. . .
hwcumfifl
z.-. i: 2...?
SA...- l“:.‘i§l i
.s‘.L.l‘,-i1f ‘lt, :
. :3?!
. A, .2. z...
. 9, . E
>1 :5...
I...

2 ‘t\tlvl
fivl!fl\h
. D-

3?".
‘13»

 

v1)“

I...“ littlaiito

.\ if}!!! 3.51.13‘:
0 .Ixxlg. |u

2. 293%
‘

lo. cl
5.15:1!!!
tn.

1

Kiwis:
.

Ax:
‘11.;x...lv..?£‘uulih
.Iv‘ «0]. viii
.P‘uflw! 169!!!
I\ 3. 3:..." 1!. .r..h..i..nu.§1a«4u.tli
.\v 61.5.1 :- 9... g
5... navavuoin
I; I“!

. 9

I. u tug-.15.

$3. I} 1.31:5...
KEV...

2 ‘I‘LW
iIVnmuqv;I2!-ll
.32.. x 5534
a. ‘ 98055.}!

I.’ Ailtli‘
. 4.: six-8.19.. . A
. pvpiv‘. ‘ -«I‘ .1
.z‘: u¢§‘unl‘ll|l

.. JalaWJf...‘
vs .5»

.. 1;} I u :7... v. u v.1. V . ‘15:...UEKE 4r .. s: ...‘
. .‘Im a ,wyfle.‘ a .3 mun“. _. .3... . W 5.41%?
S uraégn ., rawrfl ”filmymwwwpyr

. . 1 . r

5...... V;

4 .
E

I'.'I

 

 

 

LIDnhn I
Michigm’l State
University

 

This is to certify that the
dissertation entitled

DETECTING INTRACELLULAR METABOLITES AND THE
RESULTING CELL FUNCTION BY MERGING
MICROFLUIDIC AND MICROTITRE PLATE
TECHNOLOGIES.

presented by

Nicole Villiere Tolan

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Chemistry

 

 

@

V’ Major Professor’s Signature

6-29 *0;

 

Date

MSU is an Affirmative Action/Equal Opportunity Employer

——7¥—7m—7 —7 7 7 7 7 7,7 7 7 7 A

 

LIBRARY
Michigan State
University

 

 

 

--—.-n-4---v---.-------—--------—----—---.-.--r-‘--—7--—

m
7*...“—

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KIIProleccaF’rosICIRCIDateDmindd

 

.4.—

DETECTING INTRACELLULAR METABOLITES AND THE RESULTING CELL
FUNCTION BY MERGING MICROFLUIDIC AND MICROTITRE PLATE
TECHNOLOGIES.

By

Nicole Villiere Tolan

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Chemistry

2009

ABSTRACT
DETECTING INTRACELLULAR METABOLITES AND THE RESULTING CELL
FUNCTION BY MERGING MICROFLUIDIC AND MICROTITRE PLATE
TECHNOLOGIES.
By

Nicole Villiere Tolan

Recent advancements have been made in our group to better mimic the blood
vessels found in vivo in order to obtain a multicellular system that incorporates the
relevant aspect of flow into an in vitro platform. By combining flow channels that
approximate these vessels in vivo with an array of wells onto a single microfluidic device,
we demonstrate an enabling system capable of merging microfluidic and microtitre plate
technologies. This platform is used to monitor various aspects of the in vivo system that
control vasodilation as it pertains to diabetes mellitus and the complications associated
with the disease. This research has focused on the specific role of the erythrocyte or red
blood cell (RBC) as it relates to hyperglycemic conditions and antioxidant status of the
cell. The role of RBCS in controlling vascular caliber is discussed, detailing the concepts
of RBC deformability and how glucose metabolism and antioxidant status affects
adenosine triphosphate (ATP) release and its ability to stimulate subsequent nitric oxide
(N 0) production, a known vasodilator.

First, this device is used to measure both intracellular glucose metabolites and the
resulting effect on RBC function. Specifically, this device is employed to monitor the
antioxidant status of the cell while Simultaneously determining the effect on ATP release,

a known stimulus of NO production in endothelial cells, which is a major determinant in

vasodilation. Here, the affects of inhibitors on the pentose phosphate and polyol
pathways for glucose metabolism, as well as the influence of hyperglycemic conditions,
on the RBC antioxidant ability and the resulting ATP release are shown.

This multicellular device, incorporating flowing RBCS and a cultured
endothelium, is used to determine the mechanisms of action for various stimuli of RBC-
derived ATP release and the subsequent stimulation on endothelium-derived NO
production. This plays an important role in the ability to monitor the final effect these
substances have on the mechanism for NO production as it pertains to the complications
associated with diabetes, Specifically, hypertension and cardiovascular disease.

Moreover, preliminary results illustrating the importance of such a flow-based
device for monitoring the specific mechanism of action for two pharmaceuticals,
pentoxyfilline (trental) and iloprost, are presented. Only by employing a flow-based
system could the mechanism for trental—induced ATP release be shown effective,
speculating that the increase vasodilation Observed in viva, acts by way of mechanical
deformation-induced ATP release.

This research demonstrates the need for incorporating such a flow-based system
into the current technologies used for drug discovery in order to combat the exponential
increase in healthcare costs and the ability to implement a more individualized method
for clinical testing. With appropriate dimensions, identical to the microtitre plates
utilized today, automation and detection methods for drug discovery and clinical

diagnosis processes could be implemented.

DEDICATION

To all of those that have affected my life in one way or another,

to help me end up here, where I am today.

ACKNOWLEDGMENTS

I would like to acknowledge Dr. David Green and Dr. Cliff Harris of Albion
College for their undying devotion to undergraduate education; it is to them I owe all
future aspirations from that point. My interest in the sciences only began upon their
teachings and has continued to grow with their guidance. To my family, I owe the utmost
appreciation and recognition of their support, as my education has always, and always
will be, a priority.

I owe much gratitude to my graduate advisor, Dr. Dana Spence, as well as my
past and present lab-mates in the Spence group. Without the opportunity to apply myself
to an unbelievably worthy area of research, I do not feel that I would have put forth
enough dedication to continue on to obtain my degree. It is to Dr. Spence I owe the most
appreciation, he has made a large contribution to the sciences and I can only foresee his
findings to play an integral role in the medicinal arena in the future. To all the students
of the Spence Group, past and present, I have enjoyed working with such an intelligent,
hard working, and self-motivated group. To those I gain great knowledge from,
especially Luiza and Ajith, I look forward to future discussions of our research as we
move onto new areas of study. Your help and support has been essential for me.

Finally to my husband Nick, your love and patience, support and thoughtfulness,
has enabled me to achieve this degree. It is to you I owe the most important moment in

my life, thus far. Together, three as one, we are unbreakable.

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................... viii
LIST OF FIGURES ....................................................................................................... ix
CHAPTER 1: INTRODUCTION ................................................................................. 1
1.1 Diabetes Mellitus ........................................................................................ 2
1.1.1 Three Main Classifications of Diabetes ....................................... 5
1.1.2 Complications Associated with the Diabetes ............................... 10
1.1.3 Current Therapies and Treatments ............................................... 13
1.2 RBCs: More than Oxygen Carriers ............................................................. 15
1.2.1 General Structure and Function ................................................... 16
1.2.2 Mechanical Deformation-Mediated ATP Release ....................... 20
1.2.3 G-Protein Coupled Signal Transduction Pathway ...................... 23
1.2.4 Endotheliu‘rn-Derived Relaxing Factor: NO ................................ 26
1.2.5 ATP-Mediated NO Production .................................................... 28
1.2.6 Glucose Uptake and Metabolism ................................................. 31
List of References .............................................................................................. 35

CHAPTER 2: MERGING FLOW INTO MICROTITRE TECHNOLOGY

2.1 Microfluidic Devices: General History and Applications .......................... 41
2.1.1 Material Advantages and Disadvantages ..................................... 43
2.1.2 Fabrication Methods in PDMS .................................................... 47
2.1.3 Photolithography: Obtaining a Silicon Wafer Master ................. 51
2.1.4 Sofi Lithographic Techniques to for Replica Molds .................... 55
2.2 PDMS uFA ................................................................................................. 56
2.3 Implementing Automation Techniques ....................................................... 57
2.4 Integrating Fluid Flow: Circulation System Mimic .................................... 59
2.5 ADMET: Drug Testing Application ........................................................... 65
2.6 Diabetes and Pre—Diabetes Screening ......................................................... 66
List of References .............................................................................................. 70

CHAPTER 3: PERSONALIZED METABOLIC ASSESSMENT

3.1 Metabolic Syndrome: Competing Pathways in Diabetes ............................ 76
3.2 PPP: Production of GSH ............................................................................. 77
3.3 Polyol Pathway: SORB Formation and Diabetic Complications ............... 80
3.4 Glucose Metabolism and the Affect on ATP Release ................................ 81
3.5 Experimental ............................................................................................... 82
3.5.1 Obtaining Isolated RBCS ............................................................. 83
3.5.2 Standard Measurements of NADPH, GSH, SORB and ATP ...... 85
3.5.3 The affect of G6PD and AR Inhibition ........................................ 100

vi

3.6 Results and Discussion ............................................................................... 105

3.6.1 Quantitative Determination within the RBC ................................ 106
3.6.2 Evaluating the Diabetic-simulated RBC ...................................... 110
List of References .............................................................................................. 114

CHAPTER 4: ADDRESSING ENDOTHELIUM WITH BLOOD COMPONENTS

4.1 Multiple Cell Types with the uFA: Cell-Cell Interactions ......................... 121
4.2 Iloprost: Prostacyclin Receptor-Mediated ATP Release ............................ 122
4.3 Experimental ............................................................................................... 125
4.3.1 Determining Iloprost-Stimulated ATP Release ........................... 126
4.3.2 Endotheliurn Culture on the uFA ................................................. 128
4.3.3 Fluorescence Determination of Intracellular NO ......................... 131
4.4 Results and Discussion: Multicellular Communication .............................. 138
List of References .............................................................................................. 145

CHAPTER 5: FUTURE DIRECTION AND APPLICATIONS

5.1 uFA: Future Applications ........................................................................... 148
5.2 Experimental ............................................................................................... 150
5.2.1 Hypoxia-Induced ATP Release: Effect on NO ............................ 150
5.2.2 C-peptide: RBC-Mediated NO Production .................................. 153
5.2.3 Required Aspect of Shear: Trental vs. Iloprost ............................ 154
5.5 Results and Discussion ............................................................................... 156
5.5.1 Vasodilation Mimic for Mechanisms of Action .......................... 162
List of References .............................................................................................. 164

vii

Table 1.1

Table 2.1

Table 2.2

LIST OF TABLES

Degree of glycemic ranges present in each individual type of diabetes
from normal blood glucose levels to a hyperglycemic condition, which
coincides with insulin dependence as a result fiom impaired glucose
tolerance or fasting glucose levels. Both type 1 and type 2 diabetics
often require insulin for blood glucose control, suggesting a resistance
to insulin as well as a reduction in insulin secretory response. ................ 4

Table from Becker et al. detailing the properties of silicon, glass,
technical thermoplastics, thermoset polymers and elastomers as it
pertains to their use in fabrication of microfluidic devices ....................... 46

Comparison of various aspects associated with microfluidic and micro-

titre plate technologies and the combination of these advantages
associated with the microfluidic array. ..................................................... 60

viii

LIST OF FIGURES

Figure 1.1 Illustration of the formation of advanced glycation endproducts
(AGES) and reactive oxygen species result from maintained
hyperglycemic conditions. This image is modified from Stephens et
a1, linking oxidative products to the cellular processes that are
speculated to result in the complications associated with diabetes ........... 19

Figure 1.2 G-protein coupled signal transduction pathway involving either the
mechanical deformation or pharmaceutical-induced ATP release as
either the G-protein coupled receptor (GPCR) or the prostacyclin
receptor (IPR) activates G-protein (Gs). Upon activation, adenylyl
cyclase (AC) will convert intracellular ATP to 3’5’-cyc1ic adenosine
monophosphate (CAMP), stimulating the phosphorylation of the cystic
fibrosis transmembrane regulator (CFTR) by protein kinase A (PKA).... 24

Figure 1.3 Illustration of the sandwich experiment conducted by Furchgott and
Zawadzki where the endothelium-derived relaxing factor (EDRF),
later to be determined as nitric oxide (NO), from the endothelial cells
of one rabbit aortic strip to the smooth muscle cells that were exposed
on the second strip, devoid of the endothelial lining. Here, the
exposed smooth muscle cells, acting as a biological detector,
underwent relaxation from the biologically active vasodilator
transferred. ................................................................................................ 27

Figure 1.4 ATP-stimulated vasodilation mechanism and the production of nitric
oxide (N 0) resulting from ATP binding to the P2y receptor located on
the endothelial lining of the vascular system. Upon activation of
endothelial nitric oxide synthase (NOS) by calcium-dependent
calmodulin (Ca2+-CaM), L-arginine is converted to L-citrulline and
NO which can then diffuse to the smooth muscle cells, binds to soluble
guanylate cyclase (sGC) and stimulates the production of cyclic
guanylate monophosphate (cGMP) which will reduce intracellular
Ca2+ and result in vasodilation. ................................................................. 29

Figure 1.5 Illustration taken from Messana et al illustrating the integral role
hemoglobin (Hb) plays in glucose metabolism to form either
antioxidants via the pentose phosphate pathway in the case of high
oxygen stress (HOS) or the production of ATP through glycolysis in
the case of low oxygen stress (LOS) ......................................................... 32

Figure 2.1 Photolithography technique employed, illustrating the steps of a) spin

coating, b) pre- and post-baking, c) transparency mask alignment, (1)
UV exposure, and e) development of SU-8 50 negative photoresist, to

ix

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

Figure 3.1

fabricate a raised featured silicon wafer master for subsequent replica
molding. .................................................................................................... 53

Illustration of process for master wafer production indicating the
multiple steps required to obtain a raised feature master. Here, the
spin-coating process results in approximately 100 pm thick photoresist
coating on the 4” silicon wafer, which is then exposed to UV light
though the negative transparency mask. After inducing
polymerization of the photoresist below the transparent portions of the
transparency mask, developer is used to remove the unpolymerized
photoresist lefi remaining on the silicon substrate leaving only the
desired raised-features behind ................................................................... 54

Flow-based microfluidic 96-well plate fabricated using the standard
dimensions from two individual PDMS layers therrnocured around a
polycarbonate membrane. Here, sample inlets and waste ports are
introduced through the bottom layer of PDMS, comprising the network
of microfluidic channels, each addressing in parallel a series of
triplicate array-like wells that are contained in the top layer of PDMS.... 62

Illustration of possible integrations of microfluidics into drug
discovery processes as discussed by Dittrich et al. highlighting the
individual applications in generation of target IN Ds through chemical
synthesis on-chip, as well as lead identification and optimization
utilizing HTS and cell sorting techniques. ................................................ 64

Schematic drawing of the uFA used for the determination of
intracellular RBC metabolites and separately, for determining ATP-
stirnulated intracellular NO production within cultured bPAECs. Here,
two layers of PDMS, containing parallel microfluidic channels and
array-like wells, are sealed around a track-etched polycarbonate
(TEPC) membrane. ................................................................................... 68

Illustration of the pentose phosphate and polyol pathways available for
the metabolism of glucose within the red blood cell. Glucose is
normally metabolized through the pentose phosphate pathway where
glucose is phosphorylated by hexokinase (HK) to form glucose-6-
phosphate (G6P) which is then oxidized to 6-phosphoglucono-5-
lactone (6GP) upon the reduction of NADP+ to NADPH. NADPH is
an essential cofactor for the production of reduced glutathione (GSH)
from its disulfide form (GS-SG), but it is also consumed in the polyol
pathway where aldose reductase (AR) converts glucose to sorbitol.
Upon the oxidation of sorbitol to form fructose by sorbitol
dehydrogenase (SD), NADP+ is reduced to form NADH ......................... 78

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

Figure 3.8

Experimental set up used to simultaneously monitor intracellular
metabolites as well as the extracellular deformation-induced ATP
release from RBC samples. Here, a displacement syringe pump is
used to deliver solutions through the underlying microfluidic channels
of the uF A while the steromicroscope images both fluorescence
intensities and chemiluminescence produced in the determination of
NADPH, GSH, SORB and ATP. .............................................................. 84

Illustration of centrifuged whole blood where the red blood cells are
separated from the platelet-rich plasma (PRP) and buffy coat that
contain the leukocytes and platelets .......................................................... 86

Illustration of the reaction used to determine the concentration of
NADPH in various standard and sample solutions employing the uF A.
Here resazurin, in the presence of diaphorase, is oxidized to the red-
fluorescent resorufin as NADPH is reduced to NADPL Resorufin has
excitation and emission wavelengths of 563 and 587 nm respectively. 88

Fluorescence image of the 6-channel uF A used to determine the
concentration of NADPH found within a 0.7% lysed RBC solution by
simultaneously monitoring the fluorescence intensities from five
NADPH standards ranging in concentration from 12.5 to 200 M.
Here, each standard and the single RBC sample solutions are flowed
through the underlying microfluidic channels for 20 min while the
resazurin contained within the wells above is reduced to the red-
fluorescent resorufin product as NADPH is oxidized to form NADP+. 89

Reaction mechanism for the determination of GSH in various standard
and sample solutions where MCB and GSH form a fluorescent product
(MCB—GSH) with excitation and emission wavelengths of 370 nm and
478 nm in the presence of glutathione-s-transferase (GST). .................... 91

Fluorescence image of the 6-channel uFA used to determine the
concentration of GSH found within a 0.7% lysed RBC solution by
simultaneously monitoring the fluorescence intensities from five GSH
standards ranging in concentration from 2.5 to 40 1.1M. As with the
NADPH determination, each standard and the single RBC sample
solutions are flowed through the underlying microfluidic channels for
20 min while the monochlorobiamine and glutathione-s-transferase
(MCB-GST) contained within the wells above forms the fluorescent
MCB-GSH product. .................................................................................. 92

Reaction mechanism of the luciferin-catalyzed chemiluminescent

reaction employed here for the determination of mechanical
deformation induced-ATP release. ........................................................... 94

xi

Figure 3.9 Chemiluminescence image, without the excitation source employed,
for the determination of ATP within standard solutions as well as from
the mechanical deformation of a 0.7% RBC sample with and without
the incubation of diamide, a known cell-membrane stiffener. Here the
luciferin and luciferase (L/L) solutions are pipetted into the 18 wells
that comprise the uF A while the solutions are flowed through the
microfluidic channels below. .................................................................... 95

Figure 3.10 Fluorescence image of the simultaneous detection of both NADPH and
GSH within standard sample solutions as well as within a 0.7% lysed
RBC solution with and without the inhibition of G6PD by
dehydroepandroesterone (DHEA). As before, solutions were flowed
through the underlying microfluidic channels and both the resazurin
(RES) and monochlorobiamine and glutathione-s-transferase (MCB-
GST) were pipetted into only one column in order to detect both
analytes from within the same solution. Images were obtained as
before except that the resorufin and DAPI filter cubes were employed
to select the appropriate wavelengths for each analyte, NADPH and
GSH, respectively. .................................................................................... 97

Figure 3.11 Images for the determination of intracellular NADPH and GSH as well
as for the mechanical deformation-induced ATP release from RBC
samples. The first six channels, employ fluorescence spectroscopy
where two columns of wells contain RES and MCB fluorescence
probes while the last six channels, employ chemiluminescence where
all three columns contain the UL reaction mixture. ................................. 99

Figure 3.12 Schematic drawing of the experimental setup used to determine RBC-
derived ATP release in a continuous flow system employing 75 um
internal diameter tubing. Here the luciferin and luciferase enzymatic
assay was employed, monitoring the chemiluminescence produced that
is directly proportional to the concentration of ATP released fiom
various samples. ........................................................................................ 102

Figure 3.13 ATP release from 7% RBCS with the inhibition of G6PD using DHEA,
resulting in a decrease in the concentration of deformation-induced
ATP release in 50 um diameter microbore tubing. Upon the pre-
incubation of RBC with a 5 mM solution of ascorbic acid (AA), a
known aldose reductase inhibitor, the reduction in ATP release due to
the inhibition of G6PD with DHEA is opposed ........................................ 104

Figure 3.14 Images for the determination of intracellular NADPH and GSH as well
as for the mechanical deformation-induced ATP release from RBC
samples. The first six channels, employ fluorescence spectroscopy
where two columns of wells contain RES and MCB fluorescence

xii

probes while the last six channels, employ chemiluminescence where
all three columns contain the UL reaction mixture. ................................. 106

Figure 3.15 Concentrations of (a) NADPH, GSH, and (b) ATP determined for
lysed and nonlysed RBC samples. Data representing figure 3.10, the
concentrations of NADPH and GSH for 0.7% RBC samples were
determined to be 31.06 i 4.12 and 22.5 d: 2.47 uM, respectively; the
concentrations decreased for each upon the incubation of DHEA to
22.9 :1: 4.63 and 14.6 i 3.1 uM, respectively. The concentration of
ATP release within the flow-based uF A was determined to be 0.54 :i:
0.04 uM for a 7% RBC sample which was reduced to 0.34 :1: 0.04 M
when incubated with diamide. Error bars represent the SEM for the
number of rabbits indicated in the figures. ............................................... 108 ‘

Figure 3.16 Data representing the intracellular a) metabolites and b) resulting
affect on cellular function as expressed though mechanical
deformation-induced RBC-derived ATP release. The fluorescence
intensity for the metabolites (NADPH, GSH, SORB) and the
chemiluminescence produced for monitoring the cellular function
(ATP release) ' are reported as average intensities of three areas
sampled within each well for RBCs in various glucose PSS solutions
(5-10 mM) as well as with the incubation of ascorbic acid (AA) and
dihydroxyepiandrosterone (DHEA) where the error reported represents
the standard deviation. .............................................................................. 110

Figure 3.17 Data representing the fluorescence intensities for monitoring the
intracellular concentrations of each metabolite, NADPH, GSH and
SORB when incubated in various glucose (GLU) containing PSS
solutions (0-25 mM). The fluorescence intensities are reported as the
average of three sampled areas within each well and the error is
reported as the standard deviation ............................................................. 112

Figure 4.1 Cross-sectional illustration of the uF A used to monitor the cellular
communication between RBCS transversing through the underlying
microfluidic channels with the cultured bPAECs within the array-like
wells. Here, RBC-derived ATP release is able to diffuse to the
bPAECs though the TEPC membrane, stimulating the production of
intracellular NO which is imaged using the fluorescence probe DAF-

