saturated fat consumption

and health remains a matter of debate. It was recently reported that replacing dietary saturated fat

with polyunsaturated fat reportedly reduced CVD mortality by 30%, i.e. a similar level of

reduction as statins, but not when replaced with carbohydrates (Sacks et al., 2017).

Contrastingly, another analysis concluded that total fat and type of fat, i.e. saturated and

unsaturated fats, were not significantly associated with myocardial infarction or CVD mortality

and saturated fat, in fact, had an inverse association with stroke (Dehghan et al., 2017).

Overall, a better understanding of lipid digestion and metabolism will help to elucidate the

nuances of dietary fat consumption on human health and disease risk. This thesis focuses on the

rarely considered possibility that the physical properties of dietary TAGs contribute to their

digestibility, bioaccessibility and absorption. Interest in the role of lipid physical properties on

digestion and metabolism dates back several decades (Michalski et al., 2013). However, there are

many unexplored questions on this topic, as well as unclear and conflicting evidence. Relatively

3

a lot more focus has been directed towards understanding the influence of lipid composition

versus physical properties on digestion, metabolism, and health. This includes extensive studies

of the relationships between FA type/positional distribution and TAG composition, on

postprandial lipemia, i.e. the increase in plasma TAG observed after a fat-containing meal. These

studies have largely relied on making comparisons between saturated and unsaturated fatty acid

rich meals. The possibility that the associated differences in physical properties could contribute

to differences in metabolism is rarely considered.

Studies of lipid physical properties affecting digestibility, absorption and postprandial lipemia

have not always yielded consistent results. For example, in vivo (Berry, Miller, & Sanders, 2007;

Hall et al., 2014; Robinson et al., 2009) lipids with a higher solid fat content (SFC) at body

temperature (37 C) were digested slower than lipids with comparatively lower SFC at 37 C,

pointing towards differences in lipid digestion kinetics. In agreement, animal studies have

reported excretion of undigested solid fat after rats consumed a higher solid fat containing rodent

diets (Bergstedt, Hayashi, Kritchevsky, & Tso, 1990; Kaplan & Greenwood, 1998; Wang,

Corwin, Jaramillo, Wojnicki, & Coupland, 2011). More recently, research has been spurred on

by advances and availability of in vitro gastrointestinal digestion methodology to investigate

digestive mechanisms. Some in vitro studies have shown similar trends of lower digestibility for

crystalline lipids (Golding et al., 2011; Nik, Langmaid, & Wright, 2012b). Contrastingly, other

reports have shown no difference in digestibility and metabolism of diets with varying SFCs

(Wang, Wang, Spurlock, & Wang, 2015).

Importantly, it has been stated that lipid digestion is efficient and complete in the healthy adult

human gastrointestinal tract, regardless of lipid physicochemical properties (Mu & Høy, 2004).

4

Assuming this to be true, there are many instances of conditions or diseases in which lipid

digestion is sub-optimal. Moreover, emerging evidence points to relationships between the

kinetics of postprandial lipemia and endotoxemia through the absorption of gut-microbiota

derived lipopolysaccharide and partitioning of lipids between storage versus beta oxidation

(Laugerette, Vors, Peretti, & Michalski, 2011; Vors et al., 2013). Lastly, although steatorrhea is

diagnosed when >5% of lipids is excreted in stool and associated with GI upset, it is conceivable

higher melting lipids are retained as crystalline solids and excreted, unnoticed (Kaplan &

Greenwood, 1998). There have been promising advancements in the field of stable isotope

techniques, especially techniques developed to recover [1,1,1-13C3]tripalmitin and [1,1,1-

13C3]triolein from stools, that have potential to be used in lipid digestibility studies (Gabert et

al., 2011). However, such studies are expensive and would themselves rely on using TAG

species with different melting temperatures.

Therefore, addressing basic questions which remain about TAG digestibility could close the

research gap about lipid physical properties on digestion, metabolism and postprandial lipemia,

with wide-ranging implications. A better understanding on this topic could help to explain

previous apparent contradictions or discrepancies and allow for manipulations in food lipid

properties to formulate lipids with improved functionalities and health benefits. For example, this

could be to optimize the release of encapsulated molecules within a particular region of the

gastrointestinal tract or to induce feelings of satiety or to promote or minimize lipolysis, without

negative side effects. The forthcoming sections will provide a brief overview of lipid digestion in

healthy adult humans, the growing body of evidence pointing to differences in solid and liquid

lipid digestion, and techniques used to study lipid digestibility.

5

1.2 Brief overview of human lipid digestion

Human digestion is a complex series of processes involving physicochemical and structural

changes that facilitate chemical and physical breakdown and absorption of ingested foods and

nutrients, including lipids. Lipid digestion can potentially take place in the oral cavity although

the chemical degradation of lipids from lingual lipases is insignificant in adult humans, although

relevant for human infants and the rodents (Mu & Høy, 2004). The main changes due to oral

processing of lipids is structural breakdown, depending on the characteristics of the ingested

foods, specifically dispersal and initial emulsification of bulk lipids into oil droplets or

potentially emulsion droplets coalescence and flocculation upon interactions with biopolymers in

saliva (McClements, Decker & Park, 2009).

The extent of gastric lipolysis can range from 10-30 % and should not be neglected, as it can set

the stage for subsequent lipolysis. The optimal activity of gastric lipases at low pH results in the

formation of free fatty acid (FFA) and diacylglycerols (DAGs) due to sn-1 or -3 specificity (Mu

& Høy, 2004). In healthy adult humans, gastric processing promotes duodenal digestion through:

promoting lipid droplet disruption, increasing solubilization of lipolysis products, stimulating

hormonal release, increasing binding of co-lipase, and increasing activity of pancreatic lipase

(McClements et al., 2009). Motility in the gastric compartment through peristalsis and antral

contraction waves facilitates further physical breakdown of ingested foods, increasing the surface

area for gastro-duodenal digestion (Bornhorst & Singh, 2014). Gastric emptying is a critical

parameter for lipid digestion and the gastric residence time can depend on the quantity, physical

state, and structure of the lipids ingested (McClements et al., 2009).

6

Figure 1.1 A schematic summarizing the main steps involved in triacylglycerol digestion in the

small intestine. FATP: fatty acid transport protein (Adapted from Goodman, 2010).

Release of the partially digested “chyme” consisting of emulsified lipids, i.e. TAG, DAG,

monoacylglycerols (MAGs), FFA, cholesterol, fat soluble vitamins, etc., into the duodenum

induces the release of bile and phospholipids from the gallbladder and/or the liver and the

duodenal synthesized hormone ‘cholecystokinin’ that induces release of pancreatic lipase

7

(McClements et al., 2009). Sodium bicarbonate secretion from the pancreas increases the pH of

the chyme from 1-3 to 5.8-6.5, although this can vary considerably between individuals

(McClements et al., 2009). To facilitate lipase activity, bile salts and phospholipids further

emulsify the lipids to increase the interfacial surface area (Mu & Høy, 2004). Since interfacial

area is required for lipase activity, smaller emulsion droplets at the start of duodenal digestion

tend to have increased lipolysis than larger droplets (Armand et al., 1999). Recently, an MRI

study reported larger structures of solid fat were released from the gastric compartment into the

duodenum without being physically broken down (Steingoetter et al., 2015). The lack of increase

in the postprandial plasma TAG in that case indicated that the solid fat masses passed through

the duodenum without being “detected” and hence digestion by pancreatic lipase due to lack of

hormonal response (Steingoetter et al., 2015) and agrees with similar previous reports (Golding

& Wooster, 2010; Golding et al., 2011). Furthermore, apart from the structural modifications, the

presence of solid lipids is also known to alter duodenal lipolysis by altering lipase accessibility

due to crystalline interfaces (McClements et al., 2009).

Once bile salts are adsorbed at a lipid droplet interface, co-lipase facilitates the access of

pancreatic lipase as the higher surface activities of bile salts tend to displace pancreatic lipase

from the interface. Pancreatic lipase accounts for the majority of dietary TAG lipolysis (Mu &

Høy, 2004; Wilde & Chu, 2011). The sn-1 & -3 specificity of pancreatic lipase results in

formation of FFA and 2-MAG which are more hydrophilic than the TAGs and tend to

accumulate at the emulsion droplet interface, impeding continued lipolysis (Wilde & Chu, 2011).

When long chain saturated FAs located at sn-1 and -3 are released from TAGs, a tendency

towards producing crystalline phases at the presence of high divalent cations, i.e. calcium and

magnesium soap formation, can impede the absorption of these FA (McClements et al., 2009;

8

Mu & Høy, 2004). Differences in soap formation have complicated physical state comparisons

and interpretations of studies with long chain saturated versus saturated fatty acids.

Briefly, FA absorption occurs once these molecules have moved through the unstirred water

layer of the intestinal lumen to the enterocytes (Ros, 2000). Absorbed small and medium chain

length FAs, <C8 and C8 to C12 respectively, enter in to the portal vein and the systemic

circulation via the liver (McClements et al., 2009). Longer chain FAs are re-assembled into

TAGs in the enterocytes and packed into enterocyte derived chylomicron lipoproteins, secreted

to the lymphatic system and bypass the liver (McClements et al., 2009). Chylomicrons are

transported through systemic circulation and degraded by lipoprotein lipases at different

destination tissues, i.e. muscle and adipose tissue (Hussain, 2014). As such, changes in plasma

chylomicron parameters, i.e. size and composition, are indicators of postprandial responses to

dietary fat digestion. Undigested and/or unabsorbed dietary TAGs are excreted in the feces.

Interestingly, animal studies have shown excretion of solid fat, without obvious discomfort in the

animal (Wang et al., 2011). In humans, significant excretion of liquid oil can cause a sense of

urgency and oily stools, i.e. a condition known as steatorrhea, but questions of whether excreted

solid fat could be excreted without issues, remain.

1.3 Saturated fat and impact on health

Excessive fat consumption is linked with postprandial lipemia which is the elevation in plasma

TAG levels after consuming a fat rich meal and is considered as an independent risk factor for

CVD (Kolovou et al., 2005). Dissimilar to consistent findings on polyunsaturated fat being

linked to reduced CVD risk (Mozaffarian & Wu, 2011) and trans-fat being linked to detrimental

cardiovascular health (Crupkin & Zambelli, 2008), the impact of saturated fat on CVD is

9

inconsistent and a matter of ongoing debate. Findings from a prospective cohort study reported

that the total fat and the type of fat were not associated with CVD, myocardial infarction, CVD

mortality and saturated fat had an inverse relation to stroke (Dehghan et al., 2017). Completely

contradictory to these findings, a presidential advisory from American Heart Association

summarized that, in randomized controlled trials that lowered the intake of dietary saturated fat

and replaced with polyunsaturated vegetable oil reduced CVD by ~30% which was similar to the

reduction achieved by statin treatment (Sacks et al., 2017). Apart from predicted methodological

discrepancies in these and similar studies (Nettleton et al., 2017), it is speculated that the

differences in the saturated fat physical state could potential contribute to the reported

differences in these studies. In fact, it was previously reported in a review focussing on the

impact of types of saturated fat and CVD that different species of saturated fat have different

effects on serum lipids and lipoproteins (Micha & Mozaffarian, 2010). That meta-analysis of

randomized controlled trials, summarized that total and LDL cholesterol were raised by lauric,

myristic, and palmitic acids, but not by stearic acid, when carbohydrate was used as a reference

(Micha & Mozaffarian, 2010). Interestingly, the speculated physical state impacts of saturated fat

and dietary lipids in general, on lipid metabolism have most recently been mainly incidental

based on findings from human clinical trials seeking to understand the metabolic fate of

interesterified fat (Berry, 2009). However, the fatty acid compositional differences in

interesterified test fats limits the cleanest investigation physical states.

1.4 Techniques to compare lipid digestibility

Various multi-level tools and models can be used to investigate the impact of lipid physical state

on lipid digestibility. Broadly speaking, digestion models can be categorized as in vitro, in vivo

animal and in vivo human models. There are many types and variations of in vitro digestion

10

models and these can be of varying complexities (Guerra et al., 2012). More simple models like

the pH stat and static mono-compartmental gastroduodenal models measure the liberation of FAs

via titratable acidity or FFA concentrations. More complex systems like the TNO-TIM dynamic

gastroduodenal models not only measure FFA release, but also estimated lipid absorption and

excretion using membranes and employ closely controlled conditions, i.e. dynamic pH levels,

removal of nutrients mimicking absorption, gastro-duodenal motility, etc. that increase their

physiological relevance (Guerra et al., 2012; Minekus, Marteau, Havenaar, & Huis in’t Veld,

1995). Importantly, in vitro models can provide insightful visual observations on structural

modifications that occur during lipid digestion. Since the contribution of the oral and the gastric

phase to chemical breakdown of lipids is considerably smaller, some very simple in vitro

digestion models have consisted of only the duodenal phase, although this has become less

common.

Animal models have also been widely used to study the impact of lipid physical state on

digestion and metabolism (Michalski et al., 2013), including mice (Wang et al., 2015), rats

(Bergstedt et al., 1990; Wang et al., 2011), and guinea pigs (Asselin et al., 2004). Human studies

are considered “the golden standard”, although there are ethical challenges and limitations in

terms of the data that can be collected. Normal lipid metabolism as observed in healthy

individuals is altered in ‘at disease risk’ and ‘with disease’ populations (Lairon, Lopez-Miranda,

& Williams, 2007). Therefore, different populations have been utilized in lipemia studies,

depending on the study objective. Study participants have been healthy young males and/or

females, postmenopausal females, populations ‘at disease risk’, populations ‘with disease’, etc.

Interestingly, even within the healthy young populations, females (pre-menopausal) and males

metabolize lipids distinctly differently (Dias, Moughan, Wood, Singh, & Garg, 2017), so

11

participant selection to meet a study’s objective and minimize variability requires careful

consideration.

In vitro models range from very crude to highly complex. All aim to simulate conditions of the

gastrointestinal tract, drawing on studies where these parameters have been published, but none

can perfectly mimic conditions. It is important to consider physiological relevance and the

accuracy of findings obtained with in vitro models. On this note, despite a huge increase in

publications using in vitro digestion tools to investigate lipids and other nutrients, very few

studies have attempted to correlate findings among different digestion models. Further, interest

towards more in vitro and in vivo correlations on digestibility and postprandial responses of high

melting, high SFC fats vs low melting, low SFC fats were being made (Golding et al., 2011).

However, there are very few studies that show direct correlation of models that studied digestion

of the same test lipid. This approach can be very useful in interpreting and validating the finding

from one model using another model.

1.5 Investigating digestibility differences between solid vs liquid lipids

1.5.1 Interesterified (IE) lipids

Investigating the specific role of lipid physical versus chemical properties in digestibility,

gastrointestinal processing and metabolism is challenging, as these two properties are tightly

interrelated. IE lipids offer a good context in which to consider this challenge. Interesterification

(chemical randomization or enzymatic) of lipids has become a popular processing strategy that

allows manipulation of lipid composition, more specifically the stereo specific positioning of FA

within TAG molecules. This is performed to achieve so-called structured lipids for nutritional

purposes and, for the food industry, lipids with desirable physical properties, including those

12

with lower or higher plasticity and SFC, given these are determined by the range of melting

temperatures of the TAG species present. A fat or fat blend and the resulting IE lipid will have

similar FA compositions, but different TAG compositions.

The alterations in TAG composition make it challenging to isolate any specific impacts of

physical properties, specifically, and vice versa. However, IE dietary lipids are a framework for

considering the direct and indirect impacts of SFC on digestion and absorption. Previous studies

where SFC differences were speculated to be responsible for absorption differences were based

on IE and native palm oil (Berry, Woodward, Yeoh, Miller, & Sanders, 2007; Sanders, Filippou,

Berry, Baumgartner, & Mensink, 2011), IE and native shea blends (Berry et al., 2007), non-IE

high oleic sunflower oil and canola stearin blends and their IE versions (Robinson et al., 2009),

non-IE and IE palm stearin and palm kernel oil (Hall, Iqbal, Li, Gray, & Berry, 2017), IE and

native palm olein (Hall et al., 2014), and different palmitic and oleic acid containing synthesized

TAGs using esterification (Wang et al., 2015). In most of these cases, the observations about

physical property differences were incidental. For example, in a human study, it was shown that

after consuming a randomized and an unrandomized shea blends with SFC ~41% and 22%

respectively, PPL was 53 % lower for the randomized shea blend and speculations were drawn

that lower PPL was due to the higher SFC (Berry et al., 2007). It was proposed that poor

emulsification and poor accessibility of lipases at the interface of the lipids with high SFC

interfered with the digestion breakdown resulting in a slower rise in plasma TAG.

Evidence from interesterification studies reignited interest in clarifying the role of TAG SFC,

specifically on metabolic response, when it was observed that higher lipid blends with a SFC at

physiological temperature showed a slower postprandial lipemic response compared to the lower

13

or no SFC at the same temperature (Berry et al., 2007; Hall et al., 2017; Sanders et al., 2011).

The difference in SFC for some studies were 0-4.8% (Sanders et al., 2011), 0-15.2 % (Berry et

al., 2007), 0-22 % (Berry et al., 2007), and 5.4-18.6 % (Robinson et al., 2009). These were also

differences in TAGs and, in some cases, minor lipid compositions that complicate interpretation

and comparison of the findings. Indeed, changes in the metabolic fate of these structurally

modified lipids are attributable to both changes in physical properties and FA stereospecific

position that affects their digestion and absorption related to sn-specific lipases and preferential

absorption of sn-2 monoacylglycerol species. However, by controlling for FA composition,

studies of IE lipids can offer insights into the role of physical properties during lipid digestion. In

general, there is a trend towards the presence of a higher SFC at physiological temperature being

associated with lower plasma TAG increase. These incidental findings of differences in IE lipids

helped rationalize more recent in vitro (Bonnaire et al., 2008; Guo, Bellissimo, & Rousseau,

2017; Hart, Lin, Thilakarathna, & Wright, 2018; Nik et al., 2012b; Nik, Langmaid, & Wright,

2012a) and in vivo (Wang et al., 2015) studies specifically targeting the impact of SFC on

digestibility and PPL.

A human study by our research group investigated the acute metabolic effects after consuming a

non-IE (NIE) high oleic sunflower oil and canola stearin blend (70:30) and its chemically (CIE)

and enzymatically (EIE) IE blends by healthy individuals (Robinson et al., 2009). The SFCs of

these lipid blends at 37 C were 18.6, 5.6, and 5.4 % for NIE, CIE, and EIE, respectively

showing a significant difference in SFC among the IE vs non-IE blends. The stearic acid

containing high melting trisaturated TAGs (SSS or tristearin) were also different in the blends

where, NIE, CIE, and EIE had 23.3, 0.5, and 0.4 wt% respectively. Interestingly, there was no

increase in serum stearic acid after ingesting NIE but an increase in serum oleic. It was

14

speculated that the high melting tristearin was not readily emulsified in the gastrointestinal

system and thereby, was not available for digestion breakdown or subsequent absorption,

resulting in lower serum stearic acid concentration (Robinson et al., 2009). However, this was

speculative and possible physical property differences being responsible for the observed lipemic

response trends were not investigated. Therefore, the first study in this thesis (Chapter 3)

investigated the in vitro digestion of this same blend and the chemically and enzymatically IE

blends of canola oil and canola stearin and integrated these results with those of the human

postprandial study (Robinson et al., 2009). There have not been any in vitro studies reported in

literature that using IE lipids as the model fat to-date. Therefore, this thesis aimed to explore the

mechanistic role of SFC in lipid digestibility for IE lipids using in vitro digestion methods and to

compare the results to postprandial lipemia data for the same IE stearic acid-rich blends (Chapter

3).

1.5.2 Dispersed lipids

Many food and natural health products are oil-in-water emulsions. Emulsions have become a

common framework in which to study the influence of lipid properties on digestibility. Our

group previously compared emulsions with lipids of different physical states in terms of

digestive processing and encapsulated bioactive release (Nik et al., 2012a, 2012b). This work

showed slower lipolysis kinetics for the solid lipid particles compared to the liquid lipid particles

(Nik et al., 2012b) and slower bioactive release in emulsion particles containing solid fat (Nik et

al., 2012a). Very recently, a series of emulsions with different SFCs achieved by combining

soybean oil and hydrogenated soybean oil reported an inverse correlation between the SFC and

rate of lipid digestion (Guo, Bellissimo et al., 2017). Previously, canola oil emulsions stabilized

with sodium stearoyl lactylate (SSL) consisting of ~25 % SFC (added hydrogenated vegetable

15

oil) showed slower rate of lipolysis as measured by titratable acidity compared to the canola oil

emulsion with no added solid fat (Golding et al., 2011). Also the extent of lipolysis was lower for

the canola oil emulsion with the added solid fat (Golding et al., 2011). These and most other

dispersed systems (Guo, Bellissimo, et al., 2017; Jiao et al., 2019) utilized different SFCs for

comparison and have generally shown that the presence of solid fat can slow the rate and/or the

extent in vitro lipolysis and FA bioaccessibility. Emulsion SFC has, in fact, been proposed as a

strategy to induce gastric structuring in the acidic environment by affecting coalescence of

droplets, with potential subsequent interference with lipolysis as the presence of solid fat can

lead to partial coalescence that result large solid agglomerates with very low surface area

(Golding & Wooster, 2010). The decrease in surface area can affect bile salt and pancreatic

lipase accessibility due to the solid interfacial area, limiting the lipolysis breakdown of the solid

agglomerates (Guo, Bellissimo, et al., 2017). While all of these works constitute important

advancements, they all rely on comparisons of physical state or SFC based on samples with

different underlying compositions. Samples also differed in terms of parameters such as

morphology, charge and size, i.e. other particle properties that can alter digestive processes and

may therefore have confounded the effects of physical state. It is very difficult to specifically

isolate the role of SFC. As such, minimizing possible confounding influences was one of the

aims of this thesis.

To better isolate physical property differences, Bonnaire and colleagues (Bonnaire et al., 2008),

novely proposed and used dispersed systems with different physical states that had similar

composition using the undercooling method. When a 10% tripalmitin (PPP) - 0.9% sodium

dodecyl sulphate (SDS) containing dispersed system was hot homogenized at a temperature

above its melting temperature, cooled below its crystallization temperature and held at 37 C

16

(below its melting temperature of ~60 C), the formed crystallized emulsion droplets remained

solid. When a portion of hot homogenized emulsion was alternatively undercooled from 70 C to

37 C, it retained the liquid state of the droplets without being crystallized. Using this so-called

undercooling method, for the first time, the researchers were able to compare compositionally

identical emulsions with different physical states: solid and liquid emulsion droplets, in terms of

in vitro digestibility. In agreement with previous study findings based on differences in SFC, the

findings from this novel approach showed slower lipolysis kinetics for the emulsion with solid

particles compared to the liquid particles (Bonnaire et al., 2008). Despite being a landmark

publication, there were some limitations worth noting. Firstly, possible confounding influences

from droplet shape or morphology were not accounted for. Further, the digestion model included

only a 2 h duodenal stage, omitting simulation of gastric conditions. Therefore, our group

extended that work with the same system (Huynh & Wright, 2018), focussing on these gaps.

Accordingly, the solid and the liquid emulsion droplets were found to have different

morphologies, i.e. the liquid droplets were spherically shaped and solid droplets were rod

shaped. As such, the observed differences in lipolysis (faster for liquid droplets, in agreement

with Bonnaire et al., 2008) could not be solely attributed to the physical state. Further, after

exposing the emulsion systems to a more physiologically relevant digestion model which

included a simulated gastric phase, the liquid emulsion system was observed to partially

crystallize, highlighting the impact of gastric phase modifications on duodenal lipolysis.

These findings (Bonnaire et al., 2008, Huynh and Wright, 2018) with compositionally equivalent

systems need to be extended to other TAG systems and validated using in vivo systems. This

requires working with emulsions with no confounding influences from other particle

characteristics that are produced using food permissible ingredients and at levels safe for human

17

consumption. To address this gap, we (Hart et al., 2018) developed an undercooled emulsion

system with 15% cocoa butter with solid and the liquid emulsion droplets with the same

spherical morphology and particle size and charge. Again, the solid droplets had a lower

duodenal lipolysis following exposure to simulated gastric conditions. Moreover, these findings

confirmed previous reports that, although similar extents of lipolysis were ultimately achieved

for both the solid and liquid states, the lipolysis kinetics differed. A major limitation, however,

was that investigations had to be carried out at 25 C (not body temperature 37 C) in order to

maintain the desired physical states of the cocoa butter droplets. As such, the gap of a food-grade

system to compare solid versus liquid emulsions in vivo response remained. This thesis addresses

that gap.

Although undercooled emulsion systems are not without limitations, they offer a valuable but

rarely used (Bonnaire et al., 2008; Hart et al., 2018; Huynh & Wright, 2018) experimental

approach. These dispersed lipid systems have the potential to be developed without confounding

influences from other particle characteristics like the size, charge or morphology as well, if the

lipid substrate and emulsifier/s and temperature are carefully selected. Theoretically, such a

system would enable to study solely the impact of lipid physical state on lipid digestion. To this

date, such a food grade lipid system has not been studied under in vitro or in vivo conditions. As

such, this thesis aimed to build on the limited previous work using undercooled emulsion

droplets to relate TAG physical properties with digestibility and absorption. This was done

initially using in vitro digestion methods (Chapter 4) and subsequently in a randomized human

postprandial lipemia study (Chapter 5).

18

1.6 Summary

In summary, lipid digestion, absorption, and metabolism are important physiological processes

that impact health and disease management. However, there is still limited knowledge on many

aspects of lipid digestion. Among numerous factors that determine the digestibility and

metabolic fate of lipids, very limited focus has been given to the potential impact of lipid

physical state, i.e. SFC. However, there is a growing body of research evidence pointing to

differences in digestion of liquid vs solid dietary lipids that is possible by using different

research tools. More knowledge development in this area will enable careful manipulation of the

physicochemical properties of dietary lipids as a promising approach to formulate dietary lipids

with desired nutritional qualities and may help to explain some of the apparent discrepancies and

reports about the role of dietary lipids in health and disease.

19

CHAPTER 2

RATIONALE, OBJECTIVES, AND HYPOTHESES

20

2.1 Rationale

There has been renewed interest in understanding the possible effects of lipid physical properties

on their digestion and metabolism, as it might help to explain apparent discrepancies about the

roles of lipids in health and disease and enable the development of products that capitalize on

any effects. It is clear that properties beyond lipid composition can contribute significantly to

metabolism although this is an emerging area and there are still many gaps that warrant further

investigation.

IE lipids are commercially relevant systems requiring additional efforts to clarify their metabolic

and health consequences (Mensink et al., 2016). Whether or not SFC-induced changes might

alter digestibility and subsequently the metabolism of these lipids is unknown. Moreover, as a

model with the same FA content, but differing TAG species, work with IE lipids can contribute

to answering questions about if and how TAG physical property differences influence digestive

processes and if these have consequences for absorption. Many food lipids exist in emulsion

format and there has been much effort made to clarify the role emulsion properties (i.e. droplet

size, charge, interfacial properties, etc.) in determining digestion, but very little addressing the

role of solid fat content. Given the possibility to temper dispersed lipids in the undercooled state,

these offer the possibility for investigations with compositionally similar dispersed systems,

although very little work has been done. That which has been done involves confounding

influences from other particle properties and only with in vitro methods. As such, there is a need

for studies with undercooled emulsions with equivalent particle properties, with careful

application of in vitro methods that support a mechanistic understanding and a human study to

support the proof of concept that compositionally different emulsions containing different solid

fat contents may have differences in terms of postprandial lipemia.

21

2.2 Objectives and Hypotheses

Overall, this thesis aims to investigate the role of TAG physical state (i.e. liquid versus solid,

SFC) on digestion and postprandial metabolism, building on work in two areas with relevance to

the food supply, i.e. IE lipids and tempered emulsion droplets and using both in vitro and in vivo

experimental methods. It was hypothesized that the presence of more solid fat would generally

reduce the lipid digestibility resulting in a slower and/or lower release of FFAs, compared to

liquid oil or those with less solid fat. The overall aim was addressed through three lines of

investigation, corresponding to Chapters 3-5. The specific objectives of each study/chapter and

the associated hypotheses are as below.

2.2.1 Specific objective 1 and hypothesis (for Chapter 3)

To investigate the in vitro lipolysis of lipids with equivalent compositions, but different SFCs at

37 C, i.e. a stearic acid-rich non-IE lipid blend and its IE versions and to compare the findings

with a previous human postprandial lipemic study of the same lipid blends in order to evidence

the role of SFC changes in digestibility, to support a mechanistic understanding, and to explore

in vitro-in vivo correlations. It was hypothesized that the non- IE lipid blend with a higher solid

fat content would show a slower FFA release and a lower FFA bioaccessibility compared to the

IE lipid blends and positively correlate with the human study findings.

2.2.2 Specific objective 2 and hypothesis (for Chapter 4)

To investigate the in vitro digestibility of emulsions with identical compositions and similar

particle characteristics, but different SFCs, i.e. undercooled liquid and crystalline solid emulsion

droplets. Specifically, the intent was for these emulsions to be formulated using food grade

ingredients and differing only in SFC in order to improve on previous work in the area where

22

differences in other particle properties were also observed. It was hypothesized that the

digestibility of the undercooled liquid emulsion droplets would be higher and would result in

faster in vitro lipolysis kinetics compared to the crystalline solid emulsion droplets.

2.2.3 Specific objective 3 and hypothesis (For Chapter 5)

To investigate the same undercooled liquid and crystalline solid droplets as in the previous

chapter (i.e. compositionally identical with no confounding influences from other particle

characteristics) in a postprandial lipemic response human study. It was hypothesized that the

undercooled liquid emulsion droplets would be associated with an earlier rise in postprandial

plasma TAG increase compared to the crystalline solid emulsion droplets.

23

CHAPTER 3

INVESTIGATIONS OF IN VITRO BIOACCESSIBILITY FROM INTERESTERIFIED

STEARIC AND OLEIC ACID-RICH BLENDS

A very similar version of this chapter has been published in Food and Function.

Thilakarathna, S.H., Rogers, M., Lan, Y., Huynh, S., Marangoni, A.G., Robinson, L.E., and

Wright, A.J. (2016) Investigations of in vitro bioaccessibility from interesterified stearic and

oleic acid-rich blends.’, Food and Function. Royal Society of Chemistry, 7(4), pp. 1932–40. doi:

10.1039/c5fo01272d.

Co-authorship statement

This study was designed by S.H. Thilakarathna and A.J. Wright. Y. Lan operated the TIM-1

model for each digestion run under the supervision of M. Rogers in whose research facility the

TIM-1 was housed. S. Huynh assisted with TIM-1 digestion runs and free fatty acid analysis of

the samples on site. S.H. Thilakarathna performed all sample extraction and analysis (gas

chromatography, thin layer chromatography, melting profile characterization), data analysis and

manuscript writing. A.G. Marangoni collaborated through providing differential scanning

calorimetry use and manuscript review. L.E. Robinson and A.J. Wright reviewed the manuscript.

24

3.1 Abstract

Interesterification was previously found to impact stearic acid absorption in a randomized cross-

over study, when human volunteers consumed a 70:30 wt% high-oleic sunflower and canola

stearin blend (NIE) compared to the same blend which had undergone either chemical (CIE) or

enzymatic (EIE) interesterification. In this research, in vitro lipid digestion, bioaccessibility, and

changes in undigested lipid composition and melting behavior of these same test fats were

investigated using the dynamic, multi-compartmental TIM-1 digestion model and compared with

the previous human study. Overall, TIM-1 bioaccessibility was higher with interesterification

(p<0.05). Oleic acid bioaccessibility was also higher than stearic acid bioaccessibility for NIE,

and vice versa for the IE blends (p<0.05). Stearic acid was more concentrated in the undigested

triacylglycerols (TAG) from NIE, corresponding to a relatively higher melting temperature of the

undigested lipids. The results confirm the impact of TAG composition, fatty acid position and/or

physical properties on lipid digestion. TIM-1 bioaccessibility was linearly correlated

(R2=0.8640) with postprandial serum TAG concentration in the human study. Therefore, the in

vitro digestion model offered predictive insights related to the impacts of lipid interesterification

on absorption.

