Microfluidic study of plasticity in pancreatic beta cell heterogeneity

15 1.11 Generating solute concentration gradients in microfluidic channels.. Our results show that a net GLUT-2/glucokinase activity is heterogeneous and does not map to extracellular gl

MICROFLUIDIC STUDY OF PLASTICITY IN PANCREATIC β-CELL

HETEROGENEITY

TAN CHERNG-WEN, DARREN

(B.Sc (Hons.), NUS)

A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY GRADUATE PROGRAMME IN BIOEGNINEERING YONG LOO LIN SCHOOL OF MEDICINE NATIONAL UNIVERSITY OF SINGAPORE

2009

ACKNOWLEDGMENTS

The author wishes to thank A/P Partha Roy for his invaluable guidance and generous support in the course of this work

The author also wishes to express deep appreciation for the technical guidance provided

by A/P Michael Raghunath and A/P Dieter Trau Appreciation is also due A/P Hanry

Yu, A/P Lanry Yung and A/P Zhang Yong for their helpful insights and suggestions during the Qualifying Examination

Critical technical assistance was provided by Mr Shashi Ranjan, Mr Ganesh Balasubramanian, Mr Vedula Sri Ram Krishna and Mr Tok Wee Lee, without which many of the obstacles in the course of this work would never have been surmounted Finally, the author wishes to express infinite gratitude to his wife, who shared these five years of long days and endless nights, whose patience, tolerance, encouragement and unflagging support provided him with the fuel to persevere and see the task to its end

TABLE OF CONTENTS

Table of Contents ii

SUMMARY vi

LIST OF TABLES ix

LIST OF FIGURES x

ABBREVIATIONS xiv

PART I 1

Chapter 1 2

Introduction and Literature Survey 2

1.1 Introduction 2

1.2 The islets of Langerhans 2

1.3 Insulin secretion and glucose homeostasis 3

1.4 Glucose sensing and activation of -cell function 4

1.5 Effect of glucose on first phase insulin secretion 9

1.6 Effect of glucose on second phase insulin secretion 9

1.7 -cell heterogeneity 10

1.8 The hypothesis 12

1.9 Three questions raised to address the hypothesis 13

1.10 Microfluidic channel cell culture 15

1.11 Generating solute concentration gradients in microfluidic channels 16

1.12 Organisation of the thesis 19

Chapter 2 21

Development of the hydrogel-assisted gradient generator (HAGG) 21

2.1 The proposed design 21

2.2 Microfluidic gradient generators 22

2.3 Overview of microfluidic system structure 28

2.4 Overview of fabrication process 30

2.5 Master mold fabrication 30

2.6 Micromolding 31

2.7 Microfluidic channel surface treatment with TPM 32

2.8 Hydrogel formation 33

2.9 Developing and disinfecting 39

Chapter 3 40

Application of the HAGG 40

3.1 Gradient generation 40

3.2 Cell culture in the HAGG 48

3.3 Gradient generation in the presence of cells 55

3.4 Selecting assays to address the three questions 58

3.4.1 Metabolism of 2-NBDG 59

3.4.2 Glucose-induced quinacrine secretion 60

3.4.3 Ca 2+ -induced intracellular Fura-2/acetoxymethylester fluorescence 60

3.5 -TC-6 characterization 61

3.5.1 -TC-6 metablism of 2-NBDG 61

3.5.2 Glucose-induced quinacrine secretion in -TC-6 64

3.5.3 Fura-2/AM staining of -TC-6 67

3.5.4 -TC-6 glucose uptake rate 70

3.6 Addressing the three questions 70

3.6.1 Ascertaining mercury lamp intensity variation 79

3.6.2 Net GLUT2 transporter/glucokinase activity under an imposed extracellular glucose gradient 80

3.6.3 Intracellular vesicle density after exposure to an imposed extracellular glucose gradient 87

3.6.4 Intracellular vesicle density after exposure to an imposed extracellular glucose gradient and glucose challenge 91

Chapter 4 95

Understanding the results 95

4.1 Microfluidic channel design 95

4.2 Transient-diffusion model of mass transport in the HAGG 96

4.3 Understanding gradient generation in the HAGG 100

4.3.1 Transient diffusion of 2-NBDG 100

4.3.2 Steady-state diffusion of 2-NBDG 102

4.3.3 Applying the model for experimental design 103

4.4 Cell culture in the HAGG 106

4.5 Gradient generation in presence of cells 110

4.6 -TC-6 characterization 113

4.7 Concerns over the use of UV excitation 118

4.8 Addressing the three questions 119

Chapter 5 124

Conclusion and future directions 124

PART II 128

Chapter 6 129

Microfluidic competition assay via equilibrium binding 129

6.1 Analyte detection 129

6.2 Competition assays 129

6.3 Microfluidic channel-based competition assays 130

6.4 Convection-diffusion model of microfluidic competition assay 130

6.5 Materials and Methods 134

6.6 Labeling of proteins with fluorescein 135

6.7 Conjugation of anti-insulin IgG with biotin 136

6.8 Microchannel fabrication 137

6.9 Microchannel surface blocking and IgG immobilization 137

6.10 Flow cell fabrication 137

6.11 Microchannel binding assays 138

6.12 Results 139

6.12.1 Comparison of elution profiles for aprotinin-FITC and insulin-FITC 139

6.12.2 Determination of insulin-FITC equilibrium dissociation constant, K d,L 140

6.12.3 Determination of unlabeled insulin equilibrium dissociation constant, K 142

6.13 Discussion 146

6.14 Conclusion 152

Literature cited 153

Appendix 1 – Fluorescein labeling kit 160

Appendix 2 – Chemical structure of reagents used 161

Appendix 3 – INSTECH P625/10K.143 peristaltic pump 162

Appendix 4 – Lists of P-values for the 2-sample t-tests 163

Appendix 5 – Abstracts of work presented at various conferences 164

within the context of this hypothesis and addressed with the help of two specific assays

