Citation
Regulation of coupled beta-cell electrical dynamics

Material Information

Title:
Regulation of coupled beta-cell electrical dynamics
Creator:
Westacott, Matthew Joel ( author )
Place of Publication:
Denver, Colo.
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file : ;

Subjects

Subjects / Keywords:
Islands of Langerhans ( lcsh )
Diabetes ( lcsh )
Pancreatic beta cells ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis Ph.D.--University of Colorado Denver, 2017.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: Adobe Reader.
Statement of Responsibility:
by Matthew Joel Westcottt.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
on10596 ( NOTIS )
1059653964 ( OCLC )
on1059653964

Downloads

This item has the following downloads:


Full Text
REGULATION OF COUPLED /3-CELL ELECTRICAL DYNAMICS
by
MATTHEW JOEL WESTACOTT B.S., Colorado School of Mines, 2010 M.S., University of Denver, 2012
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Bioengineering Program
2017


This thesis for the Doctor of Philosophy degree by Matthew Joel Westacott has been approved for the Bioengineering Program by
Emily Gibson, Chair Richard Benninger, Advisor Diego Restrepo Tim Lei
Date:July 29, 2017


Westacott, Matthew Joel (Ph.D., Bioengineering)
Regulation of Coupled /3-Cell Electrical Dynamics Thesis directed by Assistant Professor Richard Benninger
ABSTRACT
Pancreatic /3-cells are the only cell type responsible for secreting insulin; a hormone necessary for maintaining glucose homeostasis. /3-cells are a naturally heterogeneous in expression of genes regulating glucose metabolism and electrophysiology; leading to a population of cells with unique electrical dynamics. Electrical communication between /3-cells through connexin36 gap junctions act to synchronize electrical activity and coordinate insulin secretion dynamics to promote efficacy of glucose clearing. In models of diabetes there is disruption to the coordinated electrical dynamics within pancreatic islets. However, what role functional subpopulations play in mediating this disruption or the extent to which this disruption occurs in human islets is poorly understood.
The objectives of this thesis is to determine (1) develop a computational multicellular model of pancreatic islet electrophysiology and test if it can recapitulate two aspects of Cx36 mediated electrical dysfunction (2) determine what role subpopulations play in controlling electrical activity and coordination within pancreatic islets (3) test if age and type2 diabetes in humans correlate with decreases to electrical coordination within islets. To recapitulate in-vivo models of pancreatic islet dysfunction we expanded on an established single-cellular computational model of /3-cell electro-


physiology by generating a coupled /3-cell network with islet architecture. I found the model was able to correctly predict critical loss in electrical activity and coordination; mediated by the electrical coupling between /3-cells. To determine the role of functional subpopulations I created a transgenic mouse model with /3-cell specific expression of Channelrhodopsin-2 and implemented a protocol of spatiotemporal [Ca2+] activation combined with two-photon imaging of metabolic activity which showed subpopulations of cells were spatially orientated and showed preferential control over electrical activity and dynamics within islets. Lastly we tested the hypothesis that advanced age and type2 diabetes correlate with decreases to electrical coordination within human pancreatic islets through imaging of [Ca2+] activity and developing a novel image analysis algorithm to quantify electrical coordination. I found that both advanced age and history of type2 diabetes correlated with significant decreases to electrical activity and coordination within human islets. Furthermore, we found that a Cx36 activator could restore electrical coordination within aged and type2 diabetic human islets.
The results presented herein provide insight into regulating factors of coordinated electrical dynamics between /3-cells and show potential for a novel therapeutic to recover pancreatic /3-cell function.
The form and content of this abstract are approved. I recommend its publication.
Approved: Richard Benninger
IV


ACKNOWLEDGMENTS
This work was supported by NIH Grants R01 DK102950, R01 DK106412; and Juvenile Diabetes Research Foundation Grant 5-CDA-2014-198-A-N (to R.K.P.B.), and F31 DK107043 (to M.J.W.). Microscopy was performed through the use of the University of Colorado Anschutz Medical Campus Advanced Light Microscopy Core (P30 NS048154, UL1 TR001082), islet isolation were performed in the Barbara Davis Center Islet Core (P30 DK057516) under IACUC protocal (B-95814(07) 1D), and simulations were performed on the JANUS supercomputer at the University of Colorado through support by the NSF (CNS-08217944). Islet perifusion experiments were performed through the Islet Procurement and Analysis core of the Vanderbilt Diabetes Research and Training Center (DK20593) in the Alvin Powers Lab (Vanderbilt). We thank Matthew Merrins (University of Wisconsin) and Laura Pyle (University of Colorado) for discussion and statistical advice.
v


TABLE OF CONTENTS
CHAPTER
I. BACKGROUND AND INTRODUCTION.......................................... 1
Glucose Control and Diabetes............................................ 1
Pancreatic Islets....................................................... 3
Electrophysiology and Stimulus Secretion Coupling of /3-Cells........... 4
/3-Cells as electrochemical oscillators ............................... 11
Electrical coupling between /3-cells................................... 12
Heterogeneity of /3-Cells and Relevance Diabetes....................... 15
II. MATERIALS AND METHODS............................................... 18
Mouse Models........................................................... 18
Islet Culturing........................................................ 18
Microscopy and Calcium Imaging......................................... 18
Psuedoislet Aggregation................................................ 20
[Ca2+] Coordination ................................................... 20
ChR2 Activation Analysis............................................... 21
FRAP................................................................... 22
Immunofluoresence ..................................................... 22
Insulin ELISA.......................................................... 23
Classification Learners................................................ 25
III. DEVELOPMENT OF COMPUTATIONAL MODEL ...................................26
VI


Introduction................................................................ 26
Model Description .......................................................... 30
Model Verification.......................................................... 33
Conclusions ................................................................ 45
IV. SPATIALLY ORGANIZED SUB-POPULATIONS OF CELLS CONTROL ELECTRICAL ACTIVITY AND DYNAMICS ACROSS THE ISLET OF LANGER-
HANS........................................................................ 49
Introduction................................................................ 49
Results..................................................................... 51
Local ChR2 activation reveals regions of varying excitability ..........51
Beta-cell metabolic activity controls variations in ChR2 stimulated [Ca2+]i: 58 Multicellular model links metabolic activity and ChR2 stimulated [Ca2+] 63
Spatial organization to NAD(P)H and ChR2 responses..................... 67
Wave origin correlates with lower metabolic activity and excitability ... 70 Intrinsic oscillatory frequency controls calcium wave propagation .... 72
Cytokine Mediated Disruption of Electrical Activity ................... 75
Conclusions ................................................................ 79
Metabolically active sub-populations of cells control excitability......79
Spatial organization of cell heterogeneity..............................81
Lower metabolic, high frequency cells initiate propagating calcium waves 83 Cytokine Mediated Disruption of Electrical Activity ....................85
vii


V. AGE ASSOCIATED DECLINE IN ELECTRICAL SYNCHRONIZATION
87
Introduction.......................................................... 87
Results............................................................... 89
Quantification of Electrical Activity and Coordination............89
Age Predicts Decline in [Ca2+] Coordination in Human Islets.......92
Age-dependent decline in insulin secretion........................95
Cx36 gap junction activation recovers age-dependent [Ca2+] decline 101 [Ca2+] and Cx36 gap junction function is preserved in aged mice .103 Conclusions .......................................................105
VI. CA2+METRICS IDENTIFY TYPE2 DIABETIC HUMAN DONORS......................110
Introduction..........................................................110
Results................................................................111
Generating Classifiers to Predict Type2 Diabetes..................111
Modafinil Increases Electrical Coordination in Human Type2 Islets ... 116 Conclusions .......................................................118
VII. DISCUSSION AND FUTURE DIRECTIONS.....................................119
Functional Role of Subopulations.......................................119
Loss of Ca2+Coordination with Advanced Age and Type2 Diabetes..........121
Concluding Remarks.....................................................124
BIBLIOGRAPHY..............................................................126
viii


LIST OF TABLES
TABLE
1. List of Parameters used in Cha-Noma Model..............................31
2. Parameters used to generate classification model of diabetes incidence 111
3. Summary of performance for varying classification learners.............113
IX


LIST OF FIGURES
FIGURE
1. Circuit Diagram of Excitable Cell ...................................... 6
2. Metabolism of glucose generates ATP and NADH........................ 8
3. Glucose stimulated insulin secretion .................................. 10
4. Calcium and Insulin Show In-Phase Oscillations...................... 12
5. Cx36 Couples/3-cells................................................ 14
6. Hodgkin-Huxley Model .................................................. 27
7. Diagram of Cha-Noma Model........................................... 29
8. Glucose Activated Ca2+ Oscillations in simulated islet..............33
9. Penetrance of Kir6.2[AJV30)jK:i85Ql-GFP.............................34
10. Phase transition in excitability.................................... 36
11. Link between phase transitions in Ca2+ and physiological parameters . 37
12. Coupled oscillator model describes experimental phase transitions ... 39
13. Mean-field theory anaolog of /3-cell network activity...............40
14. Model of synchronization disruption.................................42
15. Loss of electrical synchronization with microparticle incorporation ... 43
16. Loss of electrical synchronization in-silico........................44
17. ChR2 activation creates local Ca2+elevation in pancreatic islets .... 53
18. ChR2 activation is consistent through time..........................54
19. Quantification of ChR2-YFP Expression in Disassociated Islets.......56
20. Electrical excitability shows no temporal/spatial biasing ..............57
x


21. Activation of non ChR2 expressing islets.................................58
22. Metabolic Dependence on ChR2 Activation................................. 60
23. Spatial variations in metabolic activity control electrical activity.....62
24. Spatial domains in an islet model recapitulates ex-vivo excitability ... 65
25. Resting Membrane Potential Links Cellular Excitability to ChR2 Activation 67
26. Spatial analysis of NAD(P)H and ChR2-activated Ca2+ responses ... 69
27. Correlation between wave origin and metabolic activity.................. 71
28. Oscillatory Frequency................................................... 73
29. Acute Treatment of Pro-inflammatory on ChR2 Function.................... 76
30. Cytokine Treatment Disrupts Excitability................................ 77
31. Cytokine Treatment Disrupts Exctiability at High Glucose................ 78
32. Human Islet Electrical Activity......................................... 91
33. Age Predicts Decline in Human Coordination...............................93
34. Further characterization of Ca2+ dynamics in human islets with age 94
35. Age-dependent decline in Cx36 gap junction coupling in human islets 96
36. Age-dependent decline in insulin secretion dynamics......................98
37. Age Dependent Decline to Insulin Secretion ..............................99
38. [Ca2+ Activity Predicts Decline in Insulin Secretion....................100
39. Cx36 Activators Rescue Ca2+Function.....................................102
40. Preservation of Cx36 gap junction function and Ca2+ coordination in
aged mice...............................................................104
41. BMI binary clasisfication of Diabetes...................................112
XI


42. Linear Classification Performance........................................114
43. Single predictor performance.............................................114
44. Loss of electrical coordination in T2 islets.............................116
45. Modafinil Treatment of Human Type2 Islets................................117
xii


CHAPTER I
BACKGROUND AND INTRODUCTION Glucose Control and Diabetes
Glucose control is one system which is tightly regulated by a negative feedback loop to provide the necessary energy for all other functions within the body. Complex sugars in carbohydrates are broken down into the simple sugar glucose which is transported into the circulatory system. As circulating glucose levels rise a feedback signal is needed to bring glucose levels back down by promoting peripheral tissue to increase glucose uptake and in turn use it for energy. Pancreatic /3-cells are the only cell type which produce and secrete insulin; a hormone necessary for maintaining glucose homeostasis by inducing uptake of circulating blood glucose to peripheral tissue, by upregulatating glyconeogenisis, and by suppressing gluco-neogensis (Bano, 2013) and in healthy individuals, glucose levels are kept within a narrow range between 60mg/dl to 150mg/dl. Insulin acts on peripheral tissue through acting on the insulin receptor triggering a cascade via the tyrosine kinase pathway to increase transportation of glucose transporters (GLUT) to the plasma membrane; increasing the net flux of glucose from the circulatory system to surrounding tissue. Glucose is transported into cells where it enters the metabolic pathway to create ATP to be used as the primary energy source in muscular and adipose tissue or be stored and glycogen in the liver.
1


Diabetes is characterized by a state of chronic hyperglycemia, excessive high blood glucose levels, due to a failure of /3-cells to secrete the appropriate levels of insulin either due to autoimmune disease and destruction of /3-cells (type 1) or by insulin resistance and /3-cell dysfunction (type 2). This can be characterized when fasting glucose levels are above 126mg/dl (7mM) when non-diabetic levels are below 108mg/dl(6mM) or if a 2 hour post-pranadial is above 200mg/dl (11mM) when nondiabetic levels are below 140mg/dl (7.8mM). The loss of proper glucose control leads to numerous complications through damages to nervous and circulatory systems as hyperglycemia in prolonged duration is toxic to cells (Kawahito, 2009). If left unchecked this damage will lead to cardiovascular disease, strokes, neuropathy, and retinopathy. Life expectancy of typel diabetics is estimated to be 20 years younger than healthy individuals and diabetes is rated as the 7th most commons cause of death in the United States. While rates for most diseases are decreasing with time, diabetes typel (Davis, 2001) and type2 are increasing at an alarming rate showing 61% incidence between 1990 and 2001 (Nathan et al., 2007). Approximately 85% of diabetics are type2 and nearly 30 million Americans and 410 million people worldwide have some form of diabetes a rising from 108 million in 1980 with the majority of new cases coming from Asia and Africa (Organization et al., 2016). Typically in type 1 diabetes there is a multi-year progression from an initial event to trigger autoimmunity until loss of sufficient/3-cell mass leads to hyperglycemia. After diagnosis there is usually little pharmaceutical intervention available and insulin injection will be required for life. In type 2 diabetes there are remaining functional /3-cells which are
2


unable to meet required insulin demands caused by a combination of insulin resistance and /3 cell dysfunction. Pharmaceutical intervention is a preferred treatment with type2 diabetes along with diet and exercise changes. The most common pharmacological agents will either target glucose handling within the liver and peripheral tissue or directly target pancreatic /3-cells to potentiate insulin secretion by targeting specific channels or receptors. There is growing investigation into pharmacological agents which target alternative pathways within /3-cells including targeting glucose transporters in the kidneys and insulin sensitizes (Wagman and Nuss, 2001). However, additional theraputics are needed to provide higher efficacy of glucose control. Pancreatic Islets
/3-Cells are not randomly distributed throughout the pancreas, rather they exist in discrete micro-organs known as pancreatic islets (islets of Langerhans). Pancreatic islets are small (hundreds of microns) multicellular systems composed of 5 endocrine cell types held in pseudospherical structures containing a few hundred to several thousand cells. The other cell types, a, 6, e pp-cells secrete hormones which aid in glucose homeostasis by either counteracting or enhancing the effect of insulin secretion. /3-cells comprise the majority of cells within islets, typically between 50%-80% with variability between species (Cabrera et al., 2006; Kilimnik et al., 2012; Stefan et al., 1982).
Along with /3-cell content the cytoarchitecture of the islet, how cells are arranged, is different between species. In mice, /3-cells occupy the core of the islet and other cell types are located on the periphery. In humans, /3-cells are located preferentially next
3


to each other however the arrangement appears to be more stochastic in nature with /3-cell-/3-cell contact probabilities that are islet size dependent (Cabrera et al., 2006). These cytoarchitecrual differences are thought to contribute to distinct electrical dynamics between species, which are well characterized in mice but less so in humans. Studies suggest that the cytoarchitecture is age dependent and that islets in young humans is more similar to that of mice(Gregg et al., 2012); logically one would expect the electrical dynamics of islets in younger humans to be similar to that of mice however this has not been fully quantified. The combined mass pancreatic islets is typically represents 1-2% of the pancreas volume (lonescu-Tirgoviste et al., 2015) however, is variable depending on metabolic demand which is increased with obesity (Morioka and Kulkarni, 2010) and pregnancy (Rieckand Kaestner, 2010) and a failure of /3-cell expansion is associated with type 2 diabetes. Recent work also suggests the cytoarchitecure of human islets changes with type 2 diabetes (Morioka and Kulkarni, 2010) with decreased levels of/3-cell-/3-cell contacts in type 2 diabetic islets (Kilimniket al., 2011).
Electrophysiology and Stimulus Secretion Coupling of /3-Cells
Pancreatic /3-cells are a subset of electrically excitable cells within the pancreatic
islet of Langerhans and act on blood glucose levels to control regulated insulin secretion (Rorsman, 1997). The intracellular concentration of ions including Ca2+ ,
Na+, Cl- are lower than the extracellular space and intracellular concentration of K+ is held at a higher concentration that extracellular in /3-cells. The separation of charges across the cellular membrane creates an electrical field (or membrane po-
4


tential) which is described by the Nernst potential:
Ej
RT I0
--In
ZF L
(D
where R represents the gas constant, F represents Faradys constant, and Z represents the charge of the ion, I. At standard conditions the membrane potential is then proportional the the natural log of the ratio of ions outside to inside of the cell. For /3-cells the Nernst potential is approximately +40mV,+40mV,-35mV,-75mV for Ca2+
, Na+, Cl-, and K+ respectively (Drews et al., 2013; Ashcroft and Rorsman, 1989). Analogous to a simple RC circuit, the separation of charges creates across a boundary is acts as a capacitor and the resistors are channels which are selective to either Ca2+ Na+, Cl-, and K+. If ions are allowed to flow across the plasma membrane we interpret this as an increase in conductance, or a decrease in resistance, to channels specific to that ion.
5


Vm
Figure 1. Circuit Diagram of Excitable Cell Excitable cells (as /3-cells) have an analagous representation as simple RC cirucits where the cell membrane represents the capacitor (Cm), the channels represent resistors GNa,K,ca, and the voltage sources as the reverse potential for each ion ENa,ca,K-
The membrane potential becomes dependent on the level of current that flows across the plasma membrane:
^ = -(Ik + lea + INa + Ici + (2)
at
where the current of a specific ion is proportional to:
Ii = 9i* (Vin Ei) (3)
6


where gt is the conductance, or inverse resistance, of a channel specific to ion i, and Vm and Ei as the membrane potential and Nernst potential for ion i. A specific ion will continue to move across the membrane until the membrane potential reaches the specific Nernst potential. After which the flow of ions will stop and reverse if the membrane potential continues to go beyond the Nernst potential. Alternatively the conductance of the channel can be reduced to 0(S), i.e. channel closure, which will cease current flow of that ion. The conductance of a channel can be regulated by the membrane potential itself, voltage gated channels, or it can be regulated by binding of certain molecules, nucleotides, ligands, light, and other stimuli.
The hallmark of excitable cells is to use current flow to accomplish a task; in muscle tissue, including the heart, current flow of [Ca2+] controls contraction/elongation of muscle fibers to perform mechanical work. In /3-cells, as glucose levels elevate it is taken up by /3-cells through GLUT-2 transporters (mice) as well as GLUT-1 (humans). Glucose undergoes phosphorolation by glucokinase; a class of hexokinases with Kd of 6mM100mg/dl, making it more sensitive to subtle changes to blood glucose values than typical hexokinases (lynedjian, 2008). Phosphorolation of glucose to glucose-6-phosphate is the first step in the metabolic pathway which results in production of ATP and elevation of the ATP to ADP ratio. Briefly, glucose-6-phosphate is converted to fructose-6-phosphate (F6P) by glucose phosphate isomerase. F6P is phosphorolated by phosphofructokinase to create fructose-1,6-bisphosphate (F1,6BP). Two additional steps are required which creates glyceraldehyde 3-phosphate (GADP) which is final molecule before the pay off phase where a net gain of ATP and Nicoti-
7


Glucose
^ Glucokinase
Glucose-6-Phosphate
^ Phosphoglucose isomerase Frn ctose-6- Phosp hate
Phosphofructokinase

Fructose-1,6-bisphosphate
fructose-bisphosphate aldolase
triosephosphate isomerase
D-Glyceraldehyde 3-phosphate NAD+
glyceraldehyde phosphate NADH f dehdryogenase
ADP i
X Phosphoglycerate Kinase
ATP
NAD
ADP
T Phosphoglycerate mutase ADP ATP
T \7
2-Phosphoglycerate phosphoenolpyruvate
NADH
ATP
t
acetyl-CoA
f pyruvate dehydrogenase
pyruvate
enolase
pyruvate kinase
Figure 2. Metabolism of glucose generates ATP and NADH. Representative metabolic pathway in the /3-cell. Glucose is taken in through glucose transporters, phorosphorlated by glucokinase, and enters the metabolic pathway to generate ATP and NADH.
naminde adenine dinucleotide (NAD(P)H) molecules are created and subsequently result in the creation of pyruvate. Pyruvate is converted through the oxidative decarboxylation pathway to create acetyl Co A which enters the Krebs cycle to create additional ATP and NAD(P)H (Figure 2).
ATP sensitive potassium channels (KATP) contain the Kir6.2 binding site which binds ATP and acts to close the channel (Craig et al., 2008). Channel closure triggers a small depolarizing current which brings the membrane potential from -60mV
8


up to -45mV to trigger activation of voltage gated calcium channels (VDCCs) and sodium channels. This triggers a larger membrane depolarization and inducing large increase in cytosolic calcium levels ([Ca2+]^) (Drews et al., 2013). The rise in [Ca2+k is necessary to trigger insulin secretory vesicles to fuse with the plasma membrane and release insulin into the peripheral tissues and blood stream (Kasai et al., 2014) and also helps to regulate the production of insulin through [Ca2+k dependent transcription pathways (Lawrence, 2001). This pathway is referred to as stimulus secretion coupling in /3-cells where an increase in glucose (stimulus) is coupled to the activity of the metabolic pathway and to regulation of KATP channels.
9


GLUT-2
Glucose
Triose-P
Glycerol-P _
. Pyruvate
NADH
ATP -A K+
ATP \ (+)
4* ATP J
C02 i
. Ca2+
Ca2+
H||t
Insulin Secretion
Figure 3. Glucose stimulated insulin secretion. Uptake of glucose from GLUT channels enters the metabolic pathway to generate ATP causing KATP channel closure and membrane depolarization. Adapted from (Jensen et al., 2008)
Defects to the stimulus secretion coupling pathway can increase the risk or directly cause type2 diabetes including Katp channel dysfunction which linked to increased risk of type 2 diabetes (Gloyn et al., 2003) and removal of KATP function will cause overt diabetes (neonatal) (Koster et al., 2002) by interfering with primary initiator of Ca2+ elevation. Similarly defects to glucokinase activity (MODY II) will prevent sufficient phoshorloation of glucose and limit ATP production preventing KATP channel
10


closure. Neonatal diabetes can either resolve in time, or in half of the cases lead to
permanent neonatal diabetes (De Leon and Stanley, 2016).
/3-Cells as electrochemical oscillators
[Ca2+k levels are not continually elevated inside /3-cells after increases in glucose and subsequent activation of VDCCs. Indeed, chronic elevation of [Ca2+k can trigger cellular apoptosis pathways (Zhou et al., 2015). Due to rectifying potassium currents, depletion of ATP, and the membrane will be hyperpolarized and [Ca2+k will be brought back down to resting levels 100nM by uptake of [Ca2+]^ to the endoplasmic reticulum (E.R.) and via pump mechanisms to shuttle [Ca2+k to the outside of the cell. Due to the closure of KATP channels as long as ATP levels are sufficiently high, the membrane potential will rise again to VDCC activation range and the process will begin again. [Ca2+k and insulin secretion are oscillatory in nature and show in phase oscillations in isolated islets (Ravier et al., 2005) which scale proportionately to the frequency of [Ca2+] oscillation (Nunemaker et al., 2009). The study of [Ca2+k dynamics within /3-cells is then surrogate metric for insulin secretion activity and dynamics.
11


A
Cx36 + / +

C Cx36-I-
Dz
B Cx36 + / +
r 200
E
05
-100
0 10 D Cx36-/-
20
30
G12 + Fk

100 J= 2
50 2
o
Time (min)
10
Time (min)
i 20
Figure 4. Calcium and Insulin Show In-Phase Oscillations. In the absense of Cx36 insulin and [Ca2+]; oscillations are abolished. Adapted from (Ravier et al., 2005)
Persistent elevation of glucose, above 6mM, will cause /3-cells to show electrical
oscillations until extra-cellular glucose is brought down to 5mM (Figure 4).
Electrical coupling between /3-cells
Insulin secretion is regulated by extensive extra-cellular (Newsholme et al., 2010; Nyman et al., 2010; Ballian and Brunicardi, 2007; Rodriguez-Diaz and Caicedo, 2014) and intracellular cell-cell communication within pancreatic islets (Jain and Lammert, 2009). Paracrine signaling of glucagon, insulins counter hormone, will potentiate insulin secretion (Curry, 1970) whereas signaling of somatosatatin from (5-cells will attenuate insulin secretion(Strowski, 2000). Similarly, sympathetic and parasym-
12


pathetic nervous system innervation of pancreatic islets can lead regulation of insulin secretion (Thorens, 2014). /3-cells are coupled with Connexin36 gap junctions (Cx36) a cationic specific channel which allows the passage of cations and small charged molecules between adjacent cells. Cx36 form plaques at /3-cell junctions to facilitate high conductance between adjacent cells. While single channel conductance of Cx36 has been estimated to be 5pS (Moreno, 2004) the cumulative conductance between /3-cells has been measured to be 100-300pS (Benninger et al., 2008; Moreno, 2004). At elevated glucose levels /3-cells exhibit oscillatory electrical activity which is synchronized via Cx36 (Nlend et al., 2006). Although islet cytoar-chitecture plays a role in the level of synchronization (Hraha et al., 2014a) in mice, where /3-cells show high co-localization in the interior of the islet, Cx36 promotes the majority of /3-cells to electrically oscillate in phase (Benninger et al., 2008). Insulin oscillations in isolated mouse islets show in phase oscillations with electrical/[Ca2+]^ oscillations and similarly plasma insulin levels oscillate over several minutes.
13


Figure 5. Cx36 Couples /3-Cells.Model of Cx36 function in pancreatic islets. Glucose [G] enters the ceil through glucose transporters which enters the metabolic pathway to increase the ATP/ADP ratio. The elevated ATP/ADP ratio triggers closure of ATP sensitive potasiumm channels KATp which triggers a slight membrane depolarization to activate voltage gated calcium channels (VDCCs). [Ca2+] enters an individual cell from VDCCs, travels through Cx36 channels triggering depolarizing currents which activate VDCCs in neighboring cells. Coupling through Cx36 promotes the synchronization the heterogeneous oscillations (different color cells) into one cosistent frequency and phase.
Disruption to Cx36 via knockout leads to a loss of coordinated electrical oscillations and insulin oscillations (Ravier et ai., 2005). Similarly, plasma insulin oscillations are abolished in Cx36_/_ mice leading to glucose intolerance a pre-diabetic condition characterized by impaired glucose. Upon glucose challenge Cx36_/_ mice exhibited higher total glucose levels over a two hour period and had significantly reduced
14


first phase insulin secretion, but similar second phase insulin secretion. (Head et al., 2012b) In Db/Db mice, a mouse model of type2 diabetes, ex-vivo islets show loss of Cx36 function, coordinated [Ca2+k dynamics, and a loss of pulsatile insulin secretion. Similarly, in human type2 diabetes there is a loss of pulsatile insulin secretion (Lang et al., 1981; Lin et al., 2002) and a loss of pusatile insulin secretion is associated with hyperglycemia (Matthews et al., 1983; Matveyenko et al., 2012; Meier et al., 2013). [Ca2+] waves within islets are then surrogate metric for coordinated insulin secretion response and as such studying the biophysical mechanisms of [Ca2+] waves may play a role in determining how they are compromised in models of diabetes.
Heterogeneity of /3-Cells and Relevance Diabetes
Electrical dynamics in dissociated /3-cells, in the absence of electrical coupling by
Cx36, show distinct patterns of oscillations: repetitive spiking, transient depolarizations interlaced with rapid fast spikes, and cyclical plateau depolarization and hypero-larization periods (Kinard et al., 1999). Similarly, the insulin secretion responses and overall insulin content of dissociated /3-cells were also described in 3 groups of GFP expression driven by the insulin promoter (Katsuta et al., 2012) which had a positive relationship with insulin secretion. Glucokinase expression was also diverse among /3-cells (rat) leading to a distribution of glucose conditions before individual /3-cells would show electrical activity (Pipeleers, 1992).
Pancreatic islets are then composed of a heterogeneous population of /3-cells with varying electrical responses to glucose and the levels of insulin secretion. Cx36 func-
15


tion in individual pancreatic islets is functionally heterogeneous, although the role of Cx36 is to promote frequency and phase locking of a heterogeneous population of /3-cells to average out heterogeneity and give robust responses to glucose ensuring that cells with differential responses to glucose are either recruited to activity in the presence of high glucose or are rapidly silenced when glucose levels drop to fasting levels (Speier et al., 2007). Curiously, a functional role of heterogeneity may be to promote synchronization (Montaseri and Meyer-Hermann, 2016) and to control spatiotemporal [Ca2+]^ wave dynamics (Benninger et al., 2014). Recent work using a reporter gene of Fltp (Flattop) showed that /3-cells in mice could be divided into two distinct populations of maturity, and molecular and physiological features. Fltp negative cells were observed to have genes related to the functional properties of the glucose stimulated insulin secretion pathway, however had higher replication rate (Bader et al., 2016). Although many of these studies occur in disassociated islets, recently Rutter et. al. used halorhodopsin in a transgenic mouse model and a photoactivable sulfonuera in human islets to observe a population of /3-cells with high functional connectivity. These electrical silencing of the Hub cells silenced the global electrical activity of the islet (Johnston et al., 2016a).
In human islets, /3-cells were recently described as existing in 4 sub-populations based upon cell surface markers (Dorrell et al., 2016). These 4 identified sub-populations showed varying degrees of glucose stimulated insulin secretion even though the levels of insulin RNA and total insulin protein content were consistent across the subpopulations. Interestingly, in type2 diabetics the relative distribution of these subpop-
16


ulations is altered indicating that sub-populations may play a role in susceptibility in the pathophysiology of the disease (Kilimnik et al., 2011).
17


CHAPTER II
MATERIALS AND METHODS
Mouse Models
All animals were used under University of Colorado Anschutz Medical Campus IACUC approved protocol (B-9581407(1D)). Mice were held in a temperature controlled setting with a 12 hour light/dark cycle and given access to food and water continuously. /3-cell specific expression of ChR2-YFP was done through Cre-Lox mediated recombination by crossing a /3-cell specific Cre recombinase line, Pdx-qyq6Tuv/j jax (014647) with a ChR2(H134R) mouse model expressed on the Rosa26 locus (F{OSA26SorttnS2(CAG~COP4*H1S4R/EYFP)Hze/J JAX: 024109). Cx36"/- mice were generated as previous descried (Degen et al., 2004). Genotyping was verified
through qPCR (transetyx).
Islet Culturing
Human islets were obtained from the Integrated Islet Distribution Program (Table S1) during years 2013-2016 (Ca2+, static insulin, gap junction measurements) or 2012-2016 (perifusion measurements). Islets were cultured in CMRL at 37C, 5% C02, for 24-48 hours prior to imaging or insulin secretion assays. Mouse islets were
cultured in RMPI media.
Microscopy and Calcium Imaging
ChR2 expressing isolated islets were mounted on 35mm glass bottom dishes in imaging solution (125mM NaCI, 5.7mM KCI, 2.5mM CaCI2, 1.2mM MgCI2, 10mM Hepes, and 0.1% BSA, pH 7.4) and imaged using a Zeiss LSM780 system (Carl
18


Zeiss, Oberkochen, Germany) with 20X 0.8NA PlanApocrat objective at 37C in the presence of 2mM, 5mM, 11mM glucose. For [Ca2+] measurements, islets were loaded with 3pm Rhod-2 AM (AAT Bioquest, Sunnyvale, California) for 30 minutes at 37C in imaging solution. Rhod-2 was excited using a 561 nm solid state laser, which minimizes ChR2m34K activation (Prigge et al., 2012), and fluorescence emission detected at 580nm-650nm using a multi-anode PMT spectral detector. Images were acquired at 1frame/sec. (scan time 650ms) at 20pm depth from the bottom of the islet. ChR2 activation was achieved using a 458nm Ar+ laser line. Activation regions in the islet were defined using the Zen software bleaching module and each activation region set to scan 25 times over 1s. Time-series were recorded alternating between 10 Rhod-2 images followed by a ChR2 activation sequence, and repeated for 40 seconds. NADH(P)H autofluorescence was imaged under 2-photon excitation using a tunable mode-locked Ti:sapphire laser (Chameleon, Coherent, Santa Clara, California) set to 710nm, and fluorescence emission detected at 400nm-450nm using the internal detector. Z-stacks of 6-7 images were acquired spanning 5pm depth. Human islets were loaded with 4p M Fluo-4 for 90 minutes at room temperature and imaged on an inverted Nikon Eclipse-TI wide-field microscope using a 20x 0.75NA objective, at 37C. Images were acquired 1 frame/sec using 490nm/525nm emission filter. Islets from batches in which cell viability was <80%, or islets with absent Fluo4-AM signal or significant drift were excluded from subsequent analysis.
19


Psuedoislet Aggregation
MIN6 /3-cells were grown and maintained in a T75 flash in growth media (DMEM) containing 4.5g/l glucose, 10% fetal bovine serum, 50ug/ml streptomycin, 50U/ml_ penicillin, 0.5ug/ml Fungizone, 1mM sodium pyruvate and 50uM beta-mercatoethanol. Cells were re-suspended in media and were seeded into 48 well hydrogel microwell devices (Bernard et al., 2012) at 500,000 cells per 0.5ml_ media. The cells were centrifuged in the device at 160rcf for 2.5 minutes and placed in an orbital shaker for 2 hours in 37C 5%C02. 5 days post culture aggregates were passed over a 40pm strainer to remove small aggregates or particles and the remaining large aggregates cultured in 35mm dishes. Polystyrene particles 10pm in size were sterilized in ethanol and treated with fetal bovine serum and were added to the cell suspension at desired
concentrations prior to aggregate seeding.
[Ca2+] Coordination
Human [Ca2+] images were analyzed in MATLAB to assess activity and coordination. A 4-pixel averaging filter was first applied. A peak-detection algorithm (Yoder, 2015) recorded the locations and amplitudes of oscillations across all pixels of the islet. Regions in which no significant peaks could be detected were determined to be inactive. Coordinated regions were segmented based on the coincident presence of the time-points of each detected peak. This was followed by cross-correlation analysis between time courses of each subregion. If the correlation coefficient was above 85% the two sub-regions were considered to be highly coordinated and merged into a larger region.
20


ChR2 Activation Analysis
All analysis was performed using custom scripts in MATLAB (Mathworks) [Ca2+] measurements were quantified by a peak detection algorithm testing for ChR2 stimulated action potentials on a pixel by pixel basis. A pixel was defined as active if [Ca2+] peaks were identified at the time of ChR2 stimulation and no more than 30% of the peaks identified could result from non ChR2 stimulated events: in this way we could quantify areas that may have been poorly stained with [Ca2+] indicator. Areas that could not be stimulated while activating the whole islet were not included in the analysis of subregion stimulation. NAD(P)H response for each quadrant was calculated by averaging the intensity across the z-stack fluorescence in that quadrant and calculating the percentage change between 11mM glucose and 5mM glucose.
To quantify [Ca2+] oscillatory frequency in Cx36-/- islets, the peak detection algorithm from above was used to record the spatial locations and times of all oscillations across an islet. Individual cells were located by segmented regions based on coincident time-points of detected oscillations. NAD(P)H responses were calculated by mapping each cellular region onto the NAD(P)H images and the response for each cellular location was calculated. For immunofluorescence images, analysis for hormone staining and ChR2-YFP expression was restricted to single dissociated cells identified as single nuclei locations: doublet, triplet or aggregates of cells were excluded form analysis.
21


FRAP
Cx36 function was measured using Fluorescence Recovery After Photobleaching (FRAP), as previously described (Farnsworth et al., 2014). Islets were loaded with 12uM Rhodamine-123 for 30 minutes at37C in imaging solution. Rhodamine-123 was excited using a 488nm Ar+ laser line and fluorescence emission detected at 500nm-580nm. 2 baseline images were initially recorded; a region of interest was then photobleached for 30 seconds achieving on average a 44% decrease in fluorescence; and images were then acquired every 15 seconds for 6 minutes. The fluorescent recovery curve can be modeled by the following:
I(t) = (loo ip)(l e~kt) + Ip (4)
where \p and l^ represent the intensity after bleaching and at the end of imaging, respectively. The recovery rate, k can be solved by rearranging equation 4.
e
kt
Ip I(t)
Tqo Ip
+ 1
(5)
Where the negative slop of the natural log of equation 5 is equal to the recovery rate. Immunofluoresence
Isolated isolates were dissociated using Accutase (Sigma-Aldrich, St. Louis, Missouri) and plated into 8 chamber Lab-Tek dishes treated with 804G cell matrix. 24 hours after plating cells were fixed using 8% paraformaldehyde in PBS for 10 min-
22


utes. Antigen retrieval was first applied to the cells using 0.05% Tripsin with 10% CaCI2 in dH20 for 30 minutes at 37C. Cells were then permeabilized using 0.1% triton X-100 and 5% donkey serum in PBS for 2 hours. Mouse anti-glucagon (Abeam ab10988, Cambridge, United Kingdom) and guinea pig anti-insulin (abeam ab7842) primary antibodies were incubated with cells at a 1:500 dilution at 4C for 24 hours. Cy3 anti-guinea pig (Jackson 706-165-15, West Grove, Pennsylvania) and Alexafluor 647 anti-mouse (Jackson 715-605-150) secondary antibodies were applied to cells at a 1:500 dilution at room temperature for 2 hours. After washing cells were treated with DAPI fluoromount (Sigma-Aldrich) and imaged on a Zeiss LSM800 confocal microscope, DAPI, YFP, Cy3, Alexa647 were respectively excited with laser lines at 440nm, 488nm,561nm,640nm and detected over respectively 410-470nm, 500-545nm, 565-615nm, 640-700nm wavelength bands. In intact islets FM 4-64FX was
excited at 561 nm and detected over 600-620nm.
Insulin ELISA
Insulin secretion was determined in static assays by incubating 10 human islets/tube in Krebs-Ringer buffer (128.8mM NaCI, 5mM NaHC03, 5.8mM KCI, 1.2mM KH2P04, 2.5mM CaCI2, 1.2 mM MgS04, 10mM HEPES, 0.1% BSA, pH=7.4) with 2mM glucose for 1 h, followed by 1h incubation with either 2mM or 20mM glucose. Supernatant (secreted fraction) was collected and remaining islets were lysed with 2% TritonX-100 and frozen overnight for analysis of insulin content. Insulin was measured with a mouse ultrasensitive ELISA kit (ALPCO, ChrystalChem). Data are presented as insulin secretion normalized by content for each sample.
23


Insulin secretion dynamics were assessed by perifusion measurements, as described in (Kayton et al., 2015) in the Vanderbilt Diabetes Research Center. Only preparations which showed at least a 2-fold increase in insulin secretion from 5.6mM to 16.7mM, or 5.6mM to 16.7mM+100pM IBMX, were further analyzed. Perifusion time-courses were analyzed in MATLAB. Time-points for the addition/removal of 16.7mM glucose (IBMX) were recorded and the fold change in insulin secretion calculated using the maximum insulin secretion value in the stimulation period against either all initial baseline measurements prior to 16.7mM glucose, or the previous 4 timepoints prior to 16.7mM glucose+IBMX. At the end time-point of 16.7mM glucose (IBMX), an exponential decay a.exp(-b.t)+c was fit, with b the rate of decay.
24


Classification Learners
Classification systems were generated in MATLAB using k-fold cross validation (k=5) (Kohavi et al., 1995). Receiver Operating Characteristic (ROC) curves were generated by calculating sensitivity vs. 1-specificity curves at different cutoff values within the trained classification model (Swets et al., 1988). In order to evaluate model performance the area under the curve (AUC, or C-statistic) was calculated by numerically integrating the ROC curve using the trapezoidal method.
25


CHAPTER III
DEVELOPMENT OF COMPUTATIONAL MODEL 1
Introduction
Hodgkin-Huxley (HH) like models have been instrumental in describing electrical dynamics in multiple excitable systems. The canonical HH system relies on a set of 4 differential equations describing the current through a cell using sodium, potassium, and leak conductances as well as an injectable current.
1 = + 9Kn\Vm Vk) + gNamsh(Vm VNa) + - VI) (6)
where describes the conductance of a channel selective to ion i, and n,m,h are gating variables representing the functional voltage dependent gating kinetics of each channel. Upon sufficient injectable current, I, the system undergoes a bifurcation shift between steady membrane potential and oscillatory potential. The system will continually depolarize due to sodium current and hyperpolarize due to the rectifying potassium current (Figure 6).
1The work described in this chapter was accepted for publication on 7/16/2014 Hraha TH, Westacott MJ, Pozzoli M, Notary AM, McClatchey PM, et al.(2014) Phase Transitions in the Multi-cellular Regulatory Behavior of Pancreatic Islet Excitability. PLoS Comput Biol 10(9): e1003819. doi:10.1371 /journal.pcbi. 1003819
26


0.0
-0.6 + 0
200
Time (s)
400
Figure 6. Membrane potential response as a function of injected current in HH model. With increasing injected current the membrane potential of the simulated cell increases until a threshold value upon which the system undergoes hopf bifurcation and a steady oscillatory solution will exist
/3-cells were found to generate action potentials in response to glucose (Dean and Matthews, 1970) and were regulated specifically by ATP binding potassium channels
27


(Ashcroft et al., 1984). HH like models of /3-cell electrical activity were first developed in the early 1980s (Chay and Keizer, 1983). The model consisted of a calcium activated potassium channel, a voltage gated potassium and calcium channel, a leaky conductance for sodium and chloride, and an injectable current. The membrane potential is then represented as:
dV
Cm. = {9i<,Ca+9 k, Vm) (VmEK)+2gCa Vm (Vm-ECa)+gL(Vrn-VL)+I (7)
Later Magnus and Keizer began modelling medabolic pathways such as glycolyis, oxidative phosphorloation, and the TCA cycle (Magnus and Keizer, 1998b). Fridy-land et. al began to include seperate currents for [Ca2+] dynamics in the ER etc. The Cha-Noma model is an adaptation of model derived by Fridyland et. al. (Fridlyand et al., 2003). With minor changes to metabolic components were adapted to more recent literature.
28