25

3.2 Introduction

Interesterification re-arranges fatty acid (FA) within and between triacylglycerol (TAG)

molecules using either a chemical or enzymatic catalyst (Marangoni & Rousseau, 1995). As a

result, it can be used to achieve lipids with desired melting ranges and physical properties for

various food applications (Lee, Akoh, & Lee, 2008). IE lipids are of particular interest because

of the ability to use saturated FA-based hard stocks to achieve solid consistency and plasticity

without the need for trans FA from partially hydrogenated sources. Interesterification can also be

used to produce lipids with desired nutritional functionalities, such as for infant formulas or

enteral and parenteral products (Osborn & Akoh, 2002). According to human infant and animal

studies, changes in FA distribution, especially long chain saturated FA, can impact lipid

absorption based on the selectivity of pancreatic TAG lipase for positions sn-1 and 3 (Berry,

2009; Karupaiah & Sundram, 2007). This activity leaves intact 2-monoacylglycerols (Reis,

Holmberg, & Watzke, 2009), with the constituent long chain saturated FA having a higher

tendency for absorption (Berry, 2009; Berry & Sanders, 2005), For example, in infants and

animal studies, stearic and palmitic acids were found to be better absorbed when located at the

sn-2 position (Berry, 2009). In contrast, adult human studies have indicated efficient saturated

FA absorption, regardless of FA positional distribution (Hunter, 2001; Karupaiah & Sundram,

2007). Positional distribution has also been related to delayed chylomicron clearance of saturated

FA, a scenario that may lead to postprandial lipemia and increased cardiovascular risk (Berry,

2009). Therefore, further knowledge about the metabolic fate of structured lipids, based on

saturated FA, is required.

We previously investigated differences in serum TAG and FA response in an acute human study

(Robinson et al., 2009), where participants consumed stearic-rich IE fats either as a blend (NIE)

26

or after the blend was chemically (CIE) or enzymatically (EIE) interesterified. There was no

significant effect of interesterification on postprandial serum TAG response in the non-obese

individuals (BMI< 30 and/or waist circumference < 102 cm, n=10), although differences in

postprandial serum FA composition indicated increases in oleic and stearic acid concentrations

with CIE and EIE versus NIE. These differences in FA composition were also observed in the

obese group of participants (BMI> 30 and/or waist circumference > 102 cm, n=11), along with

significantly higher values of TAG AUC for CIE versus NIE. Only serum oleic acid increased

following consumption of the NIE blend by both groups. It was speculated that the differences in

physical properties, resulting from interesterification, contributed to the differences in

postprandial FA composition between NIE and IE fats. Indeed, the importance of physical

properties on lipid digestibility and absorption has long been suggested (Livesey, 2000), but

remains equivocal, despite recent interest (Berry & Sanders, 2005; Bonnaire et al., 2008;

Michalski et al., 2013; Wang et al., 2015), To advance knowledge about the health effects of

dietary lipids, a better understanding of the interplay between FA positional distribution, TAG

physical properties, digestibility, and absorption is necessary.

In vitro digestion studies have become popular for investigating dynamics occurring in the

gastrointestinal tract and for investigating the potential health implications of food composition

and structure (Bonnaire et al., 2008; Nik, Wright, & Corredig, 2011; Speranza et al., 2013).

Although not without limitations, the approach avoids some of the ethical, safety, and logistical

issues associated with collecting samples from human participants (Armand et al., 1999; Kalantzi

et al., 2006; Normén et al., 2006; Robertson & Mathers, 2000). Among the most advanced in

vitro systems is the dynamic, multi compartmental, TIM-1 (TNO, Zeist, The Netherlands). The

TIM-1 consists of stomach, duodenal, jejunal, and ileal compartments, all maintained at 37 ºC,

27

and can be coupled with the TIM-2 unit which simulates dynamics in the large intestine

(Venema, Nuenen, Ellen G. van den Heuvel, Pool, & Vossen, 2003), It is computer controlled to

maintain predetermined volumes and concentrations of digestive fluids and pH in each of glass

compartments which consists of inner flexible silicone jackets which are contracted and relaxed

using external pressure changes in order to mimic peristaltic movements (Minekus et al., 1995;

Ribnicky et al., 2014). To simulate intestinal absorption, the jejunal and ileal digestates are

passed over semi-permeable 50 nm pore size membranes (Spectrum Milikros modules M80S-

300-01P) through which the so-called bioaccessible molecules pass (Minekus et al., 1995;

Ribnicky et al., 2014). The ileal efflux therefore contains unabsorbed meal components which, in

vivo, would be passed to the colon (Ribnicky et al., 2014). The ability to collect samples for

analysis by a variety of techniques is an advantage of utilizing in vitro models like TIM-1 to gain

a mechanistic understanding of digestive events. The TIM-1 system was developed based on in

vivo data (Guerra et al., 2012; Minekus et al., 2005) and good agreement has been reported with

the results of human studies (Gervais et al., 2009; Minekus et al., 2005).

28

Figure 3.1 A schematic of the TNO intestinal model (TIM)-1. (1) food inlet, (2) jejunum

filtrate, (3) ileum filtrate, (4) ileal colorectal valve, (5a) gastric compartment, (5b) duodenal

compart- ment, (5c) jejunum compartment, (5d) ileal compartment, (6a) hollow fiber membrane

from jejunum, (6b) hollow fiber membrane from ileum, (7a and 7b) secretion pumps (Adapted

from (Speranza et al., 2013).

The purpose of this study was to investigate the in vitro lipid digestibility and FA

bioaccessibility of IE (CIE and EIE) and NIE stearic and oleic acid-rich lipids using the TIM-1 in

vitro digestion model and to compare the results with the observed differences in postprandial

blood lipids from our previous human study (Robinson et al., 2009) .

3.3 Methodology

3.3.1 Materials

Pancrex V powder (lipase activity = 25 000 units/g, protease activity = 1400 units/g, and amylase

activity = 30 000 units/g) and Rhizopus lipase (150 000 units/mg F-AP-15) were obtained from

Paines & Byrne Ltd. (Surrey, UK) and Amano Enzyme Inc. (Nagoya, Japan), respectively. Fresh

pig bile (Farm to Fork, Warren, NJ, USA) was collected, pooled and aliquoted into single

portions and stored at -20 °C until use (Speranza et al., 2013a). Trypsin from bovine pancreas

(7500 N-α-benzoyl-L-arginine ethyl ester (BAEE) units/mg, T9201), pepsin from porcine gastric

mucosa (2500 units/mg protein, P7012), α-amylase type II-A from Bacillus species (≥1500

units/mg protein A-6380) and all other chemicals utilized with the TIM-1 were purchased from

Sigma Aldrich (St. Louis, MO, USA). High oleic sunflower oil (76% oleic acid) was purchased

from Vegetol 80RBD, Acatris, Minneapolis, MN, USA, and fully hydrogenated canola stearin

(88% stearic acid) was purchased from CanAmera, Oakville, ON, Canada. Sodium methoxide

29

and Candida antarctica lipase (Novozym 435, immobilized on polyacrylic resin) were purchased

from Sigma, St Louis, MO, USA and, Novozymes Biopharma US Inc., NC, USA, respectively.

The non-esterified FA analysis kits (NEFA-HR2) were purchased from Wako Pure Chemical

Industries (Wako Diagnostics, VA, USA).

3.3.2 The test fats

NIE, CIE, and EIE were prepared as previously described (Robinson et al., 2009). Briefly, NIE

was a 70:30 wt% blend of high oleic sunflower oil and fully hydrogenated canola stearin. This

NIE blend was then either chemically interesterified (CIE) using sodium methoxide (0.3 wt%) or

enzymatically interesterified (EIE) using Candida antarctica lipase at 5 wt%.

3.3.3 LC/MS regiospecific analysis of the test fats

The regiospecificity of CIE and EIE TAG were analyzed by liquid chromatography/mass

spectrometry (LC/MS). Analyses were done using a Shimadzu LC/MS system consisting of a

Nexera UHPLC coupled to an 8030 triple quadrupole mass spectrometer. The mobile phase was

pumped at a gradient of 1 mL/min through three Agilent ChromSpher 5 lipids columns (5 μm,

4.6 × 250 mm) connected in series and maintained at 30 °C. The mobile phase, where solvent A

was heptane-2-propanol-acetonitrile (99.8:0.1:0.1) and solvent B was heptane-2-propanol-

acetonitrile (99:1:1), was held for 15 min at 5% B, then ramped to 20% B over 55 min, followed

by a change to 72%B over 50 min. The mobile phase was returned to the original composition at

135 min and held there for 60 min before the next injection. The column eluate was directed

unsplit into the mass spectrometer’s APCI source. Positive ion mass spectra were recorded as Q3

scans from m/z 450 to 1050 at a scan rate of 1363 μ/sec. Argon was used as the collision gas at

230 kPa. The APCI source temperature, heat block, and desolvation line were maintained at 325,

30

200, and 250 °C, respectively, and nitrogen was used as the nebulizing and drying gas, at flow

rates of 3 and 5 L/min, respectively. Samples were 5 mg/mL in mobile phase solvent A and the

injection volume was 0.4 μL.

3.3.4 Meal preparation

Meals consisting of NIE, CIE, and EIE for the TIM-1 in vitro digestions were prepared to

resemble those consumed by the participants in our human study (Robinson et al., 2009). The

non-obese participants had an average body weight of 86.2 ± 2.4 kg (BMI< 30 kg/m2 and/or

waist circumference < 102 cm) and consumed three slices of toasted white bread (120 g,

Wonder™, 285 calories, 3 g fat (0.6 g saturated), 1.5 g fibre, and 10.5 g protein), 500 g of water

and 1g test fat/kg body mass. Meal proportions for the TIM-1 experiments were calculated

accordingly, such that the 100 g meal consisted of 17.0, 12.2, and 70.8 g of toast, fat and water,

respectively, and the TIM-1 parameters were programmed to be representative of the

gastrointestinal environment of a healthy adult under fed state conditions. As per Robinson et al.

(Robinson et al., 2009) the day before a TIM-1 digestion, each fat was melted to erase the crystal

memory and stored overnight at 4 °C. White bread (Wonder™, Weston Bakeries, Toronto, ON;

Ingredients: enriched white wheat flour, water, sugar/glucose-fructose, yeast, vegetable oil (soya

bean and/or canola), salt, defatted soya flour, calcium propionate, stearoyl-2-lactylate,

monoacylglycerols) was toasted, cooled, and the test fat spread as per the above ratio. Toast with

fat was then combined with room temperature water in a Magic Bullet™ blender (Homeland

Housewares®, Los Angeles, CA, USA) and pulsed 20 times to mimic the mastication process and

enable loading into TIM-1.

3.3.5 TIM-1 in vitro digestions

31

Addition information about the TIM-1 methodology used, including a schematic of the system,

can be found in Speranza et al. (Speranza et al., 2013a). A small intestinal electrolyte solution

(SIES: 85.6 mM NaCl, 8.01 mM KCl, 2.25 mM CaCl2) and a 7% pancreatin solution (from

Pancrex V powder) were prepared. Prior to introducing the meal into the TIM-1, duodenal start

residue (15 g SIES, 30 g fresh porcine bile, 2 mg trypsin solution, 15 g of 7% pancreatin

solution), jejunal start residue (40 g SIES, 80 g fresh porcine bile, 40 g of 7% pancreatin

solution), and ileal start residue (160 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 of gastric enzyme solution (600 U/mL pepsin and 40 U/mL

Rhizopus lipase in gastric electrolyte solution: 4.8 g/L NaCl, 2.2 g/L KCl, 0.22 g/L CaCl2,). 100

g of the meal mixture, prepared as above, was weighed and combined with 95 g of gastric

electrolyte solution (without the gastric enzymes). After adding 11 mg of amylase and 50 g of

water to this meal mixture, the pH was adjusted to 5.5 (by drop-wise addition of 0.1M HCl).

After one minute, this pH-adjusted meal mixture was then immediately introduced to the TIM-1

gastric compartment, along with 50 g of water-rinse. The experiment was carried out using

conditions representative of the fed state and TIM-1 was operated as previously described

(Speranza et al., 2013a). A preprogrammed computer protocol controlled the secretion of 1M

HCl to the gastric compartment and regulated gastric emptying, intestinal transit times, and pH.

Fluid secretions had the following flow rates; duodenal secretions: fresh porcine bile at 0.5

mL/min, 7% pancreatin solution at 0.25 mL/min, and SIES at 3.2 mL/min; jejunal secretion:

SIES and 10% fresh porcine bile at 3.2 mL/min; ileal secretion: SIES at 3.0 mL/min. The

duodenal, jejunal, and ileal compartments were maintained at pH 6.5, 6.8, and 7.2, respectively,

by controlled secretion of a 1 M sodium bicarbonate solution.

32

3.3.6 Sample collection

The jejunal and ileal dialysates (containing the bioaccessible components) and ileal efflux

(reflecting the undigested and/or unabsorbed contents passing to colon) were collected at 30, 60,

90, 120, 180, 240, 300, and 360 min from the time of introducing the meal to the TIM-1. The

collected dialysates were cooled on ice, weighed and stored at -20 °C for subsequent analysis.

Collected ileal efflux samples for each time point were pooled and chilled on ice. Each test fat

was run in duplicate. The dialysates were analyzed for total bioaccessible FFA and the

bioaccessibility of each major FA, as described below. The ileal efflux was analyzed for FA

composition of the separated TAG and FFA fractions and melting behaviour of the extracted

lipids was investigated.

3.3.7 Total FFA measurements

The total FFA concentration was determined for each bioaccessible (i.e. dialyzed) lipid fraction

at each time point. Samples were extracted into acidic hexane and FFA quantified using the

commercial NEFA enzymatic kit with spectrophotometric analysis at 550 nm (Spectramax plus,

Molecular Devices Corporation, CA, USA) (Eldemnawy, Wright, & Corredig, 2015).

3.3.8 FA composition of bioaccessible lipids

The FA composition of the bioaccessible TIM-1 jejunal and ileal lipids (i.e. combined lipolysis

and absorption) were determined by GC, as previously described, using C17:0 as an internal

standard (Anderson, MacLennan, Hillyer, & Ma, 2014). Briefly, total lipids were extracted by

the Folch method (Folch, Lees, & Stanley, 1957) with transesterification using 14% boron

trifluoride in methanol (Sigma Aldrich, St. Louis, MO, USA). Methyl esters were re-suspended

in hexane and analyzed using an Agilent 7890AGC equipped with flame ionization detection

33

(Agilent Technologies Inc., DE, USA). For comparison with the Robinson et al. human study

(Robinson et al., 2009), relative concentrations of palmitic, stearic, oleic, and linoleic acids are

reported.

3.3.9 FA composition of ileal efflux lipids

Lipids were extracted from thawed (at 4 °C) ileal efflux samples (Bligh & Dyer, 1959).

Extracted lipids were saponified, and resuspended in 100 μL hexane and spotted on thin layer

chromatography (TLC) plates (Anderson et al., 2014) using a solvent mixture of 80:20:1

petroleum ether:ethyl ether:glacial acetic acid. After visualization under UV light and in

comparison with standards, bands for TAG and FFA were identified, scraped, methylated and the

lipids were analyzed after re-suspending in hexane. There were no mono- or diacylglycerol

bands visible in the efflux samples (data not shown), suggesting these species were digested

and/or bioaccessible. The background biliary contribution was assumed similar across all

treatments by maintaining consistent conditions. Relative concentrations of palmitic, stearic,

oleic, and linoleic acids were determined, as above.

3.3.10 Melting behavior by differential scanning calorimetry (DSC)

Lipids were extracted (Bligh & Dyer, 1959) from the ileal efflux samples and 5-10 mg were

weighed and sealed in alodined aluminum DSC pans. Samples of the extracted lipids and NIE,

EIE, and CIE were melted at 80 °C for 30 min (in the sealed pans) and crystallized overnight at 4

°C. Melting and re-crystallization behaviors were analyzed using a Q2000 model (TA

Instruments, Mississauga, ON, Canada) and peak temperatures determined using the system

software (TA Instruments Universal Analysis 2000 software, TA Instruments, ON, Canada). The

34

experimental protocol consisted of an initial temperature of 5 °C followed by heating at 5 °C/min

to 85 °C, isothermal conditions for 3 min and cooling at 5 °C/min to 5 °C.

3.3.11 Statistical analysis

Two-way analysis of variance (ANOVA) was used to compare lipolysis and FA composition

between samples and Tukey’s test was performed for multiple comparisons testing using

GraphPad Prism (v 6.0e, GraphPad Software, San Diego, CA). Normality of the data was

confirmed before each analysis using D’Agostino and Pearson Omnibus normality testing.

Pearson correlation analysis was used to compare the TIM-1 in vitro bioaccessibility (%) with

serum TAG concentration (mmol/L) obtained from the non-obese group in the human study

(Robinson et al., 2009). Data is reported as mean ± SEM and p < 0.05 was considered

statistically significant.

3.4 Results and Discussion

3.4.1 TIM-1 digestion: lipid bioaccessibility

The FA and TAG compositions of NIE, CIE, and EIE are shown in Table 3.1. Importantly, they

are consistent with the composition reported in the human study (Robinson et al., 2009) and

show that the FA composition of the three test fats was similar, although they varied

considerably in terms of TAG species.

35

Table 3.1 Fatty acid and triacylglycerol compositions of NIE, CIE, and EIE.

NIE CIE EIE

Fatty acid species (wt%)

16:0 5.1 ± 0.1 5.0 ± 0.0 5.1 ± 0.0

18:0 27.9 ± 0.0 30.0 ± 0.0 29.1 ± 0.0

18:1n-9 58.9 ± 0.1 56.4 ± 0.1 57.3 ± 0.1

18:2n-6 4.6 ± 0.0 4.7 ± 0.0 4.7 ± 0.0

96.5 ± 0.0 96.2 ± 0.2 96.2 ± 0.0

Triacylglycerol species (wt%)a NIE CIE EIE

OOS 4.2 ± 0.1 34.8 ± 2.2 b 34.1 ± 0.7 b

OOO 36.7 ± 0.9 25.3 ± 3.3 26.4 ± 0.3

SOS nd 14.0 ± 0.8 b 11.2 ± 0.5 b

OOL 21.4 ± 0.4 6.8 ± 3.2 7.5 ± 0.2

PLO 2.3 ± 0.0 3.9 ± 0.3 4.1 ± 0.3

OPO 3.7 ± 0.1 2.3 ± 0.0 3.1 ± 0.7

PPP 2.4 ± 0.2 2.7 ± 0.2 2.4 ± 0.3

POS nd 2.8 ± 0.3 1.7 ± 0.5

PPS nd 1.3 ± 0.1 1.1 ± 0.4

SSP 2.2 ± 0.1 0.5 ± 0.4 0.2 ± 0.1

SSS 23.8 ± 0.4 0.8 ± 0.3 0.6 ± 0.2

Total 96.4 ± 3.8 95.0 ± 3.4 92.2 ± 3.4

P: palmitic, S: stearic, O: oleic acid, L: Linoleic. nd: not detected.

Figure 3.2 shows the TIM-1 lipid bioaccessibility obtained for NIE, CIE, and EIE based on the

determination of FA concentrations of the jejunal and ileal dialysate samples throughout the 6 h

digestion protocol. There was a trend towards higher bioaccessibility for the IE versus NIE fats.

Statistically significant differences were observed between NIE and EIE (but not CIE) starting at

36

120 min. Thereafter, NIE had significantly lower lipid bioaccessibility than the IE fats (p<0.05).

Ultimately, the bioaccessibility achieved at 6 h was also higher for the IE blends (p<0.05)

compared with the NIE sample, but the two esterified blends did not differ (p>0.05). Lipids rich

in SSS are notoriously waxy and difficult to work with. The NIE sample was visibly less evenly

distributed throughout the TIM-1 compartments than the IE samples. Notably, at 240 minutes

and beyond, the variability in the NIE, TIM-1 bioaccessibility data was also much higher than

for the IE samples (Figure 3.2). This may correlate with the high melting temperature of SSS (i.e.

73.1 °C (Small, 1991), present in NIE and the fact that TIM-1 has a gastric emptying half-life of

70 min. (Minekus et al., 2005) Therefore, the physical state of NIE visibly impacted TIM-1

digestion dynamics.

Figure 3.2 Bioaccessibility (%) of NIE, CIE, and EIE during 6 h in vitro TIM-1 digestion (n=2).

Data reported as mean ± SEM. #Significant treatment differences (p<0.05) observed between

NIE and EIE. *Significant treatment differences (p<0.05) observed between NIE and both IE

blends. Overall ANOVA p=0.013 for time × test fat interaction.

0 100 200 300 4000

20

40

60

Time (min)

Bio

acce

ssib

ility (

%)

NIE CIE EIE

*

**

*

#

37

Lipid digestion is generally considered to be high (~95 %) under healthy physiological

conditions (Mu & Høy, 2004). That said, the issue of stearic acid digestibility and absorption is a

matter of long and considerable debate (Livesey, 2000). It was previously reported that the

impact of stearic acid on plasma lipids, red blood cells and platelet FA composition occurs at a

slower rate compared to short chain saturated FA (Dougherty, Aliman, & Iacono, 1995). In

contrast, a higher bioaccessibility coefficient was observed for stearic acid compared to oleic

acid in a previous TIM-1 study involving conjugated linoleic acid-rich milk and milk emulsions

(Gervais et al., 2009). SSS digestibility, specifically, is often regarded as being low (Michalski et

al., 2013), although estimates differ and the evidence for this is largely based on animal models.

In one rat study, stearic acid digestibility from unblended SSS was reportedly 0.15 g/g stearic

acid (Livesey, 2000). In another rat study, lipid absorption of SSS, OOO and safflower oil was

73, 98 and 97%, respectively (Feldman et al., 1979), providing evidence of lower absorption of

the highly saturated FA. The issue of and stearic acid digestibility and absorption is additionally

complicated by the fact that it can depend on dose as well as method of sample preparation

(Livesey, 2000), including factors such as high calcium levels in feed which may contribute to

soap formation. For example, as measured in lymph output in rats, SSS digestibility was

improved by blending with OOO (Karupaiah & Sundram, 2007). This may occur through

disruption of the crystalline networks (Livesey, 2000) and potential formation of eutectics.

Lipid bioaccessibility was related to TAG species present and melting behaviour. It is

hypothesized that SSS crystallinity, at body temperature, limits lipase accessibility in part,

because bile salt emulsification of solid fats is limited (Livesey, 2000). Figure 3.3a shows the

presence of a single melting peak for NIE at 62.0 ± 0.0 °C and a range of melting events for the

IE fats. In CIE, melting events were observed at 15.0 ± 1.4, 32.0 ± 0.0, and 43.0 ± 1.4 °C. In

38

EIE, melting events were observed at 2.0 ± 2.8 and 26.0 ± 2.8 °C. According to Table 3.1,

although all fats contained approximately 30% stearic acid, the IE blends contained a much more

diverse group of TAG species than NIE, with the predominant TAG species (~ 40%) being OOS

and OSO. These TAG have melting temperatures of 24.0 and 24.2 °C (Small, 1991) i.e. below

37 °C. Furthermore, NIE had a solid fat content of 18.6% at 37 °C while the IE samples

contained only 5.4 - 5.6% solid fat at this temperature (Robinson et al., 2009). Therefore, the

presence of ~ 25% SSS contributed to a much higher solids content for NIE versus the IE blends

and these differences in physical properties were associated with differences in in vitro digestion

handling and observed lower lipid bioaccessibility from NIE. It has been speculated that physical

properties of saturated FA-rich blends, including some IE samples, may contribute to attenuated

postprandial lipemia (Karupaiah & Sundram, 2007; Sundram, Karupaiah, & Hayes, 2007). For

example, Berry and Sanders (Berry & Sanders, 2005) speculated that shea butter, a native fat

with high proportion of stearic acid in the sn-1/3 positions resulted in a lower lipemic response

compared to high oleic sunflower oil partly because of associated differences in physical

properties (Berry & Sanders, 2005). In our previous human study (Robinson et al., 2009), no

significant differences were observed between values of incremental TAG area under the curve

(57.0 ± 44.4, 83.3 ± 83.8, and 116.6 ± 80.9 mmol/L*6h for NIE, CIE, and EIE, respectively;

p>0.05) for non-obese individuals (Robinson et al., 2009). Of note, in the obese study

participants, the TAG AUC values were significantly higher for the CIE (242.4 ± 145.2

mmol/L*6h) versus NIE (105.8 ± 116.8 mmol/L*6h), but not for NIE versus EIE (193.2 ± 130.1

mmol/L*6h) or CIE versus EIE (Robinson et al., 2009). The overall tendency was for lower

lipemia after NIE versus the IE fats. It is possible the small sample size and duration of the

human study (6 h) may have precluded the ability to detect significant differences among the test

39

fat treatments. Notably, there are reports that plasma TAG concentrations can remain elevated

beyond 6 h postprandial, particularly in the case of higher melting fats (Sanders et al., 2011;

Vors et al., 2013). The difference among the obese and non-obese groups could relate to

differences in post-absorptive lipid handling, given that insulin resistance can lead to elevated

postprandial TAG related to competition for hepatic receptors between chylomicrons and very

low density lipoproteins that impede the uptake of chylomicron remnants (Kolovou et al., 2005).

Regardless, the in vitro experiments support that the physical properties of the NIE and IE blends

impact lipemia, suggesting a role for the physical properties in impacting digestive processing

and lipemia. Wang et al. (Wang et al., 2015) recently attempted to clarify the relative

contributions of TAG structure and solid fat content on lipid metabolism in mice fed IE lipids

rich in palmitic and oleic acids for six weeks. Solid fat content, at the levels examined (i.e. 0-

21.9 % at 37 ºC), did not impact lipid metabolism, although TAG positional distribution had an

impact on fasting non-esterified FA and glucose concentrations.

Figure 3.3 DSC thermograms of (a) NIE, CIE, and EIE fats and (b) of extracted lipids from the

NIE, CIE, and EIE ileal efflux from the TIM-1 digestion simulator (n=3).

0 20 40 60 80-0.8

-0.6

-0.4

-0.2

0.0

Temperature (°C)

Hea

t F

low

(W

/g)

NIE CIE EIE(a)

0 20 40 60 80-0.6

-0.4

-0.2

0.0

Temperature (°C)

(b)

40

Gas chromatographic FA analysis was performed on the bioaccessible lipids obtained from the

jejunal and ileal dialysates. These results are presented in Table 3.2. The analysis was focused on

palmitic, stearic, oleic, and linoleic acids, as these were the major FA present in the test fats and

were analyzed in the human study (Robinson et al., 2009). However, the discussion focuses on

stearic and oleic acids, given the in vitro biliary contributions of palmitic and linoleic acids

which are the main saturated FA and unsaturated FA, respectively in bile phospholipids (van

Berge Henegouwen, van der Werf, & Ruben, 1987). With CIE and EIE, the bioaccessibility of

stearic acid was significantly higher than for oleic acid. In contrast, the bioaccessible lipids from

NIE were significantly higher in oleic acid (p<0.05). These trends may be related to differences

in the stereospecific positioning of stearic acid between the test fats. Regiospecific TAG analysis

of EIE and CIE by LC/MS indicated ratios in the range of 2:1 for OOS : OSO and 1:2.5 for SOS

: SSO (wt%, data not shown). Therefore, combined with the data in Table 3.1, it is estimated that

25–27% of the IE sample TAG had stearic acid located at sn-2. Although there was similarly

around 25-29% stearic acid located at sn-2 in NIE, approximately 25.5% of this was in the form

of SSS and SSP. Therefore, there was limited digestibility and release of stearic acid from these

molecules which have relatively high melting temperatures (i.e. 73.1 and 65.2, respectively ºC

(Small, 1991). It is commonly accepted that FA located at the sn-2 position remain preferentially

esterified during digestion, enter the bile salt mixed micelles as part of 2-monoacylglycerols, and

are absorbed by the enterocytes (Mu & Høy, 2004). Animal and human infant studies (Berry,

2009; Karupaiah & Sundram, 2007) have indicated that stearic acid located at sn-1/3 can have

lower absorption because of the possible formation of insoluble FA soaps with calcium

(particularly with a calcium-rich diet (Mattson, Nolen, & Webb, 1979)) and release of insoluble

acylglycerols after they are hydrolyzed by pancreatic TAG lipase (Livesey, 2000). As reviewed

41

by Livesey (Livesey, 2000), in a rat study, stearic acid hydrolysis was higher for OSO and OOS

(99.2 and 96.4%, respectively) than from SOS and OSS (70.5 and 60.1%, respectively) due to

the release of poorly digestible acylglycerol intermediates. It was concluded that, under

physiologically relevant concentrations of calcium and magnesium ions, this factor may

contribute more to stearic acid digestibility than the formation of insoluble soaps. (Livesey,

2000) The results of the present study support that stearic acid positional distribution in IE lipids

favors higher overall lipid bioaccessibility, and specifically stearic acid bioaccessibility,

compared to NIE.

Table 3.2 Fatty acid (palmitic, stearic, oleic, and linoleic acids) concentration (cumulative,

wt%) of bioaccessible lipids following 6 h TIM-1 digestion of NIE, CIE, and EIE1

NIE CIE EIE

Palmitic acid (16:0) 29.5 ± 0.2a 29.1 ± 0.6a 32.9 ± 0.6b

Stearic acid (18:0) 18.0 ± 0.4a 20.5 ± 0.1b 20.0 ± 0.1b

Oleic acid (18:1n-9) 22.9 ± 0.7a 18.4 ± 0.6b 17.0 ± 0.8b

Linoleic acid (18:2n-6) 29.6 ± 0.3a 32.0 ± 0.1b 30.1 ± 0.2a

100.0 100.0 100.0

1 Mean ± SEM.

a,b Values within each row with different superscript letters differ significantly (p<0.05).

In our previous human study (Robinson et al., 2009), postprandial increases in stearic and oleic

acids were observed following consumption of EIE and CIE, but only serum oleic acid

concentration increased with NIE. The trends observed for stearic acid bioaccessibility with the

TIM-1 were similar, although oleic acid bioaccessibility was higher for NIE than EIE and CIE

42

(p<0.05). Oleic acid in NIE was predominantly in the form of OOO, followed by OOL. In

contrast, mixed TAG species of mono-, di-, and tri-olein were present in the IE blends (Table

3.1). Previously, the rate of TAG removal from TAG-rich lipoproteins in healthy

normolipidemic participants was reported to be faster for OOO, followed by OOL and OOS.

(Abia et al., 2001) Therefore, the observed higher bioaccessibility of oleic acid from NIE can be

attributed to its positional distribution.

3.4.2 TIM-1 digestion: unabsorbed lipids in the ileal efflux

Samples of the TIM-1 ileal efflux were analyzed to characterize components of the digestate,

which are expected to reach the colon (Figure 3.4). Lipids in the ileal efflux reflect components

that have been digested, but not absorbed as well as some undigested components. Therefore, in

Figure 3.4, the TAG constitutes undigested lipids, while the FFA reflect those FA, which were

hydrolyzed but not bioaccessible. The undigested TAG of the three test fats consisted

predominantly of stearic acid followed by palmitic acid. The undigested NIE TAG were also

significantly higher in stearic acid than the IE blends (i.e. 86.5 ± 0.3, 61.1 ± 2.8, and 44.3 ± 1.7%

for NIE, CIE, and EIE, respectively, p<0.05). Also, the undigested NIE TAG contained

significantly less oleic acid compared to the IE blends, further confirming the relatively higher

bioaccessibility of oleic acid from NIE. The amount of stearic acid in the NIE FFA fraction was

24 ± 0.4%. Collectively, this indicates that the predominant form of stearic acid excretion was in

the form of TAG, suggesting a high level of undigested lipids. When lipid excretion is high (i.e.

more than 7 g daily based on a 100 g fat diet), the condition known as steatorrhea occurs. Lipid

malabsorption can result in high levels of undigested TAG in the intestines (Ros, 2000). That

said, we postulate that fecal excretion of undigested saturate-rich lipid species may go undetected

based on their solid consistency. This is consistent with a rat feeding study which reported 6.6 ±

43

0.3, 16.8 ± 1.5, and 79.1 ± 1.0% fat present in the animals’ feces, with no observed signs of

gastrointestinal distress when rats were fed feed with liquid TAG, solid TAG or a solid alkane

emulsion, respectively (Wang et al., 2011).

Figure 3.4 Palmitic, stearic, oleic, and linoleic acid composition (wt %) of triacylglycerol (a)

and free fatty acid (b) fractions of lipids remaining in the 6 h pooled TIM-1 efflux from NIE,

CIE and EIE (mean ± SEM, n=2). Different letters indicate significant differences within each

FA between test fats in each graph (p<0.05).

Most stearic acid present in the CIE and EIE ileal efflux samples was present as TAG versus

FFA, indicating a propensity for lower lipolysis of the TAG species present. The constituent

SOS, SSO, SPO, and PSO in the IE blends have melting temperatures above physiological

temperature (i.e. 37 °C) and their solid state may limit pancreatic access (Berry, 2009). In the

CIE and EIE efflux, FFA consisted of 24-26% stearic acid, which may not have been absorbed

because it is known to form calcium soaps (Osborn & Akoh, 2002). Due to their higher melting

temperature, TAG species such as SOS and SSO may undergo less digestion due to their solid

state at 37 °C compared to molecules such as OSO and OOS that will be in the liquid state.