The two assays are tested by designing an in vitro microfluidic system that allows the

establishment and control of glucose gradients over the micro-scale length of islets An immortalized insulin secreting cell line derived from transgenic mice, -TC-6 is applied in this test The gradient generating section of the microfluidic system comprises a source microchannel and a sink microchannel aligned in parallel These are connected by 24 perpendicular cross-channels, and converge at the end of the series of cross-channels to form a single eluent channel The main and cross-channels are about 105 m and 310 m

in width respectively Each cross-channel is about 317 m in length All the microchannels are 50 m in height The final product includes poly(ethylene glycol)

height of 50 m and a width of 82 m This microsystem is designated the assisted gradient generator or HAGG The fluorophore Alexa Fluor 488 and a fluorescent glucose analog (2-NBDG) are used as probes to illustrate the generation of stable, reproducible and linear, probe concentration gradients in the absence of cells A method is developed for estimating the diffusivity and hydrogel permeability of a solute

hydrogel-from in situ imaging data Concentration gradients are also generated in the presence of TC-6 to demonstrate the compatibility of the system for our study

-The three questions are concerned with the establishment of microscale glucose gradients over a population of -TC6 cells to determine the resultant mapping of (a) 2-NBDG accumulation, (b) insulin storage, and (c) insulin secretion The two specific assays test for

-cell response by probing for net GLUT-2/glucokinase activity that leads to intracellular 2-NBDG fluorescence, and intracellular vesicle density revealed by quinacrine fluorescence Our results show that (a) net GLUT-2/glucokinase activity is heterogeneous and does not map to extracellular glucose gradient, (b) quinacrine uptake

is homogeneous with no mapping of insulin storage to glucose gradient, and (c) insulin secretion is not influenced by the imposed glucose gradient

In the second part of this dissertation, a microfluidic competition assay valid for the case

of equilibrium binding between a receptor and competing ligands is developed A mathematical model describes the transient, convection-dispersion of solutes, undergoing equilibrium binding to immobilized receptors, while entrained in a low Reynolds number incompressible fluid flowing through a microchannel The proposed method involves monitoring the elution profile of a reference molecule and ligand in the presence of a competitor The time difference between the two breakthrough curves provides a measure of the unknown concentration of the competitor Theoretical results illustrate the general method for determining the equilibrium dissociation constant (K d) of the ligand and competitor, as well as the competitor concentration Experimental data is

presented for the binding of fluorescein-labeled insulin and unlabeled insulin to a monoclonal antibody It is found that the unlabeled insulin binds with higher affinity (K d

= 0.17 M) than the labeled insulin (K d = 0.76 M) The potential advantages of the method and further improvements in the model are discussed

the three questions 71Table 4 Concentrations of glucose used for the control and test samples in the HAGG

assays 76Table 5 Table showing the molecular weight of various probes and reagents 108

Table 6 Experimentally measured values of equilibriumm constants K d,L and K d,I and

other related data 146

LIST OF FIGURES

Figure 1 Diagram showing the Embden-Meyerhof-Parnas Pathway 5

Figure 2 Consensus model of glucose-induced insulin secretion 6

Figure 3 Class I microfluidic gradient generators (wide main channels, narrow cross-channels) 22

Figure 4 Modified Class I microfluidic gradient generator (Sub-class A) 23

Figure 5 Class II microfluidic gradient generators (narrow main channels, wide cross-channels) 23

Figure 6 Class III microfluidic gradient generators (cross-channel barriers) 24

Figure 7 Flow impingement in Class II generators 25

Figure 8 Diagram of the proposed microfluidic gradient generator 27

Figure 9 Class IV microfluidic gradient generators 28

Figure 10 Two-dimensional structure of the finalized microfluidic system, drawn using AutoCAD software 29

Figure 11 Words on the tube of glue are clearly visible behind a 12 mm thick block of PDMS 30

Figure 12 Phase contrast image of the microfluidic system showing the main microfluidic channels and cross-channels 32

Figure 13 Diagrams showing (a) hydrogel photomask with PDMS coating and indentation over the features, and (b) PDMS coating separated from the photomask 36

Figure 14 Microfluidic channels aligned to photomask as seen under a phase-contrast microscope during the process of photo-gelation 37

Figure 15 Phase contrast image of the microfluidic system showing the main microfluidic channels, cross-channels and hydrogel brackets 38

Figure 16 Oblique diagram of hydrogels (grey) in the cross-channels 38

Figure 17 Phase contrast image of prototype HAGG with continuous hydrogel barrier (Sub-Class B) 40

Figure 18 Acrylic adaptor lid (left) and cradle (right) 41

Figure 19 Assembled HAGG 42

Figure 20 (Left) Scaffold for buffer reservoirs (Right) Set-up for gradient generation 43

Figure 21 Diagram showing the setup of the HAGG for gradient generation without cells 43

Figure 22 Fluorescence images of 1 M Alexa Fluor 488 (left) and 10 M 2-NBDG (right) concentration gradients in the HAGG captured with ACT-1 software 45

Figure 23 Analysis of Alex Fluor 488 fluorescence intensity using ImagePro Plus software 45

Figure 24 Fluorescence intensity profiles plotted using Microsoft Excel (Blue, R2=0.9990) 1 M Alexa Fluor 488 and (Yellow, R2=0.9976) 10 M Alexa Fluor 488 as the source 46

Figure 25 Fluorescence intensity profiles plotted using Microsoft Excel (Pink, R2=0.9982 from distance=0.2 to 0.8) 1 M 2-NBDG and (Blue, R2=0.9933 from distance=0.2 to 0.8) 10 M 2-NBDG as the source 47