F M DA. ..................................................................................................... 123

Figure 4.2 Illustration of the mechanism by which iloprost stimulates the release
of ATP from the RBC through the signal transduction pathway first
proposed by Olearczyk et al. by binding to the inisitol phosphate
receptor (IPR) which activates the g-protein (Gs) cascading events ......... 124

xiii

Figure 4.3

Figure 4.4

Figure 4.5

Figure 4.6

Figure 4.7

Figure 4.8

Schematic drawing of the experimental setup used, incorporating a t-
channel microfluidic device with the methodology first employed by
our group, to determine RBC-derived ATP release in a continuous
flow system. Here the luciferin and luciferase enzymatic assay was
employed, monitoring the chemiluminescence produced that is directly
proportional to the concentration of ATP released from various
samples ...................................................................................................... 127

Schematic illustration of the method employed to prepare a cell
suspension for the culture of bPACEs within the uFA. Here key steps
are highlighted: removal of the endothelial growth medium, replaced
with HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) for
the added buffering capacity during cell culture. The cells are
removed from the flask by incubation with trypsin/EDTA solution
followed by trypsin neutralizing solution (TNS) with vigorous
pipetting. This solution is centrifuged at 2200 rpm for 5 min at 25 °C,
upon the removal of the supernatant, the bPAEC pellet is resuspended
in equilibrated media with repeated pipetting to obtain a homogenous
cell solution. .............................................................................................. 130

Mechanism detailing the process by which 4-arnino-5-methylamino-
2’7’-difluorofluoroescein (DAF-FM) diacetate enters the cell, whereby
allong for the intracellular detection of NO production. Here, the
cleavage of the diacetate groups from DAF -F M DA occurs by an
esterase as the molecule enters the cell, further reaction of DAF-FM
with NO forms the fluorescent benzotriazole derivate with excitation
and emission wavelengths of 495 nm and 515 nm, respectively. ............. 133

Fluorescence microscopy experimental setup used to monitor the
intracellular production of NO within cultured bPAECs contained
within the wells of the uFA as a result from ATP released from the
RBCS transvering through the underlying microfluidic channels in this
flow-based system. Here, the mercury arc (Hg-Arc) lamp provides the
light that is selectively filtered by a FITC filter cube for the excitation
of the benzotrazole derivative of DAF-FM and NO which is then
imaged using a dichroic camera and Microsuite computer software ........ 134

Initial microfluidic array used to monitor the fluorescence intensity
corresponding to the intracellular production of NO as RBCS with and
without the incubation of 1 uM iloprost are pumped through the
underlying microfluidic channels. ............................................................ 136

Fluorescence images of the six channel uFA used for the
determination of intracellular NO resulting from the ATP released
from various RBC and control samples. Specifically, columns A and
B contain cultured bPAECs with and without the presence of

xiv

intracellular DAF -F M DA, respectively, while column C contains cell-
free wells. .................................................................................................. 137

Figure 4.9 Raw data obtained for the real-time analysis of the chemilurninesence
produced as a function of ATP concentration from 1 uM iloprost, 7%
RBCS, 7% RBCS with 250 nM glybenclamide and 1 uM iloprost, and
7% RBCS with 1 uM iloprost by placing the microfluidic t-channel
chip on a PMT housed within a dark box. ................................................ 139

Figure 4.10 Data representation for the quantitative determination of ATP released
fi'om 7% RBCS in the presence and absence of 1 uM iloprost and 250
nM glybenclamide using the t-channel microfluidic chip and the

luciferin and luciferase based enzymatic assay. Average
concentrations of ATP in micromolar amounts are reported where the
error is represented by the standard error of the mean .............................. 141

Figure 4.11 Representative data for the initial determination of intracellular NO
production within bPAECs that are cultured within the wells of a two
channel uFA through which a 7% RBC sample with and without the
incubation of a 1 uM iloprost solution were flowed. ................................ 142

Figure 4.12 Data representing the intracellular NO production within bPAECs
cultured in the wells of a six channel uFA, here samples were flowed
through the underlying microfluidic channels for 30 min prior to
obtaining fluorescence images. The samples included PBS buffer, 1
uM iloprost, 7% RBCS, RBCS with 1 uM iloprost, RBCS with 1 mM
glybenclamide, and RBCS with both 1 mM glybenclamide and 1 uM
iloprost (Al-A6 respectively). .................................................................. 143

Figure 5.1 Fluorescence image of the six-channel uFA for the determination of
intracellular NO production within the bPACEs as a result of hypoxia.
Here, the RBC-derived ATP release from each sample stimulates the
production of NO which is detected using DAF-FM DA and the
Olympus MVX stereomicroscope fitted with a mercury-arc lamp for
excitation and a F ITC filter cube with excitation and emission
wavelengths of 470 nm and 525 nm respectively. .................................... 152

Figure 5.2 Fluorescence image of the six-channel uFA used to monitor the
intracellular production of NO within the bPACEs cultured within the
wells of the device due to the incubation of RBCS with C-peptide.
Here, the RBC-derived ATP release from each sample stimulates the
production of NO which is detected using DAF-FM DA and the
Olympus MVX stereomicroscope fitted with a mercury-arc lamp for
excitation and a F ITC filter cube with excitation and emission
wavelengths of 470 nm and 525 nm respectively. .................................... 155

XV

Figure 5.3 Schematic drawing of the experimental set up for determining the
RBC-derived ATP release from samples incubated with trental and or
iloprost to demonstrate their opposing mechanisms of action. Here,

100 uL of the luciferin and luciferase (L/L) solution is combined in a
cuvette with 100 uL of each sample. The chemiluminescence is
measured as before where the voltage is measured in real time and
plotted at a rate of 10 measurements per second for 30 seconds onto a
computer using software that was written in-house .................................. 157

Figure 5.4 Normalized fluorescence intensity obtained from the fluorescence
image of the six-channel uFA used to determine the NO produced
within bPAECs cultured within the array of wells upon stimulation of
various RBC samples flowing through the underlying microfluidic
channels. Here, these samples include 3.5% RBCS (1.00 i 0.21),
RBCS in deoxygenated buffer (2.13 :l: 0.07), and four deoxygenated
samples with prior treatment of 100 uM (1.52 :t 0.33) and 200 uM
(0.91 i 0.07) diarnide and 20 uM (1.27 :1: 0.13) and 40 uM (1.10 :I:
0.11) niflumic acid. Here the intensities are reported as averages of
each well, providing triplicate measurements, and the error is reported
as the standard error of the mean. ............................................................. 158

Figure 5.5 Initial results of the fluorescence intensity observed when addressing
bPAECs within the uFA with RBC samples alone (1.00 :t 0.05) and
with the incubation of metal-activated C-peptide (1.31 d: 0.05), 1 mM
glybenclamide (0.94 :I: 0.11), and 1 mM glybenclamide and metal-
activated C-peptide sequentially (1.10 :1: 0.20). Here the intensities are
reported as averages of each well, providing triplicate measurements,
and the error is reported as the standard error of the mean ....................... 160

Figure 5.6 Summarized data for the measurement of RBC-derived ATP release in
a non-flow system employing the luciferin and luciferase assay. Here,
ATP concentrations in micromolar are reported for 3.5% RBCS
(0.0981 :1: 0.0063 uM), with the incubation of each 5 uM trental
(0.1126 i 0.0218 uM), 1 uM iloprost separately (0.3265 1- 0.0702 uM)
and combined with 5 uM trental (0.4628 :1: 0.0935 uM). ......................... 161

xvi

CHAPTER 1: INTRODUCTION

The motivation behind the work presented here is the incorporation of analytical
methods to determine biologically relevant processes and mechanisms. Specifically,
advancements have been made to better mimic the in vivo resistance vessels by
employing an in vitro platform in order to conduct various analytical measurements on
blood analytes. Specific attention has been given to perform complete and appropriate
sample handling, analytical measurements, and accurate data interpretation of biological
analyses of bulk cell samples.l Here, biological systems related to diabetes and the main
complications associated with the disease are monitored and discussed in great detail.
Research has been performed and methods have been developed to monitor various
intracellular metabolites as well as intercellular communication between multiple cell
types, which may play a key role in complications associated with diabetes, especially
hypertension.

Microfluidic technology has fallen short of fulfilling the promises of integrating
high throughput analyses within a flow-based platform. It is speculated that this may be
due to a lack of standard dimensions which are, currently in place for microtitre plates.
Standardized dimensions has allowed for the incorporation of automation technologies,
such as liquid handling and plate readers. The goal of this work is to incorporate the
requirements for individual assays onto one microfluidic device, examining multiple
glucose metabolism pathways in a high throughput system, while also monitoring overall
cell function as it pertains to cell-to-cell communication. The research aims were to

demonstrate the integration of multiple pathways involved in glucose metabolism,

providing more information as to the mechanism of cell function that contributes to
diabetic complications.

The development of a high throughput, simultaneous method of addressing
various aspects of a single RBC sample will allow for a specific diagnosis of the glucose
metabolism malfunction that results in inadequate blood flow that is typically observed in
diabetic patients. The specific aims concentrated on evaluating both the pentose
phosphate and polyol pathways by inhibiting glucose 6-phosphate dehydrogenase
(G6PD) and aldose reductase (AR), while monitoring the resulting levels of reduced
nicotinamide adenine dinucleotide phosphate (NADPH), glutathione (GSH), sorbitol
(SORB) as well as the resulting released ATP. This research also demonstrates the
ability to then monitor on a separate microfluidic array (uFA) the intracellular NO
production within cultured bPAECs as a result of pharmaceutically and shear-induced

ATP release from the RBC.

1.1 DIABETES MELLITUS

According to the American Diabetes Association (ADA), diabetes is classically
presented with symptoms of polyuria (excessive urination), polydipsia (excessive thirst),
and unexplained weight loss. Diabetes is classified as a group of metabolic disorders that
may also be presented with polyphagia (excessive hunger), blurred vision, growth
impairment, ketoacidosis, nonketotic hyperosmolar syndrome and often times results in
susceptibility to infection. It is caused by individual instances of reduced insulin
secretion or insulin efficacy, or a combination of the two and it is often difficult to

determine the original cause of hyperglycemia (high blood sugar). The two accepted

methods for diagnosing diabetes are the fasting plasma glucose test (FPG), where no
caloric intake is sustained for 8 hrs prior to glucose measurement, and the oral glucose
tolerance test (OGTT), where plasma glucose levels are measured two hours after
administering 1.75 g glucose per kg body weight. Diabetes is confirmed when FPG
levels are above 126 mg/dL or when plasma glucose levels are at or above 200 mg/dL for
an OGTT. Another common marker of diabetes is the glycosylated hemoglobin (AlC)
test. Levels less than 7% are associated with microvascular disease prevention and
indicates the amount of hemoglobin glycated due to elevated plasma glucose levels.2 An
AlC level of 6% is synonymous to an average blood plasma glucose level of
approximately 120 mg/dL and is used to gauge control of blood glucose in diabetes
patients.3

Pre-diabetes, as described by the Center for Disease Control and Prevention
(CDC), is a condition in which a person has elevated risk of developing type II diabetes,
heart disease and stroke. Pre-diabetes is indicated by resulting FPG and OGTT levels
between 100-125 mg/dL and 140-199 mg/dL, respectively. The resulting impaired
fasting glucose (IFG) and impaired glucose tolerance (IGT) are associated with metabolic
syndrome and typically accompanied by normal levels of AlC. In 2007, the CDC
reported that 23.6 million people in the United States were affected by diabetes and
concluded that 17.9 million were diagnosed, leaving 5.7 million undiagnosed and
unaware of their disease.4 The ADA outlines the various degrees of blood glucose levels
present in the three main classifications of diabetes, type 1, type 2, and gestational

diabetes, shown in table 1.1.5

.omcommoc @8053 53%: E gouge»: a
we :03 8 533m 9 853.32 a mcflmowmzm 405:8 3.82m wooE Sm 5155 2.532 auto 83236 N cab can _ cab 50m .232
3.0on waumfl 8 8:828 882w commas: 80¢ :32 a ma "vengeance 5:55 2:3 82058 £033 £2558 288.3323:
m 8 £32 382m @003 R88: Set moconflu me 093 3:239: :08 E “:8on mews“; omfioobm mo 8hon #4 033,—.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n u 3:539 3:033va

1 39¢. omuooam 550

l t m 25

n w — can.

3.133 8% 7539 no,“ mafia—.3 $2 .2 -2 V
mew—«=78 533 war—1&2 5.35 F358 82 38: MM. am WEE
G o .9 m c . EH coax—awe.— omeoflm flan—.82 / weak
8
95:02 83:39 8:823. 3820 gush—EMF momSm /
«geek—whofifim mgoobwogoz

 

 

 

 

 

1.1.1. THREE MAIN CLASSIFICATIONS OF DIABETES

Insulin-dependentdiabetes mellitus (IDDM), commonly known as type 1 diabetes
and previously referred to as juvenile-onset diabetes, is associated with a lack of insulin
production, a hormone necessary for glucose clearance from the blood stream. Type 1
diabetes represents only about 5-10% of all the reported cases of diabetes. It typically
develops in children, however, it can occur at any age, ultimately resulting in the
necessity of insulin therapy for survival. In the case of type 1 diabetes, insulin is no
longer produced by the beta cells (B-cells) of the pancreas due to the autoimmune

destruction of these cells.5'7

The trigger for this autoimmune response leading to the
damage of the B-cells has been explained as a polygenetic dispositions resulting from
multifactoral environmental effects, which may include viral infection.9'” The B-cells
are contained within the islets of Langerhans, which are located in the pancreas, serving
to produce and secrete insulin into the blood stream in response to elevated blood glucose
levels.

It is thought that the susceptibility to autoimmune destruction of the B-cells is due
to upregulation of both Major Histocompatability Complex (MHC) class I molecules and
interferions as a result of viral infection-induced inflammation. The MHC is a region of
the human genome that is responsible for encoding for proteins that are involved in the
immune system, including autoimmunity. Interferions are proteins produced by the cells
that comprise the immune system and are responsible for cell signaling that is necessary

for the destruction of foreign bodies or malfunctioning cells which commonly include

viruses, parasites, and tumor cells.

It has been reported5 that 85-90% of all patients diagnosed with type 1 diabetes
presents with at least one autoantibody marker, either to insulin, glutamic acid
decarboxylase (GAD) or tyrosine phosphatases (IA-2 and IA 20). GAD and tyrosine
phosphatases are class of enzymes that have been of particular interest to research of
diabetes because these enzymes are typically targeted by autoantibodies in diabetes
patients. There are many genetic predispositions for B-cell destruction leaving type 1
diabetics prone to various other autoimmune disorders such as, but not limited to,
Grave’s disease, Hashimotos’s thryroiditis, Addington’s disease, vitiligo, celiac sprue,
autoimmune hepatitis, myasthenia gravis, and finally, pernicious anemias The rate of B-
cell destruction is variable, typically vigorous in infants and children, where as a much
slower reduction of B-cell mass is observed in adults.

As B-cell mass decreases and insulin production is reduced, ketoacidosis,
typically symptomatic of type 1 diabetes, occurs as the formation of ketone bodies are
produced via ketosis. Here, the body resorts to metabolism of fatty acids and the
deamination of amino acids, resulting in an eventual acidification of the blood leading to
the symptoms associated with diabetes, as well as nausea and vomiting, abdominal pain,
shortness of breath, sweet fruit- or acetone-scented breath as well as confusion. The
amount of insulin secreted is typically determined by indirectly measuring the amount of
C-peptide within the peripheral circulation.

C-peptide has been used as a means to classify diabetes type because it is secreted
in equimolar amounts with insulin from the pancreatic beta cells. Proinsulin is produced
in the B-cells and is comprised of insulin and C-peptide, which is the C-chain that

connects the A- and B-chains of insulin. Comprised of 31 amino acids, C-peptide is

cleaved from the active form of the hormone as it is secreted into the pancreatic islets. It
is measured to determine insulin secretion for several reasons, but importantly, because
its concentration within the peripheral circulation reflects the original proinsulin
production due to increased half-life, compared to insulin, which is typically extracted by
the liver due to its bioactivity.

Non-insulin dependent diabetes mellitus (NIDDM) or type 2 diabetes, previously
described as adult-onset diabetes, is classified as an insulin resistance, where the insulin
that is produced is not properly utilized by the cells comprising the peripheral tissues of
the body. This lack of insulin activity results in hyperglycemia, increasing both FPG and
AlC levels. Type 2 diabetes accounts for approximately 90 to 95% of all cases of
diabetes and can also result in a reduction of relative insulin secretion as the disease
progresses, further requiring antihyperglycemic medication or insulin along with an
insulin sensitizer.

While autoimmune B-cell destruction does not occur in type 2 diabetes, B-cell
dysfimction due to hyperglycemia may be responsible for the reduced insulin secretion

12 Measuring

and can be at least partially restored upon lowering blood glucose levels.
the concentration of B-cells within the pancreas often varies across different methods,
however, the concentration of B-cells in the pancreas of type 2 diabetic patients is
characteristically reduced by approximately 40-60% when compared to those found
within the pancreas of healthy non-diabetic patients.13 This decrease is thought to be due

to an upset of the balance between the loss (through apoptosis) and regeneration (through

replication or differentiation of precursors) of B-cells. Commonly, patients diagnosed

with type 2 diabetes are obese or have an increased percentage of body fat that is
associated with insulin resistance.

Hyperglycemia typically presents in a more gradual fashion in type 2 diabetes and
rarely results in ketoacidosis, although when it does occur, it typically is due to an onset
of another illness or infection. Unlike type 1 diabetes, the genetic profiles that are
associated with type 2 diabetes are more complex and are currently obscurely defined.
The risk of developing type 2 diabetes is increased with age, obesity, lack of physical
activity and is also elevated among individuals of certain racial or ethnic background
(Hispanic/Latino Americans, Native Hawaiians, American Indians, Asian Americans,
Afiican Americans and Pacific Islanders).5

Gestational diabetes mellitus (GDM) is defined as any sort of glucose intolerance
that is first diagnosed during pregnancy and is described as a disorder of carbohydrate

metabolism.14

The resulting glucose intolerance and lack of insulin secretion typically
only occurs during the term of pregnancy and are characteristically alleviated upon
delivery. However, the risk of sustaining glucose intolerance in the form of type 2
diabetes, even after child birth, is elevated in comparison to non-GDM individuals.15
GDM is more frequently diagnosed in women of African American, Hispanic/Latino
American, and American Indian ethnic background. It was reported that on average, 4%
of all pregnancies in the US are complicated by GDM, although it represents 90% of all
pregnancy complicationss. GDM is usually diagnosed within the third trimester of
pregnancy where diet modification and or insulin therapy can control hyperglycemia.

Treatment for hyperglycemia during pregnancy is necessary to avoid complications in the

infant as well as reduce the risk of spontaneous abortion.

Diabetes outside of these three main classifications include maturity-onset
diabetes of youth (MODY), genetic predisposition to ineffective insulin action,
endocrinopathies, drug— or chemical-induced diabetes, and those affected by damage to,
or disease of the pancreas, representing an estimated 1-5% of all cases of diagnosed
diabetes. In the case of MODY, patients are affected by a monogenetic defect in B-cell
function that is typically displayed as hyperglycemia before the age of 25 years. Here,
there are minimal or no defects in insulin action and the impaired insulin secretion is
most commonly due to a mutation in chromosome 12 (the hepatic transcription factor
(HNF)-lor), or a mutation in chromosome 7p (the glucokinase gene).5 In the past, the
genetic defect in insulin action was described as “type A” diabetes, and is a metabolic
disorder of hyperinsulinemia and modest hyperglycemia, resulting in severe diabetes.

Endocrinopathies, or diseases that are dependent upon the endocrine gland, may
result in an insulin resistance due to elevated levels of various hormones such as grth
hormone, cortisol, glucagon and epinephrine. Here, the diabetic condition is typically
found with some sort of pre—existing defect in insulin action and once hormone levels are
balanced, the hyperglycemic condition is resolved.5

Various drugs, such as nicotinic acid, glucocorticoids and oc-interferons have been
reported to impair insulin effectiveness. Numerous infections and viruses have also been
correlated with an insulin inactivity, the most common of which are, congenital rubella,
coxsackievirus B, cytomegalovirus, adenovirus and mumps. Finally, damage to the
pancreas either through pancreatitis, physical trauma, infection, or pancreatic carcinoma,
has been shown to result in insufficient insulin production and hyperglycemia. It has

been noted5 that this damage need not be extensive to greatly diminish insulin secretion

and must involve another pathway besides physical reduction in B-cell mass because

cancer of only a small portion of the pancreas has been shown to result in diabetes.

1.1.2 COMPLICATIONS ASSOCIATED WITH THE DIABETES

Diabetic patients often suffer from one or more related complications including,
but not limited to, blindness, kidney damage, nerve damage, periodontal disease,
pregnancy complications, heart disease, stroke, and cardiovascular disease. Detailed
information regarding the complications associated with diabetes and their US statistics
are derived from the Center for Disease Control and Prevention’s Diabetes National Fact
Sheet for 2007.4

Diabetic retinopathy is the leading cause of blindness in diabetics 20-74 years of
age and was reported to result in 12,000 to 24,000 new cases of blindness each year. It is
classified as non-inflammatory damage to the retina, resulting from dysfunction and
eventually bleeding from the capillaries of the eye. It has been reported that nearly all
patients diagnosed with type 1 diabetes and more than 60% of those diagnosed with type
2 diabetes will be diagnosed with retinopathy within 20 years of diagnosis.16

Nephropathy, or kidney disease, is also predominantly associated with diabetes
as it is the leading cause for renal failure, eventually resulting in the requirement for
chronic dialysis or kidney transplant. By monitoring the concentration of albumin (the
most abundant blood plasma protein) and creatinine (produced as creatinine phosphate
from the muscle, is broken down) excretion in the urine, the severity of renal failure can

be deterrrrined.l7

10

It was estimated in 2007 that 60-70% of people with diabetes have mild to severe
peripheral neuropathy, or nerve damage resulting in numbness or pain in the hands and
feet. In severe cases, neuropathy can progress to form ulcers where the need for
amputation occurs. It was reported that more than 60% of nontraumatic amputations are
associated with diabetes-related neuropathy. Autonomic neuropathy, results in
gastrointestinal and sexual dysfunction. Periodontal (gum) disease has been reported to
affect 33% of diabetics where it is twice as common in young adults, and three times
more likely to occur as poorly controlled A1C levels increase (>9%).