IcaV iKDr ItRPM IsOC IKCa(BK) IKCa(SK) IkATP IbNSC
Figure 7. Current sources in Cha-Noma Model. Glucose is metabolized through rates of glycolysis to generate NAD(P)H (Re) and ADP is converted to ATP. Adapted from (Cha et al., 2011a).
Briefly, the Cha-Noma model contains 18 coupled differential equations which model ATP generation from glycolysis and oxidative phosphorlation from [Fridylad 2005]7.
The membrane potential is described by 11 current sources including voltage activated calcium channel, KATP channel, delayed rectifying potassium current, sodium potassium and sodium calcium pumps, calcium activated potassium currents, store operated current and plasma membrane calcium pump currents.
29


a
dV
dt
IcaV + ItRPM + IsOC + IbNSC + 1KDr + IkCci(BK) + IkCci(SK) + Ik ATP + InciK + Ino-Co, + I PMC A
(8)
Model Description
In order to accurately model electrical dynamics within pancreatic /3-cells the single cell model was expanded into 1000 heterogeneous cells. While other groups have implemented multi-cellular models using 10x10x10 matrix (Silva et al., 2014) of cells we chose to use a sphere packing algorithm (Skoge et al., 2006) in order to generate a random grouping of cells. First by assembling a cubic packing of 4000 spheres then we choose the first 1000 cells closest to the center of the packing assembling a spherical structure. With this we know the coordinates of the center of each cell. Next in order to calculate each cells nearest-neighbor (in order to assign functional connectivity) we calculated the distance between each cell center. If this distance was less than 1,4x the diameter of a cell we considered those cells to be functionally connected. This gave a Gaussian distribution of connectivity where each cell was on average connected to 5.52.5 cells. In order to introduce heterogeneity into the system we took 10 parameters from each cell (summarized in table 1) and assigned Gaussian distributions to that property. This either made cells more/less electrogenic with varying oscillations frequencies and duty cycles.
30


Table 1. List of Parameters used in Cha-Noma Model. Parameters used for islet model. Heterogeneity is based on a Gaussian distribution, unless otherwise indicated, with a standard deviation as a percentage of the mean value ( Independent Variable Description Value
Cm Cell Capacitance 6.158pF
volj Cytosolic Volume 764fl
VOler Endoplasmic Reticulum Volume 280fl
fi Cytosolic Ca2+ Buffer Strength 0.01
fer ER Ca2+ Buffer Strength 0.025
PcaV Converting factor for lCav 48.9 pA mM'1
P| GkCA(BK) Conductance of I Kca(si<) 2.13 pA mV'1 (10%)
PkCA(SK) Converting factor of Ikcbcsk) 0.2 pA mM'1
PbNSC Converting factor of lbNSC 0.00396 pA mM'1
Psoc Converting factor of lSOc 0.00764 pA mM'1
Ko.5ER Half activation cone. Of Ca2+ in ER 0.003 mM
Gk(ATP) Max conductance of Ikatp 2.57 pA mV'1 (25%)
Gcoup Average conductance of Cx36 0.12 pS (40%)
PNaK Max amplitude of lNaK 350 Pa ms
PNaCa Max amplitude of lNaCa 204 pA (10%)
PpMCA Max amplitude of IPMca 1.56 pA
PsERCA Max pump rate of Ca2+ into ER 0.065 fl ms'1 (10%)
PRel Converting factor for ER Ca2+ release 0.76 fl ms1 (10%)
kglc Rate constant for glycolysis 0.000076 (25%)
l^pox Rate constant of p-oxidation 0.0000063 ms r(10%)
POp Max rate of ATP production 0.0005 ms1 (10%)
[ATPtoJ Total ATP species 4mM(10%)
Latp Rate Const, of Ca2+ind. Ca2+ consumption 0.00062 ms'1
kATP,Ca Rate Const, of Ca2+ dep. ATP consumption 0.187 mM'1 ms'1
Kadpj Rate Const, of ADPf to ADPb 0.0002 ms'1
Lad p, b Rate Const, of ADPb to ADPf 0.00002 ms'1
31


To model Cx36 current between cells we introduced a current proportional to the transmembrane potential between two connected cells:
Qcoup * (9)
Where i and j are each coupled cell and gcoup is the average conductance between cells. This value has been measured to be 100pS, however rather than assigning a gaussian distrubtion of conductivity we uesd a gamma distribution following Farnsworth et al. 2014 (Farnsworth et al., 2014). We implemented this coupled model in C++ and solved the system using a forward Euler method using a time step of 150ps. Simulations were run on the JANUS cluster (University of Colorado Boulder)
Individual /3-cells show a wide range of responses to glucose (see introduction) which is represented in the model when we set the coupling current to 0. However, when we functionally couple the cells the simulated islets show coordinated [Ca2+] waves when glucose levels are above 6.5mM and are functionally quiescent below 5mM glucose (Figure 8).
32


Figure 8. Glucose activated [Ca2+] oscillations in simulated islet [Ca2+] oscillations in WT (top) Cx36-/- (middle) as a function of glucose concentration (bottom). WT islets show abrupt change from quiescent to oscillatory near 8mM glucose. Alternatively Cx36-/- islets have contain cells which show oscillations at lower glucose concentration.
Model Verification
To validate the coupled models effectiveness at simulating properties of ex-vivo islets we chose a model of neonatal diabetes. Neonatal diabetes is a relatively rare disease, effecting only 1 in 200,000 individuals. This condition is associated with a mutation with the ATP regulated potassium channel (KATP channel) responsible for controlling the initial depolarization in response to elevated glucose. While there are multiple mutations associated with this channel (Pinney et al., 2008), many decrease the sensitivity of the channel to ATP on the Kir6.2 subunit of the channel and render them hyper-active requiring higher levels of ATP to close and depolarize the mem-
33


brane. We generated a mouse model with a mutation on the ATP binding domain of the KATP channel (Kir6.2[AAr30K185Q]-GFP) under inducible Cre^-recombinase which would only be expressed upon tamoxifen injection. The mutation would express mosaic qualities, showing expression in only a certain fraction of /3-cells within islets, the percentage of which were dependent on the number of injections given to the mouse. Each islet could now have between 0%-60% mutated channels, dependent on the number of injections given to the mouse (Figure 9). We define Pexc as 1-%mutated cells where Pe£CC=100% or 1 as WT and Pexc=0% as islet wide expression of Kir6.2[AiV30K185Q]-GFP
P
1 exc
Kir6.2[AN30'K185Q]
100% (WT) 90% 80% 45%
Figure 9. Penetrance of Kir6.2[AiV30K185Q]-GFP Varying doses of Tamoxifen (0-3) injections show 0%-60% of cells with positive expression, controlling the overall excitability of the islet.
The effect the mutant KATP channels had on a cells individual electrical activity were to silence it, regardless of glucose levels Making it unresponsive to even high glucose levels (20mM). As /3-cells are electrically coupled, in a islet with moscaic ex-
34


pression of hyperactive KATP channels there would be a mix of electrically inactive and active cells. The inactive KATP mutant cells act as an inhibiting current sink, preventing the neighboring, electrically active cells, from completely depolarizing.
To image cytosolic [Ca2+], islets were loaded with 4pM FuraRed-AM at room temperature for 90 minutes and imaged in a spinning-disk confocal microscope. Images were acquired at 1 frame/sec using a 488 nm diode laser for excitation and a 580-655 nm long-pass filter for emission. Kir6.2[AJV30)jK:i85Ql-GFP was imaged at 488nm and aquired at 500nm-520nm (CHECK).
If an islet had a small fraction of cells with mutant KATP channels (< 10%) the islet would still be able to show oscillations at 20mM glucose however the overall activity would show a small linear decline (Figure 101,11). However at a critical number of inactive cells (>20%) the current associated with closure of KATP channels in un-muated cells would sink to those cells with mutated KATP channels causing a rapid global collapse of electrical activity throughout the islet (Figure 10il I). After this critical value there would again be a linear decay in activity as a function of KATP mutant cells (Figure 101V).
35


0 60 120 180 240 300
Time (Seconds)
Figure 10. Phase transition in excitability A) Percent cells showing [Ca2+]i elevations as a function of number of excitable cells, as determined by lack of GFP and thus. Kir6.2[AiV30K185C?] expression (i.e. Pexc = 1-% GFP) B) Representative [Ca2+]i data for islets indicated in A, from regions of wild-type (I), pre-critical (II), critical (III) and post-critical (IV) levels of Pexc. Left: Areas of activity are highlighted in red and scale bars represent 50/i,m. Right: Representative time-courses of normalized Fu-raRed calcium dye fluorescence for cells within each islet, where vertical scale bar indicates 20% change in fluorescence. Red time-courses are determined to be active, black time-courses are determined to be inactive.
Insulin secretion is tightly regulated by [Ca2+] levels within /3-cells so we anticipated insulin secretion to follow similar critical behavior as [Ca2+] activity in Kir6.2 mutant mice. We performed an insulin ELISA on islets from (Figure 10) using the recorded GFP values from [Ca2+] micrsocopy. Insulin secretion followed a similar critical trend (Figure 12) to [Ca2+] activity.
36


*
100
80
60
40
0
20 40 60 80 100
100 (%GFP)
0
0-79% 80-99% 100%
100 (%GFP)
B £ 0.6
*
0.6
*
O)
c
Q_
20 40 60 80 100
100 (%GFP)
0-79% 80-99% 100%
100 (%GFP)
Figure 11. Link between phase transitions in [Ca2+]i and physiological parameters. A) Percent cells showing [Ca2+]i elevations averaged over islets from each Kir6.2[AiV30K185Q]-expressing mouse as a function of Pexc (100%-%GFP). Right: Mean(s.e.m.) for data binned to wild-type, pre- and post-critical ranges as determined by %GFP indicates significant difference (p<0.0001) between data as indicated. B) Plasma insulin levels from each mouse as a function of Pexc. Right: Mean(s.e.m.) for data binned as in A. indicates significant difference (p<0.05) between data as indicated.
We tested if we could reproduce this critical behavior in-silco by mutating the ATP response kinetics of the KATP channel. The current through the KATP chanel was described as (Magnus and Keizer, 1998a):
Ikatp 9k{atp)Pk{atp){V Vk)
(10)
37


where g represents the open channel conductance and poK(ATP represents the ATP regulated open probability described by:
POk(ATP)
0.08(1 + + 0.89(l^fl)2
1 _L ([APP]\ 2/-I I 0.45[ADP] [ATP] v V 0.01 / T 0.026 ' 0.05 /
(11)
To model expression of Kir6.2[AJV30)/f 185Q] the open probability was modified to in-
clude a constant current term:
pOK(ATP)Mut 7 (pK(ATP)) + (1 7) (12)
where 7 was set equal to 0.5. This term contributes an addition to the open probability following results from previous characterizing of glucose vs IKatp response curves (Ashcroft, 2005; Hattersley and Ashcroft, 2005; Koster et al., 2005). Cells in-silico were spatially randomly assigned as either WT or mutant K(ATP) currents. Individual cells were described as electrically active if their membrane potential reach values over -45mV. As with ex-vivo results, the coupled dynamical oscillator model predicted similar critical behavior (Figure 12). Initially there is a linear decay in activity at sub-crtical values (0%-12%) followed by a rapid dropoff in electrical activity near 20% mutated cells and a linear decay in supra-critical values. Similarly, to Figure 10, [Ca2+] duty cycle decreases and oscillatory frequency increases as a function of mutant cells (Figure 12 B).
38


A
IV III II i
mJUIJl-Ul
JULU4
xaxxxjii
XXXJ_XX1
IV
I
IIXXXJZ.
0 60 120 180 240 300
Time (Seconds)
Figure 12. Coupled oscillator model describes experimental phase transitions.A) Percent cells showing [Ca2+]i elevations in simulated islets as a function of fraction of excitable cells (Pexc), as set by the % cells lacking ATP-insensitivity. Solid line represents mean of simulation results generated from 5 random number seeds, dashed lines represents 95% confidence intervals of simulations. B) Representative simulated [Ca2+]i time-courses for parameters indicated in A, from regions of wild-type (I), pre-critical (II), critical (III) and post-critical (IV) behavior, as in Figure 11. Vertical scale bar indicates 20% change in simulated [Ca2+]i. Red time-courses are determined to be active, black time-courses are determined to be inactive. C) Percent cells showing [Ca2+]i elevations in simulated islet as a function of number of excitable cells (Pexc) for varying mean gap junction conductance values. Filled squares indicate experimental data from Kir6.2[AAr30K185Q]-expressing islets in Figure 10.
Interestingly, the level of Cx36 coupling predicts different responses to increases in the number of mutant cells (Figure 12C). While 120pS was used as the best-fit to ex-vivo data (black dots) by decreasing Cx36 conductance we anticipate a less criti-
39


cal behavior. This model of pancreatic islet dysfunction follows the analogy of mean-field theory in the classical ising model in ferromagnetism (Figure 13).
4 Inactive Cell
0 Katp Mutant
If Active Cell
^ATP Wild-Type
^ATP Pre-Critical (Pexc > 85%)
Excitability / Composition 2mM 20mM
Resulting Activity 20mM
Figure 13. Mean-field theory analogy of /3-cell network activity. The excitability of constituent units (i.e. their glucose sensitivities) and the resulting network activity takes into account coupling and different experimental perturbations. A) Cells of wild-type islets are inexcitable at low glucose (2 mM) and ail excitable at high glucose (20 mM), therefore wiid-type islets are respectively fully inactive and fully active. Cells expressing KirS^^30*'185^ are glucose-insensitive and constitutively inexcitable. When Kire^^30*'185^1 penetrance is <15% (Pexc>0.85) there are insufficient inexcitable ceils to suppress global activity, and so coupling leads to inexcitable ceils being recruited to be active. Flowever, when KirS^^30*'185^1 penetrance rises above 15% (Pexc<0.85), global quiescence ensues where the majority of normally excitable ceils are rendered inactive.
/3-cells are given simple up/down parity where inactive or unable to be active are represented by down and large down arrows, respectively. Healthy ceils are inactive
40


at 2mM glucose and transition to active at 20mM glucose. At the pre-critical stage of mutation there is a small subpopulation of cells which are unable to become electrical active on their own, however do to electrical coupling they can be recruited to be active at elevated glucose levels. At the post critical stage there are sufficient numbers of inactive cells, which have higher hyperpolarizing current than surround depolarizing cells, such that do coupling the entire system is brought down into an inactive state.
Weve examined how whole islet electrical activity undergoes a phase transition when adding electrically inexcitable cells. Due to the coupled nature of islets we hypothesized that islets may undergo a similar phase transition in electrical syn-crhonization if instead of including inexcitable cells we include functionally uncoupled cells. These electrically uncoupled cells could consist of multiple sources including alternative islet cell types (such as a or 6 cells) or /3-cells which do not express Cx36. Alternatively, another source of uncoupled cells would be to include small micro-particles which incorporate into the structure of the islet creating a functional scaffold but void of any electrophysiolgy properties. The latter was accomplished through the incorporation of polysterene microparticles in re-aggreated pseudoislets. These psuedoislets are created by aggregating a /3-cell line (MIN6 Cells) into a spherical architecture (Figure 14).
Pseudoislets were stained with Fluo-4AM to monitor [Ca2+] activity and imaged in 20mM glucose to monitor [Ca2+] oscillations. Three to four islets of each particle
41


seeding were imaged and [Ca2+] coordination calculated using the algorithm described in the methods. The percentage of microparticles within islets was caciuiated by thresholding the fluorescent signal from the microparticles and calculating the net area of fluorescence at the imaging plane.
Figure 14. Model of Synchronization disruption.Three dimension pseudoislets are formed inside microwelis consisting of green /3-cells or red microparticles. Beyond a limit of microparticles we expect the islet pseudoisiet would be beyond a critical coupling threshold and electrical activity would no longer show synchronous oscillations. Scale bar indicates 1C%im.
42


We observed that pseudoislets with low percentage of microparticles display [Ca2+] oscillations identical to that of mouse islets (Figure 15) and although there was dropoff at higher levels of seeding (i.e. >20%) islets were still coordinated overall. However at 38% seeded particles there was a rapid dropoff in [Ca2+] coordination. Indicating that between 20% and 40% seeding there exists the critical coupling limit within pseudoislets.
Figure 15. Loss of electrical synchronization with microparticle incorporation (Left) [Ca2+] Activity in pseudoislets as a function of microparticle incorporation. (Right) Quantification of the largest area of coordinated [Ca2+] activity normalized to islet size (Max. coordinated area) averaged over psuedoislets as a function of microparticle incorporation.
To replicate this experiments in-silico we modified our coupled multicellular model to instead of expressing a certain fraction of mutant, in-excitable cells, we uncoupled a variable fraction of cells from their neighbors. This creates an islet with effective coupling holes, similar to the pseudo-islets. At high fraction of holes the oscillations appear to uncouple (Figure 16B). The overall islet synchronization was calculated by first taking the Hilbert transformation of each cells membrane potential time-course. The Hilbert transform gives an analytical representation of the input signal, returning
43


a complex answer containing the phase information of of each timecourse. The instantaneous synchronization factor (r) at each timepoint could be calculated with the following:
r(t) = T jh exp(i0(t)j)
3=1
(13)
On >
E
re -20
C
0)
0) c
re
xi E
0)
-60-
-80
V

WU
w
u
Time (m)
Time (m)
o
re
li-
re
o
re
o
re
>
Figure 16. Model of Synchronization DisruptionThree dimension pseudoislets are formed inside microwells consisting of green /3-cells or red microparticles. Beyond a limit of microparticles we expect the islet pseudoislet would be beyond a critical coupling threshold and electrical activity would no longer show synchronous oscillations
44


Where 0(t)j represents the phase of each cell at timepoint t and N represents the number of cells (not including uncoupled cells). A time-average of r gives the overall synchronization in the time-course where 1 represents complete coupling and 0 represents no coupling. Our findings show that at 0% uncoupled cells (holes) there islet shows very high levels of coupling (Figure 16 C) near 0.9. The synchronization decreases with a slight linear trend until 40% after which there is an critical transition
to low synchronization.
Conclusions
The findings of these studies show that expanding a well-established single cell model of pancreatic /3-cells electrical activity into a representative multi-cellular pancreatic islet accurately recapitulates whole islet electrical activity. The model features several attributes which contribute to its function and strengths heterogeneity in electrical and metabolic properties between /3-cells which give cells varying degrees of electrogenicity in response to glucose (Figure 8) which has been described previous (Pipeleers, 1992). Islet architecture has been well characterized (Cabrera et al., 2006) we created a random spherical packing of cells to most accurately represent cells in their in-vivo state and while some studies have used cubic packing of cells (Silva et al., 2014) others used ordered hexagonal closed packing (Nittala et al., 2007), none have used random packing to create unique islets with every simulation. With our random spherical packing we assign a Cx36 coupling current unique to each cell and following a distribution shows trends similar a right-skewed Gaussian, or gamma, distribution (Farnsworth et al., 2014). At glucose levels <8mM the
45


coupled islet is quiescent while an electrically uncoupled islet contains populations of cells which show electrical activity. In the coupled state we observe a robust transition to an oscillatory state at 8mM glucose, consistent with ex-vivo observations of electrical activity in islets. The behavior of our model recapitulates islet function, we next tested how the model could predict dysfunction in two unique situations that center around electrical coupling between /3-cells.
We created an inducible transgenic mouse model showing mosaic expression of a KATP channel insensitive towards ATP through Kir6.2 mutation [cite some people] (Figure 9) where increasing the dosage of tamoxifen increases the percentage of cells expressing this channel. Cells expressing | 46


incorporating cells with high KATP channel open probability by adding a constituitive open probability to the KATP. At low levels of mutant channel expression islets show a small decrease in electrical activity. At 15% mutant cells the simulated islet undergoes a similar phase transition where it rapidly transitions to quiescent.
We next examined how deleterious effects to electrical coupling, through the inclusion of non-coupling cells alter the coordinated electrical dynamics in aggregated pseudoislets (Figure 14) and paralleled this in the multicellular model by randomly disconnecting populations of cells (Figure 16). While both models showed critical loss in synchronization there was a discrepancy in the level of uncoupled cells seen between the pseudoislets and the multicellular model. The former showed loss in synchronization at uncoupled cells at levels exceeding >20% whereas the multicellular model the decrease was right shifted towards >50%. The discrepancy may result from a few differences between the pseudoislets and model. We set the average coupling conductance between cells at 120pS with a typical cell functionally connected to 5 surrounding cells. Less is known of the functional coupling between pseudoislet aggregates and expression of Cx36. A lower level of functional coupling may help explain this discrepancy and could be examined further by directly quantifying Cx36 function using a fluorescence recovery after photobleaching assay (Farnsworth et al., 2014). Furthermore, the extent to which microparticles disrupt electrical coupling may show a stronger effect than uncoupling similar numbers of cells within the multicellular model possibly through mechanical interaction.
47


In summary, these two studies verify the single cell electrophysiology model based on Cha et al. can be expanded to a coupled multicellular model that recapitulates two aspects of pancreatic islet dysfunction. The former based upon critical loss the electrical activity caused by electrically inactive /3-cells coupled to those of normal electrical function whereas the latter demonstrates pancreatic islets show critical loss in electrical synchronization upon the inclusion of non-coupling cells. Furthermore, these results show how that a small subpopulation of cells with altered function can exert disproportionate control over the electrical activity of pancreatic islets.
48


CHAPTER IV
SPATIALLY ORGANIZED SUB-POPULATIONS OF CELLS CONTROL ELECTRICAL ACTIVITY AND DYNAMICS ACROSS THE ISLET OF
LANGERHANS2
Introduction
The emergent properties of multicellular systems have led to increased study of the architecture and biological heterogeneity in governing spatiotemporal dynamics (Breakspear and Stam, 2005; Pandit and Jalife, 2013). However, the inherent complexity in multicellular systems often renders them difficult to study in their intact state. As a result, systems are often broken apart into more manageable pieces and their coupled functions and dynamics inferred from individual component properties (Katsuta et al., 2012; Strogatz, 2001). While subpopulations of systems are identifiable using these methods, their exact role in controlling dynamics and system regulation in the intact state is lost. New techniques that can study structures in their intact state, in order to preserve spatial information of signaling dynamics, will therefore help elucidate the role of cellular subpopulations in complex systems. One such system that shows complex multicellular regulation, yet has a tractable scale for studying with cellular imaging and computer modeling approaches, is the islet of Langerhans;
where dysfunction to the islet generally causes diabetes. Individual /3-cells when
2The results presented in the section were submitted on 2/14/17 and a revised manuscript submitted on 6/8/17. Westacott, M.J., Ludin, N.W., Benninger, R.K.P., Spatially organized sub-populations of cells control electrical activity and dynamics across the islet of Langerhans. Biophysical Journal
49


electrically isolated show large variance in glucose sensitivity metabolic activity electrical dynamics and insulin secretion (Piston et al., 1999; Zhang et al., 2003; Jetton and Magnuson, 1992); and this is distinct from the uniform coordinated electrical and secretory response in the intact islet. Understanding how cell variability contributes to the overall function of the islet, and how the presence of sub-populations of cells affects islet function in pathogenic conditions is poorly understood. For example small populations of inexcitable cells that show altered KATP activity can suppress activity across the islet in the presence of coupling (Speier et al., 2007; Hraha et al., 2014b; Rocheleau et al., 2004). This causes a sharp transition between global activity and global quiescence that can result from small changes in glucose levels and KATP activity, and is physiologically important (Stozer et al., 2013). The coordinated dynamics of electrical activity show spatial heterogeneity or Small World properties in common with other biological systems, where like-dynamics are restricted to subregions of the islet (Johnston et al., 2016a; Benninger et al., 2008). Propagating calcium waves that mediate synchronization of [Ca2+]i oscillations also initiate from subregions of the islet; which has been suggested as a pacemaker region defined by local excitability (Benninger et al., 2008, 2014). Despite observations suggesting that sub-populations of cells within the islet may affect several aspects of coordinated function, discovering their presence, characterizing their intrinsic behavior, and understanding how they contribute to coordinated islet function is not well characterized. To study how functional sub-populations of cells may be distributed throughout multicellular structures such as the islet, and the role they may play in affecting global
50


function, we are limited in ways to acutely perturb function and measure resultant responses. Here we apply optogenetics by using /3-cell specific Channelrhodopsin-2 (ChR2) expression (Nagel et al., 2003) and time-dependent laser-scanning confocal activation to define the electrical regulation of sub-populations of cells within the islet. We combine this with quantitative fluorescence microscopy measurements of /3-cell function and a multicellular computational model of coupled /3-cell electrophysiology. With this, we test if islets of Langerhans contain discrete functional subpopulations and determine what effect these subpopulations have on controlling the coordinated
electrical regulation and electrical dynamics.
Results
Local ChR2 activation reveals regions of varying excitability
We first examined the spatial dependence of ChR2 activation and local membrane
depolarization on [Ca2+]i regulation across the islet. ChR2-YFP was expressed in /3-cells under Cre-recombinase control (Figure 17A), through a Pdx-Cre line which shows early expression and lacks mosaicism (Hingorani et al., 2003). ChR2-YFP expressing cells were distributed throughout the islet (Figure 19). Greater than 80% of insulin positive cells expressed ChR2-YFP, whereas less than 10% of glucagon positive cells and less than 10% of non-glucagon/insulin positive cells expressed ChR2-YFP (Figurel9). At basal (5mM) glucose levels, activating ChR2 at 451 nm with an optical power of 0.1 mW periodically across the whole islet led to coincident elevations of [Ca2+]i across 84%5% of the islet, rising immediately after excitation and subsequently decayed exponentially with a rate constant of 0.5s-1 (Figure 17B),
51


consistent with previous findings (Reinbothe etal., 2014). Activating ChR2 across subregions of 25% of the islet (quadrants) elevated [Ca2+]i across the window of ChR2 activation and extending outside of the activation window, but not across the whole islet (Figure 17C). Upon ChR2 activation across 25% of the islet varying levels of elevated [Ca2+]i were observed to extend outside of the activation window (Figure 17D); suggesting that some islet regions were more readily activated by ChR2. On average the region that provided the least [Ca2+]i stimulation activated 15% of the islet, which we refer to as ChR2 activated area, whereas the region that provided the most [Ca2+]i stimulation activated 30% of the islet. Relative to the mean ChR2 activated area for an islet, the minimum ChR2 activated area was 40% lower and the maximum ChR2 activated area 40% higher (Figure 17E). ChR2-YFP expressing cells were distributed throughout the islet (Figure 19). Activation of smaller single cell-sized regions also elevated [Ca2+]i outside of the activation window (Figure 17F), although with high variability (Figure 17G): 50% of cell-sized regions showed negligible activation of [Ca2+]l outside the window and 50% of cell-sized regions showed significant activation with a small population showing substantial activation of over 50% of the islet. Overall, the amount of ChR2 activated area increased with the size of the illumination window (Figure 17H): Illuminating 100% of the islet led to a ChR2 activated area of 84%5%; illuminating 50% of the islet led to a ChR2 activated area of 41%5%; illuminating 25% of the islet led to a ChR2 activated area of 21%4%; and illuminating single cell regions ( 5% of the islet) led to a ChR2 activated area of 16%3%.
52


A
B
Activation Window Local Activation Depolarization Global Activation
F
V ^Ni !

G
H
Time (s) Chr2 Activated Area
1.0-
Whole Semicircle Quadrant Single Ceil
Figure 17. ChR2 activation creates local [Ca2+]i elevation in pancreatic islets. (A) Model of spatial activation of ChR2 in pancreatic islets. Activation regions within the islet are defined over which ChR2 is activated by 458nm illumination. Activation of ChR2 leads to depolarization and opening of voltage gated calcium channels. (B)
An activation region defined over the whole islet (left, green dashes) generates islet wide [Ca2+]i influx, detected over the majority of the islet (left, orange). [Ca2+]i influx increases rapidly following ChR2 activation, as measured by Rhod-2 fluorescence (right). (C) An activation region defined over a quadrant of the islet (left, green dashes) generates local [Ca2+]i elevation which extends outside of the activation region (left, orange). More distant areas of the islet show no [Ca2+]i influx (right). Scale bar in B,C indicates 2% change in fluorescence. (D) ChR2 activated [Ca2+] resulting from quadrant regions of activation, presented in rank order. The ChR2-activated area where [Ca2+]i is elevated from activating ChR2 in a quadrant region (as in C) is normalized to the ChR2-activated area from activating ChR2 over the whole islet (as in B). (E) Intra-islet variation in ChR2 activated [Ca2+] relative to the islet average. The relative area of ChR2-activated [Ca2+]i over each quadrant (as in D) was expressed relative to the mean ChR2-activated [Ca2+]i of each islet, and sorted from least to most. (F) Activation of smaller single-cell regions within islets similar to (C). [Ca2+] is elevated within and outside the activation region (right). (G) Distribution of ChR2 activated area in single-cell activation regions (n=96 regions). (H) ChR2 activated area of [Ca2+]i elevation upon varying sizes of activation region. Scale bar in B,C,F indicates 2% change in fluorescence. Data in D,E averaged over n=25 islets from 6 mice; data in G averaged over n=96 regions, 26 islets from 5 mice. Data in l-K averaged over n=10 islets from 2 mice. *** indicates p<0.001. Scale bar in images represents 100)Ltm.
53


Under repeated activation of ChR2 in quadrant sub regions, those regions initially with the least ChR2 activated area continued to show on average the least ChR2 activated area over 30 minutes (Figure 18A-C). Similarly, upon sequential ChR2 acti-
vation of single cell-sized regions, those cells that initially showed negligible activation of [Ca2+] continued to show significantly lower ChR2 activated area compared to
those cells that initially showed high activation of [Ca2+] (Figure 18D).
A B c
Min Activation Max Activation
D
I U.UU I I----
Quadrant Quadrant t=0 Min T=10 Min
Min Activation Max Activation
Figure 18. ChR2 activation is consistent through time(l) Repeated activation of one quadrant at 0, 10, 20, 30 minutes shows (J) Intra-islet variation in ChR2 activated [Ca2+] as in E, at time 0 and after 30 minutes of repeated ChR2 stimulation, sorted from least to most based on time 0 measurements. (K) As in I, for minimum and maximum quadrants at 0, 10, 20, 30 minutes during ChR2 stimulation.
(L) ChR2 activated area upon single-cell activation at time 0 and after 10 minutes of repeated ChR2 stimulation, sorted based on time 0 measurements. Scale bar indicates 100/zm
The duration of light exposure influences the amount of ChR2 activated area (Figure 20). when illuminating 25% of the islet, shorter activations of 100ms resulted in [Ca2+]i elevating in less of the islet compared to the standard pulse durations of 1s as used above; whereas longer pulse durations of 10s showed no difference to the pulse durations of 1s. Despite the difference in ChR2 activated area when shorter pulses were used, there was a consistent relationship between ChR2 acti-
54


vated area with pulse length. Greater than 80% of insulin positive cells expressed ChR2-YFP, whereas less than 10% of glucagon positive cells and less than 10% of non-glucagon/insulin positive cells expressed ChR2-YFP (Figure 19B,C). These regions of varying ChR2 activated [Ca2+]i were independent of experimental parameters including stimulation order, position within the microscope field of view. Although ChR2-YFP Fluorescence was spatially variable (not shown) ChR2-YFP expression (fluorescence) did not play a role in this regional variability (Figure 20A,C,E). Regions of the islet that showed a greater ChR2-activated area under standard 1 s pulse durations also showed a greater ChR2-activated area under shorter 100ms pulse durations (Figure 20H), indicating that the intra-islet variability in ChR2-activated area is independent of the ChR2 activation protocol. Therefore within islets, at basal glucose there exist sub-regions of the islet over different spatial scales that upon depolarization can effectively recruit neighboring regions of the islet to show elevated [Ca2+]i; and this is independent of experimental protocols and occurs over several spatial scales from islet sub-regions to single cell-sized regions.
55


A
FM 4-64FX
ChR2-YFP
B
Insulin
ChR2-YFP
Merge
Insulin +
Glucagon - +
DAPI + + +
n=2474 n=308 n=185
Insulin +
Glucagon - +
DAPI + + +
n=2474 n=308 n=185
Figure 19. Quantification of ChR2-YFP Expression in islets(A) Image within intact islet of plasma membrane labeled with FM 4-64FX (left); image within intact islet of ChR2-YFP distribution (middle); together with merged image (right). Expression of ChR2-YFP is high throughout the islet. (B) Representative immunofluorescence images of insulin (top, left), YFP (top, middle), and merged image (top, right); or representative immunofluorescence images of glucagon (bottom, left), YFP (bottom, middle), and merged image (bottom, right), all in dissociated cells from the islet.
(C) Quantification of percentage insulin positive cells, glucagon positive cells or insulin and glucagon negative cells that express ChR2-YFP (left); and quantification of all cells (dapi positive) that are insulin positive, glucagon positive or are insulin and glucagon negative (right).
56


B
Quadrant Activation Temporal Order
<
o 0.2-
20 30 40 50
ChR2-YFP Fluorescence A.U.
.11
1ft
Short Medium Long
(~100ms) (~1s) (~10s)
Illumination Duration
Quadrant Activation Temporal Order
Time (s)
0.4-
ns
£ _
< fl) 0.3 o %
3 ns -
> E 0.2
{*= 0.6765 p= 0.0122
0.10 0.15
ChR2 Activated Area (Short Pulse)
Figure 20. Electrical excitability shows no temporal/spatial biasing (A)ChR2 activated area relative to the islet average sorted according to the temporal order of ChR2 quadrant activation. (B) NAD(P)H response relative to the islet average sorted according to the temporal order of ChR2 quadrant activation. (C)ChR2 activated area relative to the islet average binned according to the spatial position in the field of view of ChR2 quadrant activation. (D)NAD(P)H response relative to the islet average binned according to the spatial position in the field of view of ChR2 quadrant activation. (E) Scatterplot and linear regression (+/-95% confidence intervals) for the ChR2 activated area within an islet against the YFP fluorescence averaged across the region of activation. (F) Representative time-courses of [Ca2+]i with different durations of ChR2 activation (long=10s, medium=1s, short=100ms). (G) Quantification of ChR2 activated area as a function of ChR2 activation duration. (H) Scatterplot of ChR2 activated area in quadrants with medium pulse illumination times (1 s) vs short illumination times ( 100ms). Data in A-D averaged over n=25 islets, data in E, n=27 quadrants and G,H, n=8 quadrants.
57


Activation of non ChR2-YFP expressing islets did not lead to substantial areas where our analysis algorithm detected activation (Figure 21) indicating that only with ChR2-YFP expression do we observe significant influxes of [Ca2+] consistent with previous findings (Reinbothe et al., 2014).
Figure 21. Activation of non ChR2 expressing islets. Islets expression only PDX-Cre did not show significant areas of activation when stimulating the whole islet.
Beta-cell metabolic activity controls variations in ChR2 stimulated [Ca2+]i:
Single /3-cells dissociated from the islet are heterogeneous in the activity and dynamics of many factors underlying insulin secretion (Pipeleers et al., 1994; Pipeleers, 1992). Islet regions containing /3-cells with increased excitability could result in less ChR2-mediated depolarization required to depolarize neighboring regions and elevating [Ca2+]i. To test whether regions of altered /3-cell function affect the ChR2 activated area we measured glucose-metabolism via two-photon imaging of NAD(P)FI, alongside ChR2 activation and [Ca2+]i imaging; where heterogeneity in /3-cell glucose metabolism has been reported (Piston et al., 1999). To examine variations in intra-islet NAD(PFI) responses, we first sorted regions by ascending levels of NAD(P)FI activity such that each islet had a region of minimum and maximum metabolic activ-
o
o
0 10 20 30 40 50 60
Time (s)
58


ity. NAD(P)H levels over 2mM, 5mM, 11mM glucose were relatively consistent between islets and quadrant regions of each islet (Figure. 22A). However, there were significant variations in the NAD(P)H response from 5mM to 11mM glucose within each (Figure. 22B).
59


A
B
C
jn-
0 -r -r
Quadrant Quadrant
Min NAD(P)H Max NAD(P )H
11mM Glucose 5mlvi Glucose 2mlvl Glucose
Quadrant Quadrant
Kin NAD(P)H Max NAD(P)H
0.3-
0.3-I---1---------1--------1---------1----
Quadrant Quadrant
Kin NAD(P)H Max NAD(P)H
D
2mM Glucose
5mM Glucose
H
2m M to 5mM Response
5mM Glucose
2mMto11mM Response
J
2mM to 5mM Response
<0.1 >0.1 Activated Area
<0.1 >0.1 Activated Area
Figure 22. Metabolic Dependence on ChR2 Activation(A) NAD(P)H levels at 2mM, 5mM, 11mM glucose by quadrants, sported in ascending rank order. (B) NAD(P)H response between 5 and 11mM glucose, in quadrant regions of the islet defined by ChR2 activation regions, sorted in rank order. (C) Intra-islet variation in NAD(P)H response relative to the islet average. The NAD(P)H response over each quadrant (as in B) was expressed relative to the mean NAD(P)H response over the whole islet, and sorted from least to most. (D) ChR2 activated area relative to the islet average resulting from quadrant regions of activation at 2mM glucose, for each corresponding rank-ordered NAD(P)H level over the region at 2mM glucose. (E) ChR2 activated area relative to the islet average resulting from quadrant regions of activation at 2mM glucose, for each corresponding rank-ordered NAD(P)H response over the region between 2mM and 5mM glucose. (F) As in D for ChR2 activated area at 2mM glucose, for corresponding NAD(P)H response between 2mM and 11mM glucose. (G)
As in D for ChR2 activated area at 5mM glucose, for corresponding NAD(P)H level at 5mM glucose. (H) As in D for ChR2 activated area at 5mM glucose, for corresponding NAD(P)H response between 2mM and 5mM glucose. (I) NAD(P)H level at 5mM glucose for single cell activation regions in which ChR2 activated area was low (<0.1), high (>0.1). (J) As in H for NAD(P)H response between 2mM and 5mM glucose. Data is displayed as mean (s.e.m) Data in a-c averaged over n=25 islets from 6 mice; data in d-h averaged over n=10 islets from 3 mice; data in i,j averaged over n=96 regions, 26 islets from 5 mice. **** indicates p<0.0001 comparing experimental groups indicated. Students T-Test was used in I,J to calculate statistical significance.
60


On average within an islet, the difference between the quadrant-region of highest and lowest NAD(P)H response was 35%5% (Figure 22C). We compared quadrant-regions of NAD(P)H levels and response with ChR2 activated area at 2mM glucose and 5mM glucose. ChR2 activated area at 2mM glucose did not vary significantly with NAD(P)H levels at 2mM glucose nor did it vary with the NAD(P)H response from 2mM to 5mM glucose or from 2mM to 11 mM glucose (Figure 22D-F). ChR2 activated area at 5mM glucose did not vary with NAD(P)H levels at 5mM glucose nor with the NAD(P)H response from 2mM to 5mM glucose (Figure. 22I,J). However, it did vary significantly and substantially with the NAD(P)H response from 5mM to 11 mM glucose (Figure 23B): Regions with the lowest NAD(P)H response between 5mM and 11mM glucose had a ChR2 activated area 17% less than the islet average, whereas region of highest NAD(P)H response showing a ChR2 activated area 14% greater than the islet average. This indicates that at basal glucose spatial heterogeneity in glucose metabolism within islets leads to spatial variations in the control of electrical activity. We also tested if the link between ChR2 activated area and NADH(P)H was consistent within single cell regions. Cellular regions in which the ChR2 activated area was above the median area (10%) showed significantly greater NADP(H) response from 5mM to 11mM glucose than cellular regions in which the ChR2 activated area was below the median area (Figure 23). Thus, single cell locations showing high ChR2 activated area also had correspondingly higher NAD(P)H responses. To determine the influence of varying KATP activity, we performed similar measurements with the addition of diazoxide, a KATP activator. In the presence of diazox-
61


ide greater laser power was needed to elevate [Ca2+]i and achieve a similar ChR2-activated area. Despite this, at 5mM glucose the variation in ChR2 activated area still
varied significantly with the NAD(P)H response from 5mM to 11 mM glucose (Figure 23D), as in the absence of diazoxide.
5mM to 11mM Response
5mM Glucose+ISOjjM Diazoxide
<0.1 >0.1 Quadrant Quadrant
Chr2 Activated Area Min NAD

H Max NAD

H
5mM to 11mM Response
Figure 23. Spatial variations in metabolic activity control electrical activity. (A) Representative images of NAD(P)H autofluorescence at glucose levels indicate, or change in NAD(P)H autofluorescence between indicated glucose levels. (B) Mean (ibs.e.m.) ChR2 activated area at 5mM glucose for corresponding rank order quadrants of NAD(P)H response between 5mM and 11 mM glucose. (C) Mean (ibs.e.m) NAD(P)H response between 5mM and 11 mM glucose for single cell activation regions in which ChR2 activated area was low (<0.1), high (>0.1). (D) As in B for ChR2 activated area at 5mM glucose plus 150^M diazoxide, for corresponding NAD(P)H response between 5mM and 11mM glucose plus 150^M diazoxide. Data in B averaged over n=25 islets from 6 mice; data in C averaged over n=96 regions, 26 islets from 5 mice; data in D averaged over n=10 islets from 3 mice.* indicates p<0.05, ** indicates p<0.01. Scale bar indicates 100^m
62


Multicellular model links metabolic activity and ChR2 stimulated [Ca2+] Variations in ChR2-activated [Ca2+]i correlated with variations in metabolic activity,
over a range of spatial scales. /3-cells are intrinsically heterogeneous in excitability however the organization of this heterogeneity within the islet and how it may impact function is unknown. We asked whether specific spatial distributions of /3-cell excitability were required, or whether random distributions were sufficient to describe our experimental observations. We implemented a multicellular islet model incorporating /3-cell heterogeneity, which has previously described how the balance of gap junction coupling and KATP-regulated /3-cell excitability controls islet function (Hraha et al., 2014c). We also included a four state ChR2 current module (Nikolic et al., 2009), and represented metabolic heterogeneity by varying glucokinase activity (GK) with a distribution that matched that observed experimentally for NAD(P)H (Patterson et al., 2000). The metabolic activity of the /3-cell is determined by the flux of GK activity and thus glycolysis described by:
Jglyc = kgiyC- 2.5 * , KmATp ([^etot ~ [^e]) (14)
1 + M 1 I ATP]
Where [G] is the glucose concentration, Kgiyc is GK activity modeled with Gaussian heterogeneity (see Table 1), [ATP] is the concentration of ATP, and [Re^] and [Re] are the total pyrimidine nucleotide and NADH concentration respectively. We first divided the simulated islet into spatial partitions, each containing 200 cells with a randomly defined subset of the variation in metabolic activity (glucokinase activity) (Fig-
63


ure 24A). At sub oscillatory glucose conditions, quadrant-regions were activated by ChR2, followed by an increase to 11 mM glucose to induce [Ca2+]i oscillations (Figure 24B). Activating quadrant-regions within the simulated islet led to [Ca2+]i rises within the activation window and with significant intra-islet variation in the ChR2 activation area (Figure 24C), consistent with experimental albeit with less variability (Figure 24D). Simulated measurements of glucokinase activity over quadrant regions showed intra-islet variation that was comparable to the variation in experimentally-measured NAD(P)H, where the difference between the quadrant-region of highest and lowest glucokinase activity was 36%2% (Figure 24E). At 2mM glucose, the ChR2 activated area did not vary significantly with metabolic activity (Figure 24F), consistent with experimental measurements (Figure. 22D-F). However, at 5mM glucose the ChR2 activated area did vary significantly with metabolic activity (Figure 24G), consistent with experimental measurements (Figure 23B); where quadrant-regions with minimum glucokinase activity showed the lowest ChR2 activated area and regions with maximum glucokinase activity showed the highest ChR2 activated area.
64