Figure 26 Portable cell culture system showing the two media reservoirs and acrylic cradle containing the HAGG 49

Figure 27 System disassembled to reveal the INSTECH P625/10K.143 peristaltic pump 50

Figure 28 Polystyrene slides cut from culture flasks 51

Figure 29 Phase contrast images of -TC-6 in (left to right) a T-75 culture flask, the HAGG system at t = 24 hrs and t = 48 hrs after system assembly, the cells at super-confluence 53

Figure 30 -TC-6 cells cultured in basal medium, devoid of glucose, for 24 hrs 54

Figure 31 -TC-6 cells in HAGG showing clear main channels 55

Figure 32 Portable cell culture system on microscope stage for sample observation 56

Figure 33 (a) Fluorescence and (b) bright-field images of 2-NBDG-loaded -TC-6 cells in a T-75 flask (c) Fluorescence and (d) bright-field images of -TC-6 cells in active solution phase gradient showing intracellular 2-NBDG fluorescence 57

Figure 34 Cummulative cross-channel fluorescence of 2-NBDG resulting from an extracellular 2-NBDG gradient 58

Figure 35 -TC-6 cells in 24-well plate loaded with NBDG showing heterogeneous 2-NBDG uptake and metabolic activity 63

Figure 36 -TC-6 cells in T-75 flask loaded with quinacrine 65

Figure 37 Release profile of quinacrine secreted from the -TC-6 cells in 24-well static culture 66

Figure 38 Phase contrast image showing area of cell death after quinacrine secretion 67

Figure 39 -TC-6 cells in T-25 flask loaded with Fura-2/AM showing increased intracellular fluorescence following a glucose challenge 69

Figure 40 Flow chart of assays to be performed in the HAGG 74

Figure 41 The HAGG configured for gradient conditioning 75

Figure 42 -TC-6 cell monolayer within a cross-channel 78

Figure 43 Fluctuations in mercury lamp intensity 80

Figure 44 (a) Brightfield, (c) fluorescence prior to 2-NBDG loading, (e) fluorescence after 2-NBDG loading, images of -TC-6 cells in control cross-channels (b) Brightfield, (d) fluorescence prior to NBDG loading, (f) fluorescence after 2-NBDG loading, images of -TC-6 cells in test cross-channels 82

Figure 45 Intracellular fluorescence intensity profiles for cells in HAGG 2-NBDG assays 86

Figure 46 Intracellular fluorescence intensity profiles for cells in HAGG quinacrine assays prior to glucose challenge 90

Figure 47 Intracellular fluorescence intensity profiles for cells in HAGG quinacrine assays after glucose challenge 94

Figure 48 Diagram of transverse section through a cross-channel showing the hydrogels as well as cultured cells 97

Figure 49 Steady-state fluorescence intensity profiles for 1 µM 2-NBDG (blue, R2=0.9982 from X = 0.2 to 0.8) and 10 µM 2-NBDG (light blue, R2=0.9933 from X = 0.2 to 0.8) in the source stream 104

Figure 50 Transient diffusion in cross-channel, experimental fluorescence data (Green and Blue dots), and model results (Green and Blue solid lines) for Sh 1 = Sh 2 = 2 105

Figure 51 Fluorescence image of (a) insulin-FITC, (b) BSA-FITC and (c) IgG-FITC in source stream, PBSA in sink stream 107

Figure 52 Loss of cell viability due to shear stress in a cross-channel with one hydrogel barrier structurally-compromised 110

Figure 53 Results of mathematical simulation of gradient generation in the presence of cells 112

Figure 54 Simulation of concentration gradient for hydrogels with Sh 1 =2 and Sh 2=0.33 126

Figure 55 Diagram showing competition of ligand (L) and competitor (I) for receptors (R) 131

Figure 56 (Top) Microchannel with inlet tube and eluent tube connected to fiber optic

cable probe (Bottom) Diagram of microchannel setup (indicated by white dashed box) 138Figure 57 Breakthrough curves obtained from control experiments showing three sets of

overlapping aprotinin-FITC and insulin-FITC profiles from three experiments 140Figure 58 Average concentration at the outlet (<L>z=1) normalized with the inlet

concentration (L0) as a function of dimensionless time, from experimental

measurement of K d,L 141Figure 59 Binding curves obtained from the numerical simulation showing the change of

T50% with L0/Kd,L for two different affinity constants 142Figure 60 Experimentally obtained breakthrough curves for determination of Kd,I 143Figure 61 Fluorescence images showing (a) bright field image of microchannel, (b)

fluorescence image of microchannel without captured insulin-FITC, and (c) fluorescence image of microchannel with captured insulin-FITC 143Figure 62 Competition assay binding curves of T50% (normalized with maximum T50%

in the absence of competitor) as a function of L0/Kd,L for different values of

I0/Kd,I. 145Figure 63 Effect of immobilization density on T50% 148Figure 64 Competition assay binding curves of T50% normalized with maximum T50%

in the absence of competitor as a function of L0/Kd,L for different values of

I0/Kd,I 150

ABBREVIATIONS

2-NBDG – 2-[N-(7-nitrobenz-2-oxa-1,3-diaxol-4-yl)amino]-2-deoxyglucose ADP – Adenosine diphosphate