Approximately 5-10% of pregnancy complications due to GDM result in serious
birth defects and an increase in the number of spontaneous abortions resulting from
poorly controlled glycemia prior to conception or during the first trimester of pregnancy.
As in the case of GDM, as with diabetes in general, there has been a reported increased
incidence of vascular disease that is presented as endothelial dysfunction, increased
oxidative stress, and general over expression of inflammatory responses. '8

Oxidative stress is defined as a disruption in the balance between the production
of reactive oxygen species (ROS) and the antioxidant defenses. Superoxide anion (02'),
hydrogen peroxide (H202), hydroxyl radical (OH‘) and peroxynitrate (OONO') are all
examples of ROS that have been correlated to the oxidation of membrane lipids that are
thought to play an integral role in the development of atherosclerosis.” Research has
indicated increased concentrations of ROS may impair NO signaling which could then
reduce endothelium-dependant vasodilation, increase platelet aggregation and the

expression of adhesion molecules leading to neutrophil migration to the endothelimn and

proliferation of smooth muscle cells.

11

There are long-term complications that are associated with the continued
hyperglycemic conditions that are found in diabetes, and often results in a combination of
the complications associated with hypertension. Along with hypertension, or high blood
pressure, are increased occurrences of cardiovascular, peripheral arterial and
cerebrovascular disease. The complications typically associated with type 2 diabetes are
more frequently due to micro- and macrovascular disorders. Microvascular disease is a
condition of the resistance vessels where plaque formation, constriction of the arteries, or
damage to the lining of these vessels can occur.

It has also been reported that there is a substantial increased risk for
cardiovascular disease in people with type 1 and type 2 diabetes compared to healthy
controls.“19 Moreover, among those diagnosed with diabetes, at least 65% will die from
a heart attack or stroke. According to the CDC, cardiovascular diseases include various
forms of heart and blood vessel dysfunctions such as high blood pressure, coronary heart

disease, stroke, and heart failure.4 A heart attack, or myocardial infarction, is described

as the blockage of blood flow to a portion of the heart and upon lack of oxygen (02),
often results in permanent damage to the heart muscle. The blockage of proper blood
flow is typically a disturbance of plaque buildup in a coronary artery, which is due to
atherosclerosis.

Atherosclerosis is defined as the narrowing of the coronary arteries due to a
combination of buildup of plaque, deposits of fat, cholesterol and platelets within the
vessels comprising the circulation system, and hardening of the arteries due to
calcification. Stroke, resulting in a loss of brain function is due to a disturbance of blood

flow to the brain which results either from a blood clot or bleeding into the tissue of the

12

brain from a break in the blood vessel. It has been suggested in the literature that an
increase in oxidative stress, as with patients diagnosed with diabetes and sustained
hyperglycemia, is directly correlated with an adverse affect on the vascular system and an
increase in atherosclerosis.”

Dyslipidemia, a condition in which there is an increased level of lipids found in
the circulation system, often arises due to prolonged elevated levels of insulin, as in the
case of type 2 diabetes. The elevated levels of lipids within the blood stream combined
with hypertension, increases the risk of type 2 diabetics to have two to five times higher
risk of cardiovascular disease.4 Uncontrolled hyperglycemia has been associated with
biochemical imbalances that can lead to diabetic ketoacidosis as well as hyperosmolar

(nonketotic) comas and put individuals at higher risk of illnesses, increasing the

prognosis and likely hood of mortaility.

1.1.3 CURRENT THERAPIES AND TREATMENTS

It has been reported4 that patients diagnosed with type 1 or type 2 diabetes,
combined, take insulin only (14%), insulin and oral medication (13%), oral medication
only (57%) or neither (16%). The US department of Health and Human Services recently
compared the various oral antihyperglycemics classifying each as biguanides,
sulfonylureas, thiazolidinediones (TZDs), and finally, alpha-glucosidase inhibitors.
These oral medications all work to reduce blood glucose levels by either inhibiting the
glucose production in the liver or carbohydrate absorption in the small intestine or by

increasing insulin secretion or glucose uptake. These medications more effectively

13

decrease hyperglycemia, which is monitored by AlC levels, by using combinatorial
therapies.

The Diabetes Control and Complications Trail (DCCT)20 involved 1,500 type 1
diabetic patients and evaluated the affect of strict blood glucose control, maintaining
levels approximate to those found in healthy non-diabetics, on complications such as
retinopathy, neuropathy, and cardiovascular disease. It was determined that when
maintaining norrnoglycemic levels by increased frequency in insulin injections, as
compared to conventional treatment protocols, there was a correlated decrease in
retinopathy, microalbuminuria and clinical neur0pathy. It was also reported that the main
side effect of stringent regulation of glucose levels was severe hypoglycemia, 11 people
died, and 32 people were suspended from the clinical trial due to an evaluation deeming
continued participation hazardous.

In a separate, but similar study, the Epidemiology of Diabetes Interventions and
Complications (EDIC) research group performed a long-term (17 year) follow up study
of the original DCCT.l7 It was determined that long-term risk of cardiovascular disease,
including the incidence of non-fatal myocardial infarction, stroke, and death from
cardiovascular disease, was reduced by more than 50% with the use of the previously
adopted (DCCT) intensive treatment protocol when compared to the conventional
treatment protocol. Specifically, there was a significant increase of developing
cardiovascular disease in those patients administered conventional treatment with age,
duration of diabetes, presence of retinopathy, current smoking status, body mass index,

lipoprotein cholesterol, AlC, and albumin secretion levels.

14

1.2 RBCS; MORE THAN JUST OXYGEN CARRIERS

While one of the main functions of the RBC is to transport 02 to the tissues
surrounding the macro- and microvasculature system and uptake the carbon dioxide
(CO2) produced as waste from respiring cells, this highly specialized cell may also be
involved in glucose clearance from the blood stream, and functions as a component
necessary for vasodilation. The RBC, as reviewed by Bossi and Giardina,2| is a
simplistic biconcave cell that lacks a nucleus and intracellular organelles. Commonly
referred to as a bag of hemoglobin (Hb), the RBC contains three main classes of this most
abundant tetramer, the major adult Hb (HbA), the minor adult Hb (HbA2) and fetal Hb
(HbF) each of which are comprised of a pair of or and B subunits, except for HbF which
contains two or and two y subunits. Both the or and [3 subunits contain a highly conserved
hydrophobic heme pocket that binds to an iron atom through four pyrrole nitrogens and
one axial bond to a histidyl residue of the polypeptide chain. The sixth bond to the
ferrous iron atom (F e”) is in the axial position and serves as the O2 binding site. There
are two separate intersubunit contacts, the al -[31 and the orl -B2, where the differing
subunit connections are hydrophobic in nature as compared to the two identical chain
interactions, being polar.

The ability of hemoglobin to bind to O2 is characterized by a sigrnoidal binding
curve that results from cooperative binding between each of the four 02 binding sites.
There are two separate conformations of hemoglobin, the tense (T) state deoxygenated
structure and the relaxed (R) oxygenated structure. The T-state has a low affinity for O2
and a high affinity for molecules found in the bloodstream such as carbon dioxide,

organic phosphates, and protons. The alkaline Bohr effect, provoking the release of

15

protons above a pH of 6, results in Hb readily binding to O2, shifting the conformation to
the high-affinity relaxed R-state. The binding of Hb to the CO2 that is produced from
respiring tissues is inversely, yet equally, as important as the transport of 02. The
transport of CO2 is achieved through either a direct binding to Hb or indirect method of

hydrating CO2, both of which are reversed in the lungs.

1.2.1 GENERAL STRUCTURE AND FUNCTION

When transversing the microvasculature system, red blood cells (RBCS), with
inner diameters of 5-8 pm, are able to undergo cellular deformation to pass through
capillaries with internal diameters ranging from 3-8 pm.”23 This cellular deformability
is highly dependent on the cytoskeletal membrane proteins as well as the surface area to
volume ratio. The RBC membrane is comprised of approximately 52% protein, 40%
lipids, and 8% carbohydrates, all of which are important for the biconcave disc shape that
allows the RBC to undergo considerable deformation without affecting cellular function
or vitality. The lipid membrane is composed of a bilayer where the core is hydrophobic
from the two fatty acid chains facing in towards one another, while the hydrophilic polar
head groups face outward, towards the lumen, and inward, towards the cytoplasm of the
cell.

There are over 100 different proteins that make up the RBC membrane, and
include spectrin, adducing, protein 4.1, and ankyrin. These proteins are oriented such as
to maintain membrane asymmetry, which is only compromised by membrane lipids that
can undergo transverse diffusion and undergo outward translocation. The outer leaflet is

mostly composed of phosphatidylcholines and sphingomyelins, where as the inner leaflet

16

of the cell has highest concentrations of phosphatidylserines and
phosphatidylethanolarnines. It is the job of both the ATP-dependant translocases and the
arrrinophospholipid translocase, to revert back any phospholipids and
aminophospholipids, respectively, that undergo outward translocation, playing an
important role in apoptosis and cellular signaling for phagocytosis. Apoptosis, or
programmed cell death, of the RBC can occur through the process of erythrophagocytosis
where various morphological changes to the RBC result in identification and ingestion by
macrophages. This process is mediated by oxidative stress, activating scramblase-l,
controlling outward translocation of phosphatidylserines, and inhibiting
arninophospholipid translocase.24 It has been stated” that any sort of gene mutation or
oxidative damage that results in a defect of any of the membrane proteins involved in the
cellular membrane will change the cell volume and affect cellular deformation and
influence premature cell death.

A significant positive correlation between membrane lipid peroxidation, induced
through in vitro hyperglycemic conditions, and osmotic fragility has also been reported.25
The relationship between these compromising affects on RBCS were shown to be directly
dependent on oxidant stress due to the reversibility upon addition of fluoride, a known
glucose metabolism inhibitor, vitamin E, a known antioxidant, para-
chloromercurobenzoate and metyrapone, both known to inhibit the oxidizing manner of
the cytochrome P-450 system, as well as dimethylfuran, diphenylamine, and thiourea, all
serving as 02 radical scavengers.

It has been demonstrated that there is a significant increase in glycosylation of

RBC membrane proteins obtained from diabetic patients in comparison to healthy non-

17

diabetic controls.26 Among the membrane proteins affected, spectrin was shown to
undergo the most oxidative damage resulting from the free radicals generated when
proteins were subjected to sustained hyperglycemic conditions. Spectrin comprises the
bulk of the membrane skeleton of the RBC and it is the hexagonal lattice of the tetramer,
associated with integral actin filaments, which allows for a rugged, yet flexible cellular
membranezz’24 Studies have shown that structural abnormalities of the RBCS obtained
from people with type 2 diabetes include distorted distribution of spectrin moieties and
membrane-bound insulin receptor.27 Spectrin is a polypeptide that when subjected to
oxidation, resulting in the formation of disulfide bonds between chains, has been shown
to reduce shear-based deformation.

In the case of type 2 diabetes the hyperglycemic conditions as illustrated in figure
1.1, result in an increase in oxidant stressors leading to glucose oxidation, protein
glycation and an increase in advanced glycation endproducts (AGES).19’28 The
morphological changes that take place in the RBC due to oxidant stress can lead to
decreased deformability, increased membrane viscosity, increased aggregation,
membrane lipid peroxidation, and phosphatidylserine extemalization. Scanning electron
microscopy was used to determine the morphological changes of RBCs obtained from
type 2 diabetics when compared to healthy donors, illustrating both the acanthocytes,
typically observed in patients with hyperinsulinemia, as well as cup forms, which are

27 The rearrangement and

often found in patients with cardiovascular disease.
redistribution of spectrin throughout the cytoskeleton of RBCS obtained from patients

with diabetes was observed using fluorescence spectroscopy and these findings suggest

18

 

Hyperglycemia

l

Maillard reaction

I

Schiff base

i rearrangement

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Amadori product (Fructoselysine)

 

1 oxidation/glycosylation
AGES

A

TLipid peroxidation T Platelet aggregation T 02' AGE modified albumin

 

 

 

 

Figure 1.1 Illustration of the formation of advanced glycation endproducts (AGES)
and reactive oxygen Species result from maintained hyperglycemic conditions. This
image is modified from Stephens et al, linking oxidative products to the cellular
processes that are speculated to result in the complications associated with diabetes.

19

that oxidative damage could lead to better understanding the severity of the associated
complications.

It has been demonstrated that there is a significant reduction in the deformation of
RBCS obtained from patients with diabetes when compared to those obtained from
healthy, non-diabetic controls.23‘26 Within the circulation, RBCS exhibit non-Newtonian
flow and concentrate in the center of the circulation vessels thereby minimizing the
resistance to flow. This phenomenon is known as the Fahraeus-Lindqvist effect.29 This
effect leads to a decrease in the viscosity of the RBC solution as it passes through
decreasing vessels. It has been the main explanation for the ability of the RBC, with
approximate inner diameters ranging from 5-8 pm, to undergo mechanical deformation
against one another within the vasculature system, having inner diameters ranging
anywhere from 75 to 100 um.

Under significant shear stress, the RBCS form ellipsoids from which the
elongation, a measure of deformability, can be derived.22 The deformability of RBCS
subjected to various thiol reactive SH-reagents, resulting in disulfide formation between
membrane proteins, was determined by Bossi and Gardinia. It was concluded that
diamide, specific for oxidizing SH groups in spectrin, only induces a small fraction of
dimerization, yet lends to a 50% decrease in RBC deformation. This exemplifies that
even a small change in the number of disulfide bonds, from the native form of spectrin,

will greatly affect the shear resistance of the RBC membrane.

1.2.2 MECHANICAL DEF ORMATION-MEDIATED ATP RELEASE

20

Research has shown that RBCS play an essential role in the circulation, not only
in carrying 02 throughout the body, but also in the mechanism of vasorelaxation of the
systemic and micro circulation. It has been suggested that the ability of the RBC to
communicate the need for 02 from the tissue to the vasculature, increasing blood flow
and ultimately the supply of O2 is mediated by the ability of the RBC to release adenosine
triphosphate (ATP).3O'33 It has been well established that the ability of RBCS to release
ATP upon mechanical deformation is a necessary factor for the stimulation of nitric oxide
(NO) synthesis in the endothelium lining the luminal surface of the vasculature which
will be detailed in a subsequent chapter.

It was shown that in isolated perfused rabbit lungs, the presence of rabbit RBCS
was required in order to induce the stimulation of NO, resulting in vasodilation.34 Further
validating a relationship between RBC-derived ATP release and the control of vascular
caliber, studies have shown that upon interluminal addition of 10 uM ATP into the
arterioles of a hamster retractor muscle, significant vasodilation occurred30 and
separately, studies have also been performed, demonstrating that the presence of RBCS
was necessary for a resulting increase in vascular caliber under low 02 tension
conditions.33

It has been concluded that the RBCS ability to release low micromolar
concentrations of ATP, as a result of low 02 tension as well as a change in pH, will

30’” thereby altering the distribution of

increase vascular caliber and increase blood flow,
RBCS throughout the vascular system and ultimately increasing the supply of O2 to

localized tissues where it is in needed. Further studies, applying ATP to the intraluminal

side of arterioles and the resulting vasodilation, demonstrated that the release of low

21

nanomolar amounts of ATP from RBCS in vivo is responsible for the vasodilation
mechanism of resistance vessels.31

ATP release from RBCS has been previously measured in a continuous flow
system by employing the luciferin/luciferase assay, which has also been employed in this

research and will be discussed in detail in subsequent chapters.35

By using various
internal diameter fused silica microbore tubing and by varying the length of tubing,
increasing ATP release was shown to correlate with the increased shear stress. This
flow-based system, employing tubing with inner diameters ranging from 25-75 pm
closely mimics the arterioles (10-100 um) found in vivo, where the primary control of
vascular resistance takes place. This experimental set up is an advancement in
comparison to previous filtration methods for measuring deformation-induced ATP
release. Filtration methods require an off-line approach to determine ATP release,
employing the luciferin/luciferase assay, inherently reducing sample throughput. Also
this method strays from an in vivo approximation since it is only applicable to
hematocrits (volume percent RBCS in buffer) above 5% and less than 20% to allow for
proper signal intensity and to prevent clogging, respectively.”36

The mechanical deformation of RBCS has also been evaluated using a pressure-
based system in which varying hematocrit RBC samples as well as chemically stiffened
RBC samples were compared to each other and correlated to the mechanically-induced
ATP release.29 By replacing the filter, as used in previously mentioned studies, with a 15
cm section of 25 um microbore tubing, the RBC deformation and resulting ATP release

was determined. As the number of RBCS within the tube increases (as with increasing

hematocrit) the RBCS are able to absorb pressure placed on the system as the flow is

22

increased to 10 uL/min. Thus, the time required for the system to reach half pressure of
equilibrium is proportional to the deformability of the RBC sample and the subsequent

deformation-induced ATP release was measured in the effluent.

1.2.3 G-PROTEIN COUPLED SIGNAL TRANSDUCTION PATHWAY

The proposed mechanism for the release of ATP from RBCS includes the
activation of the receptor-mediated membrane-bound heterotrimeric G-proteins (G5 or G)
(figure 12).”39 This occurs as either mechanical deformation of the RBC membrane
changes the conformation of the G-protein coupled receptor (GPCR) or through the
pharmaceutical stimulation as iloprost can bind to the prostacyclin receptor (IPR).” It
has been reported in the literature that activation of either GS or G; will result in increased
3’5’-cyclic adenosine monophosphate (CAMP) upon activation of adenylyl cyclase (AC),
subsequently initiating the phosphorylation of the cystic fibrosis transmembrane
conductance regulator (CFTR) by protein kinase A (PKA).40 Sprague et al. effectively
demonstrated this Signal transduction pathway by monitoring the reduced ATP release
from RBC when G8 was inhibited with pertussis toxin.

Separately, the dependence of ATP release (in response to mechanical
deformation) on the CFTR protein was studied by monitoring the ATP release from
RBCS that were treated with two chemically dissimilar CFTR inhibitors (glybenclamide
and niflumic acid) as well as RBCS obtained from cystic fibrosis (CF) patients.36 It was
demOnstrated that when CFTR is either chemically inhibited or genetically deficient, as
in the case of CF patients, there is a decrease in the concentration of ATP that is released

when RBCS are passed through filters ranging from 5-12 um in diameter. Importantly, it

23

Mechanical
Iloprost Deformation

 

Figure 1.2 G-protein coupled Signal transduction pathway involving either the
mechanical deformation or pharmaceutical-induced ATP release as either the G-protein
coupled receptor (GPCR) or the prostacyclin receptor (IPR) activates G-protein (6,).
Upon activation, adenylyl cyclase (AC) will convert intracellular ATP to 3’5’-cyclic
adenosine monophosphate (CAMP), stimulating the phosphorylation of the cystic fibrosis
transmembrane regulator (CFTR) by protein kinase A (PKA).

24

was also demonstrated here by monitoring red cell transit time (RCTT), that there was no
decrease in RBC deformability, concluding that the ATP release is directly dependant on
the action of CFTR. It is also worth noting that the concentration of intracellular ATP
within RBCS was determined to be Statistically similar whether they were obtained from
healthy controls or CF patients, as well as, after chemical treatment. Here unpublished
data also demonstrated that the incubation with either glybenclamide or niflumic acid did
not directly affect the potassium-sensitive ATP channels by showing that inhibition with
cromakalim did not affect the RBC deformability or the concentration of deformation-
induced ATP release.

It has been shown that ATP release from the RBCS of people with type 2

. 4
d1abetes,37’ 1

primary pulmonary hypertension,42 and cystic fibrosis36 is significantly less
in comparison to those RBCS obtained from healthy controls. Throughout the literature it
has been reported that there is a substantial decrease in G, expression in numerous animal

4347 including rats that have been administered streptozotocin,

models of diabetes,
resulting in B-cell destruction. A decreased expression of G, in the platelets and RBCS
from type 2 diabetics,37 decreasing vascular circulation and inadequate 02 delivery
throughout the circulation system, is thought to be the reason for hypertension
complications that are associated with the disease, resulting in poor wound healing that
often times requires limb amputations.40 It was reported that there is a decreased CAMP
generation and ATP release from RBCS as AlC levels increase in response to mastoparan

7 (activates 0,), providing a possible explanation for the increase incidence of vascular

complications as Al C levels increase in type 2 diabetes.

25

1.2.4 ENDOTHELIUM-DERIVED RELAXING FACTOR: NO

In 1998 Furchgott, Ignarro and Murad were jointly awarded the Nobel Prize in
Physiology or Medicine for demonstrating that NO is a Signaling molecule in the
cardiovascular system. Specifically, Furchgott and Zawadzki48 discovered that in order
to induce rabbit thoracic aorta relaxation through the perfusion of acetylcholine (ACh),
the endothelial lining within the aortic ring was required. This demonstrated that ACh
elicits a response of the endothelial lining that in turn, induces smooth muscle relaxation
and that without the mediation of the endothelium, ACh will induce a constriction of the
smooth muscle. The sandwich experiment, as illustrated in figure 1.3 was used to
transfer the biologically active substance to a biodetector. Specifically two rabbit aortic
strips, one devoid of the endothelial lining, were placed lumen-side together, the resulting
vasodilation in the strip without endothelial cells was due to the endothelium derived
relaxing factor (EDRF) produced in the strip with the intact lining. Several other
experiments were explained in a review by Moncada49 where the EDRF from the
endothelium of either an isolated rabbit aortas or cultured endothelium, was transferred to
rabbit aortic strips, vascular rings or canine coronary arteries, inducing vasodilation in the
ladder biological detector. It was then determined that NO was the endogenous
vasodilator that was responsible for the resulting relaxation upon the perfusion of various
agonists such as ACh and bradykinin.50

NO has been reported throughout the literature to not only elicit relaxation of
smooth muscle cells comprising the cardiovascular system, but also demonstrated to
inhibit platelet aggregation, adhesion to endothelium, as well as revert platelet activation

and disaggregate thrombosis formations.51 Thrombosis is the process by which the body

26

Smooth
Muscle Cells

2
C

 

"‘1 t I 5" 4 ' r I l "‘1'

h . a. . .4 > A l . A n .1 MI
.11.. . z '33- ~. .. r. ., v ..m x“ .
- ‘ _ _ r, ; . ‘.. ti.

Smooth Muscle
lindotlrelial Relaxation
Cells

Figure 1.3 Illustration of the sandwich experiment conducted by
Furchgott and Zawadzki where the endothelium-derived relaxing factor
(EDRF), later to be determined as nitric oxide (NO), from the
endothelial cells of one rabbit aortic strip were exposed to the smooth
muscle cells on the second strip, devoid of the endothelial lining. Here,
the exposed smooth muscle cells, acting as a biological detector,
underwent relaxation from the biologically active vasodilator
transferred.