800~| 5mM [G]
I-1 400'
11mM[G]
Mill
h _AH Sig.
hi
Quadrant Min Activation
Quadrant Max Activation
Time (min)
.£ E
Min GCK Rate
Max GCK Rate
F
2mM Glucose
0-3-1
V
S -0.1-
j-j -0.2-01
-0.3-1----1----------1---------1----------1----
Quadrant Quadrant
Min GK Rate Max GK Rate
01
-0.3-1---,---------,----------,---------,----
Quadrant Quadrant
Min GK Rate Max GK Rate
5mM Glucose
+10% KATP Open Conductance
<3 OJ 0.3-i 1 1
< T3 0.2- r~i
H = 0.1- i i i 1 ___
< s. eg a. c o -0.1- -0.2-
0>
Quadrant Quadrant
Min GK Rate Max GK Rate
Figure 24. Spatial domains in an islet model recapitulates ex-vivo excitability (A) False-color map showing cellular glucokinase rate over a simulated islet, where sub-regions of similar heterogeneity is applied. (B) Representative time-course of [Ca2+] in 5 cells of a simulated islet following ChR2 protocol. (C) ChR2 activated area, in terms of number of cells activated, by quadrant in rank order. [Ca2+]=170nM was used as a cutoff for active/inactive cells. (D) Intra-islet variation in ChR2 activated area relative to the simulated islet average. The relative area of ChR2-activated [Ca2+] over each quadrant was expressed relative to the mean ChR2-activated [Ca2+] of each simulated islet, and sorted from least to most. (E) Intra-islet variation in GCK rate relative to the islet average. (F) ChR2 activated area relative to the islet average at 2mM glucose, for each corresponding quadrant GCK rate. (G) ChR2 activated area relative to the islet average at 5mM glucose, for each corresponding quadrant GCK rate at 5mM glucose. (H) ChR2 activated area relative to the islet average at 5mM glucose, for each corresponding ascending rank-ordered quadrant GCK rate under the addition of 10% open probability to KATP channels in all cells.
(I) As in A, where random spatial distribution of heterogeneity is applied. (J) As in D, for simulated islets with randomly distributed heterogeneity. (K) As in G, for simulated islets with randomly distributed heterogeneity. Data in C-H,J-K averaged over n=30.
65


At 5mM glucose, with the addition of a 10% KATP channel open probability to model diazoxide application (Hraha et al., 2014c), the ChR2 activated area still varied significantly with metabolic activity (Figure 24H), again consistent with experimental measurements (Figure 23D). To test the requirement for heterogeneity to be distributed into spatial domains of excitability we distributed variations in glucokinase activity throughout the islet without any spatial organization (Figure 24I). However, there was no consistent relationship between variations in ChR2 activated [Ca2+]i and variations in glucokinase activity (Figure 24J,K). This supports that cellular heterogeneity has some spatial organization, rather than being a purely random distribution. Membrane depolarization controls [Ca2+]i elevation. We examined whether metabolic heterogeneity had a similar impact in exerting varying depolarization in neighboring cells as with exerting varying [Ca2+]i elevations (Figure 25A). Heterogeneity in membrane depolarization was observed under a similar ChR2 protocol, and this showed similar correspondence with metabolic activity. Notably regions with greater ChR2 activated area also showed a higher resting membrane potential, consistent with greater metabolic activity, ATP production and KATP channel closure ( Figure 25B-D).
66


rt 0.2-,
* -0.3-1-----------i------
Wave Origin (Quadrant)
------1-----
Wave End (Quadrant)
Figure 25. Resting Membrane Potential Links Cellular Excitability to ChR2 Activation^) Representative time courses of membrane potential from 4 cells within a simulated islet with spatially ordered metabolic activity (as in Figure 24). (B) ChR2 activated area of membrane depolarization from a, for each quadrant of the islet in ascending rank order. Vm=-40mV was used as a cutoff between active and inactive cell during ChR2 activation. (C) ChR2 activated area relative to the islet average, for each quadrant of ascending GCK rate (metabolic activity). (D) Resting membrane potential at 5mM glucose, for each corresponding ascending rank-order quadrant of ChR2 activated area. Data in B-E average over n=30 simulated islets. (E) ChR2 activated area relative to the islet average in quadrants of wave origin and wave end.
Spatial organization to NAD(P)H and ChR2 responses
In order to test whether metabolic activity is spatially organized within the islet, as
predicted by the islet model, we first calculated the absolute differences in the NAD(P)H response as a function of separation distance. For several baseline reference points within an islet, the change in NAD(P)H between 5mM and 11mM glucose was averaged over expanding areas of radius Ar (Figure 26A). The resultant A NAD(P)H response curve increased slowly as a function of radial distance (Figure 26B), indicating that areas of the islet distant from a reference point showed on average a more
67


different NAD(P)H response than areas close to the reference point. Replicate analysis of simulated islets with the two spatial distributions (Figure 24A,I) showed a similar slow increasing response curve for the spatially organized distribution, whereas a random organization showed no significant change with distance (Figure 26C). These findings were replicated by using pairwise differences in either small areas of NAD(P)H response for experimental measurements (Figure 24D,E) or cellular glu-cokinase activity for simulations (Figure 24F). The spatial scale of the experimentally measured NAD(P)H response matched the spatially ordered heterogeneity in glucose metabolism in simulated islets (Figure 26B-F), supporting the prediction that subregions of metabolic activity exist that can impact the spatial variations in excitability. To test whether there exists regions of similar excitability we computed pairwise differences in the ChR2 activated area between each single cell region within an islet (Figure 26G). These pairwise differences were sorted by the distance between the regions (Ar). Pairwise differences in closer (Ar<50pm) regions were significantly less different than more distant (Ar>50pm) regions (Figure 26H). Similar results were observed upon computing pairwise differences in the ChR2 activated area between each single cell region within the simulated islet (Figure 261). Therefore, cells in closer proximity show more similar ChR2-activaiton of [Ca2+] consistent with the islet model and indicating excitability is also spatially ordered.
68


Ar<50 pm AT>50 pm
Norm. Separation
I
Ar<0.3 £f> 0.3 Ar<0.3 #>0.3
Norm. Separation
Figure 26. Spatial analysis of NAD(P)H and ChR2-activated [Ca2+] responses(A) Representative NAD(P)H intensity image, displaying how spatial correlations in the NAD(P)H response were quantified. The ANAD(P)H response was calculated by the absolute difference was taken between an initial NAD(P)H response measurement (purple), and the NAD(P)H response averaged over an outwardly expanding region (black circles) with radius Ar. (B) ANAD(P)H response as a function of radial distance Ar normalized to the ANAD(P)H at the largest Ar. (C) Equivalent analysis as that in B calculated from GCK rates in simulated islets with random spatial distribution of GCK rate (grey) or subregions of similar GCK rates (black). (D) Representative image showing how spatial differences in NAD(P)H response were quantified. The absolute differences in NAD(P)H response were calculated for several small regions (purple circles) separated by distance Ar, in a pairwise manner. (E) Mean absolute differences in NAD(P)H response between regions as a function of spatial separation Ar normalized to the difference in response at maximal separation. (F) Equivalent analysis as that in E calculated from GCK rates in simulated islets with random spatial distribution of GCK rate or subregions of similar GCK rates. Dashed lines in B,C,E,F indicate a distance of 6 cells. (G) Representative map of ChR2-activated [Ca2+]i in an islet at 5mM glucose (orange, blue), where activation regions are applied individually over a single-cell. (H) Pairwise absolute differences in ChR2 activated [Ca2+] area averaged by close (Ar<50^m) and far (Ar>50^m) differences; and normalized to the average ChR2 activated area for each islet. (I) As in H calculated from single cell ChR2-activated [Ca2+] in simulated islets with subregions of similar or random GCK rates. Students T-Test was used in H-l.
69


Wave origin correlates with lower metabolic activity and excitability
Coordinated [Ca2+]i oscillations within islets at elevated glucose are synchronized
by propagating waves which originate in defined regions of the islet (Benninger et al., 2014). We hypothesized that the observed spatially ordered metabolic heterogeneity may control these [Ca2+]i wave dynamics at elevated glucose. To investigate how heterogeneity in /3-cell function correlates with propagating calcium waves, we measured [Ca2+]i oscillations and wave propagation alongside ChR2-activaiton and NAD(P)H measurements or Cx36 permeability measurements. Using phase analysis (see methods), propagating waves of [Ca2+]i elevation were observed that had a consistent spatial origin and propagated across the islet in 2 seconds (Figure 27A,B), consistent with prior measurements (Benninger et al., 2008, 2014). Interestingly, the ChR2 activated area at 5mM glucose in the quadrant of the wave origin was significantly less than the ChR2 activated area in the quadrant of the wave end, with a difference of 31%12% (Figure 27C). Consistent with this, the NAD(P)H response from 5mM-11mM glucose in the region of the wave origin was significantly less than the NAD(P)H response at the wave end and the whole islet average (Figure 27D). Furthermore, the [Ca2+]i oscillation amplitude at the wave origin was less than at the wave end (Figure 27E). Therefore, surprisingly, spatial heterogeneity in glucose metabolism within the islet leads to sub-regions of lower metabolic activity and excitability that appear to control the origin of propagating calcium waves.
70


A
B
a>
O' -0.4-
Wave Origin Wave End
(Quadrant) (Quadrant)
Wave End (Quadrant)
Wave Origin (Quadrant)
1
Wave End (Quadrant)
Wave Origin
i-------------------1---------------------1----------------------1
0 20 40 60
Time (sec)
------1-------------------1--------- 0J---------------------^------------------------1-----
Wave Origin Wave End Wave Origin Wave End
(Quadrant) (Quadrant) (Quadrant) (Quadrant)
Figure 27. Calcium wave origin corresponds to less excitable and metabolically active regions in ex-vivo and in-silico islets (A) Representative phase map of [Ca2+] oscillations within an islet, as calculated through Fourier analysis, which indicates a wave emerging in region of minimum phase (dark blue) and terminating in a region of maximum phase (dark red). (B) Representative time-courses from A, showing phase lag of [Ca2+] waves. (C) ChR2 activated area of [Ca2+] elevation relative to the islet average, in quadrant of wave origin and wave end. (D) NAD(P)H response relative to the islet average, in selected regions of minimum and maximum phase, as indicated in A. (E) [Ca2+] oscillation amplitude (normalized to average Rhod-2AM fluorescence) in quadrants of wave origin and wave end. (F) Representative false color map of [Ca2+] wave in simulated islet, as in A. (G) Representative time-courses from F, showing phase lag of [Ca2+] waves. (FI) ChR2 activated area in quadrants of wave origin and wave end in simulated islet, as in C. (I) GCK rate in quadrants of wave origin and wave end in simulated islet, as in D. (J) [Ca2+] oscillation amplitude in quadrants of wave origin and wave end in simulated islet, as in E. Data in C-E averaged over n=16 islets from 4 mice. Data in H-J averaged over n=30 simulated islets. Vertical scale bars indicate 2% change in fluorescence or 100nM change in [Ca2+]i in experiment or simulated islets, respectively. Scale bar indicates 100/^m
71


Intrinsic oscillatory frequency controls calcium wave propagation
The experimental observation and model agreement that calcium waves originate
in regions of low metabolic activity is not immediately intuitive. To examine the link between metabolic activity and [Ca2+]i oscillation frequency we analyzed cells in islets from Cx36_/_ mice, which lack gap junction conductance and [Ca2+]i oscillation synchronization (Ravier et al., 2005; Benninger et al., 2008) (Figure 28A). There was a significant negative correlation between cellular [Ca2+]i oscillatory frequency and NAD(P)H response (Figure 28B), where cells with higher metabolic activity had slower [Ca2+] oscillations. Upon treating Cx36-/- islets with the glucokinase inhibitor mannoheptulose we observed a significant increase in the mean oscillatory frequency of those cells that remained active from 24mHz2mHz to 32mHz4mHz (Figure 28C). Therefore decreased metabolic activity increases [Ca2+]i oscillation frequency.
72


Figure 28. Oscillatory frequency is modulated by metabolic activity(A) Representative false color map of /3-cell [Ca2+]i oscillatory frequency in a Cx36-/- islet. (B) Relationship between NAD(P)H response between 5mM and 11 mM glucose and [Ca2+] oscillation frequency of individual /3-cells within Cx36-/- islets, together with linear regression (+/-95% confidence intervals). (C) Mean (s.e.m.) [Ca2+] oscillation frequency before (-) and after (+) addition of 5mM D-mannoheptulose in individual (3-cells within Cx36-/- islets. (D) Representative time-courses of 2 uncoupled /3-cells in the islet model with high or low GCK rate. (E) Pearson correlation coefficient for the effect of heterogeneity in the indicated parameters on [Ca2+] oscillation frequency within the islet model. (F) Natural oscillation frequency of /3-cells averaged over regions corresponding to wave origin and wave end, in electrically uncoupled islets that are simulated to have subregions of similar GCK rates. Scale bar indicates 100pm.
To examine the link between metabolic activity, oscillation frequency and wave origin in the islet model we examined how the parameters used to represent cellular heterogeneity impacted the natural oscillation frequency in electrically uncoupled cells. The rate constant of glucokinase representing glucose metabolism showed the highest absolute (negative) correlation with oscillation frequency (Figure 28D,E). This is consistent with less metabolically active areas having a higher intrinsic oscillation
73


frequency, in accordance with experimental measurements. In the absence of coupling with spatially organized metabolic activity cells corresponding to the wave origin showed significantly faster oscillation frequencies than cells corresponding to the region of the wave end (Figure 28F). To examine the relationship between oscillation frequency and wave origin in a generalized fashion we used a model of Kuramoto oscillators which has been used to study the synchronization of oscillatory subunits in chemical, biological, and physical systems (32,38). We assigned unique natural oscillation frequencies to cellular units according to a Gaussian distribution and distributed these into spatial partitions of similar natural frequencies (Figure 28G). The coupled system converged frequency with spatially-dependent phase lags between oscillators (Figure 28H). The average natural oscillation frequency in regions of signal origin was significantly higher than in regions of signal end (Figure 281), indicating regions of higher natural oscillatory frequency control the direction in which signals propagate in a generalized oscillator model.
74


Cytokine Mediated Disruption of Electrical Activity
In type2 diabetes and obesity there is an up-regulation of adipocyte derived pro-
inflammatory cytokines (Boden, 2008). Previous studies have shown that pro-inflammatory cytokine cocktails disrupt [Ca2+] oscillations in pancreatic islets (Dula et al., 2010) and decrease Cx36 function (Farnsworth et al., 2015a). We tested how a pro-inflammatory cytokine cocktails (IL-1/3 (0.5ng/ml), TNF-o:(1ng/ml), INF-7(10ng/ml) disrupts spatial excitability. Islets were treated for 1 hour in the concentrations above to induce acute dysfunction (Farnsworth et al., 2015b). We first examined if coordinated oscillations were disrupted in cytokine treated islets (Figure 29A-C). Although oscillations appeared to be slightly more irregular in cytokine treated islets (Figure 29B) than the control islets (Figure 29A) we did not observe significant decreases to electrical activity or coordination with cytokine treatment (Figure 29C-D). ChR2-YFP expression was not disrupted with acute cytokine treatment (Figure 29 E) nor the ability for ChR2 to show global [Ca2+] influx upon activation (Figure 29F). Indicating that any changes to spatial excitability would not be due to ChR2 function.
75


Timers)
0 50 100 150 200
Timers)
Cytokines Cytokines
Figure 29. Acute Treatment of Pro-inflammatory on ChR2 Function(A) Top: WT islets show coordinated oscillations across the islet. After treatment for 1 hr with pro-inflammatory cytokine cocktail (bottom) we observe coordinated, but slightly dysfunctional oscillations. (B-E) Cytokine treatment does not inhibit [Ca2+] activity, electrical coordination, CHR2 expression, or ChR2 function. Scale bar indicates 100^m.
We next tested how spatial excitability was disrupted with acute cytokine treatment (Figure 30). There was a significant reduction (60%) in the metabolic NAD(P)FI response between 5mM and 11mM glucose with cytokine treatment (Figure 30A). Using quadrant ChR2 stimulation protocols where the mean activated area was 21 2%,
76


however with cytokine treatment the mean quadrant activated area decreased to 0.152% additionally the amplitude associated with ChR2 activation decreased by
368%. We examined the distribution of ChR2 activated area in single cell activation
domains (Figure 30D).
+
Cytokines
c
1.5n
I---------------------------------------1
+
Cytokines
E
5mM to 11mM Response
+
Cytokines
Single Ceil
** Untreated
Figure 30. Cytokine treatment disrupt electrical excitability(A) Metabolic response between 5mM and 11mM glucose in cytokine treated islets. (B) Activation of quadrant size regions in cytokine treated islets and amplitude (C) of ChR2 activated [Ca2+] influx. (D) Distribution of single-cell sized activation regions in untreated and cytokine treated islets.
77


In untreated islets we observe a large distribution of ChR2 activated area. Indicating some single cell regions show higher control of surrounding regions. Upon cytokine treatment there is a significant shift in this distribution and high-controlling single cell domains are no longer observed.
Next we determined the effect of acute cytokine treatment on the ability of ChR2
activation of single cell regions to entrain global islet dynamics (Figure 31).
A
C 10s Period Driving
Untreated Treated
B
O
C
O
W
£
o
3
Ll_
<
T3
O
£
10s ChR2 Activation
0 50 100 150 200 250 300
Time (s)
20s DrivingPeriod
Untreated Treated
E 30sDrivingPeriod
Untreated Treated
Figure 31. Cytokine treatment disrupt electrical excitability (A) Activation of singlecell sized regions at 11mM glucose and (B) global islet response at 10s activation intervals. (C) Entrainment of higher oscillatory frequencies in untreated and cytokine treated islets at 11mM glucose. Scale bar indicates 100^m.
At 11 mM glucose where islets show coordinated oscillatory [Ca2+]. Activation of single cell domains entrain electrical dynamics globally throughout the islet (Figure 31A-C). With Addition of cytokine treatment single cell regions lose the ability to control whole-islet electrical dynamics (Figure 31C).
78


Conclusions
Islets of Langerhans show complex multi-cellular regulation through the electrical coupling between functionally heterogeneous cells. How different sub-populations of cells affect overall function is poorly understood. To dissect the functional architecture of the islet we applied optogenetics, quantitative fluorescence microscopy and computer modeling, to detect functional sub-populations of cells cross the islet and characterize how they control overall cellular excitability and dynamics. We discovered that: 1) metabolically active sub-populations of cells exert greater control over the electrical response of neighboring cells; 2) cellular heterogeneity with respect to glucose metabolism and excitability is spatially organized; 3) metabolically less-active cells that show greater intrinsic oscillation frequency act as pacemakers to direct calcium wave propagation and synchronize oscillations; 4) modeling dysfunction
conditions associated with diabetes decreases cellular control of neighboring cells. Metabolically active sub-populations of cells control excitability
Heterogeneous /3-cell function has been observed in /3-cells dissociated from the
islet; and prior studies have (Pipeleers, 1992; Jorns et al., 1999) suggested /3-cell heterogeneity impacts islet function (Benninger et al., 2014; Kinard et al., 1999; Sher et al., 2003). Examining the role of heterogeneity in the intact islet is challenging due to the high coordination of the electrical response. However, by using ChR2 and sequential stimulation of individual cells or sub-regions, we could identify heterogeneous electrical responses in the intact islet. There exist sub-regions or subpopulations of cells that are more able to recruit surrounding cells to become active
79


(Figure 17). This control is independent of experimental parameters (Figure 20) indicating an intrinsic property of the cell that can control excitability. Furthermore, islets in which oscillations were entrained at 11 mM glucose returned to pre-entrainement activity post ChR2 activation (Figure 31B) indicating that prolonged ChR2 activation did not show major changes to function due to optical heating of cells. Notably, in the intact islet 50% of cells are unable to recruit other cells to be active at 5mM glucose, indicating 50% of cells show a low electrical response. This is similar to observations of [Ca2+]i responses in islets lacking Cx36 and electrical coupling, where 50% of cells at 5mM glucose show no [Ca2+]i response (Benninger et al., 2011b). A major question is what factors underlie this heterogeneity in the intact islet. We observed that regions of high or low electrical control are correlated respectively with regions of high or low metabolic activity. The heterogeneity in metabolic control of electrical activity has a strong effect as it persists under diazoxide application (Figure 23). This suggests heterogeneity in factors such as KATP density, whose effect would be expected to be increased relative to metabolic activity under diazoxide application, have reduced effect, which is also supported by modelling results (Figure 24). A recent study using eNpHr3, an alternative optogenetic tool to silence populations of cells, also demonstrated there exists sub-populations of cells within the intact islet that are important for maintaining electrical control (Johnston et al., 2016b). Notably these cells had elevated glucokinase levels suggesting increased metabolic activity, which is consistent with our results. However, an important difference is that our study indicates metabolically active sub-populations of cells that exert electrical control do
80


not correspond to pacemakers that initiate wave propagation (see below). Through using an NCAM marker, subsets of/3-cells (/3high and /3low) have also been identified that show differential [Ca2+]i and ATP elevation responses to glucose (Karaca et al., 2009). /3low cells expressed GLUT2 and GCK significantly lower than /3high cells, which is also similar to the less metabolically active sub-populations we observe that are unable to recruit neighboring cells to elevate [Ca2+]i. The relationship we observed at basal (5mM) glucose conditions was also predicted by our in-silico model (Figure 24F). At lower (2mM) glucose levels, we did not observe a relationship between heterogeneity in metabolic activity and electrical control both experimentally and in the islet model, suggesting that control over excitability may be dominated by
other factors, which requires further investigation.
Spatial organization of cell heterogeneity
Analysis of NAD(P)H responses and our computer model results support the existence of spatial organization, and we observed consistent results between the organization of metabolic activity and ChR2 stimulated [Ca2+] (Figure 19). We also observed regions of variability that exist over a range of spatial scales from quadrants of the islet to single cell regions (Figure 26). This is consistent with detailed examination of [Ca2+]i dynamics revealing the islet to obey Small World network principles (Stozer et al., 2013; Markovic et al., 2015); where dynamics are more similar in localized-regions of the islet. A remaining question is how these domains arise? One possibility is that cells within a domain originate from the same progenitor. Indeed, clusters of /3-cells in the islet originate from a common progenitor suggesting that
81


properties of the progenitor may affect properties of offspring /3-cells (Desgraz and Herrera, 2009). Another possibility is one of self-organization, where the excitability of one cell may impact that of other cells. Therefore, more excitable cells that originate randomly may lead to neighboring cells to become more excitable. The ability of /3-cell function and coupling to impact the organization of other /3-cells in the islet has been predicted, and differs in type2 diabetes (Striegel et al., 2015). A final possibility is that extrinsic factors such as blood flow architecture or innervation may pattern or functionally segment the islet into functional regions (Rodriguez-Diaz et al., 2011; Nyman et al., 2010). Several studies have identified sub-populations of/3-cells and their molecular basis (Katsuta et al., 2012; Karaca et al., 2009; Dorrell et al., 2016; Bader et al., 2016) which includes heterogeneity in genes underlying glucose metabolism, electrical regulation and gap junction coupling. Our findings suggest these /3-cell sub- populations will have spatial organization and be preferentially ordered together. As discussed above, low NCAM surface expression marks a subset of /3-cells (/3low) that show reduced [Ca2+]i and ATP elevation to glucose (Karaca et al., 2009). In a rat model of type2 diabetes the proportion /3low-cells increased. Similarly the Wnt effector fltp marks proliferative sub-populations of cells that preferentially expand in obesity or pregnancy, and which have reduced metabolic activity and Cx36 expression (Bader et al., 2016). The plasticity of the islet to vary between these sub-populations under pathological conditions would impact how depolarizations propagate throughout the islet to regulate [Ca2+]i and insulin secretion. Islets are electrically quiescent below 6mM glucose and transition to being electrically ac-
82


tive above this level. We predict that if less metabolically active, inexcitable cells are spatially grouped and this population grows, the presence of islet-wide depolarization will be compromised to impact insulin secretion and glucose homeostasis. Nevertheless, the precise spatial organization and the role of sub-populations in islet function and hormone release during conditions of diabetes, as well as in development and pregnancy still remains to be fully determined.
Lower metabolic, high frequency cells initiate propagating calcium waves
Synchronized [Ca2+]i oscillations within islets enhance first phase and second phase
pulsatile insulin release and insulin action (Meier et al., 2005; Stozer et al., 2013; Nunemaker et al., 2005). Propagating calcium waves mediate this synchronization and consistently emerge from sub-regions of the islet (Benninger et al., 2008). We experimentally observed that these regions of wave initiation are regions of lower metabolic activity (Figure 22) and faster natural oscillation frequency. In models of pulsed coupled oscillators, the fastest oscillator sets the pace (Mirollo and Strogatz, 1990). Prior modeling studies predicted that wave initiation occurs in regions with higher metabolic activity and faster natural oscillation frequency (Benninger et al., 2014). These studies used a model where oscillation frequency increases with increased glucose metabolism (Bertram et al., 2000). We showed experimentally in electrically uncoupled cells that a higher oscillation frequency is observed in subpopulations of cells that show a lower NAD(P)H response, and following metabolic inhibition (Figure 27). Thus lower levels of glucose metabolism that are sufficient to elevate [Ca2+] result in higher oscillation frequencies, and this is further observed in
83


our islet model (Figure 28) (Cha et al., 2011a). While Cx36 inhibitors exist, they are weak and not specific; therefore we cannot acutely decouple /3-cells to determine within the same islet if cells at the wave origin have intrinsically higher oscillatory frequency. Nevertheless, by implementing a generalized Kuramoto oscillator model (Breakspear et al., 2010) and through examining our islet model (Figure 27), we show that wave initiation is driven by spatial locations of high natural oscillatory frequency which physiologically is determined by a lower-than-average metabolic activity (Figure 28). In the islet model, the link between metabolic activity and intrinsic oscillation frequency was the strongest, further suggesting glucose metabolism may be the dominating factor in affecting [Ca2+]i oscillation frequency. Analysis of the /3-cell model has shown that KATP closure and slow CaV inactivation determines the termination of an oscillation and thus oscillation frequency, with ER Ca2+ buffering also playing a role (Cha et al., 2011a,b). KATP, CaV and SERCA conductance/activity are all ATP dependent. Thus high glucose metabolism will reduce the termination of oscillations by CaV and KATP, and increase the oscillation period and lower the frequency. The presence of intrinsic metabolic oscillations has also been suggested (Ren et al., 2013), which may enhance pacemaking action. Further work is needed to determine other dynamical effects of spatial domains. For example, in a system of Kuramoto oscillators where oscillators of similar frequency are spatially grouped, the network synchronization is optimal under weaker coupling regimes compared to systems with random spatial assignment of frequency (Freitas et al., 2015), and a recent modelling study using a similar coupled /3-cell model showed that greater
84


heterogeneity amongst/3-cells increased synchronization (Montaseri and Meyer-Hermann, 2016) and cells with greater oscillatory frequency may promote synchronization (Jalan et al., 2015). Thus islet dysfunction following disruptions to Cx36 gap junction coupling, which can occur in several conditions linked to the development of diabetes (Ravier etal., 2005; Carvalho et al., 2012b; Farnsworth et al., 2015a), may be protected by spatially grouping cells of similar excitability and oscillation frequency.
Cytokine Mediated Disruption of Electrical Activity
Acute treatment using the pro-inflammatory cytokine cocktail decreases Cx36 function in islets and prolonged treatment significantly disrupts insulin secretion and coordinated [Ca2+] activity.(Farnsworth et al., 2015b). Here we characterize what effect acute treatment has on decreasing the ability for subpopulations to control electrical activity and dynamics within islets. As results from Figure 29 indicate we have no significantly disrupted coordinated [Ca2+] oscillations nor the ability for islets to be globally activated with ChR2 stimulation. However, we do observe signifcant decreases to electrical excitability (Figure 30A) and a corresponding decrease to quandrant activated area (Figure 30B) indicating that the loss of a metabolic response may be increasing the hyperpolarized state and thus limiting the extent to which ChR2 stimulated action potentials may propagate. These results are similarly confirmed in a high gluocse state where [Ca2+] oscillations are present (Figure 31). While [Ca2+] coordination has not been significantly disrupted in cytokine treated islets, small regions are not longer capable of entraining oscillatory behavior (Figure 31C) indicating that
85


even in an acutely dysfunctional state there is a loss of cellular control over islet-wide dynamics.
86


CHAPTER V
AGE ASSOCIATED DECLINE IN ELECTRICAL SYNCHRONIZATION 3 Introduction
There is a significant decline in glucose tolerance and increase in the risk of type2 diabetes with advancing age in humans (Karve and Hayward, 2010a). Reduced glucose tolerance in ageing results from both reduced insulin sensitivity and reduced insulin secretion in response to elevated glucose (Nathan et al., 2007; Gumbiner et al., 1989; Basu etal., 2003; Chang and Halter, 2003). /3-cell proliferation and proliferative capacity declines with age, which can partly explain the reduced insulin secretion and risk for type2 diabetes (Kushner, 2013; Gunasekaran and Gannon, 2011). However /3-cell function is also altered upon ageing (lozzo et al., 1999; Ihm et al., 2006). /3-cells respond to glucose by elevated mitochondrial respiration and ATP generation; membrane depolarization and action potential generation; increased intracellular free-calcium ([Ca2+]); and insulin granule exocytosis. Several of these steps, including mitochondrial respiration (Gregg et al., 2016; Helman et al., 2016; Li et al., 2014) ATP generation (Gregg et al., 2016; Helman et al., 2016), and [Ca2+] handling (Gregg et al., 2016; Li et al., 2014; Avrahami et al., 2015) are altered in /3-cells from aged mice and humans. However, there have been conflicting results. For example
in human islets, improved insulin secretion upon ageing or in senescent /3-cells has
3The results in this chapter were accepted to publication on 6/1/2017: Westacott, M.J., Farnsworth, N.L., St. Clair, J.R., Poffenberger, G., Heintz, A., Hart., N.J., Powers, A.C., Benninger, R.K.P, Age-dependent decline in the coordinated Ca2+ and insulin secretory dynamics in human pancreatic islets. Diabetes
87


been reported (Helman et al., 2016), yet others have reported declines in insulin secretion with age (Ihm et al., 2006; Gregg et al., 2016; Li et al., 2014; Fritsche et al., 2002). Differences between mouse and human /3-cells responses to ageing have also been reported (Gregg et al., 2016; Avrahami et al., 2015). Therefore much remains to be understood regarding how/3-cell function is altered upon ageing. /3-cells within the islets of Langerhans do not function autonomously. There is extensive communication between /3-cells and with other cell types (Benninger et al., 2011 b; Konstantinova et al., 2007; van der Meulen etal., 2015; Lernmark, 1974; Hashimoto et al., 2014; Rodriguez-Diaz et al., 2012) that is important for how/3-cells function within the islet. Gap junctions formed from connexin36 (Cx36) electrically couple (3-cells (Ravier et al., 2005; Benninger et al., 2008; Serre-Beinier et al., 2009) which coordinates the oscillatory [Ca2+] response to elevated glucose and regulates the dynamics of insulin secretion (Ravier et al., 2005; Benninger et al., 2008). This includes enhancing the first phase and second phase pulses, where a loss of Cx36 gap junction coupling leads to glucose intolerance (Head et al., 2012b). Notably gap junction coupling and coordinated [Ca2+] dynamics are disrupted in models of obesity or type2 diabetes (Carvalho et al., 2012b; Hodson et al., 2013; Farnsworth et al., 2015a; Ravier et al., 2002) suggesting a role in the pathogenesis of diabetes (Rutter and Hodson, 2013). However, the effect of ageing on intra-islet communication and the regulation of insulin secretion has been poorly examined. This includes examining how gap junction coupling and coordinated [Ca2+] may be affected by ageing. Furthermore, the role of gap junction coupling and coordinated [Ca2+] in human islet
88


Full Text

PAGE 1

REGULATIONOFCOUPLED -CELLELECTRICALDYNAMICS by MATTHEWJOELWESTACOTT B.S.,ColoradoSchoolofMines,2010 M.S.,UniversityofDenver,2012 Athesissubmittedtothe FacultyoftheGraduateSchoolofthe UniversityofColoradoinpartialfulllment oftherequirementsforthedegreeof DoctorofPhilosophy BioengineeringProgram 2017

PAGE 2

ThisthesisfortheDoctorofPhilosophydegreeby MatthewJoelWestacott hasbeenapprovedforthe BioengineeringProgram by EmilyGibson,Chair RichardBenninger,Advisor DiegoRestrepo TimLei Date:July29,2017 ii

PAGE 3

Westacott,MatthewJoelPh.D.,Bioengineering RegulationofCoupled -CellElectricalDynamics ThesisdirectedbyAssistantProfessorRichardBenninger ABSTRACT Pancreatic -cellsaretheonlycelltyperesponsibleforsecretinginsulin;ahormonenecessaryformaintainingglucosehomeostasis. -cellsareanaturallyheterogeneousinexpressionofgenesregulatingglucosemetabolismandelectrophysiology;leadingtoapopulationofcellswithuniqueelectricaldynamics.Electricalcommunicationbetween -cellsthroughconnexin36gapjunctionsacttosynchronize electricalactivityandcoordinateinsulinsecretiondynamicstopromoteefcacyof glucoseclearing.Inmodelsofdiabetesthereisdisruptiontothecoordinatedelectricaldynamicswithinpancreaticislets.However,whatrolefunctionalsubpopulations playinmediatingthisdisruptionortheextenttowhichthisdisruptionoccursinhumanisletsispoorlyunderstood. TheobjectivesofthisthesisistodeterminedevelopacomputationalmulticellularmodelofpancreaticisletelectrophysiologyandtestifitcanrecapitulatetwoaspectsofCx36mediatedelectricaldysfunctiondeterminewhatrolesubpopulationsplayincontrollingelectricalactivityandcoordinationwithinpancreaticislets testifageandtype2diabetesinhumanscorrelatewithdecreasestoelectricalcoordinationwithinislets.Torecapitulate in-vivo modelsofpancreaticisletdysfunction weexpandedonanestablishedsingle-cellularcomputationalmodelof -cellelectroiii

PAGE 4

physiologybygeneratingacoupled -cellnetworkwithisletarchitecture.Ifoundthe modelwasabletocorrectlypredictcriticallossinelectricalactivityandcoordination; mediatedbytheelectricalcouplingbetween -cells.TodeterminetheroleoffunctionalsubpopulationsIcreatedatransgenicmousemodelwith -cellspecicexpressionofChannelrhodopsin-2andimplementedaprotocolofspatiotemporal[Ca 2+ ] activationcombinedwithtwo-photonimagingofmetabolicactivitywhichshowedsubpopulationsofcellswerespatiallyorientatedandshowedpreferentialcontrolover electricalactivityanddynamicswithinislets.Lastly,wetestedthehypothesisthatadvancedageandtype2diabetescorrelatewithdecreasestoelectricalcoordination withinhumanpancreaticisletsthroughimagingof[Ca 2+ ]activityanddevelopinga novelimageanalysisalgorithmtoquantifyelectricalcoordination.Ifoundthatboth advancedageandhistoryoftype2diabetescorrelatedwithsignicantdecreasesto electricalactivityandcoordinationwithinhumanislets.Furthermore,wefoundthat aCx36activatorcouldrestoreelectricalcoordinationwithinagedandtype2diabetic humanislets. Theresultspresentedhereinprovideinsightintoregulatingfactorsofcoordinated electricaldynamicsbetween -cellsandshowpotentialforanoveltherapeutictorecoverpancreatic -cellfunction. Theformandcontentofthisabstractareapproved.Irecommenditspublication. Approved:RichardBenninger iv

PAGE 5

ACKNOWLEDGMENTS ThisworkwassupportedbyNIHGrantsR01DK102950,R01DK106412;andJuvenileDiabetesResearchFoundationGrant5-CDA-2014-198-A-NtoR.K.P.B., andF31DK107043toM.J.W..Microscopywasperformedthroughtheuseofthe UniversityofColoradoAnschutzMedicalCampusAdvancedLightMicroscopyCore P30NS048154,UL1TR001082,isletisolationwereperformedintheBarbaraDavis CenterIsletCoreP30DK057516underIACUCprotocalB-95814D,andsimulationswereperformedontheJANUSsupercomputerattheUniversityofColorado throughsupportbytheNSFCNS-08217944.IsletperifusionexperimentswereperformedthroughtheIsletProcurementandAnalysiscoreoftheVanderbiltDiabetes ResearchandTrainingCenterDK20593intheAlvinPowersLabVanderbilt.We thankMatthewMerrinsUniversityofWisconsinandLauraPyleUniversityofColoradofordiscussionandstatisticaladvice. v

PAGE 6

TABLEOFCONTENTS CHAPTER I.BACKGROUNDANDINTRODUCTION....................1 GlucoseControlandDiabetes........................1 PancreaticIslets...............................3 ElectrophysiologyandStimulusSecretionCouplingof -Cells........4 -Cellsaselectrochemicaloscillators....................11 Electricalcouplingbetween -cells......................12 Heterogeneityof -CellsandRelevanceDiabetes..............15 II.MATERIALSANDMETHODS.........................18 MouseModels................................18 IsletCulturing.................................18 MicroscopyandCalciumImaging.......................18 PsuedoisletAggregation...........................20 [Ca 2+ ]Coordination.............................20 ChR2ActivationAnalysis...........................21 FRAP.....................................22 Immunouoresence.............................22 InsulinELISA.................................23 ClassicationLearners............................25 III.DEVELOPMENTOFCOMPUTATIONALMODEL..............26 vi

PAGE 7

Introduction..................................26 ModelDescription..............................30 ModelVerication...............................33 Conclusions.................................45 IV.SPATIALLYORGANIZEDSUB-POPULATIONSOFCELLSCONTROLELECTRICALACTIVITYANDDYNAMICSACROSSTHEISLETOFLANGERHANS.....................................49 Introduction..................................49 Results....................................51 LocalChR2activationrevealsregionsofvaryingexcitability......51 Beta-cellmetabolicactivitycontrolsvariationsinChR2stimulated[Ca 2+ ]i:58 MulticellularmodellinksmetabolicactivityandChR2stimulated[Ca 2+ ].63 SpatialorganizationtoNADPHandChR2responses.........67 Waveorigincorrelateswithlowermetabolicactivityandexcitability...70 Intrinsicoscillatoryfrequencycontrolscalciumwavepropagation....72 CytokineMediatedDisruptionofElectricalActivity...........75 Conclusions.................................79 Metabolicallyactivesub-populationsofcellscontrolexcitability.....79 Spatialorganizationofcellheterogeneity................81 Lowermetabolic,highfrequencycellsinitiatepropagatingcalciumwaves83 CytokineMediatedDisruptionofElectricalActivity...........85 vii

PAGE 8

V.AGEASSOCIATEDDECLINEINELECTRICALSYNCHRONIZATION....87 Introduction..................................87 Results....................................89 QuanticationofElectricalActivityandCoordination..........89 AgePredictsDeclinein[Ca 2+ ]CoordinationinHumanIslets......92 Age-dependentdeclineininsulinsecretion...............95 Cx36gapjunctionactivationrecoversage-dependent[Ca 2+ ]decline..101 [Ca 2+ ]andCx36gapjunctionfunctionispreservedinagedmice....103 Conclusions.................................105 VI.CA2+METRICSIDENTIFYTYPE2DIABETICHUMANDONORS......110 Introduction..................................110 Results....................................111 GeneratingClassierstoPredictType2Diabetes............111 ModanilIncreasesElectricalCoordinationinHumanType2Islets...116 Conclusions.................................118 VII.DISCUSSIONANDFUTUREDIRECTIONS.................119 FunctionalRoleofSubopulations.......................119 LossofCa2+CoordinationwithAdvancedAgeandType2Diabetes.....121 ConcludingRemarks.............................124 BIBLIOGRAPHY.................................126 viii

PAGE 9

LISTOFTABLES TABLE 1.ListofParametersusedinCha-NomaModel..............31 2.Parametersusedtogenerateclassicationmodelofdiabetesincidence111 3.Summaryofperformanceforvaryingclassicationlearners.......113 ix

PAGE 10

LISTOFFIGURES FIGURE 1.CircuitDiagramofExcitableCell....................6 2.MetabolismofglucosegeneratesATPandNADH............8 3.Glucosestimulatedinsulinsecretion..................10 4.CalciumandInsulinShowIn-PhaseOscillations............12 5.Cx36Couples -cells..........................14 6.Hodgkin-HuxleyModel.........................27 7.DiagramofCha-NomaModel......................29 8.GlucoseActivatedCa2+Oscillationsinsimulatedislet.........33 9.PenetranceofKir6.2 [ N 30 ;K 185 Q ] -GFP.................34 10.Phasetransitioninexcitability......................36 11.LinkbetweenphasetransitionsinCa2+andphysiologicalparameters.37 12.Coupledoscillatormodeldescribesexperimentalphasetransitions...39 13.Mean-eldtheoryanaologof -cellnetworkactivity...........40 14.Modelofsynchronizationdisruption...................42 15.Lossofelectricalsynchronizationwithmicroparticleincorporation...43 16.Lossofelectricalsynchronization in-silico ................44 17.ChR2activationcreateslocalCa2+elevationinpancreaticislets....53 18.ChR2activationisconsistentthroughtime...............54 19.QuanticationofChR2-YFPExpressioninDisassociatedIslets.....56 20.Electricalexcitabilityshowsnotemporal/spatialbiasing........57 x

PAGE 11

21.ActivationofnonChR2expressingislets................58 22.MetabolicDependenceonChR2Activation...............60 23.Spatialvariationsinmetabolicactivitycontrolelectricalactivity.....62 24.Spatialdomainsinanisletmodelrecapitulatesex-vivoexcitability...65 25.RestingMembranePotentialLinksCellularExcitabilitytoChR2Activation67 26.SpatialanalysisofNADPHandChR2-activatedCa2+responses...69 27.Correlationbetweenwaveoriginandmetabolicactivity.........71 28.OscillatoryFrequency..........................73 29.AcuteTreatmentofPro-inammatoryonChR2Function........76 30.CytokineTreatmentDisruptsExcitability.................77 31.CytokineTreatmentDisruptsExctiabilityatHighGlucose........78 32.HumanIsletElectricalActivity......................91 33.AgePredictsDeclineinHumanCoordination..............93 34.FurthercharacterizationofCa2+dynamicsinhumanisletswithage..94 35.Age-dependentdeclineinCx36gapjunctioncouplinginhumanislets.96 36.Age-dependentdeclineininsulinsecretiondynamics..........98 37.AgeDependentDeclinetoInsulinSecretion..............99 38.[Ca2+ActivityPredictsDeclineinInsulinSecretion...........100 39.Cx36ActivatorsRescueCa2+Function.................102 40.PreservationofCx36gapjunctionfunctionandCa2+coordinationin agedmice................................104 41.BMIbinaryclasiscationofDiabetes..................112 xi

PAGE 12

42.LinearClassicationPerformance....................114 43.Singlepredictorperformance......................114 44.LossofelectricalcoordinationinT2islets................116 45.ModanilTreatmentofHumanType2Islets...............117 xii