ATP – Adenosine trisphosphate

BSA – Bovine serum albumin

GLUT – Glucose transporter

GLUT2 – Glucose transporter 2

HAGG – Hydrogel-assisted gradient generator

IgG – Immunoglobulin G

IPA – Isopropyl alcohol / isopropanol

MWCO – Molecular weight cut-off

NADH – Reduced nicotinamide adenine dinucleotide

PBS – Phosphate-buffered saline

PDMS – Poly(dimethyl) siloxane

PEG-DA – Poly(ethylene glycol) diacrylate

SNARE – Soluble N-ethylmaleimide-sensitive fusion protein-attachment protein

receptor

SV40 – Simian Virus 40

TPM – Trichlorosilyl propyl methacrylate

UV – Ultra-violet

P A R T I

1.2 The islets of Langerhans

The islets of Langerhans are discrete clusters of endocrine cells found dispersed

within the tissue of the pancreas Each islet comprises four cell types – -cells, cells, -cells and PP cells The organization of these cell types in the islet shows species-specificity, with rodent islets typically presenting the -cells forming a core cluster, surrounded by a mantle comprising the -, - and PP cells2 Primate islets, on the other hand, do not present a core-mantle structure, but have the cells in a random distribution1 Each islet is innervated by parasympathetic nerves and, in large islets, is supplied by blood vessels2 Islet vasculature is thought to arise from the invasion of the mantle and penetration of the core by an arteriole, which then branches radially into capillaries to exit the islet at the periphery3, 4 Thereafter, the capillaries meet to form a venous network, which ultimately drains into the hepatic portal veinibid This ensures that the islet secretions are carried

-directly to the liver, the major target of islet hormone regulation, via the shortest

vascular route available

1.3 Insulin secretion and glucose homeostasis

Each islet is an organ playing a crucial role in glucose homeostasis in the body 2 Glucose uptake from the circulation and subsequent metabolism are finely regulated by the orchestrated secretion of hormones by the different cell types comprising each islet The binding of glucose to the glucose transporter, GLUT2

on the surface of -cells triggers a signaling cascade that results in insulin secretion into the tissue fluid and, more importantly, the islet vasculature through which it reaches the target cells2-4 Binding of insulin to the insulin receptor in target cells then induces glucose uptake and metabolism The main target of insulin action is the hepatocyte, which actively takes up excess glucose from the circulation upon insulin stimulation, and converts it into the storage form, glycogen This effectively removes excess glucose from the circulation, so preventing hyperglycemia-induced cell trauma, and also ensures that a source of glucose exists in the absence of glucose intake in the form of food In addition to glucose homeostasis, insulin also regulates protein and lipid metabolism, tissue growth, development and survival, as well as glucose uptake and metabolism by skeletal muscle cells2, 5 Insulin action is also implicated in wound healing and fracture repair6, 7

The pancreatic -cells, on the other hand, secrete the hormone glucagon, whose actions on glucose homeostasis antagonize those of insulin Glucagon stimulates glycogen conversion to glucose and inhibits secretion of insulin from -cells2 As such, where insulin action tends to reduce systemic glucose concentration, glucagon action tends to increase it Maintenance of systemic glucose concentration is thus achieved by the concerted effect of these two hormones Pancreatic -cells produce somatostatin, which also serves to modulate the activities of the - and -cells However, the role of PP cells, which secrete the

function typically leads to loss of glucose homeostasis, resulting in disorders such

as diabetes mellitus8, 9 To date, much of the research in diabetes mellitus is centered upon understanding, augmenting or replacing -cell function

1.4 Glucose sensing and activation of -cell function

Pancreatic -cells play the crucial role of modulating systemic glucose concentration in order to maintain euglycemia10 This is one of the most important homeostatic processes since it ensures a constant supply of energy to the brain in the only form that it can utilize – glucose To be able to fulfill this task, pancreatic -cells must have a means to sense the systemic glucose concentration, as well as to respond appropriately to this information

Sensing ambient glucose must involve interaction of the glucose molecules with the pancreatic -cells When presented with a glucose challenge -cells would endocytose the glucose, so initiating a transduction process that ultimately triggers the fusion of insulin granule vesicles with the cell membrane

In pancreatic -cells, glucose is primarily endocytosed by the GLUT2 cell surface receptor The intracellular enzyme glucokinase then begins the enzymatic

processing of glucose via the Embden-Meyerhof-Parnas (EMP) pathway which

results in the production of pyruvate (Figure 1)11, 12 This process also produces adenosine trisphosphate (ATP) causing a shift in the adenosine diphosphate (ADP) to ATP ratio This can be represented in a simple reaction equation as follows:



Glucose + 2NAD +2ADP2P i 2NADH2Pyruvate2ATP2H2O2H

Hexokinase / Glucokinase

Phosphoglucose isomerase

ATP ADP

Phosphofructokinase-1

Aldolase

Triosephosphate isomerase

Glyceraldehyde phosphate dehydrogenase

3-2 x ADP

2 x ATP

Phosphoglycerate kinase

This increase in intracellular ATP then activates ATP-sensitive potassium ion channels, causing them to close The resultant membrane depolarization then activates voltage-gated L-type calcium channels which open, allowing Ca2+ ions entry to the cytoplasm This calcium ion influx then serves as the signal triggering exocytosis of insulin13, 14 (Figure 2)

Figure 2 Consensus model of glucose-induced insulin secretion

In this scheme, two molecules and their activities have been proposed as the putative glucose sensing mechanism in pancreatic -cells While some groups cite glucose uptake activity by the GLUT2 receptor as the possible candidate15, 16, others propose that it is in fact, the rate-limiting enzymatic activity of glucokinase,

Glucose-6-phosphate

Pyruvate Glycolysis Glucokinase

Oxidative phosphoryalation

the first of the EMP-pathway enzymes9, 17, 18 A means to resolve this dispute may

be to study which, of the two, is the rate-limiting activity in glucose-induced insulin release from pancreatic -cells