27

recruits platelet aggregation in order to stop hemorrhaging from lesions in the vascular
system. Research in our group has focused on monitoring the NO that is produced by the
endothelium or platelets themselves and how it can prevent platelet activation and

aggregation.52'53

1.2.5 ATP-MEDIATED NO PRODUCTION

Once ATP is released from the RBC, it can bind to the P2y purinergic receptor
that is found on the membrane of the endothelial cells that line the luminal surface of
circulatory vessels. Subsequently, as illustrated in figure 1.4, NO is produced when
vascular endothelium nitric oxide synthase (eNOS) converts L-arginine to L-Citrulline
upon activation of the Ca2+ dependent calmodulin protein, which is bound to the eNOS
enzyme.49‘50‘54 NO can diffirse back out into the circulation where it can affect other cell
types of the vascular system, or it can also diffuse through the underlying smooth muscle
cells where it binds to soluble guanylate cyclase (sGC) and stimulates the production of
cyclic guanylate monophosphate (cGMP).55 This CGMP has been shown to reduce the
concentration of intracellular Ca2+, preventing crossover of myosin and actin, ultimately
inducing smooth muscle relaxation.54-‘56

Smooth muscle contraction and relaxation is mediated via a Cay-dependent
mechanism that involves the phosphorylation of myosin light chain (MLC) by MLC
kinase.57 During contraction, an intracellular increase of Ca2+, binding with the acidic

protein calmodulin, activates that MLC kinase. Upon phosphorylation, the cycling of the

myosin cross-bridges with actin is initiated with the release of energy from ATPase

28

 

£03268? E :32 98 +~m0 53—089: 0262 53 533 $208 ucmngmonmocofi 22.39% 0:93 we accosted
05 85233 28 603 omfioxo Beingsw 0328 8 $55 .m=8 28:8 509% 2: 8 gate 55 :8 533
oz can oamzshmoé 8 85250 mm oEEwSJ .Emo-+~mov 51608.8 “coecoaocéamofio 3 $96 085:? 02x0
FEE 3205205 we c2358. com: Scum? 830m? 05 mo waE: 33585 2: co c882 8&8“: in 2: 8 weaves
”2?. 88m mes—3.2 57C 0238 2.53 we nouosuoa 2: was Emmamsoofi messages, noun—zfismihra. .2 char—

 

 

main—cm
.aucsooafluaam
.33anqu
102.532

. r
a

29

activity resulting in the contraction of the smooth muscles that line the walls of the
resistance vessels.

Amperometric determination of NO produced from ATP stimulated bovine
pulmonary artery endothelial cells (bPAECs) that were cultured within a section of fused
silica microbore tubing demonstrated that NO production was dependent on the
concentration of ATP pumped through the tubing.58 The cultured endothelium is a close
mimic of the lining of the lumen found in vivo and it was also demonstrated that shear
alone did not cause synthesis and release of NO into the effluent. Fluorescence-based
determination of intracellular NO production has been demonstrated by our group both in

- 5
non-flow4| and flow-based experiments. 9’60

In the non-flow system, bPAECs were
cultured in tissue culture flasks and the fluorescence intensity of intracellular NO
production was monitored as increasing concentrations of ATP (IO-100 uM) were
introduced to the flask. The observed intensity was Shown to be statistically different
from the fluorescence intensity resulting from buffer alone and basal levels of
intracellular NO production.

The resistance vessels found in vivo were better mimicked in the flow-based
system where microfluidic devices were used to determine the concentration of NO
production as various agonists, bradykinin and ATP, and L-nitroarginine methyl ester (L-
NAME) were pumped through the Channel and over the endothelium.59 L-NAME
inhibits the production of NO from the endothelial cell lining the vascular system within
rabbit lungs, such that, there was an inhibited vasodilation when blood is perfused
through this vasculature“ and has also been shown clinically to result in increased

54

hypertension. In order to demonstrate the ability of RBC-mediated ATP to induce

30

endothelium NO in vitro, a microfluidic device that is discussed in detail here will be
presented.60 Collectively, the resulting effects of diabetes are thought to decrease blood
flow in the microcirculation, inducing an increase is tissue hypoxia and possibly the

mechanism for increased incidence of hypertension and cardiovascular disease.

1.2.6 GLUCOSE UPTAKE AND METABOLISM

Cellular metabolism, while essential for catabolic and anabolic reactions, also
plays an important role in diabetic complications due to its involvement with blood
glucose levels. While the RBC is absent of intracellular organelles and unable to perform
protein or lipid synthesis, nor oxidative phosphorylation, there is still a requirement for
energy production in order to maintain ATP dependent membrane pumps and osmotic
stability. It was also suggested in the literature that the oxygenation state of Hb affects
specifically RBC glucose metabolism.”‘62 It is thought that the oxygenation state of Hb
directly affects the glucose metabolism through glycolysis or the pentose phosphate
pathway (PPP) by affecting the interaction of band 3 proteins with either glycolytic
enzymes (redirecting to PPP shunt) or deoxygenated Hb (metabolism through glycolysis)
as illustrated in figure 1.5 from Messana et al. During periods of high oxidative stress
with high levels of R-state oxygenated Hb, the RBC membrane Shifts glucose metabolism
through the PPP, increasing antioxidant ability of the cell via nicotinamide adenine
dinucleotide phosphate (NADPH) and glutathione (GSH) production.

The two main processes for glucose metabolism in the RBC are glycolysis, and to
a lesser extent, the PPP. It is through the latter that glucose is shunted, when the need for

antioxidant production is required to protect the cell from oxidative stress, for which the

31

HOS
glucose 33p

 

 

ATP

 

lactate
. ATP

 

J

l
lactate

L05
glucose 33p

 

 

 

 

 

,.--- GBP
NADPH +1,
4 t
\~_..II.
\
j; 2,3-DPG
C02 " ‘~. _____ ,I’ ATP
lactate
t
I
l
lactate

Figure 1.5 Illustration taken from Messana et al illustrating the integral role
hemoglobin (Hb) plays in glucose metabolism to form either antioxidants via the
pentose phosphate pathway in the case of high oxygen stress (HOS) or the production of
ATP through glycolysis in the case of low oxygen stress (LOS).

RBC is largely susceptible to. Oxidative stress can compromise the vitality of the cell
and inhibit the proper function of a number of enzymes and membrane proteins. In this
instance, glucose metabolism is redirected to the PPP where glucose—6-phosphate
dehydrogenase (G6PD) converts glucose-6-phosphate to 6-phosphogluconolactone
(6GP), producing NADPH, which is a major proton donor for a number of other
enzymatic processes including those of GSH as well as catalase, both of which are
involved in peroxide detoxification. It has been reported that the normal amount of
glucose metabolism through the PPP under normal conditions only operates at 1/60th of
its potential,21 indicating the great reductive potential this system is capable of under high
incidence of oxidative stress.

RBCS of diabetic patients have been shown to release a lower concentration of
ATP under mechanical deformation when compared to RBCS of non-diabetic patients.4|
Vasodilation is dependent upon the amount of ATP released from RBCS within the
circulation, which has been Shown to be directly affected by the antioxidant system. It is
speculated that the reduced release of ATP from RBCS of diabetic patients is linked to the
reduced deformability of the cell membrane due to an increase in oxidative stress under
hyperglycemic conditions.

Diabetes and the associated complications are subject for the research presented
here, as it pertains to the ability of the RBC to mediate vasodilation of the vascular
system. By employing a flow-based microfluidic device that incorporates microtitre plate
technology, the relationship between glucose metabolism within the RBC and the

resulting affect on cellular function is examined. Specifically, the antioxidant status of

33

the cell is determined by quantifying various intracellular glucose metabolites while

simultaneously determining the affect on RBC-derived ATP release.

34

10.

11.

REFERENCES

Spence, D.M. Bioanalytical challenges for analytical chemists. Analyst 129, 102-
104 (2004).

Executive Summary: Standards of Medical Care in Diabetes - 2009. Diabetes
Care 32, S6-SlZ (2009).

Nathan, D.M., Singer, D.E., Hurxthal, K. & Goodson, J .D. The clinical
information value of the glycosylated hemoglobin assay. N Engl J Med 310, 341-
346 (1984).

CDC Centers for Disease Control and Prevention. National Diabetes Fact Sheet:
General information and National Estimates on Diabetes in the United States,
2007. US Department of Health and Human Services (2008).

Anon Diagnosis and classification of diabetes mellitus. Diabetes Care 31, S55-
S60(2008)

Castano, L. & Eisenbarth, G.S. Type-1 Diabetes: A Chronic Autoimmune Disease
of Human, Mouse, and Rat. Annu Rev Immunol 8, 647 (1990).

Roep, B.O. Islet Autoreactive CD8 T-cellS in Type 1 Diabetes. Diabetes 57,
1156-1156 (2008).

Eisenbarth, G.S. Molecular aspects of the etiology of type I diabetes mellitus. J
Diabetes Complications 7, 142-150 (1993).

Filippi, C.M. & von Herrath, M.G. Viral trigger for type 1 diabetes: pros and
cons. Diabetes 57, 2863-2871 (2008).

Cudworth, A.G. Type I diabetes mellitus. Diabetologia 14, 281-291 (1978).

von Herrath, M. Can we learn from viruses how to prevent type 1 diabetes? The
role of viral infections in the pathogenesis of type 1 diabetes and the development
of novel combination therapies. Diabetes 58, 2-11 (2008).

35

12.

l3.

14.

15.

l6.

17.

18.

19.

20.

21.

22.

Nathan, D.M. Initial Management of Glycemia in Type 2 Diabetes Mellitus. N
Engl J Med 347, 1342-1349 (2002).

Bonner-Weir, S. & O'Brien, T.D. Islets in type 2 diabetes: in honor of Dr. Robert
C. Turner. Diabetes 57, 2899-2904 (2008).

Hod, M. & Yogev, Y. Goals of metabolic management of gestational diabetes: is
it all about the sugar? Diabetes Care 30, Sl80-S187 (2007).

Kitzmiller, J.L., Dang-Kilduff, L. & Taslimi, M.M. Gestational diabetes after
delivery: short-term management and long-term risks. Diabetes Care 30, S225-
S235(2007)

Fong, D.S. et al. Diabetic retinopathy. Diabetes Care 26, 226-229 (2003).

EDIC Intensive Diabetes Treatment and Cardiovascular Disease in Patients with
Type 1 Diabetes. N Engl J Med 353, 2643-2653 (2005).

Carpenter, M.W. Gestational diabetes, pregnancy hypertension, and late vascular
disease. Diabetes Care 30, 8246-5250 (2007).

Stephens, J .W., Khanolkar, M.P. & Bain, SC. The biological relevance and
measurement of plasma markers of oxidative stress in diabetes and cardiovascular
disease. Atherosclerosis 202, 321-329 (2009).

DCCT The effect of intensive treatment of diabetes on the development and
progression of long-term complications in insulin-dependent diabetes mellitus.
The Diabetes Control and Complications Trial Research Group. N Engl J Med
329, 977-986 (1993).

Bossi, D. & Giardina, B. Red cell physiology. Mal Aspects Med 17, 117-128
(1996).

Fischer, T.M., Haest, C.W.M., Stoehr, M., Kamp, D. & Deuticke, B. Selective
alteration of erythrocyte deformability by sulfhydryl reagents. Evidence for an
involvement of spectrin in membrane shear elasticity. Biochim Biophys Acta 510,
270-282 (1978).

36

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

MacRury, S.M., Small, M., Anderson, J., MacCuish, A.C. & Lowe, G.D.
Evaluation of red cell deformability by a filtration method in type 1 and type 2

diabetes mellitus with and without vascular complications. Diabetes Res 13, 61-
65 (1990).

Delaunay, J. The molecular basis of hereditary red cell membrane disorders.
Blood Rev 21, 1-20 (2007).

Jain, S.K. Hyperglycemia can cause membrane lipid peroxidation and osmotic
fragility in human red blood cells. J Biol Chem 264, 21340-21345 (1989).

Schwartz, R.S., Madsen, J.W., Rybicki, A.C. & Nagel, R.L. Oxidation of spectrin
and deformability defects in diabetic erythrocytes. Diabetes 40, 701-708 (1991).

Straface, E. et al. Structural Changes of the Erythrocyte as a Marker of Non-

Insulin-Dependent Diabetes: Protective Effects of N-Acetylcysteine. Biochem.
Biophys. Res. Commun. 290, 1393-1398 (2002).

Gaskin, R.S., Estwick, D. & Peddi, R. G6PD deficiency: its role in the high
prevalence of hypertension and diabetes mellitus. Ethn Dis 11, 749-754 (2001).

Fischer, D.J., Torrence, N.J. & Spence, D.M. Development of a biomimetic
circulatory system using chips fabricated from PDMS. Abstracts of Papers, 225th
ACS National Meeting, New Orleans, LA, United States, March 23-27, 2003,
BIOT-099 (2003).

Ellsworth, M.L., Forrester, T., Ellis, C.G. & Dietrich, H.H. The erythrocyte as a
regulator of vascular tone. Am J Physiol Heart Circ Physiol 269, H2155-216l
(1995).

McCullough, W.T., Collins, D.M. & Ellsworth, M.L. Arteriolar responses to
extracellular ATP in striated muscle. Am J Physiol Heart Circ Physiol 272,
H1886-H1891 (1997).

Collins, D.M., McCullough, W.T. & Ellsworth, M.L. Conducted vascular
responses: communication across the capillary bed. Microvasc Res 56, 43-53
(1998)

37

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

Dietrich, H.H., Ellsworth, M.L., Sprague, R.S. & Dacey, R.G., Jr. Red blood cell
regulation of microvascular tone through adenosine triphosphate. Am J Physiol
Heart Circ Physiol 278, H1294-H1298 (2000).

Sprague, R.S., Ellsworth, M.L., Stephenson, A.H. & Lonigro, A.J. ATP: the red
blood cell link to NO and local control of the pulmonary circulation. Am J Physio]
Heart C irc Physiol 271, H271 7-H2722 (1996).

Sprung, R., Sprague, R. & Spence, D. Determination of ATP release from
erythrocytes using microbore tubing as a model of resistance vessels in vivo. A nal
Chem 74, 2274-2278 (2002).

Sprague, R.S., Ellsworth, M.L., Stephenson, A.H., Kleinhenz, M.E. & Lonigro,
A.J. Deformation-induced ATP release from red blood cells requires CFTR
activity. Am J Physiol Heart C irc Physio] 275, Hl726-l732 (1998).

Sprague, R.S., Stephenson, A.H., Bowles, E.A., Stumpf, M.S. & Lonigro, A.J.
Reduced expression of Gi in erythrocytes of humans with type 2 diabetes is

associated with impairment of both CAMP generation and ATP release. Diabetes
55, 3588-3593 (2006).

Olearczyk, J.J., Stephenson, A.H., Lonigro, A.J. & Sprague, R.S. Receptor-
mediated activation of the heterotrimeric G-protein Gs results in ATP release
from erythrocytes. Med Sci Monit 7, 669-674 (2001).

Olearczyk, J .J., Stephenson, A.H., Lonigro, A.J. & Sprague, R.S. Heterotrimeric
G protein Gi is involved in a signal transduction pathway for ATP release from
erythrocytes. Am J Physiol Heart C irc Physiol 286, H940-945 (2004).

Sprague, R.S., Stephenson, A.H. & Ellsworth, M.L. Red not dead: signaling in
and from erythrocytes. Trends Endocrinol Metab 18, 350-355 (2007).

Carroll, J. et al. An altered oxidant defense system in red blood cells affects their
ability to release nitric oxide-stimulating ATP. Mo] Biosyst 2, 305-311 (2006).

Sprague, R.S., Stephenson, A.H., Ellsworth, M.L., Keller, C. & Lonigro, A.J.
Impaired release of ATP from red blood cells of humans with primary pulmonary
hypertension. Exp Biol Med 226, 434-439 (2001).

38

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

Hashim, S., Liu, Y.Y., Wang, R. & Anand-Srivastava, M.B. Streptozotocin-
induced diabetes impairs G-protein linked signal transduction in vascular smooth
muscle. Mol Cell Biochem 240, 57-65 (2002).

Matsuda, N. et al. Differential gene transcriptional regulation of Gi isoforms and
Gs protein expression in diabetic rat hearts. Naunyn Schmiedebergs Arch
Pharmacol 361, 53-60 (2000).

Hashim, S., Li, Y., Nagakura, A., Takeo, S. & Anand-Srivastava, M.B.
Modulation of G-protein expression and adenylyl cyclase signaling by high
glucose in vascular smooth muscle. Cardiovasc Res 63, 709-718 (2004).

Moxham, C.M. & Malbon, C.C. Insulin action impaired by deficiency of the G-
protein subunit Gi[alpha]2. Nature 379, 840-844 (1996).

Gawler, D., Milligan, G., Spiegel, A.M., Unson, C.G. & Houslay, M.D. Abolition
of the expression of inhibitory guanine nucleotide regulatory protein Gi activity in
diabetes. Nature 327, 229-232 (1987).

Furchgott, R.F. & Zawadzki, J.V. The obligatory role of endothelial cells in the
relaxation of arterial smooth muscle by acetylcholine. Nature 288, 373-376
(1980)

Moncada, S., Palmer, R.M.J. & Higgs, E.A. Nitric oxide: physiology,
pathophysiology, and pharmacology. Pharmacol Rev 43, 109-142 (1991).

Moncada, S., Palmer, R. & Higgs, E. The discovery of nitric oxide as the
endogenous nitrovasodilator. Hypertension 12, 365-372 (1988).

Palmer, R.M.J., Ashton, D.S. & Moncada, S. Vascular endothelial cells synthesize
nitric oxide from L-arginine. Nature 333, 664-666 (1988).

Carroll, J.S., Ku, C.-J., Karunarathne, W. & Spence, D.M. Red Blood Cell
Stimulation of Platelet Nitric Oxide Production Indicated by Quantitative
Monitoring of the Communication between Cells in the Bloodstream. Anal Chem
79, 5133-5138 (2007).

39

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

Ku, C.-J., Karunarathne, W., Kenyon, 8., Root, P. & Spence, D. Fluorescence
Determination of Nitric Oxide Production in Stimulated and Activated Platelets.
Anal Chem 79, 2421-2426 (2007).

Tuteja, N., Chandra, M., Tuteja, R. & Misra, M.K. Nitric oxide as a unique
bioactive signaling messenger in physiology and pathophysiology. J Biomed
Biotechno], 227-237 (2004).

Murad, F. Discovery of some of the biological effects of nitric oxide and its role
in cell Signaling. Biosci Rep 24, 452-474 (2004).

Sobey, C.G. Potassium Channel function in vascular disease. Arterioscler T hromb
Vasc Biol 21, 28-38 (2001).

Webb, R.C. Smooth Muscle Contraction and Relaxation. Advan. Physio]. Edu. 27,
201-206 (2003).

Kotsis, D.H. & Spence, D.M. Detection of ATP-induced nitric oxide in a
biomimetic circulatory vessel containing an immobilized endothelium. Anal
Chem 75, 145-151 (2003). ‘

Oblak, T.D.A., Root, P. & Spence, D.M. Fluorescence Monitoring of ATP-
Stimulated, Endothelium-Derived Nitric Oxide Production in Channels of a
Poly(dimethylsiloxane)-Based Microfluidic Device. Anal Chem 78, 3193-3197
(2006)

Genes, L.I., Tolan, N.V., Hulvey, M.K., Martin, R.S. & Spence, D.M. Addressing
a vascular endothelium array with blood components using underlying
microfluidic Channels. Lab Chip 7, 1256-1259 (2007).

Sprague, R.S. et al. Effect of L-NAME on pressure-flow relationships in isolated
rabbit lungs: role of red blood cells. Am J Physiol Heart Circ Physiol 269, H1941-
H1948 (1995).

Messana, I. et al. Human erythrocyte metabolism is modulated by the OZ-linked
transition of hemoglobin. FEBS Letters 390, 25-28 (1996).

40

CHAPTER 2: MERGING FLOW INTO MICROTITRE TECHNOLOGY

2.] MICROFLUIDIC DEVICES: GENERAL HISTORY AND APPLICATIONS

The main advantage associated with the miniaturization of analytical processes by
employing microscale total analysis systems (uTAS)l is to be able to perform fast,
accurate and inexpensive analyses in a high throughput manner. Ideally these systems
integrate sample and reagent introduction, the ability to move these solutions on-Chip,
often times requiring mixing of liquids or gases, incorporating purification processes
when Chemical synthesis is employed, and finally the ability to detect various analytes on
a Single platform. These systems are highly dependent on the developments in the
fabrication of microelectronics and microelectromecharrical systems (MEMS), providing
a means to easily integrate fluid flow into a similar Silicon or glass device. 24
Microfluidics, as put best by Hansen and Quake,5 integrates the high levels of wet
laboratory functionality on a single chip by manipulating solutions in volumes ranging
from nano- to femtoliters. The employment of uTAS is highly dependent on microfluidic
technology, leading to advantages over standard systems that include the reduction of
consumed sample and reagents, faster analysis times, increased sensitivity, disposability
and portability.2’3

The first demonstration of microfluidic technology was the development of a gas
chromatograph on a Silicon wafer in 1979 at Stanford University by Terry et al.6
Miniaturized analytical systems have been primarily employed Since then using

microfluidic technology in the areas of capillary electrophoresis (CE), gas

chromatography (GC), liquid Chromatography (LC), and mass spectrometry (MS) to

41

name a few of the earliest applications. Along with the MEMS platforms that were
typically fabricated in silicon and glass, there were additional motivations for the
development of microfluidics and the resulting uTAS that are employed today as
reviewed by Whitesides.7 First, molecular analysis was integrated into a miniaturized
system to employ GC, high-perforrnance liquid chromatography (HPLC) and CE,
allowing for an increased sensitivity and resolution that was not obtained with standard
bench top instrumentation. Secondly, there was motivation for the development of
portable uTAS in order to combat chemical and biological weapons in the field, where
they posed military and terrorist threats. Finally, the requirement for high throughput
analysis, as well as increased sensitivity and resolution for DNA sequencing during the
19803 for genetic research, was enhanced with the employment of microfluidic systems.
Microfluidic technologies were developed in the four research labs of Manz,
Harrison, Ramsey and Mathies. Manz has demonstrated the ability to perform free-flow

1’2'8’9 Harrison’s group has

electrophoresis and synchronized cyclic CE devices.
developed the ability to integrate immunoassays and chemical reactions in organic
solvents on-chip.1°'15 Ramsey and Jacobson are best known for incorporating various
elements for on-chip sample pre- and post-treatrnents for CE, PCR, and DNA
sequencing.16 Additional work includes the ability to study enzyme kinetics, perform
electrospray for subsequent mass spectrometry (MS), and carry out miniaturized open
channel electrochromatography and micellar electrokinetic capillary Chromatography, all
in microfluidic devices.”22 Finally, Mathies developed the ability to perform DNA
sequencing and restriction assays for genotyping, using multichannel analysis with the

incorporation of multicolor detection.”27

42

With advanced fabrication techniques and component integration, there are
currently an expansive number of applications for microfluidic technology. Some of the
most prominent include the integration in biochemical and chemical processes, analytical
systems, biomedical devices and systems for fundamental research.”29 Biochemical
applications have enabled on-chip cell growth, counting, and sorting, in order to detect
various biological species and have also been employed in genomic studies. Additional
applications include the time-resolved studies of protein folding and highly efficient
protein crystal growth.5 This is performed in a high throughput approach by monitoring
the effect of various conditions such as pH, ionic strength, composition, and
concentration of solvents on crystalline growth. By performing protein crystallization
within microfluidics, subsequent damage through handling is reduced and electrokinetic
manipulation, vesicle encapsulation and the incorporation of mechanical valves have
been demonstrated.30

Systems biology, the integration of complex biological systems to understand
cellular mechanisms, is strongly dependent on the ability to study both single cells and
bulk cellular samples using microfluidic platforms}1 The nature of microfluidics
provides a means to mimic physiological conditions and apply Specific experimental
conditions for studying intra- and intercellular molecular interactions involved in cellular

events.