NIE CIE EIE0

25

50

75

100

Fat

ty a

cid c

once

ntr

atio

n (

%)

A

BB

a

bb

a

b

c

xxy

y

(a)

NIE CIE EIE

Oleic acid Linoleic acid

A

BC

a

bc

xy y

Palmitic acid Stearic acid(b)

44

Stearic acid positional distribution can also come into play for these lower melting TAG, in that

stearic acid at sn-2 (as in OSO) can be better absorbed from the sn-2 monoacylglycerol, which

forms during digestion compared to stearic acid at sn-1 or 3 that is released by pancreatic lipase,

but then forms a calcium soap. Previous reports also suggest that fats, mainly composed of long

chain FA, are associated with slower rates of gastric emptying in humans, compared to short or

medium chain FA (Mu & Porsgaard, 2005) enabling slower, and more efficient, fat digestion. A

limitation of in vitro digestion models is the inability to replicate such digestion dynamics.

Figure 3.3b shows that lipids extracted from the ileal efflux following digestion of NIE, EIE or

CIE have several melting events. Two peak melting events were observed in NIE at 33 ± 0.0 and

53 ± 1.4°C. Single, distinct melting peaks were observed for CIE and EIE at 48 ± 1.4 and 36 ±

0.0 °C, respectively. Therefore, lipid fractions from the NIE and CIE efflux were solid at body

temperature. In contrast, the EIE lipids from the efflux were in the liquid state at 37 °C. This may

relate to the fact that, despite similar TAG profiles, EIE contained more partial glycerides (i.e.

6.6 versus 2.0% DAG in EIE versus CI, respectively, and 0% in NIE) (Robinson et al., 2009).

Chemical interesterification is a random process and, since Candida antarctica lipase (Novozym

435, Novozymes Biopharma US Inc., NC, USA) is also non-specific, both processes are

expected to result in TAG species of similar composition. It would be interesting to investigate

the EIE ileal efflux fat further to understand the reasons for the low bioaccessibility of TAG with

relatively lower melting temperatures. However, the very low fat content (<0.2%) in the efflux

makes it challenging to extract sufficient lipids for further analysis. There also is an unavoidable

contribution from the biliary lipids that impact the melting properties of the ileal efflux fat,

although this applies similarly to all three fats tested.

45

3.4.3 Correlations between TIM-1 and human study results

When a correlation analysis was performed between the TIM-1 lipid bioaccessibility data and the

healthy non-obese participants’ serum postprandial TAG concentration at each corresponding

time point, a significant positive correlation (R2=0.8640; p< 0.05) was observed (Figure 3.5).

TIM-1 bioaccessibility was significantly different between IE and NIE after starting at 180 min

of in vitro digestion, and for both IE fats thereafter. Although not significantly different, in vivo,

serum TAG concentrations for the IE fats were higher than for NIE (p>0.05) during the 6 h

period. These observations suggest that IE fats may be absorbed more rapidly and warrant

further investigation. There was no correlation found between the serum TAG AUC values, i.e.

an indicator of cumulative absorption, and the TIM-1 total lipid bioaccessibility at 360 min

(R2=0.7588; p=0.3269) with these same healthy participants. This relates to the fact that

differences were observed between the fats in the TIM-1 study, but there were no significant

differences observed in vivo in terms of serum TAG AUC values between NIE, EIE and CIE

(Robinson et al., 2009). Of note, there was a trend towards higher blood lipids for CIE and EIE

versus NIE in the healthy study participants. These differences are consistent with TIM-1 results

showing lower digestibility and bioaccessibility of the NIE versus IE lipids. The efficiency of

human lipid metabolism, significant intra-individual variability and/or low participant numbers

might explain this apparent discrepancy. Support for this can also be found in the data from the

obese human study participants. Because the TIM-1 conditions utilized reflect those of healthy

individuals (Minekus et al., 1995), comparisons with that group of participants were prioritized.

However, it is interesting that, in the obese participants, serum TAG concentrations over time

and TAG AUC were higher for CIE versus NIE (Robinson et al., 2009). The TIM-1 results

would predict this lower impact on postprandial lipemia following consumption of NIE versus

46

the IE blends. However, in contrast to the data from healthy participants, a significant correlation

was not observed between the serum TAG concentrations over time and TIM-1 lipid

bioaccessibility for the obese participants (R2=0.1723; p=0.0613). Also, as with the healthy

participants, serum TAG AUC following consumption of the lipids was not correlated with TIM-

1 total lipid bioaccessibility at 360 min (R2=0.3591; p=0.5909). Lastly, both the human and

TIM-1 results reflect differences in terms of FA absorption between the IE and NIE fats, i.e.

stearic acid absorption and bioaccessibility were higher with consumption of the CIE and EIE

compared with NIE.

Figure 3.5 Correlation analyses between TIM-1 6 h bioaccessibility (%, error bars indicate the

SEM, n=2) and human study 6 h postprandial serum TAG concentration of non-obese subjects

(mmol/L, n=10) fed the same test fats.

3.5 Conclusions

The objective of this work was to investigate the impact of TAG structure and physical

properties of NIE and CIE and EIE blends with similar FA composition on in vitro digestive

processing and in comparison, with previously observed postprandial lipemic effects. The results

confirm previous evidence that TAG molecular structure and melting point impact lipid

1.0 1.2 1.4 1.6 1.80

20

40

60

Serum TAG (mmol/L)

TIM

-1 B

ioac

cess

ibil

ity (

%)

R2 = 0.8640p<0.0001

47

digestion. Moreover, they help to validate the use of the TIM-1 in modeling and predicting

human digestion, as the TIM-1 bioaccessibility data showed good linear correlation with human

serum TAG concentration. The lower concentrations of SSS in the IE versus NIE blends and the

corresponding lower melting temperatures and solid fat contents at 37 °C were associated with

higher in vitro TIM-1 bioaccessibility and lower variability. Therefore, physical properties and

composition of the fat blends impacted digestion. The presence of higher proportions of

undigested stearic acid in the ileal efflux of NIE indicates the lower SSS digestibility and the

limited TAG digestibility of stearic acid-containing molecules was shown by the presence of 44-

61% (relative percentage) stearic acid in the IE undigested TAG. Results from the TIM-1

concerning the effects of interesterification mirrored the trend towards higher lipid absorption

from the IE blends versus NIE. The TIM-1 experiments indicated potential differences in lipid

absorption between the IE and NIE, supporting that sophisticated in vitro models can aid in the

study of mechanisms impacting the digestion of structured lipids.

3.6 Acknowledgements

The authors would like to acknowledge Dr. John Carney, Solazyme Inc. for the LC/MS TAG

analysis and Lois Xinjie Lin for assistance with the TIM-1. We also acknowledge the financial

support of Canada’s Natural Sciences and Engineering Research Council.

3.7 Bridge to Chapter 4

IE lipids are a model fat to consider the impacts of SFC on digestive processing and metabolism

given that IE and non-IE lipid blends have similar fatty acid chemical compositions. However,

they also have different TAG positional distributions which makes it impossible to distinguish

between differences in digestibility based on enzyme positional specificity since FA

48

stereospecific position is a determinant factor for the metabolic fate of the resultant FA and the

2-MAG molecules (Berry, 2009; Karupaiah & Sundram, 2007) and physical properties, i.e.

melting temperature. Bonnaire and colleagues (Bonnaire et al., 2008) took the novel approach of

tempering emulsions to contain undercooled liquid and crystalline solid emulsion droplets of

identical composition. This was a unique strategy to address a critical research gap although

there was no follow up work reported. Furthermore, since the tripalmitin emulsions were

formulated using non-food grade ingredients that system was not suitable for in vivo

investigations. We also identified morphological differences between the liquid and partially

crystalline droplets (Huynh & Wright, 2018). Therefore, in the next chapter, we prioritized

finding a combination of food grade lipid and emulsifier that could be tempered to contain

undercooled and crystalline droplets for in vitro digestion investigations to isolate the effects of

lipid physical state on digestive processing.

49

CHAPTER 4

CRYSTALLINITY AND ASSOCIATED GASTRIC COALESCENCE ATTENUATE

PALM STEARIN EMULSION DROPLET IN VITRO LIPOLYSIS

This chapter has been published in the Journal of Agricultural and Food Chemistry.

“Adapted with permission from (Thilakarathna, S. H. and Wright, A. J. (2018) ‘Attenuation of

Palm Stearin Emulsion Droplet in Vitro Lipolysis with Crystallinity and Gastric Aggregation’,

Journal of Agricultural and Food Chemistry, 66(39), pp. 10292–10299. Doi:

10.1021/acs.jafc.8b02636). Copyright (2018) American Chemical Society.”

Co-authorship statement

S.H. Thilakarathna and A.J. Wright designed the experiments. S. H. Thilakarathna conducted the

laboratory experiments, sample analysis, data analysis and manuscript writing. A.J. Wright

supervised and guided with the experimentation, data interpretation and manuscript review.

50

4.1 Abstract

Emulsions with partially crystalline solid (SE) and undercooled liquid (LE) droplets with

equivalent droplet sizes (centering ~416 nm), surface charge (~-56 mV), and spherical

morphologies were prepared by hot microfluidization based on 10% palm stearin and 0.4% Span

60. Lipid crystallinity attenuated early in vitro gastroduodenal lipolysis (p<0.05), both with and

without inclusion of a gastric phase (p<0.05). Gastric exposure, in particular acidic pH, led to

partial coalescence in SE, flocculation and partial crystallization in LE, and attenuated the rate and

extent of lipolysis in both samples. In vitro shear conditions further impacted colloidal stability,

particularly for SE, with implications for digestibility. Although lipid crystallinity consistently

attenuated early lipolysis, SE gastric phase partial coalescence had a relatively greater impact on

digestibility than did droplet physical state. These findings support that a complex interplay exists

between droplet physical state, colloidal properties, and digestion conditions, combining to impact

emulsion in vitro lipolysis.

Key words: emulsions, physical state, in vitro digestion, lipolysis, colloidal stability.

51

4.2 Introduction

Dietary lipid digestion has wide ranging implications for health and disease. Specific impacts of

lipid physical properties, i.e. physical state, solid fat content (SFC), polymorphism, etc. in

determining digestibility and absorption are rarely considered, but this is an area of renewed

interest (McClements et al., 2009; McClements, Decker, Park, & Weiss, 2008; Michalski et al.,

2013). Indeed, there is in vitro (Guo, Bellissimo, et al., 2017; Thilakarathna et al., 2016), animal

(Bergstedt et al., 1990; Kaplan & Greenwood, 1998; Wang et al., 2015), and human (Berry et al.,

2007; Hall et al., 2014; Robinson et al., 2009) study evidence that fatty acid release and/or

bioavailability can be slower or lower for lipids with relatively higher melting temperatures or

higher SFCs compared with those that are liquid at body temperature (i.e. 37 °C). For example,

canola stearin solid lipid nanoparticles stabilized with either Poloxamer 188 or Tween 20 had

slower lipolysis rates and were hydrolyzed to lower extents compared to canola oil emulsions

prepared with the same emulsifiers (Nik et al., 2012a). Similarly, in whey protein stabilized

emulsions containing different proportions of liquid soybean oil and high melting fully

hydrogenated soybean oil, decreasing rates of lipid digestion were evidenced with increasing

SFC (Guo, Bellissimo, et al., 2017).

Unfortunately, in the above-mentioned and other studies, lipid physical property differences

were confounded by various other physicochemical differences, including composition (i.e. fatty

acid, triacylglycerol (TAG), or emulsifier), particle size, morphology, and/or surface charge.

Bonnaire et al. (Bonnaire et al., 2008) uniquely addressed this by comparing the in vitro

digestibility of undercooled liquid versus crystalline solid tripalmitin and sodium dodecyl

sulphate emulsion droplets achieved by tempering the samples differently after hot high pressure

homogenization. Huynh and Wright (Huynh & Wright, 2018) replicated their finding that

52

lipolysis was more extensive for the liquid versus solid droplets when a gastric phase was

included in the in vitro model and also evidenced the induction of a small amount of crystallinity

in the undercooled droplets. The solid and liquid particles were also found to have different

morphologies, meaning that the comparison of physical state was partly confounded by

differences in interfacial area. This is important given that TAG lipolysis is an interfacial

process. The same trend of attenuated early lipolysis for the liquid state was also recently

observed in comparisons of undercooled liquid and solid cocoa butter emulsion droplets

formulated using Span 60 and Tween 60 (Hart et al., 2018). In this case, the droplets had

equivalent compositions, size distributions, surface charges, and morphologies, although

experiments were carried out at 25C versus 37 C, in order to maintain the differences in droplet

crystallinity.

In the present study, we aimed to extend our understanding of how lipid physical state impacts

digestibility using emulsions with undercooled liquid (LE) and crystalline solid (SE) droplets at

37 °C in which particle size, morphology and surface charge were similar and using only food

permissible ingredients and levels (Health Canada, 2016). Initial experiments were carried out

with various lipids (i.e. palm oil, coconut oil, fully hydrogenated cottonseed oil, fully

hydrogenated canola oil, coconut oil, and palm stearin and blends thereof) and emulsifiers (i.e.

Tween 60, Span 60, Tween 80, Span 80, soy lecithin, and blends thereof). It was very

challenging to find a combination where droplets could be stabilized both as undercooled liquid

droplets and with contrastingly high SFC (data not shown). However, samples with 10 wt% palm

stearin and 0.4% Span 60 were successfully tempered into undercooled liquid and partially

crystalline droplets. Span 60 was partly investigated based on reports it can delay some TAG

polymorphic transitions (Aronhime, Sarig, & Garti, 1987; Aronhime, Sarig, & Garti, 1988) and

53

therefore, might not promote crystallization in the undercooled system. Herein, we provide a

characterization of these palm stearin-0.4% Span 60 emulsions and detail their behaviour during

exposure to simulated duodenal conditions (with and without a gastric phase), aimed at

addressing the hypothesis that TAG physical state, specifically, attenuates lipolytic digestibility.

Experiments were also conducted with higher impact shear conditions (i.e. inclusion of glass

beads), without pepsin, and at pH 7 to determine these relative contributions to colloidal

stability, crystallinity, and subsequent lipid digestion.

4.3 Materials & Methods

4.3.1 Materials

Palm stearin (PS) and sorbitan monostearate (Span 60) were generously provided by Bunge Oils

Inc. (Bradley, IL, USA) and Croda Canada Ltd. (Vaughan, ON, Canada), respectively. The

major FAs of PS were palmitic acid at ~58% and monounsaturated oleic acid at ~27%, based on

gas chromatography. (Anderson et al., 2014) POP (26.1 0.1%), PPP (23.1 0.1%), and POO

(13.8 0.1%) were the main TAGs present, based on HPLC analysis described below. For the in

vitro gastro-duodenal experiments, pancreatin from porcine pancreas (4 × USP, contains trypsin,

amylase, lipase, ribonuclease and protease), porcine bile extract (composition as previously

reported: (Hur, Decker, & McClements, 2009) hyodeoxycholic acid (1-5%), deoxycholic acid

(0.5-7%), cholic acid (0.5-2%), glycodeoxycholic acid (10-15%), and taurodeoxycholic acid (3-

9%), ~49% bile salts), pepsin (from porcine stomach mucosa with activity of 1020 units mg−1

protein), and pyrogallol (99%, A.C.S. reagent) were purchased from Sigma Aldrich (St. Louis,

MO, USA).

4.3.2 Triacylglycerol composition analysis

54

The TAG composition of bulk PS was determined using a Waters Alliance model 2690 high

performance liquid chromatography with a refractive index detector (Waters model 2410,

Waters, Milford, MA, USA). Chromatographic separation of the diacylglycerols and TAGs was

achieved using a Waters xbridge C18 (Waters Limited, Mississauga, ON, Canada) column

(4.6 mm × 250 mm internal diameter with 5 μm particle size) and identification made by

comparison with TAG internal standard retention times. Isocratic elution with a flow rate of

1 mL/min of degassed acetone/acetonitrile 60/40 (v/v) was applied and column and detector

temperatures were set at 40 °C. Data were analyzed using Millenium32 (K&K Testing, LLC,

Decatur, GA, USA).

4.3.3 Preparation and analysis of palm stearin emulsions

A 10% (w/v) palm stearin in water emulsion was prepared using 0.4% (w/v) sorbitan

monostearate (Span 60) as the emulsifier. Span 60 is a non-ionic emulsifier with an HLB value

of 4.7 (Griffin, 1949). In brief, to prepare 100 mL of the emulsion, 10 g of PS and 0.4 g of Span

60 were melted at 80 °C for 30 min. A coarse emulsion was prepared by adding 80 °C deionized

water (DI) and mixing with a hot handheld homogenizer at 12,000 rpm for 1 min (Ultra Turax,

Ika T18 Basic, Germany). The emulsion was then transferred to the hot hopper of a

microfluidizer (M-110EH, Microfluidics, MA, USA) and hot homogenization carried out at 125

MPa for 5 passes with the piping immersed in a 95 °C water bath. Half of the hot homogenate

was transferred to a glass jar warmed at 80 °C and the sample then directly placed at 37 °C to

produce the liquid emulsion droplets (i.e. LE, stored for a maximum of three days). The

remainder was transferred to a glass jar previously cooled at 5 °C and then placed in an ice-water

bath for 20 min before storage at 5 °C (for a maximum of seven days) until 20-30 min before

analyses when it was moved to 37 °C. This sample containing solid particles is referred to as

55

SE. Strict temperature control was maintained throughout sample preparation, storage, and

analysis, including warming utensils to 37 °C and minimizing temperature fluctuations, in order

to maintain the intended differences in physical state, which was confirmed routinely by

differential scanning calorimetry, as below.

4.3.4 Particle size and -potential

Particle size of the emulsion samples was determined by laser diffraction using a Mastersizer

2000S (Malvern Instruments Inc., Southborough, MA, USA). Particle size distributions and

volume-weighted (D4,3) and surface-weighted (D3,2) mean diameters were measured. Emulsion

droplet charge was measured using a particle electrophoresis instrument (Zetasizer Nano ZS,

Malvern Instruments Inc., Southborough, MA, USA). In order to minimize multiple scattering

effects, emulsion samples were diluted 1:250 with DI water. For both techniques, values of 1.45

and 1.33 were used as the refractive indices of PS and water, respectively.

4.3.5 Melting and re-crystallization behavior

The melting and crystallization behavior of bulk PS and each batch of the emulsions were

analyzed in duplicate by differential scanning calorimetry (DSC, Q2000 model (TA Instruments,

Mississauga, ON, Canada). In brief, 5-10 mg of emulsion was weighed and sealed in pre-

weighed alodined aluminum DSC pans (TA Instruments, Mississauga, ON, Canada). Samples

were held in the DSC at 37 °C for 3 min, then heated at 5 °C/min to 80 °C, held for 3 min and

then cooled at 5 °C/min to 0 °C. To confirm complete crystallization, SE samples at 37 °C were

cooled from 37 to 0 °C at 5 °C/min. Enthalpies and peak onset and maximum temperatures were

determined using the system software (TA Instruments Universal Analysis 2000 software, TA

Instruments, ON, Canada).

56

4.3.6 Lipid polymorphism

The polymorphism of the SE droplets before and during gastric and duodenal digestion, as well

as PS tempered similarly as SE, was investigated using a MultiFlex x-ray diffractometer

(Rigaku, Japan). Scans were performed from 15 to 30° at 0.3 °/minute on ~ 1 mL of sample

placed on an x-ray diffraction (XRD) glass slide pre-heated and maintained at 37 °C. The

instrument was operated at 40 kV and 44 mA with copper as the x-ray source ( = 1.54 Å) and

angle slits of 0.5, 0.5 and 0.3 mm. Peak positions were determined based on Bragg’s law using

MID’s Jade 9.0 software (Rigaku, Japan).

4.3.7 Solid fat content (SFC)

The SFC of bulk PS was measured according to AOCS official method Cd 16b-93 using a

Bruker Minispec PC/20 series pulsed nuclear magnetic resonance spectrophotometer (Bruker

Spectrospin, Milton, Canada).

4.3.8 Emulsion droplet morphology

Emulsion droplet morphology was observed by light microscopy. A drop of emulsion was

placed on a glass slide pre-heated to 37 C and a pre-heated (37 C) coverslip was carefully

placed on top, being careful to ensure no air bubbles were present. The sample was then

observed with an Olympus BH light microscope (Olympus, Tokyo, Japan) equipped with a Sony

Camera (Sony Corporation, Japan) and Image Capture software under polarized and bright field

conditions.

4.3.9 In vitro digestion during exposure to gastric and duodenal conditions

57

The two emulsion samples were subjected to an in vitro digestion model consisting of simulated

gastric and duodenal phases, as previously described (Lin & Wright, 2018). All fluids, along

with the amber jars used, were warmed to 37 °C prior to use. Briefly, 5 mL of emulsion sample

was added to 5 mL of simulated gastric fluid (2000 U/mL pepsin and 12.6 mg/mL pyrogallol as

an antioxidant) and incubated at pH 3, 37 °C for 2 h in a shaking incubator at 250 rpm to

simulate the gastric conditions. Ten milliliters of simulated duodenal fluid were then added to

start the duodenal phase and the final digestion mixture consisted of 2000 U/mL pancreatic

lipase, 10 mg/mL bile extract (~10 mM bile salts), and 3.8 mg/mL phospholipids (~8 mM). The

pH of the final mixture was adjusted to 7 by dropwise addition of 1 M sodium hydroxide and the

duodenal digestion phase was carried out for 4 h. Variations in the experimental conditions were

applied in order to understand the impact of specific factors. Firstly, only the duodenal phase was

utilized (i.e. without any preceding gastric conditions). Secondly, glass beads (four, each 10 mm

in diameter) were added to each digestion jar to change the shear conditions. Lastly, when the

gastric stage was included, experiments with two different pH levels (i.e. 3 and 7), as well as in

the absence and presence of pepsin, were conducted.

Digestate analysis

Duodenal digestate samples were collected at 2, 5, 10, 15, 30, 45, 60, 90, 120, 180, and 240 min

and extracted into acidic hexane. FFA concentrations were determined using a non-esterified FA

kit (NEFA-HR2, Wako Diagnostics, VA, USA) with spectrophotometric analysis at 550 nm

(Spectramax plus, Molecular Devices Corporation, CA, USA) (Lin & Wright, 2018). Digestate

samples were inspected visually and DSC and particle size analyses performed at the end of the

gastric phase and during the duodenal phase. Because of extensive aggregation during the acidic

gastric phase, SE could not be sampled representatively.

58

4.3.10 Data and statistical analysis

At least 3 separate experiments were carried out with freshly prepared samples and at least 3

analytical replicates were performed for each. Normality of data was confirmed with d’Agostino-

Pearson omnibus testing and data analysis was performed either by t-tests or ANOVA using

GraphPad Prism (GraphPad Software, San Diego, CA, USA) and a significance level of P<0.05.

Results are reported as mean SD.

4.4 Results and Discussion

4.4. 1 Compositional, physical property and morphological analyses

Different physical states were evidenced for LE and SE using DSC (Figure 4.1a & b). In LE

there was no melting event, confirming there was no crystalline fat present, whereas SE had a

peak melting temperature of 52.6 0.4 °C, indicating the presence of solid fat. There was no

crystallization observed when SE was directly cooled in the DSC from 37 to 0 ° C (data not

shown). Also, the melting enthalpy of SE (i.e. 4.5 0.5 J/g) was ~10% of that for bulk PS (i.e.

44.7 4.7 J/g), providing crude evidence that SE lipid droplets achieved maximum attainable

crystallinity for an emulsion with 10% lipid phase. For reference, according to pNMR analysis

(Figure 4.1c inset), the SFC of palm stearin at 37 °C was 33.17 0.05 %. Also, when PS was

tempered using the same method of producing SE, it had a peak melting temperature of 54.0

0.3 °C and re-crystallization peaks at 30.2 0.0 °C and 7.7 0.4 °C (Figure 4.1c). Bulk Span 60

had a melting peak at 57.9 0.7 °C and re-crystallization peaks at 51.8 0.4 and 44.8 0.6°C

(data not shown). According to XRD, both SE and the tempered bulk PS contained the

polymorph, with a dominant reflection at 4.6 Å and weaker ones at 3.9 and 3.7 Å (Figure 4.1d).

59

Figure 4.1 Representative DSC thermograms of melting and crystallization of (a) Liquid (LE)

and (b) solid (SE) droplets, (c) bulk palm stearin (PS) tempered as per SE with PS SFC curve

inset, (d) X-ray diffractograms for bulk PS and SE. SFC Data indicate mean SD, n=3.

Figure 4.2 shows that the LE and SE particle size distributions perfectly overlaid each other, with

monomodal distributions centering around ~ 416 nm. Values of D3,2 and D4,3 also were not

significantly different (p>0.05) between LE (0.36 0.01 and 0.49 0.01 m, respectively) and

SE (0.31 0.03 and 0.46 0.03 m, respectively). -potential values for LE (-56.5 2.7 mV)

and SE (-56.6 4.8 mV) were statistically similar, as well (p>0.05). Although Span 60 is a non-

0 20 40 60 80

-1

0

1

2

Temperature (°C)

Hea

t F

low

(W

/g)

0 20 40 60 80

-0.4

0.0

0.4

Temperature (°C)

Hea

t F

low

(W

/g)

15 20 25 300

4000

8000

12000Bulk PSSE

4.6 Å

3.9 Å

3.7 Å

2q (degrees)

Inte

nsi

ty (

Count)

0 20 40 60 80

-0.4

0.0

0.4

Temperature (°C)

(c)

(a)

(d)

(b)

0 10 20 30 40 50 600

20

40

60

Temperature (°C)

Soli

d f

at

con

ten

t (%

)

60

ionic emulsifier, the higher negative droplet -potential values observed were believed to be

contributed by the ionic impurities in the emulsifier and/or molecules like FFA present in palm

stearin. Based on visual observation and particle size analysis, LE and SE were found to be

gravitationally stable during storage at 37 °C for 3 days and at 5 °C for 7 days, respectively (data

not shown). Therefore, LE (stored at 37 °C) was always analyzed and used within 3 days of

preparation. SE samples were stored for a maximum of 7 days at 5 °C and equilibrated at 37 °C

for 30 min prior to use.

Figure 4.2 Representative particle size distributions of LE and SE. Data indicate mean SD.

LE and SE were observed under bright field (Figure 4.3a & c) and polarized (Figure 4.3b & d)

light and both emulsions contained spherically shaped droplets. Under polarized light, the typical

Maltese cross-like pattern as observed for birefringent crystalline lipids were obvious for the SE

droplets. Some birefringence was also evident in the LE samples, mostly at the interface,

suggesting interfacial organization of the Span 60 emulsifier. Indeed, when canola oil emulsions

0.01 0.1 1 10 100 1000 100000

2

4

6

8

10 LE SE

D3,2 (mm): 0.36±0.01 0.31±0.03

D4,3 (mm): 0.49±0.01 0.46±0.03

z-potential: -56.5±2.7 -56.6±4.8

(mV)

Particle size (mM)

Volu

me

%

61

were similarly prepared with Span 60, the same interfacial shell was observed (Figure 4.3e & f),

pointing to the presence of organized Span 60 at the LE droplet interface versus TAG

crystallization. Span 60 was similarly shown to form a shell around the droplets in water-in-oil

emulsions containing canola oil (90%) and fully hydrogenated canola oil (8.5%) (Tran, Green,

Ghosh, & Rousseau, 2017). The impact of an emulsifier on TAG nucleation and crystal growth

depends on molecular interactions and degree of structural complementarity (Judith Schlichter

Aronhime et al., 1988; Douaire et al., 2014). In bulk cocoa butter, Span 60 was recently found to

phase separate, crystallize, and effectively slow down TAG nucleation and crystal growth

(Sonwai, Podchong, & Rousseau, 2017). In the present study, Span 60 and its stearyl chains,

evidently did not induce LE droplet crystallization during the 3 days of storage at 37 °C, despite

these TAGs being in a thermodynamically metastable state. In summary, LE and SE droplets

were completely liquid versus partially crystalline, respectively, but had the same size

distribution, surface charge, and morphology. This eliminates these factors as confounding

influences in subsequent experiments.

62

Figure 4.3 Microstructure of (a & b) LE, (c & d) SE, and (e & f) 10% canola oil-0.4% Span 60

emulsion droplets under (a, c, & e) bright field and (b, d, & f) polarized light conditions.

10 µm 10 µm

10 µm 10 µm

10 µm 10 µm

(a) (b)

(d) (c)

(f) (e)

63

4.4.1 In vitro duodenal lipolysis without exposure to a gastric phase

In vitro duodenal lipolysis experiments were initially performed without the gastric phase

(Figure 4.4a). Accordingly, both LE and SE reached the maximum achievable lipolysis by 60

min (p<0.05), with higher lipolysis observed for LE than SE at 5 minutes (p<005), indicating a

faster initial rate. The FFA release area under the curve (AUC) was the same for LE and SE,

reflecting similar extents of lipolysis over the 4 h digestion period (p>0.05). Therefore, the liquid

state facilitated lipolysis of the emulsion droplets, but did not change the ultimate digestibility.

No obvious particle aggregation was observed for either system when digestive fluids were

added, suggesting electrostatic repulsion remained (Fredrick, Walstra, & Dewettinck, 2010).

According to Figure 4.5a, slightly larger particle sizes (more so for LE) were observed by 15

minutes into the duodenal digestion, likely related to coalescence, given the extensive lipolysis

already achieved by this point. There also appears to be a shift towards smaller particles in SE,

but not LE, potentially related to differences in the rates of initial lipolysis and also to a higher

tendency for LE droplets, being liquid, to coalesce during digestion and for SE particles to

reduce in size as they are hydrolyzed. When LE digestate samples were analyzed by DSC there

was no evidence that the undercooled droplets had crystallized (i.e. no shift in baseline, data not

shown).

However, with this high level of dilution and the rapid lipolysis observed, it may not have been

possible to detect small amounts of solid fat. In summary, when samples of undercooled LE and

partially crystalline SE (which were equivalent in particle size, morphology, and zeta potential)

were exposed to conditions simulating the upper duodenal environment, both were rapidly and

extensively hydrolyzed, with slightly, but significantly, faster initial digestion of the liquid

droplets. To the best of our knowledge, this is the first such investigation comparing droplets

64

with such similar particles. The results lend strong support to the potential for kinetic differences

in lipid digestion based specifically on whether TAGs are present as liquids or solids and based

on overall SFC.

0 60 120 180 2400

20

40

60

80

100(b)

p<0.05

Lip

oly

sis

(%)

0 60 120 180 2400

20

40

60

80

100 (c)

p<0.05

Time (min)

Lip

oly

sis

(%)

0 60 120 180 2400

20

40

60

80

100 (a)

LE SE

Lip

oly

sis

(%)

65

Figure 4.4 Free fatty acid release during (a) duodenal digestion without the gastric phase, (b)

duodenal digestion with the gastric phase, and (c) gastro-duodenal digestion with added glass

beads (d=10mm). Data indicate mean SD, n=3. Significant differences (p<0.05) were observed

in (b) up to 60 min and (c) up to 240 min. The lipolysis data was fitted to a first order kinetics

model for graphing purposes with r2 values obtained for LE and SE, respectively, were (a)

0.7771 and 0.8254, (b) 0.9231 and 0.9935, (c) 0.9675 and 0.9825.

4.4.2 In vitro duodenal lipolysis following exposure to a gastric phase

LE and SE were next exposed to a more physiologically relevant approach by mimicking gastric

conditions before the duodenal phase. Although Span 60 is considered to be colloidally stable

under mildly acidic conditions (Croda Europe Ltd, 2009), particle size increased for both

samples (Figure 4.5b & c), likely related to decreases in droplet electrostatic repulsion in the

highly acidic environment. Hydrogen ion shielding of negatively charged emulsion droplet

interfaces can lead to decreased charge repulsion, and thereby droplet aggregation and emulsion

destabilization (Comas, Wagner, & Tomás, 2006; Huynh & Wright, 2018; Mantovani,

Cavallieri, Netto, & Cunha, 2013). LE easily redispersed with mixing and upon exposure to the

duodenal fluids, indicating flocculation at the gastric stage. In contrast, SE was extensively

clumped and resisted dispersal in the duodenal fluids, pointing to the likelihood of partial

coalescence occurring with gastric stage aggregation (Figure 4.5c). According to pNMR, the

palm stearin contains 33.17 0.05 % solids at 37 °C. This is within the range where partial

coalescence is expected, especially under conditions of shear when droplet collisions are

promoted (Fredrick et al., 2010). Other studies have evidenced that solid fat contributes to

differences in gastric colloidal stability (Golding & Wooster, 2010; Golding et al., 2011). For

66

example, in emulsions formulated with acid-susceptible whey protein isolate and sodium

caseinate, those consisting of higher solid fat contents tended to destabilize more under gastric

conditions in vitro (Guo, Bellissimo, et al., 2017) and in vivo (Steingoetter et al., 2015),

respectively. In a human study, phase separation of a Span 80-stabilized emulsion was observed

in the stomach with extensive fat layering and there were associated effects on satiety (Marciani

et al., 2009).