GLUT2 belong to a family of five structurally-related glucose transporters that facilitate diffusion of glucose across the mammalian cell membrane15, 16 This particular isoform is found in the -cells of the islets as well as the hepatocytes of the liver, and has been reported to be responsible for 80% of the glucose transport phenomena in -cells GLUT2 is described to have low affinity (K m  15 – 20 mM)19, but high capacity, for glucose, so ensuring that its activity is dose-dependent upon the glucose concentration, and yet not rate-limited by glucose concentrations in the physiological range In addition, this activity is itself modulated by glucose stimulation, which regulates the expression of the GLUT2

receptor Yasuda and co-workers showed how GLUT2 gene transcription is

up-regulated when rat islets have been cultured for 24 hrs in 11.1 mM glucose compared to those cultured in 5.5 mM glucose16 A similar trend was reported by

Inagaki and co-workers in HIT-T15 cells In addition, Inagaki had shown that this glucose-dependent increase in GLUT2 mRNA levels parallels the glucose-

dependent increase in the 2-DG – a non-metabolisable analog of glucose – uptake rate Furthermore, they had shown that cells cultured in media deprived of glucose for 24 hrs resulted in 2-DG uptake rates that were undetectable using the methods that they had employed Paradoxically, incubation of the cells in media supplemented with 5.5 mM glucose and then challenging them at supra-physiological concentrations (greater than 11.1 mM) also resulted in a drastic reduction in glucose uptake The reasons for this phenomenon have yet to be elucidated

The data seem to support the hypothesis that glucose uptake activity might be the

phosphorylation occurs at a higher rate than glucose transport in HIT-T15 cells,

so making glucose uptake the rate-limiting step, the same is not the case for islets, where the glucose transport rates are higher than those of phosphorylation and metabolism15, 17

Matschinsky and co-workers, on the other hand, published an informative review wherein glucokinase enzymatic characteristics were presented as evidence to support the hypothesis that this molecule serves as the glucose sensor in pancreatic -cells17 Glucokinase is one among four hexokinases related homologously and probably evolutionarily This enzyme is the first, and the rate-limiting one, that glucose encounters in the EMP pathway Glucose is converted

to glucose-6-phosphate by glucokinase

Not only does glucokinase have the lowest affinity for glucose (K m  10 mM) among the glycolytic enzymes, it also shows cooperative kinetics with its substrate, glucose9 This cooperativity has a Hill coefficient of about 1.7 This latter character appears to center about the set point of 5 mM glucose, which would make glucokinase activity highly sensitive to deviations in glucose concentration from normoglycaemiaibid Glucokinase activity has been shown to increase up to five-fold, with increasing concentrations of ambient glucose from 3

mM to 30 mM in islet tissue extractsibid Changes in glucokinase activity have also been shown to affect the glucose stimulatory threshold of -cellsibid Finally, polymorphisms in the glucokinase gene, has been linked to maturity-onset diabetes of the young (MODY), further demonstrating how changes in glucokinase activity might affect glucose sensing and pancreatic -cell response to

a glucose challengeibid The fact that this enzyme activity is functionally tuned to perceive perturbations to homeostatic glucose concentrations, as well as the fact that the loss of its activity appears to result in failure to sense changes in

finely-systemic glucose concentration support the proposition that glucokinase is the glucose sensor in pancreatic -cells15, 17

1.5 Effect of glucose on first phase insulin secretion

Regardless of whether GLUT2 or glucokinase activity serves as the sensing mechanism, endocytosis of glucose ultimately induces the release of insulin from granules in pancreatic -cells

Barg and co-workers describe the granules of secretory cells as comprising the reserve pool, the readily releasable pool (RRP) and the immediately releasable pool (IRP)20 Each group is distinguished by the maturity of the insulin granules

as well as by how quickly the granules can be released upon stimulation by a secretagogueibid Unlike the reserve pool which has yet to undergo several reactions to become releasable, the RRP and IRP consist of granules that are

already physically docked with the plasma membrane via trans-SNARE (soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor)

complexesibid As such, other related molecules such as SNAP-25, syntaxin and the calcium sensor synaptotagmin are thought to be involved with the calcium ion-triggered insulin secretion process in pancreatic -cellsibid Furthermore, Barg presents evidence suggesting that a large fraction of the RRP granules are also tightly linked to the Ca2+ channels in pancreatic -cells, and that their activation is responsible for the first-phase of insulin releaseibid

1.6 Effect of glucose on second phase insulin secretion

Upon exocytosis of the existing pool of insulin granules, further release of insulin

is limited by the glucose-dependent biosynthesis of this hormone Metabolism of glucose has been shown to induce a preferential increase in insulin biosynthesis over other proteins14, 21, 22 Hence glucose must have a role in regulating insulin

The insulin gene promoter contains three transcription factor binding sites designated A3, E1 and C1, which are glucose-responsive22 When stimulated, these are bound by the transcription factors, Pdx-1 (pancreatic and duodenal homeobox-1), NeuroD1 (neurogenic differentiation 1), MafA (V-maf musculoaponeurotic fibrosarcoma oncogene homologue A) respectivelyibid

Studies suggest that these factors are activated by glucose via post-translational modifications such as O-linked glycosylation and phosphorylation, as well as their

translocation to the nucleusibid These three transcription factors have been identified as key components in the regulation of glucose-stimulated insulin biosynthesisibid They have differing potencies, but act in a coordinated manner to synergistically promote insulin gene expression Studies suggest that they act in combinations of the following ways: (i) by binding to other transcription factors, such as RNA polymerase II, and promoting their binding to the insulin gene promoter; (ii) by recruiting various histone modifying proteins, such as the histone acetyltransferase p300, to the promoter; (iii) by upregulating the transcription and translation of other transcription factors or proteins involved in the glucose-stimulated insulin secretion machinery; or (iv) by repressing factors that inhibit insulin biosynthesis and recruiting repressors of these inhibitors Co-regulation has also been proposed between Pdx-1, NeuroD1 and MafAibid