2.1.1 MATERIAL ADVANTAGES AND DISADVANTAGES
Microfluidic devices have been fabricated in many substrates such as silicon,

glass, and polymers like poly(methyl methacrylate) (PMMA) and poly(dimethylsiloxane)

43

(PDMS).3’28’32’33 Silicon micromachining is typically utilized in mass production
applications, especially for MEMS devices where the resulting features range from 1 pm
into the millimeter range. Silicon is opaque in the ultraviolet and visible region
(UV/V IS), rendering it incompatible with many optical detection techniques, and
relatively expensive. Glass, while transparent in the UV/V IS region, is an amorphous
material, making it difficult to etch vertical side walls.28 Together, the major
disadvantages associated with using silicon and glass for microfluidic devices are that the
fabrication processes are slow and complicated, clean room facilities are required for
each unit fabricated and the bonding techniques are time consuming. Also, the high
temperatures that are required for bonding are not well tolerated by various surface
Chemistry modifications that may be desired.4 The material properties of each, their
fragility, cost, and incompatibility with various detection schemes, increased the need for
new polymeric substrates.

Structures composed of silicon or glass are rigid, making moving parts such as
valves difficult to integrate. By integrating valves, pumps, mixers, filters and necessary
interconnects, uTAS systems are defined by the ability to convert large analytical devices
to small miniaturized platforms that are often disposable. uTAS systems incorporate
sample acquisition, pre- and post-treatrnent, separation, and detection all on a single
device without external components for the analysis of liquid or gas samples.3 Polymers
are less expensive, more rugged, and their fabrication is faster and cheaper by employing
the processes of casting, embossing and injection molding.34

Polymers have many advantages over both Silicon and glass, a few of which are

their case of use, numerous fabrication and sealing techniques not requiring a Clean room,

44

rapid fabrication in comparison, and economically advantageous both in the cost of
material and reusability. Non-traditional materials such as hydrogels, plastics and
elastomers allow for fast and economical, passive and active microfluidic devices.5 As
outlined in table 2.1, Becker et a].34 compare various attributes of silicon, glass, technical
thermoplastics, thermoset polymers and elastomers detailing how they each perform in
various fabrication processes for microfluidic devices.

Technical thermoplastics, such as poly(methyl methacrylate) (PMMA),
polycarbonate (PC), and polyetheretherketone (PEEK), become malleable at elevated
temperatures and can be remolded by reheating. Although these materials were among
the first employed for the fabrication of microfluidic devices, much attention is now
being directed to cycloolefin polymers and copolymers for their optical properties.
Thermoset polymers, on the other hand unlike technical thermoplastics, undergo a
Chemical reaction when cured at elevated temperatures. They represent duroplastic
materials, such as the photoresists used in photolithography and polyimide, which is
frequently used in the fabrication of microelectronics. These materials are often referred
to as resins that cure at elevated temperatures forming a rigid, inflexible replica mold.
Elastomers, as classically defined, exhibit at least 200% elastic elongation due to their
composition of long entangled molecular Chains. These materials, the most common of
which is PDMS, are cured at relatively low temperatures and make up the bulk of
materials currently used for the fabrication of microfluidic devices.

There are numerous advantages to fabricating microfluidic devices in PDMS
including the properties associated with optical transparency (from 240 to 1100 nm),

permeability to gases, non-toxic properties to mammalian cells, and the ability to

45

000300 0423405045 .40 505004.55 5 cm: .505 8 050500 5 mm 50503040 000 505305 0000585
.mosmflqogofi 4005408 awe—w 500:7. .40 005.8005 05 52500 40 B Hoxoom 50¢ 005000 030.4. fim 030,—.

 

 

304 55002 5502:4504 52-55002 55002 84.5 $4,902

545 beam 54.5 .5002 £34 0 2 .355 050.0 48300

>64 "@55532 554005.304 515-5232 :wf 3:5me

. . . . . . 5028 0554.5

05 .05 if .05 855 5523532

:05 55 :05 .05 .05 0:35 2%

304 55002 5535.304 :53 swam bins; 2550 050.4.

304 be.) swam 554005-304 :wi :53 35%: 235582

:5805Smam 49am 5am 2:055 52.50 505250535

awam 55002 swam 550042 5cm .3 050034

0352. mm .0” $82 0352. am .0” 582 am .0” .52 an .0335 on .35: 4.2.858”, 03380
5032?. be:

r 5530 .m:_::0.40:55 5250:5823 “45:30 1
5:55 i . . l . momm000.am:_._50_5w
$205805: .mczmau 5550050 40: $5280 40>.» be 05 40>, .

@5305 50500.54

exam :5502 Amen 55005- chm—Wu 55.005- chasm 2020045505042

$050.55 505300 4805.55. 05mm .Um .SZEA $204 man—O 500:5

3:530:55 30550.4.

 

46

fabricate structures on the micro scale at relatively low temperatures.28'29 These
characteristics allow for UVN IS absorption as well as fluorescence detection without
interference from the substrate. The ability to incorporate multiple cellular components
allows for the fabrication of an in vitro platform that can closely replicate the in viva

system.

2.1.2 FABRICATION METHODS IN PDMS

The fabrication method chosen to prepare a microfluidic device depends upon the
material used, the complexity of the desired design, the compatibility between the
solutions intended for use within the device, and the application for flexible academic or
commercial mass production. The fabrication processes for PDMS-based devices are
discussed here, focusing primarily on the techniques of photolithography and replica
molding, or casting of PDMS.

Photolithography is a technique used to obtain microscale structures from
thermoset polymer resins, which cross-link under exposure to light, on various substrates.
Rapid prototyping, as first described by the Whitesides group,3 is the fabrication of a
master wafer using photolithography for subsequent replica molding in PDMS. A high-
resolution commercial image setter is used to create a negative transparency of channels
designed using a computer software program. This transparency, with approximately 20
um resolution, can be used in contact photolithography to create a positive relief of
photoresist on a silicon wafer, serving as a master for subsequent replica molding with
polymeric materials. While the channel dimensions possible are much smaller (500 nm)

using a chrome mask in contact photolithography, the transparencies are much more

47

economical, not only in cost, but also in the time required to obtain the mask, lending
flexibility to the channel designs necessary for method development.28

The term sofi lithography refers to any non-lithographic technique, such as PDMS
casting, where the resolution of the resulting replica mold is dependent upon the
resolution of the master wafer from which it is molded, not by the optical diffraction limit
of the incident light. When fabricating microfluidic devices using PDMS and replica
molding, there are many advantages associated with various aspects of design and sealing
that are specific to the experimental methodology employed. Flexibility in the design of
the microfluidic charmels comes in the ability to design any arrangement of various
dimensions through the hi gh-resolution printing of transparency masks.

In order to direct flow throughout the device, the channels have to be sealed either
reversibly or irreversibly to another substrate. Reversible scaling is possible due to
hydrophobic conformal contact, displaying between a replica molded layer of PDMS and
a second layer of glass, silicon, or another PDMS slab.3 Due to the flexibility of the
material, PDMS is able to seal to another layer despite minor imperfections. These
sealing properties of PDMS make it a fast process and it occurs at room temperature.
There is no residue left behind on the substrate, allowing for reusability of the device
through repeated resealing events that are water tight, but don’t withstand high pressures
(>5 psi).28 Due to the hydrophobic nature of the native PDMS, it is difficult to wet the
channels fully with aqueous solutions, resulting in bubbles and adsorption of other
hydrophobic moieties onto the surface of the device.

Irreversible sealing is possible with microfluidic devices fabricated in PDMS by

employing plasma sealing. Plasma sealing involves the oxidation of two layers of PDMS

48

and has been developed as a means to irreversibly enclose the channels with a second
layer of PDMS, oxidized polystyrene, glass, silicon or silicon oxide. Plasma, partially
charged ionized gas that is composed of an equal number of independent positive and
negative charges occupying the same space, results in the formation of silanol groups on
the surface of the PDMS.3 Once sealed, the device can withstand much higher pressures
(30-50 psi) than a reversible seal and cannot be separated. This technique, while fast and
strong, is a permanent, requiring precise alignment and an additional step in fabrication
of the device. The PDMS surfaces become negatively charged, which is much more
compatible with aqueous solutions, allowing for the integration of electroosmotic flow
(EOF) and chemical separations using neutral or basic solutions.

EOF within a microfluidic device can be generated by applying an electrical
potential across the length of the channel. Molecules can then be separated based on their

electrophoretic mobilities towards the cathode.35

Inducing flow in an ionic liquid will
allow for the integration of pumping, mixing, and injection techniques without the need
for any mechanical components. Jacobson et al. have reported on a technique for sample
and reagent parallel and serial mixing within an EOF-based microfluidic device, where
the voltage control is simplified.22 A single voltage source is employed and by
combining the termination reservoirs of various buffers, samples and analysis channels,
the complex external hardware such as programmable power supplies is not needed to
obtain desired voltage division. More recent work in J acobson’s lab has demonstrated

the ability to create spatial and temporal gradients with varying slopes and offsets that are

easily altered, changing gradient composition in both pressure and EOF systems.36

49

When employing EOF, the electrical double layer (EDL) becomes prominent
within microfluidic systems, as compared to conventional systems, due to the high
surface area to volume ratio in these miniaturized devices.35 This allows for biological
assays that are impracticable or impossible on the macroscale to benefit from the
increasing interaction between the sample and the surface of the channel wall or sensing
element. By employing EOF within the micrometer footprint of the channels, there is
elimination of the parabolic-flow profile that is observed with pumping. This causes the
sample to move throughout the channel in a plug, reducing the sample broadening, and
results in the increased separation resolution observed within microfluidic devices.

Multilayer soft lithography allows for the integration of valves and pumps by
employing a thin elastomer layer to separate planar channel structures. Specifically,
mechanical elastomeric valves allow for precise fluid control where complete sealing of
multiple layers allows for low actuation forces, small space requirements with negligible

dead volumes and flow rates of approximately 3 nL/secf"5

Implementation of these
structures increases the sophistication of assays that can be employed where the sample
solution within a specific portion of the device can be adjusted to sensitivity.

Various manipulations such as pneumatic valving4, sample injection techniques”,
and sample separation2 are all possible using such PDMS-based devices. Peristaltic and
membrane pumps are possible by incorporating multilayer microfluidic fabrication with
the use of elastomeric polymers.29 In both techniques, actuation of pump channels in an

adjacent layer can occur using a variety of materials such as water, air, the properties

associated with the expansion of hydrogel, or magnetic field forces.

50

When studying the unique physics of fluid flow and mass transport on the
microscale, when compared to those occurring within macroscale channels, many
advantages are realized. First, there is an absence of turbulent flow, meaning that when
solutions are pumped through microfluidic channels they adopt laminar flow.7 This is
due to the reduced Reynolds number (Re):

Re = puIJu (l)
where p = fluid density (g/cm3), u = characteristic velocity (cm/s), L = characteristic
channel dimension (cm), and u = viscosity (g/cm-sec) and is the ratio of the relative
inertial forces to solution viscosity, which is minimized for the flow of water in
microfluidic channels.5

In macroscopic fluids, Re is large, meaning that the inertia overpowers the effect
of the solution viscosity resulting in turbulent flow. In contrast, within microscale
systems, this number is reduced because here, the viscosity is the major contributing
component. With laminar flow, two samples introduced into a single channel will flow
adjacent to one another, only mixing by diffusion across the defined interface. Increased
mass transport to the channel walls is exhibited, resulting in a parabolic-flow profile
which gives advantageous fluid-flow properties in the case of separation science.
However, it is a major obstacle to overcome when considering the need to combine and
mix various solutions on-chip, which have been the motivation to integrate mixing

components.

2.1.3 PHOTOLITHOGRAPHY: OBTAINING A SILICON WAFER MASTER

51

In order to obtain master wafers for soft lithography techniques and fabricate
PDMS replica molds, photolithography was used as previously established,3’23’29 and
figure 2.1 highlights the key steps in the process. First, 4 mL (1 mL per inch substrate) of
degassed SU8-50 negative photoresist (MicroChem, Newton, MA) is spin coated onto a
4” silicon wafer (Silicon Inc., Boise, Idaho) using a dual-step program to obtain a
thickness of approximately 100 um, corresponding to final channel depth. The spread
cycle, ramped at 100 rpm/sec to 500 rpm and held for 10 sec is used to cover the entire
substrate surface while the spin cycle, 300 rpm/sec to 1000 rpm for 30 sec will provide a
uniform coating of the silicon wafer. In order to evaporate the solvent and densify the
film, the coated wafer is then placed on a digital hot plate (VWR, West Chester, PA) at
95 °C for 15 min.

The desired channel widths and patterns are first created by specifying their exact
dimensions using Freehand software, which is then printed on a negative transparency
mask with 5080 dpi resolution (Pageworks, Cambridge, MA). This transparency mask,
which is adhered to a piece of glass, is placed transparency-side down on top of the
photoresist-coated silicon wafer. As illustrated in figure 2.2, a flood source (Newport,
Irvine, CA) provides near ultraviolet light (with a wavelength range of 350 to 400 nm)
which passes through the transparent regions of the mask, comprising the channel pattern,
polymerizing the photoresist below. An energy dosage of 550 J see”1 is provided by
exposing the substrate for 31.19 sec, leading to an optimal exposure and preventing
uneven overexposure of the top of the photoresist layer, resulting in t-topping. A post
exposure bake for 5 min at 95 °C is performed to cross-link only the exposed portions of

the photoresist, which corresponds to the desired channel pattern. The unexposed

52

G .058

.w5205 002000 50340033 .48 500005 00.403 500:6 08500.4
@045 0 30250.4 8 40400000043 0>uaw0a om w-Dm 40 505904050 6 58

n

050050 >3 3

n

5050540 04005 50008050:

“won 05.. -05 3 n@5000 5% Q .40 3000 06 535002: 4005350 03354000 Eamcmofizoeosm fin 05w;

 

53

.mohsaoméommfi 35.0w 2.: Eco 3162 06533 coomzm 2: so chBEE
t2 362823 wowtoaboqfi. 05 2682 8 33 fl 8&293 £me mocoammawb 05 mo mcomton
EoSchab 2: 323 67,8893 05 mo dowwntoaboa wEQfiE 59x £38 58.8%:«5 oifimoa
05 swab: Em: >3 9 womomxo :05 mm £033 5‘33 coommm av 05 no wcgoo 562893 x25
E: o3 bofifiioamm E 3:58 mmoooa 9558-5“? 05 .203 .589: 833% 3&2 a £890 3
@839: 3on 292:8 05 @3865 8326on 5mm? Emma: How 8305 .«o 8:832: N.N can—urn

agnosuonm vacuum can ans—Z 5855 55:5 av 3.8
canton—boggy Q2959 33.—323 cute—From 5:35:255‘ :w=< 37.2325 230-555

 
 

Em: >:

54

photoresist is then developed away using propylene glycol monomethyl ether acetate as
the silicon wafer is submerged with agitation for approximately 3 min or until only minor
amounts of unpolymerized photoresist is remaining on the substrate. The wafer is then
removed from the developer bath and rinsed with acetone, then isopropyl alcohol, and
dried with a stream of compressed, clean, dry air. The master wafer obtained is then used

for replica molding employing previously established soft—lithographic techniques.28

2.1.4 SOFT LITHOGRAPHIC TECHNIQUES TO OBTAIN REPLICA MOLDS

The microfluidic array (uFA) fabricated here is comprised of two individual
layers of PDMS that are irreversibly sealed around a track-etched polycarbonate
membrane (TEPC) membrane allowing underlying microfluidic channels to deliver
solutions to a microtitre plate-like array of wells. A 12-channel master wafer, obtained
using the photolithography techniques described above, was used in replica molding to
obtain the bottom layer of the uFA. In this instance, a degassed mixture of 20:1 Sylgard
184 PDMS (DOW Corning, Midland, MI), where the ratio refers to the amount of
elastomeric base to curing agent, is poured onto the silicon wafer and partially cured for
20 min at 75 °C. Once removed fiom the silicon wafer, the inlets for each channel were
punched through this layer of PDMS using a 20 gauge luer stub adaptor (BD, Sparks,
MD) while waste ports were generated using a 1/8” hollow punch (Fastenal, Winona,
MN). A second master wafer, absent of any raised channel features, is used to obtain the
top layer of the “FA in which subsequent array-like wells are punched through. This
layer was produced by pouring a degassed mixture of 5:1 PDMS onto the blank master

and as before, partially baking for 20 min at 75 °C. The array of wells was obtained in

55

the top layer by using the 1/8” hollow punch to produce various wells that are then
aligned over the parallel channels once the membrane is placed in between the PDMS.
The uFA is completed by further baking these layers together for 30 min at 75 °C. A
displacement syringe pump is used to introduce samples from 500 uL gas-tight syringes
(Hamilton, Reno, NV) through Tygon tubing (Saint-Gobain PPL, Courbevoie, France)
that are fitted with 15 mm 23 gauge hypodermic steel tubing bent at 90 degrees (New
England Small Tube, Litchfield, NH). This steel tubing is placed in each of the inlets
created in fabrication in order to deliver solutions directly into the channels. Solutions
can be flowed through the underlying channels for an optimized time allowing for
diffusion through the polycarbonate membrane, where the time chosen is specific for the
type of experimental methodology desired. Fluorescence microscopy is employed to
monitor the fluorescence intensity observed in each well and the fluorescence images are
analyzed for pixel intensities, which are dependent upon the concentrations of analytes of
interest within the flowing solutions below. By including various fluorescence probes or
cultured cells within the wells, this pFA serves as a precursor for future integration of a
more realistic flow-based in vitro platform into the already available microtitre
technologies, including automation, which is currently employed in drug research

laboratories.

2.2 PDMS uFA
To better understand the metabolic processes of glucose that may lead to a
decreased antioxidant system, a microfluidic device is used to address various glucose

metabolism pathways of interest. Exploiting the benefits of sofi lithography techniques

56

and the rapid replica molding associated with PDMS, the multi-dimensional pFA detailed
in this research38 is fabricated in a relatively short period of time. Moreover, the
microscale dimensions of the channels of the uF A approximate resistance vessels in vivo,
providing a more realistic mimic of the in vivo environment, while including an important
semi-preparative sampling technique all on chip.

A microfluidic device, composed of two individual PDMS layers that are
irreversibly sealed around a TEPC (GE Osmonics, Minnetonka, MN) is obtained by
simply aligning the layers prior to full polymerization in the baking process.39 Similar

(132’3 7’40"” in comparison to traditional

methods of thermocuring have been describe
plasma sealing. The actual fabrication of this device depends upon the application, the
top layer is composed of a variable number of 1/8” wells that contain either fluorescent
reaction solutions or cultured bPAECs while the bottom layer includes individual
channels (approximately 200 um wide by 100 pm in depth) providing a means to deliver
sample solutions and blood components under the wells above. The TEPC membrane
pore inner diameters are selected to optimize the interaction of the analyte(s) flowing
through the underlying microfluidic channels for the individual experiment of interest.

The channel-containing PDMS layer is replica molded from a silicon wafer master that is

obtained using standard photolithographic techniques.

2.3 IMPLEMENTING AUTOMATION TECHNIQUES
The automation and various technologies already available for micro-titre plates
will allow for this uFA to handle simultaneous and variable sample volumes all while

maintaining the currently accepted dimensions of micro-titre plates. Techniques for drug

57

development utilizing micro-titre plate technology have the advantage of incorporating
commercial automation capabilities for high-throughput screening (HTS). Complex
biological sample preparation, introduction of various sample volumes simultaneously,
and integrating various instrumentation for sample separation, identification, and
quantification are various tasks that can be performed with automated instruments. Due
to the acceptance of a standard dimension for the micro-titre plate, numerous instruments
have been developed on this scale leading to various advancements in the automation and
detection methods for this technology. Houston and Banks42 reviewed the
transformations that were taking place in drug discovery in 1997 and discussed the
implementation of automation for various HTS techniques, from dry compound storage
and handling to solid and liquid combinatorial chemistry, for developing possible drug
candidates. As noted, the requirements for integrating HTS automation into the drug
research field requires homogeneous assays (devoid of separation) that use innovative
assay designs, focusing on cellular and biochemical in vitro assays. The lead rationale is
to construct potential drug targets, and hope that a candidate interacts with one of the
more than 100 targets studied. Increased competition and exponentially increasing
healthcare costs have led researches to now focus on changing the serendipity into
certainty and this is where the uFA may be beneficial.

An evolutionary process has been implemented in current drug discovery
techniques to increase assay density and miniaturize by converting automation techniques
from the standard 96-well plate to 384-well plates and beyond. One possible reason why
microfluidics hasn’t been incorporated into drug discovery is because of the lack of

application of these automation techniques due to varying dimensions. The advancement

58

in automation to include reliable small volume dispensing, effective mixing at such small
volumes, and increased detection schemes and sensitivity suggests that the uFA holds
great possibilities for high-throughput and automated clinical testing and research by
combining all of the advantages listed for each system in table 2.2.

The development of a device that includes multiple cell types, coupled with the
fluid flow forces that are found within the circulation, results in a tool that more closely
mimics cell-to-cell communication that may exist in the bloodstream. For instance,
overcoming hypertension through vasodilation (resulting from the communication
between multiple cell types) can now be investigated more closely using this in vitro
platform. Also, the current plate design limits the ability to incorporate both adherent and
non-adherent cells such as those found within the vascular system (endothelial cells,
white and red blood cells, platelets, etc). This limits the ability to realistically mimic and
monitor the toxicity of a drug candidate on a particular cell-type of interest while

simultaneously monitoring the effect on non-targeted cell-types.