PAGE 13

CHAPTERI BACKGROUNDANDINTRODUCTION GlucoseControlandDiabetes Glucosecontrolisonesystemwhichistightlyregulatedbyanegativefeedback looptoprovidethenecessaryenergyforallotherfunctionswithinthebody.Complexsugarsincarbohydratesarebrokendownintothesimplesugarglucosewhich istransportedintothecirculatorysystem.Ascirculatingglucoselevelsriseafeedbacksignalisneededtobringglucoselevelsbackdownbypromotingperipheral tissuetoincreaseglucoseuptakeandinturnuseitforenergy.Pancreatic -cells aretheonlycelltypewhichproduceandsecreteinsulin;ahormonenecessaryfor maintainingglucosehomeostasisbyinducinguptakeofcirculatingbloodglucose toperipheraltissue,byupregulatatingglyconeogenisis,andbysuppressinggluconeogensisBano,2013andinhealthyindividuals,glucoselevelsarekeptwithina narrowrangebetween60mg/dlto150mg/dl.Insulinactsonperipheraltissuethrough actingontheinsulinreceptortriggeringacascadeviathetyrosinekinasepathway toincreasetransportationofglucosetransportersGLUTtotheplasmamembrane; increasingthenetuxofglucosefromthecirculatorysystemtosurroundingtissue. GlucoseistransportedintocellswhereitentersthemetabolicpathwaytocreateATP tobeusedastheprimaryenergysourceinmuscularandadiposetissueorbestored andglycogenintheliver. 1

PAGE 14

Diabetesischaracterizedbyastateofchronichyperglycemia,excessivehighblood glucoselevels,duetoafailureof -cellstosecretetheappropriatelevelsofinsulin eitherduetoautoimmunediseaseanddestructionof -cellstype1orbyinsulin resistanceand -celldysfunctiontype2.Thiscanbecharacterizedwhenfastingglucoselevelsareabove126mg/dlmMwhennon-diabeticlevelsarebelow 108mg/dlmMorifa2hourpost-pranadialisabove200mg/dlmMwhennondiabeticlevelsarebelow140mg/dl.8mM.Thelossofproperglucosecontrolleads tonumerouscomplicationsthroughdamagestonervousandcirculatorysystemsas hyperglycemiainprolongeddurationistoxictocellsKawahito,2009.Ifleftunchecked thisdamagewillleadtocardiovasculardisease,strokes,neuropathy,andretinopathy.Lifeexpectancyoftype1diabeticsisestimatedtobe20yearsyoungerthan healthyindividualsanddiabetesisratedasthe7thmostcommonscauseofdeath intheUnitedStates.Whileratesformostdiseasesaredecreasingwithtime,diabetestype1Davis,2001andtype2areincreasingatanalarmingrate-showing 61%incidencebetween1990and2001Nathanetal.,2007.Approximately85%of diabeticsaretype2-andnearly30millionAmericansand410millionpeopleworldwidehavesomeformofdiabetes-arisingfrom108millionin1980withthemajority ofnewcasescomingfromAsiaandAfricaOrganizationetal.,2016.Typicallyin type1diabetesthereisamulti-yearprogressionfromaninitialeventtotriggerautoimmunityuntillossofsufcient -cellmassleadstohyperglycemia.Afterdiagnosis thereisusuallylittlepharmaceuticalinterventionavailableandinsulininjectionwillbe requiredforlife.Intype2diabetesthereareremainingfunctional -cellswhichare 2

PAGE 15

unabletomeetrequiredinsulindemandscausedbyacombinationofinsulinresistanceand celldysfunction.Pharmaceuticalinterventionisapreferredtreatment withtype2diabetesalongwithdietandexercisechanges.Themostcommonpharmacologicalagentswilleithertargetglucosehandlingwithintheliverandperipheral tissueordirectlytargetpancreatic -cellstopotentiateinsulinsecretionbytargeting specicchannelsorreceptors.Thereisgrowinginvestigationintopharmacological agentswhichtargetalternativepathwayswithin -cellsincludingtargetingglucose transportersinthekidneysandinsulinsensitizesWagmanandNuss,2001.However,additionaltheraputicsareneededtoprovidehigherefcacyofglucosecontrol. PancreaticIslets -Cellsarenotrandomlydistributedthroughoutthepancreas,rathertheyexistin discretemicro-organsknownaspancreaticisletsisletsofLangerhans.Pancreatic isletsaresmallhundredsofmicronsmulticellularsystemscomposedof5endocrine celltypesheldinpseudosphericalstructurescontainingafewhundredtoseveral thousandcells.Theothercelltypes, ,pp-cellssecretehormoneswhichaidin glucosehomeostasisbyeithercounteractingorenhancingtheeffectofinsulinsecretion. -cellscomprisethemajorityofcellswithinislets,typicallybetween50%-80% withvariabilitybetweenspeciesCabreraetal.,2006;Kilimniketal.,2012;Stefan etal.,1982. Alongwith -cellcontentthecytoarchitectureoftheislet,howcellsarearranged,is differentbetweenspecies.Inmice, -cellsoccupythecoreoftheisletandothercell typesarelocatedontheperiphery.Inhumans, -cellsarelocatedpreferentiallynext 3

PAGE 16

toeachotherhoweverthearrangementappearstobemorestochasticinnaturewith -cell-cellcontactprobabilitiesthatareisletsizedependentCabreraetal.,2006. Thesecytoarchitecrualdifferencesarethoughttocontributetodistinctelectricaldynamicsbetweenspecies,whicharewellcharacterizedinmicebutlesssoinhumans. Studiessuggestthatthecytoarchitectureisagedependent-andthatisletsinyoung humansismoresimilartothatofmiceGreggetal.,2012;logicallyonewouldexpecttheelectricaldynamicsofisletsinyoungerhumanstobesimilartothatofmice howeverthishasnotbeenfullyquantied.Thecombinedmasspancreaticisletsis typicallyrepresents1-2%ofthepancreasvolumeIonescu-Tirgovisteetal.,2015 however,isvariabledependingonmetabolicdemandwhichisincreasedwithobesityMoriokaandKulkarni,2010andpregnancyRieckandKaestner,2010and afailureof -cellexpansionisassociatedwithtype2diabetes.Recentworkalso suggeststhecytoarchitecureofhumanisletschangeswithtype2diabetesMorioka andKulkarni,2010-withdecreasedlevelsof -cell-cellcontactsintype2diabetic isletsKilimniketal.,2011. ElectrophysiologyandStimulusSecretionCouplingof -Cells Pancreatic -cellsareasubsetofelectricallyexcitablecellswithinthepancreatic isletofLangerhansandactonbloodglucoselevelstocontrolregulatedinsulinsecretionRorsman,1997.TheintracellularconcentrationofionsincludingCa 2+ Na + ,Cl )]TJ/F36 11.9552 Tf 11.415 -4.339 Td [(arelowerthantheextracellularspaceandintracellularconcentrationof K + isheldatahigherconcentrationthatextracellularin -cells.Theseparationof chargesacrossthecellularmembranecreatesanelectricaleldormembranepo4

PAGE 17

tentialwhichisdescribedbytheNernstpotential: E I = RT ZF ln I o I i whereRrepresentsthegasconstant,FrepresentsFarady'sconstant,andZrepresentsthechargeoftheion,I.Atstandardconditionsthemembranepotentialisthen proportionalthethenaturallogoftheratioofionsoutsidetoinsideofthecell.For -cellstheNernstpotentialisapproximately+40mV,+40mV,-35mV,-75mVforCa 2+ ,Na + ,Cl )]TJ/F36 11.9552 Tf 8.092 -4.339 Td [(,andK + respectivelyDrewsetal.,2013;AshcroftandRorsman,1989. AnalogoustoasimpleRCcircuit,theseparationofchargescreatesacrossaboundaryisactsasacapacitorandtheresistorsarechannelswhichareselectivetoeither Ca 2+ ,Na + ,Cl )]TJ/F36 11.9552 Tf 8.092 -4.339 Td [(,andK + .Ifionsareallowedtoowacrosstheplasmamembrane weinterpretthisasanincreaseinconductance,oradecreaseinresistance,tochannelsspecictothation. 5

PAGE 18

Figure1. CircuitDiagramofExcitableCellExcitablecellsas -cellshavean analagousrepresentationassimpleRCcirucitswherethecellmembranerepresentsthecapacitorC m ,thechannelsrepresentresistorsG Na;K;Ca ,andthevoltage sourcesasthereversepotentialforeachionE Na;Ca;K Themembranepotentialbecomesdependentonthelevelofcurrentthatowsacross theplasmamembrane: dV m dt = )]TJ/F18 11.9552 Tf 10.693 0 Td [( I K + I Ca + I Na + I Cl + I ::: =Cm wherethecurrentofaspecicionisproportionalto: I i = g i V m )]TJ/F21 11.9552 Tf 13.748 0 Td [(E i 6

PAGE 19

where g i istheconductance,orinverseresistance,ofachannelspecictoioni,and V m and E i asthemembranepotentialandNernstpotentialforioni.Aspecicion willcontinuetomoveacrossthemembraneuntilthemembranepotentialreaches thespecicNernstpotential.Afterwhichtheowofionswillstopandreverseifthe membranepotentialcontinuestogobeyondtheNernstpotential.Alternativelythe conductanceofthechannelcanbereducedto0S,i.e.channelclosure,whichwill ceasecurrentowofthation.Theconductanceofachannelcanberegulatedbythe membranepotentialitself,voltagegatedchannels,oritcanberegulatedbybinding ofcertainmolecules,nucleotides,ligands,light,andotherstimuli. Thehallmarkofexcitablecellsistousecurrentowtoaccomplishatask;inmuscletissue,includingtheheart,currentowof[Ca 2+ ]controlscontraction/elongation ofmuscleberstoperformmechanicalwork.In -cells,asglucoselevelselevateit istakenupby -cellsthroughGLUT-2transportersmiceaswellasGLUT-1humans.Glucoseundergoesphosphorolationbyglucokinase;aclassofhexokinases withK d of6mM 100mg/dl,makingitmoresensitivetosubtlechangestobloodglucosevaluesthantypicalhexokinasesIynedjian,2008.Phosphorolationofglucose toglucose-6-phosphateistherststepinthemetabolicpathwaywhichresultsinproductionofATPandelevationoftheATPtoADPratio.Briey,glucose-6-phosphate isconvertedtofructose-6-phosphateF6Pbyglucosephosphateisomerase.F6Pis phosphorolatedbyphosphofructokinasetocreatefructose-1,6-bisphosphateF1,6BP. Twoadditionalstepsarerequiredwhichcreatesglyceraldehyde3-phosphateGADP whichisnalmoleculebeforethepayoffphasewhereanetgainofATPandNicoti7

PAGE 20

Figure2. MetabolismofglucosegeneratesATPandNADH.Representative metabolicpathwayinthe -cell.Glucoseistakeninthroughglucosetransporters, phorosphorlatedbyglucokinase,andentersthemetabolicpathwaytogenerateATP andNADH. namindeadeninedinucleotideNADPHmoleculesarecreatedandsubsequently resultinthecreationofpyruvate.PyruvateisconvertedthroughtheoxidativedecarboxylationpathwaytocreateacetylCoAwhichenterstheKrebscycletocreate additionalATPandNADPHFigure2. ATPsensitivepotassiumchannelsK ATP containtheKir6.2bindingsitewhich bindsATPandactstoclosethechannelCraigetal.,2008.Channelclosuretriggersasmalldepolarizingcurrentwhichbringsthemembranepotentialfrom-60mV 8

PAGE 21

upto-45mVtotriggeractivationofvoltagegatedcalciumchannelsVDCCsand sodiumchannels.Thistriggersalargermembranedepolarizationandinducinglarge increaseincytosoliccalciumlevels[Ca 2+ ] i Drewsetal.,2013.Therisein[Ca 2+ ] i isnecessarytotriggerinsulinsecretoryvesiclestofusewiththeplasmamembrane andreleaseinsulinintotheperipheraltissuesandbloodstreamKasaietal.,2014 andalsohelpstoregulatetheproductionofinsulinthrough[Ca 2+ ] i dependenttranscriptionpathwaysLawrence,2001.Thispathwayisreferredtoasstimulussecretioncouplingin -cellswhereanincreaseinglucosestimulusiscoupledtotheactivityofthemetabolicpathwayandtoregulationofK ATP channels. 9

PAGE 22

Figure3. Glucosestimulatedinsulinsecretion.UptakeofglucosefromGLUTchannelsentersthemetabolicpathwaytogenerateATP-causingKATPchannelclosure andmembranedepolarization.AdaptedfromJensenetal.,2008 Defectstothestimulussecretioncouplingpathwaycanincreasetheriskordirectly causetype2diabetesincludingK ATP channeldysfunctionwhichlinkedtoincreased riskoftype2diabetesGloynetal.,2003andremovalofKATPfunctionwillcause overtdiabetesneonatalKosteretal.,2002byinterferingwithprimaryinitiatorof Ca 2+ elevation.SimilarlydefectstoglucokinaseactivityMODYIIwillpreventsufcientphoshorloationofglucoseandlimitATPproductionpreventingK ATP channel 10

PAGE 23

closure.Neonataldiabetescaneitherresolveintime,orinhalfofthecasesleadto permanentneonataldiabetesDeLe onandStanley,2016. -Cellsaselectrochemicaloscillators [Ca 2+ ] i levelsarenotcontinuallyelevatedinside -cellsafterincreasesinglucose andsubsequentactivationofVDCCs.Indeed,chronicelevationof[Ca 2+ ] i cantriggercellularapoptosispathwaysZhouetal.,2015.Duetorectifyingpotassiumcurrents,depletionofATP,andthemembranewillbehyperpolarized-and[Ca 2+ ] i will bebroughtbackdowntorestinglevels100nMbyuptakeof[Ca 2+ ] i totheendoplasmicreticulumE.R.andviapumpmechanismstoshuttle[Ca 2+ ] i totheoutsideof thecell.DuetotheclosureofKATPchannelsaslongasATPlevelsaresufciently high,themembranepotentialwillriseagaintoVDCCactivationrangeandtheprocesswillbeginagain.[Ca 2+ ] i andinsulinsecretionareoscillatoryinnatureandshow inphaseoscillationsinisolatedisletsRavieretal.,2005whichscaleproportionatelytothefrequencyof[Ca 2+ ]oscillationNunemakeretal.,2009.Thestudyof [Ca 2+ ] i dynamicswithin -cellsisthensurrogatemetricforinsulinsecretionactivity anddynamics. 11

PAGE 24

Figure4. CalciumandInsulinShowIn-PhaseOscillations.IntheabsenseofCx36 insulinand[Ca 2+ ] i oscillationsareabolished.AdaptedfromRavieretal.,2005 Persistentelevationofglucose,above6mM,willcause -cellstoshowelectrical oscillationsuntilextra-cellularglucoseisbroughtdownto5mMFigure4. Electricalcouplingbetween -cells Insulinsecretionisregulatedbyextensiveextra-cellularNewsholmeetal.,2010; Nymanetal.,2010;BallianandBrunicardi,2007;Rodriguez-DiazandCaicedo,2014 andintracellularcell-cellcommunicationwithinpancreaticisletsJainandLammert, 2009.Paracrinesignalingofglucagon,insulin'scounterhormone,willpotentiate insulinsecretionCurry,1970whereassignalingofsomatosatatinfrom -cellswill attenuateinsulinsecretionStrowski,2000.Similarly,sympatheticandparasym12

PAGE 25

patheticnervoussysteminnervationofpancreaticisletscanleadregulationofinsulinsecretionThorens,2014. -cellsarecoupledwithConnexin36gapjunctions Cx36-acationicspecicchannelwhichallowsthepassageofcationsandsmall chargedmoleculesbetweenadjacentcells.Cx36formplaquesat -celljunctions tofacilitatehighconductancebetweenadjacentcells.WhilesinglechannelconductanceofCx36hasbeenestimatedtobe5pSMoreno,2004thecumulativeconductancebetween -cellshasbeenmeasuredtobe100-300pSBenningeretal., 2008;Moreno,2004.Atelevatedglucoselevels -cellsexhibitoscillatoryelectrical activitywhichissynchronizedviaCx36Nlendetal.,2006.AlthoughisletcytoarchitectureplaysaroleinthelevelofsynchronizationHrahaetal.,2014ainmice, where -cellsshowhighco-localizationintheinterioroftheislet,Cx36promotesthe majorityof -cellstoelectricallyoscillateinphaseBenningeretal.,2008.Insulin oscillationsinisolatedmouseisletsshowinphaseoscillationswithelectrical/[Ca 2+ ] i oscillationsandsimilarlyplasmainsulinlevelsoscillateoverseveralminutes. 13

PAGE 26

Figure5. Cx36Couples -Cells.ModelofCx36functioninpancreaticislets.Glucose [G]entersthecellthroughglucosetransporterswhichentersthemetabolicpathway toincreasetheATP/ADPratio.TheelevatedATP/ADPratiotriggersclosureofATP sensitivepotasiummchannelsK ATP whichtriggersaslightmembranedepolarization toactivatevoltagegatedcalciumchannelsVDCCs.[Ca 2+ ]entersanindividualcell fromVDCCs,travelsthroughCx36channelstriggeringdepolarizingcurrentswhich activateVDCCsinneighboringcells.CouplingthroughCx36promotesthesynchronizationtheheterogeneousoscillationsdifferentcolorcellsintoonecosistentfrequencyandphase. DisruptiontoCx36viaknockoutleadstoalossofcoordinatedelectricaloscillations andinsulinoscillationsRavieretal.,2005.Similarly,plasmainsulinoscillationsare abolishedinCx36 )]TJ/F22 7.9701 Tf 7.593 0 Td [(= )]TJ/F36 11.9552 Tf 11.415 -4.338 Td [(miceleadingtoglucoseintolerance,apre-diabeticcondition characterizedbyimpairedglucose.UponglucosechallengeCx36 )]TJ/F22 7.9701 Tf 7.594 0 Td [(= )]TJ/F36 11.9552 Tf 11.415 -4.339 Td [(miceexhibitedhighertotalglucoselevelsoveratwohourperiodandhadsignicantlyreduced 14

PAGE 27

rstphaseinsulinsecretion,butsimilarsecondphaseinsulinsecretion.Headetal., 2012bInDb/Dbmice,amousemodeloftype2diabetes,ex-vivoisletsshowlossof Cx36function,coordinated[Ca 2+ ] i dynamics,andalossofpulsatileinsulinsecretion.Similarly,inhumantype2diabetesthereisalossofpulsatileinsulinsecretion Langetal.,1981;Linetal.,2002andalossofpusatileinsulinsecretionisassociatedwithhyperglycemiaMatthewsetal.,1983;Matveyenkoetal.,2012;Meier etal.,2013.[Ca 2+ ]waveswithinisletsarethensurrogatemetricforcoordinatedinsulinsecretionresponseandassuchstudyingthebiophysicalmechanismsof[Ca 2+ ] wavesmayplayaroleindetermininghowtheyarecompromisedinmodelsofdiabetes. Heterogeneityof -CellsandRelevanceDiabetes Electricaldynamicsindissociated -cells,intheabsenceofelectricalcouplingby Cx36,showdistinctpatternsofoscillations:repetitivespiking,transientdepolarizationsinterlacedwithrapidfastspikes,andcyclicalplateaudepolarizationandhyperolarizationperiodsKinardetal.,1999.Similarly,theinsulinsecretionresponsesand overallinsulincontentofdissociated -cellswerealsodescribedin3groupsofGFP expressiondrivenbytheinsulinpromoterKatsutaetal.,2012whichhadapositive relationshipwithinsulinsecretion.Glucokinaseexpressionwasalsodiverseamong -cellsratleadingtoadistributionofglucoseconditionsbeforeindividual -cells wouldshowelectricalactivityPipeleers,1992. Pancreaticisletsarethencomposedofaheterogeneouspopulationof -cellswith varyingelectricalresponsestoglucoseandthelevelsofinsulinsecretion.Cx36func15

PAGE 28

tioninindividualpancreaticisletsisfunctionallyheterogeneous,althoughtheroleof Cx36istopromotefrequencyandphaselockingofaheterogeneouspopulationof -cellstoaverageoutheterogeneityandgiverobustresponsestoglucose-ensuringthatcellswithdifferentialresponsestoglucoseareeitherrecruitedtoactivityin thepresenceofhighglucoseorarerapidlysilencedwhenglucoselevelsdroptofastinglevelsSpeieretal.,2007.Curiously,afunctionalroleofheterogeneitymaybe topromotesynchronizationMontaseriandMeyer-Hermann,2016andtocontrol spatiotemporal[Ca 2+ ] i wavedynamicsBenningeretal.,2014.Recentworkusinga reportergeneofFltpFlattopshowedthat -cellsinmicecouldbedividedintotwo distinctpopulationsofmaturity,andmolecularandphysiologicalfeatures.Fltpnegativecellswereobservedtohavegenesrelatedtothefunctionalpropertiesoftheglucosestimulatedinsulinsecretionpathway,howeverhadhigherreplicationrateBader etal.,2016.Althoughmanyofthesestudiesoccurindisassociatedislets,recently Rutteret.al.usedhalorhodopsininatransgenicmousemodelandaphotoactivable sulfonuerainhumanisletstoobserveapopulationof -cellswithhighfunctional connectivity.Theseelectricalsilencingofthe'Hub'cellssilencedtheglobalelectricalactivityoftheisletJohnstonetal.,2016a. Inhumanislets, -cellswererecentlydescribedasexistingin4sub-populations baseduponcellsurfacemarkersDorrelletal.,2016.These4identiedsub-populations showedvaryingdegreesofglucosestimulatedinsulinsecretioneventhoughthelevelsofinsulinRNAandtotalinsulinproteincontentwereconsistentacrossthesubpopulations.Interestingly,intype2diabeticstherelativedistributionofthesesubpop16

PAGE 29

ulationsisalteredindicatingthatsub-populationsmayplayaroleinsusceptibilityin thepathophysiologyofthediseaseKilimniketal.,2011. 17

PAGE 30

CHAPTERII MATERIALSANDMETHODS MouseModels AllanimalswereusedunderUniversityofColoradoAnschutzMedicalCampus IACUCapprovedprotocolB-9581407D.Micewereheldinatemperaturecontrolledsettingwitha12hourlight/darkcycleandgivenaccesstofoodandwater continuously. -cellspecicexpressionofChR2-YFPwasdonethroughCre-Lox mediatedrecombinationbycrossinga -cellspecicCrerecombinaseline,PdxCre 6 Tuv=J JaxwithaChR2H134RmousemodelexpressedontheRosa26 locusROSA26Sor tm 32 CAG )]TJ/F22 7.9701 Tf 7.594 0 Td [(COP 4 H 134 R=EYFP Hze=J JAX:024109.Cx36 )]TJ/F22 7.9701 Tf 7.594 0 Td [(= )]TJ/F36 11.9552 Tf 11.415 -4.338 Td [(mice weregeneratedaspreviousdescriedDegenetal.,2004.Genotypingwasveried throughqPCRtransetyx. IsletCulturing HumanisletswereobtainedfromtheIntegratedIsletDistributionProgramTable S1duringyears2013-2016Ca2+,staticinsulin,gapjunctionmeasurementsor 2012-2016perifusionmeasurements.IsletswereculturedinCMRLat37 o C,5% CO2,for24-48hourspriortoimagingorinsulinsecretionassays.Mouseisletswere culturedinRMPImedia. MicroscopyandCalciumImaging ChR2expressingisolatedisletsweremountedon35mmglassbottomdishesin imagingsolutionmMNaCl,5.7mMKCl,2.5mMCaCl2,1.2mMMgCl2,10mM Hepes,and0.1%BSA,pH7.4andimagedusingaZeissLSM780systemCarl 18

PAGE 31

Zeiss,Oberkochen,Germanywith20X0.8NAPlanApocratobjectiveat37 o Cinthe presenceof2mM,5mM,11mMglucose.For[Ca 2+ ]measurements,isletswereloaded with3 mRhod-2AMAATBioquest,Sunnyvale,Californiafor30minutesat37 o Cin imagingsolution.Rhod-2wasexcitedusinga561nmsolidstatelaser,whichminimizesChR2 H 134 R activationPriggeetal.,2012,anduorescenceemissiondetectedat580nm-650nmusingamulti-anodePMTspectraldetector.Imageswere acquiredat1frame/sec.scantime650msat20 mdepthfromthebottomofthe islet.ChR2activationwasachievedusinga458nmAr+laserline.Activationregions intheisletweredenedusingtheZensoftware`bleachingmodule'andeachactivationregionsettoscan25timesover1s.Time-serieswererecordedalternating between10Rhod-2imagesfollowedbyaChR2activationsequence,andrepeated for40seconds.NADHPHautouorescencewasimagedunder2-photonexcitation usingatunablemode-lockedTi:sapphirelaserChameleon,Coherent,SantaClara, Californiasetto710nm,anduorescenceemissiondetectedat400nm-450nmusingtheinternaldetector.Z-stacksof6-7imageswereacquiredspanning5 mdepth. Humanisletswereloadedwith4 MFluo-4for90minutesatroomtemperatureand imagedonaninvertedNikonEclipse-TIwide-eldmicroscopeusinga20x0.75NA objective,at37 o C.Imageswereacquired1frame/secusing490nm/525nmemissionlter.Isletsfrombatchesinwhichcellviabilitywas < 80%,orisletswithabsent Fluo4-AMsignalorsignicantdriftwereexcludedfromsubsequentanalysis. 19

PAGE 32

PsuedoisletAggregation MIN6 -cellsweregrownandmaintainedinaT75ashingrowthmediaDMEM containing4.5g/lglucose,10%fetalbovineserum,50ug/mlstreptomycin,50U/mL penicillin,0.5ug/mlFungizone,1mMsodiumpyruvateand50uMbeta-mercatoethanol. Cellswerere-suspendedinmediaandwereseededinto48wellhydrogelmicrowelldevicesBernardetal.,2012at500,000cellsper0.5mLmedia.Thecellswere centrifugedinthedeviceat160rcffor2.5minutesandplacedinanorbitalshakerfor 2hoursin37C5%CO2.5dayspostcultureaggregateswerepassedovera40 m strainertoremovesmallaggregatesorparticlesandtheremaininglargeaggregates culturedin35mmdishes.Polystyreneparticles10 minsizeweresterilizedinethanol andtreatedwithfetalbovineserumandwereaddedtothecellsuspensionatdesired concentrationspriortoaggregateseeding. [Ca 2+ ]Coordination Human[Ca 2+ ]imageswereanalyzedinMATLABtoassessactivityandcoordination.A4-pixelaveraginglterwasrstapplied.Apeak-detectionalgorithmYoder, 2015recordedthelocationsandamplitudesofoscillationsacrossallpixelsofthe islet.Regionsinwhichnosignicantpeakscouldbedetectedweredeterminedtobe inactive.Coordinatedregionsweresegmentedbasedonthecoincidentpresenceof thetime-pointsofeachdetectedpeak.Thiswasfollowedbycross-correlationanalysisbetweentimecoursesofeachsubregion.Ifthecorrelationcoefcientwasabove 85%thetwosub-regionswereconsideredtobehighlycoordinatedandmergedinto alargerregion. 20

PAGE 33

ChR2ActivationAnalysis AllanalysiswasperformedusingcustomscriptsinMATLABMathworks[Ca 2+ ] measurementswerequantiedbyapeakdetectionalgorithmtestingforChR2stimulatedactionpotentialsonapixelbypixelbasis.Apixelwasdenedas`active'if [Ca 2+ ]peakswereidentiedatthetimeofChR2stimulationandnomorethan30% ofthepeaksidentiedcouldresultfromnonChR2stimulatedevents:inthiswaywe couldquantifyareasthatmayhavebeenpoorlystainedwith[Ca 2+ ]indicator.Areas thatcouldnotbestimulatedwhileactivatingthewholeisletwerenotincludedinthe analysisofsubregionstimulation.NADPHresponseforeachquadrantwascalculatedbyaveragingtheintensityacrossthez-stackuorescenceinthatquadrant andcalculatingthepercentagechangebetween11mMglucoseand5mMglucose. Toquantify[Ca 2+ ]oscillatoryfrequencyinCx36 )]TJ/F22 7.9701 Tf 7.594 0 Td [(= )]TJ/F36 11.9552 Tf 11.415 -4.339 Td [(islets,thepeakdetectionalgorithmfromabovewasusedtorecordthespatiallocationsandtimesofalloscillations acrossanislet.Individualcellswerelocatedbysegmentedregionsbasedoncoincidenttime-pointsofdetectedoscillations.NADPHresponseswerecalculatedby mappingeachcellularregionontotheNADPHimagesandtheresponseforeach cellularlocationwascalculated.Forimmunouorescenceimages,analysisforhormonestainingandChR2-YFPexpressionwasrestrictedtosingledissociatedcells identiedassinglenucleilocations:doublet,tripletoraggregatesofcellswereexcludedformanalysis. 21

PAGE 34

FRAP Cx36functionwasmeasuredusingFluorescenceRecoveryAfterPhotobleaching FRAP,aspreviouslydescribedFarnsworthetal.,2014.Isletswereloadedwith 12uMRhodamine-123for30minutesat37 o Cinimagingsolution.Rhodamine-123 wasexcitedusinga488nmAr+laserlineanduorescenceemissiondetectedat 500nm-580nm.2baselineimageswereinitiallyrecorded;aregionofinterestwas thenphotobleachedfor30secondsachievingonaveragea44%decreaseinuorescence;andimageswerethenacquiredevery15secondsfor6minutes.Theuorescentrecoverycurvecanbemodeledbythefollowing: I t = I 1 )]TJ/F21 11.9552 Tf 13.749 0 Td [(I p )]TJ/F21 11.9552 Tf 13.748 0 Td [(e )]TJ/F22 7.9701 Tf 7.593 0 Td [(kt + I p whereI p andI 1 representtheintensityafterbleachingandattheendofimaging, respectively.Therecoveryrate, k canbesolvedbyrearrangingequation4. e )]TJ/F22 7.9701 Tf 7.593 0 Td [(kt = I p )]TJ/F21 11.9552 Tf 13.749 0 Td [(I t I 1 )]TJ/F21 11.9552 Tf 13.748 0 Td [(I p +1 Wherethenegativeslopofthenaturallogofequation5isequaltotherecoveryrate. Immunouoresence IsolatedisolatesweredissociatedusingAccutase R Sigma-Aldrich,St.Louis,Missouriandplatedinto8chamberLab-Tekdishestreatedwith804Gcellmatrix.24 hoursafterplatingcellswerexedusing8%paraformaldehydeinPBSfor10min22

PAGE 35

utes.Antigenretrievalwasrstappliedtothecellsusing0.05%Tripsinwith10% CaCl2indH20for30minutesat37 o C.Cellswerethenpermeabilizedusing0.1% tritonX-100and5%donkeyseruminPBSfor2hours.Mouseanti-glucagonAbcam ab10988,Cambridge,UnitedKingdomandguineapiganti-insulinabcamab7842 primaryantibodieswereincubatedwithcellsata1:500dilutionat4 o Cfor24hours. Cy3anti-guineapigJackson706-165-15,WestGrove,PennsylvaniaandAlexauor 647anti-mouseJackson715-605-150secondaryantibodieswereappliedtocellsat a1:500dilutionatroomtemperaturefor2hours.Afterwashingcellsweretreated withDAPIuoromountSigma-AldrichandimagedonaZeissLSM800confocal microscope,DAPI,YFP,Cy3,Alexa647wererespectivelyexcitedwithlaserlines at440nm,488nm,561nm,640nmanddetectedoverrespectively410-470nm,500545nm,565-615nm,640-700nmwavelengthbands.InintactisletsFM4-64FXwas excitedat561nmanddetectedover600-620nm. InsulinELISA Insulinsecretionwasdeterminedinstaticassaysbyincubating10humanislets/tube inKrebs-Ringerbuffer.8mMNaCl,5mMNaHCO3,5.8mMKCl,1.2mMKH2PO4, 2.5mMCaCl2,1.2mMMgSO4,10mMHEPES,0.1%BSA,pH=7.4with2mMglucosefor1h,followedby1hincubationwitheither2mMor20mMglucose.Supernatantsecretedfractionwascollectedandremainingisletswerelysedwith2% TritonX-100andfrozenovernightforanalysisofinsulincontent.InsulinwasmeasuredwithamouseultrasensitiveELISAkitALPCO,ChrystalChem.Dataarepresentedasinsulinsecretionnormalizedbycontentforeachsample. 23

PAGE 36

Insulinsecretiondynamicswereassessedbyperifusionmeasurements,asdescribedinKaytonetal.,2015intheVanderbiltDiabetesResearchCenter.Only preparationswhichshowedatleasta2-foldincreaseininsulinsecretionfrom5.6mM to16.7mM,or5.6mMto16.7mM+100 MIBMX,werefurtheranalyzed.Perifusion time-courseswereanalyzedinMATLAB.Time-pointsfortheaddition/removalof16.7mM glucose IBMXwererecordedandthefoldchangeininsulinsecretioncalculated usingthemaximuminsulinsecretionvalueinthestimulationperiodagainsteitherall initialbaselinemeasurementspriorto16.7mMglucose,ortheprevious4timepoints priorto16.7mMglucose+IBMX.Attheendtime-pointof16.7mMglucose IBMX, anexponentialdecaya.exp-b.t+cwast,withbtherateofdecay. 24

PAGE 37

ClassicationLearners ClassicationsystemsweregeneratedinMATLABusingk-foldcrossvalidation k=5Kohavietal.,1995.ReceiverOperatingCharacteristicROCcurveswere generatedbycalculatingsensitivityvs.1-specicitycurvesatdifferentcutoffvalues withinthetrainedclassicationmodelSwetsetal.,1988.Inordertoevaluatemodel performancetheareaunderthecurveAUC,orC-statisticwascalculatedbynumericallyintegratingtheROCcurveusingthetrapezoidalmethod. 25

PAGE 38

CHAPTERIII DEVELOPMENTOFCOMPUTATIONALMODEL 1 Introduction Hodgkin-HuxleyHHlikemodelshavebeeninstrumentalindescribingelectrical dynamicsinmultipleexcitablesystems.ThecanonicalHHsystemreliesonasetof4 differentialequationsdescribingthecurrentthroughacellusingsodium,potassium, andleakconductancesaswellasaninjectablecurrent. I = C m dV m dt + g K n 4 V m )]TJ/F21 11.9552 Tf 13.749 0 Td [(V k + g Na m 3 h V m )]TJ/F21 11.9552 Tf 13.749 0 Td [(V Na + g l V m )]TJ/F21 11.9552 Tf 13.748 0 Td [(Vl where g i describestheconductanceofachannelselectivetoion i ,andn,m,hare gatingvariablesrepresentingthefunctionalvoltagedependentgatingkineticsof eachchannel.Uponsufcientinjectablecurrent,I,thesystemundergoesabifurcationshiftbetweensteadymembranepotentialandoscillatorypotential.Thesystem willcontinuallydepolarizeduetosodiumcurrentandhyperpolarizeduetotherectifyingpotassiumcurrentFigure6. 1 Theworkdescribedinthischapterwasacceptedforpublicationon7/16/2014Hraha TH,WestacottMJ,PozzoliM,NotaryAM,McClatcheyPM,etal.PhaseTransitionsintheMulti-cellularRegulatoryBehaviorofPancreaticIsletExcitability.PLoS ComputBiol10:e1003819.doi:10.1371/journal.pcbi.1003819 26

PAGE 39

Figure6. MembranepotentialresponseasafunctionofinjectedcurrentinHH model.Withincreasinginjectedcurrentthemembranepotentialofthesimulatedcell increasesuntilathresholdvalueuponwhichthesystemundergoeshopfbifurcation andasteadyoscillatorysolutionwillexist -cellswerefoundtogenerateactionpotentialsinresponsetoglucoseDeanand Matthews,1970andwereregulatedspecicallybyATPbindingpotassiumchannels 27

PAGE 40

Ashcroftetal.,1984.HHlikemodelsof -cellelectricalactivitywererstdeveloped intheearly1980sChayandKeizer,1983.Themodelconsistedofacalciumactivatedpotassiumchannel,avoltagegatedpotassiumandcalciumchannel,aleaky conductanceforsodiumandchloride,andaninjectablecurrent.Themembranepotentialisthenrepresentedas: C m dV dt = g K;Ca + g K;Vm V m )]TJ/F21 11.9552 Tf 10.741 0 Td [(E K +2 g Ca;Vm V m )]TJ/F21 11.9552 Tf 10.741 0 Td [(E Ca + g L Vm )]TJ/F21 11.9552 Tf 10.742 0 Td [(V L + I LaterMagnusandKeizerbeganmodellingmedabolicpathwayssuchasglycolyis, oxidativephosphorloation,andtheTCAcycleMagnusandKeizer,1998b.Fridylandet.albegantoincludeseperatecurrentsfor[Ca 2+ ]dynamicsintheERetc.The Cha-NomamodelisanadaptationofmodelderivedbyFridylandet.al.Fridlyand etal.,2003.Withminorchangestometaboliccomponentswereadaptedtomore recentliterature. 28

PAGE 41

Figure7. CurrentsourcesinCha-NomaModel.Glucoseismetabolizedthrough ratesofglycolysistogenerateNADPHReandADPisconvertedtoATP.Adapted fromChaetal.,2011a. Briey,theCha-Nomamodelcontains18coupleddifferentialequationswhichmodel ATPgenerationfromglycolysisandoxidativephosphorlationfrom[Fridylad2005]7. Themembranepotentialisdescribedby11currentsourcesincludingvoltageactivatedcalciumchannel,KATPchannel,delayedrectifyingpotassiumcurrent,sodium potassiumandsodiumcalciumpumps,calciumactivatedpotassiumcurrents,store operatedcurrentandplasmamembranecalciumpumpcurrents. 29

PAGE 42

C m dV m dt = I CaV + I TRPM + I SOC + I bNSC + I KDr + I KCa BK + I KCa SK + I KATP + I NaK + I NaCa + I PMCA ModelDescription Inordertoaccuratelymodelelectricaldynamicswithinpancreatic -cellsthesingle cellmodelwasexpandedinto1000heterogeneouscells.Whileothergroupshave implementedmulti-cellularmodelsusing10x10x10matrixSilvaetal.,2014ofcells wechosetouseaspherepackingalgorithmSkogeetal.,2006inordertogeneratearandomgroupingofcells.Firstbyassemblingacubicpackingof4000spheres thenwechoosetherst1000cellsclosesttothecenterofthepacking-assembling asphericalstructure.Withthisweknowthecoordinatesofthecenterofeachcell. Nextinordertocalculateeachcells'nearest-neighborinordertoassignfunctional connectivitywecalculatedthedistancebetweeneachcellcenter.Ifthisdistance waslessthan1.4xthediameterofacellweconsideredthosecellstobefunctionallyconnected.ThisgaveaGaussiandistributionofconnectivitywhereeachcellwas onaverageconnectedto5.5 2.5cells.Inordertointroduceheterogeneityintothe systemwetook10parametersfromeachcellsummarizedintable1andassigned Gaussiandistributionstothatproperty.Thiseithermadecellsmore/lesselectrogenic withvaryingoscillationsfrequenciesanddutycycles. 30

PAGE 43

Table1. ListofParametersusedinCha-NomaModel.Parametersusedforislet model.HeterogeneityisbasedonaGaussiandistribution,unlessotherwiseindicated,withastandarddeviationasapercentageofthemeanvalue =% 31

PAGE 44

TomodelCx36currentbetweencellsweintroducedacurrentproportionaltothe transmembranepotentialbetweentwoconnectedcells: I i;j = g coup V i )]TJ/F21 11.9552 Tf 13.748 0 Td [(Vj Whereiandjareeachcoupledcelland g coup istheaverageconductancebetween cells.Thisvaluehasbeenmeasuredtobe 100pS,howeverratherthanassigningagaussiandistrubtionofconductivityweuesdagammadistributionfollowing Farnsworthetal.2014Farnsworthetal.,2014.Weimplementedthiscoupledmodel inC++andsolvedthesystemusingaforwardEulermethodusingatimestepof 150 s.SimulationswererunontheJANUSclusterUniversityofColoradoBoulder Individual -cellsshowawiderangeofresponsestoglucoseseeintroduction whichisrepresentedinthemodelwhenwesetthecouplingcurrentto0.However, whenwefunctionallycouplethecellsthesimulatedisletsshowcoordinated[Ca 2+ ] waveswhenglucoselevelsareabove6.5mMandarefunctionallyquiescentbelow 5mMglucoseFigure8. 32

PAGE 45

Figure8. Glucoseactivated[Ca 2+ ]oscillationsinsimulatedislet[Ca 2+ ]oscillations inWTtopCx36 )]TJ/F22 7.9701 Tf 7.594 0 Td [(= )]TJ/F36 11.9552 Tf 11.415 -4.338 Td [(middleasafunctionofglucoseconcentrationbottom.WT isletsshowabruptchangefromquiescenttooscillatorynear8mMglucose.AlternativelyCx36 )]TJ/F22 7.9701 Tf 7.594 0 Td [(= )]TJ/F36 11.9552 Tf 11.415 -4.338 Td [(isletshavecontaincellswhichshowoscillationsatlowerglucose concentration. ModelVerication Tovalidatethecoupledmodels'effectivenessatsimulatingpropertiesof ex-vivo isletswechoseamodelofneonataldiabetes.Neonataldiabetesisarelativelyrare disease,effectingonly1in200,000individuals.Thisconditionisassociatedwitha mutationwiththeATPregulatedpotassiumchannelKATPchannelresponsiblefor controllingtheinitialdepolarizationinresponsetoelevatedglucose.Whilethereare multiplemutationsassociatedwiththischannelPinneyetal.,2008,manydecrease thesensitivityofthechanneltoATPontheKir6.2subunitofthechannelandrender themhyper-active-requiringhigherlevelsofATPtocloseanddepolarizethemem33

PAGE 46

brane.WegeneratedamousemodelwithamutationontheATPbindingdomain oftheKATPchannelKir6.2 [ N 30 ;K 185 Q ] -GFPunderinducibleCre ER -recombinase whichwouldonlybeexpressedupontamoxifeninjection.Themutationwouldexpressmosaicqualities,showingexpressioninonlyacertainfractionof -cellswithin islets,thepercentageofwhichweredependentonthenumberofinjectionsgivento themouse.Eachisletcouldnowhavebetween0%-60%mutatedchannels,dependentonthenumberofinjectionsgiventothemouseFigure9.WedeneP exc as 1-%mutatedcellswhereP exc =100%or1asWTandP exc =0%asisletwideexpressionofKir6.2 [ N 30 ;K 185 Q ] -GFP. Figure9. PenetranceofKir6.2 [ N 30 ;K 185 Q ] -GFPVaryingdosesofTamoxifen-3 injectionsshow0%-60%ofcellswithpositiveexpression,controllingtheoverallexcitabilityoftheislet. TheeffectthemutantKATPchannelshadonacellsindividualelectricalactivity weretosilenceit,regardlessofglucoselevelsMakingitunresponsivetoevenhigh glucoselevelsmM.As -cellsareelectricallycoupled,inaisletwithmoscaicex34