1.7 -cell heterogeneity

Despite the plethora of studies investigating glucose-mediated -cell activity, there remain phenomena in -cell function that are poorly understood An example is the physiological and functional heterogeneity that exists among, and even within, individual pancreatic -cells and within islets23-27 This heterogeneity

is ultimately reflected in the difference in the ability of -cells to sense glucose and to secrete insulin under this stimulationibid It can also manifest itself as inter-cell differences in morphology, such as differences in intracellular insulin density

glucose stimulation, rates of insulin biosynthesis, rates of metabolic redox shifts, etc17, 28

Studies have demonstrated morphological heterogeneity in the form of glucokinase distribution in the various classes of -cell organization21, 23, 29 It has been found that glucokinase tended to be uniformly distributed in individual -cells, but showed heterogeneous distribution in rat islets Furthermore, in -cells that are in contact with a blood vessel, there appears to be concentration of glucokinase in the cytoplasm closest to the vasculature29

Our specific interest, however, is in -cell functional heterogeneity characterized

by the difference in glucose stimulatory thresholds This heterogeneity results in only a sub-population of -cells in any given islet releasing insulin upon stimulation by a specific glucose concentration It has been shown in rat islets that a certain hierarchy in sensitivity prevails, where -cells of a higher sensitivity are apparently found in the medullary (core) regions, near the origin of islet vascularisation26, 27, 29 Consistent with this distribution of glucose sensitivity, upon glucose stimulation, -cells from the medullary region of rat islets respond first, followed by those from the cortical (peripheral) regions as the glucose concentration increases29

Nonetheless, secretion of insulin occurs in synchronization, so suggesting that communication and coordination occur between the activated -cells Indeed, coordinated -cell function has been shown to be dependent on -cell-to--cell

contact, and communication via gap junctions24, 30-36 It has been reported that loss

of -cell-to--cell contact is correlated to a reduction in responsiveness to glucose stimulation, and that rodents deficient in the gap junction protein connexin 36



Most interestingly, -cell heterogeneity has been shown to be plastic25, as is evident from the loss of insulin and glucokinase immunostaining heterogeneity from the medullary to the cortical regions of large rat islets under conditions of starvation29 Reducing the extracellular glucose concentration appeared to increase the intracellular concentrations of insulin, an indicator of a decrease in glucose metabolism and insulin secretion More interesting is the observation that these changes tended to occur from the medullary to the cortical regions of large rat islets These changes were reversed in order when the rats were re-fedibid Small islets and isolated -cells, on the other hand, tend to be homogenous for insulin immunostainingibid

1.8 The hypothesis

These observations suggest directionality in -cell functional and morphological heterogeneity Considering the fact that the large islets possess a more mature vascular network, and that the vasculature is thought to be organized such that the arteriole supplying a large islet penetrates all the way to the core, from which

it then brachiates toward the periphery forming the islet capillaries3, 4, 37, it is possible that the islet vasculature might be a means by which this directionality is established and maintained The fact that glucose actively modulates pancreatic -cell response, and that the blood glucose concentration in islet vasculature would decrease from the core to the periphery, further suggest that it is this glucose concentration gradient that is responsible for the phenomenon This hypothesis is corroborated by the fact that small islets and isolated -cells, which have little or immature vasculature of their own, tend to show less -cell heterogeneity29, 37 The fact that -cell-to--cell interaction is crucial for the resolution of this functional variation33 suggests that gap junction communications, in the context

of the glucose concentration gradient, may mediate the spatial patterning of -cell functional heterogeneity

In view of these previous experimental findings, the following hypothesis was proposed: that the establishment of spatial gradients of extracellular glucose and promotion of -cell interactions are adequate for development

of glucose concentration-mapped functional heterogeneity in -cell populations

1.9 Three questions raised to address the hypothesis

The exhaustive verification of this hypothesis is beyond the scope of this thesis However, specific questions that may contribute to its validation were addressed Broadly, I propose to evaluate the functional re-distribution of a heterogeneous population of -TC-6 cells, in response to an imposed extracellular glucose concentration gradient -TC-6 is a murine SV40 T-antigen-transformed pancreatic -cell line, selected because it retains major functional characteristics

of pancreatic -cells, in addition to being easier to maintain in culture than the primary cells38 This gradient must be generated over a domain of 300 to 500 m

to simulate the in vivo length scales in which these cells would normally exist

These cells would then be exposed to a glucose challenge to determine if the spatial pattern of -TC-6 function in response to the challenge has changed The following three questions will be addressed specifically:

i) Does an imposed extracellular glucose gradient result in mapping of net GLUT2/glucokinase activity in a heterogeneous population of -TC-6 cells?

ii) Is intracellular insulin granule density correlated with an imposed extracellular

glucose gradient?

iii) Are changes in intracellular insulin granule density upon a glucose challenge

Two assays were selected for probing net GLUT2/glucokinase activity and insulin granule density The uptake and metabolism of the glucose analog 2-NBDG has been used extensively in glucose uptake and metabolism studies17, 18, 39-

41 Endocytosis of this fluorophore is facilitated by transport via cell surface

GLUT receptors It competes with glucose for the same transport receptors and has thus been used as a surrogate for measuring glucose uptake rates Once endlol, neceocytosed, it is metabolised by glucokinase, as well as other enzymes of the Embden-Meyerhof-Parnas pathway to yield non-fluorescent products As such, the resultant intracellular 2-NBDG fluorescence in a loaded cell is due to, and is a measure of, net GLUT/glucokinase activity -TC-6 is able to endocytose and metabolilse 2-NBDG (See section 3.5.1 -TC-6 metablism of 2-NBDG.) This cell line expresses primarily GLUT2 receptors38 and the resultant intracellular 2-NBDG fluorescence is heterogeneous in a random population of the cells, reflecting the heterogeneous net GLUT2/glucokinase activity This demonstrates the suitability of the 2-NBDG uptake and metabolism assay for probing net GLUT2/glucokinase activity in -TC-6