2.4 INTEGRATING FLUID FLOW: CIRCULATION SYSTEM MIMIC

Although automation of micro-titre plate technology can reduce the time required
to perform redundant analyses, there are limitations of such systems when applied to drug
discovery. For example, current micro-titre plate technology results in a static system
that is unable to mimic the flow-based vascular system observed in viva. Many of the
advantages of microfluidic devices are associated with the ability to integrate a more lab-
on-a—chip approach onto a single device such as the ability to perform on-chip cell

culture, sample injection and separation, and incorporate gaseous interfaces and valving

59

 

Microfluiclic l\slicro-t1tre Plate

. . Microfluichc Ana‘s
Teclmologies Teclmologies '

 

'\

Flow-based System J
11L to 11L V'ohuues
| .ah-ou-a-( ‘hi p lutegrati on
Eliminate Matrix
Interferences
Multicellular Capabilities
Multiple [Detection Schemes
High 'l'lu'oughput. Aualysns

\\\\'\

Simultanemls Detection

\ \\'\

Current Clinical Diagnostic
Tool
Automation Systems i/

«xxssxxs'x

Table 2.2 Comparison of various aspects associated with microfluidic and micro-titre
plate technologies and the combination of these advantages associated with the
microfluidic array.

60

techniques. The ability to better mimic the in vivo circulation by incorporating the flow
of blood components, coupled with simultaneous detection and laboratory automation in
place for micro-titre plates, suggests that the microfluidic array presented here will allow
for improved mechanistic drug research studies.

By integrating the microfluidic channels and maintaining the dimensions of the
standard micro-titre plate as illustrated in figure 2.3, clinical diagnosis, disease
progression, drug research and efficacy studies could now include the very relevant
aspect of blood flow. Here, the potential of such a device is demonstrated with two
examples; an application to more individualized medicine, as well as the application of
this device for determining drug efficacy and mechanisms of action. This device may
improve the manner by which drugs are prescribed to patients by first providing a means
to study personalized drug efficacy on various cell types. It is anticipated that the uFA
may lead to higher success rates of prescribing the right drug for the right patient on the
first attempt. Specifically, in order to prescribe the most effective drug and at an
effective dosage, this device will allow clinicians to actually monitor the effect on various
cell-types even before the drug is administered to the patient. For example, this device
has used fluorescence microscopy to determine concentrations of multiple metabolites
present within the RBC and here, its application towards using this device for
personalized medicine is presented.

Separately, this device has demonstrated its improved in vitro measurement
capabilities involving iloprost, a pharmaceutical drug reported to improve blood flow,43
and how its mechanism of action is determined by incorporating multiple cell types onto

a single device. This device is employed to detect multiple components fiom various

6]

Waste Ports
Sample Inlets

r

(9’
G5
(9,
(129'
C9
@

’I I’ 4 , ; I I I; 1
I r , , r r
z I a I r z I I I

5:1

 

Polycarbonate
Membranes

Figure 2.3 Flow-based microfluidic 96-well plate fabricated using the standard
dimensions from two individual PDMS layers thermocured around a polycarbonate
membrane. Here, sample inlets and waste ports are introduced through the bottom
layer of PDMS, comprising the network of microfluidic channels, each addressing in

parallel a series of triplicate array-like wells that are contained in the top layer of
PDMS.

62

RBC samples in less than 20 minutes and will enable improved in vitro experimentation.
By integrating flow into a high throughput and automated platform, drug efficacy that is
dependent upon shear-induced deformation (such as pentoxyfilline, with brand name
Trental) can now be accurately analyzed using a controlled in vitro platform. Our group
previously determined44 that the ability of Trental to increase the ATP release fi'om a
RBC sample was dependent upon the RBC being subjected to flow-induced deformation.
In the absence of flow, it was established that Trental did not significantly increase the
concentration of ATP released from the RBCs. Therefore, only by monitoring this drug
in a flow system could the possible mechanism of vasodilation action (ATP release) be
determined. Furthermore, unpublished results in our group have shown that hydroxyurea,

“’46 also behaves in the same

a current treatment for patients with sickle cell disease,
manner, requiring mechanical deformation of the RBC to substantially increase the
release of ATP. As outlined in table 2.2,38 the combination of features found in
microfluidic technologies and micro-titre based measurements are utilized with the
microfluidic array that is described in this work.

Most, if not all, in vitro methods involving drug discovery do not incorporate the
flow of blood that is found in vivo. Recently, the role of microfluidics in drug discovery
has been discussed47 and focuses on each individual step involved, including compound
generation through chemical synthesis, target identification, as well as lead identification
and optimization, using HTS and cell sorting techniques, as illustrated in figure 2.4.
Here, the numerous advantages microfluidic technologies have over conventional liquid

handling techniques are outlined and the implementation of miniaturized robotic systems,

including the introduction of high-density micro-titre plates as well as nanoliter

63

 

 

 

 

T d
arget , Compoun Lead Identification Lead Optimization
Identification Generation
cell culture 2D & 3D microctometry, cell
Microfluidic . ’ . . sorting, small volume . .
srngle cell micromixers, . . . cell and tissue studies
Methods . . liquid handling and
studies rmcroreactors .
generation
organic drug discovery, protein drug efficacy,
. . metabolomics, synthesis, crystallization, molecular pharmacological
Apphcations . . . . . . .
proteorrncs combinatorial evolution, compound profiling, toxrcrty
chemistry library screening testing, chemotaxis

 

 

 

 

 

Figure 2.4 Illustration of possible integrations of microfluidics into drug discovery
processes as discussed by Dittrich et al. highlighting the individual applications in
generation of target investigational new drugs through chemical synthesis on-chip, as well
as lead identification and optimization utilizing HTS and cell sorting techniques.

64

 

dispensing systems. The pF A presented here is expected to play an important role in the

in vitro testing period of drug development prior to entering clinical trials.

2.5 ADMET: DRUG TESTING APPLICATION

Collectively, the advantages associated with microfluidic technologies and micro-
titre plate automation may enable advancements in biotechnology associated with
improved mimics of in vivo processes. For example, this device is an initial step toward
incorporating the flow-based systems of microfluidics with already available micro-titre
plate technology to monitor the properties of absorption, distribution, metabolism,
excretion, and toxicity (ADMET) on a more realistic in vitro platform.3"47’48 The Food
and Drug Administration’s Center for Drug Evaluation and Research has reported that it
requires anywhere fiom 0.8 to 1.7 billion dollars for one drug candidate to reach the

market"9

. Recently, it was estimated that the costs associated with bringing a new drug to
market were due to poorly designed clinical trials.50 In addition, recent reports of hidden
marketing campaigns have made the clinical trial phase of drug discovery inefficient,
resulting in trivial information that doesn’t address relevant concerns. It was reported51
that an estimate of 46% of investigational new drugs (INDs) fail in clinical and pre-
clinical development due to insufficient efficacy and an additional 40% of INDs fail due
to safety reasons. It has been speculated by the FDA52 that the recent pipeline problem,
resulting in a progressive slow down in innovative medical therapies, could be
circumvented by employing new knowledge-based biomedical research. Kola and

Landis53 describe that one of the most effective ways for drug development companies to

reduce attrition rates is to pursue more applicable models that provide strong evidence for

65

proof of mechanisms, targeting specifically, efficacy and toxicity properties. It was also
considered by Hood et. al in 2004, that the reason for the overall decrease in
pharmaceutical research and development and the decade-long decrease in INDs in late-
stage clinical trials was due to the inability for current drug research to fully incorporate
the complexity of biological systems. Integrating systems biology into drug discovery, or
approaching analytically the complexities involved in a biological system, the overall
understanding of predisposition to disease, its diagnosis, and prevention, would lead to
earlier diagnoses with overall disease comprehension for individualized drug therapy
ultimately resulting in preventative drugs.54 Many aspects of the drug discovery process,
including the mechanism of action of certain drugs, solubility of the drug candidate, and
toxicity studies could be examined using the described uF A, a multicellular system that
involves actual blood flow, prior to the drug candidate entering into the clinical trial
phase of development. This would allow researchers to discover any unforeseen
complications earlier in the drug development stage, thereby reducing the number of
drugs that reach clinical trials only to fail and add to the cost associated with testing

unpromising drug candidates.

2.6 DIABETES AND PRE-DIABETES SCREENING

In continuance with the complications linked with diabetes, we have recently
reported that the ability of RBCS obtained fiom people with type 2 diabetes to release
ATP is decreased in comparison to control subjects and that this decrease is associated
with oxidant stress in the RBC.”56 As a stimulus of NO production in the endothelium,

RBC-derived ATP has been implicated as a possible determinant in the control of blood

66

flow in vivo. Another report57 has also shown a decrease in ATP release from the RBCS
of people with diabetes that may be correlated with an increase in the glycation of
proteins involved in the ATP release pathway. Finally, it has also been shown that RBCS
in the absence of metal-activated C-peptide, also release less ATP than those RBCS in the
presence of the peptide.58 Summarily, the various ATP release pathways suggest that
multiple factors or metabolic processes may affect ATP release from the RBC and some
of these processes are described here in detail.

While studies are still ongoing in our labs and others to identify those factors
affecting such cellular phenomena as ATP release from RBCS, until recently59 there was
no device available to perform an assessment of metabolic status and cellular function,
simultaneously. Such a device, as illustrated in figure 2.5, has been beneficial because it
not only enables the detection of possible biomarkers for a particular disease (e.g., low
ATP release fi'om the RBCs of people with diabetes), but also provides information
towards a therapeutic intervention and the root cause of the decreased ATP as it pertains
to glucose metabolism in RBCS from diabetic patients. Unfortunately, there are
challenges associated with measuring multiple analytes and cell function simultaneously
because the measurement of the analytes typically requires cell lysis while cell function
would be better studied on the intact cell. The current approach to solving this problem
is to compartmentalize the required measurements to different areas on a device, thus
facilitating the measurements while at the same time enabling the potential for high
throughput analyses. By monitoring both intracellular components of the RBC while
simultaneously evaluating cell function, the uFA presented here has applications for

determining diabetes as well as the earlier stages, pre-diabetes. This device is described

67

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

/ Q
/ , O
’ O
.’ (J
’ O
,
.’ 20:1
Polycarbonate
Membrane

 

Array-Like Wells

Figure 2.5 Schematic drawing of the uFA used for the
determination of intracellular RBC metabolites and
separately, for determining ATP-stimulated intracellular NO
production within cultured bPAECs. Here, two layers of
PDMS, containing parallel microfluidic channels and array-
like wells, are sealed around a track-etched polycarbonate
(TEPC) membrane.

68

to not only determine the level of shear-induced ATP release from RBCS, as cell
function, but also, the ability to simultaneously determine intracellular components that

may indicate a metabolic disorder.

69

10.

REFERENCES

Manz, A., Graber, N. & Widmer, H.M. Miniaturized total chemical analysis
systems: A novel concept for chemical sensing. Sens Actuators B Chem 1, 244-
248 (1990).

Manz, A. et al. Planar chips technology for miniaturization and integration of
separation techniques into monitoring systems. Capillary electrophoresis on a
chip. J Chromatogr 593, 253-258 (1992).

Duffy, D.C., McDonald, J.C., Schueller, O.J.A. & Whitesides, G.M. Rapid
Prototyping of Microfluidic Systems in Poly(dimethylsiloxane). Anal Chem 70,
4974-4984 (1998).

Quake, S.R. & Scherer, A. From micro- to nanofabrication with soft materials.
Science 290, 1536-1540 (2000).

Hansen, C. & Quake, S.R. Microfluidics in structural biology: smaller, faster...
better. Curr Opin Struct Biol 13, 538-544 (2003).

Terry, S.C., Jerrnan, J .H. & Angell, J.B. A gas chromatographic air analyzer
fabricated on a silicon wafer. IEEE Trans Electron Devices 26, 1880-1886 (1979).

Whitesides, G.M. The origins and the fiiture of microfluidics. Nature 442, 368-
373 (2006).

Seiler, K., Harrison, D.J. & Manz, A. Planar glass chips for capillary
electrophoresis: repetitive sample injection, quantitation, and separation
efficiency. Anal Chem 65, 1481-1488 (1993).

Kopp, M.U., De Mello, A.J. & Manz, A. Chemical amplification: continuous-
flow PCR on a chip. Science 280, 1046-1048 (1998).

Fluri, K., Fitzpatrick, G., Chiem, N. & Harrison, D.J. Integrated Capillary
Electrophoresis Devices with an Efficient Postcolumn Reactor in Planar Quartz
and Glass Chips. Anal Chem 68, 4285-4290 (1996).

70

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

Salimi-Moosavi, H., Tang, T. & Harrison, D.J. Electroosmotic Pumping of
Organic Solvents and Reagents in Microfabricated Reactor Chips. J Am Chem Soc
119, 8716-8717 (1997).

Harrison, DJ. et al. Micromachining a miniaturized capillary electrophoresis-
based chemical analysis system on a chip. Science 261, 895-897 (1993).

Liang, Z. et al. Microfabrication of a Planar Absorbance and Fluorescence Cell
for Integrated Capillary Electrophoresis Devices. Anal Chem 68, 1040-1046
(1996)

Chiem, N. & Harrison, D.J. Microchip-based capillary electrophoresis for
immunoassays: analysis of monoclonal antibodies and theophylline. Anal Chem
69, 373-378 (1997). ‘

Li, P.C.H. & Harrison, D.J. Transport, Manipulation, and Reaction of Biological
Cells On-Chip Using Electrokinetic Effects. Anal Chem 69, 1564-1568 (1997).

Waters, L.C. et al. Multiple Sample PCR Amplification and Electrophoretic
Analysis on a Microchip. Anal Chem 70, 5172-5176 (1998).

Hadd, A.G., Raymond, D.E., Halliwell, J.W., Jacobson, S.C. & Ramsey, J.M.
Microchip device for performing enzyme assays. Anal Chem 69, 3407-3412
(1997)

Ennakov, S.V., Jacobson, S.C. & Ramsey, J.M. Computer simulations of
electrokinetic transport in microfabricated channel structures. Anal Chem 70,
4494-4504 (1998).

Kutter, J .P., Jacobson, S.C., Matsubara, N. & Ramsey, J .M. Solvent-Prograrnmed
Microchip Open-Channel Electrochromatography. Anal Chem 70, 3291-3297
(1998)

Jacobson, S.C., Culbertson, C.T., Daler, J.E. & Ramsey, J.M. Microchip
Structures for Submillisecond Electrophoresis. Anal Chem 70, 3476-3480 (1998).

Khandurina, J., Jacobson, S.C., Waters, L.C., Foote, R.S. & Ramsey, J.M.
Microfabricated Porous Membrane Structure for Sample Concentration and
Electrophoretic Analysis. Anal Chem 71, 1815-1819 (1999).

71

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

Jacobson, S.C., McKnight, T.E. & Ramsey, J.M. Microfluidic Devices for
Electrokinetically Driven Parallel and Serial Mixing. Anal Chem 71, 4455—4459
(1999)

Woolley, A.T. & Mathies, R.A. Ultra-high-speed DNA sequencing using
capillary electrophoresis chips. Anal Chem 67, 3676-3680 (1995).

Woolley, A.T., Lao, K., Glazer, A.N. & Mathies, R.A. Capillary electrophoresis
chips with integrated electrochemical detection. Anal Chem 70, 684-688 (1998).

Simpson, P.C. et al. High-throughput genetic analysis using microfabricated 96-
sarnple capillary array electrophoresis microplates. Proc Natl Acad Sci U S A 95,
2256-2261 (1998).

Liu, S., Shi, Y., Ja, W.W. & Mathies, R.A. Optimization of High-Speed DNA
Sequencing on Microfabricated Capillary Electrophoresis Channels. Anal Chem
71, 566-573 (1999).

Paegel, B.M., Grover, W.H., Skelley, A.M., Mathies, R.A. & Joyce, G.F.
Microfluidic Serial Dilution Circuit. Anal Chem 78, 7522-7527 (2006).

McDonald, J .C. et al. Fabrication of microfluidic systems in
poly(dimethylsiloxane). Electrophoresis 21, 27-40 (2000).

McDonald, J .C. & Whitesides, G.M. Poly(dimethylsiloxane) as a Material for
Fabricating Microfluidic Devices. Acc Chem Res 35, 491-499 (2002).

Hong, J.W. & Quake, S.R. Integrated nanoliter systems. Nat Biotech 21, 1179-
1183 (2003).

Breslauer, D.N.L., P. J.; Lee, L. P. Microfluidics-based systems biology. Mol
BioSyst 2, 97—1 12 (2006).

Becker, H.G.C. Polymer microfabrication methods for microfluidic analytical
applications. Electrophoresis 21, 12-26 (2000).

Dolnik, V.L., S.; Jovanovich S. Capillary electrophoresis on microship.
Electrophoresis 21, 41-54 (2000).

72

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

Becker, H. & Giirtner, C. Polymer microfabrication technologies for microfluidic
systems. Anal Bioanal Chem 390, 89-111 (2008).

Barz, D.P.J. & Ehrhard, P. Model and verification of electrokinetic flow and
transport in a micro-electrophoresis device. Lab Chip 5, 949-958 (2005).

Amarie, D., Glazier, J .A. & Jacobson, S.C. Compact Microfluidic Structures for
Generating Spatial and Temporal Gradients. Anal Chem 79, 9471-9477 (2007).

Li, M.W., Huynh, B.H., Hulvey, M.K., Lunte, S.M. & Martin, R.S. Design and
Characterization of Poly(dimethylsiloxane)-Based Valves for Interfacing

Continuous-Flow Sampling to Microchip Electrophoresis. Anal Chem 78, 1042-
1051 (2006).

Tolan, N.V., Genes, L.I. & Spence, D.M. Merging Microfluidics with Microtitre
Technology for More Efficient Drug Discovery. JALA 13, 275-279 (2008).

Genes, L.I., Tolan, N.V., Hulvey, M.K., Martin, R.S. & Spence, D.M. Addressing
a vascular endothelium array with blood components using underlying
microfluidic channels. Lab Chip 7, 1256-1259 (2007).

Wu, H., Huang, B. & Zare, R.N. Construction of microfluidic chips using
polydimethylsiloxane for adhesive bonding. Lab Chip 5, 1393-1398 (2005).

Unger, M.A., Chou, H.-P., Thorsen, T., Scherer, A. & Quake, S.R. Monolithic
microfabricated valves and pumps by multilayer soft lithography. Science 288,
113-116 (2000).

Houston, J.G. & Banks, M. The chemical-biological interface: developments in
automated and miniaturised screening technology. Curr Opin Biotechnol 8, 734-
740 (1997).

Schillinger, E., Krais, T., Lehmann, M. & Stock, G. Iloprost. Am J Cardiovasc
Drugs 209-231 (1986).

Carroll, J.S., Ku, C.-J., Karunarathne, W. & Spence, D.M. Red Blood Cell
Stimulation of Platelet Nitric Oxide Production Indicated by Quantitative

Monitoring of the Communication between Cells in the Bloodstream. Anal Chem
79, 5133-5138 (2007).

73

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

Ballas, S.K., Dover, G.J. & Charache, S. Effect of hydroxyurea on the rheological
properties of sickle erythrocytes in vivo. Am J Hematol 32, 104-111 (1989).

Athanassiou, G., Moutzouri, A., Kourakli, A. & Zoumbos, N. Effect of
hydroxyurea on the deformability of the red blood cell membrane in patients with
sickle cell anemia. Clin Hemorheol Microcirc 35, 291-295 (2006).

Dittrich, P.S. & Manz, A. Lab-on-a-chip: microfluidics in drug discovery. Nat
Rev Drug Discov 5, 210-218 (2006).

Cohen, C.M. A path to improved pharmaceutical productivity. Nat Rev Drug
Discov 2, 751-753 (2003).

Galson, S. in Piper J affi'ay 17th Annual Healthcare Conference (FDA Center for
Drug Evaluation and Research, New York, NY; 2005).

Malakoff, D. Spiraling Costs Threaten Gridlock. Science 322, 210-213 (2008).

Varma Manthena, V.S. et al. Biopharrnaceutic classification system: a scientific

fiamework for pharrnacokinetic optimization in drug research. Curr Drug Metab
5, 375-388 (2004).

FDA (US. Department of Health and Human Serives Food and Drug
Administration, 2004).

Kola, I. & Landis, J. Opinion: Can the pharmaceutical industry reduce attrition
rates? Nat Rev Drug Discov 3, 711-716 (2004).

Hood, L. & Perlmutter, RM. The impact of systems approaches on biological
problems in drug discovery. Nat Biotechnol 22, 1215-1217 (2004).

Carroll, J. et al. An altered oxidant defense system in red blood cells affects their
ability to release nitric oxide-stirnulating ATP. Mol Biosyst 2, 305-311 (2006).

Subasinghe, W. & Spence, D.M. Simultaneous determination of cell aging and
ATP release from erythrocytes and its implications in type 2 diabetes. Anal Chim
Acta 618, 227-233 (2008).

74

57.

58.

59.

Sprague, R.S., Stephenson, A.H., Bowles, E.A., Stumpf, M.S. & Lonigro, A.J.
Reduced expression of Gi in erythrocytes of humans with type 2 diabetes is
associated with impairment of both cAMP generation and ATP release. Diabetes
55, 3588-3593 (2006).

Meyer, J .A., Froelich, J .M., Reid, G.E., Karunarathne, W.K.A. & Spence, D.M.
Metal-activated C-peptide facilitates glucose clearance and the release of a nitric
oxide stimulus via the GLUTl transporter. Diabetologia 51, 175-182 (2008).

Tolan, N.V., Genes, L.I., Subasinghe, W., Raththagala, M. & Spence, D.M.
Personalized Metabolic Assessment of Erythrocytes Using Microfluidic Delivery
to an Array of Luminescent Wells. Anal Chem 81, 3102-3108 (2009).

75

CHAPTER 3: PERSONALIZED METABOLIC ASSESSMENT

3.1 METABOLIC SYNDROME: COMPETING PATHWAYS IN DIABETES
Metabolic syndrome (MetS) is generally accepted, as the presence of, first and
foremost, insulin resistance, central adiposity or obesity, dyslipidemia in the form of
elevated triglycerides and decreased high-density lipoprotein (HDL) cholesterol, as well
as hypertension. Increased risk of cardiovascular disease and developing diabetes has
been reported to be anywhere from 2-5 times more likely with the diagnosis of MetS and

it has been shown to be a statistical determinant of micro- and macrovascular

complications.l These vascular complications, retinopathy, neuropathy, nephropathy and

cardiovascular disease are also risks associated with diabetes, 2'5 firrther exemplifying the

strong link with MetS.