Figure 4.5 Representative graphs showing particle size distributions of duodenal digestion of

LE and SE (a) at 15 min duodenal digestion without gastric phase and (b) at 2 h gastric and 15

min duodenal digestion with gastric phase, and visual observations of LE and SE at (c) 2 h

gastric and (d) end of duodenal (4 h) phase following gastric phase.

Undercooled emulsion droplets are metastable and susceptible to the induction of crystallization.

Partial crystallization was previously observed for 10% tripalmin-SDS undercooled emulsion

0.01 0.1 1 10 100 1000 100000

5

10LE-15m-D

SE-15m-D

(a) Digestive fluids

Particle Size (mm)

Volu

me

%

0.01 0.1 1 10 100 1000 100000

5

10LE-120m-G

LE-15m-D(b) SE-120m-G

SE-15m-D

Particle Size (mm)

(c) (d)

67

droplets during in vitro digestion (Huynh & Wright, 2018). In this study, when LE digestate was

analyzed by DSC at the end of the gastric phase, a small melting event was observed at 54.3

0.5 °C (data not shown). Unfortunately, the extent of this crystallization is difficult to quantify.

For crude comparison, the melting enthalpy for this event was 45.4 6.1 % compared with the

calculated result of what the SE melting enthalpy would be at an equivalent level of dilution.

This suggests a significant proportion of the LE lipids crystallized in the presence of acid,

potentially aggravated by shear leading to increased collision frequency. According to Figure

4.5b, some of the LE aggregation was reversible, i.e. LE droplets partly redispersed by 15

minutes into duodenal phase (i.e. peaks centered around 1 and 10 m were observed). Therefore,

despite LE partial crystallization, partial coalescence in these samples was minimal compared to

SE, potentially related to the low SFC. Assuming the crystallinity induced in LE during the

gastric exposure was ~45% SFC (see above), based on PS containing 33% SFC at 37 C, the

droplet solids content for this emulsion would be ~ 14.8%, i.e. potentially too low for partial

coalescence. Differences in the location and type of crystals formed might also play a role, given

the very different nucleation conditions when SE was quench cooled to crystallize (Fredrick et

al., 2010). Interestingly, when samples of LE and aggregated SE duodenal digestates were

analyzed by DSC, slightly higher end of melt temperatures (i.e. ~58 C) were consistently

observed, compared with those drawn at the end of the gastric phase (i.e. 52-54 C; p<0.05). This

corresponds to the melting temperature of Span 60 (i.e. 57.9 0.7 °C), suggesting interfacial

displacement of the emulsifier during lipolysis.

Figure 4.4b shows that, when the gastric phase was included, the initial rate of duodenal lipolysis

was much faster for LE versus SE (i.e. ~3 x faster within the first 5 minutes, p<0.05) and that

lipolysis was significantly higher for LE versus SE up to 60 min (p<0.05). This is despite the

68

partial crystallinity observed in LE. Similar levels of lipolysis were reached for SE and LE by 1.5

and 4 h, (p>0.05), although lipolysis was, overall, relatively low for both samples. Comparing

with Figure 4.4a when the gastric phase was omitted, LE and SE duodenal lipolysis was ~ 60 %

lower (Figure 4.4b). Incomplete digestion can occur with in vitro methods because of insufficient

enzyme activity or the interfacial accumulation of products of digestion. However, the above

results without the gastric phase support that the conditions utilized were capable of fully

hydrolyzing these samples. Rather, the results point to differences in interfacial area as being a

critical factor.

Gastric phase destabilization, including with the presence of solid fat (Day et al., 2014), was

previously shown to contribute to attenuated lipid digestion (Mackie, 2017). In this study, the

fact that LE partially crystallized during the gastric phase allows the comparison of duodenal

lipolysis for completely liquid (0% SFC) dispersed droplets (i.e. LE without gastric phase –

Figure 4.4a), partially crystalline (~15% SFC) partially aggregated droplets (i.e. LE after gastric

phase – Figure 4.4b), partially crystalline (~33% SFC) dispersed droplets (i.e. SE without gastric

phase- Figure 4.4a), and partially crystalline (~33% SFC) highly aggregated partially coalesced

droplets (i.e. SE after gastric phase – Figure 4.4b). Specifically, the partially crystalline (~15%

SFC) LE had a comparatively lower rate and extent of lipolysis during early duodenal digestion

compared to both the LE and SE systems which remained dispersed because they were never

exposed to the gastric phase (p<0.05). This points to the substantial impact of colloidal

destabilization, amplified by the presence of solid fat, on digestibility. In summary, when well-

dispersed, liquid droplets were hydrolyzed more readily than those containing solid fat.

Moreover, the presence of solid fat in the emulsions was associated with more extensive gastric

phase aggregation, specifically partial coalescence, and associated attenuated lipolysis. In vivo,

69

these systems would be expected to have different rates of gastric emptying (Steingoetter et al.,

2015), although such dynamics could not be controlled for using the current model.

4.4.3 Role of gastric parameters in determining duodenal lipolysis

To further investigate gastric phase destabilization, samples were also incubated without pepsin

and at pH 7, instead of pH 3. In the case of pH 7 (with pepsin) SE showed no visible

aggregation. Particle size analysis (Figure 4.6a & b) also shows that LE and SE were colloidally

unchanged in the presence of pepsin at pH 7. However, at pH 3, both with and without pepsin,

increases in particle size were observed for LE and SE, albeit to different extents (Figure 4.6).

Whenever LE was exposed to pH 3, flocculation was visibly observed, but samples redispersed

with mixing. SE aggregation was much more extensive, both with and without pepsin at pH 3.

Indeed, acidic pH was a major factor in emulsion droplet aggregation. In terms of LE liquid state

stability, regardless of pH (3 or 7) or the presence or absence of pepsin, similar levels of

crystallinity were induced (p>0.05, data not shown). Therefore, LE was vulnerable to partial

crystallization, independent of acid-induced flocculation. Separate experiments did show that LE

crystallization was not induced by the introduction of shear (250 rpm orbital shaking for 1 h) or

dilution with DI water (data not shown). The differences in the extent of gastric phase

destabilization of solid versus liquid Span 60-containing emulsions point to interactions between

solid fat content, emulsifier behavior, and types of crystallinity, that warrant further

investigations.

70

Figure 4.6 Gastric phase impact on (a) LE and (b) SE particle size distribution. With

pepsin at pH 3, without pepsin at pH 3, with pepsin at pH 7, undigested LE,

undigested SE.

Static digestion models use different levels and types of shear, i.e. orbital, horizontal, or rotary

shaking, stirring, and physical impact via a solid matrix (glass beads), etc. to mimic gastric

mixing and motility (Eldemnawy et al., 2015; Kong & Singh, 2008; O′Sullivan, Davidovich-

Pinhas, Wright, Barbut, & Marangoni, 2017), although achieving physiological relevance is

difficult. To facilitate mixing and add impact forces (Kong & Singh, 2008), glass beads

(d=10mm) were added to each sample jar. Doing so exaggerated the differences in colloidal

properties between SE and LE at end of the gastric phase. Specifically, LE remained dispersed,

but even larger densely packed aggregates of SE were observed, visually. The initial SFC of the

SE droplets, decrease in droplet electrostatic repulsion, and increased shear forces during gastric

exposure potentially enhanced the collision frequency and capture efficiency, increasing the

susceptibility of SE for partial coalescence (Fredrick et al., 2010). The SE aggregates formed

resisted redispersal throughout duodenal digestion. In terms of lipolysis, the presence of the

0.01 0.1 1 10 100 1000100000

5

10 (a)LE

Particle Size (mm)

Volu

me

%

0.01 0.1 1 10 100 1000 100000

5

10 (b)SE

Particle Size (mm)

71

beads attenuated early fatty acid release in both LE and SE (p<0.05). Also, LE lipolysis was

significantly higher than for SE throughout the digestion (Figure 4.4c; p<0.05). At 45 min, for

example, it was ~90% higher. Therefore, the higher impact forces resulted in attenuated lipolysis

of the SE particles but had relatively minimal impact on LE. This was presumably related to

massive differences in available surface area during duodenal digestion and underscores the

critical role that aggregation induced during the gastric phase can play in lipid digestibility.

In summary, this study investigated the differences in in vitro digestive lipolysis between

partially crystalline solid versus undercooled liquid emulsion droplets with equivalent

composition, size, charge, and morphology and based on the emulsifier Span 60. Early lipolysis

was attenuated by the presence of TAG crystallinity in SE and with particle destabilization,

aggravated by the presence of solid fat, during exposure to simulated gastric conditions.

Moreover, the observed differences in early stage lipolysis support the general hypothesis that

solid fat is less readily digested than liquid oil, pointing to possible kinetic differences, even if

lipolysis is ultimately complete. Indeed, although lipid digestion in healthy adult humans is

generally considered to be complete (Mu & Høy, 2004), there is in vivo evidence pointing to

differences in digestion kinetics based on physical state. For example, rises in plasma (Berry et

al., 2007; Robinson et al., 2009) and chylomicron (Hall, Brito, Huang, Wood, Filippou, Sanders,

& Berry, 2014) TAG were slower during the postprandial period when lipids with solid fat or

higher solid fat contents at 37 C, were consumed, even though similar levels of lipemia were

eventually achieved. Differences in lipid digestion kinetics may also impact the release location

of encapsulated molecules, inductions of hormonal feedback loops related to satiety, and change

the profile of the lipemic response (Golding & Wooster, 2010; Marciani et al., 2009; Steingoetter

et al., 2017). The current results also support that in vitro digestion experimental parameters, in

72

particular the inclusion of a gastric phase and impact shear conditions, influence the results

obtained. The use of a static model lacking an oral phase and gastric lipase are limitations of the

work. Additionally, although comparing crystalline droplets to undercooled droplets of identical

composition enables a unique comparison minimizing confounding influences, the situation does

not exactly resemble that of typical liquid dietary lipids, given the thermodynamic instability of

the LE system and the fact that partial crystallization of LE was induced. Further investigations

are required to better understand how lipid physical properties impact colloidal and lipolytic

behavior during digestion, including through validation with human research.

4.5 Bridge to Chapter 5

In Chapter 4, compositionally identical undercooled liquid and partially crystalline emulsion

droplets were found to differ in terms of their digestibility, i.e. this was when solid fat was

present. That study was successful in eliminating differences in particle properties, i.e. size and

shape. Although TAG SFC has been related to in vivo lipemia, these findings have been based

on lipids with differing FA compositions to achieve the different levels of solids or incidental,

i.e. for IE lipids (Berry et al., 2007; Hall et al., 2014). Never has an in vivo study been reported

where compositionally equivalent lipids were compared to isolate the effects of physical state.

Moreover, the in vivo behaviour of an undercooled emulsion has never been reported. Therefore,

an acute meal postprandial lipid study was undertaken for the same samples discussed in Chapter

5. The aim was to investigate whether the observed slower in vitro lipolysis and lower

digestibility for solid versus liquid TAGs would be associated with similar observations in

postprandial lipemic response.

73

CHAPTER 5

EMULSION DROPLET CRYSTALLINITY MODESTLY ATTENUATES

POSTPRANDIAL PLASMA TRIACYLGLYCEROL RESPONSE IN HEALTHY

MALES: A RANDOMIZED DOUBLE-BLINDED CROSSOVER ACUTE MEAL STUDY

Surangi H. Thilakarathna, Samar Hamad, Amanda Cuncins, Melissa Brown, and

Amanda J. Wright

Co-authorship statement

Amanda J. Wright was the principle faculty investigator and Surangi Thilakarathna was the lead

study coordinator. Surangi H. Thilakarathna and Amanda J. Wright designed the human study

and obtained the ethical and other required approvals. Surangi H. Thilakarathna, Samar Hamad,

Amanda Cuncins, Melissa Brown contributed to participant recruitment, treatment emulsion

preparation, and study visit duties. Surangi H. Thilakarathna conducted the sample and data

analysis and wrote the manuscript. Amanda J. Wright guided with data interpretation. All co-

authors reviewed the manuscript.

74

5.1 Abstract

Background: The presence of solid fat in dietary lipids is assumed to influence digestibility and

postprandial lipemic response (PPL), although studies to-date are limited. In fact, the possibility

that TAG physical state differences, specifically, could alter their absorption is rarely considered.

Objective: This study aimed to investigate whether the presence of solid fat versus undercooled

liquid oil specifically, plays a role in determining PPL by comparing emulsion droplets with the

same compositions and colloidal properties and differing only in terms of physical state.

Methods: Emulsions were carefully tempered to contain identically sized (P>0.05), charged

(P>0.05), and shaped (spherical) undercooled liquid (LE) versus partially crystalline solid (SE,

33.20.1% solid fat at 37 C) emulsion droplets based on 10% palm stearin and 0.4% sorbitan

monostearate. 15 healthy fasting adult males consumed 500 mL of each emulsion on separate

occasions and plasma triacylglycerol (TAG) concentrations, particle size of the plasma

chylomicron-rich fraction (CMRF), and fatty acid composition of the TAGs from the CMRF

were investigated in a randomized double-blinded crossover acute meal study.

Results: Higher postprandial TAG change from baseline (P=0.08), peak concentration (P=0.04),

and iAUC0-6h (P=0.03) were evident for LE versus SE. LE also increased significantly from

baseline (P<0.05) earlier than SE (i.e. from 2 versus 3 h) pointing to an earlier rise in plasma

TAG for lipids in the liquid state. Postprandially, the proportions of palmitic, stearic, oleic, and

linoleic acid in the CMRF-TAG shifted towards those of palm stearin by 4 h but did not differ

between SE and LE (P>0.05). Nor were there differences in CMRF size based on droplet

physical state (P>0.05).

75

Conclusions: The presence of differences in the absorption of SE versus LE, despite their

identical compositions and colloidal properties, is evidence that TAG physical state can

contribute to PPL, with solid fat having an attenuating influence.

Key words: emulsions, solid fat content, healthy males, postprandial lipemia, plasma

triacylglycerol response, chylomicron rich fraction, acute meal study.

76

5.2 Introduction

Cardiovascular disease (CVD) remains the leading cause of death in the United States (Benjamin

et al., 2018) and worldwide (WHO, 2017), with evidence linking postprandial lipemia and

impaired clearance of chylomicron remnants with increased atherosclerotic plaque formation

(Sanders, 2003) and overall disease risk (FAO, 2008). Moreover, there is emerging evidence that

postprandial lipemia is associated with inflammation (Laugerette et al., 2011) and postprandial

handling of digested fat, i.e. tendency for storage vs. beta oxidation (Vors et al., 2013) that can

influence metabolic syndrome and CVD risk. As such, there is strong impetus to understand the

nuances of how dietary lipids are digested and absorbed, including a renewed focus on the role

of lipid physicochemical properties (Dias et al., 2017). It has long been speculated that dietary

lipids are digested at different rates and possibly to differing degrees, based on differences in

triacylglycerol (TAG) melting temperature leading to differences in the proportion of solid fat

versus liquid oil present at body temperature (i.e. SFC at 37 °C). The presence of TAG

crystallinity can impair lipolysis by decreasing pancreatic lipase (Golding et al., 2011) and bile

salt accessibility, thereby limiting enzyme access and emulsification (Guo, Bellissimo, et al.,

2017; Guo, Ye, Bellissimo, Singh, & Rousseau, 2017). Stearic acid absorption is also attenuated

by the formation of insoluble calcium and magnesium soaps in the gastrointestinal tract (Mu &

Høy, 2004). This complicates the ability to discern whether TAG digestibility, specifically is

different based on physical state. Many animal studies concluding that higher melting lipids and

those with relatively higher SFCs at body temperature are digested relatively slowly and are

poorly absorbed have utilized stearic acid-containing lipids (as reviewed by Michalski et al.,

2013). However, in general, in vitro and in vivo animal and human studies support that higher

melting dietary lipids and those with higher SFC at 37 C are hydrolyzed more slowly and/or

77

have lower postprandial lipemic kinetics compared to relatively lower melting lipids with lower

or no SFC at the same temperature. Nonetheless, some of the evidence is conflicting. For

example, some animal studies have reported excretion of undigested solid fat after rats consumed

a higher solid fat containing rodent diets (Bergstedt et al., 1990; Kaplan & Greenwood, 1998;

Wang et al., 2011), whereas others report no difference in digestibility and metabolism of diets

with varying SFCs (Wang et al., 2015).

Some of the evidence for differences in absorption based on TAG physical state comes indirectly

from studies with IE lipids that have related changes in SFC around body temperature (i.e. 37

C) to differences in lipemic response. For example, lower plasma TAG increases and lower

plasma TAG AUC have been observed for IE vs non-IE lipids with higher SFCs at 37C. This

included 24 versus 21 % SFC (at 35 C) for a blend and the IE blend of palm stearin and palm

kernel (80:20) (Hall et al., 2017); 22 % SFC for unrandomized shea blend vs liquid high oleic

sunflower oil (Berry et al., 2007) and 15.2% SFC in IE palm oil vs 3.6% SFC in native palm oil

(Berry et al., 2007). Our group previously showed similarly higher serum levels of stearic acid

after consumption of an IE versus non-IE stearic acid-rich blend based on tristearin (SSS)

(Robinson et al., 2009). Differences in TAG digestibility, specifically, for these stearic-rich

lipids were subsequently supported by the observation of lower stearic acid in vitro

bioaccessibility and more undigested stearic acid in the effluent TAG fraction corresponding to

undigested TAGs using the TIM-1 digestion simulator (Thilakarathna et al., 2016).

Other in vitro digestion studies using oil-in-water emulsions have yielded results supporting that

samples with higher SFCs at 37 C tend to have lower digestibility, in terms of lipolysis kinetics

and lower FA bioaccessibility in vitro (Golding et al., 2011; Guo, Bellissimo, et al., 2017; Nik et

78

al., 2012b). However, each of these comparisons between SFC differences was achieved using

lipids with differing FA compositions and/or FA positional distributions. As such, differences

based on the proportion of solid versus liquid fat present were confounded by compositional and

potentially other physical differences that could alter digestion, e.g. positional distribution related

to pancreatic lipase sn-1,3 specificity and droplet size related to interfacial area.

In a ground breaking study, Bonnaire and colleagues (Bonnaire et al., 2008) novelly used

tempering to formulate and compare the in vitro digestibility of compositionally identical

tripalmitin - sodium dodecyl sulphate emulsions containing droplets in the crystalline versus

undercooled liquid states. Our group replicated their finding of attenuated in vitro lipolysis for

the crystalline emulsion droplets, but also identified confounding differences in particle shape

(Huynh & Wright, 2018). Shape differences were subsequently eliminated working with cocoa

butter – sorbitan monostearate (Span 60)/ polyoxyethylene sorbitan monostearate (Tween 20)

emulsions and the lipolysis trend was confirmed, albeit working at non-physiological 25 C to

maintain the contrasting droplet SFCs (Hart et al., 2018). Most recently we reported (Chapter 4,

Thilakarathna and Wright, 2018) on the comparison of palm stearin – Span 60 (both food

permissible)-based emulsions, where the partially crystalline and undercooled liquid droplets

were identical in terms of composition, size, shape, and zeta potential at 37 C. In vitro lipolysis

of the solid emulsion droplets was slower compared to those in the liquid state (Thilakarathna &

Wright, 2018). The differences were slight when only duodenal digestion was simulated. With

inclusion of acidic gastric conditions, and especially when mixed with glass beads, the lipolytic

differences between the undercooled and crystalline droplets were more dramatic, related to

partial coalescence in the case of droplets containing solid fat. This was consistent with previous

reports that gastric structuring alters emulsion digestibility and postprandial TAG response

79

(Golding et al., 2011). The undercooled droplets were also monitored for the development of

crystallinity and, at the end of the gastric phase, found to contain ~ 45.4 ± 6.1% relative to the

system tempered originally to be maximally crystalline (i.e. ~15.1% versus 33.2% at 37 C,

Thilakarathna and Wright, 2018). Therefore, at the stage of duodenal digestion, LE and SE

differed on the basis of SFC.

In all, substantial evidence supports that lipid SFC, specifically, plays a role in gastrointestinal

lipolysis. Whether or not this translates into differences in postprandial lipemia remains to be

established, with very limited work, to date. Specifically, no in vivo studies have accounted for

confounding differences in lipid composition between test meals when investigating differences

in physical state. Therefore, this randomized acute meal human study was carried out with the

objective of comparing the palm stearin emulsions from our previous in vitro study

(Thilakarathna & Wright, 2018) in terms of postprandial lipemic response, i.e. plasma TAG and

non-esterified fatty acid (NEFA) concentrations, the size distribution and FA composition of the

TAGs present in the chylomicron-rich fraction (CMRF) from the plasma. It compares, for the

first time, the lipemic response for emulsions which differ only in terms of their SFC. It was

hypothesized that the emulsion containing liquid undercooled droplets (i.e. 0% SFC) would be

associated with a higher plasma TAG area under the curve compared to the system tempered to

contain droplets with crystalline fat (i.e. 33.2% SFC).

5.3 Materials and Methods

5.3.1 Study design

This study was a double-blind randomized cross-over acute meal study, where participants

attended two study visits separated by at least a six-day washout period. The treatment beverages

were emulsions at 37 C containing either liquid or partially crystalline droplets. All study

80

materials, including blood and plasma samples and data were identified by study visit number (1

& 2) and a 3-digit participant number, assigned based on order of enrollment, by a researcher not

directly involved in sample or data analysis. That same individual generated a random allocation

sequence for treatment order assignment, by flipping a coin and allocating participant numbers in

sequential order, i.e. 1-20. Participant allocation was concealed from participants and the lead

study coordinator who enrolled participants (after approval of the principal investigator), served

the treatments, and analyzed samples and data. To support this, both SE and LE emulsions were

prepared one day before each visit and the researcher who generated the treatment order

allocation covertly transferred the appropriate emulsion either from the 37 °C incubator or 5 °C

fridge to the 37 °C immersion precision cooker in the clinical centre (see below) on each study

visit morning. Unblinding to treatment was done after concluding sample and data analysis.

5.3.2 Participants

This study was approved by the University of Guelph Human Research Ethics Board, Ontario,

Canada (REB # 18-01-005, Appendix 1), registered at clinicaltrials.gov (NCT03515590), and

completed between May and October 2018. Written informed consent (Appendix 2) was

obtained from eligible participants prior to study commencement. Fifteen healthy male

participants were recruited from the University of Guelph and surrounding from Guelph

community through advertisements (Appendix 3) and emails through the University of Guelph.

The inclusion criteria were: generally healthy males from age 18-55 yrs, BMI 18-27 kg/m2,

blood pressure < 130/80 mm Hg, fasting plasma cholesterol level <5.2 mmol/L, plasma

triacylglycerol level <1.7 mmol/L and plasma glucose level <5.6 mmol/L, no history of any

major diseases or medical conditions, not taking any prescription medications, no antibiotic use

during the past 3 months or planning to take in the next 3 months, no significant weight change

81

(>10 %) during the past 3 months, not a regular smoker or user of recreational drugs, non to

moderate alcohol drinkers (< 5 drinks per sitting and <14 drinks total per week), not having any

allergies or anaphylactic reactions to food or other materials, not an elite or training athlete, no

sensitivity to the emulsion ingredients (palm stearin, span 60, the artificial sweetener sucralose

and artificial vanilla extract), or natural health products that interfere with postprandial lipid

metabolism. Participants were initially screened through a brief telephone interview (Appendix

4) and eligible individuals were invited for an in-person screening visit after a 12 h overnight

fast, for confirmation of eligibility, including paper questionnaires (Appendix 5) about health

history, dietary habits, physical activity, and dietary supplement use and determination of height,

weight, waist circumference, seated blood pressure, blood lipid profile and glucose by

fingerprick (Alere Cholestech LDX analyzer, Alere Inc., MA, USA) by a study coordinator.

5.3.3 Study visit protocols

Prior to each study visit, participants were asked to maintain their usual lifestyle habits, with

some changes in the 48 h period leading up to the study visit: refrain from strenuous physical

activities, consuming caffeine containing food and beverages, taking alcohol, any over-the-

counter medications or supplements. A standard low-fat pre-study frozen dinner meal, dessert

and evening snack were provided for participants to consume the evening prior to the study visit.

They were requested to not consume any food or drink thereafter but encouraged to drink plenty

of water throughout the fast, up to 30 min prior to consuming the treatment emulsion.

Participants arrived at the HNRU the following morning between 08:30-09:00 after a 12 h

overnight fast. After confirming compliance to protocol through a questionnaire, an intravenous

catheter was inserted into the participant’s forearm by a trained phlebotomist and a fasting blood

sample drawn into EDTA containing vacutainer tubes. Participants consumed one jar of test

82

emulsion (125 mL, all held at 37 C) at a time and completed drinking the total 500 mL (4 125

mL) within 10 min. Hourly blood samples were collected throughout the 6 h postprandial period.

No food or drinks, including water, were allowed during this period and participants remained

seated, except for short washroom breaks and an IV saline line maintained their hydration during

the postprandial period. After concluding each study visit day, participants were provided with a

post-trial sandwich meal of their choice. There were no instances of gastrointestinal upset

reported during the 6 h postprandial period. Twenty four hours after leaving each study visit,

participants completed check-in questionnaires to assess differences in gastrointestinal symptoms

associated with consumption of SE and LE; (1) Are you experiencing any feelings of illness or

discomfort after consuming the study treatments yesterday? If yes, please describe. (2) Did you

have any unusual gastrointestinal symptoms since the study visit yesterday? (i.e. nausea,

belching, bloating, gas, discomfort, etc.? If yes, please describe. (3) Did you have any bowel

movements after the study visit yesterday? If yes, how many? (4) Did you feel the bowel

movement was different from your regular bowel movements, i.e. sense of urgency, texture,

appearance, odour, etc.? (6) In particular, did you notice that your stools appeared to be fatty or

oily? If yes, please describe.

5.3.4 Emulsion preparation

Emulsion samples were always prepared the day prior to a study visit using 10 % (w/v) palm

stearin (Bunge Oils Inc., Bradley, IL, USA) and 0.4 % (w/v) sorbitan monostearate (Span 60;

Croda Canada Ltd., Vaughan, ON, Canada) in food grade DI water (obtained from the University

of Guelph Food Science Pilot Plant), as described previously (Thilakarathna & Wright, 2018).

Palmitic and oleic acids (~58 and 27wt%, respectively) and POP (26.1 0.1%), PPP and POO

83

(23.1 0.1% and 13.8 0.1wt%, respectively) were the main FAs and TAGs present in the PS

(Thilakarathna & Wright, 2018). To prepare 500 mL of the emulsion, 50 g of PS and 2.0 g of

Span 60 were melted at 80 °C for 30 min. A coarse emulsion was formed by combining this with

80 °C food grade deionized water (DI) sweetened with 2 g of Sucralose (Splenda) and flavored

with 5 drops of artificial vanilla (Club House) and mixing with a hot handheld homogenizer at

12,000 rpm for 1 min (Ultra Turax, Ika T18 Basic, Germany). This coarse emulsion was then

transferred to the microfluidizer (M-110EH, Microfluidics, MA, USA) hot hopper and passed 5

times at 125 MPa with the piping immersed in a 95 °C water bath. The emulsion was portioned

into four glass jars with ~125 mL emulsion in each. For LE, the hot homogenate was transferred

to 4 glass jars warmed to 80 °C and the samples directly placed in an incubator set at 37 °C. To

prepare SE, the hot homogenate was transferred to 4 glass jars previously cooled at 5 °C and then

immediately placed in an ice-water bath for 20 min before storage in a 5 °C fridge until the

following morning. On each study day morning, the glass jars containing either LE or SE were

transferred from either the incubator or fridge to a water bath maintained at 37 °C using an

immersion precision cooker (Anova Culinary, CA, USA) and held for 60 min (i.e. sufficient time

for the SE samples to reach 37 °C, data not shown). Strict temperature control was maintained

throughout sample preparation, storage, and analysis, including warming utensils to 37 °C and

minimizing temperature fluctuations in order to maintain the intended differences in physical

state. Particle sizing and physical state were confirmed routinely by differential scanning

calorimetry and laser diffraction, as below. Whereas LE was also liquid prior to consumption, SE

contained 33.20.1% solid fat at 37 C, with a melting temperature of 53.40.2 C.

5.3.5 Emulsion droplet characterization

84

LE and SE samples were characterized for the physical state, particle size, and particle charge

according as described previously (Thilakarathna & Wright, 2018). In brief, particle size

distributions, volume-weighted (D4,3) and surface-weighted (D3,2) mean diameters were

determined by laser diffraction using a Mastersizer 2000S (Malvern Instruments Inc.,

Southborough, MA, USA) and zeta potential determined after emulsion dilution 1:250 with DI

water using a particle electrophoresis instrument (Zetasizer Nano ZS, Malvern Instruments Inc.,

Southborough, MA, USA). For both techniques, values of 1.45 and 1.33 were used as the

refractive indices of PS and water, respectively. Melting and crystallization behaviors were

analyzed by differential scanning calorimetry (DSC, Q2000 model (TA Instruments,

Mississauga, ON, Canada). In brief, 5-10 mg of emulsion was weighed and sealed in pre-

weighed alodined aluminum DSC pans (TA Instruments, Mississauga, ON, Canada). Samples

were held in the DSC at 37 °C for 3 min, then heated at 5 °C/min to 80 °C, held for 3 min and

then cooled at 5 °C/min to 0 °C. To confirm complete crystallization, SE samples at 37 °C were

cooled from 37 to 0 °C at 5 °C/min. Enthalpies and peak onset and maximum temperatures were

determined using the system software (TA Instruments Universal Analysis 2000 software, TA

Instruments, ON, Canada).

5.3.6 Blood sample collection and analysis

Blood samples were collected into EDTA-containing vacutainer tubes (Becton Dickinson and

Company, NJ, USA) and immediately centrifuged at 4 C for 10 min at 3000 rpm. Separated

plasma was aliquoted and frozen at -80 C for later biochemical analysis, within 6 months. Blood

samples were drawn every hour (0-6 h) for plasma TAG and NEFA analysis and every 2 hours

(0, 2, 4, & 6 h) for analyzing the size and FA composition of the chylomicron-rich fraction

85

(CMRF). To separate the CMRF, 3 mL of separated plasma was layered with 2.5 mL 1.006 g/mL

saline solution and ultracentrifuged (Sorvall WX Ultra 80, ThermoFisher Scientific, NC, USA)

for 23 min at 4 C at 65,000 rpm (relative centrifugal force 3.9105 g at rmax, Beckman Coulter,

IN, USA). Afterwards, the top cloudy layer was removed and immediately analyzed for CMRF

size and a portion was frozen at -20 C for future determination of FA composition by gas

chromatography.

Plasma TAG (TAG assay kit, Wako Chemicals USA Inc., Richmond, VA, USA) and NEFA

(NEFA-HR(2)), Wako Diagnostics, Richmond, VA, USA) were analyzed using colorimetric

assay kits according to manufacturer instructions. The intra- and inter-assay variabilities (CV%)

for TAG and NEFA were 4.4 and 8.1 % and 3.7 and 8.5%, respectively. CMRF particle size was

measured by laser diffraction (Zetasizer Nano ZS, Malvern Instruments Inc., Southborough, MA,

USA). The frozen CMRF was thawed at 5 C and, after separating the TAG fraction by thin

layer chromatography, the FA composition of the TAG fraction analyzed by GC (Thilakarathna

et al., 2016). FA composition is reported as a relative percentage of the total FA present.

5.3.7 Statistical analysis

A sample size calculation determined the need for 13 participants to complete the study based on

90 % power and p<0.05 to detect a minimum peak postprandial plasma TAG concentration

difference of 0.1mmol/L by a pairwise comparison with a 0.1 mmol/L SD of differences using

G*Power software (Version 3.1.9.3., Universität Düsseldorf, Germany). This is consistent with

previous similar study designs (Berry et al., 2007; Berry et al., 2007; Hall et al., 2017). To allow

for the possibility of a 10% dropout rate, 15 participants were sought, although 17 were initially

86

enrolled. 15 participants finished the study and data from all 15 was included in all analyses,

unless otherwise stated.