PAGE 47

pressionofhyperactiveKATPchannelstherewouldbeamixofelectricallyinactive andactivecells.TheinactiveKATPmutantcellsactasaninhibitingcurrentsink,preventingtheneighboring,electricallyactivecells,fromcompletelydepolarizing. Toimagecytosolic[Ca 2+ ],isletswereloadedwith4 MFuraRed-AMatroomtemperaturefor90minutesandimagedinaspinning-diskconfocalmicroscope.Imageswereacquiredat1frame/secusinga488nmdiodelaserforexcitationanda 580nmlong-passlterforemission.Kir6.2 [ N 30 ;K 185 Q ] -GFPwasimagedat 488nmandaquiredat500nm-520nmCHECK. IfanislethadasmallfractionofcellswithmutantKATPchannels < 10%theislet wouldstillbeabletoshowoscillationsat20mMglucosehowevertheoverallactivitywouldshowasmalllineardeclineFigure10I,II.Howeveratacriticalnumberof inactivecells > 20%thecurrentassociatedwithclosureofKATPchannelsinunmuatedcellswouldsinktothosecellswithmutatedKATPchannels-causingarapid globalcollapseofelectricalactivitythroughouttheisletFigure10III.AfterthiscriticalvaluetherewouldagainbealineardecayinactivityasafunctionofKATPmutant cellsFigure10IV. 35

PAGE 48

Figure10. PhasetransitioninexcitabilityAPercentcellsshowing[Ca 2+ ]ielevationsasafunctionofnumberofexcitablecells,asdeterminedbylackofGFPand thus.Kir6.2 [ N 30 ;K 185 Q ] expressioni.e.Pexc=1-%GFPBRepresentative[Ca 2+ ]i dataforisletsindicatedinA,fromregionsofwild-typeI,`pre-critical'II,`critical'III and`post-critical'IVlevelsofPexc.Left:Areasofactivityarehighlightedinredand scalebarsrepresent50 m.Right:Representativetime-coursesofnormalizedFuraRedcalciumdyeuorescenceforcellswithineachislet,whereverticalscalebar indicates20%changeinuorescence.Redtime-coursesaredeterminedtobeactive,blacktime-coursesaredeterminedtobeinactive. Insulinsecretionistightlyregulatedby[Ca 2+ ]levelswithin -cellssoweanticipated insulinsecretiontofollowsimilarcriticalbehavioras[Ca 2+ ]activityinKir6.2mutant mice.WeperformedaninsulinELISAonisletsfromFigure10usingtherecorded GFPvaluesfrom[Ca 2+ ]micrsocopy.Insulinsecretionfollowedasimilarcriticaltrend Figure12to[Ca 2+ ]activity. 36

PAGE 49

Figure11. Linkbetweenphasetransitionsin[Ca 2+ ]iandphysiologicalparameters.APercentcellsshowing[Ca 2+ ]ielevationsaveragedoverisletsfromeach Kir6.2 [ N 30 ;K 185 Q ] -expressingmouseasafunctionofPexc%-%GFP.Right: Mean s.e.m.fordatabinnedtowild-type,pre-andpost-criticalrangesasdeterminedby%GFP. indicatessignicantdifferencep < 0.0001betweendataas indicated.BPlasmainsulinlevelsfromeachmouseasafunctionofPexc.Right: Mean s.e.m.fordatabinnedasinA. indicatessignicantdifferencep < 0.05 betweendataasindicated. Wetestedifwecouldreproducethiscriticalbehavior in-silco bymutatingtheATP responsekineticsoftheKATPchannel.ThecurrentthroughtheKATPchanelwas describedasMagnusandKeizer,1998a: I KATP = g K ATP po K ATP V )]TJ/F21 11.9552 Tf 13.748 0 Td [(V K 37

PAGE 50

where g representstheopenchannelconductanceand po K ATP representstheATP regulatedopenprobabilitydescribedby: po K ATP = 0 : 08+ 2[ ADP ] 0 : 01 +0 : 89 [ ADP ] 0 : 01 2 1+ [ ADP ] 0 : 01 2 + 0 : 45[ ADP ] 0 : 026 + [ ATP ] 0 : 05 TomodelexpressionofKir6.2 [ N 30 ;K 185 Q ] theopenprobabilitywasmodiedtoincludeaconstantcurrentterm: po K ATP Mut = po K ATP + )]TJ/F21 11.9552 Tf 13.749 0 Td [( where wassetequalto0.5.ThistermcontributesanadditiontotheopenprobabilityfollowingresultsfrompreviouscharacterizingofglucosevsI KATP response curvesAshcroft,2005;HattersleyandAshcroft,2005;Kosteretal.,2005.Cells insilico werespatiallyrandomlyassignedaseitherWTormutantKATPcurrents.Individualcellsweredescribedaselectricallyactiveiftheirmembranepotentialreach valuesover-45mV.Aswith ex-vivo results,thecoupleddynamicaloscillatormodel predictedsimilarcriticalbehaviorFigure12.Initiallythereisalineardecayinactivityatsub-crticalvalues%-12%followedbyarapiddropoffinelectricalactivity near20%mutatedcellsandalineardecayinsupra-criticalvalues.Similarly,toFigure10,[Ca 2+ ]dutycycledecreasesandoscillatoryfrequencyincreasesasafunctionofmutantcellsFigure12B. 38

PAGE 51

Figure12. Coupledoscillatormodeldescribesexperimentalphasetransitions.A Percentcellsshowing[Ca 2+ ]ielevationsinsimulatedisletsasafunctionoffraction ofexcitablecellsPexc,assetbythe%cellslackingATP-insensitivity.Solidlinerepresentsmeanofsimulationresultsgeneratedfrom5randomnumberseeds,dashed linesrepresents95%condenceintervalsofsimulations.BRepresentativesimulated[Ca 2+ ]itime-coursesforparametersindicatedinA,fromregionsofwild-type I,`pre-critical'II,`critical'IIIand`post-critical'IVbehavior,asinFigure11.Verticalscalebarindicates20%changeinsimulated[Ca 2+ ]i.Redtime-coursesare determinedtobeactive,blacktime-coursesaredeterminedtobeinactive.CPercentcellsshowing[Ca 2+ ]ielevationsinsimulatedisletasafunctionofnumberofexcitablecellsPexcforvaryingmeangapjunctionconductancevalues.Filledsquares indicateexperimentaldatafromKir6.2 [ N 30 ;K 185 Q ] -expressingisletsinFigure10. Interestingly,thelevelofCx36couplingpredictsdifferentresponsestoincreasesin thenumberofmutantcellsFigure12C.While120pSwasusedasthebest-tto ex-vivo datablackdotsbydecreasingCx36conductanceweanticipatealesscriti39

PAGE 52

calbehavior.ThismodelofpancreaticisletdysfunctionfollowstheanalogyofmeaneldtheoryintheclassicalisingmodelinferromagnetismFigure13. Figure13. Mean-eldtheoryanalogyof -cellnetworkactivity.Theexcitabilityof constituentunitsi.e.theirglucosesensitivitiesandtheresultingnetworkactivity takesintoaccountcouplinganddifferentexperimentalperturbations.ACellsofwildtypeisletsareinexcitableatlowglucosemMandallexcitableathighglucose mM,thereforewild-typeisletsarerespectivelyfullyinactiveandfullyactive.Cells expressingKir6.2 [ N 30 ;K 185 Q ] areglucose-insensitiveandconstitutivelyinexcitable. WhenKir6.2 [ N 30 ;K 185 Q ] penetranceis < 15%Pexc > 0.85thereareinsufcientinexcitablecellstosuppressglobalactivity,andsocouplingleadstoinexcitablecells beingrecruitedtobeactive.However,whenKir6.2 [ N 30 ;K 185 Q ] penetrancerises above15%Pexc < 0.85,globalquiescenceensueswherethemajorityofnormally excitablecellsarerenderedinactive. -cellsaregivensimpleup/downparitywhereinactiveorunabletobeactiveare representedbydownandlargedownarrows,respectively.Healthycellsareinactive 40

PAGE 53

at2mMglucoseandtransitiontoactiveat20mMglucose.Atthepre-criticalstageof mutationthereisasmallsubpopulationofcellswhichareunabletobecomeelectricalactiveontheirown,howeverdotoelectricalcouplingtheycanberecruitedtobe activeatelevatedglucoselevels.Atthepostcriticalstagetherearesufcientnumbersofinactivecells,whichhavehigherhyperpolarizingcurrentthansurrounddepolarizingcells,suchthatdocouplingtheentiresystemisbroughtdownintoaninactive state. We'veexaminedhowwholeisletelectricalactivityundergoesaphasetransition whenaddingelectricallyinexcitablecells.Duetothecouplednatureofisletswe hypothesizedthatisletsmayundergoasimilarphasetransitioninelectrical syncrhonization ifinsteadofincludinginexcitablecellsweincludefunctionallyuncoupledcells.Theseelectricallyuncoupledcellscouldconsistofmultiplesourcesincludingalternativeisletcelltypessuchas or cellsor -cellswhichdonotexpress Cx36.Alternatively,anothersourceofuncoupledcellswouldbetoincludesmall micro-particleswhichincorporateintothestructureoftheislet-creatingafunctional scaffoldbutvoidofanyelectrophysiolgyproperties.Thelatterwasaccomplished throughtheincorporationofpolysterenemicroparticlesinre-aggreatedpseudoislets. Thesepsuedoisletsarecreatedbyaggregatinga -celllineMIN6CellsintoasphericalarchitectureFigure14. PseudoisletswerestainedwithFluo-4AMtomonitor[Ca 2+ ]activityandimagedin 20mMglucosetomonitor[Ca 2+ ]oscillations.Threetofourisletsofeachparticle 41

PAGE 54

seedingwereimagedand[Ca 2+ ]coordinationcalculatedusingthealgorithmdescribedinthemethods.Thepercentageofmicroparticleswithinisletswascaclulated bythresholdingtheuorescentsignalfromthemicroparticlesandcalculatingthenet areaofuorescenceattheimagingplane. Figure14. ModelofSynchronizationdisruption.Threedimensionpseudoisletsare formedinsidemicrowellsconsistingofgreen -cellsorredmicroparticles.Beyonda limitofmicroparticlesweexpecttheisletpseudoisletwouldbebeyondacriticalcouplingthresholdandelectricalactivitywouldnolongershowsynchronousoscillations. Scalebarindicates100 m. 42

PAGE 55

Weobservedthatpseudoisletswithlowpercentageofmicroparticlesdisplay[Ca 2+ ] oscillationsidenticaltothatofmouseisletsFigure15andalthoughtherewasdropoff athigherlevelsofseedingi.e. > 20%isletswerestillcoordinatedoverall.However at38%seededparticlestherewasarapiddropoffin[Ca 2+ ]coordination.Indicatingthatbetween20%and40%seedingthereexiststhecriticalcouplinglimitwithin pseudoislets. Figure15. LossofelectricalsynchronizationwithmicroparticleincorporationLeft [Ca 2+ ]Activityinpseudoisletsasafunctionofmicroparticleincorporation.Right Quanticationofthelargestareaofcoordinated[Ca 2+ ]activitynormalizedtoislet size`Max.coordinatedarea'averagedoverpsuedoisletsasafunctionofmicroparticleincorporation. Toreplicatethisexperiments in-silico wemodiedourcoupledmulticellularmodel toinsteadofexpressingacertainfractionofmutant,in-excitablecells,weuncoupled avariablefractionofcellsfromtheirneighbors.Thiscreatesanisletwitheffective couplingholes,similartothepseudo-islets.Athighfractionofholestheoscillations appeartouncoupleFigure16B.Theoverallisletsynchronizationwascalculatedby rsttakingtheHilberttransformationofeachcell'smembranepotentialtime-course. TheHilberttransformgivesananalyticalrepresentationoftheinputsignal,returning 43

PAGE 56

acomplexanswercontainingthephaseinformationofofeachtimecourse.Theinstantaneoussynchronizationfactorrateachtimepointcouldbecalculatedwiththe following: r t = 1 N N X j =1 exp i t j Figure16. ModelofSynchronizationDisruptionThreedimensionpseudoisletsare formedinsidemicrowellsconsistingofgreen -cellsorredmicroparticles.Beyonda limitofmicroparticlesweexpecttheisletpseudoisletwouldbebeyondacriticalcouplingthresholdandelectricalactivitywouldnolongershowsynchronousoscillations 44

PAGE 57

Where t j representsthephaseofeachcellattimepoint t and N representsthe numberofcellsnotincludinguncoupledcells.Atime-averageofrgivestheoverallsynchronizationinthetime-coursewhere1representscompletecouplingand0 representsnocoupling.Ourndingsshowthatat0%uncoupledcellsholesthere isletshowsveryhighlevelsofcouplingFigure16Cnear0.9.Thesynchronization decreaseswithaslightlineartrenduntil40%afterwhichthereisancriticaltransition tolowsynchronization. Conclusions Thendingsofthesestudiesshowthatexpandingawell-establishedsinglecell modelofpancreatic -cellselectricalactivityintoarepresentativemulti-cellularpancreaticisletaccuratelyrecapitulateswholeisletelectricalactivity.Themodelfeaturesseveralattributeswhichcontributetoit'sfunctionandstrengths-heterogeneity inelectricalandmetabolicpropertiesbetween -cellswhichgivecellsvaryingdegreesofelectrogenicityinresponsetoglucoseFigure8whichhasbeendescribed previousPipeleers,1992.IsletarchitecturehasbeenwellcharacterizedCabrera etal.,2006wecreatedarandomsphericalpackingofcellstomostaccuratelyrepresentcellsintheirin-vivostateandwhilesomestudieshaveusedcubicpackingof cellsSilvaetal.,2014othersusedorderedhexagonalclosedpackingNittalaetal., 2007,nonehaveusedrandompackingtocreateuniqueisletswitheverysimulation.WithourrandomsphericalpackingweassignaCx36couplingcurrentunique toeachcellandfollowingadistributionshowstrendssimilararight-skewedGaussian,orgamma,distributionFarnsworthetal.,2014.Atglucoselevels < 8mMthe 45

PAGE 58

coupledisletisquiescentwhileanelectricallyuncoupledisletcontainspopulations ofcellswhichshowelectricalactivity.Inthecoupledstateweobservearobusttransitiontoanoscillatorystateat 8mMglucose,consistentwithex-vivoobservations ofelectricalactivityinislets.Thebehaviorofourmodelrecapitulatesisletfunction, wenexttestedhowthemodelcouldpredictdysfunctionintwouniquesituationsthat centeraroundelectricalcouplingbetween -cells. Wecreatedaninducibletransgenicmousemodelshowingmosaicexpressionof aKATPchannelinsensitivetowardsATPthroughKir6.2mutation[citesomepeople]Figure9whereincreasingthedosageoftamoxifenincreasesthepercentage ofcellsexpressingthischannel.CellsexpressingKir6.2 [ N 30 ;K 185 Q ] areunableto showelectricalactivityatsupra-thresholdglucosevaluesmM.Howeverdueto couplingcurrentthroughCx36theymayshow[Ca 2+ ]oscillationsifsufcientdepolarizingcurrentispassedthroughcouplednon-mutantcells.At 15%cellularexpressionof[thatthing]thecouplingcurrentofWTcellsisunabletorecrcuitsurrounding mutatedcellstodepolarize-forcingtheisletintoaquiescentstateFigure10and suppressinsulinsecretionFigure11.OursimulationsshowthatmodulatingCx36 conductivitydecreasesthecriticalthresholdforinactivationFigure13.Indeed,abolishingCx36couplingwithahomozygousknockoutrescuesinsulinsecretionNguyen etal.,2014athighlevelsofmutantKATPpenetranceanda50%Cx36knockout usingaheterozygouslossoffunctiondecreasesthecriticalthresholdNotaryetal., 2016.Wetestedifthemulticellularmodelwouldshowsimilarphasetransitionafter 46

PAGE 59

incorporatingcellswithhighKATPchannelopenprobabilitybyaddingaconstituitive openprobabilitytotheK ATP .Atlowlevelsofmutantchannelexpressionisletsshow asmalldecreaseinelectricalactivity.At 15%mutantcellsthesimulatedisletundergoesasimilarphasetransitionwhereitrapidlytransitionstoquiescent. Wenextexaminedhowdeleteriouseffectstoelectricalcoupling,throughtheinclusionofnon-coupling'cells'alterthecoordinatedelectricaldynamicsinaggregated pseudoisletsFigure14andparalleledthisinthemulticellularmodelbyrandomly disconnectingpopulationsofcellsFigure16.Whilebothmodelsshowedcritical lossinsynchronizationtherewasadiscrepancyinthelevelofuncoupledcellsseen betweenthepseudoisletsandthemulticellularmodel.Theformershowedlossin synchronizationatuncoupledcellsatlevelsexceeding > 20%whereasthemulticellularmodelthedecreasewasrightshiftedtowards > 50%.Thediscrepancymay resultfromafewdifferencesbetweenthepseudoisletsandmodel.Wesettheaveragecouplingconductancebetweencellsat120pSwithatypicalcellfunctionally connectedto5surroundingcells.Lessisknownofthefunctionalcouplingbetween pseudoisletaggregatesandexpressionofCx36.Alowerleveloffunctionalcouplingmayhelpexplainthisdiscrepancyandcouldbeexaminedfurtherbydirectly quantifyingCx36functionusingauorescencerecoveryafterphotobleachingassayFarnsworthetal.,2014.Furthermore,theextenttowhichmicroparticlesdisrupt electricalcouplingmayshowastrongereffectthanuncouplingsimilarnumbersof cellswithinthemulticellularmodel-possiblythroughmechanicalinteraction. 47

PAGE 60

Insummary,thesetwostudiesverifythesinglecellelectrophysiologymodelbased onChaetal.canbeexpandedtoacoupledmulticellularmodelthatrecapitulates twoaspectsofpancreaticisletdysfunction.Theformerbaseduponcriticallossthe electricalactivitycausedbyelectricallyinactive -cellscoupledtothoseofnormal electricalfunctionwhereasthelatterdemonstratespancreaticisletsshowcriticalloss inelectricalsynchronizationupontheinclusionofnon-couplingcells.Furthermore, theseresultsshowhowthatasmallsubpopulationofcellswithalteredfunctioncan exertdisproportionatecontrolovertheelectricalactivityofpancreaticislets. 48

PAGE 61

CHAPTERIV SPATIALLYORGANIZEDSUB-POPULATIONSOFCELLSCONTROL ELECTRICALACTIVITYANDDYNAMICSACROSSTHEISLETOF LANGERHANS 2 Introduction Theemergentpropertiesofmulticellularsystemshaveledtoincreasedstudyof thearchitectureandbiologicalheterogeneityingoverningspatiotemporaldynamicsBreakspearandStam,2005;PanditandJalife,2013.However,theinherent complexityinmulticellularsystemsoftenrendersthemdifculttostudyintheirintact state.Asaresult,systemsareoftenbrokenapartintomoremanageablepiecesand theircoupledfunctionsanddynamicsinferredfromindividualcomponentproperties Katsutaetal.,2012;Strogatz,2001.Whilesubpopulationsofsystemsareidentiableusingthesemethods,theirexactroleincontrollingdynamicsandsystemregulationintheintactstateislost.Newtechniquesthatcanstudystructuresintheirintact state,inordertopreservespatialinformationofsignalingdynamics,willtherefore helpelucidatetheroleofcellularsubpopulationsincomplexsystems.Onesuchsystemthatshowscomplexmulticellularregulation,yethasatractablescaleforstudying withcellularimagingandcomputermodelingapproaches,istheisletofLangerhans; wheredysfunctiontotheisletgenerallycausesdiabetes.Individual -cellswhen 2 Theresultspresentedinthesectionweresubmittedon2/14/17andarevised manuscriptsubmittedon6/8/17.Westacott,M.J.,Ludin,N.W.,Benninger,R.K.P., Spatiallyorganizedsub-populationsofcellscontrolelectricalactivityanddynamics acrosstheisletofLangerhans. BiophysicalJournal 49

PAGE 62

electricallyisolatedshowlargevarianceinglucosesensitivity,metabolicactivity,electricaldynamicsandinsulinsecretionPistonetal.,1999;Zhangetal.,2003;Jetton andMagnuson,1992;andthisisdistinctfromtheuniformcoordinatedelectricaland secretoryresponseintheintactislet.Understandinghowcellvariabilitycontributes totheoverallfunctionoftheislet,andhowthepresenceofsub-populationsofcells affectsisletfunctioninpathogenicconditionsispoorlyunderstood.Forexample smallpopulationsofinexcitablecellsthatshowalteredKATPactivitycansuppress activityacrosstheisletinthepresenceofcouplingSpeieretal.,2007;Hrahaetal., 2014b;Rocheleauetal.,2004.Thiscausesasharptransitionbetweenglobalactivityandglobalquiescencethatcanresultfromsmallchangesinglucoselevelsand KATPactivity,andisphysiologicallyimportantSto zeretal.,2013.Thecoordinated dynamicsofelectricalactivityshowspatialheterogeneityor`SmallWorld'propertiesincommonwithotherbiologicalsystems,wherelike-dynamicsarerestrictedto subregionsoftheisletJohnstonetal.,2016a;Benningeretal.,2008.Propagating calciumwavesthatmediatesynchronizationof[Ca 2+ ]ioscillationsalsoinitiatefrom subregionsoftheislet;whichhasbeensuggestedasapacemakerregiondened bylocalexcitabilityBenningeretal.,2008,2014.Despiteobservationssuggesting thatsub-populationsofcellswithintheisletmayaffectseveralaspectsofcoordinated function,discoveringtheirpresence,characterizingtheirintrinsicbehavior,andunderstandinghowtheycontributetocoordinatedisletfunctionisnotwellcharacterized.Tostudyhowfunctionalsub-populationsofcellsmaybedistributedthroughout multicellularstructuressuchastheislet,andtheroletheymayplayinaffectingglobal 50

PAGE 63

function,wearelimitedinwaystoacutelyperturbfunctionandmeasureresultantresponses.Hereweapplyoptogeneticsbyusing -cellspecicChannelrhodopsin-2 ChR2expressionNageletal.,2003andtime-dependentlaser-scanningconfocal activationtodenetheelectricalregulationofsub-populationsofcellswithintheislet. Wecombinethiswithquantitativeuorescencemicroscopymeasurementsof -cell functionandamulticellularcomputationalmodelofcoupled -cellelectrophysiology. Withthis,wetestifisletsofLangerhanscontaindiscretefunctionalsubpopulations anddeterminewhateffectthesesubpopulationshaveoncontrollingthecoordinated electricalregulationandelectricaldynamics. Results LocalChR2activationrevealsregionsofvaryingexcitability WerstexaminedthespatialdependenceofChR2activationandlocalmembrane depolarizationon[Ca 2+ ]iregulationacrosstheislet.ChR2-YFPwasexpressedin -cellsunderCre-recombinasecontrolFigure17A,throughaPdx-Crelinewhich showsearlyexpressionandlacksmosaicismHingoranietal.,2003.ChR2-YFPexpressingcellsweredistributedthroughouttheisletFigure19.Greaterthan80%of insulinpositivecellsexpressedChR2-YFP,whereaslessthan10%ofglucagonpositivecellsandlessthan10%ofnon-glucagon/insulinpositivecellsexpressedChR2YFPFigure19.AtbasalmMglucoselevels,activatingChR2at451nmwithan opticalpowerof0.1mWperiodicallyacrossthewholeisletledtocoincidentelevationsof[Ca 2+ ]iacross84% 5%oftheislet,risingimmediatelyafterexcitationand subsequentlydecayedexponentiallywitharateconstantof0.5s-1Figure17B, 51

PAGE 64

consistentwithpreviousndingsReinbotheetal.,2014.ActivatingChR2across subregionsof25%oftheisletquadrantselevated[Ca 2+ ]iacrossthewindowof ChR2activationandextendingoutsideoftheactivationwindow,butnotacrossthe wholeisletFigure17C.UponChR2activationacross25%oftheisletvaryinglevelsofelevated[Ca 2+ ]iwereobservedtoextendoutsideoftheactivationwindow Figure17D;suggestingthatsomeisletregionsweremorereadilyactivatedbyChR2. Onaveragetheregionthatprovidedtheleast[Ca 2+ ]istimulationactivated15%of theislet,whichwerefertoas`ChR2activatedarea',whereastheregionthatprovidedthemost[Ca 2+ ]istimulationactivated30%oftheislet.Relativetothemean ChR2activatedareaforanislet,theminimumChR2activatedareawas40%lower andthemaximumChR2activatedarea40%higherFigure17E.ChR2-YFPexpressingcellsweredistributedthroughouttheisletFigure19.Activationofsmaller singlecell-sizedregionsalsoelevated[Ca 2+ ]ioutsideoftheactivationwindowFigure17F,althoughwithhighvariabilityFigure17G:50%ofcell-sizedregionsshowed negligibleactivationof[Ca 2+ ]Ioutsidethewindowand50%ofcell-sizedregions showedsignicantactivationwithasmallpopulationshowingsubstantialactivation ofover50%oftheislet.Overall,theamountofChR2activatedareaincreasedwith thesizeoftheilluminationwindowFigure17H:Illuminating100%oftheisletledto aChR2activatedareaof84% 5%;illuminating50%oftheisletledtoaChR2activatedareaof41% 5%;illuminating25%oftheisletledtoaChR2activatedarea of21% 4%;andilluminatingsinglecellregions5%oftheisletledtoaChR2activatedareaof16% 3%. 52

PAGE 65

Figure17. ChR2activationcreateslocal[Ca 2+ ]ielevationinpancreaticislets.A ModelofspatialactivationofChR2inpancreaticislets.Activationregionswithinthe isletaredenedoverwhichChR2isactivatedby458nmillumination.Activationof ChR2leadstodepolarizationandopeningofvoltagegatedcalciumchannels.B Anactivationregiondenedoverthewholeisletleft,greendashesgeneratesislet wide[Ca 2+ ]iinux,detectedoverthemajorityoftheisletleft,orange.[Ca 2+ ]iinuxincreasesrapidlyfollowingChR2activation,asmeasuredbyRhod-2uorescenceright.CAnactivationregiondenedoveraquadrantoftheisletleft,green dashesgenerateslocal[Ca 2+ ]ielevationwhichextendsoutsideoftheactivationregionleft,orange.Moredistantareasoftheisletshowno[Ca 2+ ]iinuxright.Scale barinB,Cindicates2%changeinuorescence.DChR2activated[Ca 2+ ]resultingfromquadrantregionsofactivation,presentedinrankorder.TheChR2-activated areawhere[Ca 2+ ]iiselevatedfromactivatingChR2inaquadrantregionasinC isnormalizedtotheChR2-activatedareafromactivatingChR2overthewholeislet asinB.EIntra-isletvariationinChR2activated[Ca 2+ ]relativetotheisletaverage.TherelativeareaofChR2-activated[Ca 2+ ]iovereachquadrantasinDwas expressedrelativetothemeanChR2-activated[Ca 2+ ]iofeachislet,andsortedfrom leasttomost.FActivationofsmaller`single-cell'regionswithinisletssimilartoC. [Ca 2+ ]iselevatedwithinandoutsidetheactivationregionright.GDistributionof ChR2activatedareainsingle-cellactivationregionsn=96regions.HChR2activatedareaof[Ca 2+ ]ielevationuponvaryingsizesofactivationregion.Scalebarin B,C,Findicates2%changeinuorescence.DatainD,Eaveragedovern=25islets from6mice;datainGaveragedovern=96regions,26isletsfrom5mice.DatainI-K averagedovern=10isletsfrom2mice.***indicatesp < 0.001.Scalebarinimages represents100 m. 53

PAGE 66

UnderrepeatedactivationofChR2inquadrantsubregions,thoseregionsinitially withtheleastChR2activatedareacontinuedtoshowonaveragetheleastChR2activatedareaover30minutesFigure18A-C.Similarly,uponsequentialChR2activationofsinglecell-sizedregions,thosecellsthatinitiallyshowednegligibleactivationof[Ca 2+ ]continuedtoshowsignicantlylowerChR2activatedareacomparedto thosecellsthatinitiallyshowedhighactivationof[Ca 2+ ]Figure18D. Figure18. ChR2activationisconsistentthroughtimeIRepeatedactivationof onequadrantat0,10,20,30minutesshowsJIntra-isletvariationinChR2activated[Ca 2+ ]asinE,attime0andafter30minutesofrepeatedChR2stimulation,sortedfromleasttomostbasedontime0measurements.KAsinI,forminimumandmaximumquadrantsat0,10,20,30minutesduringChR2stimulation. LChR2activatedareauponsingle-cellactivationattime0andafter10minutes ofrepeatedChR2stimulation,sortedbasedontime0measurements.Scalebarindicates100 m ThedurationoflightexposureinuencestheamountofChR2activatedareaFigure20.whenilluminating25%oftheislet,shorteractivationsof100msresulted in[Ca 2+ ]ielevatinginlessoftheisletcomparedtothe`standard'pulsedurations of1sasusedabove;whereaslongerpulsedurationsof10sshowednodifference tothepulsedurationsof1s.DespitethedifferenceinChR2activatedareawhen shorterpulseswereused,therewasaconsistentrelationshipbetweenChR2acti54

PAGE 67

vatedareawithpulselength.Greaterthan80%ofinsulinpositivecellsexpressed ChR2-YFP,whereaslessthan10%ofglucagonpositivecellsandlessthan10%of non-glucagon/insulinpositivecellsexpressedChR2-YFPFigure19B,C.TheseregionsofvaryingChR2activated[Ca 2+ ]iwereindependentofexperimentalparametersincludingstimulationorder,positionwithinthemicroscopeeldofview.Although ChR2-YFPFluorescencewasspatiallyvariablenotshownChR2-YFPexpression uorescencedidnotplayaroleinthisregionalvariabilityFigure20A,C,E.Regions oftheisletthatshowedagreaterChR2-activatedareaunderstandard1spulsedurationsalsoshowedagreaterChR2-activatedareaundershorter100mspulsedurationsFigure20H,indicatingthattheintra-isletvariabilityinChR2-activatedareais independentoftheChR2activationprotocol.Thereforewithinislets,atbasalglucose thereexistsub-regionsoftheisletoverdifferentspatialscalesthatupondepolarizationcaneffectivelyrecruitneighboringregionsoftheislettoshowelevated[Ca 2+ ]i; andthisisindependentofexperimentalprotocolsandoccursoverseveralspatial scalesfromisletsub-regionstosinglecell-sizedregions. 55

PAGE 68

Figure19. QuanticationofChR2-YFPExpressioninisletsAImagewithinintact isletofplasmamembranelabeledwithFM4-64FXleft;imagewithinintactisletof ChR2-YFPdistributionmiddle;togetherwithmergedimageright.Expressionof ChR2-YFPishighthroughouttheislet.BRepresentativeimmunouorescenceimagesofinsulintop,left,YFPtop,middle,andmergedimagetop,right;orrepresentativeimmunouorescenceimagesofglucagonbottom,left,YFPbottom, middle,andmergedimagebottom,right,allindissociatedcellsfromtheislet. CQuanticationofpercentageinsulinpositivecells,glucagonpositivecellsorinsulinandglucagonnegativecellsthatexpressChR2-YFPleft;andquanticationof allcellsdapipositivethatareinsulinpositive,glucagonpositiveorareinsulinand glucagonnegativeright. 56

PAGE 69

Figure20. Electricalexcitabilityshowsnotemporal/spatialbiasingAChR2activatedarearelativetotheisletaveragesortedaccordingtothetemporalorderof ChR2quadrantactivation.BNADPHresponserelativetotheisletaveragesorted accordingtothetemporalorderofChR2quadrantactivation.CChR2activated arearelativetotheisletaveragebinnedaccordingtothespatialpositionintheeld ofviewofChR2quadrantactivation.DNADPHresponserelativetotheisletaveragebinnedaccordingtothespatialpositionintheeldofviewofChR2quadrant activation.EScatterplotandlinearregression+/-95%condenceintervalsforthe ChR2activatedareawithinanisletagainsttheYFPuorescenceaveragedacross theregionofactivation.FRepresentativetime-coursesof[Ca 2+ ]iwithdifferentdurationsofChR2activationlong=10s,medium=1s,short=100ms.GQuantication ofChR2activatedareaasafunctionofChR2activationduration.HScatterplotof ChR2activatedareainquadrantswithmediumpulseilluminationtimes1svsshort illuminationtimes100ms.DatainA-Daveragedovern=25islets,datainE,n=27 quadrantsandG,H,n=8quadrants. 57

PAGE 70

ActivationofnonChR2-YFPexpressingisletsdidnotleadtosubstantialareaswhere ouranalysisalgorithmdetectedactivationFigure21indicatingthatonlywithChR2YFPexpressiondoweobservesignicantinuxesof[Ca 2+ ]consistentwithprevious ndingsReinbotheetal.,2014. Figure21. ActivationofnonChR2expressingislets.IsletsexpressiononlyPDX-Cre didnotshowsignicantareasofactivationwhenstimulatingthewholeislet. Beta-cellmetabolicactivitycontrolsvariationsinChR2stimulated[Ca 2+ ]i: Single -cellsdissociatedfromtheisletareheterogeneousintheactivityanddynamicsofmanyfactorsunderlyinginsulinsecretionPipeleersetal.,1994;Pipeleers, 1992.Isletregionscontaining -cellswithincreasedexcitabilitycouldresultinless ChR2-mediateddepolarizationrequiredtodepolarizeneighboringregionsandelevating[Ca 2+ ]i.Totestwhetherregionsofaltered -cellfunctionaffecttheChR2activatedareawemeasuredglucose-metabolismviatwo-photonimagingofNADPH, alongsideChR2activationand[Ca 2+ ]iimaging;whereheterogeneityin -cellglucosemetabolismhasbeenreportedPistonetal.,1999.Toexaminevariationsin intra-isletNADPHresponses,werstsortedregionsbyascendinglevelsofNADPH activitysuchthateachislethadaregionofminimumandmaximummetabolicactiv58

PAGE 71

ity.NADPHlevelsover2mM,5mM,11mMglucosewererelativelyconsistentbetweenisletsandquadrantregionsofeachisletFigure.22A.However,therewere signicantvariationsintheNADPHresponsefrom5mMto11mMglucosewithin eachFigure.22B. 59

PAGE 72

Figure22. MetabolicDependenceonChR2ActivationANADPHlevelsat2mM, 5mM,11mMglucosebyquadrants,sportedinascendingrankorder.BNADPH responsebetween5and11mMglucose,inquadrantregionsoftheisletdenedby ChR2activationregions,sortedinrankorder.CIntra-isletvariationinNADPHresponserelativetotheisletaverage.TheNADPHresponseovereachquadrantas inBwasexpressedrelativetothemeanNADPHresponseoverthewholeislet, andsortedfromleasttomost.DChR2activatedarearelativetotheisletaverage resultingfromquadrantregionsofactivationat2mMglucose,foreachcorrespondingrank-orderedNADPHlevelovertheregionat2mMglucose.EChR2activated arearelativetotheisletaverageresultingfromquadrantregionsofactivationat2mM glucose,foreachcorrespondingrank-orderedNADPHresponseovertheregion between2mMand5mMglucose.FAsinDforChR2activatedareaat2mMglucose,forcorrespondingNADPHresponsebetween2mMand11mMglucose.G AsinDforChR2activatedareaat5mMglucose,forcorrespondingNADPHlevel at5mMglucose.HAsinDforChR2activatedareaat5mMglucose,forcorrespondingNADPHresponsebetween2mMand5mMglucose.INADPHlevel at5mMglucoseforsinglecellactivationregionsinwhichChR2activatedareawas low < 0.1,high > 0.1.JAsinHforNADPHresponsebetween2mMand5mM glucose.Dataisdisplayedasmean s.e.mDataina-caveragedovern=25islets from6mice;dataind-haveragedovern=10isletsfrom3mice;dataini,javeraged overn=96regions,26isletsfrom5mice.****indicatesp < 0.0001comparingexperimentalgroupsindicated.StudentsT-TestwasusedinI,Jtocalculatestatisticalsignificance. 60

PAGE 73

Onaveragewithinanislet,thedifferencebetweenthequadrant-regionofhighest andlowestNADPHresponsewas35% 5%Figure22C.WecomparedquadrantregionsofNADPHlevelsandresponsewithChR2activatedareaat2mMglucose and5mMglucose.ChR2activatedareaat2mMglucosedidnotvarysignicantly withNADPHlevelsat2mMglucosenordiditvarywiththeNADPHresponsefrom 2mMto5mMglucoseorfrom2mMto11mMglucoseFigure22D-F.ChR2activated areaat5mMglucosedidnotvarywithNADPHlevelsat5mMglucosenorwiththe NADPHresponsefrom2mMto5mMglucoseFigure.22I,J.However,itdidvary signicantlyandsubstantiallywiththeNADPHresponsefrom5mMto11mMglucoseFigure23B:RegionswiththelowestNADPHresponsebetween5mMand 11mMglucosehadaChR2activatedarea17%lessthantheisletaverage,whereas regionofhighestNADPHresponseshowingaChR2activatedarea14%greater thantheisletaverage.Thisindicatesthatatbasalglucosespatialheterogeneityin glucosemetabolismwithinisletsleadstospatialvariationsinthecontrolofelectricalactivity.WealsotestedifthelinkbetweenChR2activatedareaandNADHPH wasconsistentwithinsinglecellregions.CellularregionsinwhichtheChR2activatedareawasabovethemedianarea%showedsignicantlygreaterNADPH responsefrom5mMto11mMglucosethancellularregionsinwhichtheChR2activatedareawasbelowthemedianareaFigure23.Thus,singlecelllocationsshowinghighChR2activatedareaalsohadcorrespondinglyhigherNADPHresponses. TodeterminetheinuenceofvaryingKATPactivity,weperformedsimilarmeasurementswiththeadditionofdiazoxide,aKATPactivator.Inthepresenceofdiazox61

PAGE 74

idegreaterlaserpowerwasneededtoelevate[Ca 2+ ]iandachieveasimilarChR2activatedarea.Despitethis,at5mMglucosethevariationinChR2activatedareastill variedsignicantlywiththeNADPHresponsefrom5mMto11mMglucoseFigure 23D,asintheabsenceofdiazoxide. Figure23. Spatialvariationsinmetabolicactivitycontrolelectricalactivity.A RepresentativeimagesofNADPHautouorescenceatglucoselevelsindicate,or changeinNADPHautouorescencebetweenindicatedglucoselevels.BMean s.e.m.ChR2activatedareaat5mMglucoseforcorrespondingrankorderquadrantsofNADPHresponsebetween5mMand11mMglucose.CMean s.e.m NADPHresponsebetween5mMand11mMglucoseforsinglecellactivationregionsinwhichChR2activatedareawaslow < 0.1,high > 0.1.DAsinBfor ChR2activatedareaat5mMglucoseplus150 Mdiazoxide,forcorresponding NADPHresponsebetween5mMand11mMglucoseplus150 Mdiazoxide.Data inBaveragedovern=25isletsfrom6mice;datainCaveragedovern=96regions, 26isletsfrom5mice;datainDaveragedovern=10isletsfrom3mice. indicates p < 0.05, indicatesp < 0.01.Scalebarindicates100 m 62

PAGE 75

MulticellularmodellinksmetabolicactivityandChR2stimulated[Ca 2+ ] VariationsinChR2-activated[Ca 2+ ]icorrelatedwithvariationsinmetabolicactivity, overarangeofspatialscales. -cellsareintrinsicallyheterogeneousinexcitability, howevertheorganizationofthisheterogeneitywithintheisletandhowitmayimpact functionisunknown.Weaskedwhetherspecicspatialdistributionsof -cellexcitabilitywererequired,orwhetherrandomdistributionsweresufcienttodescribe ourexperimentalobservations.Weimplementedamulticellularisletmodelincorporating -cellheterogeneity,whichhaspreviouslydescribedhowthebalanceof gapjunctioncouplingandKATP-regulated -cellexcitabilitycontrolsisletfunction Hrahaetal.,2014c.WealsoincludedafourstateChR2currentmoduleNikolic etal.,2009,andrepresentedmetabolicheterogeneitybyvaryingglucokinaseactivityGKwithadistributionthatmatchedthatobservedexperimentallyforNADPH Pattersonetal.,2000.Themetabolicactivityofthe -cellisdeterminedbytheux ofGKactivityandthusglycolysisdescribedby: J glyc = k glyc 1 1+ K G [ G ] 2 : 5 1 1+ K mATP [ ATP ] [ Re tot )]TJ/F18 11.9552 Tf 13.748 0 Td [([ Re ] Where[G]istheglucoseconcentration, K glyc isGKactivitymodeledwithGaussian heterogeneityseeTable1,[ATP]istheconcentrationofATP,and[Re tot ]and[Re] arethetotalpyrimidinenucleotideandNADHconcentrationrespectively.Werstdividedthesimulatedisletintospatialpartitions,eachcontaining200cellswitharandomlydenedsubsetofthevariationinmetabolicactivityglucokinaseactivityFig63

PAGE 76

ure24A.Atsuboscillatoryglucoseconditions,quadrant-regionswereactivatedby ChR2,followedbyanincreaseto11mMglucosetoinduce[Ca 2+ ]ioscillationsFigure24B.Activatingquadrant-regionswithinthesimulatedisletledto[Ca 2+ ]irises withintheactivationwindowandwithsignicantintra-isletvariationintheChR2activationareaFigure24C,consistentwithexperimentalalbeitwithlessvariabilityFigure24D.Simulatedmeasurementsofglucokinaseactivityoverquadrantregions showedintra-isletvariationthatwascomparabletothevariationinexperimentallymeasuredNADPH,wherethedifferencebetweenthequadrant-regionofhighest andlowestglucokinaseactivitywas36% 2%Figure24E.At2mMglucose,the ChR2activatedareadidnotvarysignicantlywithmetabolicactivityFigure24F, consistentwithexperimentalmeasurementsFigure.22D-F.However,at5mMglucosetheChR2activatedareadidvarysignicantlywithmetabolicactivityFigure 24G,consistentwithexperimentalmeasurementsFigure23B;wherequadrantregionswithminimumglucokinaseactivityshowedthelowestChR2activatedarea andregionswithmaximumglucokinaseactivityshowedthehighestChR2activated area. 64

PAGE 77

Figure24. Spatialdomainsinanisletmodelrecapitulatesex-vivoexcitabilityA False-colormapshowingcellularglucokinaserateoverasimulatedislet,wheresubregionsofsimilarheterogeneityisapplied.BRepresentativetime-courseof[Ca 2+ ] in5cellsofasimulatedisletfollowingChR2protocol.CChR2activatedarea,in termsofnumberofcellsactivated,byquadrantinrankorder.[Ca 2+ ]=170nMwas usedasacutoffforactive/inactivecells.DIntra-isletvariationinChR2activated arearelativetothesimulatedisletaverage.TherelativeareaofChR2-activated [Ca 2+ ]overeachquadrantwasexpressedrelativetothemeanChR2-activated [Ca 2+ ]ofeachsimulatedislet,andsortedfromleasttomost.EIntra-isletvariationinGCKraterelativetotheisletaverage.FChR2activatedarearelativetothe isletaverageat2mMglucose,foreachcorrespondingquadrantGCKrate.GChR2 activatedarearelativetotheisletaverageat5mMglucose,foreachcorresponding quadrantGCKrateat5mMglucose.HChR2activatedarearelativetotheisletaverageat5mMglucose,foreachcorrespondingascendingrank-orderedquadrant GCKrateundertheadditionof10%openprobabilitytoKATPchannelsinallcells. IAsinA,whererandomspatialdistributionofheterogeneityisapplied.JAsinD, forsimulatedisletswithrandomlydistributedheterogeneity.KAsinG,forsimulated isletswithrandomlydistributedheterogeneity.DatainC-H,J-Kaveragedovern=30. 65