Various therapies include the modification of lifestyle, as it pertains to healthy
dietary restrictions as well as the increase in daily activity. Pharmotherapies have
concentrated on treating these conditions by the administration of insulin, metformin, 0t-
glucosidase inhibitors, glucagon-like peptides (GLP-l), receptor agonists and dipeptidyl-
peptidase IV (PPP-IV) enzyme inhibitors.6 There is generally an observed weight gain
associated with these treatment options, which is also associated with worsening insulin
resistance. Clinical studies have demonstrated the physical toll MetS plays on the

vascular system, specifically, the increased vascular thickness in the arterial wall and also
wall stiffness,7 resulting in an increased risk of cardiovascular complications.

By employing the uFA that is first introduced in chapter 2, this research focuses

on the use of the flow-based system, providing a high throughput platform, to better

76

understand the relationship of the associated risk factors of MetS, specifically as it
pertains to oxidant stress and hypertension. This device allows for the simultaneous
determination of intracellular glucose metabolites within a RBC sample, while also

monitoring the resulting affects on cell function, as expressed by the ability to release

ATP. Furthermore, this work is an extension of the research performed in our groupg’10

to link the intracellular antioxidants produced through glucose metabolism in a dynamic
measurement scheme that is important for measuring the effect of oxidant stress on
cellular function. In this work, the relationship between the pentose phosphate and the
polyol pathways for glucose metabolism are investigated, as it pertains to the
complications exhibited by diabetic patients by using this in vitro model.

The most relevant application for this uFA is in pharmaceutical drug discovery, as
previously discussed in chapter 2, specifically monitoring the properties of absorption,

“-13

distribution, metabolism, excretion and toxicity (ADMET) and the implementation of

already available microtitre plate technology. Development of the ability to efficiently
screen multiple samples for various metabolic functions will allow for a more specialized

treatment of individual diabetic patients.

3.2 PPP: PRODUCTION OF GSH

Illustrated in figure 3.1, as glucose metabolism is shified to the pentose phosphate
pathway (PPP) to provide antioxidant defense for the cell, it is first phosphorylated by
hexokinase (HK) to produce glucose-6-phosphate (G6P). Glucose-6-phosphate
dehydrogenase (G6PD) will then oxidize G6P, forming 6-phosphogluconolactone (6GP)

upon the reduction of NADP+ to NADPH. NADPH is consumed in the production of

77

 

cuzou

Glucose

O OH
OH

Figure 3.1 Illustration of the pentose phosphate and polyol pathways available for the
metabolism of glucose within the red blood cell. Glucose is normally metabolized
through the pentose phosphate pathway where glucose is phosphorylated by hexokinase
(HK) to form glucose-6-phosphate (G6P) which is then oxidized to 6-phosphoglucono-
8-lactone (6GP) upon the reduction of NADP+ to NADPH. NADPH is an essential
cofactor for the production of reduced glutathione (GSH) from its disulfide form (GS-
SG), but it is also consumed in the polyol pathway where aldose reductase (AR)
converts glucose to sorbitol. Upon the oxidation of sorbitol to form fructose by sorbitol
dehydrogenase (SD), NAD” is reduced to form NADH.

78

reduced GSH, the most abundant non-enzymatic antioxidant, by way of glutathione

l4-l7

reductase (GSHR). Reduced GSH is formed from oxidized glutathione (GSSG),

which contains a disulfide bond between two GSH molecules, while NADPH is oxidized
back to NADP+.

The enzyme activity of G6PD has been shown to be lower in patients with
diabetes, leading to a reduced conversion of G6P to 6GP in the PPP.18 Likewise, there is

a reduction in the concentration of NADPH that is coproduced, leading to a reduced
concentration in GSH and antioxidant ability within the RBCS obtained from diabetic

patients, which is well documented throughout the literaturelg'23 Our group has shown

that with the onset of oxidative stress and with decreased levels of GSH, a reduction in

the release of ATP occurs which is thought to be due to a decrease in membrane
. . . 3
flexrbilrty.
The flexibility of the cellular membrane is restricted when insulted with oxidant

stressors such as peroxides, superoxides, and free radicals that oxidize spectrin instead of

the biological alternative GSH.16,18,19,24

Without sufficient GSH available, oxidant
stressors will oxidize spectrin, introducing disulfide bonds and ultimately stiffen the cell
membrane.”27 Irnportantly, our group and others have shown that RBCs from diabetic
patients release less ATP in comparison to RBCs from healthy, non-diabetic controls,

when these cells are subjected to mechanical deformation.8 This is thought to be due, in

part, to a reduction in the ability of the cells to fight off oxidant insults, resulting in
decreased membrane flexibility when subjected to oxidant stressors. Further work in our

lab has shown that the reduced level in RBC-derived ATP release coincides with a

79

decrease of intracellular NADPH concentrations10 as well as reduced levels of GSH

within RBCS from diabetics.28

Together, these results indicate that RBCs obtained from diabetic patients have
reduced intracellular NADPH, the necessary precursor for GSH production, rendering the
cell susceptible to oxidant stress. This oxidant stress has been shown to decrease the
RBC-derived ATP release, which is a major contributor for endothelium-derived NO
production and vasodilation action. This decrease in NO is thought to result in reduced
vasodilation and, ultimately, increased hypertension and prevalence of the observed

complications associated with diabetes.

3.3 POLYOL PATHWAY: SORB FORMATION AND DIABETIC COMPLICATIONS

In the case of chronic hyperglycemia when blood glucose levels are elevated,
aldolase reductase (AR), which typically has a low affinity for glucose, initiates the
polyol pathway and converts excess glucose to sorbitol (SORB). This process requires
the oxidation of NADPH to form NADP+ which reduces the substrate available for GSH
production in the competing PPP. SORB is then metabolized to fructose upon the
reduction of NAD+ to NADH with the coenzyrne sorbitol dehydrogenase (SD). The
buildup of SORB within the nerve and lens tissues, due to its reduced permeability
through the cell membrane, results in osmotic stress on the cell as the body tries to
equilibrate the concentration gradient over the membrane. This results in structure and

function abnormalities that is thought to be the cause for neuropathy and retinopathy

complications exhibited by patients with diabetes.29'3l

80

The research presented here attempts to demonstrate that increased metabolism of
glucose through the polyol pathway may lead to a decrease in antioxidant ability of the
RBC, resulting in reduced deformability upon oxidant attack. This may play an
important role in the observed complications associated with diabetes, because a reduced
deformability of the RBC has been well established to result in decreased shear-induced

ATP release. ATP is a known stimulus for the production of endothelium-derived NO,
which plays an important role in determining vascular caliber.”36 It is speculated that a

reduced NO production from the endothelial lining of the vascular system could result in
reduced vasodilation, which may decrease the ability of RBCs to deliver 02 to the tissues

surrounding the microvascular system.

3.4 GLUCOSE METABOLISM AND AFFECT ON ATP RELEASE

It has been speculated that reduced flexibility in the RBC membrane is due to the
malfiinction of the PPP, namely reduced G6PD activity.8’10 However, a re-examination

is necessary because of the complex involvement of both the pentose phosphate and

polyol pathways in glucose metabolism, specifically, the intersection of the shared
coenzyrne, NADPH. The most recently reported methodg’lo of determining G6PD

activity by quantifying the intracellular NADPH concentration, also simultaneously

accounts for the activity of AR, which oxidizes NADPH to form SORB. The proposed
reduction of G6PD activity in RBCs from diabetic patients8 may actually be due to an

increase in the polyol pathway and AR activity. The quantification of NADPH

concentration cannot distinguish the enzyme activities of AR and G6PD individually;

81

however it can effectively measure the amount of NADPH available for the reduction of
GSSG to GSH.

The more recent “redox hypothesis” focuses on the competing pentose phosphate
and polyol pathways, specifically their intersection of the required NADPH substrate.3 7'41
Research is needed to determine if the previously found reduction of G6PD activity is the

foundation for the occurrence of diabetic complications or if the increased flux of glucose

metabolism through the polyol pathway plays a vital role.

3.5 EXPERIMENTAL
Previous experimentsg’9 have utilized fluorescence detection of NADPH and GSH

by employing non-flow fluorometer. Here, these same reactions will be monitored using
fluorescence microscopy within a flow system, specifically the itFA, as diffusion of
cellular components through the track-etched polycarbonate (TEPC) membrane react
with isolated fluorescence probe solutions within the array of wells above. RBC samples
and standard solutions are delivered to each row of wells that are loaded with the
corresponding fluorescence probes for the analytes of interest, namely NADPH, GSH,
SORB and ATP. In order to effectively measure the enzyme activities of G6PD, GSHR,
AR and SD, concentrations of NADPH, GSH, SORB and ATP are quantified in
individual wells by fluorescence and chemiluminescence detection. RBC samples are
flowed at an optimized flow rate and time through the underlying microfluidic channels
such as to maximize the sensitivity of the assay.

Using a fluorescence microscope, the fluorescence intensities are measured by

analyzing each fluorescence image for average pixel intensities. The fluorescence

82

intensities are proportional to the concentrations of each metabolite present in the
solutions other than ATP, which will utilize microscopy to image the chemiluminescence
intensity produced without employing the excitation source. The experimental set up for
the determination of intracellular metabolites and the simultaneous detection of ATP
release is illustrated in figure 3.2. The entire uF A can be imaged by taking advantage of
the low magnification capabilities of the Olympus MVX fluorescence steromicroscope
(Center Valley, PA). The MVX microscope utilizes a mercury-arc lamp fitted with a
carousel of various filter cubes allowing for fast and instantaneous wavelength selection.
The microfluidic delivery of the samples to the wells is important both analytically
and physiologically, because the ATP release from the RBCs is stimulated by flow-
induced shear stress. By incorporating microfluidic delivery of the intact cells to the
wells, a portion of the cellular oxidant status and cellular function can be monitored
simultaneously. Moreover, calibration and measurement of the analytes are performed in
the presence and absence of various inhibitors on a single device. Furthermore,
simultaneous imaging of the wells that are addressed by the underlying channels
represent a starting point for attempts to merge microfluidic and microtitre-plate

technology.

3.5.1 OBTAINING ISOLATED RBCS

All reagents were fi'om Sigma Chemical (St. Louis, MO) unless otherwise noted.
The RBCS used in these studies were obtained from male New Zealand White rabbits
following the protocols approved by the Animal Investigation Committee at Michigan

State University. As previously described,4246 rabbits (2.0-2.5 kg) were anesthetized

83

 

CCD Camera ——>

Hg-Arc Lamp —> '

viii

 

 

   

Microfluidic Array \3‘ .
t- i - a: - "
Syringe Pump Stereomicroscope Data Analysis

Figure 3.2 Experimental set up used to simultaneously monitor intracellular
metabolites as well as the extracellular deformation-induced ATP release from RBC
samples. Here, a displacement syringe pump is used to deliver solutions through the
underlying microfluidic channels of the uFA while the steromicroscope images both
fluorescence intensities and chemiluminescence produced in the determination of
NADPH, GSH, SORB and ATP.

84

using ketamine (8 mL/kg, i.m.) and xylazine (1 mg/kg i.m.) followed by pentobarbital
sodium (15 mg/kg i.v.). Rabbits are then ventilated with room air, at a rate of 20
breaths/min using a tidal volume of 20 mL/kg, by placing a cannula in the trachea. A
catheter is then placed into the carotid artery for the administration of heparin (500 units,
i.v.) prior to the exsanguination and collection of approximately 80 mL whole blood,
which is obtained through the same catheter.

Whole blood is then centrifuged at 500 x g at 25 °C for 10 min at which time, the
platelet-rich plasma and buffy coat were collected for other experimentation, as
illustrated in figure 3.3. Here, the illustration shows the separation of the platelet-rich
plasma and buffy coat, containing platelets and leukocytes, from the RBCS. The
remaining RBCS were resuspended and washed three times in a physiological salt
solution (PSS) (containing in mM, 4.7 KCl, 2.0 CaClz, 140.5 NaCl, 12 MgSO4, 21.0
tris(hydroxymethyl)aminomethane, 5.6 glucose with 5% bovine serum albumin at a final
pH of 7.4). All samples were prepared, the final solution’s hematocrit is determined for

subsequent dilutions, and experiments were performed within 8 hrs of RBC collection.

3.5.2 STANDARD MEASUREMENTS OF NADPH, GSH, SORB AND ATP

A single fluorescence image can be used to simultaneously provide the
information necessary to construct a standard calibration curve as well as quantify the
concentration of NADPH, GSH, and SORB contained in, as well as the concentration of
ATP released fiom a single RBC sample. The microfluidic array employed here
significantly reduces the time required to make multiple measurements when compared

to a single measurement detection method. Sample preparation and fabrication of the

85

PRP and Buffy Coat
(Leukocytes and Platelets)

 

 

Red Blood Cells “

 

 

Figure 3.3 Illustration of centrifuged whole blood where
the red blood cells are separated from the platelet-rich
plasma (PRP) and buffy coat that contain the leukocytes
and platelets.

86

microfluidic array aside, the actual time required for the entire analysis including the
incubation required for the analytes in the channels to react with the luminescent probes

immobilized within each well, is approximately 20 minutes.
As previously reportedg, the activity of G6PD was monitored by measuring the

fluorescence intensity of red-fluorescent resorufin that was produced when NADPH,
found within lysed RBC samples, was incubated with a resazurin reaction mixture
containing diaphorase (Vybrant Cytotoxicity Assay Kit, Invitrogen Corp., Carlsbad, CA).
This reaction mechanism is illustrated in figure 3.4, and although the concentration of
NADPH is not exclusive to G6PD as previously discussed, it can however, be directly
linked to the concentration available for the production of GSH for antioxidant defense.
Standard solutions of NADPH in phosphate buffered saline (PBS) solution
(containing in g/L, 0.144 KH2P04, 9.00 NaCl, 0.795 NazPO4 (anhydrous), with a final pH
of 7.4 :1: 0.1) ranging from 12.5 to 200 uM and a single 0.7% lysed RBC sample solution,
as illustrated in figure 3.5, were pumped through the underlying microfluidic channels for
20 min at a flow rate of 1.0 uL/min. These standard solutions were prepared from a 1.0
mM stock solution (8.3 mg dissolved and diluted with PBS to 10 mL). The RBC sample
flowed through channel 6, was prepared by diluting the washed RBCs to a 0.7%
hematocrit in deionized and distilled water (DDW, 18.1 M!) resistivity) which results in
complete lysis, thus enabling the intracellular concentrations of the NADPH and GSH to
be determined. The resazurin fluorescence probe (30 uM) was pipetted into each of the
18 individual wells of the uFA, which was reduced by NADPH to produce the
fluorescent resorufin product having excitation and emission wavelengths of 563 and 5 87

nm respectively. Fluorescence images were obtained using the Olympus MVX macro

87

,£1: 31>; M... ———~,£1” £10, +

Resazurin Resorufin

Figure 3.4 Illustration of the reaction used to determine the concentration of NADPH
in various standard and sample solutions employing the uF A. Here resazurin, in the
presence of diaphorase, is oxidized to the red-fluorescent resorufin as NADPH is
reduced to NADP+. Resorufin has excitation and emission wavelengths of 563 and
587 nm respectively.

88

200 pl“ NADPH

100 11M NADPH

50 ill“ NADPH

25 “M NADPII

12.5 ”M NADPH

0.7% Lyserl RBC 5

 

Figure 3.5 Fluorescence image of the 6-channel uFA used to
determine the concentration of NADPH found within a 0.7% lysed
RBC solution by simultaneously monitoring the fluorescence
intensities from five NADPH standards ranging in concentration from
12.5 to 200 uM. Here, each standard and the single RBC sample
solutions are flowed through the underlying microfluidic channels for
20 min while the resazurin contained within the wells above is
reduced to the red-fluorescent resorufin product as NADPH is
oxidized to form NADPL

89

zoom fluorescence microscope fitted with a resorufin filter cube that has maximum
excitation and emission wavelengths of 570 nm and 585 nm respectively. The images
were then analyzed for average pixel intensities, averaging each of the three wells thus
enabling triplicate measurements, and plotted versus the known concentration of NADPH
within each standard solution.

The methodology developed in our group to determine the concentration of GSH
within a RBC sample9 was modified slightly, however the enzymatic process is the same

as illustrated in figure 3.6, only that the semi-preparative properties of the membrane
within the uF A omitted the need to perform standard addition. The fluorescence probe
mixture of 500 M monochlorobimane and 8 U/mL glutathione-S-transferase (from
equine liver, Sigma G6511) (MCB-GST) were pipetted into the 18 array-like wells and
upon reaction, the fluorescence intensity of the GSH-MCB product with excitation and
emission wavelengths of 370 nm and 478 nm respectively was determined. Here, the

GST was added to aid in the binding of GSH to the MCB probe, significantly reducing
the time required for analysis.9 Figure 3.7 illustrates the fluorescence image obtained as

standard solutions of GSH in PBS ranging from 2.5 to 40 uM and a single 0.7% lysed
RBC sample solution were pumped through the underlying microfluidic channels for 20
min at a flow rate of 1.0 uL/min. These standard solutions were prepared from a 10 mM
stock solution (0.0307 g in 10 mL PBS). The microscope was fitted with a 4’,6’-
diamidino-2-phenylindole (DAPI) filter cube having excitation and emission wavelengths
of 372 nm and 456 nm respectively. The images were analyzed for fluorescence
intensity, averaging the three wells for each channel, and plotted against the known GSH.

concentrations within solution to obtain the standard working curve.

90

GST
Hscé: iglCHa +GSH ——~> H302;E fé~CH3 +HC1

CIHZC GSHZC

MCB MCB-GSH
Figure 3.6 Reaction mechanism for the determination of GSH in various standard
and sample solutions where MCB and GSH form a fluorescent product (MCB-GSH)

with excitation and emission wavelengths of 370 nm and 478 nm in the presence of
glutathione-S-transferase (GST).

91

2.5 pM GSH

SuMGSH

10 pM GSH

20 ”M GSH

40 pM GSH

0.7% Lysed RBCS

 

Figure 3.7 Fluorescence image of the 6-channel uFA used to
determine the concentration of GSH found within a 0.7% lysed
RBC solution by simultaneously monitoring the fluorescence
intensities from five GSH standards ranging in concentration from
2.5 to 40 uM. As with the NADPH determination, each standard
and the single RBC sample solutions are flowed through the
underlying microfluidic channels for 20 min while the
monochlorobimane and glutathione-S-transferase (MCB-GST)
contained within the wells above forms the fluorescent MCB-GSH
product.

92

Using the luciferin/luciferase enzyme-based reaction, detailed in figure 3.8 and as

lO,43-45,47—49

previously employed using microbore tubing, an ATP calibration curve was

constructed by monitoring the chemiluminescence produced, which is directly
proportional to the concentration of ATP within each standard solution. To prepare the
luciferin/luciferase mixture used to measure the ATP by chemiluminescence, 5 mL of
DDW were added to a vial containing luciferase and luciferin (FLE-SO, Sigma). In order
to enhance the sensitivity of the assay, 2 mg of luciferin were added to the vial.

This assay was used to determine the ATP concentration released from 7% RBC
sample solutions with and without the addition of diamide, a known cell-membrane

24,50

stiffener. As illustrated in figure 3.9, ATP standard solutions ranging from O.25-1.0

M in PSS and the RBC samples are flowed through the microfluidic channels, reacting
with the luciferin/luciferase solution contained in each of the 18 wells that comprise the
uF A. The ATP standard solutions were prepared from a 100 uM stock solution (27.6 mg
in 500 mL DDW) in PSS. In addition, two 7% RBC samples were prepared by diluting
the washed RBCS to 7% in PSS; however, one of these RBC samples is incubated in PSS
containing 20 uM diamide. The diamide incubation occurred for 15 min prior to starting
flow within the array in order to maximize the effect of the diamide solution and because

the RBCS are able to recover from the oxidant attack and regain membrane flexibility to
near initial levels within 30 to 40 min.8 The excitation lamp was not employed for the

chemiluminescence determination of ATP and the images were obtained without a
wavelength selection filter cube. Images were obtained with a 5 sec exposure time in

order to maximize the chemiluminescence intensity from each well. The pixel intensities

93

   

~ 11 1 I1
\ / + HO—P—P—P
.0 > < J.- J. J,
Luciferin ATP
Firefly Lucifcrase NH2

+ Mg2+

.0 S

   

Adenyl-lucifen'n

02 Luminescence 0” 0”

NHrj/jhzo + co2

Oxyluciferin

'0

Figure 3.8 Reaction mechanism of the luciferin-catalyzed chemiluminescent reaction
employed here for the determination of mechanical deformation induced-ATP release.

94

0.25 ”M ATP

0.50 uMATP

' 0.75 M ATP

1.0 uMATP

7% RBCS

7% RBCS + 20 uM Diamide

 

Figure 3.9 Chemiluminescence image, without the excitation source employed, for
the determination of ATP within standard solutions as well as from the mechanical
deformation of a 7% RBC sample with and without the incubation of diamide, a
known cell-membrane stiffener. Here the luciferin and luciferase (L/L) solutions are
pipetted into the 18 wells that comprise the uFA while the solutions are flowed
through the microfluidic channels below.

95

for the standard solutions were used for linear regression analysis in order to determine
the concentration of ATP release from the two non-lysed RBC samples.

The calibration curves for both NADPH and GSH were performed simultaneously
by combining NADPH standards ranging from 25 to 200 uM and GSH standards ranging
from 5 to 40 uM together into four individual standard solutions. Combining NADPH
and GSH in the same standard solutions, while monitoring the fluorescence intensities
produced in the two separate columns of wells, allowed for the construction of two
separate calibration curves and only required three channels. Each metabolite was
effectively measured in the presence of the other within each standard solution and from
the same RBC sample. Here, the standard preparation was identical to that described
above; however, the incremental standard solutions of NADPH and GSH were prepared
in the same volumetric flask.

As illustrated in figure 3.10, the first 6 channels of the lZ-channel array contain
the resazurin probe solution which was only loaded in the first column of wells, while the
MCB-GST probe solution was loaded into the second column of wells. The NADPH and
GSH present in each standard solution reacts with their respective fluorescent probe
immobilized within wells above, allowing for the determination of both NADPH and
GSH within each solution pumped through the underlying microfluidic channels. Here,
the standard samples and a 0.7% lysed RBC sample were pumped through the underlying
channels for 20 min at which point, the average pixel intensities were used to construct a
calibration curve to determine the concentration of both NADPH and GSH within the

same RBC sample.