Data was analyzed using SPSS software (version 25, IBM Corp., NY, USA) and expressed as

mean SEM unless otherwise specified. Postprandial plasma TAG change, incremental area

under the curve (iAUC), and plasma NEFA change were calculated after adjusting for individual

participant fasting (baseline) values at each study visit. Postprandial changes in CMRF size and

FA composition of the CMRF TAGs were calculated from baseline. All data were normally

distributed according to the Shapiro-Wilk normality test. Statistical analysis of the postprandial

changes from baseline for plasma TAG, NEFA, CMRF size and FA composition of the TAG

fraction were performed using general linear model (GLM) for repeated measures analysis of

variance (ANOVA) with emulsion droplet physical state and time as within subject factors. The

Greenhouse-Geisser correction was applied for data that did not satisfy the sphericity assumption

and Bonferroni adjustment was applied for post hoc analyses. iAUC values for plasma TAG and

NEFA were calculated using GraphPad Prism software (version 5.0a, San Diego, CA, USA)

using the trapezoidal rule and iAUC differences between LE and SE explored using pairwise t-

tests. Unless otherwise stated, pairwise testing between SE and LE was bi-directional, i.e. 2-

tailed. However, 1-tailed t-tests were used to compare TAG response, given the main hypothesis

and evidence that lipolysis was attenuated for SE versus LE. CMRF-TAG changes in terms of

the proportion of each major FA present in the emulsion (i.e. palmitic, oleic, linoleic and stearic

acids) were analyzed using one-way ANOVA with Bonferroni multiple comparison test.

Pearson’s chi squared test was performed to study the associations between post-study visit

gastrointestinal symptoms for LE and SE. A cut-off level of significance of P<0.05 (two-tailed)

87

was used for all the tests and trends were discussed when P values fell between 0.05 and

<0.10.

5.4 Results

5.4.1 Participant characteristics

The flow of participants through the study is shown in the Figure 5.1. From 49 phone-screened

participants, 29 were screened in-person. Of the 20 eligible to continue, 3 declined to participate

and treatment randomization was allocated for 17 individuals. Subsequently, 2 participants

withdrew from the study, i.e. one declined to continue even before the first study visit and the

other withdrew due to difficulties in obtaining the fasting blood sample at the first study visit.

Table 5.1 shows the fasting baseline characteristics of the 15 healthy male participants who

completed the study.

88

Figure 5.1 Participant flow through the study

Completed the study and data

available for analysis (n=15)

Expressed initial interest (n=121)

Did not follow up with

study coordinators (n=72)

Phone Screened (n=49)

In-person screened (n=29)

Eligible to participate (n=20)

Assigned as study participants and

treatment randomization allocated (n=17)

Did not meet inclusion criteria (n=20)

Did not meet inclusion criteria for

fasting blood criteria (n=9)

Declined to participate (n=3)

Excluded (n=2)

(declined to participate n=1,

difficulty to draw blood n=1)

En

roll

men

t A

lloca

tion

F

oll

ow

up

an

d

An

aly

sis

89

Table 5.1 Characteristics of the 15 healthy male participants at baseline1

Mean SD

Age (yr) 27.5 5.7

Height (cm) 175.3 6.1

Weight (kg) 73.9 7.7

BMI (kg/m2) 24.1 2.5

Waist circumference (cm) 84.8 6.0

Diastolic blood pressure (mm Hg): 74.2 7.9

Systolic blood pressure (mm Hg) 120.3 8.2

Total cholesterol (mmo/L) 4.4 0.6

Triacylglycerols (mmol/L) 0.9 0.3

LDL (mmol/L) 2.8 0.8

HDL (mmol/L) 1.1 0.2

Glucose (mmol/L) 5.0 0.4

5.4.2 Test emulsion characteristics

The melting and crystallization behaviour, emulsion droplet size and charge of the palm stearin-

Span 60 in water emulsions after the addition of the artificial sweetener and flavor were similar

to those reported for without the flavouring, as extensively detailed previously (Thilakarathna &

Wright, 2018). LE did not show any melting event when heated from 37 to 80 °C, indicating the

absence of any crystallinity. SE had a peak melting event at 53.39 ± 0.35 °C (Figure 5.2a&b). It

was confirmed to be maximally crystalline by cooling the sample from 37 to 0 °C and not

observing any further TAG crystallization. The SFC of bulk palm stearin (Figure 5.2d) was

33.17±0.05 % at 37 °C. This is consistent with the melting enthalpies observed (i.e. 4.5±0.5 J/g

for SE and 44.7±4.7 J/g for bulk PS) given that SE is a 10% oil in water emulsion. LE and SE

had monomodal size distributions and perfectly overlaid each other, centering around 416 nm

90

(Figure 5.2c). The surface weighted (LE: 0.36 ± 0.00 and SE: 0.35 ± 0.00 m, P=0.32) and

volume weighted (LE: 0.50 ±0.01 and SE: 0.47 ± 0.02 m, P=0.13) mean diameters and

potential values (LE: 54.9±0.7 SE:53.0±0.4 mV, P=0.41) were also statistically similar.

Figure 5.2 Characteristics of the undercooled liquid (LE) and crystalline solid (SE) emulsions.

Representative DSC thermograms of LE (a) and SE (b) showing the melting behavior and

particle size distributions (c) and the solid fat content of bulk palm stearin (d).

5.4.3 Postprandial plasma TAG responses

Fasting plasma TAG concentrations (baseline, t=0) were similar between both study visits

(P=0.57). Postprandial changes in plasma TAG concentration and iAUC from fasting are shown

0 20 40 60 80

-0.4

0.0

0.4

Temperature (°C)

Hea

t F

low

(W

/g)

0 20 40 60 80

-0.4

0.0

0.4

Temperature (°C)

53.39 ± 0.35 °C

0 403710 20 30 50 600

20

40

60

33.17±0.05 %

Temperature (°C)

Soli

d f

at c

onte

nt

(%)

0.01 0.1 1 10 100 1000 100001000000

2

4

6

8

10

LE SE

Particle size (mm)

Dis

trib

uti

on %

(a) (b)

(c) (d)

91

in Figure 5.3. Accordingly, plasma TAGs increased significantly with time (Ptime=0.001) and

trended towards SE being low compared to LE (Pphysical state=0.08) and there was no interaction

between physical state and time (Pphyical state x time =0.77). iAUC values for plasma TAG change

from fasting were statistically different (P=0.03 and P=0.06 for the one and two-tailed

comparisons, respectively) based on physical state (1.950.39 and 1.450.31 mmol/L*h, for LE

and SE, respectively. Overall, Figure 5.3 shows plasma TAG peak around 4 hours for both LE

and SE. Indeed, values of mean TAG time-to-peak for LE (4.60.32 h) and SE (4.70.36 h) were

similar (P= 0.36 and P=0.72 for the one and two-tailed comparisons, respectively). However,

based on individual TAG peak concentrations (ranged between 2-6 h), LE was higher than SE

(1.470.19 and 1.200.15 mmol/L, respectively; P=0.036 and P=0.07 for the one and two-tailed

comparisons, respectively). Trends observed for the change from baseline values for plasma

TAG between 1-6 h (P=0.08) and 2-6 h (P=0.07) consistently support that there was a lower

plasma TAG increase for SE versus LE. Comparisons of iAUC values between LE and SE

during 1-6 h also indicated statistical differences between LE and SE (i.e. P=0.03 and P=0.05 for

iAUC1-6h one and two-tailed comparisons, respectively). Furthermore, compared with LE, there

was a delay in terms of when SE plasma TAG concentration deviated from baseline, i.e. LE

increased significantly from baseline at 2 h (T0h:0.68 mmol/L, T2h:0.91 mmol/L, P=0.025 and

P=0.049 for one and two-tailed comparisons, respectively), whereas SE only deviated

significantly from baseline at 3 h (T0h:0.62 mmol/L, T3h:0.90 mmol/L, P=0.006 and P=0.012 for

one and two-tailed comparisons, respectively). This supports the hypothesis of delayed plasma

TAG response with consumption of lipids in the solid state. All in all, increases in plasma TAG

and iAUC values were higher for LE compared to SE. In every comparison, the hypothesized

attenuations were observed for lipids in the partially solid versus undercooled liquid state at

92

either a 95 or 90% confidence interval. The observed differences, although small, point to a role

for TAG physical state in determining postrprandial TAG response, given that LE and SE were

identical in all ways other than SFC, i.e. composition and colloidal properties. All results indicate

a significant attenuating effect on plasma TAG response when solid fat was present. This is

despite the similar colloidal properties and composition and is in line with the previous in vitro

evidence for the same emulsion system (Thilakarathna & Wright, 2018).

Figure 5.3 (a) Mean change in plasma TAG concentration from baseline and (b) TAG iAUC0-6h

and iAUC2-6h after participants consumed 50 g of 10 % palm stearin and 0.4% span 60 in-water

emulsion either as undercooled liquid (LE) or crystalline (SE) emulsion droplets. Error bars

represent SEM. Change from fasting plasma values were analyzed by general linear model for

-1 1 2 3 4 5 6

-0.2

0.2

0.4

0.6

0.8LE SE

Time (h)Chan

ge

in P

lasm

a T

AG

(m

mol/L

)

0

1

2

3LE SE

iAUC 0-6 h iAUC 1-6 h

P=0.034P=0.030

iAU

C P

lasm

a T

AG

(m

mol/L

*h)

(a)

(b)

93

repeated measures ANOVA with emulsion droplet physical state (LE and SE) and time (0-6 h) as

within subject factors: Pphysical state=0.08, Ptime=0.001, Pphysical state time=0.77. The P values stated in

the iAUC are one-tailed. The two-tailed P value for TAG iAUC0-6h and iAUC1-6h were P=0.06

and P=0.07, respectively. TAG: triacylglycerol

5.4.4 Postprandial plasma NEFA response

The postprandial plasma NEFA responses after consuming the two emulsions are shown in

Figure 5.4. There was a significant effect of time (P<0.0001), but no statistical difference based

on droplet physical state (P=0.72) or a physical state time interaction (P=0.88). There was a

visible difference in plasma NEFA between LE and SE up to 3 h, i.e. it decreased further below

baseline for LE than SE, but this was not statistically different (P=0.82).

Figure 5.4 Mean change in plasma NEFA concentration change from baseline after 15

participants consumed 50 g of 10 % palm stearin and 0.4% span 60 in-water emulsion with either

undercooled liquid (LE) or crystalline (SE) emulsion droplets. Error bars represent SEM.

Change from fasting plasma values were analyzed by general linear model for repeated measures

0 1 2 3 4 5 6-0.1

0.0

0.1

0.2

0.3

LE SE

Time (h)

NE

FA

conce

ntr

aton (

mm

ol/L

)

94

ANOVA with emulsion droplet physical state (LE and SE) and time (0-6 h) as within subject

factors: Pphysical state=0.72, Ptime <0.0001, Pphysical state time=0.88.

5.4.5 Changes in postprandial CMRF

Figure 5.5 (a-d) shows the proportional changes in the FA isolated from the CMRF-TAG

fraction from fasting and at 2, 4, and 6 h postprandially. This is based on the major FA present in

palm stearin, i.e. palmitic (57.50.5%), oleic (27.20.2%), linoleic (6.50.1%) and stearic

(5.60.2%) acids. There was no effect of emulsion droplet physical state on the proportional

changes between LE and SE (P>0.05). There were, however, significant time effects observed

for palmitic and linoleic (P<0.0001) acids, but not oleic (P=0.09) or stearic (P=0.40) acids. The

proportions of palmitic and linoleic acids in the original emulsion and in the fasting CMRF-TAG

were significantly different (P<0.05), but the proportions of oleic and stearic acids were similar

(P>0.05). By 4 h postprandially, the plasma CMRF-TAG FA composition shifted towards that of

the emulsion (Figure 5.5e). The change in proportion of stearic acid was not statistically

significant (P=0.92).

CMRF particle size increase significantly (P<0.0001) overtime, peaking by the postprandial

plasma TAG peak of ~4 h, with similar size changes observed between the emulsions (P=0.79)

and no physical state x time interaction (P=0.89).

95

Figure 5.5 CMRF mean concentration of (a) & (c) saturated and (b) & (d) unsaturated fatty

acids, (e) proportional FA composition in the initial emulsion, CMRF-TAG fraction at fasting

state and 4 h postprandial consuming LE and SE, and (f) CMRF particle size change through the

6 h postprandial period. Error bars represent SEM. Differences between the proportion of FAs

in the CMRF-TAG fraction and size at 0, 2, 4, and 6 h were separately analyzed by general linear

model for repeated measures ANOVA with emulsion droplet physical state (LE and SE) and

time (0-6 h) as within subject factors. In (e), for FA where P<0.05, columns having different

0 2 4 625

30

35

40

45

LE SE

Pphysical state =0.65

Ptime <0.0001

Pphysical state ´ time=0.51Rel

ativ

e co

nce

ntr

atio

n (

%)

0 2 4 62.5

5.0

7.5

Pphysical state =0.54

P time =0.40

Pphysical state ´ time=0.30

Time (min)

Rel

ativ

e co

nce

ntr

atio

n (

%)

0 2 4 625

30

35

40

Pphysical state =0.95

Ptime =0.09

Pphysical state ´ time=0.25Rel

ativ

e co

nce

ntr

atio

n (

%)

0 2 4 610

15

20

Pphysical state =0.85

Ptime <0.0001

Pphysical state ´ time=0.21

Time (min)

Rel

ativ

e co

nce

ntr

atio

n (

%)

(a) C16:0 (b) C18:1

(c) C18:0 (d) C18:2

Emulsion Fasting LE at 4 h SE at 4 h0

20

40

60 C16:0 C18:1

C18:0 C18:2

A

B

CC

ab

aab

b

x

yz xz

Rel

ativ

e co

nce

ntr

atio

n (

%)

0 2 4 6

60

80

100

120

Pphysical state =0.79

Ptime <0.0001

Pphysical state ´ time=0.89

Time (h)

Par

ticl

e si

ze (

mm

)

(e) (f)

96

letters for the same FA indicate a significant difference at 95% CI (Bonferroni multiple

comparisons test).

5.4.6 Post-study visit gastrointestinal and bowel habits

Participants’ self-reported gastrointestinal habits for the 24 hours following each study visit were

compared to explore any associations between lipid physical state and gastrointestinal tolerance

(Figure 5.6). In general, gastrointestinal tolerance for both samples was high. No incidence of

any illness was reported, at either study visit or during the 24 hour periods after leaving the

HNRU. This supports the suitability, in general, of utilizing these high fat emulsion meals to

investigate postprandial metabolism. None of the comparisons were statistically significant (i.e.

gastrointestinal symptoms (2(1)=1.007, P=0.32), number of bowel movements (2(1)=2.81,

P=0.42), bowel habits (2(1)=3.39, P=0.07) and stool texture (2(1)=2.2, P=0.14). However, it

was surprising to see qualitative differences between SE and LE, considering the treatments were

identical in composition. 24 h after consuming LE and SE, 3 (e.g. gas and bloating) versus 1 (e.g.

belching and gas) participant reported any GI symptoms (Figure 5.6a). There were also

differences observed between LE and SE in terms of bowel habits (Figure 5.6b, c) and stool

texture (Figure 5.6d). 13/15 participants had a bowel movement (10, 2, and 1 participants had 1

2, and 3 bowel movements, respectively) after consuming LE and 31% of these reported bowel

movements were described as being “different” from usual. Specifically, 4 participants reported

a greater sense of urgency or unusually fatty stools and 7 reported differences from usual stool

texture (i.e. 6 reporting more oily and softer stools, and one reporting stool as being firmer). In

contrast, after consuming SE, only 9 participants reported having a bowel movement (8 and 1

participants had 1 and 2 bowel movements, respectively) and these were reported to be similar to

97

usual, with no unusual symptoms. From the 9 participants that reported bowel movements after

SE, 7 reported the stool texture was similar to their normal stools. Only two participants reported

different stool textures with SE, i.e. one reported softer and one reported firmer stool. Again,

while not statistically significant, these trends in gastrointestinal symptoms are particularly

interesting given that participants consumed the exact same amount and type of lipid on both

occasions.

Figure 5.6 Self-reported gastrointestinal symptoms and bowel habits for the 24 h period after

participants’ each study visit. (a) Did you have any unusual gastrointestinal symptoms since the

study visit yesterday? (b) Did you have any bowel movements after the study visit yesterday? If

yes, how many? (c) If yes to b, did you feel the bowel movement was different from your regular

bowel? (d) If yes to b, did you notice that your stools appeared to be fatty or oily? For LE, n=15

and SE, n=14.

Abs

ent

Prese

nt

Abs

ent

Prese

nt0

5

10

15LE SE

Num

ber

of

par

tici

pan

ts

Usu

al

Diff

eren

t

Usu

al

Diff

eren

t0

5

10

15

Num

ber

of

par

tici

pan

ts

0 1 2 3 0 1 2 30

5

10

15

Nor

mal

Oily

Firmer

Nor

mal

Oily

Firmer

0

5

10

15

(a) (b)

(d)(c)

c2(1)=1.007

P=0.32

c2(1)=2.81

P=0.42

c2(1)=3.39

P=0.07

c2(1)=2.2

P=0.14

98

5.4.7 Correlation between previous in vitro and human study results

Correlation analysis was performed between the previously reported (Thilakarathna and Wright,

2018) 4 hr in vitro lipolysis results and the current in vivo 1-4 h postprandial plasma TAG

concentrations for LE and SE. For these purposes, analysis was performed both on the pooled

(LE and SE) data and for these emulsions independently. Accordingly, the pooled LE and SE in

vitro and in vivo data showed a significant positive linear correlation (Figure 5.7 (a), r2=0.86,

P=0.007). When the emulsions were separately analyzed, LE showed a positive non-statistically

significant correlation (r2=0.81, P=0.19) and SE showed a significantly strong positive linear

correlation (r2=0.96, P=0.046) between the in vitro and in vivo data.

99

Figure 5.7 Correlation analysis between in vitro lipolysis (%, n=3) and baseline adjusted plasma

TAG concentrations (mmol/L, n=15) (a) for both LE and SE as a group (b) LE and SE

separately, during 1-4 h in vitro and in vivo digestions. Error bars represent SEM.

5.5 Discussion

The objective of this study was to investigate the impact of undercooled liquid versus partially

crystalline emulsion droplets on postprandial lipemia. The emulsion systems studied were unique

40 50 60

0.0

0.2

0.4

0.6

Pearson r=0.86 P=0.007 B

asel

ine

adju

sted

pla

sma

TA

G c

once

ntr

atio

n (

mm

ol/L

)

40 50 60

0.0

0.2

0.4

0.6

LE SE

Pearson r 0.81 0.96 P 0.19 0.046

In vitro lipolysis (%)

Bas

elin

e ad

just

ed

pla

sma

TA

G c

once

ntr

atio

n (

mm

ol/L

)

(a)

(b)

100

in having identical compositions and other particle properties investigated, effectively isolating

the effect of droplet SFC on postprandial lipemia. To our understanding, this is the first-time

such a study has been pursued and also the first report of in vivo effects of an undercooled

emulsion system. Given the similarities between the systems and efficiencies of human

digestion, large differences in postprandial lipemia or digestibility were not anticipated. As such,

we were open to accepting any differences in postprandial lipemia as evidence that SFC plays a

role in postprandial lipemia. We also considered that, although P<0.05 is an important basis for

our statistical inference, there is rationale for not strictly adhering to the P<0.05 cut off, as

statistical significance is not equivalent to scientific significance (Wasserstein & Lazar, 2016).

However, there were several results pointing to an attenuated lipemic response for SE

specifically supporting the hypothesis of attenuated response for TAGs in the crystalline state,

both at the 0.05 and 0.1 significance levels. Overall, the data point to small, but statistically

significant differences and trends between LE and SE.

Beyond statistical significance, is the matter of biological significance. In terms of TAG

response, a postprandial peak plasma TAG increase of ~0.3-0.5 mmol/L is considered to be

clinically significant based on increasing the hazard ratios for coronary heart disease mortality in

men from the Norwegian Counties Study (Lindman, Veierød, Tverdal, Pedersen, & Selmer,

2010). The mean plasma TAG peak at 4 h was 0.2 mmol/L lower for SE than LE. When

individual plasma TAG curves were considered, the peak plasma TAG concentration occurred

between 2-6 h such that the mean value between SE and LE differed by 0.27 mmol/L. This is

less than but approaching the 0.3 mmol/L attenuation considered clinically significant. It

supports the fact that, by manipulating the SFC in a dietary lipid, even without significantly

altering the composition, it is possible to reduce postprandial lipemia. Furthermore, the reduction

101

in iAUC0-6h for SE was in the order of ~26% and approaching the level considered clinically

relevant, i.e. a ~30% reduction (Hall et al., 2014). The amount of fat in this study, i.e. 50 g, is

considered to be a moderate fat load (30-50 g of fat in a single meal) within the range of where a

dose dependent increase in postprandial lipemia is expected (Lairon et al., 2007) and contributes

to a notable acute postprandial lipemic effect (Sanders, 2003). Similar fat loads have been used

in previous studies with larger differences observed between treatments in plasma TAG response

than seen in the present study. This is likely related to the very nuanced differences between SE

and LE. The test fats studied previously were mainly IE and non-IE fats and blends and

therefore, differences in both composition and physical state would have contributed to

differences in plasma TAG. Further, some IE lipid studies that have reported larger differences in

plasma TAG between different SFC treatments utilized a much higher fat load (Hall et al., 2014).

Elevated postprandial plasma TAG levels are recognized as a risk factor for developing CVD

(Kolovou et al., 2011). It remains to be established whether the extent of postprandial lipemia,

peak plasma TAG concentration or the duration, i.e. extended postprandial lipemia period, is

more detrimental as the greatest risk factor for CVD (Hall et al., 2014). Previous literature

supports that liquid lipids tend to show earlier onsets of postprandial lipemia and higher

postprandial lipemic responses of relatively short duration, compared to solid lipids which show

delayed onset and lower, but extended postprandial lipemic response. Specifically, this was

observed in studies comparing IE palm oil with 15.2% SFC vs high oleic sunflower oil with 0%

SFC (Berry et al., 2007) and IE vs non-IE palm stearin palm kernel oil blend with 24 and 21 %

SFC at 35 C, respectively (Hall et al., 2017). In agreement, LE trended towards an earlier and

exhibited a higher postprandial lipemia peak pointing to more rapid and greater digestibility of

liquid lipids compared to those in the solid state. This is in agreement with the in vitro digestion

102

results for the same systems (Thilakarathna and Wright, 2018). Although studying the individual

postprandial lipemia curves indicates that both systems had similar time to peak values (i.e. ~4.6-

4.7 h), the overall change in plasma TAG as observed in the Figure 5.3a pointed to extended

postprandial lipemia for SE. This difference in absorption could be mainly attributed to the lower

and slower digestibility of solid fat (Berry et al., 2007). The fact that SE remained elevated at 6 h

also underscores the importance of considering a longer investigational period (Lairon et al.,

2007), especially for higher melting lipids.

We previously reported that colloidal destabilization and apparent partial coalescence of SE

occurred during exposure to simulated gastric conditions and that this was associated with

retarded TAG digestibility because of decreased surface area (Thilakarathna and Wright, 2018).

As such, the presence of SE was associated with reductions in both the amount of interface

available for lipolytic activity, as well as a solid interface that may have interfered with lipase

accessibility as previously shown for compositionally equivalent undercooled and partially

crystallized droplets (Huynh & Wright, 2018). In line with these in vitro observations, the lower

in vivo digestibility observed for SE in this study was anticipated. Apart from altered lipolysis in

SE, the colloidal instability evident during the in vitro gastric conditions was expected to occur,

to some extent, in vivo, potentially delaying gastric emptying. This could offer another

explanation to the observed delay in the plasma TAG increase after SE and is under

investigation, separately. Under in vivo gastric conditions, extensively destabilized and phase

separated solid fat containing emulsions systems reportedly emptied the liquid portion faster and

showed a blunted and delayed increase in plasma TAG in vivo (Steingoetter et al., 2015).

The major limitation of this study is the fact the liquid state in LE was thermodynamically

unstable and therefore susceptible to crystallization. Our previous in vitro findings showed that

103

acid in the gastric phase induced partial crystallization in LE but not dilution or shear conditions.

After gastric exposure, the SFC of LE increased from 0 up to ~15% (of the lipid phase). This

contrasted with the 33% SFC present in the SE lipids (Thilakarathna & Wright, 2018).

Therefore, while we anticipate there was some degree of LE crystallization in the gastrointestinal

tracts of participants in the human study, it is impossible to know to what extent. Even between

individuals, fasting gastric pH level can vary considerably, e.g. ranging from a pH 1-3 (Maio &

Carrier, 2011). Moreover, our previous work on the same emulsions showed gastric phase shear

conditions exaggerated the difference between LE and SE lipolysis and reduced overall lipolysis

compared to when shear was not applied (Thilakarathna & Wright, 2018). Therefore, while

partial LE crystallization was expected during digestion, it is a limitation that we cannot

conclude to what extent this occurred.

There were no differences observed between LE and SE in terms of chylomicron size. Size

increases followed a similar pattern, peaking at 4 h for both LE and SE. This is indicative that LE

and SE transported similar fat loads during the postprandial period. Lipoprotein remodeling

during circulation, especially chylomicrons, resulting in small dense chylomicron remnant

particles (Martins, Mortimer, Miller, & Redgrave, 1996) is a risk factor for CVD because of

associated prolonged residence in the circulation and reduced liver uptake. CMRF particle

number was not determined in this study. Previously rates of chylomicron clearance were

reportedly similar for fats with different SFCs as measured by the lipoprotein lipase and hepatic

lipase activities, indicating digestibility and rate of absorption played a larger role than the rates

of chylomicron clearance from circulation (Berry et al., 2007).

In agreement with previous reports, the FA composition of the CMRF-TAG was altered from

fasting and reflected that of the ingested lipid by 2 h, and especially by 4 h. Comparing between

104

fasting and 4 h postprandially, significantly more palmitic was present in plasma after the

treatments. Interestingly, with SE, by 4h, the proportion of oleic acid present was significantly

lower than at fasting. This could be related to clumping of solid fat during the gastric phase

possibly trapping liquid oil present and limiting lipase access, thereby, making oleic acid

(concentrated in the liquid fraction) less available for absorption. In the present study, the oleic

acid proportion was similar to that of fasting levels by 6 h postprandial indicating possible

digestion breakdown of the larger solid fat clumps (data not shown).

As a secondary outcome measure, participant gastrointestinal symptoms were monitored after

each study visit. Interestingly, no unusual bowel habits were reported by participants after

consuming SE, although this was not the case for LE. With LE, participants experienced more

gas, belching and fatty, oily stools. Steatorrhea refers to the excretion of excess amounts of

liquid oil in the stool and this can be associated with a sense of urgency and fatty stools (Levy,

2015). However, whether this applies equally to liquid oils or solid fats is unclear. The trends

between SE and LE do support the theory that, when solid fat is excreted, it may go “unnoticed”

without inducing unusual bowel habits or symptoms (Kaplan & Greenwood, 1998). Although

lipid digestion in healthy adult humans is considered efficient and complete, few studies have

specifically addressed whether lipids with higher melting temperatures are excreted in greater

proportions. In his review, Livesey (2000) argued that poorer digestibility of tristearin based on

its high melting temperature (73 °C) and associated crystallinity in the gastrointestinal tract

explained reportedely poor absorption for some stearic acid-rich fats which had previously been

attributed to higher dietary intakes. In summary, although minor, there were differences in the

gastrointestinal habits reported by participants after consuming SE and LE that are in agreement

with speculation that the presence of crystallinity, which can attenuate digestibility and

105

absorption independent of compositional differences, may be associated with undetected lipid

excretion compared with liquid oil.

Very few studies have attempted to correlate in vitro and in vivo findings and even fewer studies

have made comparisons based on the same test fat systems. Overall, the in vitro percent lipolysis

and the in vivo plasma TAG increases over time were strongly positively correlated signifying

the relatability of in vitro and in vivo studies and offering consistent interpretations among

studies. When LE and SE were considered separately, LE did not show a linear relationship

while SE did. One limitation of static digestion models is the lack of digestion product removal,

such that the reactions reach a lipolysis plateau and this is not consistent with human physiology.

Further, when careful attention is paid, postprandial plasma TAG change in SE had not reached a

plateau by 6 h (Figure 5.3a) supporting the hypothesis that the larger clumps of solid fat is

broken down slower than the dispersed but flocculated LE droplets. This study was not extended

beyond 6 h which is a limitation and if continued up to 8 h, could potentially show whether SE

plasma TAG plateaued and/or dropped.

In conclusion, compositionally identical undercooled liquid and partially crystalline emulsion

droplets showed minor differences in postprandial TAG responses when ingested by healthy

men, pointing to a specific influence of SFC on postprandial lipemia. The results indicate

differences in digestibility due to differences in the SFC, possibly partly related to gastric phase

partial coalescence based on our previously published in vitro observations with the same

emulsion systems. The presence of solid fat had an attenuating effect on postprandial lipemia.

Although minor, this brings new perspective to appreciating how saturated fats with melting

temperatures above 37 °C may contribute to disease risk and how SFC can be tailored for

metabolic response, warranting further exploration.

106

107

CHAPTER 6

INTEGRATED DISCUSSION

108

6.1 Preamble

Excessive dietary lipid consumption has been linked to the increased risk of developing

numerous chronic diseases and conditions, including CVD, type II diabetes, metabolic

syndrome, and obesity. Moreover, atherosclerosis is now recognized as a postprandial

phenomenon and postprandial lipemia, rather than fasting plasma TAG levels, is now considered

a better independent risk factor for CVD. The amount of fat consumed, lipid composition, e.g.

saturated vs unsaturated fat, and lipid structure and form, e.g. bulk vs. emulsified lipids, are main

contributors to postprandial lipemia. However, recent incidental findings have also pointed

towards possible contributions from the associated lipid physical state differences and renewed

attention to the topic. To-date, understanding of the specific relevance of SFC for lipid

digestibility and postprandial metabolism is limited and there are discrepant results and opinions

among experts. A better understanding of lipid digestive processes, digestibility, absorption, and

postprandial metabolism, in relation to how lipid physico-chemical properties determine these

physiological processes, is needed. This is particularly true given the complex interplay now

recognized between lipid structure, digestibility, inflammation, i.e. postprandial endotoxemia,

and chronic disease risk.

This thesis, overall, sought to investigate the contributions of TAG physical state, i.e. solid fat

content, to digestive processing and postprandial metabolism, focussing mainly on IE bulk lipids

and tempered emulsion droplets, using multilevel research tools. These two lipid systems were

targeted because of their relevance to the food supply. Moreover, these two lipid test systems

have equivalent or similar compositions which helps to eliminate possible confounding

influences that might otherwise obscure the effects of TAG physical properties. It was

109

hypothesized that the presence of more solid fat would, in general, reduce lipid digestibility and

result in a slower and/or lower release of FFAs compared to the lipids with less solid fat.

6.2 Summary of major findings, strengths and limitations

Study 1 (Chapter 3) investigated the in vitro lipolysis of lipids with equivalent fatty acid

compositions, but different TAG species and SFCs at 37 C, i.e. a stearic acid-rich non-IE lipid

blend and its interesterified versions. This was done using a dynamic gastrointestinal TIM-1

digestion model and the findings were compared to those from a previous human postprandial

lipemic study of the same lipid blends (Robinson et al., 2009). The objectives were to investigate

the role of SFC in impacting the digestibility to support a mechanistic understanding and to

explore in vitro-in vivo correlations. The findings from that study agreed with the hypothesis that

the non-IE lipid blend with a higher SFC would show a slower FFA release and a lower FFA

bioaccessibility compared to the IE lipid blends and positively correlate with the human study

findings. The samples containing tristearin were also found to remain undigested in the TIM-1

effluent, supporting that high melting TAGs resist lipolysis. To the best of our knowledge, the

physical properties of IE lipids have never been studied under such in vitro conditions and

studies exploring in vitro-in vivo correlations of the same lipid systems are very rare, but

necessary, both to validate the use of in vitro models and in vitro findings. Therefore, this work

contributes to an understudied area and supports that SFC changes with interesterification

contribute to differences in digestibility and subsequently absorption. A limitation of the work

was that the test lipids were equivalent in terms of fatty acids, but differed in TAG species, as

well as physical state.