The other assay selected is the quinacrine secretion assay This involves loading the seceretory vesicles of -TC-6 with the fluorophore quinacrine, which accumulates preferentially in acidified intracellular compartments14 Quinacrine has been shown to co-locate, and to be secreted proportionally, with the hormone insulin in primary pancreatic -cells14 -TC-6 cells are also able to be stain with quinacrine and upon exposure to glucose, loses intracellular quinacrine fluorescence in a dose-dependent manner paralleling its insulin release profile (See section 3.5.2 Glucose-induced quinacrine secretion in -TC-6.) As such, staining the cells with quinacrine would allow us to probe the intracellular insulin granule densities of -TC-6, while the secretion assay would allow us to track glucose-induced loss of insulin granule density

These two assays would have to be performed in a specialized test platform for addressing our three questions The development, fabrication and application of this testing platform are the main aims of this body of work

1.10 Microfluidic channel cell culture

To date, studies that address the nutrient-dependent nature of -cell functional heterogeneity, as well as the role of cell-cell interactions in this context, tend to employ discrete, isotropic, changes in extracellular glucose concentration, rather than a continuous concentration gradient Since, conventional methods of cell culture cannot mimic the glucose concentration gradients over the distances typical of an islet, it is necessary to employ an alternative system of cell culture, such as microfluidic channels

The use of microfluidic channels allows us to mimic characteristics of the in situ

microenvironment of a cell more closely42-44 These characteristics include specific cell densities and specific interactions between differing populations of cells Microfluidic channels are, in essence, cell culture chambers with at least two

tissue-of their dimensions in the order tissue-of microns, which can be perfused with growth medium at a controlled flow rate The relatively small volume of the microfluidic channels allows more precise control over maintenance of osmolarity, pH and temperature in the perfusing medium, without the use of complicated apparatusesibid This small volume also allows the use of much fewer cells for culture studies, making it suitable for situations where the cells of interest are scarce or limited This, in turn, would incur a much lighter burden on resources dedicated to cell culture supplies and expensive reagents

The use of microfluidic channels for cell culture allows nutrition by a combination of perfusion and diffusion, a condition that closely mimics that in vascularised tissue This is particularly pertinent to this project, since it involves

the presence of a perfusion system, a requirement fulfilled by blood vessels Futhermore, microfluidic channels can be designed to have dimensions that will allow us to mimic the length scales typical of tissue microenvironments This

would allow us to set up cellular microenvironments that would mimic that in vivo, thus making our model more pertinent to in vivo phenomena Finally, various

microfluidic phenomena can be exploited to manipulate mass transport in the microfluidic channels, so making it possible to generate continuous solute concentration gradients

1.11 Generating solute concentration gradients in microfluidic channels

It is necessary that a continuous glucose concentration gradient be generated over

a monolayer of pancreatic -cells, or a transformed derivative of such, in order to study changes in the glucose-dependent response Early systems for generating solute concentration gradients employed collagen, fibrin or agarose hydrogels as barriers to diffusion of signal molecules over the test domains45-48 Typically, the signal molecule solution is dispensed into wells made in the hydrogel matrices These then diffuse out of the wells and into the surrounding material Others like the Micropipette, Boyden Chamber/Transwell, as well as Zigmond and Dunn Chamber assays rely upon the molecules diffusing in solution from a source region to one of lower concentration49 These have limited success in establishing solute concentration gradients since the gradients generated are not stable in either time or spaceibid Furthermore, the concentration gradients can only be established over distances far in excess of the length-scales commonly encountered in biological systems, so compromising their relevance to the study

of natural biological processes

Microfluidic systems designed to create solute concentration gradients can address the need for establishing gradients along the length scales of biological environments A search of the scientific literature reveals several methods of

creating solute concentration gradients using phenomena that are sometimes peculiar to microfluidics systems Some systems have been used to generate discrete, or stepwise, changes in solute concentration primarily through the use of controlled distribution of the solutes50, while others employ diffusive mixing in addition to controlled distribution51-53 to achieve continuous gradients of solute concentration Of the latter, the diffusion of solutes across laminar flows of differing solute concentration, in so-called T-sensor systems, is used to establish the gradients52, 53 Microinjections54 and electrokinetic processes55, 56, wherein an applied voltage is employed to effect the solute distribution, have also been used

to attain localized, microscale continuous changes in solute concentration

Microfluidic systems based on the early hydrogel modalities have a hydrogel positioned above or in these microfluidic channels serving as a diffusion barrier

A reservoir then serves as a solute source while a separate, but similar reservoir serves as the sink Diffusion of the solute from the reservoir then occurs across the hydrogel, and hence the microfluidic channel, towards the sink This establishes a transient concentration gradient across the necessary length scales without the need for perfusion through the microfluidic channels57 However, such systems suffer from the transience of the concentration gradient, which quickly degrades as an equilibrium concentration of the solute is established between the source and the sink49, 57 As such, their use is restricted to the short period corresponding to this transient stability, or requires constant re-establishment of the source and sink conditions49, 58

An alternative means of maintaining the source and sink conditions is to provide

a constant perfusion of source and sink material through the microfluidic systemibid The previously-mentioned T-sensor type systems exploit the fact that two contacting laminar streams flowing through a microfluidic channel does not

one stream across to the other49 A concentration gradient of the solute is thus established from the source stream, transversely through the microfluidic channel, to the sink stream Concentration gradients so-established are temporally stable, so long as the source and sink concentrations of the solute are maintainedibid However, they are not spatially constant since the gradient will change continuously along the axis of perfusionibid Use of these systems is therefore confined to the regions wherein the appropriate concentration gradient existsibid Furthermore, the necessity of maintaining perfusion means that any cells

in the microfluidic channels will be exposed to shear stress, whose effects on cell behaviour might prove confounding49, 59 In addition, this constant perfusion would also deplete any important autocrine or paracrine signals from the vicinity

of the cells, which may again result in a confounding element being introduced into the experimental analysisibid