PAGE 78

At5mMglucose,withtheadditionofa10%KATPchannelopenprobabilitytomodel diazoxideapplicationHrahaetal.,2014c,theChR2activatedareastillvariedsignificantlywithmetabolicactivityFigure24H,againconsistentwithexperimentalmeasurementsFigure23D.Totesttherequirementforheterogeneitytobedistributed intospatialdomainsofexcitabilitywedistributedvariationsinglucokinaseactivity throughouttheisletwithoutanyspatialorganizationFigure24I.However,therewas noconsistentrelationshipbetweenvariationsinChR2activated[Ca 2+ ]iandvariationsinglucokinaseactivityFigure24J,K.Thissupportsthatcellularheterogeneity hassomespatialorganization,ratherthanbeingapurelyrandomdistribution.Membranedepolarizationcontrols[Ca 2+ ]ielevation.Weexaminedwhethermetabolicheterogeneityhadasimilarimpactinexertingvaryingdepolarizationinneighboringcells aswithexertingvarying[Ca 2+ ]ielevationsFigure25A.Heterogeneityinmembrane depolarizationwasobservedunderasimilarChR2protocol,andthisshowedsimilarcorrespondencewithmetabolicactivity.NotablyregionswithgreaterChR2activatedareaalsoshowedahigherrestingmembranepotential,consistentwithgreater metabolicactivity,ATPproductionandKATPchannelclosureFigure25B-D. 66

PAGE 79

Figure25. RestingMembranePotentialLinksCellularExcitabilitytoChR2ActivationARepresentativetimecoursesofmembranepotentialfrom4cellswithinasimulatedisletwithspatiallyorderedmetabolicactivityasinFigure24.BChR2activatedareaofmembranedepolarizationfroma,foreachquadrantoftheisletinascendingrankorder.Vm=-40mVwasusedasacutoffbetweenactiveandinactive cellduringChR2activation.CChR2activatedarearelativetotheisletaverage,for eachquadrantofascendingGCKratemetabolicactivity.DRestingmembrane potentialat5mMglucose,foreachcorrespondingascendingrank-orderquadrantof ChR2activatedarea.DatainB-Eaverageovern=30simulatedislets.EChR2activatedarearelativetotheisletaverageinquadrantsofwaveoriginandwaveend. SpatialorganizationtoNADPHandChR2responses Inordertotestwhethermetabolicactivityisspatiallyorganizedwithintheislet,as predictedbytheisletmodel,werstcalculatedtheabsolutedifferencesintheNADPH responseasafunctionofseparationdistance.Forseveralbaselinereferencepoints withinanislet,thechangeinNADPHbetween5mMand11mMglucosewasaveragedoverexpandingareasofradius rFigure26A.Theresultant` NADPH`responsecurveincreasedslowlyasafunctionofradialdistanceFigure26B,indicatingthatareasoftheisletdistantfromareferencepointshowedonaverageamore 67

PAGE 80

differentNADPHresponsethanareasclosetothereferencepoint.ReplicateanalysisofsimulatedisletswiththetwospatialdistributionsFigure24A,Ishowedasimilarslowincreasingresponsecurveforthespatiallyorganizeddistribution,whereas arandomorganizationshowednosignicantchangewithdistanceFigure26C. Thesendingswerereplicatedbyusingpairwisedifferencesineithersmallareas ofNADPHresponseforexperimentalmeasurementsFigure24D,EorcellularglucokinaseactivityforsimulationsFigure24F.Thespatialscaleoftheexperimentally measuredNADPHresponsematchedthespatiallyorderedheterogeneityinglucosemetabolisminsimulatedisletsFigure26B-F,supportingthepredictionthat subregionsofmetabolicactivityexistthatcanimpactthespatialvariationsinexcitability.TotestwhetherthereexistsregionsofsimilarexcitabilitywecomputedpairwisedifferencesintheChR2activatedareabetweeneachsinglecellregionwithin anisletFigure26G.Thesepairwisedifferencesweresortedbythedistancebetweentheregions r.Pairwisedifferencesincloser r < 50 mregionsweresignicantlylessdifferentthanmoredistant r > 50 mregionsFigure26H.Similar resultswereobserveduponcomputingpairwisedifferencesintheChR2activated areabetweeneachsinglecellregionwithinthesimulatedisletFigure26I.Therefore,cellsincloserproximityshowmoresimilarChR2-activaitonof[Ca 2+ ]consistent withtheisletmodelandindicatingexcitabilityisalsospatiallyordered. 68

PAGE 81

Figure26. SpatialanalysisofNADPHandChR2-activated[Ca 2+ ]responsesA RepresentativeNADPHintensityimage,displayinghowspatialcorrelationsinthe NADPHresponsewerequantied.The NADPHresponsewascalculatedby theabsolutedifferencewastakenbetweenaninitialNADPHresponsemeasurementpurple,andtheNADPHresponseaveragedoveranoutwardlyexpanding regionblackcircleswithradius r.B NADPHresponseasafunctionofradial distance rnormalizedtothe NADPHatthelargest r.CEquivalentanalysis asthatinBcalculatedfromGCKratesinsimulatedisletswithrandomspatialdistributionofGCKrategreyorsubregionsofsimilarGCKratesblack.DRepresentativeimageshowinghowspatialdifferencesinNADPHresponsewerequantied.TheabsolutedifferencesinNADPHresponsewerecalculatedforseveral smallregionspurplecirclesseparatedbydistance r,inapairwisemanner.E MeanabsolutedifferencesinNADPHresponsebetweenregionsasafunctionof spatialseparation rnormalizedtothedifferenceinresponseatmaximalseparation.FEquivalentanalysisasthatinEcalculatedfromGCKratesinsimulated isletswithrandomspatialdistributionofGCKrateorsubregionsofsimilarGCKrates. DashedlinesinB,C,E,Findicateadistanceof6cells.GRepresentativemapof ChR2-activated[Ca 2+ ]iinanisletat5mMglucoseorange,blue,whereactivation regionsareappliedindividuallyoverasingle-cell.HPairwiseabsolutedifferences inChR2activated[Ca 2+ ]areaaveragedbyclose r < 50 mandfar r > 50 m differences;andnormalizedtotheaverageChR2activatedareaforeachislet.IAs inHcalculatedfromsinglecellChR2-activated[Ca 2+ ]insimulatedisletswithsubregionsofsimilarorrandomGCKrates.StudentsT-TestwasusedinH-I. 69

PAGE 82

Waveorigincorrelateswithlowermetabolicactivityandexcitability Coordinated[Ca 2+ ]ioscillationswithinisletsatelevatedglucosearesynchronized bypropagatingwaveswhichoriginateindenedregionsoftheisletBenningeretal., 2014.Wehypothesizedthattheobservedspatiallyorderedmetabolicheterogeneitymaycontrolthese[Ca 2+ ]iwavedynamicsatelevatedglucose.Toinvestigatehow heterogeneityin -cellfunctioncorrelateswithpropagatingcalciumwaves,wemeasured[Ca 2+ ]ioscillationsandwavepropagationalongsideChR2-activaitonandNADPH measurementsorCx36permeabilitymeasurements.Usingphaseanalysissee methods,propagatingwavesof[Ca 2+ ]ielevationwereobservedthathadaconsistentspatialoriginandpropagatedacrosstheisletin2secondsFigure27A,B, consistentwithpriormeasurementsBenningeretal.,2008,2014.Interestingly,the ChR2activatedareaat5mMglucoseinthequadrantofthewaveoriginwassignificantlylessthantheChR2activatedareainthequadrantofthewaveend,witha differenceof31% 12%Figure27C.Consistentwiththis,theNADPHresponse from5mM-11mMglucoseintheregionofthewaveoriginwassignicantlylessthan theNADPHresponseatthewaveendandthewholeisletaverageFigure27D. Furthermore,the[Ca 2+ ]ioscillationamplitudeatthewaveoriginwaslessthanat thewaveendFigure27E.Therefore,surprisingly,spatialheterogeneityinglucose metabolismwithintheisletleadstosub-regionsoflowermetabolicactivityandexcitabilitythatappeartocontroltheoriginofpropagatingcalciumwaves. 70

PAGE 83

Figure27. Calciumwaveorigincorrespondstolessexcitableandmetabolicallyactiveregionsin ex-vivo and in-silicoislets ARepresentativephasemapof[Ca 2+ ] oscillationswithinanislet,ascalculatedthroughFourieranalysis,whichindicatesa waveemerginginregionofminimumphasedarkblueandterminatinginaregionof maximumphasedarkred.BRepresentativetime-coursesfromA,showingphase lagof[Ca 2+ ]waves.CChR2activatedareaof[Ca 2+ ]elevationrelativetotheislet average,inquadrantofwaveoriginandwaveend.DNADPHresponserelative totheisletaverage,inselectedregionsofminimumandmaximumphase,asindicatedinA.E[Ca 2+ ]oscillationamplitudenormalizedtoaverageRhod-2AMuorescenceinquadrantsofwaveoriginandwaveend.FRepresentativefalsecolor mapof[Ca 2+ ]waveinsimulatedislet,asinA.GRepresentativetime-coursesfrom F,showingphaselagof[Ca 2+ ]waves.HChR2activatedareainquadrantsofwave originandwaveendinsimulatedislet,asinC.IGCKrateinquadrantsofwaveoriginandwaveendinsimulatedislet,asinD.J[Ca 2+ ]oscillationamplitudeinquadrantsofwaveoriginandwaveendinsimulatedislet,asinE.DatainC-Eaveraged overn=16isletsfrom4mice.DatainH-Javeragedovern=30simulatedislets.Verticalscalebarsindicate2%changeinuorescenceor100nMchangein[Ca 2+ ]iin experimentorsimulatedislets,respectively.Scalebarindicates100 m 71

PAGE 84

Intrinsicoscillatoryfrequencycontrolscalciumwavepropagation Theexperimentalobservationandmodelagreementthatcalciumwavesoriginate inregionsoflowmetabolicactivityisnotimmediatelyintuitive.Toexaminethelink betweenmetabolicactivityand[Ca 2+ ]ioscillationfrequencyweanalyzedcellsin isletsfromCx36 )]TJ/F22 7.9701 Tf 7.593 0 Td [(= )]TJ/F36 11.9552 Tf 11.416 -4.338 Td [(mice,whichlackgapjunctionconductanceand[Ca 2+ ]ioscillationsynchronizationRavieretal.,2005;Benningeretal.,2008Figure28A.There wasasignicantnegativecorrelationbetweencellular[Ca 2+ ]ioscillatoryfrequency andNADPHresponseFigure28B,wherecellswithhighermetabolicactivityhad slower[Ca 2+ ]oscillations.UpontreatingCx36-/-isletswiththeglucokinaseinhibitor mannoheptuloseweobservedasignicantincreaseinthemeanoscillatoryfrequency ofthosecellsthatremainedactivefrom24mHz 2mHzto32mHz 4mHzFigure 28C.Thereforedecreasedmetabolicactivityincreases[Ca 2+ ]ioscillationfrequency. 72

PAGE 85

Figure28. OscillatoryfrequencyismodulatedbymetabolicactivityARepresentativefalsecolormapof -cell[Ca 2+ ]ioscillatoryfrequencyinaCx36-/-islet.BRelationshipbetweenNADPHresponsebetween5mMand11mMglucoseand[Ca 2+ ] oscillationfrequencyofindividual -cellswithinCx36-/-islets,togetherwithlinear regression+/-95%condenceintervals.CMean s.e.m.[Ca 2+ ]oscillationfrequencybefore-andafter+additionof5mMD-mannoheptuloseinindividual cellswithinCx36-/-islets.DRepresentativetime-coursesof2uncoupled -cellsin theisletmodelwithhighorlowGCKrate.EPearsoncorrelationcoefcientforthe effectofheterogeneityintheindicatedparameterson[Ca 2+ ]oscillationfrequency withintheisletmodel.FNaturaloscillationfrequencyof -cellsaveragedoverregionscorrespondingtowaveoriginandwaveend,inelectricallyuncoupledisletsthat aresimulatedtohavesubregionsofsimilarGCKrates.Scalebarindicates100 m. Toexaminethelinkbetweenmetabolicactivity,oscillationfrequencyandwaveoriginintheisletmodelweexaminedhowtheparametersusedtorepresentcellular heterogeneityimpactedthenaturaloscillationfrequencyinelectricallyuncoupled cells.Therateconstantofglucokinaserepresentingglucosemetabolismshowedthe highestabsolutenegativecorrelationwithoscillationfrequencyFigure28D,E.This isconsistentwithlessmetabolicallyactiveareashavingahigherintrinsicoscillation 73

PAGE 86

frequency,inaccordancewithexperimentalmeasurements.Intheabsenceofcouplingwithspatiallyorganizedmetabolicactivity,cellscorrespondingtothewaveoriginshowedsignicantlyfasteroscillationfrequenciesthancellscorrespondingtothe regionofthewaveendFigure28F.Toexaminetherelationshipbetweenoscillation frequencyandwaveorigininageneralizedfashionweusedamodelofKuramoto oscillatorswhichhasbeenusedtostudythesynchronizationofoscillatorysubunits inchemical,biological,andphysicalsystems,38.Weassigneduniquenatural oscillationfrequenciestocellularunitsaccordingtoaGaussiandistributionanddistributedtheseintospatialpartitionsofsimilarnaturalfrequenciesFigure28G.The coupledsystemconvergedfrequencywithspatially-dependentphaselagsbetween oscillatorsFigure28H.TheaveragenaturaloscillationfrequencyinregionsofsignaloriginwassignicantlyhigherthaninregionsofsignalendFigure28I,indicating regionsofhighernaturaloscillatoryfrequencycontrolthedirectioninwhichsignals propagateinageneralizedoscillatormodel. 74

PAGE 87

CytokineMediatedDisruptionofElectricalActivity Intype2diabetesandobesitythereisanup-regulationofadipocytederivedproinammatorycytokinesBoden,2008.Previousstudieshaveshownthatpro-inammatory cytokinecocktailsdisrupt[Ca 2+ ]oscillationsinpancreaticisletsDulaetal.,2010 anddecreaseCx36functionFarnsworthetal.,2015a.Wetestedhowapro-inammatory cytokinecocktailsIL-1 .5ng/ml,TNFng/ml,INFng/mldisruptsspatialexcitability.Isletsweretreatedfor1hourintheconcentrationsabovetoinduce acutedysfunctionFarnsworthetal.,2015b.WerstexaminedifcoordinatedoscillationsweredisruptedincytokinetreatedisletsFigure29A-C.Althoughoscillations appearedtobeslightlymoreirregularincytokinetreatedisletsFigure29Bthan thecontrolisletsFigure29Awedidnotobservesignicantdecreasestoelectrical activityorcoordinationwithcytokinetreatmentFigure29C-D.ChR2-YFPexpressionwasnotdisruptedwithacutecytokinetreatmentFigure29Enortheabilityfor ChR2toshowglobal[Ca 2+ ]inuxuponactivationFigure29F.Indicatingthatany changestospatialexcitabilitywouldnotbeduetoChR2function. 75

PAGE 88

Figure29. AcuteTreatmentofPro-inammatoryonChR2FunctionATop:WT isletsshowcoordinatedoscillationsacrosstheislet.Aftertreatmentfor1hrwithproinammatorycytokinecocktailbottomweobservecoordinated,butslightlydysfunctionaloscillations.B-ECytokinetreatmentdoesnotinhibit[Ca 2+ ]activity,electrical coordination,CHR2expression,orChR2function.Scalebarindicates100 m. Wenexttestedhowspatialexcitabilitywasdisruptedwithacutecytokinetreatment Figure30.Therewasasignicantreduction 60%inthemetabolicNADPHresponsebetween5mMand11mMglucosewithcytokinetreatmentFigure30A.UsingquadrantChR2stimulationprotocolswherethemeanactivatedareawas21 2%, 76

PAGE 89

howeverwithcytokinetreatmentthemeanquadrantactivatedareadecreasedto 0.15 2%additionallytheamplitudeassociatedwithChR2activationdecreasedby 36 8%.WeexaminedthedistributionofChR2activatedareainsinglecellactivation domainsFigure30D. Figure30. CytokinetreatmentdisruptelectricalexcitabilityAMetabolicresponse between5mMand11mMglucoseincytokinetreatedislets.BActivationofquadrantsizeregionsincytokinetreatedisletsandamplitudeCofChR2activated [Ca 2+ ]inux.DDistributionofsingle-cellsizedactivationregionsinuntreatedand cytokinetreatedislets. 77

PAGE 90

InuntreatedisletsweobservealargedistributionofChR2activatedarea.Indicating somesinglecellregionsshowhighercontrolofsurroundingregions.Uponcytokine treatmentthereisasignicantshiftinthisdistributionandhigh-controllingsinglecell domainsarenolongerobserved. NextwedeterminedtheeffectofacutecytokinetreatmentontheabilityofChR2 activationofsinglecellregionstoentrainglobalisletdynamicsFigure31. Figure31. CytokinetreatmentdisruptelectricalexcitabilityAActivationofsinglecellsizedregionsat11mMglucoseandBglobalisletresponseat10sactivation intervals.CEntrainmentofhigheroscillatoryfrequenciesinuntreatedandcytokine treatedisletsat11mMglucose.Scalebarindicates100 m. At11mMglucosewhereisletsshowcoordinatedoscillatory[Ca 2+ ].Activationof singlecelldomainsentrainelectricaldynamicsgloballythroughouttheisletFigure 31A-C.WithAdditionofcytokinetreatmentsinglecellregionslosetheabilitytocontrolwhole-isletelectricaldynamicsFigure31C. 78

PAGE 91

Conclusions IsletsofLangerhansshowcomplexmulti-cellularregulationthroughtheelectrical couplingbetweenfunctionallyheterogeneouscells.Howdifferentsub-populationsof cellsaffectoverallfunctionispoorlyunderstood.Todissectthefunctionalarchitectureoftheisletweappliedoptogenetics,quantitativeuorescencemicroscopyand computermodeling,todetectfunctionalsub-populationsofcellscrosstheisletand characterizehowtheycontroloverallcellularexcitabilityanddynamics.Wediscoveredthat:1metabolicallyactivesub-populationsofcellsexertgreatercontrolover theelectricalresponseofneighboringcells;2cellularheterogeneitywithrespect toglucosemetabolismandexcitabilityisspatiallyorganized;3metabolicallylessactivecellsthatshowgreaterintrinsicoscillationfrequencyactas`pacemakers'todirectcalciumwavepropagationandsynchronizeoscillations;4modelingdysfunction conditionsassociatedwithdiabetesdecreasescellularcontrolofneighboringcells. Metabolicallyactivesub-populationsofcellscontrolexcitability Heterogeneous -cellfunctionhasbeenobservedin -cellsdissociatedfromthe islet;andpriorstudieshavePipeleers,1992;Jornsetal.,1999suggested -cell heterogeneityimpactsisletfunctionBenningeretal.,2014;Kinardetal.,1999;Sher etal.,2003.Examiningtheroleofheterogeneityintheintactisletischallenging duetothehighcoordinationoftheelectricalresponse.However,byusingChR2 andsequentialstimulationofindividualcellsorsub-regions,wecouldidentifyheterogeneouselectricalresponsesintheintactislet.Thereexistsub-regionsorsubpopulationsofcellsthataremoreabletorecruitsurroundingcellstobecomeactive 79

PAGE 92

Figure17.ThiscontrolisindependentofexperimentalparametersFigure20indicatinganintrinsicpropertyofthecellthatcancontrolexcitability.Furthermore,islets inwhichoscillationswereentrainedat11mMglucosereturnedtopre-entrainement activitypostChR2activationFigure31BindicatingthatprolongedChR2activation didnotshowmajorchangestofunctionduetoopticalheatingofcells.Notably,inthe intactislet50%ofcellsareunabletorecruitothercellstobeactiveat5mMglucose, indicating50%ofcellsshowalowelectricalresponse.Thisissimilartoobservationsof[Ca 2+ ]iresponsesinisletslackingCx36andelectricalcoupling,where50% ofcellsat5mMglucoseshowno[Ca 2+ ]iresponseBenningeretal.,2011b.Amajor questioniswhatfactorsunderliethisheterogeneityintheintactislet.Weobserved thatregionsofhighorlowelectricalcontrolarecorrelatedrespectivelywithregions ofhighorlowmetabolicactivity.Theheterogeneityinmetaboliccontrolofelectrical activityhasastrongeffectasitpersistsunderdiazoxideapplicationFigure23.This suggestsheterogeneityinfactorssuchasKATPdensity,whoseeffectwouldbeexpectedtobeincreasedrelativetometabolicactivityunderdiazoxideapplication,have reducedeffect,whichisalsosupportedbymodellingresultsFigure24.Arecent studyusingeNpHr3,analternativeoptogenetictooltosilencepopulationsofcells, alsodemonstratedthereexistssub-populationsofcellswithintheintactisletthatare importantformaintainingelectricalcontrolJohnstonetal.,2016b.Notablythese cellshadelevatedglucokinaselevelssuggestingincreasedmetabolicactivity,which isconsistentwithourresults.However,animportantdifferenceisthatourstudyindicatesmetabolicallyactivesub-populationsofcellsthatexertelectricalcontroldo 80

PAGE 93

notcorrespondto`pacemakers'thatinitiatewavepropagationseebelow.Through usinganNCAMmarker,subsetsof -cells highand lowhavealsobeenidentiedthatshowdifferential[Ca 2+ ]iandATPelevationresponsestoglucoseKaraca etal.,2009. lowcellsexpressedGLUT2andGCKsignicantlylowerthan high cells,whichisalsosimilartothelessmetabolicallyactivesub-populationsweobservethatareunabletorecruitneighboringcellstoelevate[Ca 2+ ]i.Therelationship weobservedatbasalmMglucoseconditionswasalsopredictedbyourin-silico modelFigure24F.AtlowermMglucoselevels,wedidnotobservearelationship betweenheterogeneityinmetabolicactivityandelectricalcontrolbothexperimentally andintheisletmodel,suggestingthatcontroloverexcitabilitymaybedominatedby otherfactors,whichrequiresfurtherinvestigation. Spatialorganizationofcellheterogeneity AnalysisofNADPHresponsesandourcomputermodelresultssupporttheexistenceofspatialorganization,andweobservedconsistentresultsbetweentheorganizationofmetabolicactivityandChR2stimulated[Ca 2+ ]Figure19.Wealsoobservedregionsofvariabilitythatexistoverarangeofspatialscalesfromquadrants oftheislettosinglecellregionsFigure26.Thisisconsistentwithdetailedexaminationof[Ca 2+ ]idynamicsrevealingtheislettoobey`SmallWorldnetwork'principlesSto zeretal.,2013;Markovi cetal.,2015;wheredynamicsaremoresimilarin localized-regionsoftheislet.Aremainingquestionishowthesedomainsarise?One possibilityisthatcellswithinadomainoriginatefromthesameprogenitor.Indeed, clustersof -cellsintheisletoriginatefromacommonprogenitorsuggestingthat 81

PAGE 94

propertiesoftheprogenitormayaffectpropertiesofoffspring -cellsDesgrazand Herrera,2009.Anotherpossibilityisoneofself-organization,wheretheexcitability ofonecellmayimpactthatofothercells.Therefore,moreexcitablecellsthatoriginaterandomlymayleadtoneighboringcellstobecomemoreexcitable.Theability of -cellfunctionandcouplingtoimpacttheorganizationofother -cellsintheislet hasbeenpredicted,anddiffersintype2diabetesStriegeletal.,2015.AnalpossibilityisthatextrinsicfactorssuchasbloodowarchitectureorinnervationmaypatternorfunctionallysegmenttheisletintofunctionalregionsRodriguez-Diazetal., 2011;Nymanetal.,2010.Severalstudieshaveidentiedsub-populationsof -cells andtheirmolecularbasisKatsutaetal.,2012;Karacaetal.,2009;Dorrelletal., 2016;Baderetal.,2016,whichincludesheterogeneityingenesunderlyingglucose metabolism,electricalregulationandgapjunctioncoupling.Ourndingssuggest these -cellsub-populationswillhavespatialorganizationandbepreferentiallyorderedtogether.Asdiscussedabove,lowNCAMsurfaceexpressionmarksasubset of -cells lowthatshowreduced[Ca 2+ ]iandATPelevationtoglucoseKaraca etal.,2009.Inaratmodeloftype2diabetestheproportion low-cellsincreased. SimilarlytheWnteffectortpmarksproliferativesub-populationsofcellsthatpreferentiallyexpandinobesityorpregnancy,andwhichhavereducedmetabolicactivity andCx36expressionBaderetal.,2016.Theplasticityoftheislettovarybetween thesesub-populationsunderpathologicalconditionswouldimpacthowdepolarizationspropagatethroughouttheislettoregulate[Ca 2+ ]iandinsulinsecretion.Islets areelectricallyquiescentbelow6mMglucoseandtransitiontobeingelectricallyac82

PAGE 95

tiveabovethislevel.Wepredictthatiflessmetabolicallyactive,inexcitablecellsare spatiallygroupedandthispopulationgrows,thepresenceofislet-widedepolarization willbecompromisedtoimpactinsulinsecretionandglucosehomeostasis.Nevertheless,theprecisespatialorganizationandtheroleofsub-populationsinisletfunction andhormonereleaseduringconditionsofdiabetes,aswellasindevelopmentand pregnancystillremainstobefullydetermined. Lowermetabolic,highfrequencycellsinitiatepropagatingcalciumwaves Synchronized[Ca 2+ ]ioscillationswithinisletsenhancerstphaseandsecondphase pulsatileinsulinreleaseandinsulinactionMeieretal.,2005;Sto zeretal.,2013; Nunemakeretal.,2005.Propagatingcalciumwavesmediatethissynchronization andconsistentlyemergefromsub-regionsoftheisletBenningeretal.,2008.We experimentallyobservedthattheseregionsofwaveinitiationareregionsoflower metabolicactivityFigure22andfasternaturaloscillationfrequency.Inmodelsof pulsedcoupledoscillators,thefastestoscillatorsetsthepaceMirolloandStrogatz, 1990.Priormodelingstudiespredictedthatwaveinitiationoccursinregionswith highermetabolicactivityandfasternaturaloscillationfrequencyBenningeretal., 2014.ThesestudiesusedamodelwhereoscillationfrequencyincreaseswithincreasedglucosemetabolismBertrametal.,2000.Weshowedexperimentallyin electricallyuncoupledcellsthatahigheroscillationfrequencyisobservedinsubpopulationsofcellsthatshowalowerNADPHresponse,andfollowingmetabolic inhibitionFigure27.Thuslowerlevelsofglucosemetabolismthataresufcientto elevate[Ca 2+ ]resultinhigheroscillationfrequencies,andthisisfurtherobservedin 83

PAGE 96

ourisletmodelFigure28Chaetal.,2011a.WhileCx36inhibitorsexist,theyare weakandnotspecic;thereforewecannotacutelydecouple -cellstodetermine withinthesameisletifcellsatthewaveoriginhaveintrinsicallyhigheroscillatory frequency.Nevertheless,byimplementingageneralizedKuramotooscillatormodel Breakspearetal.,2010andthroughexaminingourisletmodelFigure27,we showthatwaveinitiationisdrivenbyspatiallocationsofhighnaturaloscillatoryfrequencywhichphysiologicallyisdeterminedbyalower-than-averagemetabolicactivityFigure28.Intheisletmodel,thelinkbetweenmetabolicactivityandintrinsicoscillationfrequencywasthestrongest,furthersuggestingglucosemetabolismmaybe thedominatingfactorinaffecting[Ca 2+ ]ioscillationfrequency.Analysisofthe -cell modelhasshownthatKATPclosureandslowCaVinactivationdeterminestheterminationofanoscillationandthusoscillationfrequency,withERCa2+bufferingalso playingaroleChaetal.,2011a,b.KATP,CaVandSERCAconductance/activity areallATPdependent.Thushighglucosemetabolismwillreducethetermination ofoscillationsbyCaVandKATP,andincreasetheoscillationperiodandlowerthe frequency.Thepresenceofintrinsicmetabolicoscillationshasalsobeensuggested Renetal.,2013,whichmayenhancepacemakingaction.Furtherworkisneeded todetermineotherdynamicaleffectsofspatialdomains.Forexample,inasystem ofKuramotooscillatorswhereoscillatorsofsimilarfrequencyarespatiallygrouped, thenetworksynchronizationisoptimalunderweakercouplingregimescompared tosystemswithrandomspatialassignmentoffrequencyFreitasetal.,2015,and arecentmodellingstudyusingasimilarcoupled -cellmodelshowedthatgreater 84

PAGE 97

heterogeneityamongst -cellsincreasedsynchronizationMontaseriandMeyerHermann,2016andcellswithgreateroscillatoryfrequencymaypromotesynchronizationJalanetal.,2015.ThusisletdysfunctionfollowingdisruptionstoCx36gap junctioncoupling,whichcanoccurinseveralconditionslinkedtothedevelopment ofdiabetesRavieretal.,2005;Carvalhoetal.,2012b;Farnsworthetal.,2015a, maybeprotectedbyspatiallygroupingcellsofsimilarexcitabilityandoscillationfrequency. CytokineMediatedDisruptionofElectricalActivity Acutetreatmentusingthepro-inammatorycytokinecocktaildecreasesCx36functioninisletsandprolongedtreatmentsignicantlydisruptsinsulinsecretionandcoordinated[Ca 2+ ]activity.Farnsworthetal.,2015b.Herewecharacterizewhateffect acutetreatmenthasondecreasingtheabilityforsubpopulationstocontrolelectrical activityanddynamicswithinislets.AsresultsfromFigure29indicatewehavenosignicantlydisruptedcoordinated[Ca 2+ ]oscillationsnortheabilityforisletstobegloballyactivatedwithChR2stimulation.However,wedoobservesignifcantdecreases toelectricalexcitabilityFigure30AandacorrespondingdecreasetoquandrantactivatedareaFigure30BindicatingthatthelossofametabolicresponsemaybeincreasingthehyperpolarizedstateandthuslimitingtheextenttowhichChR2stimulatedactionpotentialsmaypropagate.Theseresultsaresimilarlyconrmedinahigh gluocsestatewhere[Ca 2+ ]oscillationsarepresentFigure31.While[Ca 2+ ]coordinationhasnotbeensignicantlydisruptedincytokinetreatedislets,smallregions arenotlongercapableofentrainingoscillatorybehaviorFigure31Cindicatingthat 85

PAGE 98

eveninanacutelydysfunctionalstatethereisalossofcellularcontroloverislet-wide dynamics. 86

PAGE 99

CHAPTERV AGEASSOCIATEDDECLINEINELECTRICALSYNCHRONIZATION 3 Introduction Thereisasignicantdeclineinglucosetoleranceandincreaseintheriskoftype2 diabeteswithadvancingageinhumansKarveandHayward,2010a.ReducedglucosetoleranceinageingresultsfrombothreducedinsulinsensitivityandreducedinsulinsecretioninresponsetoelevatedglucoseNathanetal.,2007;Gumbineretal., 1989;Basuetal.,2003;ChangandHalter,2003. -cellproliferationandproliferativecapacitydeclineswithage,whichcanpartlyexplainthereducedinsulinsecretion andriskfortype2diabetesKushner,2013;GunasekaranandGannon,2011.However -cellfunctionisalsoaltereduponageingIozzoetal.,1999;Ihmetal.,2006. -cellsrespondtoglucosebyelevatedmitochondrialrespirationandATPgeneration;membranedepolarizationandactionpotentialgeneration;increasedintracellularfree-calcium[Ca 2+ ];andinsulingranuleexocytosis.Severalofthesesteps, includingmitochondrialrespirationGreggetal.,2016;Helmanetal.,2016;Lietal., 2014ATPgenerationGreggetal.,2016;Helmanetal.,2016,and[Ca 2+ ]handling Greggetal.,2016;Lietal.,2014;Avrahamietal.,2015arealteredin -cellsfrom agedmiceandhumans.However,therehavebeenconictingresults.Forexample inhumanislets,improvedinsulinsecretionuponageingorinsenescent -cellshas 3 Theresultsinthischapterwereacceptedtopublicationon6/1/2017:Westacott, M.J.,Farnsworth,N.L.,St.Clair,J.R.,Poffenberger,G.,Heintz,A.,Hart.,N.J.,Powers,A.C.,Benninger,R.K.P.,Age-dependentdeclineinthecoordinatedCa2+and insulinsecretorydynamicsinhumanpancreaticislets. Diabetes 87

PAGE 100

beenreportedHelmanetal.,2016,yetothershavereporteddeclinesininsulinsecretionwithageIhmetal.,2006;Greggetal.,2016;Lietal.,2014;Fritscheetal., 2002.Differencesbetweenmouseandhuman -cellsresponsestoageinghave alsobeenreportedGreggetal.,2016;Avrahamietal.,2015.Thereforemuchremainstobeunderstoodregardinghow -cellfunctionisaltereduponageing. -cells withintheisletsofLangerhansdonotfunctionautonomously.Thereisextensive communicationbetween -cellsandwithothercelltypesBenningeretal.,2011b; Konstantinovaetal.,2007;vanderMeulenetal.,2015;Lernmark,1974;Hashimoto etal.,2014;Rodriguez-Diazetal.,2012thatisimportantforhow -cellsfunction withintheislet.Gapjunctionsformedfromconnexin36Cx36electricallycouple cellsRavieretal.,2005;Benningeretal.,2008;Serre-Beinieretal.,2009which coordinatestheoscillatory[Ca 2+ ]responsetoelevatedglucoseandregulatesthe dynamicsofinsulinsecretionRavieretal.,2005;Benningeretal.,2008.Thisincludesenhancingtherstphaseandsecondphasepulses,wherealossofCx36 gapjunctioncouplingleadstoglucoseintoleranceHeadetal.,2012b.Notablygap junctioncouplingandcoordinated[Ca 2+ ]dynamicsaredisruptedinmodelsofobesityortype2diabetesCarvalhoetal.,2012b;Hodsonetal.,2013;Farnsworthetal., 2015a;Ravieretal.,2002suggestingaroleinthepathogenesisofdiabetesRutter andHodson,2013.However,theeffectofageingonintra-isletcommunicationand theregulationofinsulinsecretionhasbeenpoorlyexamined.Thisincludesexamininghowgapjunctioncouplingandcoordinated[Ca 2+ ]maybeaffectedbyageing. Furthermore,theroleofgapjunctioncouplingandcoordinated[Ca 2+ ]inhumanislet 88

PAGE 101

functionhasbeenless-wellcharacterizedcomparedtorodentislets.Hereweexaminetheintra-isletregulationof -cell[Ca 2+ ]andinsulinsecretioninisletsfromlarge numbersofnon-diabetichumandonors;aswellasinyoung-adultandagedmice. Weexaminehowgapjunctioncouplingandcoordinated[Ca 2+ ]aredisruptedbyageing;howthisimpactstheregulationofinsulinsecretion;andhowgapjunctionmodulationimpactschangesin[Ca 2+ ]andinsulinsecretion.Theresultsfromthisstudy providefurtherevidenceforadeclineinisletfunctionwithageinhumans,whichmay contributetoincreasedriskoftype2diabetes. Results QuanticationofElectricalActivityandCoordination Werstexamined[Ca 2+ ]activityinWTmice,Cx36 )]TJ/F22 7.9701 Tf 7.594 0 Td [(= )]TJ/F36 11.9552 Tf 11.415 -4.338 Td [(mice,andhumanisletsin ordertoconrmthatcustomMATLABroutinescouldproperlyidentifytheanticipated high[Ca 2+ ]coordinationinWTmouseislets,lowcoordinationinCx36-/-isletsand totest,ingeneral,how[Ca 2+ ]activityandcoordinationscaledinhumanisletsFigure32.IsletsfromWTmiceshowedfullycoordinated[Ca 2+ ]activityandcoordinationLeftImageA,DconsistentwithpriorstudiesBenningeretal.,2008.Islets fromCx36 )]TJ/F22 7.9701 Tf 7.593 0 Td [(= )]TJ/F36 11.9552 Tf 11.416 -4.338 Td [(miceshowed[Ca 2+ ]activitythatwasslightlylessthanfromWTmice, however[Ca 2+ ]activitylackedanycoordinatedareasFigure32A,rightimage,D. Thesedataindicatethatour[Ca 2+ ]activityandcoordinationalgorithmsproduceexpectedresults.Wethenexamined[Ca 2+ ]activityandcoordinationfrom54human donorsbetween2013and2016.Weobservedthat[Ca 2+ ]coordinationinsomehumandonorisletsbehavedsimilarlytoWTmouseisletsFigure32Bcoordinatedos89

PAGE 102

cillationsacrossthemajorityoftheislet.Theseobservationswereararityhowever, withonlyapproximately8%ofthedonorsshowinghighlevelsof[Ca 2+ ]coordination.Isletsfromtheremainingdonorsshowedlow[Ca 2+ ]activityandcoordination regionsFigure32C,LeftorhigherlevelsofactivitythatwererestrictedtosmalldomainsFigure32C,Right.Quanticationof[Ca 2+ ]activityacrossalldonorsshowed that63 4%oftheareashowed[Ca 2+ ]activityFigure32D,whichisconsistentwith theknownfractionof -cellsinhumanisletsCabreraetal.,2006. 90

PAGE 103

Figure32. [Ca 2+ ]CoordinationinHumanIsletsARepresentativefalsecolormap of[Ca 2+ ]activityandcoordinationinC57B6WTandCx36-/-isletsleft,withtimecoursesfromfourindividualcellsindicatedintheseisletsright.[Ca 2+ ]activityis representedbypresenceoffalsecolor,witheachcolorrepresentingaseparateregionof[Ca 2+ ]coordination,asindicatedinlegendabove.BRepresentative[Ca 2+ ] activityandcoordinationmapsinhumanisletsfromdonorswherehighcoordination similartomouseisletsisobservedleft,asinA,withtime-coursesfromfourindividualcellsindicatedintheseisletsright.CRepresentative[Ca 2+ ]activityandcoordinationmapsinhumanisletsfromdonorswherelowcoordinationisobservedleft, asinA,withtime-coursesfromfourindividualcellsindicatedintheseisletsright. DAreaof[Ca 2+ ]activitynormalizedtoisletsize`Activearea'averagedoverislets fromWTmice,Cx36-/-mice,andallhumandonors.ELargestareaofcoordinated [Ca 2+ ]activitynormalizedtoisletsize`Max.coordinatedarea'averagedoverislets fromWTmice,Cx36-/-mice,andallhumandonors.FAbsolutelargestareaofcoordinated[Ca 2+ ]activityplottedasafunctionofisletsizeleft,orbinnedbyisletsize right.GLargestareaofcoordinated[Ca 2+ ]activitynormalizedtoisletsizeplotted asafunctionofisletsizeleft,orbinnedbyisletsizeright.DatainpanelsD-Gis displayedasmean s.e.m,averagedovern=4WTmice,3Cx36-/-mice,andn=40 humandonors.VerticalscalebarsinA-Crightindicate20%uorescencechange. Scalebarindicates100 m. 91

PAGE 104

Themeansizeof[Ca 2+ ]coordinationexpressedasthelargestfractionareaMax CoordinatedAreawhichshowedcoordinatedactivitywas31 4%ofthetotalarea 36cellsFigure32E.ThiswassignicantlylessthanthecoordinatedareaofWT mice,butgreaterthanthecoordinatedareaofCx36 )]TJ/F22 7.9701 Tf 7.594 0 Td [(= )]TJ/F36 11.9552 Tf 11.415 -4.338 Td [(miceindicatingelectrical couplingispresent.Wetestedifthe[Ca 2+ ]maximumcoordinatedareashowedislet sizedependenceFigure32Fhoweveritdidnotvarysignicantlythereforewhen normalizedtoisletsizeFigure32Gshowedsomesizedependence. AgePredictsDeclinein[Ca 2+ ]CoordinationinHumanIslets Wenexttestedhow[Ca 2+ ]activityandcoordinationchangewithdonorage.[Ca 2+ ] activitydecreasedmodestlyasafunctionofageFigure33A,Left.Wethenbinned donorsaccordingtoagegreaterthanorlessthan40yearsold.[Ca 2+ ]activityin theyoungergroupwasslightlyhigherthantheoldergroup:71% 4%to54% 4% Figure33ARight.Coordinated[Ca 2+ ]activity,normalizedtoisletareaandunnormalized,showedsignicantdecreaseswithageFigure33B,C,Leftandsimilarlywhensplitbyage,theoldergrouphadreducedcoordinationcomparedtothe youngergroup24 3%,3874 485 m 2 26cellscomparedto40 4%,7437 1111 m 2 50cellsFigure33BandFigure33C. 92

PAGE 105

Figure33. [Ca 2+ ]coordinationAge-dependentdeclineininhumanislets.AArea of[Ca 2+ ]activity`activearea'normalizedtoisletsizeasafunctionofdonorage leftandaveragedoverdonorslessthanblackorgreaterthanredthemedian ageof40yearsright.BAreaofcoordinated[Ca 2+ ]activitynormalizedtoislet sizeasafunctionofdonorageleftandaveragedoverdonorslessthanorgreater thanthemedianage,asinAright.CAbsoluteareaofcoordinated[Ca 2+ ]activityasafunctionofdonorageleftandaveragedoverdonorslessthanorgreater thanthemedianage,asinAright.D[Ca 2+ ]oscillationdutycycleplateaufractionoflargestcoordinatedareaasafunctionofdonorageleftandaveragedover donorslessthanorgreaterthanthemedianage,asinAright.Inleftpanelseach datapointrepresentsasingledonor,withoutliersROUTtestindicatedbyempty circles.Solidlineindicateslinearregression,dashedlinesindicate95%condence intervals,andp-valuesindicatethesignicanceofacorrelation.Inrightpanelsdata aredisplayedasmean s.e.m.averagedovern=40donors,withp-valuesindicating thesignicanceofdifferencesbetweenindicatedgroupsStudentst-test. Consistentwiththesedata,therewasasignicantcorrelationbetween[Ca 2+ ]activityandthecoordinationof[Ca 2+ ]dynamicsovereachdonorFigure34A.Themean sizeofisletsstudieddidnotvarywithageFigure34B,excludinganisletsizeeffect. Themeandutycycleplateaufractionofthelargestcoordinatedareaineachdonor isletdidnotshowanyvariationwithage,whereitremainedat43 2%Figure33D. 93

PAGE 106

Figure34. Furthercharacterizationof[Ca 2+ ]dynamicsinhumanisletswithage A[Ca 2+ ]coordinationasafunctionof[Ca 2+ ]activitymeasuredinisletsfromeach donor.BMeanisletsizeasafunctionofdonorageleftandaveragedoverdonors lessthanblackorgreaterthanredthemedianageright.C[Ca 2+ ]oscillation periodinlargestcoordinatedareaasafunctionofdonorageleftandaveragedover donorslessthanblackorgreaterthanredthemedianageright.ForpanelsB-C lefteachdatapointrepresentsasingledonor,withoutliersROUTtestindicated byemptycircles;solidlineindicateslinearregression,dashedlinesindicate95% condenceintervals,andp-valuesindicatethesignicanceofacorrelation.ForpanelsB-Crightdataaredisplayedasmean s.e.m.averagedovern=40donors,with p-valuesindicatingthesignicanceofdifferencesbetweenindicatedgroupsStudentst-test 94