110

Moving forward, study 2 (Chapter 4) investigated the in vitro digestibility of emulsions with

identical compositions and similar particle characteristics, but different SFCs, i.e. undercooled

liquid and crystalline solid emulsion droplets. Specifically, the intent was for these emulsions to

be formulated using food grade ingredients and to differ only in terms of SFC in order to

improve on previous limited work in the area where differences in other particle properties also

existed (Bonnaire et al., 2008; Huynh & Wright, 2018). In agreement with the hypothesis, the in

vitro digestibility of the undercooled liquid emulsion droplets was higher and there was faster in

vitro lipolysis compared to the crystalline solid emulsion droplets, although the ultimate extent of

lipid digestion was similar for both systems. Interestingly, gastric structuring occurred with the

presence of solid fat when the emulsions flocculated in the presence of gastric acidity. This

destabilization was associated with significantly reduced digestibility which partly obscured the

influence of crystallinity per se on emulsion digestibility. The observation that the presence of

solid fat decreased emulsion digestibility and that rate of lipid digestion was inversely related to

the SFC agrees with other recent findings using emulsions based on soybean oil and fully

hydrogenated soybean oil (Guo, Bellissimo, et al., 2017). We also recently published work on a

compositionally identical cocoa butter emulsions that similarly showed that the presence of solid

fat impeded lipid digestibility, although, those digestion experiments were conducted at 25 C, to

maintain contrasting droplet physical states (Hart et al., 2018).

Study 3 (Chapter 5) investigated the same undercooled liquid and crystalline solid droplets as in

the previous chapter (i.e. compositionally identical with no confounding influences from other

particle characteristics) in a human postprandial lipemic study. Based on the in vitro findings, it

was predicted that these emulsions would show similar levels of total fat absorption, but the

undercooled liquid emulsion droplets would have a faster initial rise in lipemia. Agreeing with

111

this hypothesis, the emulsion with the undercooled liquid droplets showed a slight earlier

increase in plasma TAG and iAUC, pointing to differences in emulsion droplets with different

levels of solid fat. Although minor, it is meaningful that any differences were detected in PPL,

considering the emulsions were identical in all ways, other than SFC. Furthermore, the presence

of solid fat significantly delayed the plasma TAG increase from fasting compared to the

undercooled liquid emulsion droplet system. There were also minor, although not statistically

significant differences, observed in terms of post-trial gastrointestinal symptoms indicating

differences depending on physical state of the ingested lipids. The observed differences in study

point to a specific contribution of TAG physical state to postprandial metabolic response. This

study offers the cleanest evidence to date that lipid physical state, specifically, is a contributing

factor to digestibility.

The approach of undercooling emulsion droplets to retain the liquid state is an innovative

technique allowing comparisons of compositionally identical test lipid systems. However, these

droplets are thermodynamically unstable leading to considerable challenges when handling these

samples so as to avoid inducing crystallization or gravitational destabilization. Every possible

effort was made to avoid temperature fluctuations while handling the emulsions, as discussed in

Chapter 4. Further, the emulsions were prepared and stored for a maximum of one day to retain

the stability and routinely monitored for crystallinity by DSC. It is worth noting that the initial

formulation of this system was also challenging. Specifically, determining the lipid and

emulsifier ingredients and levels that allowed the formation of both reasonably stable

undercooled lipid droplets and partially crystalline droplets with a reasonably high SFC took a

lot of trial and error, rationalized based on the primary literature. Over 30 lipid and emulsifier

combinations were tested to finally develop the current palm stearin and Span 60 based

112

emulsion. The fact that partial crystallization was observed for LE in the in vitro experiments

means that this may also have occurred for LE in vivo. Unfortunately, it is impossible to know

the degree to which LE may or may not have crystallized, once ingested. This and the in vitro

evidence of partial coalescence for SE with gastric exposure, raises the possibility that LE and

SE had different colloidal behaviours in the gastrointestinal tract (despite having the same initial

size distributions and diameters) that contributed to their observed lipemic responses.

6.3 Limitations in the different digestion models used

This thesis involved in vitro digestion experiments carried out using models of varying

complexity: from a simpler static mono-compartmental digestion model with and without a

gastric phase to a more sophisticated dynamic TNO TIM digestion model to an in vivo human

model of healthy adult males. These different approaches have their inherent advantages and

limitations and the in vivo human model is considered the golden standard. The digestion

experiments with and without the gastric phase emphasized the importance of including a gastric

phase in in vitro digestion experiments and, apart from lipid hydrolysis, pointed to the significant

contribution of gastric structuring in determining digestibility of lipids, specifically in the

presence of solid fat.

In comparison to the dynamic in vitro digestion model used, the static model had several

limitations, including likely lipolysis product accumulation, absence of a physiologically relevant

absorption process, and simplistic shear conditions, i.e. type and level. The experiments that used

the static digestion model to study emulsion in vitro digestibility (Chapter 4) did not include an

oral phase, as the residence time of an emulsion in the oral phase is considerably shorter, i.e. less

than 5 seconds. However, as recommended recently, it would be ideal to include the oral phase

113

to accurately mimic all the phases of in vivo digestion (Brodkorb et al., 2019). Further, for the

gastric phase of this model, lipases were not used due to the commercial unavailability of a

lipase analogue to human gastric lipase. However, rabbit gastric lipase (rabbit gastric extract) is

now available and is advised to use for the gastric phase digestion under static in vitro conditions

(Brodkorb et al., 2019).

The dynamic model has the ability to mimic human digestion more closely, i.e. gastrointestinal

pH changes, flow rate of secretions, “absorption” of digested products, etc. However, this system

has also certain limitations in mimicking actual human digestion physiology. The gastric phase

of the dynamic TIM-1 model uses sn-1, 3 specific Rhizopus lipase where, human gastric lipase is

either sn-1 or 3 specific (Reis et al., 2009). Therefore, the lipolysis and free fatty acid release in

the TIM-1 system will potentially be an over estimation compared to actual human in vivo

digestion. It has been shown that dietary lipids with different solid fat contents differ in their rate

of human gastric emptying (Steingoetter et al., 2017). As such, the set constant gastric emptying

rate in the TIM-1 model is a limitation (Minekus et al., 1995) as it assumes similar rates of

gastric emptying across test fats with differing solid fat contents and potentially other physical

properties. Indeed in vivo- in vitro correlations were shown in this work, although, this system

needs further improvements to address the challenges associated with testing meals containing

high levels of lipid, including solid fat.

Given the ethical restrictions, expense, and inherent high individual variability, in vitro digestion

models can serve as higher throughput, less expensive, and less time-consuming alternatives for

pre-screening and to isolate treatment effects. Static and dynamic models can also allow for

sampling with analysis to investigate underlying digestive mechanisms and processes, as in

Chapters 3 and 4.

114

6.4 Future directions

It is exciting to witness the renewed interest in clarifying possible contributions of physical

properties to metabolic response, including for dietary TAGs. In the future, similar studies using

compositionally similar undercooled versus crystallized droplets should be compared but using

emulsifiers not sensitive to acidic flocculation. This would eliminate confounding influences

based on differing degrees of colloidal stability and possible partial coalescence. It is also

possible that within the same SFC, different crystal structures could potentially show different

digestibilities, based on crystal polymorphism, morphology or distribution. Such systems based

on the same chemical composition with no confounding influences from other particle properties

have not been compared thus far and are an interesting avenue for future exploration. Lastly, in

reality, meals consist of complex mixtures of nutrients resulting in nutrient-nutrient interactions

that further complicate lipid digestion. Emulsified lipids in particular, i.e. ingested in the

emulsified form or emulsified in the gastrointestinal system to facilitate digestion, can

significantly alter their characteristics, particle charge specifically, in the presence of other

nutrients and compounds, including proteins, carbohydrates, minerals, and soluble and/or

insoluble fibres. This could either exaggerate or minimize the extent of gastric destabilization

and contribute to the digestibility and absorption. Such interactions constitute a complex yet very

important area to explore. All in all, a deeper understanding of the contributions of solid fat to

digestive processing will help in developing lipid formulations with targeted release and

absorption potential, i.e. faster vs slower, with potential as a strategy for energy regulation and

weight management.

115

6.5 Conclusions

This thesis aimed to investigate the impact of the presence of solid fat on digestibility and

postprandial metabolism using compositionally similar, i.e. IE and non-IE lipid blends, and

compositionally identical, i.e. undercooled liquid and crystalline solid emulsion droplet, test lipid

systems. Overall, the presence of more solid fat was associated with decreased digestibility, as

reflected in slower in vitro FA release and lower total amount of FAs hydrolyzed, as well as a

delayed postprandial plasma TAG increase from fasting, lower peak plasma TAG, and lower

overall absorbed FAs, i.e. iAUC, in vivo. In vitro, the presence of solid fat contributed to gastric

structuring which was shown to significantly impede the subsequent rate of digestion and the

FFA release, potentially related to decreased surface area available for lipase breakdown. Lipid

digestion is expected to be more efficient and complete in vivo compared to the in vitro digestion

models. However, it remains unclear exactly how kinetic differences in lipid hydrolysis, i.e. the

rate of FFA release alter PPL e.g. the time taken to reach postprandial peak, the peak value, the

return to baseline, etc. as well as other aspects of metabolism and subsequent chronic disease

risk. This thesis offers important insights with respect to TAG crystallinity. It helps to

mechanistically evidence that the digestibility of crystalline TAGs is attenuated compared to

those in the liquid state, with implications evidenced for postprandial lipemia.

116

REFERENCE

Abia, R., Pacheco, Y. M., Perona, J. S., Montero, E., Muriana, F. J., & Ruiz-Gutiérrez, V.

(2001). The metabolic availability of dietary triacylglycerols from two high oleic oils during

the postprandial period does not depend on the amount of oleic acid ingested by healthy

men. The Journal of Nutrition, 131(June 2000), 59–65.

Anderson, B. M., MacLennan, M. B., Hillyer, L. M., & Ma, D. W. L. (2014). Lifelong exposure

to n-3 PUFA affects pubertal mammary gland development. Applied Physiology, Nutrition,

and Metabolism = Physiologie Appliquée, Nutrition et Métabolisme, 39(6), 699–706.

https://doi.org/10.1139/apnm-2013-0365

Armand, M., Pasquier, B., André, M., Borel, P., Senft, M., Peyrot, J., … Lairon, D. (1999).

Digestion and absorption of 2 fat emulsions with different droplet sizes in the human

digestive tract. American Journal of Clinical Nutrition, 70(6), 1096–1106.

Aronhime, J. Schlichter, Sarig, S., & Garti, N. (1987). Mechanistic considerations of

polymorphic transformations of tristearin in the presence of emulsifiers. Journal of the

American Oil Chemists’ Society, 64(4), 529–533. https://doi.org/10.1007/BF02636388

Aronhime, Judith Schlichter, Sarig, S., & Garti, N. (1988). Dynamic control of polymorphic

transformation in triglycerides by surfactants: The button syndrome. Journal of the

American Oil Chemists’ Society, 65(7), 1144–1150. https://doi.org/10.1007/BF02660571

Asselin, G., Lavigne, C., Bergeron, N., Angers, P., Belkacemi, K., Arul, J., & Jacques, H.

(2004). Fasting and postprandial lipid response to the consumption of modified milk fats by

guinea pigs. Lipids, 39(10), 985–992. https://doi.org/10.1007/s11745-004-1320-5

Benjamin, E. J., Virani, S. S., Callaway, C. W., Chamberlain, A. M., Chang, A. R., Cheng, S., …

Muntner, P. (2018). Heart disease and stroke statistics - 2018 update: A report from the

American Heart Association. Circulation (Vol. 137).

https://doi.org/10.1161/CIR.0000000000000558

Bergstedt, S. E., Hayashi, H., Kritchevsky, D., & Tso, P. (1990). A comparison of absorption of

glycerol tristearate and glycerol trioleate by rat small intestine. The American Journal of

Physiology, 259(3 Pt 1), G386–G393.

Berry, S. E., Miller, G. J., & Sanders, T. A. (2007). The solid fat content of stearic acid – rich

fats determines their. American Journal of Clinical Nutrition, 85, 1486–1494.

Berry, S. E E, Woodward, R., Yeoh, C., Miller, G. J., & Sanders, T. A. B. (2007). Effect of

interesterification of palmitic acid-rich triacylglycerol on postprandial lipid and factor VII

response. Lipids, 42(4), 315–323. https://doi.org/10.1007/s11745-007-3024-x

Berry, Sarah E E. (2009). Triacylglycerol structure and interesterification of palmitic and stearic

acid-rich fats: an overview and implications for cardiovascular disease. Nutrition Research

Reviews, 22, 3–17. https://doi.org/10.1017/S0954422409369267

117

Berry, Sarah E E, & Sanders, T. A. B. (2005). Influence of triacylglycerol structure of stearic

acid-rich fats on postprandial lipaemia. Proceedings of the Nutrition Society, 64, 205–212.

https://doi.org/10.1079/PNS2005422

Bligh, E. G., & Dyer, W. J. (1959). A rapid method of total lipid extraction and purification.

Canadian Journal of Biochemistry and Physiology, 37(8), 911–917.

Bonnaire, L., Sandra, S., Helgason, T., Decker, E. A., Weiss, J., & McClements, D. J. (2008).

Influence of lipid physical state on the in vitro digestibility of emulsified lipids. Journal of

Agricultural and Food Chemistry, 56(10), 3791–3797. https://doi.org/10.1021/jf800159e

Bornhorst, G. M., & Singh, P. R. (2014). Gastric digestion in vivo and in vitro: how the

structural aspects of food influence the digestion process. Annual Review of Food Science

and Technology, 5(December 2013), 111–132. https://doi.org/10.1146/annurev-food-

030713-092346

Brodkorb, A., Egger, L., Alminger, M., Alvito, P., Assunção, R., Ballance, S., … Recio, I.

(2019). INFOGEST static in vitro simulation of gastrointestinal food digestion. Nature

Protocols, 14(April). https://doi.org/10.1038/s41596-018-0119-1

Comas, D. I., Wagner, J. R., & Tomás, M. C. (2006). Creaming stability of oil in water (O/W)

emulsions: Influence of pH on soybean protein-lecithin interaction. Food Hydrocolloids,

20(7), 990–996. https://doi.org/10.1016/j.foodhyd.2005.11.006

Croda Europe Ltd. (2009). Span and Tween.

Crupkin, M., & Zambelli, A. (2008). Detrimental impact of trans fats on human health: Stearic

acid-rich fats as possible substitutes. Comprehensive Reviews in Food Science and Food

Safety, 7(3), 271–279. https://doi.org/10.1111/j.1541-4337.2008.00045.x

Day, L., Golding, M., Xu, M., Keogh, J., Clifton, P., & Wooster, T. J. (2014). Food

Hydrocolloids Tailoring the digestion of structured emulsions using mixed monoglyceride e

caseinate interfaces. Food Hydrocolloids, 36, 151–161.

https://doi.org/10.1016/j.foodhyd.2013.09.019

Dehghan, M., Mente, A., Zhang, X., Swaminathan, S., Li, W., Mohan, V., … Kumar, R. (2017).

Associations of fats and carbohydrate intake with cardiovascular disease and mortality in 18

countries from five continents ( PURE ): a prospective cohort study. The Lancet, 390,

2050–2062. https://doi.org/10.1016/S0140-6736(17)32252-3

Dias, C. B., Moughan, P. J., Wood, L. G., Singh, H., & Garg, M. L. (2017). Postprandial lipemia:

Factoring in lipemic response for ranking foods for their healthiness. Lipids in Health and

Disease, 16(1), 1–11. https://doi.org/10.1186/s12944-017-0568-5

Douaire, M., Di Bari, V., Norton, J. E., Sullo, A., Lillford, P., & Norton, I. T. (2014). Fat

crystallisation at oil-water interfaces. Advances in Colloid and Interface Science, 203, 1–10.

https://doi.org/10.1016/j.cis.2013.10.022

118

Dougherty, R. M., Aliman, A. M., & Iacono, J. M. (1995). Effects of diets containing acid on

plasma lipoprotein excretion of men1 high or low amounts of stearic fractions and fecal

fatty acid. American Journal of Clinical Nutiriion, 61, 1120–1128.

Eldemnawy, H. Y., Wright, A., & Corredig, M. (2015). A Better Understanding of the Factors

Affecting In vitro Lipolysis Using Static Mono-compartmental Models. Food Digestion,

DOI: 10.1007/s13228-015-0038-3. https://doi.org/10.1007/s13228-015-0038-3

FAO. (2008). Fats and fatty acids in human nutrition, Report of an expert consultation. FAO,

Food and Nutrition Paper (Vol. 91). https://doi.org/978-92-5-106733-8

Feldman, E. B., Russell, B. S., Schnare, F. H., Miles, B. C., Doyle, E. A., & Moretti-Rojas, I.

(1979). Effects of tristearin, triolein and safflower oil diets on cholesterol balance in rats.

The Journal of Nutrition, 109(12), 2226–2236.

Folch, J., Lees, M., & Stanley, G. H. S. (1957). A simple method for the isolation and

purifucation of total lipides from animal tissues. The Journal of Biological Chemistry, 226,

497–509.

Fredrick, E., Walstra, P., & Dewettinck, K. (2010). Factors governing partial coalescence in oil-

in-water emulsions. Advances in Colloid and Interface Science, 153(1–2), 30–42.

https://doi.org/10.1016/j.cis.2009.10.003

Gabert, L., Vors, C., Pélissier, C. L., Sauvinet, V., Porcheron, S. L., Drai, J., … Michalski, M. C.

(2011). 13C tracer recovery in human stools after digestion of a of a fat‐rich meal labelled

with [1,1,1‐13C3]tripalmitin and [1,1,1‐13C3]triolein. Rapid Communications in Mass

Spectrometry, 25(April), 2697–2703. https://doi.org/10.1002/rcm.5067

Gervais, R., Gagnon, F., Kheadr, E. E., Van Calsteren, M. R., Farnworth, E. R., Fliss, I., &

Chouinard, P. Y. (2009). Bioaccessibility of fatty acids from conjugated linoleic acid-

enriched milk and milk emulsions studied in a dynamic in vitro gastrointestinal model.

International Dairy Journal, 19(10), 574–581. https://doi.org/10.1016/j.idairyj.2009.05.001

Golding, M., & Wooster, T. J. (2010). The influence of emulsion structure and stability on lipid

digestion. Current Opinion in Colloid and Interface Science, 15(1–2), 90–101.

https://doi.org/10.1016/j.cocis.2009.11.006

Golding, M., Wooster, T. J., Day, L., Xu, M., Lundin, L., Keogh, J., & Clifton, P. (2011). Impact

of gastric structuring on the lipolysis of emulsified lipids. Soft Matter, 7(7), 3513.

https://doi.org/10.1039/c0sm01227k

Goodman, B. E. (2010). Insights into digestion and absorption of major nutrients in humans.

Advances in Physiology Education, 34, 44–53. https://doi.org/10.1152/advan.00094.2009

Griffin, W. C. (1949). Classification of surface-active agents by" HLB". Journal of the Society of

Cosmetic Chemists, 1(5), 311–326. Retrieved from

http://journal.scconline.org//pdf/cc1949/cc001n05/p00311-p00326.pdf

119

Guerra, A., Etienne-Mesmin, L., Livrelli, V., Denis, S., Blanquet-Diot, S., & Alric, M. (2012).

Relevance and challenges in modeling human gastric and small intestinal digestion. Trends

in Biotechnology, 30(11), 591–600. https://doi.org/10.1016/j.tibtech.2012.08.001

Guo, Q., Bellissimo, N., & Rousseau, D. (2017). The Physical State of Emulsified Edible Oil

Modulates Its in Vitro Digestion. Journal of Agricultural and Food Chemistry, 65(41),

9120–9127. https://doi.org/10.1021/acs.jafc.7b03368

Guo, Q., Ye, A., Bellissimo, N., Singh, H., & Rousseau, D. (2017). Modulating fat digestion

through food structure design. Progress in Lipid Research, 68(August), 109–118.

https://doi.org/10.1016/j.plipres.2017.10.001

Hall, W. L., Brito, M. F., Huang, J., Wood, L. V., Filippou, A., Sanders, T. a. B., … Berry, S. E.

E. (2014). An Interesterified Palm Olein Test Meal Decreases Early-Phase Postprandial

Lipemia Compared to Palm Olein: a Randomized Controlled Trial. Lipids, 49(9), 895–904.

https://doi.org/10.1007/s11745-014-3936-1

Hall, W. L., Brito, M. F., Huang, J., Wood, L. V., Filippou, A., Sanders, T. A. B., & Berry, S. E.

E. (2014). An Interesterified Palm Olein Test Meal Decreases Early-Phase Postprandial

Lipemia Compared to Palm Olein: a Randomized Controlled Trial. Lipids, 49(9), 895–904.

https://doi.org/10.1007/s11745-014-3936-1

Hall, W. L., Iqbal, S., Li, H., Gray, R., & Berry, S. E. E. (2017). Modulation of postprandial

lipaemia by a single meal containing a commonly consumed interesterified palmitic acid ‐ rich fat blend compared to a non ‐ interesterified equivalent. European Journal of Nutrition,

56(8), 2487–2495. https://doi.org/10.1007/s00394-016-1284-z

Hart, S. M., (Lois) Lin, X., Thilakarathna, S., & Wright, A. J. (2018). Emulsion droplet

crystallinity attenuates early in vitro digestive lipolysis and beta-carotene bioaccessibility.

Food Chemistry, 260(November 2017), 145–151.

https://doi.org/10.1016/j.foodchem.2018.03.142

Health Canada. (2016). Food Additives. In Food Additives. Retrieved from

https://www.canada.ca/en/health-canada/services/food-nutrition/food-safety/food-

additives.html

Hunter, E. J. (2001). Studies on effects of dietary fatty acids as related to their position on

triglycerides. Lipids, 36(7), 655–668. https://doi.org/10.1007/s11745-001-0770-0

Hur, S. J., Decker, E. A., & McClements, D. J. (2009). Influence of initial emulsifier type on

microstructural changes occurring in emulsified lipids during in vitro digestion. Food

Chemistry, 114(1), 253–262. https://doi.org/10.1016/j.foodchem.2008.09.069

Hussain, M. M. (2014). Intestinal lipid absorption and lipoprotein formation. Current Opinion in

Lipidology, 25(3), 200–206. https://doi.org/10.1016/j.biotechadv.2011.08.021.Secreted

Huynh, S., & Wright, A. J. (2018). Tripalmitin-Sodium Dodecyl Sulfate Emulsion Droplet

Liquid vs. Solid State Impacts in vitro Digestive Lipolysis. Journal of the American Oil

120

Chemists’ Society, 95(March), 161–170. https://doi.org/10.1002/aocs.12004

Jiao, W., Lin Li, Anling Yu, Zhao, D., Sheng, B., Aikelamu, M., … Zhang, X. (2019). In Vitro

Gastrointestinal Digestibility of Crystalline Oil-in-Water Emulsions: Influence of Fat

Crystal Structure. Journal of Agricultural and Food Chemistry, 67, 927–934. research-

article. https://doi.org/10.1021/acs.jafc.8b04287

Kalantzi, L., Goumas, K., Kalioras, V., Abrahamsson, B., Dressman, J. B., & Reppas, C. (2006).

Characterization of the human upper gastrointestinal contents under conditions simulating

bioavailability/bioequivalence studies. Pharmaceutical Research, 23(1), 165–176.

https://doi.org/10.1007/s11095-005-8476-1

Kaplan, R. J., & Greenwood, C. E. (1998). Poor Digestibility of Fully Hydrogenated Soybean

Oil in Rats : A Potential Benefit of Hydrogenated Fats and Oils 1 – 3. The Journal of

Nutrition, 128(5), 875–880.

Karupaiah, T., & Sundram, K. (2007). Effects of stereospecific positioning of fatty acids in

triacylglycerol structures in native and randomized fats: a review of their nutritional

implications. Nutrition & Metabolism, 4, 16. https://doi.org/10.1186/1743-7075-4-16

Kolovou, G. D., Anagnostopoulou, K. K., Pavlidis, A. N., Salpea, K. D., Iraklianou, S. A.,

Tsarpalis, K., … Cokkinos, D. V. (2005). Postprandial lipemia in men with metabolic

syndrome, hypertensives and healthy subjects. Lipids in Health and Disease, 4, 21.

https://doi.org/10.1186/1476-511X-4-21

Kolovou, G., P. Mikhailidis, D., Kovar, J., Lairon, D., G. Nordestgaard, B., Ooi, T.-C., …

Panotopoulos, G. (2011). Assessment and Clinical Relevance of Non-Fasting and

Postprandial Triglycerides: An Expert Panel Statement. Current Vascular Pharmacology,

999(999), 1–13. https://doi.org/10.2174/1570211213146321611

Kong, F., & Singh, R. P. (2008). A Model Stomach System to Investigate Disintegration

Kinetics of Solid Foods during gastric digestion. Food Science, 73(5), 202–210.

https://doi.org/10.1111/j.1750-3841.2008.00745.x

Lairon, D., Lopez-Miranda, J., & Williams, C. (2007). Methodology for studying postprandial

lipid metabolism.(Review)(Report). European Journal of Clinical Nutrition, 61(10), 1145.

https://doi.org/10.1038/sj.ejcn.1602749

Laugerette, F., Vors, C., Peretti, N., & Michalski, M. (2011). Biochimie Complex links between

dietary lipids , endogenous endotoxins and metabolic in fl ammation. Biochimie, 93(1), 39–

45. https://doi.org/10.1016/j.biochi.2010.04.016

Lee, J. H., Akoh, C. C., & Lee, K. T. (2008). Physical properties of trans-free bakery shortening

produced by lipase-catalyzed interesterification. JAOCS, Journal of the American Oil

Chemists’ Society, 85, 1–11. https://doi.org/10.1007/s11746-007-1155-0

Levy, E. (2015). Insights from human congenital disorders of intestinal lipid metabolism, 56,

945–962. https://doi.org/10.1194/jlr.R052415

121

Lin, X., & Wright, A. J. (2018). Pectin and gastric pH interactively affect DHA-rich emulsion in

vitro digestion microstructure , digestibility and bioaccessibility. Food Hydrocolloids, 76,

49–59. https://doi.org/10.1016/j.foodhyd.2017.06.010

Lindman, A. S., Veierød, M. B., Tverdal, A., Pedersen, J. I., & Selmer, R. (2010). Nonfasting

triglycerides and risk of cardiovascular death in men and women from the Norwegian

Counties Study. European Journal of Epidemiology, 25, 789–798.

https://doi.org/10.1007/s10654-010-9501-1

Livesey, G. (2000). The absorption of stearic acid from triacylglycerols: an inquiry and analysis.

Nutrition Research Reviews, 13, 185–214. https://doi.org/10.1079/095442200108729061

Mackie, A. (2017). Food: more than the sum of its parts. Current Opinion in Food Science, 16,

120–124. https://doi.org/10.1016/j.cofs.2017.07.004

Maio, S. Di, & Carrier, R. L. (2011). Gastrointestinal contents in fasted state and post-lipid

ingestion : In vivo measurements and in vitro models for studying oral drug delivery.

Journal of Controlled Release, 151(2), 110–122.

https://doi.org/10.1016/j.jconrel.2010.11.034

Mantovani, R. A., Cavallieri, Â. L. F., Netto, F. M., & Cunha, R. L. (2013). Stability and in vitro

digestibility of emulsions containing lecithin and whey proteins. Food & Function, 4(9),

1322. https://doi.org/10.1039/c3fo60156k

Marangoni, A. G., & Rousseau, D. (1995). Engineering triacylglycerols: The role of

interesterification. Trends in Food Science & Technology, 6, 329–335.

Marciani, L., Faulks, R., Wickham, M. S. J., Bush, D., Pick, B., Wright, J., … Spiller, R. C.

(2009). Effect of intragastric acid stability of fat emulsions on gastric emptying, plasma

lipid profile and postprandial satiety. British Journal of Nutrition, 101(06), 919.

https://doi.org/10.1017/S0007114508039986

Martins, I. J., Mortimer, B. C., Miller, J., & Redgrave, T. G. (1996). Effects of particle size and

number on the plasma clearance of chylomicrons and remnants. J. Lipid Res., 37(12), 2696–

2705. Retrieved from http://www.jlr.org/content/37/12/2696.full.pdf

Mattson, F. H., Nolen, G. A., & Webb, M. R. (1979). The absorbability by rats of various

triglycerides of stearic and oleic acid and the effect of dietary calcium and magnesium. The

Journal of Nutrition, 109(10), 1682–1687.

McClements, D. J., Decker, E. A., & Park, Y. (2009). Controlling lipid bioavailability through

physicochemical and structural approaches. Critical Reviews in Food Science and Nutrition,

49(1), 48–67. https://doi.org/10.1080/10408390701764245

McClements, D. J., Decker, E. A., Park, Y., & Weiss, J. (2008). Designing food structure to

control stability, digestion, release and absorption of lipophilic food components. Food

Biophysics, 3(2), 219–228. https://doi.org/10.1007/s11483-008-9070-y

122

Mensink, R. P., Sanders, T. A., Baer, D. J., Hayes, K. C., Howles, P. N., & Marangoni, A.

(2016). The Increasing Use of Interesteri fi ed Lipids in the Food Supply and Their Effects

on Health. Advances in Nutrition, 7, 719–729. https://doi.org/10.3945/an.115.009662.719

Micha, R., & Mozaffarian, D. (2010). Saturated fat and cardiometabolic risk factors, coronary

heart disease, stroke, and diabetes: A fresh look at the evidence. Lipids, 45(10), 893–905.

https://doi.org/10.1007/s11745-010-3393-4

Michalski, M. C., Genot, C., Gayet, C., Lopez, C., Fine, F., Joffre, F., … Raynal-Ljutovac, K.

(2013). Multiscale structures of lipids in foods as parameters affecting fatty acid

bioavailability and lipid metabolism. Progress in Lipid Research, 52(4), 354–373.

https://doi.org/10.1016/j.plipres.2013.04.004

Minekus, M., Jelier, M., Xiao, J.-Z., Kondo, S., Iwatsuki, K., Kokubo, S., … Havenaar, R.

(2005). Effect of partially hydrolyzed guar gum (PHGG) on the bioaccessibility of fat and

cholesterol. Bioscience, Biotechnology, and Biochemistry, 69(5), 932–938.

https://doi.org/10.1271/bbb.69.932

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 Intestine. Alternatives to Laboratory Animals, 23, 197–209.

Mozaffarian, D., & Wu, J. (2011). Omega-3 fatty acids and cardiovascular disease. Journal of

the American College of Cardiology, 58(20), 2047–2067.

https://doi.org/10.1016/j.jacc.2011.06.063

Mu, H., & Høy, C. E. (2004). The digestion of dietary triacylglycerols. Progress in Lipid

Research, 43, 105–133. https://doi.org/10.1016/S0163-7827(03)00050-X

Mu, H., & Porsgaard, T. (2005). The metabolism of structured triacylglycerols. Progress in Lipid

Research, 44(6), 430–448. https://doi.org/10.1016/j.plipres.2005.09.002

Nettleton, J. A., Brouwer, A., Geleijnse, J. M., & Hornstra, G. (2017). Saturated Fat

Consumption and Risk of Coronary Heart Disease and Ischemic Stroke : A Science Update.

Annals of Nutrition and Metabolism, 70, 26–33. https://doi.org/10.1159/000455681

Nik, A Malaki, Wright, A., & Corredig, M. (2011). Impact of interfacial composition on

emulsion digestion and rate of lipid hydrolysis using different in vitro digestion models.

Colloids and Surfaces B: Biointerfaces, 83(2), 321–330.

https://doi.org/10.1016/j.colsurfb.2010.12.001

Nik, Amir Malaki, Langmaid, S., & Wright, A. J. (2012a). Digestibility and β-carotene release

from lipid nanodispersions depend on dispersed phase crystallinity and interfacial

properties. Food and Function, 3(3), 234. https://doi.org/10.1039/c1fo10201j

Nik, Amir Malaki, Langmaid, S., & Wright, A. J. (2012b). Nonionic surfactant and interfacial

structure impact crystallinity and stability of beta-carotene loaded lipid nanodispersions.

Journal of Agricultural and Food Chemistry, 60(16), 4126–4135.

123

https://doi.org/10.1021/jf204810m

Normén, L., Ellegård, L., Janssen, H. G., Steenbergen, H., Trautwein, E., & Andersson, H.

(2006). Phytosterol and phytostanol esters are effectively hydrolysed in the gut and do not

affect fat digestion in ileostomy subjects. European Journal of Nutrition, 45(3), 165–170.

https://doi.org/10.1007/s00394-006-0578-y

O′Sullivan, C. M., Davidovich-Pinhas, M., Wright, A. J., Barbut, S., & Marangoni, A. G. (2017).

Ethylcellulose oleogels for lipophilic bioactive delivery – effect of oleogelation on in vitro

bioaccessibility and stability of beta-carotene. Food and Function, 8(4), 1438–1451.

https://doi.org/10.1039/C6FO01805J

Osborn, H. T., & Akoh, C. C. (2002). Structured Lipids – Novel Fats with Medical ,

Nutraceutical , and Food Applications. Comprehensive Reviews In Food Science And Food

Safety, 3, 110–120. https://doi.org/10.1111/j.1541-4337.2002.tb00010.x

Reis, P., Holmberg, K., & Watzke, H. (2009). Lipases at interfaces: a review. Advances in

Colloid and Interface Science, 147–148, 237–250. https://doi.org/10.1016/j.cis.2008.06.001

Ribnicky, D. M., Roopchand, D. E., Oren, A., Grace, M., Poulev, A., Lila, M. A., … Raskin, I.