Some researchers have combined the previous two strategies, employing perfusion to maintain the source and sink conditions, but separating the source and sink spatially, so that diffusive flux of the solute occurs through a static fluid

or hydrogel domain49, 58, 60 Mosadegh and co-workers, designed a microfluidic

system comprising two main channels in which the source and sink streams flow, connected by cross-channels of varying dimensions, filled with agarose hydrogel Using this system they were able to generate solute concentration gradients with linear as well as non-linear gradients Kamm and co-workers employed a similar strategy to develop a microfluidic model of vascular endothelial growth factor-induced angiogenesis61, 62

Li and co-workers, however, exploited the fluidic resistance generated in cross

channels with a 25-fold smaller cross-sectional area compared to the main channels, to restrict the perfusion to the main channels This allowed diffusive

flux to establish a solute concentration gradient from the source stream, through the intervening static fluid domain in the cross-channels to the sink stream

Clearly, microfluidic systems offer a wide range of solutions for our purpose It is only necessary to select and adopt an appropriate approach Since our system involves the use of glucose, an uncharged molecule, the choice of appropriate techniques is limited to those that employ controlled solute distribution and diffusive mixing Unfortunately, some of the methods discussed here require laminar flows of the solute solution over the region where the concentration gradient is to be established Since the concentration gradient must be generated over the monolayer of cells in our system, this would mean that the cells would

be exposed to fluid flow and hence, considerable shear stress Those methods that generate the gradients within hydrogels complicate presentation of the solute

to the cells, making mathematical modeling of the transport phenomena difficult Those that rely on complex microfluidic manipulation to prevent cross-flow in the cross-channels are difficult to emulate, as will be demonstrated in the next chapter An alternative means of generating the concentration gradient in a microfluidic system, based on controlled distribution and diffusive mixing, was therefore required

1.12 Organisation of the thesis

The following chapters will detail the development of our microfluidic test system as well as how it was used to address the three questions raised Chapter 2 highlights the evolution of our microfluidic gradient generator, beginning with a short presentation of the various prototypes that had been designed and tested, as well as the elements which led to the final design Details are given of the various materials and protocols used for fabricating our microfluidic gradient generator Chapter 3 describes the various assays used to first characterize the microfluidic

the choice of materials as well as assays is also presented Chapter 4 provides analyses of the results obtained from the various assays, showing how they contribute to validating or refuting our hypothesis Some concerns over the experimental protocols used are also mentioned for later improvement Chapter 5 concludes Part I by suggesting additional tests that may be carried out using our system, as well as casting this work in the larger context of studying cell behaviour regulated by signal concentration gradients Part II commences with Chapter 6 which highlights the work done in developing a microfluidic competition assay as an alternative method of studying equilibrium binding events A method for determining the various binding constants, as well as for determining analyte concentration using these data is described

2.1 The proposed design

I proposed to design and fabricate a cross-channel type microfluidic device wherein the perfusion of the glucose-containing buffers is limited to parallel flows bracketing the region of cell culture (henceforth referred to as the cell culture chamber) Diffusion of glucose from one bracket to the other, as well as its consumption, across the culture chamber would establish the concentration gradient If successful, this technique can, in fact, be extended to any other analyte or multiple analytes in the perfusing medium with appropriate modifications to the flow parameters Since the cells would ideally be exposed to minimal fluid movement during the establishment of the concentration gradient, the gradient can be set up in the presence of a population of contacting cells This can be achieved by limiting the perfusion of the glucose-containing buffers

to parallel laminar flows bracketing the cell culture chamber Complete medium perfused through the laminar streams instead, would serve as a nutrient source

when culturing the cells in the system The cells would thus be confined within a

static fluid domain throughout

2.2 Microfluidic gradient generators

Four strategies were adopted to prevent or reduce cross-flow of perfusate from

the main channels into the cross-channels These gave rise to four classes of

microfluidic gradient generators The first approach employed microfluidic

systems with main channels of width 17, 10 and 4 times greater than that of the

A subclass of generators were those with initially narrow cross-channels leading

to a wider middle chamber (Figure 4)

Figure 4 Modified Class I microfluidic gradient generator (Sub-class A)

These designs aimed to induce preferential flow of fluid in the main channels by

increasing the pressure drop, and hence resistance to flow, within the

cross-channels

The Class II microfluidic gradient generators were designed in obverse, wherein

the cross-channels were of width 6 and 10 times that of the main channels

These generators were designed to exploit the decreased fluidic pressure exerted

by a laminar stream with higher velocity upon another of lower velocity The

channels overcome what little pressure is exerted by the stream in the main

channels, so preventing their impingement

The third class of generators was designed to increase resistance to flow into the

cross-channels by the placement of physical barriers at the entrance of each

However these are not absolute barriers, but were perforated at regular intervals

to allow mass transport between the main and cross-channels Varying the

perforation size, frequency and barrier density allowed us to control the amount

of resistance presented to impinging flow from the main channels (Figure 6)

D 15

D 16

D 17

D 18

D 19

D 20

D 21

D 22

D 23

D 24

D 25

D 26

D 27

D 28

D 29

G 11

G 12

G 13

G 14

G 15

G 16

G 17

G 18

G 19

G 20

G 21

G 22

G 23

G 24

G 25

G 26

G 27

H 07

H 08

H 09

H 10

H 11

H 12

H 13

H 14

H 15

H 16

H 17

Figure 6 Class III microfluidic gradient generators (cross-channel barriers) Arrows indicate cross- channel barriers

These microfluidic gradient generators were each tested for their ability to

establish linear fluorescent probe concentration gradients as described in the