PAGE 107

WenextmeasuredhowhumanisletCx36gapjunctioncouplingvariedwithage. Cx36gapjunctioncoupling,asmeasuredbyFRAP,showedadecreasewithincreased ageFigure35A,.Whensegmentingbyage,theolderagegrouphadsignicantly reducedCx36gapjunctioncouplingcomparedtotheyoungergroupFigure35A. DonorsfromisletswhichshowedincreasedCx36gapjunctioncouplingalsoshowed increased[Ca 2+ ]coordinationFigure35B,.Thosedonorswithisletgapjunction couplingabovethemedianvalueshowedasignicant2-foldincreasein[Ca 2+ ]coordinationcomparedtodonorswithisletgapjunctioncouplingbelowthemedian valueFigure35B.Thisindicatesincreasedgapjunctionfunctioncontributesto largerareasof[Ca 2+ ]coordination.Thereforethereexistsanage-dependentdeclineinhumanislet[Ca 2+ ]coordinationasaresultofanage-dependentdeclinein gapjunctioncoupling. Age-dependentdeclineininsulinsecretion Cx36gapjunctioncouplingandcoordinated[Ca 2+ ]responsesareimportantfor regulatinginsulinsecretiondynamicRavieretal.,2005;Headetal.,2012a;Speier etal.,2007.TheresponsetimeforinsulinsecretiontoreverttobasallevelsafterloweringofglucoseismarkedlysloweruponlossofCx36inmiceSpeieretal.,2007. Howeverinsulinsecretionatelevatedglucosedoesnotsignicantlyvaryuponloss ofCx36inmiceBenningeretal.,2011b.Therefore,wemeasuredthedeclinekineticsforhumanisletsfrom80donorsbetween2012-2016duringperifusionmeasurementsfollowingelevatedglucoseandglucose+IBMXFigure36A. 95

PAGE 108

Figure35. Age-dependentdeclineinCx36gapjunctioncouplinginhumanisletsA Cx36gapjunctionfunction,asassessedthroughFRAPrateofuorescencerecoveryasafunctionofdonorageleftandaveragedoverdonorslessthanblackor greaterthanredthemedianageof40yearsright.BAbsoluteareaofcoordinated[Ca 2+ ]activityasafunctionofCx36gapjunctionfunctioninisletsfromeach donorleftorinisletsofeachdonorsthatshowslowerthanmedianorgreaterthan medianCx36gapjunctionfunctionright.Inleftpanelseachdatapointrepresentsa singledonor.Solidlineindicateslinearregression,dashedlinesindicate95%condenceintervals,andp-valuesindicatethesignicanceofacorrelation.Inrightpanelsdataaredisplayedasmean s.e.m.averagedovern=32donors,withp-values indicatingthesignicanceofdifferencesbetweenindicatedgroupsStudentst-test. Therewasnosignicantvariationinthefold-changeininsulinsecretionuponelevatedglucosewithageFigure36B,aspreviouslyreportedKaytonetal.,2015. AdditionofIBMXampliedthefold-changeininsulinsecretion,Figure36C,which 96

PAGE 109

wasindependentofage.Afterglucoseelevation,therateofdeclineininsulinsecretionfollowingresumptionofbasalglucoseshowedasignicantreductionwithage, indicatingaslowerresponsivenessoftheislettoglucosechangesFigure36D. Thesendingsshowingalackofsecretionlevelsbutalteredsecretionkineticsare consistentwiththatofreducedgapjunctioncouplingoccurringwithage.Incontrast, afterglucoseelevationwithIBMX,therateofdeclineininsulinsecretionfollowingresumptionofbasalglucoseshowednosignicantchangewithageFigure36E.In isletsfromyoungerdonors < 40years,therateofdeclineafterglucosestimulation showednodifferencewithandwithoutIBMXFigure36E.Howeverinisletsfrom olderdonors,therateofdeclineafterglucosestimulationshowedasignicantincreasewithIBMXFigure36E,totherateobservedinyoungerdonors.ThusIBMX canacutelyrecovertheage-dependentdeclineininsulinsecretiondynamics,consistentwithrecoveringanage-dependentdeclineingapjunctioncoupling.Wefurtherexaminedtheimpactofageoninsulinsecretionlevelsthroughstaticassays, forwhichage-dependentdeclinesinsecretionhavebeenreportedIhmetal.,2006; Greggetal.,2016;Lietal.,2014;Fritscheetal.,2002. 97

PAGE 110

Figure36. Age-dependentdeclineininsulinsecretiondynamicsARepresentative perifusiontime-courseoftwohumandonorsunderindicatedglucosestimulationprotocol.Thedeclinefollowing16.7mMglucoseor16.7mMglucose+IBMXisttedtoan exponentialdecaymodelgrey.BFold-changeinsulinsecretionbetween5.6mM and16.7mMasafunctionofdonorage.CFold-changeinsulinsecretioninthe presenceandabsenceofIBMX,averagedoverdonorsgreaterthanredorlessthan blackthemedianageof40years.DExponentialdecayrateofinsulinsecretion followingtransitionfrom16.7mMglucoseto5.6mMglucoseasafunctionofdonor age.EExponentialdecayrateofinsulinsecretionfollowingtransitionform16.7mM glucoseto5.6mMglucoseinthepresenceandabsenceofIBMX,averagedover donorsgreaterthanredorlessthanblackthemedianage.FordatainB,Deach datapointrepresentsasingledonor,solidlineindicateslinearregression,dashed linesindicate95%condenceintervals,andp-valuesindicatethesignicanceofa correlation.FordatainC,Edataaredisplayedasmean s.e.m.averagedovern=76 donors,withp-valuesindicatingthesignicanceofdifferencesbetweenindicated groupsStudentst-test. 98

PAGE 111

Insulinsecretionat2mMglucosedecreasedslightlywithageFigure37A.Insulin secretionat20mMglucosesignicantlydeclinedwithageFigure37B,whichwhen segmentedshoweda42%declineinsecretionindonors > 40yearscomparedto < 40years.However,therewasnochangeinthestimulationindexwithageFigure 37C.InsulincontentalsodidnotvarywithageFigure37D,andthereforedidnot explainthedeclineininsulinsecretion.Elevationsin[Ca 2+ ]triggerinsulinsecretion, andweobservedmarkedchangesin[Ca 2+ ]withageFigure33. Figure37. AgeDependentDeclinetoInsulinSecretionAInsulinsecretionat2mM glucose,asmeasuredbystaticassays,asafunctionofdonorageleftandaveragedoverdonorslessthanblackorgreaterthanredthemedianageof40years right.BInsulinsecretionat20mMglucoseasafunctionofdonorageleftand averagedoverdonorslessthanorgreaterthanthemedianageright,asinA. CStimulationindexfoldchangebetween2and20mMglucoseasafunctionof donorageleftandaveragedoverdonorslessthanorgreaterthanthemedianage right,asinA.DInsulincontentasafunctionofdonorageleftandaveragedover donorslessthanorgreaterthanthemedianageright,asinA.InA-Dleftpanels eachdatapointrepresentsasingledonor,withoutliersROUTtest,basedonpanel Bindicatedbyemptycircles.Solidlineindicateslinearregression,dashedlinesindicate95%condenceintervals,andp-valuesindicatethesignicanceofacorrelation.InA-Drightpanelsdataaredisplayedasmean s.e.m.averagedovern=43 donors,withp-valuesindicatingthesignicanceofdifferencesbetweenindicated groupsStudentst-test. 99

PAGE 112

[Ca 2+ ]activitycorrelatedwithinsulinsecretionat20mMglucoseFigure38A.Howeverneither[Ca 2+ ]coordinationnorCx36gapjunctionfunctioncorrelatedwithinsulinsecretionFigure38BandFigure38C. Figure38. Ca2+ActivityPredictsDeclineinInsulinSecretionAInsulinsecretionat 20mMglucoseasafunctionof[Ca 2+ ]activitymeasuredinisletsfromeachdonor. BInsulinsecretionat20mMglucoseasafunctionof[Ca 2+ ]coordinationmeasuredinisletsfromeachdonor.CInsulinsecretionat20mMglucoseasafunction ofCx36gapjunctionfunctionmeasuredinisletsfromeachdonor. Thislackofcorrelationbetweeninsulinsecretionlevelsunderstaticassaysand [Ca 2+ ]coordinationorgapjunctioncouplingarestillconsistentwithreducedgap junctioncouplingoccurringwithage,21,24Benningeretal.,2011b;Ravieretal., 2005;Headetal.,2012b;andsuggestthedeclineininsulinsecretionlevelsarises fromsomeotherfactor. Wenextcomparedisletsfromasubsetofdonorsinwhichinsulinsecretionmeasurementsviaperifusionandstaticassayswerebothperformed,aswellas[Ca 2+ ] imagingmeasurementsFigureS7.AmongcomparisonsofinsulinsecretionFigure S7A-C,thestrongestcorrelationwasobservedbetweenmeasurementsofSecretion Indexviaperifusionandstaticassays,albeitwithoutstatisticalsignicanceFigure S7A,p=0.14.Therewasnostatisticalsignicanceobservedcomparingperifusion 100

PAGE 113

measurementsand[Ca 2+ ]activitybypeakp=0.26,andbyAUCp=0.27.Among comparisonsofthesecretionrise/decayrateduringperifusionwith[Ca 2+ ]coordinationFigureS7F-I,thestrongestcorrelationwasbetweenthedecayrateand[Ca 2+ ] coordinationFigureS7H,p=0.06. Cx36gapjunctionactivationrecoversage-dependent[Ca 2+ ]decline AcuteIBMXtreatmentreversedage-dependentdeclinesininsulinsecretiondynamics,whereIBMXelevatesmouseisletCx36gapjunctioncoupling.ModanilactivatesCx36gapjunctioncouplinginisletsandtheCNS,37.Wetestedifacute ModaniltreatmentcouldincreaseCx36functionandrescueage-dependentdeclinesin[Ca 2+ ]coordination.TreatmentwithModanilfor1hsignicantlyincreased [Ca 2+ ]activityinisletsfromolder > 40yearsdonorsFigure39A,andslightlyincreasedactivityinisletsfromyounger < 40yearsdonors.[Ca 2+ ]coordinationwas alsosignicantlyincreasedinisletsfromolderdonors;asmeasuredbyboththesize ofthelargestcoordinatedareanormalizedbyisletsizeFigure39BandtheabsolutesizeofthelargestcoordinatedareaFigure39C.HowevernosignicantincreasewasobservedforeachmeasurementinisletsfromyoungerdonorsFigure 39B,C.Notably,Modanilincreasedthe[Ca 2+ ]coordinationinisletsfromolderdonors tothelevelinisletsfromyoungerdonorsFigure39B,C.SimilarlyModanilsignificantlyelevatedgapjunctioncouplinginisletsfromolderdonorsFigure39D,but showednosignicantincreaseinisletsfromyoungerdonors.Thusacutegapjunctionactivationrecoverstheage-dependentdeclinein[Ca 2+ ]coordinationandgap junctioncouplinginisletsfromolderdonors. 101

PAGE 114

Figure39. Modanilrecoverstheage-dependentdeclineinCx36functionand [Ca 2+ ]coordinationA[Ca 2+ ]activity`Activearea'inthepresenceandabsence ofmodanilhincubation,averagedoverdonorsgreaterthanredorlessthan blackthemedianageof40years.BAreaofcoordinated[Ca 2+ ]activitynormalizedtoisletsizeinthepresenceandabsenceofmodanilhincubation,averaged overdonorsgreaterthanorlessthanthemedianage,asinA.CAbsoluteareaof coordinated[Ca 2+ ]activityinthepresenceandabsenceofmodanilhincubation,averagedoverdonorsgreaterthanorlessthanthemedianage,asinA.D Cx36gapjunctionfunction,asassessedthroughFRAPrateofuorescencerecoveryinthepresenceandabsenceofmodanilhincubation,averagedover donorsgreaterthanorlessthanthemedianage,asinA.Dataispresentedas mean s.e.m,averagedovern=8 < 40andn=9 > 40donors,withp-valuesindicatingthesignicanceofdifferencesbetweenindicatedgroupsStudentst-test. 102

PAGE 115

[Ca 2+ ]andCx36gapjunctionfunctionispreservedinagedmice Wenexttestedifmouseisletsshowsimilardeclinesingapjunctionfunctionand [Ca 2+ ]withageingbycomparingisletsfrom24montholdC57/B6agedmicewith3 montholdyoung-adultmice,throughsimilarassays.Isletsfromagedmiceshowed dramaticallydifferentoscillationpatternsthanthoseofyoungermiceFigure40. Thiswascharacterizedbyanincreased[Ca 2+ ]dutycycleplateaufraction20% increase,Figure40Bandasubstantiallysloweroscillationperiod 21saged comparedto51 3syoung-adult,Figure40C.Howevertherewasnosignicant changeintheproportionoftheisletshowing[Ca 2+ ]activityFigure40D,[Ca 2+ ] coordinationFigure40E,ForCx36functionFigure40G.Thereforeincontrastto human,gapjunctionmediated[Ca 2+ ]activityandcoordinationispreservedduring ageinginmouse. 103

PAGE 116

Figure40. PreservationofCx36gapjunctionfunctionandCa2+coordinationin agedmiceARepresentativetime-coursesof[Ca 2+ ]averagedoverislets,each from3montholdblackand24montholdredC57Bl/6mice.Scalebarindicates 20%uorescencechange.B[Ca 2+ ]oscillationdutycycleplateaufractionin isletsfrom3monthblackand24monthredmice.C[Ca 2+ ]oscillationperiod inisletsfrom3monthand24monthmice,asinB.DAreaof[Ca 2+ ]activity`activearea'normalizedtoisletsizeinisletsfrom3monthand24monthmice,asinB. EAreaofcoordinated[Ca 2+ ]activitynormalizedtoisletsizeinisletsfrom3month and24monthmice,asinB.FAbsoluteareaofcoordinated[Ca 2+ ]activityinislets from3monthand24monthmice,asinB.GCx36gapjunctionfunction,asassessedthroughFRAPrateofuorescencerecoveryinisletsfrom3monthand24 monthmice,asinB.DatainB-Gispresentedasmean s.e.m.averagedovern=3, 3montholdmiceandn=4,24montholdmice. 104

PAGE 117

Conclusions Inthisstudywecharacterizedthecoordinated[Ca 2+ ]dynamicsinhumanislets, andtestedhowageingimpactsislet[Ca 2+ ].Ourkeyndingsarethatthecoordinated[Ca 2+ ]responseatelevatedglucoseisdisruptedinhumanisletsfromaged donors.Thisdisruptioncorrelateswithreducedgapjunctioncouplinganddisruption toinsulinsecretiondynamics.Notablythisage-dependentdisruptioncanberecoveredbyactivatorsofgapjunctioncoupling.Incontrast,nodisruptionwasobserved inisletsfromagedmice. Humanislet[Ca 2+ ]coordinationisrestrictedtosub-populationsofcells:Despite widevariability,wegenerallyobserved[Ca 2+ ]tobecoordinatedoveralimitedrange inhumanislets;whereasimilarmaximalnumberofcellswerecoordinatedirrespectiveofisletsizeFigure1.Thisisconsistentwithprior[Ca 2+ ]measurementsinhumanisletswhere[Ca 2+ ]oscillationswerereportedoversub-regionsoftheisletCabreraetal.,2006;Quesadaetal.,2006;andisalsoconsistentwithhistologicalstudiesCabreraetal.,2006;Kilimniketal.,2011.Theabsenceofsignicantincreases inCx36gapjunctioncouplingor[Ca 2+ ]coordinationinyoungerdonorsupongap junctionactivationFigure6indicatesanupperlimittothecoordinationachievablein humanislets.Interestinglyhumanisletarchitecturehasbeensuggestedtoconsistof folded2D`sheets'of -cellsand -cells:ourndingsaresimilartomeasurementsof [Ca 2+ ]in2Dclustersof -cellswherecoordinatedactivityisrestrictedtosub-regions Hrahaetal.,2014a.Overall,humanislet[Ca 2+ ]coordinationisconsistentwiththe 105

PAGE 118

electricalcouplingbetween -cellsbeinggovernedby`small-world'principles,where electricalconnectionsbetweensmallnumbersofcellsdominatebehavior,aspreviouslyshowninmouseisletsSto zeretal.,2013;Johnstonetal.,2016a. Age-dependentdeclineingapjunctioncouplingdisrupts[Ca 2+ ]coordinationand insulinsecretiondynamicsinhumans:Cx36gapjunctionchannelscoordinate[Ca 2+ ] dynamicswithinmouseislets.Thereforethedeclineincoordinated[Ca 2+ ]dynamicswithageFigure2likelyresultsfromthereducedgapjunctioncouplingwithage Figure3.Thisissupportedbytherecoveryincoordinated[Ca 2+ ]dynamicsby gapjunctionactivationFigure6.Thedeclineingapjunctioncouplingwemeasure 25%between20-60yearsislessthanthelossin[Ca 2+ ]coordination60%between20-60years.Inmouseislets,modellingstudiespredictthatbeyondacertain levelofgapjunctioncoupling,subsequentsmalldecreasesincouplingcancause largedisruptionsto[Ca 2+ ]coordinationBenningeretal.,2008;Hrahaetal.,2014a. Thearchitectureofhumanisletsresultsinfewer -cellneighbors,whichmayexplainourresultsandsuggestsagreatersusceptibilityofhumanisletsto[Ca 2+ ]disruptionuponasmallerlossofgapjunctioncoupling.Recoveryfromtheage-dependent declineincoordinated[Ca 2+ ]oscillations,andinsulinsecretiondynamicsinhuman isletswereachievedacutelybytheactivatorsofgapjunctioncoupling,Modanilor IBMX.Asthisrecoveryoccurredin < 1hhours,thisindicatestheage-dependentdeclinetoelectricalcoordinationinhumanisletsisunlikelytobeduetotranscriptional changes.Inhumanislets,Cx36transcriptionincreasesandCx43decreasesdur106

PAGE 119

ingmaturity < 2comparedto > 18years,asinmouseisletsCarvalhoetal.,2010, butthereisnosignicantchangeinCx36orCx43transcriptionduringageing40comparedto > 40years[K.Kaestner,personalcommunication].Therefore,the age-dependentdeclineingapjunctioncouplinglikelyresultsfrompost-translational regulation,andmayincludealteredphosphorylationwhichreducesCx36gapjunctioncouplinginothersystemsUrscheletal.,2006;Ivanovaetal.,2015.Alossof gapjunctioncouplingandcoordinated[Ca 2+ ]dynamicsdisruptspulsatileinsulin secretiondynamicsandglucosetoleranceHeadetal.,2012b,withthedeclinein insulinsecretionfollowingglucosereductionbeingmarkedlyslowerHeadetal., 2012b;Speieretal.,2007.Wewereunabletomeasureinsulinsecretiondynamicsfromsingleisletstoassesspulsatility.Howeverperifusionresultsdemonstrating aslowerdeclineininsulinsecretionindicateareducedglucose-responsivenessof insulinsecretion,andthussuggestsasignicantdisruptiontopulseamplitudemay alsobepresent.Indeedadisruptiontosecretionpulseamplitudein-vivo,bothonthe timescaleofsecondphasepulsatilityaswellasultradianpulseshasbeenshownto occurduringageingChangandHalter,2003.Thiscouldcontributetothereduced glucosetolerancethatoccursinagedhumans.Theseage-dependentchangesto intra-isletcommunicationaresummarizedinFigure8.Wealsoobservedsubstantialdifferencesinage-dependent[Ca 2+ ]changesbetweenhumanandmouseislets. Priorstudieshavealsoshowndifferencesinhumanisletscomparedtomouseislets duringageing,includingalackofcompensatorychangesinKATPdensityanda declineinexpressionofgenesassociatedwithisletfunctionGreggetal.,2016; 107

PAGE 120

Avrahamietal.,2015.Thisindicatesabroaderdifferenceinhowmouseandhumanisletsrespondtoageing.Themicewestudiedwerelean,whereas50%of donorsfromwhichweobtainedhumanisletswereobeseBMI > 30.Isletsfromaged micereceivingwesterndietmayresolveourdifferentndings.Howeverweobserved similarage-dependentdeclinesinislet[Ca 2+ ]inlowerandhigherBMIgroupsnot shown,suggestingthatobesitystatuscannotfullyexplainthesedifferences. Age-dependentdeclineininsulinsecretioninhumans:Wealsoobservedadecline ininsulinsecretionlevelswithageinhumanisletsunderstaticsecretionassay.Our resultsdemonstratingreducedinsulinsecretionlevels,butmaintainedstimulationindexarealsoconsistentwithpublishedndingsGumbineretal.,1989;Iozzoetal., 1999;Greggetal.,2016;Kaytonetal.,2015.Thedeclineinsecretioncorrelated withadeclinein[Ca 2+ ]activity,butnotdeclinesin[Ca 2+ ]coordinationorgapjunctioncouplingFigure38.Inmouseislets,deciencyinCx36gapjunctioncoupling doesnotsignicantlyimpactinsulinsecretionlevels,ratheritaffectinsulinsecretion dynamicsunderstimulatoryconditionsBenningeretal.,2011b;includinginsulin pulsatilityandthedeclineinsecretionfollowingresumptionofbasalglucose,causingglucoseintolerance.Thesendingssuggestotherpathwaysaredisruptedduring ageingtoimpactinsulinsecretionlevels. -cellglucosemetabolismandATPproductionarealsodisruptedwithageingGreggetal.,2016whichcouldexplainthecorrelationofsecretionlevelswith[Ca 2+ ]activitybutnotwith[Ca 2+ ]coordination.Furthermore,thedisruptiontoglucosemetabolismwithageingwouldreduceamplify108

PAGE 121

ingpathwaysandthussecondphaseinsulinsecretionHenquin,2000;Kibbeyetal., 2007.Thismayexplainwhyperifusionanalysisdoesnotshowanage-dependent declineininsulinsecretionlevelsKaytonetal.,2015,butstaticassaysdoshow suchadeclineIozzoetal.,1999;Ihmetal.,2006;Greggetal.,2016.Nevertheless,wecannotexcludethatinhumanisletstheagedependentdeclineingapjunctioncouplingand[Ca 2+ ]coordinationmayweaklyimpact[Ca 2+ ]activityandinsulin secretion,giventhecorrelationbetween[Ca 2+ ]coordinationandactivityFigure34. Thismayalsobeexplainedbysub-populationsof -cellsthatcandisproportionately affectisletfunctionJohnstonetal.,2016a;Baderetal.,2016;Dorrelletal.,2016 beinglostduringageing[S.Bonner-WeirandC.Aguayo-Mazzucato,personalcommunication]. Insummaryweobserveadeclineingapjunctionfunctionandcoordinated[Ca 2+ ], withinhumanisletsduringageing,unlikeinmice.Thisdeclinedisruptsinsulinsecretiondynamicsandthustheresponsivenessoftheislettoacuteglucosechanges. Theage-dependentdisruptiontogapjunctioncoupling,[Ca 2+ ]coordinationandinsulinsecretiondynamicscanberecoveredacutelybyactivatorsofgapjunctioncoupling.Theseresultsprovidefurtherunderstandingtothedeclineinisletfunctionwith ageinhumansthatcausesglucoseintoleranceandpredisposestoincreasedriskfor diabetes. 109

PAGE 122

CHAPTERVI CA2+METRICSIDENTIFYTYPE2DIABETICHUMANDONORS Introduction ClinicaldiabetesismostrecentlydiagnosedusingHbA1ctestingwhereatestlevel of > 6.5%indicatesclinicaldiabetesandprediabetesbetween5.6%and6.4%.Developingindicatorsandpredictorsfordiabetesisofincreasinginterestasincidence rises.BMIaloneistypicallyapoorindicatorofincidenceoftype2diabetesAlmajwal etal.,2009however,systemshavebeendevelopedthatuseacombinationofage, BMI,diet,physicalactivity,familyhistorytoidentifyriskLindstromandTuomilehto, 2003.Indeed,glucosetolerancetestsremainoneofthebestwaystoidentifyriskof developingtype2diabetesAbdul-Ghanietal.,2007;Heikesetal.,2007.Impaired glucosetoleranceisahallmarkofdiabetesandisdisruptedinCx36 )]TJ/F22 7.9701 Tf 7.594 0 Td [(= )]TJ/F36 11.9552 Tf 11.415 -4.338 Td [(miceHead etal.,2012a.Whilethereisevidencetosuggestdecreased[Ca 2+ ]coordinationin modelsofdiabetesinmiceandhumans,thereislessinformationifthisobservedin conrmedhumantype2diabeticsCarvalhoetal.,2012a;Benningeretal.,2011a; Hodsonetal.,2013;Ravieretal.,2002andassuchimplicatinglossofcoordinated [Ca 2+ ]activitytostateofdiabetesinhumansmayprovidevaluableinsightintothe pathophysiologyofthedisease. 110

PAGE 123

Results GeneratingClassierstoPredictType2Diabetes [Ca 2+ ]activityandcoordinationmetricswerecollectedfrom62humandonors,9of whichwereconrmedtype2diabetic. Table2. Parametersusedtogenerateclassicationmodelofdiabetesincidence Sampleparametersforseveraldonorsusedtogenerateclassicationsystems. Weincorporatedactivearea,maximumcoordinatedareanormalizedandunnormalizedandthesynchronizationfactoraspredictorvariablesalongwiththediseasestateusedastheresponsevariable. Age BMI ActiveArea Max.Coord.Area m 2 Sync.Factorr Disease 50 30.2 0.74 9123 0.84 0 36 36.6 0.51 1138 0.39 0 59 32.5 0.28 1570 0.056 2 45 30.5 0.42 766 0.364 2 AsBMIisapoorindicatoroftype2diabetesAlmajwaletal.,2009wersttested theefciencyofabinaryclassicationsystemusing only BMIasapredictorfordiabetesbycalculatingareceiveroperatorcurveROCandit'scorrespondingarea underthecurveAUCSwetsetal.,1988.AnAUCof1correspondstoaperfect classierandaAUCof0.5showstheclassicationisnobetterthanrandomguess. WecalculatethetruepositiverateofclassicationbycomparingthenumberofdiabeticdonorsateachBMIcutoffitaccuratelyclassiesandthefalsepositiverateby thenumberofdonorseachcutoffincorrectlyclassiesasdiabetic. 111

PAGE 124

Figure41. BMIBinaryClassicationofDiabetes.Asimplebinaryclassicationsystem,identifyingdiabeticdonorsbasedonBMIcutoffs,givespoorpredictivepowerfor incidenceoftype2diabetes. TheAUCvalueof0.63inFigure41showswhenclassifyingthediabeticstatusof thedonorBMIismarginallybetterthanguessing-andthusconrmingBMIisapoor predictorofdiabetesinhumandonors.SimilaranalysisusingAgeasthepredictor resultsinanAUCof0.73nowshownwhichalthoughhigherthanBMIisstillaspoor classier. Wenexttestedif[Ca 2+ ]metricsactiveArea,maximumcoordinatedarea,andthe synchronizationfactorrfromequation13couldbeusedtotrainanaccurateclassicationsystem.Avarietyofclassicationlearnersweretestedandtheirperformance summarycanbefoundinTable3. 112

PAGE 125

Table3. SummaryofperformanceforvaryingclassicationlearnersTheperformanceofmultipleclassicationsystemswasevaluatedusingtheareaunderthe curveorC-statisticforthereceiveroperatingcurve.Allclassicationsystemswere trainedinMATLABusingk-foldcrossvalidationk=5.Briey,thedataissplitinto5 partitionsandtheclassieristrainedondatafrom4ofthepartitionsandtestedon 1ofthepartitions.Thisisiterateduntileachpartitionisusedasthetestingdataand theresultsfromallclassiersareaveragedtocreateanestimatorofperformance. ClassicationLearner AUC LinearSVM 0.92 BaggedTrees 0.87 QuadraticSVM 0.86 LogisticRegression 0.84 LinearDiscriminant 0.83 FineKNN 0.8 SimpleTrees 0.67 Boostedtrees 0.4 CoarseKNN 0.39 MultipleclassicationalgorithmsshowedAUCvaluesof > 0.8indicatinggoodto highprociencyatseparatingdiabeticandnon-diabeticdonors.LinearsupportvectormachinesshowedthehighestAUCat0.92Table3andFigure42soweused thisclassicationsystemforsubsequentanalysis. 113

PAGE 126

Figure42. LinearClassicationPerformanceSampleperformanceoftwoclassicationalgorithmsandtheircorrespondingareaunderthecurveusingActiveArea, MaximumCoordinatedAreaandthesynchronizationfactorr. Figure43. SinglepredictorperformanceWeseparatedpredictorvariablessuchthat thelinearsupportvectorclassierwoulduse1variableinitstrainingandtesting. Kfoldcrossvalidationk=5wasusedandtheAUCoftheoftheROCcurvereport above. 114

PAGE 127

Wetestedifanyspecicindividual[Ca 2+ ]metricwasresponsibleforthevariation betweendiabeticandnon-diabeticdonorsbyindividuallytestingtheclassication performanceofeither[Ca 2+ ]activearea,[Ca 2+ ]maximumcoordinatedarea,orthe isletsynchronizationfactorFigure43. Interestingly,theareaofmaximumcoordinatedactivitywaspoorpredictorofincidenceofdiabetesscoringaAUCnearrandomguessingof0.55.[Ca 2+ ]activearea showedahigherpredictionscoreat0.75,whilethiswouldstillbeconsideredapoor classierandtheisletsynchronizationfactor,r,showedanAUCof0.92givingitvery highpredictivepower.Thesedatashowthatoverallelectricalsynchronizationofthe islet,nottheoveralllevelof[Ca 2+ ]isthebestdiscriminatorbetweendiabeticand non-diabeticdonors.Wequantiedthedecreasestoelectricalactivityandsynchronizationintype2diabeticisletscomparedtoagematched > 40yearcontrolislets Figure44.Therewassignicantdeclineinareaofelectricalactivityintype2islets ascomparedtoagematchedcontrolsFigure44A.Similarlythemaximumsizeof thelargestcoordinatedareashowsimilardecreaseFigure44B-C 115

PAGE 128

Figure44. LossofelectricalcoordinationinT2isletsAActiveareainhealthyand type2diabeticislets.BMaximalcoordinatedareainisletsfromhealthyandtype2 diabeticislets.CMaximalcoordinatedareaun-normalizedinisletsfromhealthy adtype2diabeticislets.DElectricalsynchronizationbetweenareasthatshowelectricalactivityfromhealthyandtype2islets.Dataispresentedasmean s.e.m., healthyisletsweretakenfromagematcheddonorsrespectivetotype2donorswhere meandonorageofhealthyisletswas52andmeanofthetype2donorswas51, n=19healthyandn=8type2donors. ModanilIncreasesElectricalCoordinationinHumanType2Islets WenexttestedifModanilcouldincrease[Ca 2+ ]activityandcoordinationmetrics similartotheeffectseeninolderhumandonorsFigure38.Isletsfrom6type2conrmeddonorsweretreatedwithModanil Mfor1hourandimagedusingsim116

PAGE 129

ilarmethodstoChapterV.Modanilincreasedelectricalactivityby 75%andcoordinationmetricsincreaseby 300%intreatedisletsindicatingCx36activatorscan vastlyincreasetheelectricalactivityandcoordinationinhumantype2diabeticislets. Figure45. ModanilTreatmentofHumanType2IsletsA[Ca 2+ ]activeareain isletstreatedwith/without.B,CSizeofmaximal[Ca 2+ ]coordinationareaunnormalizedandwithrespecttoisletsizeuntreatedandtreatedwithmodanil.D Synchronizationfactorrinisletstreatedwith/withoutmodanil.Dataispresented asmean s.e.m.n=6type2donors.StudentsT-Testusedtodeterminestatisticalsignicance. 117

PAGE 130

Conclusions Inthisstudyweusedopticalmicroscopyandimageanalysistoclassifythepresenceoftype2diabetesinhumandonors.Wendthat[Ca 2+ ]coordinationcancreate ahighlypredictiveclassicationmodelinconjunctionwithmodernmachinelearning techniques.Wendthatthesynchronizationfactor-theaveragecoordinationacross allareaswithinanislet,holdsmostofthepredictivepowerneededtodetermineif donorisletscamefromatype2diabetic.Interestingly,thephysicalsizeofthelargest coordinatedareaisinitselfaratherpoorpredictor-perhapsindicatingtheimage analysisalgorithmusedtosegmentregionswithinisletsintodistinctcoordinatedareasisaslightunderrepresentationofthetruecoordinatedelectricaldynamicsinan islet.NeverthelesswiththeadditionoftheCx36activatorModanilwewereableto seemarkedincreasestoelectricalactivityandcoordinationintreatedislets.While donorisletsfromcadaversarenecessarytoascertaintheelectricalcoordination withinislets-isletsdecientofCx36inmiceshownocoordinatedoscillationsand areglucoseintolerant-anmeasuredtakenbyintravenousglucosechallenge.Asintravenous/oralglucosetolerancetestsinhumansareoneofthebestindicatorsof susceptibilityfordevelopingtype2diabetesAbdul-Ghanietal.,2007;Heikesetal., 2007ourresultshintatacorrelationbetweenimpairedglucosetoleranceandloss ofcoordinated[Ca 2+ ]oscillationsintype2diabeticislets. 118

PAGE 131

CHAPTERVII DISCUSSIONANDFUTUREDIRECTIONS Theworkpresentedhasusednoveltechniquesandanalysestostudyregulatory factorsofcoupledelectricaldynamicsinpancreaticislets.Thesedatahelpshed lightonseveralunansweredquestionswithintheeldofisletbiology;includinghow subpopulationsofcellscanexertdisproportionatecontroloverelectricalactivityand showsevidencefornewmethodsoftreatingdiabetes.Furthermore,thisworkraises severalimportantquestionsthatfuturestudiescanaddress. FunctionalRoleofSubopulations Thesedatacontainedtheanalysisofelectrophysiologywithinisletsby in-silico modeling,theuseofmousemodelstoperturbelectricalactivity,anddirectobservationof humanisletelectricalactivity.Werstveriedthe in-silico modelbyrunningparallelexperimentsusinganinduciblemodelofneonataldiabetesHrahaetal.,2014c withtheroleofsubpopulationcontroloverglobalisletelectricalactivityasourprimaryinterest.Weobservedthatcellularexpressionof > 20%offunctionallyinactive cellscausesthecollapseofelectricalactivityoftheisletasawhole-andthatthiseffectisregulateddirectlybyCx36levelsNotaryetal.,2016;Nguyenetal.,2014and wewereabletorecapitulatetheeffectofhyperactiveKATPchannelactivitywithin the in-silico modelHrahaetal.,2014c.Thesedatashowthatsmallsubpopulationsofcellscanexertdisproportionatecontroloverelectricalactivityinislets-due tothepresenceofelectricalcouplingbetweencells.Wetestedifthiswassimilarin healthypancreaticisletsthroughspatiotemporallyperturbingsmallgroupsofcells 119

PAGE 132

withoptogenetics.Weobservedthattherewasahighspatialvariationtoelectrical excitability-whichwehypothesizedwasduetocellsofsimilarelectricalcharacteristicspreferentiallylocatednexttoeachotherandfoundthishypothesiswassupported byourmulticellular in-silico model.Thisorganizationshowedseveralfunctionaleffects;thatneighboringcellscouldberecruitedtoelectricalringinareasofhigher metabolicactivityandtheoppositeforregionsoflowermetabolicactivityandthat the'de-facto'pacemakerregionwascorrelatedtoregionsoflowermetabolicactivity-whichwehypothesizeduetotheoscillatoryfrequencyinthoseregionsbeing higherthanthoseofhighmetabolicactivityintheabsenceofelectricalcoupling. Interestingly,additionofpro-inammatorycytokineswhichdisruptsCx36function showsamarkedchangedinhowsmallregionsareabletocontrolisletelectricaldynamicsFarnsworthetal.,2015b.Inuntreatedisletscellsofhighmetabolicactivity andhigherChR2stimulated[Ca 2+ ]thancellsoflessermetabolicactivity.OurresultsshowedasignicantdeclinetothemetabolicresponsewhichpresumablycorrelatestoadiminishedChR2mediated[Ca 2+ ]response.Asthemetabolicactivity controlsextentofcellularexcitabilityfuturestudiesmaycharacterizethemechanistic basisofmetabolicsubpopulations.Futureworkwillalsoneedtoaddresswhateffect functionalsubpopulationsplayincontrollingfunctioninhumanpancreaticislets-in particulararetheresimilarspatialregionsofhigh/lowmetabolicactivityiscorrelated withwhole-isletsynchronizationandifsubpopulationsaredisruptedwithadvanced ageorhistoryofdiabetes.Thismaybeaccomplishedbytheuseofaphotoactivatablesulfonureawasusedinpreviousstudiestospatiotemporallyactivatesmallpop120

PAGE 133

ulationsof -cells.Uponlightactivationthismoleculeundergoesaconformational changeandbindstoKATPchannelstriggeringchannelclosureandmembranedepolarizationJohnstonetal.,2016b;Broichhagenetal.,2014.Additionally,future studieswillmodelthecoupledelectrophysiologyofhumanisletcytoarchitecture.As -cellsoccupythecoreofmouseislets,humanisletsshowquasi-randomarrangementsofcellsthroughoutisletswhichmaybeuidasafunctionofageCabrera etal.,2006;Greggetal.,2012.Modelingagedependent -cellcontactprobabilitiesandfunctionalsubpopulationsmaygivevaluableinsightintocoupledelectrical dynamicswithinhumanisletsandwemayexpectthatasthecontactprobabilities between -cellsdecreasewithage-theseisletsmaybemoresusceptibletouncoordinated[Ca 2+ ]dynamics. LossofCa2+CoordinationwithAdvancedAgeandType2Diabetes Theriskfortype2diabetesincreaseswithageinhumansKarveandHayward, 2010balthoughelucidatingfunctionalcausesisdifcult.Inpartduetopooravailabilityofhumandonorisletsandbecausecoordinatedoscillatoryactivityhasbeen difculttoquantifydirectlyduetothequasi-coordinatedelectricaldynamicsofhuman pancreaticislets.Wenexttestedhowelectricalactivityandcoordinationdecreases asafunctionofdonorageinhumanislets.Priorstudiesshowthatalthoughglucose tolerancedoesnotdecreasesignicantlyinagedmiceanditdoesinhumansKarve andHayward,2010bandthatelectricalcoordinationhasbeenlinkedtoglucosetoleranceinmiceHeadetal.,2012awhichshowsnodecreaseisoldermiceFigure 40,howeveritdoesshowsignicantdeclinesinhumansFigure33ourhypothesis 121

PAGE 134

isthatdecreasestoelectricalcoordinationinpancreaticisletsinhumanswithage isacontributingfactortotheincreaseofriskoftype2diabetes-althoughthisisdifculttoshowinalongitudinalstudyaspancreaticisletsareonlyavailablethrough cadavertissue-anddirectimagingofpancreaticislets in-vivo isinvasiveLindsay etal.,2014.Neverthelessweshowthattherearesignicantdeclinetoelectricalcoordinationwithageinhumanislets.WetestedhowtargetingCx36levelsmayimproveelectricalcoordinationinhumanisletsasapotentialtherapeuticandpreventativemeasureagainstdiabetes.ModanilisaFDAapprovedpharmaceuticalagent, usedtotreatnarcolepsyBastojiandJouvet,1988howevertheexactmechanism ofactionislessunderstoodalthoughCAMkinaseIIhasbeenimplicatedUrbano etal.,2007.AftertreatedisletsinModanilwefoundthattheelectricalcoordinationassociatedwitholderdonorswasincreasedtolevelsofyoungerdonors-showingthatfunctionalcoordinationcouldberecoveredthroughincreasestoCx36levels Farnsworthetal.,2014.Interestingly,coordinationwasnotsignicantlyincreased inyoungerdonorsimplyingthattheremaybeafunctionallimittocoordinationinhumanislets,whichisdistinctfromthefullcoordinationobservedinmouseislets.Humanisletarchitecturemayplayacrucialroleintheinabilityofhumanisletstoshow islet-wideelectricalcoordinationandindeedthishasbeenobservedtochangeas afunctionofageinhumansGreggetal.,2012.Quanticationofisletarchitecture withageinconjunctionwith[Ca 2+ ]imagingmaybecrucialindeterminingifthereisa correlationbetween cellcontactsand[Ca 2+ ]coordination. 122

PAGE 135

Giventheavailabilityofhumanisletdonorislets,andtheprevalenceoftype2diabetesamongthegeneralpopulationstudyingelectricaldynamicsintype2diabetic humanisletsrequiressignicantinvestmentintime.Weobtainedn=9isletsfrom conrmedtype2diabeticdonorsandquantiedelectricalactivity,coordination,and responsetoCx36activators.Thesedatashowsignicantdisruptiontoelectricalsynchronizationinisletsfromtype2diabeticdonorscomparedtonon-diabeticdonors whichcanberecoveredwiththeadditionofCx36activatorModanil.Whilestudies suggestdisruptedelectricalcouplinginanimalmodelsoftype2diabetesthisisthe rstquanticationofdisruptiontoelectricalsynchronizationinhumanswithtype2diabetesandourdataimpliesitmayberecoverablewithCx36activatorspotentially creatinganewmethodologyfortreatmentoftype2diabetes.Thelossofpulsatileinsulinoscillations,whichismoreeffectiveatclearingbloodglucoseandupregulating theinsulinreceptorMatveyenkoetal.,2012;Matthewsetal.,1983isassociated withtype2diabetesandsimilarlywithCx36knockoutinmice.Byrestoringcoordinatedoscillationsintype2diabeticswemayrestorepulsatileinsulinoscillationsand providegreaterefcacyinclearingbloodglucose.SimilarlyasdysfunctiontopulsatileinsulinsecretionhasbeenlinkedtodecreaseininsulinsensitivityZarkovic etal.,1999.Restoringpulsatileoscillationsmaydecreaseinsulinresistance.As thereisasegmentofthepopulationcurrentlyusingModaniltotreatvariousconditions-itwouldbelogicaltotestifthispopulationhaslowerincidenceofdiabetes orismoreglucosetolerance.Suchpreliminarystudiesmayjustifyfurthertestingfor Modanilasatype2diabetestherapeutic.Althoughtheriskfordevelopingtype2dia123