(2014). Effects of a high fat meal matrix and protein complexation on the bioaccessibility of

blueberry anthocyanins using the TNO gastrointestinal model (TIM-1). Food Chemistry,

142(2014), 349–357. https://doi.org/10.1016/j.foodchem.2013.07.073

Robertson, M. D., & Mathers, J. C. (2000). Gastric emptying rate of solids is reduced in a group

of ileostomy patients. Digestive Diseases and Sciences, 45(7), 1285–1292.

https://doi.org/10.1023/A:1005587516233

Robinson, D. M., Martin, N. C., Robinson, L. E., Ahmadi, L., Marangoni, A. G., & Wright, A. J.

(2009). Influence of interesterification of a stearic acid-rich spreadable fat on acute

metabolic risk factors. Lipids, 44(1), 17–26. https://doi.org/10.1007/s11745-008-3253-7

Ros, E. (2000). Intestinal absorption of triglyceride and cholesterol. Dietary and pharmacological

inhibition to reduce cardiovascular risk. Atherosclerosis, 151(2), 357–379.

https://doi.org/10.1016/S0021-9150(00)00456-1

Sacks, F. M., Lichtenstein, A. H., Wu, J. H. Y., Appel, L. J., Creager, M. A., Kris-Etherton, P.

M., … Horn, L. V. Van. (2017). Dietary Fats and Cardiovascular Disease. Circulation, 136,

e1–e23. https://doi.org/10.1161/CIR.0000000000000510

Sanders, Thomas A B, Filippou, A., Berry, S. E., Baumgartner, S., & Mensink, R. P. (2011).

Palmitic acid in the sn -2 position of triacylglycerols acutely influences postprandial lipid

metabolism 1 – 3. The American Journal of Clinical Nutrition, 94(2), 1433–1441.

https://doi.org/10.3945/ajcn.111.017459.1

Sanders, Tom A.B. (2003). Dietary fat and postprandial lipids. Current Atherosclerosis Reports,

5(6), 445–451. https://doi.org/10.1007/s11883-003-0034-x

124

Small, D. M. (1991). the Effects of Glyceride Structure on absorption and metabolism. Annual

Reviews, 11, 413–434.

Sonwai, S., Podchong, P., & Rousseau, D. (2017). Crystallization kinetics of cocoa butter in the

presence of sorbitan esters. Food Chemistry, 214, 497–506.

https://doi.org/10.1016/j.foodchem.2016.07.092

Speranza, A., Corradini, M. G., Hartman, T. G., Ribnicky, D., Oren, A., & Rogers, M. A.

(2013a). Influence of emulsifier structure on lipid bioaccessibility in oil-water

nanoemulsions. Journal of Agricultural and Food Chemistry, 61(26), 6505–6515.

Speranza, A., Corradini, M. G., Hartman, T. G., Ribnicky, D., Oren, A., & Rogers, M. a.

(2013b). In fluence of Emulsifier Structure on Lipid Bioaccessibility in Oil−Water

Nanoemulsions. Journal of Agricultural and Food Chemistry, 61, 6506–6515.

Steingoetter, A., Buetikofer, S., Curcic, J., Menne, D., Rehfeld, J. F., Fried, M., … Wooster, T. J.

(2017). The Dynamics of Gastric Emptying and Self-Reported Feelings of Satiation Are

Better Predictors Than Gastrointestinal Hormones of the Effects of Lipid Emulsion

Structure on Fat Digestion in Healthy Adults—A Bayesian Inference Approach. The

Journal of Nutrition, 147(4), 706–714. https://doi.org/10.3945/jn.116.237800

Steingoetter, A., Radovic, T., Buetikofer, S., Curcic, J., Menne, D., Fried, M., … Wooster, T. J.

(2015). Imaging gastric structuring of lipid emulsions and its effect on gastrointestinal

function: A randomized trial in healthy subjects. American Journal of Clinical Nutrition,

101(4), 714–724. https://doi.org/10.3945/ajcn.114.100263

Sundram, K., Karupaiah, T., & Hayes, K. C. (2007). Stearic acid-rich interesterified fat and

trans-rich fat raise the LDL/HDL ratio and plasma glucose relative to palm olein in humans.

Nutrition & Metabolism, 4, 3. https://doi.org/10.1186/1743-7075-4-3

Thilakarathna, S H, Rogers, M., Lan, Y., Huynh, S., Marangoni, A. G., Robinson, L. E., &

Wright, A. J. (2016). Investigations of in vitro bioaccessibility from interesterified stearic

and oleic acid-rich blends. Food and Function, 7(4), 1932–1940.

https://doi.org/10.1039/c5fo01272d

Thilakarathna, Surangi H, & Wright, A. J. (2018). Attenuation of Palm Stearin Emulsion Droplet

in Vitro Lipolysis with Crystallinity and Gastric Aggregation. Journal of Agricultural and

Food Chemistry, 66, 10292–10299. research-article.

https://doi.org/10.1021/acs.jafc.8b02636

Tran, T., Green, N. L., Ghosh, S., & Rousseau, D. (2017). Encapsulation of water-in-oil

emulsion droplets within crystal spheroids. Colloids and Surfaces A: Physicochemical and

Engineering Aspects, 524, 1–7. https://doi.org/10.1016/j.colsurfa.2016.12.054

van Berge Henegouwen, G. P., van der Werf, S. D., & Ruben, A. T. (1987). Fatty acid

composition of phospholipids in bile in man: promoting effect of deoxycholate on

arachidonate. Clinica Chimica Acta; International Journal of Clinical Chemistry, 165(1),

27–37. https://doi.org/10.1016/0009-8981(87)90215-4

125

Venema, K., Nuenen, M. H. M. C. van, Ellen G. van den Heuvel, Pool, W., & Vossen, J. M. B.

M. van der. (2003). The Effect of Lactulose on the Composition of the Intestinal Microbiota

and Short-chain Fatty Acid Production in Human Volunteers and a Computer-controlled

Model of the Proximal Large Intestine. Microbial Ecology in Health and Disease, 15(2–3),

94–105. https://doi.org/10.1080/08910600310019895

Vors, C., Pineau, G., Gabert, L., Drai, J., Louche-Peĺissier, C., Defoort, C., … Michalski, M. C.

(2013). Modulating absorption and postprandial handling of dietary fatty acids by

structuring fat in the meal: A randomized crossover clinical trial. American Journal of

Clinical Nutrition, 97(1), 23–36. https://doi.org/10.3945/ajcn.112.043976

Wang, X. X., Wang, T., Spurlock, M. E., & Wang, X. X. (2015). Effects of triacylglycerol

structure and solid fat content on fasting responses of mice. European Journal of Nutrition,

55(4), DOI: 10.1007/s00394-015-0972-4. https://doi.org/10.1007/s00394-015-0972-4

Wang, Y., Corwin, R., Jaramillo, D. P., Wojnicki, F. J., & Coupland, J. N. (2011). Acceptability

and digestibility of emulsions in a rat model: Effects of solid fat content and lipid type.

JAOCS, Journal of the American Oil Chemists’ Society, 88(2), 235–241.

https://doi.org/10.1007/s11746-010-1671-1

Wasserstein, R. L., & Lazar, N. A. (2016). The ASA ’ s Statement on p -Values : Context ,

Process , and Purpose. The American Statistician, 70(2), 129–133.

https://doi.org/10.1080/00031305.2016.1154108

Wilde, P., & Chu, B. (2011). Interfacial & colloidal aspects of lipid digestion. Advances in

Colloid and Interface Science, 165(1), 14–22. https://doi.org/10.1016/j.cis.2011.02.004

126

APPENDICES

Appendix 1 University of Guelph Human Research Ethics Board Certificate of Approval

127

Appendix 2 The Lipid Digestion Study Informed Consent Form

STUDY INFORMED CONSENT FORM

The Lipid Digestion Study: Impact of emulsion droplet physical state on postprandial lipemia and satiety in healthy human adult participants

You are being asked to participate in a research study conducted by members of the Department

of Human Health & Nutritional Sciences, University of Guelph. Results of this study will contribute

towards a thesis for PhD student Surangi Thilakarathna and project course credit for the other

students involved. This research is supported by an NSERC Discovery Grant held by faculty

investigator Amanda Wright, PhD.

RESEARCH CONTACT INFORMATION

If you have any questions or concerns about the research, please feel free to contact the study

personnel at any time.

Surangi Thilakarathna, MSc

Lead study coordinator

PhD student, HHNS

Email: sthilaka@uoguelph.ca

Phone: 519-824-4120 x 56314

Amanda Cuncins, BSc

Study coordinator

MSc student, HHNS

Email: acuncins@uoguelph.ca

Phone: 519-824-4120 x 56314

Amanda Wright, PhD

Faculty Investigator

Associate Professor, HHNS

Email:

ajwright@uoguelph.ca

Phone: 519-824-4120 x 54697

Samar Hamad,

Graduate study coordinator

PhD student, HHNS

Email: hamads@uoguelph.ca

Phone: 519-824-4120 x 56314

Melissa Brown, BSc

Study coordinator

MSc student, HHNS

Email: mbrown62@uoguelph.ca

Phone: 519-824-4120 x 56314

PURPOSE OF THE STUDY

A lot of research has focused on how dietary lipids impact human health. Much of this has

focused on how lipids with different compositions (i.e. saturated versus trans fatty acids or

monounsaturated versus polyunsaturated fatty acids) are absorbed and metabolized

differently, leading to differences in disease risk. However, lipids with different compositions

also differ in terms of their physical properties. This means they have different melting

128

temperatures, solid fat contents or crystal structures. Any impact of these physical property

differences, specifically, on lipid digestion and absorption are rarely considered and difficult to

investigate. This study will address this gap by comparing the absorption of solid versus liquid

lipids, using uniquely formulated emulsion beverages.

The two study emulsions that have the exact same composition but have been tempered (i.e.

held at different temperatures) to ensure that in one sample the emulsion lipid droplets are in

the liquid state and, in the other sample they are solid (i.e. crystallized). By measuring the rate

of appearance of lipids in the blood, we can determine if lipid physical state differences,

specifically, contribute to differences in lipid absorption. Because the physical state of a meal

and differences in lipid absorption can change feelings of fullness after a meal is consumed, we

will also ask participants to rate their feelings of satiety and determine gastric emptying by

having participants consume acetaminophen with each test beverages and measure its rate of

appearance in the blood. These endpoints will provide novel insights into if and how dietary

solid versus liquid dietary lipids are digested, absorbed, and contribute to food intake and

cardiovascular disease risk, differently.

PROCEDURES

This study will require two 7-hour study visits separated by a washout of at least 7 days. At each

study visit you will consume one of the flavored test emulsion beverages (i.e. the one with solid

droplets or the one with liquid droplets), in randomized order. Study visits will take place in the

Human Nutraceutical Research Unit (HNRU) of the Food Science-HHNS Addition Building, 88

McGilvray St. at the University of Guelph. On each study visit morning, you will arrive following

an overnight fast, have an intravenous catheter inserted, consume 500 mL of the flavored

emulsion beverage along with crushed and dissolved acetaminophen in water (discussed in detail

below) and have blood samples collected and provide satiety ratings over the next 6 hours. At

the conclusion, you will be provided with a Subway sub to consume before leaving the HNRU. A

study coordinator will check in with you the next day by email or phone.

PRIOR TO THE COMMENCEMENT OF THE STUDY:

Your diet and lifestyle in the 48 hours prior to a blood lipid study can impact results. For this

reason, we ask that you follow a few guidelines during this period leading up to each study visit.

For 48 hours prior to each study visit, you are asked to follow your normal daily routines in terms

of diet as well as physical activity but to avoid any strenuous physical activities, any over-the-

counter medications (other than those recommended by a physician), consumption of alcohol,

and caffeine containing beverages or food (coffee, tea, pop, chocolates, etc.). When you come

for your second study visit, you are requested to follow the same guidelines as for before the first

study visit. In particular, we ask you to consume similar foods and drinks, as the day before your

129

first visit. To support you in doing this, at your first study day, we will ask you to recall what you

ate, record this information and provide it to you so you can follow it as closely as possible prior

to your second study visit.

If you participate in this study, before each of the two study visits, you will also be asked to

undergo a 10-12 hour overnight fast where no consumption of food or drink is permitted, except

water. The dinner meal and an evening snack for the day before each study visit will be provided

to you by the researchers. The list of food choices and ingredients provided by the manufacturers

are available in Appendix 1. Please note that you are requested to consume only the dinner and

the snack provided to you and abstain from consuming any other food items. This is because

what you eat and drink (as well as your activity level) each day can impact you blood lipid

metabolism the following day. For our study, we want to minimize any confounding effects of

your previous food intake. However, during each fast, you are encouraged to drink plenty of

water. When you wake up each study visit morning, you are encouraged to have one – two tall

glasses of water. This will help ensure adequate hydration for the intravenous catheter insertion.

However, please note you are asked to stop drinking water 30 min before arriving at the HNRU.

This is to ensure that any water has drained from your stomach prior to consuming the test

beverages.

As above, you will be asked to consume 1500 mg of acetaminophen along with each treatment

beverage. This will allow us to assess the rate at which your stomach contents empty into the

small intestine since this correlates with the rate of appearance of acetaminophen in the blood.

To participate in this study, we have asked to ensure that you have previously consumed

acetaminophen, without incident and that you have never been told, by a medical professional

to avoid acetaminophen-containing products, such as Tylenol. Additionally, although the study

1500 mg of acetaminophen is below Health Canada’s maximum allowable daily dose (i.e. 4000

mg; Health Canada, 2009; https://www.canada.ca/en/health-canada/services/drugs-medical-

devices/acetaminophen.html), you should avoid taking any acetaminophen-containing

products for 48 hours prior to the study visit (and for 48 hours after the study visit). Examples

of acetaminophen-containing products are mentioned below.

AT EACH STUDY VISIT:

On each study visit day, you will be asked to come early in the morning (by ~8:30 am). A study

coordinator will greet you and ask questions about the past 24 hours, i.e. how do you feel, your

wellness, what types of food you consumed? You will then fill out the first satiety questionnaire

to rate your fullness, hunger, appetite, fatigue, and feelings of nausea.

A qualified and department-approved phlebotomist will then insert an intravenous catheter

(using a needle) into a vein in your forearm in a private sampling bay of the HNRU. The technician

130

will then draw 24 mL of blood through the catheter which will be maintained with a saline drip

and facilitate blood sampling over the next 6 hours. You will then consume 500 mL of the flavored

test beverage, 50 mL of water with crushed acetaminophen, and 50 mL rinse water, within 15

minutes. Details about the study beverages and acetaminophen method are provided below. No

additional water will be provided during the next 6 hours after product consumption.

After consuming the beverage, you will be asked to fill out a short paper questionnaire to capture

details about the taste and mouthfeel of the beverage. Every 30 minutes, you will also complete

a paper questionnaire to rate your fullness, hunger, appetite, fatigue, and nausea. Throughout

the day, you will be asked to remain seated with minimal activity. There will be magazines and

movies available but are also invited to bring a laptop with work or books to read. Blood samples

will be collected every 20 min during the first hour, then every 30 min until 3 hours and every

hour, thereafter, until 6 hours. 24 mL of blood will be drawn at 0, 2, 4, and 6-hour time points

and 6 mL of blood will be drawn at rest of the time points. A total of 138 mL of blood will be

drawn during the 6 hours (11 time points). This volume of blood is roughly one third of the blood

taken by Canadian Blood Services in a single donation (i.e. total ~ 450 mL)

(https://www.blood.ca/en/blood/faqs-whole-blood-donations).

After the 6 hours, the phlebotomist will remove the intravenous catheter, while applying

pressure and bandage with gauze and medical tape to minimize risk of bruising. A ‘Subway’

sandwich or ‘Need a Pita’ of your choice (Appendix 2) will be purchased and provided to you to

consume with a study coordinator prior to leaving the HNRU. The dinner meal and snack prior to

the second visit will be provided to take home with you.

AFTER THE STUDY VISIT:

After leaving the HNRU, you will be asked not to participate in strenuous activity for the rest of

the day. You should resume your typical dietary and lifestyle habits. You are also required not to

consume additional over-the-counter acetaminophen containing products for 48 hours after

each study visit (see list below). As below, the daily limit is 4000 mg in 24 hours and you will

consume 1500 mg at each study visit as excess consumption can lead to liver damage. You are

also asked not to consume any alcohol for the 48 hours after each study visit to minimize the risk

of liver damage which can occur when acetaminophen and alcohol are combined.

Trade names of acetaminophen-containing products to avoid:

Alka-Seltzer Plus Liquid Gels®, Dayquil®, Dimetapp®, Excedrin®, Midol®, Nyquil®,

Robitussin®, Sinutab®, Sudafed®, Tylenol®, Tylenol Cold®

24 hours after each study visit, a study coordinator will contact you by e-mail or phone, based on

your preference, and send you the 24-hour check-in questionnaire. The check-in questionnaire

131

will ask questions about how you are feeling since the study visit, including asking if you

experienced any changes in terms of gastrointestinal symptoms. The questionnaire will include

questions asking if you had bowel movements since your study visit and if this was unusual in

appearance or texture and if there were obvious signs of lipid (i.e. fatty stools). While we value

this feedback, you are reminded that participation in all aspects of the study is voluntary and you

may choose not to answer questions.

STUDY TEST PRODUCTS

At each study visit, you will be asked to consume 500 mL of a vanilla flavored emulsion beverage,

i.e. one in which the lipid droplets are solid (i.e. crystallized) and the other in which they are

liquid. Both emulsions will contain 10 % palm stearin (i.e. 50 g in the 500 mL, supplied by Bunge

Oils Inc., Bradley, IL, USA; an ingredient in cake mixes, bread, non-hydrogenated margarine, plant

based) and 0.4% sorbitan monostearate (also called span 60, i.e. 2 g in the 500 mL, Croda Canada

Ltd., Vaughan, ON, Canada; which is commonly used in whipping cream, cake mixes, and ice

cream, plant or animal based). The remainder of the beverage (i.e. ~449.6 g) will be bottled

water. To improve the palatability of the emulsion beverages, the artificial sweetener sucralose

(Two 1 g packets of Sugar Twin, Ingredients: Dextrose, Maltodextrin, Sucralose. Contains 6.0

mg sucralose per ½ Packet (0.5 g) serving) and artificial vanilla flavor (Club House, Ingredients:

water, alcohol, caramel color, artificial flavor) will be added. All the ingredients are food grade,

commercially available and will used at levels permitted by Health Canada. The beverages will be

prepared in the Food Science Department formulation kitchen, following good manufacturing

and sanitary practices. The beverages will be held at 37 C, i.e. body temperature, to maintain

their physical properties prior to serving them to participants in the HNRU.

DETAILS ABOUT THE ACETAMINOPHEN

As mentioned, acetaminophen is being used in this study to determine gastric emptying because

its rate of appearance in the blood corresponds with the rate at which liquid contents empty

from the stomach into the upper small intestine. Therefore, right after the treatment beverage,

you will also consume 1500 mg (3 extra-strength Tylenol tablets (DIN 00723908) crushed and

dissolve in 50 mL bottle water. Some information about the Extra Strength Tylenol appears

below, but additional information about this product, including risk information can be found at;

https://www.canada.ca/en/health-canada/services/drugs-medical-

devices/acetaminophen.html). After consuming the acetaminophen in water, you will be

provided another 50 mL water to rinse around your mouth, to clean your pallet, and swallow.

Extra strength Tylenol, DIN 00723908:

Medicinal ingredient: Acetaminophen (500 mg/pill)

132

Non-medicinal ingredients (alphabetical): cellulose, corn starch, hypromellose,

magnesium stearate, polyethylene glycol, sodium starch glycolate.

STUDY SAMPLE LABORATORY ANALYSIS

Intravenous blood samples from each time point will be centrifuged and the plasma will be

separated in the HNRU. Samples will be stored at -80 C until analysis of plasma triacylglycerol

concentration, free fatty acids, glucose, insulin, chylomicron size, fatty acid composition,

ApolipoproteinB48 content, and the fatty acid composition of the separated triacylglycerol

fraction from chylomicrons, plasma acetaminophen levels, and plasma satiety hormone levels by

members of the study team at the University of Guelph, in the future. All stored samples will be

analyzed within one year of sample collection. Apart from the analysis of the aforementioned

plasma measurements, samples will not be stored for any future analysis.

STUDY RESULTS PUBLICATION

Results from this study may be published but will always be presented as group data with no

ability to link data back to individuals. Your decision to be a participant in this study is voluntary

and you are free to withdraw from the study at any time. Upon withdrawal from the study or

study completion, you will be invited to complete a study exit questionnaire where you can

indicate if you would like a summary of your individual and overall study results mailed to you.

Also, please note that the results from this study will not be obtained from a licensed medical

laboratory and thus should not be used for diagnostic purposes. If you have concerns about

your results you are advised to consult a physician.

POTENTIAL RISKS AND DISCOMFORTS

There are minimum risks associated with participation in this study. The following summarizes

the potential risks:

At each of the two study visits, a phlebotomist will obtain intravenous blood samples. There is a

chance that this process could cause you some discomfort. These risks and potential discomforts

from the blood draws will be managed by having qualified and experienced personnel obtaining

your blood samples. In addition, consuming plenty of water the night before and the morning of

the study visit can facilitate the flow of blood. Any risk of bruising can be minimized by applying

pressure with gauze to the site after the catheter has been removed.

Participants will be required to drink lipid-rich emulsion beverages, in relatively large amounts

(i.e. 500 mL) and this might feel uncomfortable. This volume, however, is only slightly larger than

a Tim Horton’s large coffee (~415 mL) and less than a bottle of Gatorade (~591 mL). Participants

133

will be given encouragement to finish within 15 minutes, but also time to consume the treatment

at their own pace. The amount of fat, while relatively high (50 g) is similar to a breakfast meal of

McDonald’s sausage, egg and cheese McGriddles, hashbrowns, and a medium coffee with double

cream (~50 g fat) or less than a big Mac sandwich with medium fries and a milk shake (a total of

~60 g) or Burger King Double Whopper with cheese (~68 g fat). The emulsion beverage is not a

commercial product and might not taste very pleasant, even when flavored with vanilla extract

and the sweetener Sugar Twin.

Also, at each study visit, you will be asked to consume 1500 mg of acetaminophen suspended in

water, along with the 500 mL of the emulsion beverage in order to measure the rate of gastric

emptying. The 1500 mg dose is equivalent to three extra strength Tylenol pills, and is less than

half of the maximum recommended daily intake (4000 mg). However, you are advised to consult

a physician and notify the research team if you experience any of the following potential side-

effects: wheezing, rash, itching, increased sweating, nausea, vomiting, stomach pain, and loss of

appetite. As above, you should avoid over-the-counter acetaminophen-containing products for

48 hours prior to your study visits, and for 48 hours after your study visits to avoid over-

consumption of acetaminophen (>4000 mg in 24 h), which may lead to liver damage.

On each study visit, you will consume 50 g of fat in 500 mL emulsion beverage. This could be

more fat than you would typically consume in a meal. Therefore, you might experience

uncomfortable gastrointestinal side effects, such as nausea, fatty stools, greater sense of urgency

of bowel movements, etc. Based on the blood sampling, you are advised to avoid strenuous

activity for the remainder of each study visit day and should exercise caution in returning to

activities until you are confident any possible gastrointestinal issues have resolved.

POTENTIAL BENEFITS TO PARTICIPANTS AND/OR SOCIETY

There is very minimal personal benefit from participating in this study, although you will gain

the experience as a study participant. However, your involvement will lead to results that will

provide valuable insights into the digestion, absorption, and blood lipid changes after

consuming solid and liquid lipid emulsions.

COMPENSATION

You will be financially compensated for your participation in this study. You will receive $ 75 for

participating in each study visit for a total of $ 150 when you complete both study visits. Once

you send in the 24 h check-in questionnaire for the 2nd visit, you will receive a bonus $50. If for

any reason, you have to withdraw from the study, the cash payment will be prorated based on

your commitment. You will receive any financial compensation for the study visits in the form

134

of a cheque. For this purpose, we will collect your mailing address – to be used for this purpose

only.

COST OF PARTICIPATION

There is no direct cost for participating in this study. However, you will be responsible for any

costs related to attending your scheduled study visits at the HNRU, such as gas money, public

transportation fees, parking fees, etc.

PARTICIPATION AND WITHDRAWAL

You can choose whether to be in this study or not. If you volunteer to be in this study, you may

withdraw at any time without consequences of any kind. You may exercise the option of

removing your data from the study. You may also refuse to answer any questions you don’t want

to answer and still remain in the study. Please note that, if you decide to withdraw your data

from the study after completing both study visits, you will be able to do so within up to one week

of completing your second study visit. If information becomes available that may be relevant to

your willingness to continue participating in the trial, you will be informed in a timely manner.

The investigator may withdraw you from this research if circumstances arise that warrant doing

so.

CONFIDENTIALITY

Every effort will be made to ensure confidentiality of any identifying information that is obtained

in connection with this study. All participants will be assigned a number, and a study code will be

used. Your name will never be used in communicating results of the study. Records will be kept

on an encrypted computer and/or in a locked file cabinet in the faculty investigator’s locked

office. De-identified data and the master list of identifiable information will be kept until

publication of the results. In following these guidelines, your confidentiality will be maintained

to the best of our ability. The University of Guelph will permit trial-related monitoring, audits,

REB review, and regulatory inspection(s), providing direct access to source data/documents as

required. Results from the study may be published but will be presented as group data.

RIGHTS OF RESEARCH PARTICIPANTS

You are not waiving any legal claims, rights or remedies because of your participation in this

research study. This study has been reviewed and received ethics clearance through the

University of Guelph Research Ethics Board (REB 18-01-005). If you have questions regarding

your rights and welfare as a research participant, please contact:

Director, Research Ethics, (519) 824-4120, ext. 56606; sauld@uoguelph.ca

135

SIGNATURE OF RESEARCH PARTICIPANT

I have read the information provided for the screening visit for the study “The Lipid Study. Compare the changes in blood lipids and feelings of satiety after consumption of oil-in-water emulsions in which the droplets are in either the liquid or solid (i.e. crystalline) state, as described herein. My questions have been answered to my satisfaction, and I agree to participate in this study. I have been given a copy of this form.

Name of Participant: ____________________________________________________

(please print)

Signature of Participant: _________________________ Date: ________________

SIGNATURE OF WITNESS

Name of Witness: ________________________________________________________

(please print)

Signature of Witness: _________________________ Date: _______________

136

Appendix 1

The Lipid Physical State Study

Food Items & Ingredient Listings for Standardized Pre-trial Meals

The standard supper meal includes a choice of one entrée (options with meat from Lean Cuisine and Healthy Choice or vegetarian options from PC Blue menu), a pudding cup (Snack Pack vanilla or banana cream pie), a granola bar, and a beverage (apple, apple and grape or water). Participants will rank 5 entrées from the list provided below according to their preference and the meals will be purchased based on their market availability. For example, if a participant’s preferred choice is not available, he will receive his second-best choice. The accompanying information will be available to them at the time of their selection. Participants will be asked to consume the exact same meal before each trial day.

Choice of Entrée

“Lean Cuisine” Entrée options

1. Grilled Chicken Carbonara (280 cal; 8g fat; 4g prot; 38g CHO)

Description: Seasoned chicken in a bellavitano and parmesan sauce, topped with

crumbled bacon.

Ingredients: Cooked linguine (water, durum wheat semolina, wheat gluten) , water, cooked seasoned chicken (white chicken meat, water, soy protein isolate, modified corn and tapioca starches, seasoning, sodium phosphate, corn maltodextrin, salt) ,milk ingredients, parmesan and bellavitano cheeses (modifiedmilk ingredients, salt, microbial enzymes, lipase, cellulose) , bacon (pork, water, salt, sugar, smoke flavour, sodium phosphate, sodium erythorbate, sodium nitrite) ,soy oil, modified corn starch, salt, modifiedmilk ingredients, modified rice starch, potassium chloride, parsley flakes, lactic acid, xanthan gum, spice, flavours.

2. Grilled chicken and vegetables (250cal; 6g fat; 17g prot; 35g CHO)

Description: A blend of seasoned chicken, rigatoni, and vegetables topped with a

savoury sauce that’s accented with herbs, garlic, and cheese.

Ingredients: Cooked Rigatoni Pasta (Water, Durum Wheat Semolina, Wheat Gluten) , Water, Cooked Seasoned Chicken (White Meat Chicken, Water, Soy Protein Isolate, Modified Corn and Tapioca Starches, Corn Maltodextrin, Salt, Sodium Phosphate, Seasoning) , Tomatoes in Juice (Contain Citric Acid [Acidulant], Calcium Chloride) ,Yellow Zucchini, Broccoli, Carrots, Parmesan and Romano Cheeses (from Milk)

137

,Modified Corn Starch, Onions, Cider Vinegar, Tomato Paste, Salt, Sugar, Garlic Purée, Soy Oil, Olive Oil, Brown Sugar, Yeast Extract, Basil, Oregano, Potassium Chloride, Flavour, Spices.

3. Chicken Fettuccine (300cal; 7g fat; 20g prot; 41g CHO)

Description: Seasoned chicken, fettuccini, broccoli and zucchini tossed in a creamy

parmesan and Romano sauce. Ingredients: Cooked fettuccine pasta (water, semolina wheat flour, wheat gluten) , water, cooked seasoned chicken (white chicken meat, seasoning [salt, maltodextrin, garlic, sugar, spices, dextrose, autolyzed yeast extract, carrageenan, orange peel, onion, soy oil, chicken fat, paprika, chicken broth, dehydrated celery, flavours], soy protein isolate, modified corn starch, colour, maltodextrin, sodium phosphate, canola oil) ,yellow zucchini, broccoli, parmesan and romano cheeses, modified milk ingredients, soy oil, modified corn starch, parmesan cheese product (granular and parmesan cheeses [milk ingredients, bacterial cultures, salt, microbial enzymes], water, salt, lactic acid, citric acid [acidulant]) ,salt, rice starch, autolyzed yeast extract, mono- and diglycerides, spices, potassium chloride, xanthan gum, colour, wheat starch, flavour.

4. Chicken Parmigiana (300cal; 9g fat; 18g prot; 38g CHO) Description: tender pieces of white meat chicken topped with mozzarella and parmesan cheese, and a zesty tomato sauce. Ingredients: Pasta (Water, Durum Wheat Semolina) , Tomatoes in Juice (Contain Citric Acid [Acidulant], Calcium Chloride) , Breaded Seasoned Chicken (White Meat Chicken, Water, Wheat Flour, Soy Protein Isolate, Dextrose, Corn Maltodextrin, Soy Oil, Parmesan Cheese, Modified Milk Ingredient, Salt, Sodium Phosphate, Onion and Garlic Powder, Dried Parsley, Colour, Flavour, Seasoning) , Water, Zucchini, Part Skim Mozzarella Cheese, Tomato Paste, Onions, Sugar, Modified Corn Starch, Garlic Purée, Salt, Soy Oil, Basil, Potassium Chloride, Spices, Xanthan Gum.

5. Meat lasagna (310cal; 8g fat; 19g prot; 41g CHO)

Description: Delicious lasagna made with a hearty meat sauce and a blend of

mozzarella, parmesan and cottage cheese. Ingredients: Cooked lasagna (water, wheat semolina) , tomatoes (tomatoes, tomato juice, citric acid [acidulant], calcium chloride) ,part skim mozzarella, parmesan and cottage cheese (milk ingredients, salt, bacterial cultures, microbial enzymes, lipase, cellulose) , water, ground pork, tomato paste, onions, sugar, textured soy protein concentrate, garlic, ground beef, modified corn starch, roastedgarlic, basil, salt, flavour (soy sauce [water, soybeans, wheat, salt], autolyzed yeast extract, soy oil, soy lecithin, flavour) ,potassium chloride, spices, garlic powder, xanthan gum, onion powder, colour.