PAGE 136

betesincreaseswithage,andindeedthemajorityofourdonorswithtype2diabetes wereovertheageof40thequestionofwhatcausesthefunctionalcollapsein[Ca 2+ ] activityandcoordinationintype2diabetics.TheadditionofModanlsuggeststhat cellscanberecruitedtoactivityintype2diabeticisletswhichmayindicatethatthey arenotasdysfunctionalasthe -cellsexpressingtheKir6.2 [ N 30 ;K 185 Q ] -GFPmutationaswewouldexpectthathighercouplingwouldleadtofurthersilencingofthe isletaspriorworkwouldsuggestNotaryetal.,2016;Nguyenetal.,2014.Determiningthephysiologicalcharacteristicsofthoseindividualcellsmayhelpexplainwhy somecellsfailandothersdonot. ConcludingRemarks Thechapterspresentedinthisdissertationhaveaunderlyingthemeofinvestigatingregulatingfactorsofcoordinatedelectricaldynamicsof -cellsinpancreaticislets anddeterminewhatrolethisplaysinthepathophysiologyofdiabetes.Thiswasachieved throughseveraltechniquesandanalyses:Developamulticellularcomputational modeltobeusedinfuturestudiesanddescribetwodistinctaspectsofpancreatic isletdysfunction.Usethismodelincombinationwithspatiotemporalmanipulationof -cellelectricalactivityusingoptogeneticstoquantifyspatialorientationof subpopluations,showsubpopulationspreferentialcontrolofisletelectricaldynamics, andtesthowoptogeneticcontrolisdisruptedindysfunctionalconditions.Quantifychangestopancreaticisletelectricaldynamicswithrespecttoageanddiabetic stateinhumandonorisletsandproposeanewpotentialtherapeutictotreattype2 124

PAGE 137

diabetes.Ibelievetheseresultshaveprovidednewinsightsintocoupledpancreatic -celldynamicsanditsroleindiabetes. 125

PAGE 138

BIBLIOGRAPHY Abdul-Ghani,M.A.,Williams,K.,DeFronzo,R.A.,andStern,M..Whatisthe bestpredictoroffuturetype2diabetes? DiabetesCare ,30:1544. Almajwal,A.,Al-Baghli,N.,Batterham,M.,Williams,P.,Al-Turki,K.,andAl-Ghamdi, A..Performanceofbodymassindexinpredictingdiabetesandhypertension intheeasternprovinceofsaudiarabia. AnnalsofSaudiMedicine ,29:437. Ashcroft,F.M..ATP-sensitivepotassiumchannelopathies:focusoninsulin secretion. JournalofClinicalInvestigation ,115:2047. Ashcroft,F.M.,Harrison,D.E.,andAshcroft,S.J.H..Glucoseinduces closureofsinglepotassiumchannelsinisolatedratpancreatic -cells. Nature 312:446. Ashcroft,F.M.andRorsman,P..Electrophysiologyofthepancreatic -cell. ProgressinBiophysicsandMolecularBiology ,54:87. Avrahami,D.,Li,C.,Zhang,J.,Schug,J.,Avrahami,R.,Rao,S.,Stadler,M.B., Burger,L.,Sch ubeler,D.,Glaser,B.,andKaestner,K.H..Aging-Dependent DemethylationofRegulatoryElementsCorrelateswithChromatinStateandImproved CellFunction. CellMetabolism ,22:619. Bader,E.,Migliorini,A.,Gegg,M.,Moruzzi,N.,Gerdes,J.,Roscioni,S.S.,Bakhti, M.,Brandl,E.,Irmler,M.,Beckers,J.,Aichler,M.,Feuchtinger,A.,Leitzinger,C., Zischka,H.,Wang-Sattler,R.,Jastroch,M.,Tsch op,M.,Machicao,F.,Staiger,H., H aring,H.-U.,Chmelova,H.,Chouinard,J.A.,Oskolkov,N.,Korsgren,O.,Speier,S., andLickert,H..Identicationofproliferativeandmature -cellsintheisletsof Langerhans. Nature ,535:430. Ballian,N.andBrunicardi,F.C..IsletVasculatureasaRegulatorofEndocrine PancreasFunction. WorldJ.Surg. ,31:705. Bano,G..Glucosehomeostasis,obesityanddiabetes. BestPractice&ResearchClinicalObstetrics&Gynaecology ,27:715. Bastoji,H.andJouvet,M..Successfultreatmentofidiopathichypersomnia andnarcolepsywithmodanil. Progressinneuro-psychopharmacologyandbiologicalpsychiatry ,12:695. Basu,R.,Breda,E.,Oberg,A.L.,Powell,C.C.,DallaMan,C.,Basu,A.,Vittone, J.L.,Klee,G.G.,Arora,P.,Jensen,M.D.,Toffolo,G.,Cobelli,C.,andRizza,R.A. .MechanismsoftheAge-AssociatedDeteriorationinGlucoseTolerance: ContributionofAlterationsinInsulinSecretion,Action,andClearance. Diabetes 52:1738. 126

PAGE 139

Benninger,R.,Remedi,M.,Head,W.,Ustione,A.,Piston,D.,andNichols,C. a.Defectsinbetacellca2+signalling,glucosemetabolismandinsulinsecretioninamurinemodelofkatpchannel-inducedneonataldiabetesmellitus. Diabetologia ,54:1087. Benninger,R.K.,Hutchens,T.,Head,W.S.,McCaughey,M.J.,Zhang,M.,Marchand,S.J.L.,Satin,L.S.,andPiston,D.W..Intrinsicisletheterogeneityand gapjunctioncouplingdeterminespatiotemporalca2+wavedynamics. Biophysical Journal ,107:2723. Benninger,R.K.,Zhang,M.,Head,W.S.,Satin,L.S.,andPiston,D.W..Gap JunctionCouplingandCalciumWavesinthePancreaticIslet. BiophysicalJournal 95:5048. Benninger,R.K.P.,Head,W.S.,Zhang,M.,Satin,L.S.,andPiston,D.W.b. Gapjunctionsandothermechanismsofcell-cellcommunicationregulatebasalinsulinsecretioninthepancreaticislet. TheJournalofPhysiology ,589:5453 5466. Bernard,A.B.,Lin,C.-C.,andAnseth,K.S..Amicrowellcellcultureplatformfortheaggregationofpancreatic -cells. TissueEngineeringPartC:Methods 18:583. Bertram,R.,Previte,J.,Sherman,A.,Kinard,T.A.,andSatin,L.S..The phantomburstermodelforpancreatic -cells. BiophysicalJournal ,79:2880. Boden,G..Obesityandfreefattyacids. EndocrinologyandMetabolismClinicsofNorthAmerica ,37:635. Breakspear,M.,Heitmann,S.,andDaffertshofer,A..Generativemodelsof corticaloscillations:Neurobiologicalimplicationsofthekuramotomodel. Frontiersin HumanNeuroscience ,4. Breakspear,M.andStam,C.J..Dynamicsofaneuralsystemwithamultiscalearchitecture. PhilosophicalTransactionsoftheRoyalSocietyB:Biological Sciences ,360:1051. Broichhagen,J.,Schonberger,M.,Cork,S.C.,Frank,J.A.,Marchetti,P.,Bugliani, M.,Shapiro,A.M.J.,Trapp,S.,Rutter,G.A.,Hodson,D.J.,andTrauner,D.. Opticalcontrolofinsulinreleaseusingaphotoswitchablesulfonylurea. NatureCommunications ,5:5116. Cabrera,O.,Berman,D.M.,Kenyon,N.S.,Ricordi,C.,Berggren,P.-O.,and Caicedo,A..Theuniquecytoarchitectureofhumanpancreaticisletshasimplicationsforisletcellfunction. ProceedingsoftheNationalAcademyofSciences 103:2334. 127

PAGE 140

Carvalho,C.P.d.F.,Oliveira,R.B.,Britan,A.,Silva-Santos,J.C.,Boschero,A.C., Meda,P.,andCollares-Buzato,C.B.a.Impairedbeta-to-betacellcoupling mediatedbycx36gapjunctionsinpre-diabeticmice. AmericanJournalofPhysiology -EndocrinologyandMetabolism Carvalho,C.P.F.,Barbosa,H.C.L.,Britan,A.,Santos-Silva,J.C.R.,Boschero, A.C.,Meda,P.,andCollares-Buzato,C.B..Betacellcouplingandconnexin expressionchangeduringthefunctionalmaturationofratpancreaticislets. Diabetologia ,53:1428. Carvalho,C.P.F.,Oliveira,R.B.,Britan,A.,Santos-Silva,J.C.,Boschero,A.C., Meda,P.,andCollares-Buzato,C.B.b.Impaired -cell-cellcouplingmediatedbyCx36gapjunctionsinprediabeticmice. AJP:EndocrinologyandMetabolism 303:E144E151. Cha,C.Y.,Nakamura,Y.,Himeno,Y.,Wang,J.,Fujimoto,S.,Inagaki,N.,Earm, Y.E.,andNoma,A.a.Ionicmechanismsandca2+dynamicsunderlyingthe glucoseresponseofpancreatic cells:asimulationstudy. TheJournalofGeneral Physiology ,138:21. Cha,C.Y.,Santos,E.,Amano,A.,Shimayoshi,T.,andNoma,A.b.Timedependentchangesinmembraneexcitabilityduringglucose-inducedburstingactivity inpancreatic cells. TheJournalofGeneralPhysiology ,138:39. Chang,A.M.andHalter,J.B..Agingandinsulinsecretion. AmericanJournal ofPhysiology-EndocrinologyAndMetabolism ,284:E7E12. Chay,T.andKeizer,J..Minimalmodelformembraneoscillationsinthepancreaticbeta-cell. BiophysicalJournal ,42:181. Craig,T.J.,Ashcroft,F.M.,andProks,P..HowATPinhibitstheopenkATP channel. TheJournalofGeneralPhysiology ,132:131. Curry,D.L..Glucagonpotentiationofinsulinsecretionbytheperfusedrat pancreas. Diabetes ,19:420. Davis,A..Obesity,increasedlineargrowth,andriskoftype1diabetesinchildren. ClinicalPediatrics ,40:634. DeLe on,D.D.andStanley,C.A..Permanentneonataldiabetesmellitus. Dean,P.M.andMatthews,E.K..Glucose-inducedelectricalactivityinpancreaticisletcells. TheJournalofPhysiology ,210:255. Degen,J.,Meier,C.,VanDerGiessen,R.S.,S ohl,G.,Petrasch-Parwez,E.,Urschel, S.,Dermietzel,R.,Schilling,K.,DeZeeuw,C.I.,andWillecke,K..Expression patternoflacZreportergenerepresentingconnexin36intransgenicmice. Journalof ComparativeNeurology ,473:511. 128

PAGE 141

Desgraz,R.andHerrera,P.L..Pancreaticneurogenin3-expressingcellsare unipotentisletprecursors. Development ,136:3567. Dorrell,C.,Schug,J.,Canaday,P.S.,Russ,H.A.,Tarlow,B.D.,Grompe,M.T.,Horton,T.,Hebrok,M.,Streeter,P.R.,Kaestner,K.H.,andGrompe,M..Human isletscontainfourdistinctsubtypesof cells. NatureCommunications ,7:11756. Drews,G.,Krippeit-Drews,P.,andDufer,M..Electrophysiologyofisletcells. In IsletsofLangerhans,2.ed. ,pages1.SpringerScience+BusinessMedia. Dula,S.B.,Jecmenica,M.,Wu,R.,Jahanshahi,P.,Verrilli,G.M.,Carter,J.D.,Brayman,K.L.,andNunemaker,C.S..Evidencethatlow-gradesystemicinammationcaninduceisletdysfunctionasmeasuredbyimpairedcalciumhandling. Cell Calcium ,48-3:133. Farnsworth,N.L.,Hemmati,A.,Pozzoli,M.,andBenninger,R.K.P..Fluorescencerecoveryafterphotobleachingrevealsregulationanddistributionof connexin36gapjunctioncouplingwithinmouseisletsofLangerhans. JPhysiol 592:4431. Farnsworth,N.L.,Walter,R.L.,Hemmati,A.,Westacott,M.J.,andBenninger,R. K.P.a.Lowlevelpro-inammatorycytokinesdecreaseconnexin36gapjunctioncouplinginmouseandhumanisletsthroughnitricoxide-mediatedproteinkinase c JournalofBiologicalChemistry ,291:3184. Farnsworth,N.L.,Walter,R.L.,Hemmati,A.,Westacott,M.J.,andBenninger,R. K.P.b.LowLevelPro-inammatoryCytokinesDecreaseConnexin36Gap JunctionCouplinginMouseandHumanIsletsthroughNitricOxide-mediatedProtein Kinase JournalofBiologicalChemistry ,291:3184. Freitas,C.,Macau,E.,andViana,R.L..Synchronizationversusneighborhoodsimilarityincomplexnetworksofnonidenticaloscillators. PhysicalReviewE 92. Fridlyand,L.E.,Tamarina,N.,andPhilipson,L.H..Modelingofca2+uxin pancreatic -cells:roleoftheplasmamembraneandintracellularstores. American JournalofPhysiology-EndocrinologyAndMetabolism ,285:E138E154. Fritsche,A.,Madaus,A.,Stefan,N.,Tschritter,O.,Maerker,E.,Teigeler,A.,Haring, H.,andStumvoll,M..RelationshipsAmongAge,ProinsulinConversion,and -CellFunctioninNondiabeticHumans. Diabetes ,51Supplement1:S234S239. Gloyn,A.L.,Weedon,M.N.,Owen,K.R.,Turner,M.J.,Knight,B.A.,Hitman,G., Walker,M.,Levy,J.C.,Sampson,M.,Halford,S.,McCarthy,M.I.,Hattersley,A.T., andFrayling,T.M..Large-scaleassociationstudiesofvariantsingenes encodingthepancreatic -cellKATPchannelsubunitskir6.2KCNJ11andSUR1 129

PAGE 142

ABCC8conrmthattheKCNJ11e23kvariantisassociatedwithtype2diabetes. Diabetes ,52:568. Gregg,B.E.,Moore,P.C.,Demozay,D.,Hall,B.A.,Li,M.,Husain,A.,Wright,A.J., Atkinson,M.A.,andRhodes,C.J..Formationofahuman -cellpopulation withinpancreaticisletsissetearlyinlife. TheJournalofClinicalEndocrinology& Metabolism ,97:3197. Gregg,T.,Poudel,C.,Schmidt,B.A.,Dhillon,R.S.,Sdao,S.M.,Truchan,N.A., Baar,E.L.,Fernandez,L.A.,Denu,J.M.,Eliceiri,K.W.,Rogers,J.D.,Kimple, M.E.,Lamming,D.W.,andMerrins,M.J..Pancreatic cellsfrommiceoffset age-associatedmitochondrialdeciencywithreducedKATPChannelactivity. Diabetes ,pagedb160432. Gumbiner,B.,Polonsky,K.S.,Beltz,W.F.,Wallace,P.,Brechtel,G.,andFink,R.I. .EffectsofAgingonInsulinSecretion. Diabetes ,38:1549. Gunasekaran,U.andGannon,M..Type2DiabetesandtheAgingPancreatic BetaCell. Aging ,3:565. Hashimoto,S.,Kubota,N.,Sato,H.,Sasaki,M.,Takamoto,I.,Kubota,T.,Nakaya,K., Noda,M.,Ueki,K.,andKadowaki,T..InsulinReceptorSubstrate-2Irs2in EndothelialCellsPlaysaCrucialRoleinInsulinSecretion. Diabetes ,64:876. Hattersley,A.T.andAshcroft,F.M..Activatingmutationsinkir6.2andneonataldiabetes. Diabetes ,54:2503. Head,W.S.,Orseth,M.L.,Nunemaker,C.S.,Satin,L.S.,Piston,D.W.,andBenninger,R.K.a.Connexin-36gapjunctionsregulateinvivorst-andsecondphaseinsulinsecretiondynamicsandglucosetoleranceintheconsciousmouse. Diabetes ,61:1700. Head,W.S.,Orseth,M.L.,Nunemaker,C.S.,Satin,L.S.,Piston,D.W.,andBenninger,R.K.P.b.Connexin-36GapJunctionsRegulateInVivoFirst-and Second-PhaseInsulinSecretionDynamicsandGlucoseToleranceintheConscious Mouse. Diabetes ,61:1700. Heikes,K.E.,Eddy,D.M.,Arondekar,B.,andSchlessinger,L..Diabetesrisk calculator:Asimpletoolfordetectingundiagnoseddiabetesandpre-diabetes. DiabetesCare ,31:1040. Helman,A.,Klochendler,A.,Azazmeh,N.,Gabai,Y.,Horwitz,E.,Anzi,S.,Swisa,A., Condiotti,R.,Granit,R.Z.,Nevo,Y.,Fixler,Y.,Shreibman,D.,Zamir,A.,TornovskyBabeay,S.,Dai,C.,Glaser,B.,Powers,A.C.,Shapiro,A.M.J.,Magnuson,M.A., Dor,Y.,andBen-Porath,I..p16ink4a-inducedsenescenceofpancreaticbeta cellsenhancesinsulinsecretion. NatureMedicine ,22:412. 130

PAGE 143

Henquin,J.C..Triggeringandamplifyingpathwaysofregulationofinsulin secretionbyglucose. Diabetes ,49:1751. Hingorani,S.R.,Petricoin,E.F.,Maitra,A.,Rajapakse,V.,King,C.,Jacobetz,M.A., Ross,S.,Conrads,T.P.,Veenstra,T.D.,Hitt,B.A.,Kawaguchi,Y.,Johann,D.,Liotta,L.A.,Crawford,H.C.,Putt,M.E.,Jacks,T.,Wright,C.V.,Hruban,R.H.,Lowy, A.M.,andTuveson,D.A..Preinvasiveandinvasiveductalpancreaticcancer anditsearlydetectioninthemouse. CancerCell ,4:437. Hodson,D.J.,Mitchell,R.K.,Bellomo,E.A.,Sun,G.,Vinet,L.,Meda,P.,Li,D.,Li, W.-H.,Bugliani,M.,Marchetti,P.,Bosco,D.,Piemonti,L.,Johnson,P.,Hughes,S.J., andRutter,G.A..Lipotoxicitydisruptsincretin-regulatedhuman cellconnectivity. JournalofClinicalInvestigation ,123:4182. Hraha,T.H.,Bernard,A.B.,Nguyen,L.M.,Anseth,K.S.,andBenninger,R.K. a.DimensionalityandSizeScalingofCoordinatedCa2+DynamicsinMIN6 -cellClusters. BiophysicalJournal ,106:299. Hraha,T.H.,Westacott,M.J.,Pozzoli,M.,Notary,A.M.,McClatchey,P.M.,and Benninger,R.K.P.b.PhaseTransitionsintheMulti-cellularRegulatoryBehaviorofPancreaticIsletExcitability. PLoSComputBiol ,10:e1003819. Hraha,T.H.,Westacott,M.J.,Pozzoli,M.,Notary,A.M.,McClatchey,P.M.,and Benninger,R.K.P.c.Phasetransitionsinthemulti-cellularregulatorybehaviorofpancreaticisletexcitability. PLoSComputBiol ,10:e1003819. Ihm,S.-H.,Matsumoto,I.,Sawada,T.,Nakano,M.,Zhang,H.J.,Ansite,J.D., Sutherland,D.E.,andHering,B.J..EffectofDonorAgeonFunctionofIsolatedHumanIslets. Diabetes ,55:1361. Ionescu-Tirgoviste,C.,Gagniuc,P.A.,Gubceac,E.,Mardare,L.,Popescu,I.,Dima, S.,andMilitaru,M..A3dmapoftheisletroutesthroughoutthehealthyhumanpancreas. ScienticReports ,5:14634. Iozzo,P.,Beck-Nielsen,H.,Laakso,M.,Smith,U.,Yki-J arvinen,H.,andFerrannini, E..IndependentInuenceofAgeonBasalInsulinSecretioninNondiabetic Humans. TheJournalofClinicalEndocrinology&Metabolism ,84:863. Ivanova,E.,Yee,C.W.,andSagdullaev,B.T..Increasedphosphorylation ofCx36gapjunctionsintheAIIamacrinecellsofRDretina. FrontiersinCellular Neuroscience ,9. Iynedjian,P.B..Molecularphysiologyofmammalianglucokinase. Cellular andMolecularLifeSciences ,66:27. Jain,R.andLammert,E..Cell-cellinteractionsintheendocrinepancreas. Diabetes,ObesityandMetabolism ,11:159. 131

PAGE 144

Jalan,S.,Singh,A.,Acharyya,S.,andKurths,J..Impactofaleaderonclustersynchronization. PhysicalReviewE ,91. Jensen,M.V.,Joseph,J.W.,Ronnebaum,S.M.,Burgess,S.C.,Sherry,A.D., andNewgard,C.B..Metaboliccyclingincontrolofglucose-stimulatedinsulinsecretion. AmericanJournalofPhysiology-EndocrinologyandMetabolism 295:E1287E1297. Jetton,T.L.andMagnuson,M.A..Heterogeneousexpressionofglucokinase amongpancreaticbetacells. ProceedingsoftheNationalAcademyofSciences 89:2619. Johnston,N.R.,Mitchell,R.K.,Haythorne,E.,Pessoa,M.P.,Semplici,F.,Ferrer, J.,Piemonti,L.,Marchetti,P.,Bugliani,M.,Bosco,D.,Berishvili,E.,Duncanson, P.,Watkinson,M.,Broichhagen,J.,Trauner,D.,Rutter,G.A.,andHodson,D.J. a.BetaCellHubsDictatePancreaticIsletResponsestoGlucose. Cell Metabolism ,24:389. Johnston,N.R.,Mitchell,R.K.,Haythorne,E.,Pessoa,M.P.,Semplici,F.,Ferrer,J.,Piemonti,L.,Marchetti,P.,Bugliani,M.,Bosco,D.,Berishvili,E.,Duncanson,P.,Watkinson,M.,Broichhagen,J.,Trauner,D.,Rutter,G.A.,andHodson, D.J.b.Betacellhubsdictatepancreaticisletresponsestoglucose. Cell Metabolism ,24:389. Jorns,A.,Tiedge,M.,andLenzen,S..Nutrient-dependentdistributionof insulinandglucokinaseimmunoreactivitiesinratpancreaticbetacells. Virchows Archiv ,434:75. Karaca,M.,Castel,J.,Tourrel-Cuzin,C.,Brun,M.,G eant,A.,Dubois,M.,Catesson, S.,Rodriguez,M.,Luquet,S.,Cattan,P.,Lockhart,B.,Lang,J.,Ktorza,A.,Magnan, C.,andKargar,C..Exploringfunctional -cellheterogeneityinvivousing PSA-NCAMasaspecicmarker. PLoSONE ,4:e5555. Karve,A.andHayward,R.A.a.Prevalence,Diagnosis,andTreatmentofImpairedFastingGlucoseandImpairedGlucoseToleranceinNondiabeticU.S.Adults. DiabetesCare ,33:2355. Karve,A.andHayward,R.A.b.Prevalence,diagnosis,andtreatmentofimpairedfastingglucoseandimpairedglucosetoleranceinnondiabeticu.s.adults. DiabetesCare ,33:2355. Kasai,H.,Hatakeyama,H.,Ohno,M.,andTakahashi,N..Exocytosisinislet -cells.In IsletsofLangerhans ,pages475.SpringerScience+BusinessMedia. 132

PAGE 145

Katsuta,H.,Aguayo-Mazzucato,C.,Katsuta,R.,Akashi,T.,Hollister-Lock,J., Sharma,A.J.,Bonner-Weir,S.,andWeir,G.C..SubpopulationsofGFPmarkedmousepancreatic -cellsdifferinsize,granularity,andinsulinsecretion. Endocrinology ,153:5180. Kawahito,S..Problemsassociatedwithglucosetoxicity:Roleof hyperglycemia-inducedoxidativestress. WorldJournalofGastroenterology 15:4137. Kayton,N.S.,Poffenberger,G.,Henske,J.,Dai,C.,Thompson,C.,Aramandla,R., Shostak,A.,Nicholson,W.,Brissova,M.,Bush,W.S.,andPowers,A.C..Humanisletpreparationsdistributedforresearchexhibitavarietyofinsulin-secretory proles. AmericanJournalofPhysiology-EndocrinologyAndMetabolism 308:E592E602. Kibbey,R.G.,Pongratz,R.L.,Romanelli,A.J.,Wollheim,C.B.,Cline,G.W.,and Shulman,G.I..MitochondrialGTPregulatesglucose-stimulatedinsulinsecretion. CellMetabolism ,5:253. Kilimnik,G.,Jo,J.,Periwal,V.,Zielinski,M.C.,andHara,M..Quanticationof isletsizeandarchitecture. Islets ,4:167. Kilimnik,G.,Zhao,B.,Jo,J.,Periwal,V.,Witkowski,P.,Misawa,R.,andHara,M. .AlteredIsletCompositionandDisproportionateLossofLargeIsletsinPatientswithType2Diabetes. PLoSONE ,6:e27445. Kinard,T.,deVries,G.,Sherman,A.,andSatin,L..Modulationofthebursting propertiesofsinglemousepancreatic -cellsbyarticialconductances. Biophysical Journal ,76:1423. Kohavi,R.etal..Astudyofcross-validationandbootstrapforaccuracyestimationandmodelselection.In Ijcai ,volume14,pages1137.Stanford,CA. Konstantinova,I.,Nikolova,G.,Ohara-Imaizumi,M.,Meda,P.,Kucuera,T.,Zarbalis, K.,Wurst,W.,Nagamatsu,S.,andLammert,E..EphA-ephrin-a-mediated cellcommunicationregulatesinsulinsecretionfrompancreaticislets. Cell 129:359. Koster,J.C.,Permutt,M.A.,andNichols,C.G..Diabetesandinsulinsecretion. Diabetes ,54:3065. Koster,J.C.,Remedi,M.S.,Flagg,T.P.,Johnson,J.D.,Markova,K.P.,Marshall, B.A.,andNichols,C.G..Hyperinsulinisminducedbytargetedsuppression ofbetacellKATPchannels. ProceedingsoftheNationalAcademyofSciences 99:16992. 133

PAGE 146

Kushner,J.A..Theroleofagingupon cellturnover. JournalofClinical Investigation ,123:990. Lang,D.A.,Matthews,D.R.,Burnett,M.,andTurner,R.C..Brief,irregular oscillationsofbasalplasmainsulinandglucoseconcentrationsindiabeticman. Diabetes ,30:435. Lawrence,M.C..Regulationofinsulingenetranscriptionbyaca2+responsivepathwayinvolvingcalcineurinandnuclearfactorofactivatedtcells. MolecularEndocrinology ,15:1758. Lernmark,..Thepreparationof,andstudieson,freecellsuspensionsfrom mousepancreaticislets. Diabetologia ,10:431. Li,L.,Trifunovic,A.,Kohler,M.,Wang,Y.,Berglund,J.P.,Illies,C.,Juntti-Berggren, L.,Larsson,N.-G.,andBerggren,P.-O..Defectsin -cellca2+dynamicsin age-induceddiabetes. Diabetes ,63:4100. Lin,J.-M.,Fabregat,M.E.,Gomis,R.,andBergsten,P..Pulsatileinsulin releasefromisletsisolatedfromthreesubjectswithtype2diabetes. Diabetes 51:988. Lindsay,R.S.,Corbin,K.,Mahne,A.,Levitt,B.E.,Gebert,M.J.,Wigton,E.J., Bradley,B.J.,Haskins,K.,Jacobelli,J.,Tang,Q.,Krummel,M.F.,andFriedman, R.S..Antigenrecognitionintheisletschangeswithprogressionofautoimmuneisletinltration. TheJournalofImmunology ,194:522. Lindstrom,J.andTuomilehto,J..Thediabetesriskscore:Apracticaltoolto predicttype2diabetesrisk. DiabetesCare ,26:725. Magnus,G.andKeizer,J.a.Modelof -cellmitochondrialcalciumhandling andelectricalactivity.i.cytoplasmicvariables. AmericanJournalofPhysiology-Cell Physiology ,274:C1158C1173. Magnus,G.andKeizer,J.b.Modelofbeta-cellmitochondrialcalciumhandling andelectricalactivity.II.Mitochondrialvariables. Am.J.Physiol. ,274Pt1:C1174 1184. Markovi c,R.,Sto zer,A.,Gosak,M.,Dolen sek,J.,Marhl,M.,andRupnik,M.S. .Progressiveglucosestimulationofisletbetacellsrevealsatransitionfrom segregatedtointegratedmodularfunctionalconnectivitypatterns. Sci.Rep. ,5:7845. Matthews,D.R.,Naylor,B.A.,Jones,R.G.,Ward,G.M.,andTurner,R.C.. Pulsatileinsulinhasgreaterhypoglycemiceffectthancontinuousdelivery. Diabetes 32:617. 134

PAGE 147

Matveyenko,A.V.,Liuwantara,D.,Gurlo,T.,Kirakossian,D.,Man,C.D.,Cobelli,C., White,M.F.,Copps,K.D.,Volpi,E.,Fujita,S.,andButler,P.C..Pulsatile portalveininsulindeliveryenhanceshepaticinsulinactionandsignaling. Diabetes 61:2269. Meier,J.J.,Pennartz,C.,Schenker,N.,Menge,B.A.,Schmidt,W.E.,Heise,T., Kapitza,C.,andVeldhuis,J.D..Hyperglycaemiaisassociatedwithimpaired pulsatileinsulinsecretion:effectofbasalinsulintherapy. DiabetesObesMetab 15:258. Meier,J.J.,Veldhuis,J.D.,andButler,P.C..Pulsatileinsulinsecretiondictatessystemicinsulindeliverybyregulatinghepaticinsulinextractioninhumans. Diabetes ,54:1649. Mirollo,R.E.andStrogatz,S.H..Synchronizationofpulse-coupledbiological oscillators. SIAMJ.Appl.Math. ,50:1645. Montaseri,G.andMeyer-Hermann,M..Diversityofcoupledoscillatorscan enhancetheirsynchronization. PhysicalReviewE ,94. Moreno,A.P..Biophysicalevidencethatconnexin-36formsfunctionalgap junctionchannelsbetweenpancreaticmouse -cells. AJP:Endocrinologyand Metabolism ,288:E948E956. Morioka,T.andKulkarni,R.N..Pancreaticislet -cellfailureinobesity.In MetabolicBasisofObesity ,pages199.SpringerNewYork. Nagel,G.,Szellas,T.,Huhn,W.,Kateriya,S.,Adeishvili,N.,Berthold,P.,Ollig,D., Hegemann,P.,andBamberg,E..Channelrhodopsin-2,adirectlylight-gated cation-selectivemembranechannel. ProceedingsoftheNationalAcademyofSciences ,100:13940. Nathan,D.M.,Davidson,M.B.,DeFronzo,R.A.,Heine,R.J.,Henry,R.R.,Pratley, R.,andZinman,B..ImpairedFastingGlucoseandImpairedGlucoseTolerance. DiabetesCare ,30:753. Newsholme,P.,Gaudel,C.,andMcClenaghan,N.H..Nutrientregulationof insulinsecretionand -cellfunctionalintegrity.pages91. Nguyen,L.M.,Pozzoli,M.,Hraha,T.H.,andBenninger,R.K..Decreasing cx36gapjunctioncouplingcompensatesforoveractiveKATPChannelstorestore insulinsecretionandpreventhyperglycemiainamousemodelofneonataldiabetes. Diabetes ,63:1685. Nikolic,K.,Grossman,N.,Grubb,M.S.,Burrone,J.,Toumazou,C.,andDegenaar, P..Photocyclesofchannelrhodopsin-2. PhotochemistryandPhotobiology 85:400. 135

PAGE 148

Nittala,A.,Ghosh,S.,andWang,X..Investigatingtheroleofisletcytoarchitectureinitsoscillationusinganew -cellclustermodel. PLoSONE ,2:e983. Nlend,R.N.,Michon,L.,Bavamian,S.,Boucard,N.,Caille,D.,Cancela,J.,Charollais,A.,Charpantier,E.,Klee,P.,Peyrou,M.,Populaire,C.,Zulianello,L.,andMeda, P..Connexin36andpancreatic -cellfunctions. ArchivesofPhysiologyand Biochemistry ,112:74. Notary,A.M.,Westacott,M.J.,Hraha,T.H.,Pozzoli,M.,andBenninger,R.K.P. .Decreasesingapjunctioncouplingrecoversca2+andinsulinsecretionin neonataldiabetesmellitus,dependentonbetacellheterogeneityandnoise. PLOS ComputationalBiology ,12:e1005116. Nunemaker,C.S.,Dishinger,J.F.,Dula,S.B.,Wu,R.,Merrins,M.J.,Reid,K.R., Sherman,A.,Kennedy,R.T.,andSatin,L.S..Glucosemetabolism,isletarchitecture,andgenetichomogeneityinimprintingof[ca2+]iandinsulinrhythmsin mouseislets. PLoSONE ,4:e8428. Nunemaker,C.S.,Zhang,M.,Wasserman,D.H.,McGuinness,O.P.,Powers,A.C., Bertram,R.,Sherman,A.,andSatin,L.S..Individualmicecanbedistinguishedbytheperiodoftheirisletcalciumoscillations:Isthereanintrinsicisletperiodthatisimprintedinvivo? Diabetes ,54:3517. Nyman,L.R.,Ford,E.,Powers,A.C.,andPiston,D.W..Glucose-dependent bloodowdynamicsinmurinepancreaticisletsinvivo. AJP:Endocrinologyand Metabolism ,298:E807E814. Organization,W.H.etal..Globalreportondiabetes. Pandit,S.V.andJalife,J..Rotorsandthedynamicsofcardiacbrillation. CirculationResearch ,112:849. Patterson,G.H.,Knobel,S.M.,Arkhammar,P.,Thastrup,O.,andPiston,D.W. .Separationoftheglucose-stimulatedcytoplasmicandmitochondrialNADph responsesinpancreaticisletbetacells. ProceedingsoftheNationalAcademyof Sciences ,97:5203. Pinney,S.E.,MacMullen,C.,Becker,S.,Lin,Y.-W.,Hanna,C.,Thornton,P.,Ganguly,A.,Shyng,S.-L.,andStanley,C.A..ClinicalcharacteristicsandbiochemicalmechanismsofcongenitalhyperinsulinismassociatedwithdominantKATP channelmutations. JournalofClinicalInvestigation ,118:2877. Pipeleers,D.,Kiekens,R.,Ling,Z.,Wilikens,A.,andSchuit,F..Physiologic relevanceofheterogeneityinthepancreaticbeta-cellpopulation. Diabetologia 37S2:S57S64. 136

PAGE 149

Pipeleers,D.G..Heterogeneityinpancreaticbeta-cellpopulation. Diabetes 41:777. Piston,D.W.,Knobel,S.M.,Postic,C.,Shelton,K.D.,andMagnuson,M.A.. Adenovirus-mediatedknockoutofaconditionalglucokinasegeneinisolatedpancreaticisletsrevealsanessentialroleforproximalmetaboliccouplingeventsinglucosestimulatedinsulinsecretion. JournalofBiologicalChemistry ,274:1000. Prigge,M.,Schneider,F.,Tsunoda,S.P.,Shilyansky,C.,Wietek,J.,Deisseroth,K., andHegemann,P..Color-tunedchannelrhodopsinsformultiwavelengthoptogenetics. JournalofBiologicalChemistry ,287:31804. Quesada,I.,Todorova,M.G.,Alonso-Magdalena,P.,Beltr a,M.,Carneiro,E.M., Martin,F.,Nadal,A.,andSoria,B..Glucoseinducesoppositeintracellular ca2+concentrationoscillatorypatternsinidentied -and -cellswithinintacthumanisletsoflangerhans. Diabetes ,55:2463. Ravier,M.,Sehlin,J.,andHenquin,J..DisorganizationofcytoplasmicCa2+ oscillationsandpulsatileinsulinsecretioninisletsfromob/obmice. Diabetologia 45:1154. Ravier,M.A.,Guldenagel,M.,Charollais,A.,Gjinovci,A.,Caille,D.,Sohl,G.,Wollheim,C.B.,Willecke,K.,Henquin,J.-C.,andMeda,P..LossofConnexin36 ChannelsAlters -CellCoupling,IsletSynchronizationofGlucose-InducedCa2+ andInsulinOscillations,andBasalInsulinRelease. Diabetes ,54:1798. Reinbothe,T.M.,Sa,F.,Axelsson,A.S.,Mollet,I.G.,andRosengren,A.H.. Optogeneticcontrolofinsulinsecretioninintactpancreaticisletswith -cell-specic expressionofchannelrhodopsin-2. Islets ,6:e28095. Ren,J.,Sherman,A.,Bertram,R.,Goforth,P.B.,Nunemaker,C.S.,Waters,C.D., andSatin,L.S..SlowoscillationsofKATPconductanceinmousepancreatic isletsprovidesupportforelectricalburstingdrivenbymetabolicoscillations. AJP: EndocrinologyandMetabolism ,305:E805E817. Rieck,S.andKaestner,K.H..Expansionof -cellmassinresponsetopregnancy. TrendsinEndocrinology&Metabolism ,21:151. Rocheleau,J.V.,Walker,G.M.,Head,W.S.,McGuinness,O.P.,andPiston,D.W. .Microuidicglucosestimulationrevealslimitedcoordinationofintracellular ca2+activityoscillationsinpancreaticislets. ProceedingsoftheNationalAcademy ofSciences ,101:12899. Rodriguez-Diaz,R.,Abdulreda,M.H.,Formoso,A.L.,Gans,I.,Ricordi,C., Berggren,P.-O.,andCaicedo,A..Innervationpatternsofautonomicaxons inthehumanendocrinepancreas. CellMetabolism ,14:45. 137

PAGE 150

Rodriguez-Diaz,R.andCaicedo,A..Neuralcontroloftheendocrinepancreas. BestPractice&ResearchClinicalEndocrinology&Metabolism ,28:745 756. Rodriguez-Diaz,R.,Speier,S.,Molano,R.D.,Formoso,A.,Gans,I.,Abdulreda, M.H.,Cabrera,O.,Molina,J.,Fachado,A.,Ricordi,C.,Leibiger,I.,Pileggi,A., Berggren,P.-O.,andCaicedo,A..Noninvasiveinvivomodeldemonstrating theeffectsofautonomicinnervationonpancreaticisletfunction. Proceedingsofthe NationalAcademyofSciences ,109:21456. Rorsman,P..Thepancreaticbeta-cellasafuelsensor:anelectrophysiologist'sviewpoint. Diabetologia ,40:487. Rutter,G.A.andHodson,D.J..Minireview:IntraisletRegulationofInsulin SecretioninHumans. MolecularEndocrinology ,27:1984. Serre-Beinier,V.,Bosco,D.,Zulianello,L.,Charollais,A.,Caille,D.,Charpantier,E., Gauthier,B.R.,Diaferia,G.R.,Giepmans,B.N.,Lupi,R.,Marchetti,P.,Deng,S., Buhler,L.,Berney,T.,Cirulli,V.,andMeda,P..Cx36makeschannelscoupling humanpancreatic -cells,andcorrelateswithinsulinexpression. HumanMolecular Genetics ,18:428. Sher,E.,Giovannini,F.,Codignola,A.,Passafaro,M.,Giorgi-Rossi,P.,Volsen,S., Craig,P.,Davalli,A.,andCarrera,P..Voltage-operatedcalciumchannelheterogeneityinpancreaticcells:Physiopathologicalimplications. JournalofBioenergeticsandBiomembranes ,35:687. Silva,J.R.,Cooper,P.,andNichols,C.G..Modelingk,ATP-dependentexcitabilityinpancreaticislets. BiophysicalJournal ,107:2016. Skoge,M.,Donev,A.,Stillinger,F.H.,andTorquato,S..Packinghyperspheresinhigh-dimensionalEuclideanspaces.,74:041127. Speier,S.,Gjinovci,A.,Charollais,A.,Meda,P.,andRupnik,M.7.Cx36mediatedcouplingreduces -cellheterogeneity,connesthestimulatingglucose concentrationrange,andaffectsinsulinreleasekinetics. Diabetes ,56:1078. Stefan,Y.,Orci,L.,Malaisse-Lagae,F.,Perrelet,A.,Patel,Y.,andUnger,R.H. .Quantitationofendocrinecellcontentinthepancreasofnondiabeticand diabetichumans. Diabetes ,31:694. Sto zer,A.,Gosak,M.,Dolen sek,J.,Perc,M.,Marhl,M.,Rupnik,M.S.,andKoro sak, D..FunctionalConnectivityinIsletsofLangerhansfromMousePancreas TissueSlices. PLoSComputBiol ,9:e1002923. 138

PAGE 151

Sto zer,A.,Gosak,M.,Dolen sek,J.,Perc,M.,Marhl,M.,Rupnik,M.S.,andKoro sak, D..Functionalconnectivityinisletsoflangerhansfrommousepancreastissue slices. PLoSComputBiol ,9:e1002923. Striegel,D.A.,Hara,M.,andPeriwal,V..Thebetacellinitscluster:Stochasticgraphsofbetacellconnectivityintheisletsoflangerhans. PLOSComputational Biology ,11:e1004423. Strogatz,S.H..Exploringcomplexnetworks. Nature ,410:268. Strowski,M.Z..Somatostatininhibitsinsulinandglucagonsecretionviatwo receptorsubtypes:Aninvitrostudyofpancreaticisletsfromsomatostatinreceptor2 knockoutmice. Endocrinology ,141:111. Swets,J.A.etal..Measuringtheaccuracyofdiagnosticsystems. Science 240:1285. Thorens,B..Neuralregulationofpancreaticisletcellmassandfunction. Diabetes,ObesityandMetabolism ,16S1:87. Urbano,F.J.,Leznik,E.,andLlinas,R.R..Modanilenhancesthalamocorticalactivitybyincreasingneuronalelectrotoniccoupling. ProceedingsoftheNational AcademyofSciences ,104:12554. Urschel,S.,Hoher,T.,Schubert,T.,Alev,C.,Sohl,G.,Worsdorfer,P.,Asahara,T., Dermietzel,R.,Weiler,R.,andWillecke,K..ProteinKinaseA-mediated PhosphorylationofConnexin36inMouseRetinaResultsinDecreasedGapJunctionalCommunicationbetweenAIIAmacrineCells. JournalofBiologicalChemistry 281:33163. vanderMeulen,T.,Donaldson,C.J.,C aceres,E.,Hunter,A.E.,Cowing-Zitron,C., Pound,L.D.,Adams,M.W.,Zembrzycki,A.,Grove,K.L.,andHuising,M.O.. Urocortin3mediatessomatostatin-dependentnegativefeedbackcontrolofinsulin secretion. NatureMedicine ,21:769. Wagman,A.andNuss,J..Currenttherapiesandemergingtargetsforthe treatmentofdiabetes. CPD ,7:417. Yoder,N..PeakFinder. Zarkovic,M.,Ciric,J.,Stojanovic,M.,Penezic,Z.,Trbojevic,B.,Dresgic,M.,and Nesovic,M..Effectofinsulinsensitivityonpulsatileinsulinsecretion. Europeanjournalofendocrinology ,141:494. Zhang,M.,Goforth,P.,Bertram,R.,Sherman,A.,andSatin,L..Theca2+ dynamicsofisolatedmouse -cellsandislets:Implicationsformathematicalmodels. BiophysicalJournal ,84:2852. 139

PAGE 152

Zhou,Y.,Sun,P.,Wang,T.,Chen,K.,Zhu,W.,andWang,H.5.Inhibitionof calciuminuxreducesdysfunctionandapoptosisinlipotoxicpancreatic -cellsvia regulationofendoplasmicreticulumstress. PLOSONE ,10:e0132411. 140