“Healthy Choice” Entrée options

138

6. Meatball Marinara (280 cal; 6g fat, 18g prot; 38g CHO) Description: meatballs, whole grain penne pasta and freshly chopped spinach in a zesty Parmesan Marinara sauce. Ingredients: information not available online

7. Mediterranean Grilled Chicken (270 cal; 4g fat; 20g prot; 38g CHO) Description: seasoned chicken breast, drizzled with a tangy balsamic-garlic glaze and paired with a garden-inspired medley of broccoli, mushrooms, spinach and sun-ripened tomatoes Ingredients: cooked whole grain rigatoni pasta (water, whole grain wheat flour, durum wheat semolina, dried egg whites), cooked seasoned grilled chicken chunks (chicken breast, water, olive oil, soy protein isolate product [soy protein isolate, modified potato and corn starches, corn starch, carageenan, soy lecithin (emulsifier)], dextrose, salt, potassium and sodium phosphates, spice extractives, potassium chloride, onion, garlic, paprika, caramel), water, broccoli, mushrooms, spinach, diced tomatoes, onions, sherry wine, red peppers, balsamic vinegar, sugar, onion purée (onions, high maltose corn syrup, salt), modified corn starch, garlic purée (garlic, water, phosphoric acid, xanthan gum, sorbic acid, garlic extract), chicken base (chicken meat and chicken juices, salt, chicken fat, sugar, hydrolyzed (corn, wheat gluten and soy) protein, dried whey, maltodextrin, yeast extract, natural flavouring, disodium inosinate and disodium guanylate, natural spice extractives), extra virgin olive oil, salt, parsley, spices._x000d_ _x000d_ contains: wheat, soy, milk, egg, sulphites

8. Grilled chicken red pepper Alfredo (260 cal; 6g fat; 22g prot; 28g CHO) Description: A thick and creamy red pepper Alfredo sauce with tender, seasoned chicken breast, broccoli and red bell peppers, and topped with a sprinkle of parmesan cheese

Ingredients: cooked linguini pasta (water, enriched durum semolina, soybean oil, dried egg whites), broccoli, cooked seasoned grilled chicken chunks (chicken breast, water, olive oil, soy protein isolate product [soy protein isolate, modified corn and potato starches, corn starch, carrageenan, soy lecithin], dextrose, potassium chloride, sun dried tomatoes [sulphites], salt, sodium phosphates, dehydrated onion and garlic, spices, spice extractives, caramel), water, roasted red pepper puree (sweet red pepper, high maltose corn syrup solids, salt), red peppers, parmesan cheese, milk ingredients, modified corn starch, whey protein concentrate, red pepper sauce (vinegar, red pepper, salt), salt, dehydrated garlic, potassium chloride, locust bean gum, sodium phosphate, spices. contains: wheat, soy, milk, egg, sulphites.

9. Grilled Balsamic Chicken (320 cal; 8g fat; 21g prot; 42g CHO) Description: grilled chicken, sun-ripened tomatoes and roasted garlic, angel hair pasta and drizzled with a light and tangy balsamic vinaigrette.

139

Ingredients: cooked angel hair pasta with basil (water, enriched durum semolina, olive oil, basil, dried egg whites), cooked seasoned grilled chicken chunks (chicken breast, water, olive oil, soy protein isolate product [soy protein isolate, modified corn and potato starches, corn starch, carrageenan, soy lecithin], dextrose, potassium chloride, sun dried tomatoes [sulphites], salt, sodium phosphates, dehydrated onion and garlic, spices, spice extractives, caramel), tomatoes, water, raisin juice concentrate, olive oil, balsamic vinegar, sugar, roasted garlic, modified corn starch, spices, salt, potassium chloride, locust bean gum, diacetyl tartaric acid ester of mono and diglycerides, mono and diglycerides, corn oil, maltodextrin, dehydrated garlic, flavour._x000d_ _x000d_ contains: wheat, soy, egg, sulphites.

10. Grilled chicken marinara (270 cal, 4g fat; 20g prot; 38g CHO) Description: Grilled chicken and crisp broccoli florets, cooked whole-grain penne pasta and zesty marinara sauce.

Ingredients: cooked whole grain penne pasta (water, whole grain wheat flour, durum semolina, olive oil, dried egg whites), cooked seasoned grilled chicken chunks (chicken breast, water, olive oil, soy protein isolate product [soy protein isolate, modified corn and potato starches, corn starch, carrageenan, soy lecithin {emulsifier}], dextrose, potassium chloride, dehydrated garlic, salt, sodium phosphates, spices, spice extractives, caramel), broccoli, water, diced tomatoes (tomatoes, tomato juice, citric acid [acidulant],calcium chloride), tomato paste, parmesan cheese, brown sugar, onions, garlic, salt, modified corn starch, spices, potassium chloride, xanthan gum._x000d_ _x000d_ contains: wheat, soy, milk, egg

11. Grilled chicken linguini (290 cal; 6g fat; 20g prot; 38g CHO) Description: semolina linguini, tomato-basil sauce, grilled chicken and vegetables

Ingredients: cooked linguini (water, enriched durum semolina, olive oil, dried egg whites, soybean oil, granulated garlic), cooked grilled basil chicken chunks (chicken breast, water, olive oil, soy protein isolate product [soy protein isolate, modified corn and potato starches, corn starch, carrageenan, soy lecithin], dextrose, potassium chloride, salt, sodium phosphate, dehydrated onion and garlic, basil, spices, spice extractives, caramel), water, broccoli, yellow zucchini, red peppers, tomato concentrate, basil, modified corn starch, tomato base (fire roasted tomatoes, maltodextrin, salt, tomato paste, corn syrup, corn oil, tomato flavour, sugar, modified corn starch), cream, basil blend (basil, roasted garlic puree, garlic puree, high maltose corn syrup, olive oil, water, salt), parmesan cheese, chardonnay wine (sulphites), rendered chicken fat, salt, garlic puree, sugar, carob gum, disodium phosphate, colour, spices. _x000d_ contains: wheat, soy, milk, egg, sulphites.

12. Grilled chicken pesto (310cal; 9g fat; 20g prot; 37g CHO)

Description: pesto sauce, grilled chicken, tender rotini pasta, and a medley of spinach and zucchini

140

Ingredients: cooked pasta (water, enriched durum wheat semolina, basil, parsley, olive oil, dried egg whites), cooked seasoned chicken chunks (chicken breast, water, olive oil, soy protein isolate product [soy protein isolate, modified corn and potato starches, corn starch, carrageenan, soy lecithin], dextrose, potassium chloride, sun dried tomatoes, salt, sodium phosphates, dehydrated onion and garlic, spices, spice extracts, caramel, flavour), water, green zucchini, yellow zucchini, spinach, chardonnay wine, parmesan-style cheese (skim milk, cheese culture, salt, enzymes, cellulose), olive oil, modified corn starch, basil, garlic, chicken base (chicken meat and chicken juices, salt, chicken fat, sugar, hydrolyzed [corn,wheat gluten, soy] protein, dried whey, maltodextrin, yeast extract, natural flavoring, disodium inosinate and disodium guanylate, extracts of turmeric and annatto), butter, onion, sugar, garlic puree, evaporated cane juice, green onions, parsley, salt, spice, sea salt, seasoning (sea salt, garlic, dehydrated onions, red peppers, soybean oil), acetylated tartaric acid esters of mono and di-glycerides, mono and di-glycerides. contains: wheat, soy, milk, egg, sulphites

“PC Blue Menu” Vegetarian Entrée options

13. Roasted Vegetable Lasagna (260 Cal; 6g fat; 16g prot; 35g CHO)

Description: Layers of multigrain pasta with ricotta and partly skimmed mozzarella

cheeses, roasted red and yellow peppers, eggplant and zucchini.

Ingredients: Tomato purée (water, concentrated tomato paste), cooked multigrain pasta [water, durum wheat semolina, multigrain blend (yellow peas, chickpeas, pearled barley, oats, flaxseeds, kamut wheat, wheat fibre), dried egg-white], partly skimmed mozzarella, light ricotta whey and ricotta whey cheeses (partly skimmed milk, whey powder, cream, vinegar, carrageenan, cheese cultures, salt, microbial enzymes, cellulose), tomatoes (contain tomato juice, citric acid, calcium chloride), roasted vegetables (red and yellow bell peppers, eggplant, zucchini, onion), water, modified corn starch, carrots, concentrated apple juice (contains ascorbic acid), extra virgin olive oil, garlic, salt, red wine vinegar (contains sulphites), lemon juice from concentrate, spinach, corn starch, herbs and spices, dehydrated garlic.

14. Fettucine Alfredo (240 Cal; 8g fat; 13g prot; 30g CHO)

Description: Fettuccine and broccoli in a rich Alfredo sauce made with Parmesan and

Romano cheeses, real cream, butter and garlic.

Ingredients: Pasta (wheat), milk, water, broccoli, cream (milk, cream, sodium citrate, sodium phosphate, carrageenan), parmesan cheese, Romano cheese, butter, garlic purée (garlic, water), modified corn starch, salt, canola oil, corn maltodextrin, xanthan gum, herbs and spices.

141

Choice of Dessert (A) Snack Pack Pudding, Vanilla Description: Made with skim milk, real ground vanilla beans and all natural flavours. No preservatives added.

Nutritional Information: per 1 pudding cup serving (99 g) %DV Calories 100 Cal Fat 3 g 5 Saturates 1.5 g 8 Trans 0 g Cholesterol 0 mg 0 Sodium 130 mg 5 Carbohydrate 18 g 6 Fibre 0 g 0 Sugars 13 g Protein 0.4 g

Vitamin A 0 Vitamin C 0 Calcium 320 mg 2 Iron 2 Ingredients: Water, Skim milk from concentrate (water, concentrated skim milk), sugar, modified corn starch, modified palm oil, artificial and natural flavors, salt, sodium stearoyl-2-lactylate, color (contains Tartrazine)

(B) Snack Pack Pudding, Banana Cream Pie Description: Made with skim milk, and all natural flavours. No preservatives added.

Nutritional Information:

per 1 pudding cup serving (99 g) %DV Calories 110 Cal Fat 3.5 g 5 Saturates 1.0 g 5 Trans 0 g Cholesterol 0 mg 0 Sodium 150 mg 6 Carbohydrate 19 g

6

Fibre 1.0 g 4 Sugars 14 g Protein 0.3 g

Vitamin A 0

Vitamin C 0 Calcium 2 Iron 0

Ingredients: Water, Skim milk from concentrate (water, concentrated skim milk), sugar, modified corn starch, modified palm oil, modified corn starch, banana puree, salt, sodium stearoyl-2-lactylate, sodium phoshphate, artificial and natural flavors

142

Granola Bar Nature Valley, Chewy Trail Mix Bars, Fruit & Nut Nutritional Information:

per 1 bar (35 g) %DV

Calories 140 Cal

Fat 4 g 6

Saturates 0.5 g 2

Trans 0 g

Cholesterol 0 mg

Sodium 65 mg 3

Carbohydrate 24 g 8

Fibre 2 g 8

Sugars 7 g

Protein 3 g

Vitamin A 0

Vitamin C 0

Calcium 2

Iron 4

Ingredients: granola (whole grain rolled oats, sugar, canola oil, fructose, salt, baking soda, soy lecithin, rosemary extract), corn syrup, mixed berry flavoured fruit pieces (sugar, cranberries, blueberry juice from concentrate, grape juice concentrate, citric acid, elderberry juice concentrate, natural fruit flavours, sunflower oil), rice flour, almonds, glycerin, honey, whole oat flour, sugar, fructose, corn starch, canola oil, soy lecithin, natural flavour, barley malt extract, refined peanut oil, salt, citric acid, monoglycerides, tocopherols, rice extract, rosemary extract.

Choice of Beverage Apple Juice

(B) Apple & Grape Juice

Ingredients: pure filtered water, concentrated apple juice, less than 0.5% of: calcium citrate (calcium source), ascorbic acid (vitamin C), and potassium phosphate

143

(c) Water.

Ingredients: Contains apple, grape and pear juices from concentrate, less than 2% of each: ascorbic acid (vitamin C), calcium citrate (calcium source), citric acid, nature

144

Appendix 2

The Lipid Physical State Study

Post-trial meal options: “Need a Pita”

At the end of each study visit, we will provide you with a meal including a choice of a pita, chips and a beverage. Please select from the options listed below.

PICK A PITA

☐ Black Forest Ham

☐ Roast Beef

☐ Gyros

☐ B.L.T

☐ Garden

☐ Falafel

☐ Baba Ganoush

☐ Hummus

CHOOSE YOUR TOPPINGS VEGGIES

☐ Lettuce

☐ Spinach ☐ Tomato ☐ Onions ☐ Cucumber ☐ Pickles ☐ Green Peppers ☐ Hot Peppers ☐ Olives ☐ Alfalfa Sprouts

☐ Mushrooms ☐ Pineapple

SAUCES & DRESSINGS

☐ Tzatziki

☐ Mayo ☐ Mustard ☐ Caesar ☐ Ranch ☐ Honey Mustard ☐ Dijon Mustard

145

☐ BBQ Sauce ☐ Hot Sauce ☐ Teriyaki

☐ Secret Sauce ☐ Italian ☐ Ancho Chipotle

CHEESES

☐ Feta

☐ Cheddar ☐ Swiss ☐ Parmesan

POTATO CHIPS (Please rank your top choices from 1-3)

☐ Lay’s Original

☐ Lay’s Salt & Vinegar

☐ Lay’s BBQ ☐ Lay’s Ketchup

☐ Doritos Nacho Cheese

☐ Ruffles All Dresses ☐ Cheetos Puffs

BEVERAGES

☐ Water ☐ Apple Juice ☐ Grape Juice

146

Appendix 3 The Lipid Digestion Study Participant Recruitment Poster

147

Appendix 4 The Lipid Digestion Study Phone Screening Questionnaire

Screening Number: ____________________ REB Number: ______________________

Researcher Initials: _____________________ Principal Investigators: Amanda Wright, Ph.D.

Date: ________________________________

PHONE SCREENING PARTICIPANT ELIGIBILITY QUESTIONNAIRE

The Lipid Digestion Study: Impact of emulsion droplet physical state on postprandial

lipemia and satiety in healthy human adult participants

HNRU Study Coordinator: SCREENING #: S___________

Date: Time:

Name of caller: Sex:

Phone #: Email:

Best way to get in touch:

Hello, my name is _____ , a MSc./ Ph.D Student in Human Health and Nutritional Science at the

University of Guelph. Thank you for your interest in The Lipid Digestion Study. I am going to

provide you with some important details about the study to ensure your interest and will then ask

you a few questions to see if you are eligible for the study. This will take approximately 15-20

minutes. Do you have time right now?

If NO: When would be a better time to contact you?

If YES: Great. Please feel free to ask me questions at any time. I’ll also be checking throughout

to make sure the information presented is clear.

The study will take place at the University of Guelph, in the lab called the Human Nutraceutical

Research Unit (HNRU) where we run human nutrition studies. The purpose of the study is to

compare the changes in blood lipids and feelings of satiety after participants consume two different

lipid rich oil-in-water emulsions. The emulsions are made from the same ingredients, but differ in

148

terms of their physical states, i.e. in one, the droplets are liquid and in the other they are solid (ie.

crystalline).

At this point, are you still interested in participating in this study?

If NO: No problem, thank you for your time.

If YES: (Continue…)

If you participate in this study, we will ask that you consume these two flavored emulsion

beverages – one each - on two separate study visits. You will be asked to visit the HNRU for ~7

hours on each study visit day, following an 10-12 hour overnight fast. A trained and approved

medical technician will then insert an intravenous catheter into a vein of your forearm and draw a

blood sample. After the fasting blood sample is taken, you will drink the flavored emulsion

beverage. You will also be asked to consume 1500 mg of acetaminophen (3 extra-strength

Tylenol tablets crushed) suspended in 50 mL water. This is part of a research technique that

allows us to determine how quickly food empties from your stomach because we can determine

how quickly the acetaminophen appears in your blood.

Do you have any questions or concerns at this point?

If YES: Discuss.

If NO: (Continue…)

After you ingest the test meal, you will hang around the HNRU for 6 hours. Blood samples will

be drawn from the catheter, periodically, and you will also rate your feelings of satiety

periodically on paper questionnaires. There will be magazines and movies available, and you are

invited to bring your laptop, books, etc. from home to pass the time. Your blood samples will be

processed at the HNRU for analysis of blood triacylglycerols, free fatty acids, chylomicron size

and fatty acid composition and satiety hormone markers. These are different parameters that tell

us about how your body digested and metabolized the lipids you consumed. At the end of each

visit, the iv catheter will be removed and you will be given a Subway sub of your choice to

consume before leaving the HNRU. You will consume the second test beverage at a second

study visit after a minimum of a 7-day washout period. The study protocols will be similar to the

first study visit. You will be compensated financially, where you will receive $200 after

completing both study visits. The payments will be prorated and if you complete only one study

visit, you will receive $75. Once you have completed both visits and have submitted the 24-h

check-in questionnaire for the 2nd visit, you will receive a bonus of $50.

Are you still interested in participating in this study?

If NO: No problem, thank you for your time.

If YES: (Continue…)

Finally, if you participate in this study, we ask you to avoid alcohol, caffeinated beverages and

food and over the counter medication for 48 hours prior to the study visit. You will also be asked

to avoid strenuous physical activity 48 hour prior to each study visit. This is very important for

the study protocol because we want to maintain normal lipid digestion without these interference

factors in order to determine how different the solid and liquid emulsion droplets are digested.

149

Does this sound like a study that you would be interested in participating in? Do you wish to

continue with the pre-screening questionnaire?

IF NO: Thank you for your time.

IF YES: We will be asking you some questions regarding your health and lifestyle as, these

factors could impact our study results. If you do not feel comfortable answering any question,

you do not have to answer it. The information you do provide will be kept confidential and will

only be seen by the researchers involved in this study. Can we proceed with the questionnaire?

1. How did you hear about the Lipid Study?

☐ Poster ☐ Friend ☐ Department Email ☐ Other:

2. How old are you?

3. What is your height?

4. What is your weight?

Space for BMI Calculation (Range 18 < BMI < 26)

5. How often do you smoke tobacco or cannabis?

YES NO

6. (a) How often do you consume alcohol per week?

(b) How much alcohol do you consume at one sitting?

150

7. How would you describe your health?

Poor Fair Good Very Good Excellent

8. Do you have any diseases or medical conditions (E.g. Liver or renal

disease, cardiovascular disease – including high blood pressure,

diabetes etc.)?

If YES, please describe:

YES NO

9. Do you have any gastrointestinal disorders (Celiac disease, gluten

intolerant, lactose intolerance, irritable bowel syndrome, etc.?)

If YES, please specify:

YES NO

10. Do you have any food allergies?

If YES, please state what the allergy is and describe your reaction:

YES NO

11. Do you have any anaphylactic or life-threatening allergies? (Foods,

venoms, latex, etc.)

If YES, please specify:

YES NO

12. Are you currently taking any prescription medications?

If YES, what are they?

If YES, where they recommended to you by a healthcare professional?

YES NO

151

13. Are you currently taking any over-the-counter medications?

If YES, what are they?

If YES, where they recommended to you by a healthcare professional?

YES NO

14. Have you ever taken the medication “acetaminophen” either

prescribed or over-the-counter? (A popular trade name is Tylenol®).

(a) If YES, how often?

(b) Have you ever had a sensitivity or allergic reaction to

acetaminophen?

(c) Have you ever been advised by a medical practitioner to avoid

using acetaminophen?

YES NO

YES NO

YES NO

15. (a) Are you currently taking any natural health products or

supplements? This includes all vitamins and minerals, herbal remedies

etc.

If YES, what are they?

If YES, where they recommended to you by a healthcare professional?

(b) Are you taking any muscle building supplements?

If YES, please provide the following information:

- The brand name & main ingredient/s if known:

YES NO

YES NO

YES NO

152

- The dose (amount/day):

- For how long have you being taking it?

16. Have you taken oral antibiotics in the past 3 months, or plan to take

oral antibiotics in the next 3 months? YES NO

17. Are you currently trying to lose or gain weight? YES NO

18. Elite athletes are defined as anyone currently competing as a varsity

player (individual or team), a professional player or a national or

international level player. Training athletes are participating in intense

practice and exercise for individual or team events. Based on this

explanation, do you consider yourself an elite or training athlete?

YES NO

19. Do you have any specific dietary patterns/choices, restrictions, or any

food that you avoid?

YES NO

20. Have you ever had a sensitivity or reaction to fat/high fat in the diet?

If YES, please explain.

YES NO

21. This study involves consuming 500 mL of a flavored lipid based

emulsion which contains 50 g of fat. The amount of fat is similar to a

McDonald’s sausage, egg and cheese McGriddles, hashbrowns, and a

medium coffee with double cream (~50 g fat) and less than a

McDonald’s big Mac sandwich with medium fries and a milk shake

(~60 g fat) or a Double Whopper with cheese from Burger King (~68

g fat). Do you have any concerns about this?

YES NO

153

If YES, please explain.

22. Have you ever had problems with your gall bladder or had your gall

bladder removed?

YES NO

23. The treatment emulsion beverage will be sweetened with an artificial

sweetener “Sugar Twin” Sucralose which is commonly used for

sweetening coffee or tea without adding calories and flavored with

artificial vanilla flavor which is a commonly used food flavor. Do you

have any concerns about consuming these products?

If YES, please explain.

YES NO

24. This study requires participants to provide venous blood samples by

intravenous (IV) catheter. Do you have any concerns about this? If

they ask what this is: An intravenous catheter is a small flexible tube

that is inserted using a needle into a vein in your arm for easy flow of

blood over multiple time points.

YES NO

25. The study visits require participants to fast for 10-12 hours overnight.

Water is permitted until 30 minutes before each study visit. Would you

be comfortable with this?

YES NO

26. This study requires you to avoid consuming any acetaminophen

products 48 hour prior to and 48 hours after each study visit. Do you

have any concerns about this?

YES NO

27. This study requires you to avoid consuming alcohol 48 hours prior to

and 48 hours after each study visit. Do you have any concerns about

this?

YES NO

28. This study requires you to avoid consuming caffeine containing

beverages like coffee, tea, pop, sports drinks and food like chocolate

for 48 hours prior to the study visit. Do you have any concerns about

this?

YES NO

154

29. This study requires you to avoid strenuous physical activities for 48

hours prior to and 24 hours after each study visit. Do you have any

concerns about this?

YES NO

30. I will review the information from this questionnaire with other

members of the study team. If you seem eligible to continue in the

screening process, are you able to come to the HNRU at the University

of Guelph for a 30 minutes - 1 hr in-person screening and orientation

visit?

YES NO

31. If eligible to participate in the study, you will be asked to visit the

University of Guelph to attend two, 7 hr study visits on two week

days, separated by at least one week. Would you be able to

accommodate this request?

YES NO

This concludes the questionnaire. Do you have any questions?

The study coordinators will carefully review each participant’s information case by case, with

the faculty investigator and decide on his eligibility to participate in the study. Participants will

be contacted by phone or e-mail as per their preferred method of communication, and inform the

decision of the study group whether a particular participant is eligible or ineligible as mentioned

below.

INELIGIBLE

Based on the questionnaire, you have not met the requirements for eligibility. We hope you

will consider the HNRU for future studies. Please check the HNRU website for our current

studies. Thank you for your time and interest. Have a nice day.

● Fill out the Phone Screening on the Tracking Log and Trial Activity Checklist.

ELIGIBLE

Based on the questionnaire, you have met the initial eligibility requirements. Are you still

interested in moving further with the study?

If NO: Thank you for your time.

If YES: We would like to set up an in-person screening visit which will involve filling out a

more detailed questionnaire and give you the opportunity to learn more about the study.

We ask you to fast for 10-12 hours prior to your arrival at the HNRU so that we can take

155

a fasting finger prick blood sample. A $5 Tim Horton’s gift card will be given to you as

compensation for your in-person screening. Thank you again for your time, we will

contact you with specific details about booking this visit and send a reminder closer to the

date. Included in the email will be a map and directions for finding the HNRU, in case

you do not know where we are located. We appreciate the time you are investing in our

study. Do you have any more questions?

● Set up an in-person screening visit

● Enter the scheduled date and time of the screening visit into the Google calendar.

● Fill out the Phone Screening on the Tracking Log and Trial Activity Checklist.

● Arrange to send the potential participant a map to the HNRU and provide parking

information.

In-person screening visit scheduled for:

Date: Time: Coordinator Signature:

156

Appendix 5 The Lipid Digestion Study In-Person Screening Eligibility Questionnaire

Screening Number: ____________________ REB Number: ______________________ Researcher Initials: _____________________ Principal Investigator: Amanda Wright, Ph.D. Date: ________________________________

IN-PERSON SCREENING PARTICIPANT ELIGIBILITY QUESTIONNAIRE

The Lipid Digestion Study: Impact of emulsion droplet physical state on postprandial lipemia and satiety in healthy human adult participants

HNRU Coordinator: SCREENING #: S________

Date:

Time:

Thank you very much for your interest in this study. The purpose of this questionnaire is to gather more information about you and to ensure your safety as a participant in this study. Please feel free to NOT answer any questions that you are uncomfortable with answering and to ask the study coordinator any questions you might have.

All information provided in this questionnaire will be kept strictly confidential.

The study coordinator will measure your body weight, height, and blood pressure.

HEIGHT (cm)

WEIGHT (kg)

157

BMI (kg/m2)

BLOOD PRESSURE (mmHg)

How are you feeling today (e.g. good, tired, stressed, sick, etc)?

1. How would you describe your general health?

POOR FAIR GOOD VERY GOOD EXCELLENT

2. Have you considerably lost or gain weight during the past 3 months?

YES / NO

If YES,

(a) Was the loss/gain of weight intentional? YES /

NO

(b)Did you change your regular dietary pattern to achieve this weight gain/loss?

YES / NO

(c) Are you still following this modified dietary pattern to lose/gain weight further?

YES / NO

(d)How much weight did you lose/gain overall during the past 3 months?

3. Do you have any history of major diseases or any medical conditions (liver, renal or

heart disease, diabetes, etc)?

YES / NO

If YES, please specify:

158

4. Do you have gastrointestinal disorders (Celiac disease, lactose intolerance, irritable

bowel syndrome, gall bladder disease, etc.)?

YES / NO

If YES, please specify:

5. Do you smoke?

YES / NO

a) If NO, have you ever smoked?

YES / NO

b) If YES, how long since you quitted smoking?

6. Do you use recreational drugs?

YES / NO

7. Have you taken oral antibiotics in the past 3 months, or plan to take oral antibiotics in

the next 3 months?

YES

/ NO

8. Do you use any prescribed and/or over-the-counter medications (e.g. Tylenol,

Aspirin, Sudafed etc)?

YES / NO

If YES, please complete the following table:

PRODUCT

NAME

REASON

FOR USE

MEDICINAL

INGREDIENTS

DOSE

HOW

OFTEN

HOW

LONG

9. Do you use any dietary supplements (i.e. vitamins, protein powders, antioxidants)?

YES / NO

If YES, please complete the following table:

159

PRODUCT

NAME

REASON

FOR USE

ACTIVE

INGREDIENTS

DOSE

HOW

OFTE

N

HOW

LONG

10. This study requires that participants avoid taking non-prescribed over the counter

medications (e.g. Tylenol) for the 48 hours before a study visit.

Would you be comfortable with this? YES /

NO

11. Do you have ANY allergies (i.e. food, medications, ragweed or pollen)?

YES / NO

a) If YES, please list:

b) Please describe your type of allergic reactions: (i.e. anaphylaxis, difficulty, breathing,

swelling, etc.)

12. The two main ingredients in this study product is palm stearin and the emulsifier

sorbitan monostearate (Span 60). Do you know whether you have any sensitivity or

allergy to any of them? YES / NO

If YES, please describe what reactions:

13. Have you ever used the medication called ‘acetaminophen’? This medicinal

ingredient is found in these products.

Tylenol® and Tylenol Cold®, Alka-Seltzer Plus Liquid Gels®, Dayquil®, Dimetapp®, Excedrin®, Midol®, Nyquil®, Robitussin®, Sinutab®, and Sudafed®

YES / NO

160

If YES, (a) when did you last use it?

(b) How often do you use it?

14. Have you ever had a sensitivity or allergic reaction to acetaminophen containing

products? YES /

NO

15. Has a medical professional ever advised you to avoid taking any acetaminophen

containing products? YES /

NO

16. This study requires participants to consume the equivalent of 3 extra strength Tylenol

capsules with the study beverage, at each visit. This is 1,500 mg of acetaminophen,

where the maximum daily limit set by Health Canada is 4,000 mg. Would you be

comfortable doing this?

YES / NO

17. Because we don’t want participants to reach upper limits of acetaminophen intake,

this study requires participants to NOT consume any acetaminophen containing

product 48 hour prior and 48 hours after each study visit. Would you be comfortable

with this?

YES / NO

18. How would you describe your fitness/activity level?

VERY

LOW

LOW MODERATE HIGH VERY HIGH

19. Do you participate in any of the following types of exercises?:

a) Weight training? YES NO If YES, how often? ____________________

b) Running/jogging? YES NO If YES, how often? ____________________

c) Aerobics? YES NO If YES, how often? ____________________

d) Team sports? YES NO If YES, how often? ____________________

e) Other? (please list) YES NO

_____________________________________ How often? _____________________

_____________________________________ How often? _____________________

_____________________________________ How often? _____________________

161

20. This study requires that participants DO NOT participate in vigorous exercise for 48

hours prior to and 24 hours after the HNRU study visits (a total of 2 occasions).

Would you be comfortable with this?

YES / NO

21. Do you have unusual sleeping patterns (shift work etc.)?

YES / NO

22. How many hours of sleep do you usually get per night? 0 – 3 4 – 6 7 – 9 10+

23. Approximately how many alcoholic drinks do you consume per week?

(1 drink = 12 oz beer, 5 oz wine, or 1.5 oz hard liquor).

And at one typical sitting?

24. This study requires that participants DO NOT consume any alcoholic beverages for

48 hours prior to and 48 hours after the HNRU study visits (a total of 2 occasions).

Would you be comfortable with this?

YES / NO

25. Do you consume caffeinated beverages (coffee, tea, pop, energy drinks etc.) or food

(chocolate) frequently?

YES / NO

a) If YES, approximately how many caffeinated drinks do you consume per day?

b) If YES, please list the types of caffeinated beverages and food you consume:

26. This study requires that participants DO NOT consume any caffeinated beverages or

food for 48 hours prior to the HNRU study visits. Would you be comfortable with

this?

YES / NO

27. Are you following a vegan or vegetarian diet?

YES / NO

28. The study products are basically 500 mL of 10% fat emulsion beverages where there

would be 50 g of fat in each beverage. The amount of fat in the beverage is similar to

a breakfast meal consisting of a McDonald’s sausage, egg and cheese McGriddles,

hashbrowns, and a medium coffee with double cream (~50 g fat) or less than in a

162

meal consisting of a big Mac sandwich with medium fries and a milk shake (~60 g

fat) or a Burger King Double Whopper with cheese (~68 g fat).

(a) Are you avoiding high fat in your diet? YES /

NO

(b) Did you ever had any trouble consuming high amount of fat in one sitting?

Some high fat containing food items are: cheeseburgers, fried food like fries,

chicken nuggets, milkshakes, many desserts, shakes and specialty beverages, from

fast food chain restaurants.

YES / NO

If YES, please explain.

29. This study involves two separate study visits where, on each visit, you will be

consuming a 500 mL of an emulsion beverage consisting of 50 g of palm stearin in

bottled water within 15 minutes. This volume is slightly larger than a Tim Horton’s

large coffee (~415 mL) and less than a bottle of Gatorade (~591 mL).

Do you have any concerns about this?

YES / NO

30. The treatment beverage will be sweetened with Sugar Twin Sucralose, which is an

artificial sweetener commonly used to sweeten coffee, tea without adding extra

calories. Do you have any concerns with consuming this sweetener in the treatment

beverage?

YES / NO

31. The treatment beverage will be flavored with the artificial vanilla extract used

regularly in baking. Do you have any concerns with consuming this flavor in the

treatment beverage?

YES / NO

163

32. Have you ever had a reaction or sensitivity (i.e. indigestion, diarrhea, fatty stools) to

any fat containing products?

YES / NO

If YES, please describe:

33. This study requires participants to provide venous blood samples by intravenous (IV)

catheter. Do you have any concerns about this?

(An intravenous catheter is a small flexible tube that is inserted using a needle into a

vein in your arm for easy flow of blood over multiple time points.)

YES / NO

34. The study visits require participants to fast for 10-12 hours overnight. Water is

permitted until 30 minutes before each study visit. Would you be comfortable with

this?

YES / NO

35. Are you able to come to the HNRU at the University of Guelph for a 30 minutes - 1

hr in-person screening and orientation visit?

YES / NO

36. In this study, participants will be asked to remain at the HNRU, University of Guelph

on each study visit for approximately 7 hours (on two different days/7 days in

between). Would you be comfortable with this? YES /

NO

37. This study requires 2 visits (7 h each, minimum 7 days in between) to the HNRU at

the University of Guelph.

a) Can your schedule accommodate these visits?

YES / NO

b) Would you prefer 7-h study visits on Monday, Tuesday, Wednesday, or Thursdays?

38. Are you currently participating in any other research studies?

YES / NO

If YES, please describe:

164

________________________________

Graduate student Coordinator Signature