96

RES MCB

25 uM NADPH & 5 uM GSH

50 uM NADPH & 10 uM GSH

100 MM NADPH & 20 uM GSH

200 uM NADPH & 40 MM GSH

0.79 o Lysed RBCS

 

0.7% Lysed RBCS + 100 [AM DHEA

Figure 3.10 Fluorescence image of the simultaneous detection of both
NADPH and GSH within standard sample solutions as well as within a
0.7% lysed RBC solution with and without the inhibition of G6PD by
dehydroepandroesterone (DHEA). As before, solutions were flowed
through the underlying microfluidic channels and both the resazurin
(RES) and monochlorobimane and glutathione-S—transferase (MCB-GST)
were pipetted into only one column in order to detect both analytes from
within the same solution. Images were obtained as before except that the
resorufin and DAPI filter cubes were employed to select the appropriate
wavelengths for each analyte, NADPH and GSH, respectively.

97

On a single 12-channel uFA the determination of intracellular metabolites
(NADPH and GSH) were performed while simultaneously monitoring the cellular
function (as ATP release) from the same RBC sample. Figure 3.11 shows the
fluorescence image obtained for the determination of NADPH and GSH from within two
lysed RBC samples while simultaneously monitoring the concentration of ATP release
from two non-lysed RBC solutions, each incorporating calibration standards.

The standard samples containing NADPH (25-200 uM) and GSH (5-40 M) were
prepared as detailed above and flowed through the first four channels of the uFA. Two
0.7% lysed RBC solutions with and without the incubation (prior to lysis) with
dehydroepandroesterone (DHEA), a known G6PD inhibitor, were flowed through
channels five and six. A 1 mM stock solution of DHEA was prepared by dissolving
0.0043 g in 10 mL of PSS which was subsequently diluted to obtain a solution containing
100 uM DHEA and 7% RBCS. This solution was incubated for 30 min in order to inhibit
G6PD and effectively lower the GSH and NADPH concentrations within the cell,
simulating RBCS obtained fi'om a patient with diabetes.”54 A control sample was also
prepared, containing 7% RBC in PSS without DHEA. After the 30 min incubation, both
cell solutions were diluted to 0.7% and lysed by adding 200 uL of the 7% RBC solutions
to 1.8 mL DDW, individually.

The standard and sample solutions were pumped through the underlying
microfluidic channels for 20 min at a flow rate of 1.0 uL/min at which time, the
fluorescence images were obtained using both the resorufin and DAPI filter cubes. The
fluorescence intensities observed in each well are directly proportional to the

concentration of each NADPH and GSH within the solutions, allowing for simultaneous

98

RES MC B

25pMNADPH&5pMGSH

501d“ NADPH & 1011M GSH

100 "M NADPH & 20 [AM GSH

200nm NADPH & 40 ”M GSH

0.7% Lysed RBCS

0.7% Lysed RBCS + 100 Ml DHEA

0.25 le ATP

0.50 pm ATP

0.75 "M ATP

1.00 ”M ATP

7% RBCS

7% RB(‘s+ 20 ”M DIAMIDE

 

 

Figure 3.11 Images for the determination of intracellular NADPH and GSH as well
as for the mechanical deformation-induced ATP release from RBC samples. The
first six channels, employ fluorescence spectroscopy where two columns of wells
contain RES and MCB fluorescence probes while the last six channels, employ
chemiluminescence where all three columns contain the UL reaction mixture.

99

detection of both metabolites by simply changing the excitation and emission
wavelengths.

The ATP assay that was described above was carried out on the last six channels
of the same uF A (figure 3.11). This assay monitored cell fimction, requiring non-lysed
solutions, making it was necessary to carry out the ATP determination in separate
channels from those used to monitor NADPH and GSH. ATP standard solutions ranging
in concentration from 0.25-1.0 uM were pumped through channels seven through ten,
while two 7% non-lysed RBC samples in the presence and absence of 20 M diamide, a
known cell stiffener, were pumped through the last two channels. The pixel intensities
for the standard solutions were used for linear regression analysis in order to determine

the concentration of ATP release from the two non-lysed RBC samples.

3.5.3 THE AFFECT OF G6PD AND AR INHIBITION

The complications exhibited by patients with diabetes have been reduced with

29,55-63

continued treatment of aldose reductase inhibitors (ARIs). However, clinical trials

have revealed that pharmaceutical drugs developed for the inhibition of AR, in order to
promote the PPP for glucose metabolism and increase the intracellular level of GSH,
have serious side affects.62 These negative side effects, including hypersensitivity
reactions and liver damage, result when administering ARIs such as Fidarestat, Aldos,
Sorbinil, Tolrestat, and Statinil.

Recent studies have indicated that these side effects can be eliminated by the use
of ascorbic acid (AA), or vitamin C for the inhibition of AR?“56 It was concluded that

administration of approximately 1,000 mg of AA per day, results in a significant decrease

100

in the concentration of sorbitol found within RBCS as well as the previously reported
diabetic complications (such as neuropathy and retinopathy). It is suspected that this is
due to an increase in intracellular metabolites of the PPP and antioxidant status,
preventing the stiffening of the cellular membrane when insulted with oxidant stress and
increasing ATP release by maintaining RBC deformability.

The inhibition of G6PD will allow for the quantification of NADPH found within
RBCS due to the polyol pathway and indirectly measure the activity of AR. Here,
quantification of NADPH and GSH with alternating inhibition of AR and G6PD, will
allow for the determination of enzyme activities of each and for further conclusions to be
made regarding any enzyme deficiency of diabetic patients. It has been shown that the
inhibition of AR and the polyol pathway will redirect the metabolism of glucose though
the pentose phosphate pathway, decreasing the concentration of sorbitol and increasing
the concentration of GSH found within various cell types.”61 Preliminary research has
shown that the incubation of RBCS with AA will increase the release of ATP once treated
with DHEA, a known G6PD inhibitor, compared to a control RBC sample not incubated
with AA prior to introduction of DHEA.

The concentration of RBC-derived ATP was determined using the
chemiluminescence reaction with the luciferin/luciferase assay solution by implementing
the flow-based system illustrated in figure 3.12. A single syringe pump was used to flow
ATP standards and various RBC samples, as well as the luciferin/luciferase solution
through microbore tubing, with an internal diameter of 50 um. The samples were flowed
at a rate of 6.7 uL/min and combined in a T-junction (Upchurch Scientific, Oak Harbor,

WA) allowing for the solutions to mix before combining into another 50 cm section of

101

Dark Box

 
   
  
 
 

Luciferin/Luciferase

Standards or RBC Samples
I u... ‘..e

    

Microbore Tubing

Figure 3.12 Schematic drawing of the experimental setup used to determine RBC-derived
ATP release in a continuous flow system employing 75 um internal diameter tubing. Here
the luciferin and luciferase enzymatic assay was employed, monitoring the
chemiluminescence produced that is directly proportional to the concentration of ATP
released fi'om various samples.

102

tubing with the polyimide coating removed allowing for the chemiluminescence
produced to be detected by the PMT housed within a light excluding box. The current
collected by the PMT from the chemiluminescence produced, is converted to voltage,
measured in real time, and plotted at a rate of 10 measurements per second for 30 seconds
onto a computer using LabView software (National Instruments, Austin, TX) with a
program that was written in-house.

Figure 3.13 illustrates that the concentration of ATP released from RBCS
incubated with a 1.0 mM AA solution prior to a 100 uM DHEA solution, was greater
than that from RBCS treated with a 100 uM DHEA solution alone. A 10 mM AA stock
solution was prepared by dissolving 0.0176 g in 10 mL PSS, prior to the combination
with a second PSS solution to obtain a 1 mM AA and 20% RBC solution. This solution
was incubated for 30 min prior to combining with a final PSS solution to obtain 7%
RBCS and 100 uM DHEA which was incubated for an additional 30 rrrin prior to
determining the ATP release. The results indicate that the incubation of RBCS with AA
prior to DHEA will increase ATP release, however, further studies are necessary to
conclude that the incubation of RBCS with AA inhibits AR and promotes the pentose
phosphate pathway and the production of GSH.

Finally, recent work has been performed to include both AA and DHEA along
with various glycemic environments for the fluorescence determination of SORB”69
along with NADPH, GSH and ATP onto the uF A. As described above, the fluorescence
probe solution was pipetted into the array like wells of a single column, where here the
solution comprised of 2 U/mL SDH and 3.0 mM NAD+. The SDH within the

fluorescence probe solution converts SORB to fructose while NAD+ is reduced to NADH

103

[ATP]

 

0.4

 

 

 

7% RBCS + DHEA w/ AA + DHEA

Figure 3.13 ATP release from 7% RBCS with the inhibition of G6PD using
DHEA, resulting in a decrease in the concentration of deformation-induced ATP
release in 50 um diameter microbore tubing. Upon the pre-incubation of RBC
with a 5 mM solution of ascorbic acid (AA), a known aldose reductase inhibitor,
the reduction in ATP release due to the inhibition of G6PD with DHEA is
somewhat opposed (p < 0.05). The mean concentrations of ATP released are
reported where the error is represented as the SEM.

104

which is fluorescently active (excitation and emission wavelengths of 340 nm and 460
nm respectively). As illustrated in figure 3.14, the concentration of each NADPH, GSH,
and SORB were monitored within the first six channels of the uF A as RBC incubated
within various glucose environments, as well as, with DHEA and AA, flowed through the
underlying microfluidic channels. Here, the samples were prepared as detailed above, for
the preparation of RBCS with 1 mM AA and 100 uM DHEA separately, and combined.
Additionally, RBC samples were prepared. in PSS containing various concentrations of
glucose, which were incubated for 6 hrs before they were pumped through the underlying
rrricrofluidic channels for 20 min. Each of samples flowed through the first six channels
of the uFA were prepared as 0.7% RBCS in DDW, lysing the cells for the determination
of intracellular metabolites. The samples flowed through the last six channels were intact
RBCS that were monitored for cell function by way of mechanical deformation-induced

ATP release.

3.6 RESULTS AND DISCUSSION

When NADPH, GSH, and ATP were optimized on separate microfluidic devices
by monitoring the fluorescence intensities of the three wells for each underlying channel,
the averages were used for the linear regression analysis. This effectively represented the
variation associated with the consecutive sample analysis along the channel length and
indicated that the reproducibility of each standard fluorescence intensity was appropriate
for concentration determination of the analytes within the RBC samples. The flow rate of

1.0 uL/min was chosen along with the internal pore diameter of the TEPC membrane (0.1

105

0.7% Lyscd RBCS t 0 lel GLU PSS

+ 5 lel GLU PSS

t 10 m.\rl GLU PSS

+ lml\'l.»‘\_>‘\ + 100 MM DHEA

+ lmI\/l AA

+ 100 11th DHEA

7°15 RBCS * 0 mM GLU PSS

+ 5 ml\1 GLU PSS

1 10 mM GLU PSS

+ lml\1 AA + 100 tLIVI DHEA

l lmM AA

+ 100 1.1M DHEA

 

L/L

Figure 3.14 Images for the determination of intracellular NADPH and GSH
as well as for the mechanical deformation-induced ATP release from RBC
samples. The first six channels, employ fluorescence spectroscopy where two
columns of wells contain RES and MCB fluorescence probes while the last six
channels, employ chemiluminescence where all three columns contain the UL
reaction mixture.

106

um) to minimize the bulk fluid flow through the membrane, preventing dilution of the
fluorescence probe solution located in the wells.

For each sample analysis, background subtraction was performed and standard
solutions were incorporated onto each device in order to reproducibly and accurately
determine analyte concentrations with the RBC samples. The experimentation was
performed intentionally to allow for only a single dependant variable, the analyte
concentration, eliminating any inconsistencies within the set up, including but not limited
to, variable excitation and emission intensities resulting from the wavelength selection
filter cubes, any possible cross-talk between wavelengths, and heterogeneity of channel-
well interface surface area. Here, standard and sample solutions were analyzed
approximately 30 min afier preparation in order to keep analyte degradation within
solution minimal and consistent. More specifically, each solution was pumped through
the underlying channels in a manner that would take into account the reaction kinetics
between the analytes and the fluorescent probe solutions and the incubation time was

determined by the commencement of fluid flow.

3.6.1 QUANTITATIVE DETERMINATION WITHIN THE RBC

As shown in figure 3.15a, the 0.7% RBC sample was determined to contain on
average 31.06 i 4.12 uM NADPH and 22.55 :t 2.47 uM GSH, which upon G6PD
inhibition using DHEA, was reduced to 22.99 1: 4.63 uM NADPH and 14.60 i 3.14 1.1M
GSH, where the variance is reported as the standard error of the mean. The

concentrations of NADPH and GSH for the 0.7% lysed RBC sample without DHEA

107

40

 

a
) _ NADPH (n = 6)
1:1 GSH (n = 7)
30 —

20—

Concentrations (uM)

    

 

 

0.7% RBCS + DHEA

0.7

 

3 b)
. = 12
0.6 -_ (n )
0.5 5
0.4 f

0.3 {

Concentration (uM)

0.2 f

0.1 {

 

 

   

 

0.0' . ' " .
7% RBCs + DIAMIDE

Figure 3.15 Concentrations of (a) NADPH, GSH, and (b) ATP determined for lysed and
nonlysed RBC samples. Data representing figure 3.10, the concentrations of NADPH and
GSH for 0.7% RBC samples were determined to be 31.06 :t 4.12 and 22.5 i 2.47 uM,
respectively; the concentrations decreased for each upon the incubation of DHEA to 22.9 i
4.63 and 14.6 at 3.1 uM, respectively. The concentration of ATP release within the flow-
based uF A was determined to be 0.54 i: 0.04 M for a 7% RBC sample which was reduced
to 0.34 i 0.04 uM when incubated with diamide. Error bars represent the SEM for the
number of rabbits indicated in the figures.

108

incubation are 377 i 50 amol/RBC and 274 :1: 30 amol/RBC respectively, falling within

. . 9,16,19—21,7O
the range reported elsewhere usrng varrous methods.

Figure 3.15b illustrates the results for the quantitative determination of ATP
release and is given as average micromolar amounts where the variance is expressed as
standard error of the mean. Here, the concentration of ATP released from a 7% RBC
sample was determined to be 0.54 :1: 0.04 uM which was reduced to 0.34 i 0.04 uM when
incubated with diamide prior to analysis. ATP release, indicative of the shear forces that
are exerted on the RBCS, is specific to the channel dimensions of this particular
microfluidic device as well as the flow rate of the samples through the underlying
channels. Here, the flow rate for this device is approximately three-fold greater than that

found within 50 um i.d. microbore tubing, indicating ATP release that is notably
equivalent to previous reports.8

By incubating 7% RBC solutions in various glucose concentrations as well as
with 1 mM AA and 100 uM DHEA separately and combined, figure 3.16a represents the
preliminary data obtained from the 12-channel uFA for the determination of NADPH,
GSH, SORB and ATP release. Here, as the concentration of glucose in the PSS solution
increases there is a notable increase in the fluorescence intensity observed which is
relative to the concentrations of NADPH and SORB. This is suspected to be due to an
increase in glucose metabolism through both the pentose phosphate and polyol pathways.
The incubation of RBCS with AA prior to the inhibition of G6PD, with DHEA, shows
that there is an increase in the concentration of NADPH within the RBC, returning it to a

value consistent with untreated 0.7% RBCS.

109

 

 

 

         

 

I a) — NADPH
500 - GSH
- — SORB
E .
E) 400 —-
s
a)
0
g .
g 300 —
2 .
O
2
‘L .
200 / /
0 _
0.7% Lysed + 5mM +10mM + AA + DHEA + AA
RBCs + DHEA
3 b) I :3 ATP
(D
5 . 1,. I
8 240 - "
c _ .............
a)
U
U)
(D
.E .
E 230 - """
o .
0 I. I ‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7% RBCS +5mM +10 mM +AA +DHEA +AA

+ DHEA
Figure 3.16 Data representing the intracellular a) metabolites and b) resulting affect
on cellular function as expressed though mechanical deformation-induced RBC-
derived ATP release. The fluorescence intensity for the metabolites (NADPH, GSH,
SORB) and the chemiluminescence produced for monitoring the cellular function
(ATP release) are reported as average intensities of three areas sampled within each
well for RBCS in various glucose PSS solutions (0-10 mM) as well as with the
incubation of ascorbic acid (AA) and dihydroxyepiandrosterone (DHEA) where the
error reported represents the SEM.

110

The ATP release, as shown in figure 3.16b, indicates that as glucose
concentrations increase there is a corresponding decrease in the concentration of
mechanical deformation-induced ATP release, which is thought to be due to a stiffening
of the RBC as oxidant attack and glycosylation increases. The ATP release due to the
incubation with DHEA is shown to be reduced upon the inhibition of G6PD, which when
pre-incubated with 1 mM AA, the ATP release is slightly regained.

By expanding the concentrations of glucose within the PSS, as shown in figure
3.17, the concentration of metabolites within the 0.7% RBCs are shown to decrease when
incubated in hyperglycemic conditions (15-25 mM). In this range, there is a substantial
decrease in the intracellular concentration of NADPH and GSH. At elevated glucose
concentrations, it is suspected that as the glucose is metabolized through the polyol
pathway, the SORB is more rapidly metabolized to fructose resulting in the gradual
decrease in SORB that is shown. Future studies should include a SD inhibitor in order to
prevent fructose formation to more effectively monitor the glucose metabolism through

the polyol pathway.

3.6.2 EVALUATING THE DIABETIC-SIMULATED RBC

The determination of the concentrations of NADPH, GSH and ATP from RBC
samples treated to reflect the state of RBCs obtained from patients with diabetes is
demonstrated here. It is anticipated that this methodology will allow for the
differentiation of RBCS obtained from diabetic patients, compared to those obtained from
healthy non-diabetic controls. This uFA allows for multiple optical detection methods

(fluorescence and chemiluminescence) to be used simultaneously, while also

111

Fluorescence Intensity

 

       

 

 

80
' * - NADPH
GSH
_ - SORB
60 _ * p < 0.001
- * *
*
40 -
20 ‘ I l
0 _

OmM Glu +5mM Glu +10mM Glu +15mM Glu +20mM Glu +25mM Glu

Figure 3.17 Data representing the fluorescence intensities for monitoring the
intracellular concentrations of each metabolite, NADPH, GSH and SORB when
incubated in various glucose (GLU) containing PSS solutions (0-25 mM). The
fluorescence intensities are reported as the average of three sampled areas within
each well and the error is reported as the SEM.

112

incorporating a flow-based system without the need for sample preparation due to the
separation of the wells from the channels by a TEPC membrane. Here, simultaneous
analysis performed on a single platform may serve as a diagnostic tool for determining
the cause for hypertension complications as it relates to diabetes.71 This research
illustrates the ability to profile various analytes from within a lysed RBC solution, as well
as analytes released from intact RBCS.

Monitoring the concentrations of NADPH, GSH and ATP on a single device is an
initial attempt at merging metabolic profiling with cell function assays simultaneously.
For example, people with cystic fibrosis have RBCS that release very little ATP due to a
reduced activity of the CFTR.72 However, as mentioned above, people with diabetes
have sirrrilar ATP release patterns. Collectively, if a blood (or RBC) sample was to be
screened for ATP release and the release was found to be lower than normal, it would be
premature to conclude the subject has cystic fibrosis unless more quantitative information
about the sample was obtained. A glycosylated hemoglobin value (AlC), which is
generally higher in the poorly controlled diabetic but often normal in people with cystic
fibrosis, could be obtained to learn more about the initial screen. Therefore, coupling a
quantitative determination of ATP with another measured parameter would be useful, if

not necessary.

113

10.

11.

REFERENCES

The Metabolic Syndrome Is a Risk Indicator of Microvascular and Macrovascular
Complications in Diabetes. Diabetes Care 29, 2701-2707 (2006).

Fong, D.S. et al. Diabetic retinopathy. Diabetes Care 26, 226-229 (2003).

EDIC Intensive Diabetes Treatment and Cardiovascular Disease in Patients with
Type 1 Diabetes. N Engl J Med 353, 2643-2653 (2005).

CDC Centers for Disease Control and Prevention. National Diabetes Fact Sheet:
General information and National Estimates on Diabetes in the United States,
2007. US Department of Health and Human Services (2008).

Stephens, J .W., Khanolkar, M.P. & Bain, SC. The biological relevance and
measurement of plasma markers of oxidative stress in diabetes and cardiovascular
disease. Atherosclerosis 202, 321-329 (2009).

Joffe, D. & Yanagisawa, R.T. Metabolic syndrome and type 2 diabetes: can we
stop the weight gain with diabetes? Med. Clin. North Am. 91, 1107-1123 (2007).

Scuteri, A. et al. Metabolic syndrome amplifies the age-associated increases in
vascular thickness and stiffness. J Am Coll Cardiol 43, 1388-1395 (2004).

Carroll, J. et al. An altered oxidant defense system in red blood cells affects their
ability to release nitric oxide-stimulating ATP. Mol Biosyst 2, 305-311 (2006).

Raththagala, M., Root, P.D. & Spence, D.M. Dynamic Monitoring of Glutathione
in Erythrocytes, without a Separation Step, in the Presence of an Oxidant Insult.
Anal Chem 78, 8556-8560 (2006).

Subasinghe, W. & Spence, D.M. Simultaneous determination of cell aging and
ATP release from erythrocytes and its implications in type 2 diabetes. Anal Chim
Acta 618, 227-233 (2008).

Dittrich, P.S. & Manz, A. Lab-on-a-chip: microfluidics in drug discovery. Nat
Rev Drug Discov 5, 210-218 (2006).

114

 

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

Breslauer, D.N.L., P. J .; Lee, L. P. Microfluidics-based systems biology. Mol
BioSyst 2, 97—1 12 (2006).

Cohen, C.M. A path to improved pharmaceutical productivity. Nat Rev Drug
Discov 2, 751-753 (2003).

Meldrum, N.U. & Tarr, H.L.A. The reduction of glutathione by the Warburg-
Christian system. Biochem J29, 108-115 (1935).

Dumaswala, U.J., Zhuo, L., Jacobsen, D.W., Jain, S.K. & Sukalski, K.A. Protein
and lipid oxidation of banked human erythrocytes: role of glutathione. Free Radic
Biol Med 27, 1041-1049 (1999).

Aaseth, J. & Stoa—Birketvedt, G. Glutathione in overweight patients with poorly

controlled type 2 diabetes. Journal of Trace Elements in Experimental Medicine
13, 105-111 (2000).

Di Mascio, P., Dewez, B. & Garcia, C.R.S. Ghost protein damage by
peroxynitrite and its protection by melatonin. Braz J Med Biol Res 33, 11-17
(2000).

Beard, K.M., Shangari, N., Wu, B. & O'Brien, PJ. Metabolism, not autoxidation,
plays a role in a-oxoaldehyde- and reducing sugar-induced erythrocyte GSH
depletion: Relevance for diabetes mellitus. Mol Cell Biochem 252, 331-338
(2003)

Costagliola, C. Oxidative state of glutathione in red blood cells and plasma of
diabetic patients: in vivo and in vitro study. Clin Physiol Biochem 8, 204-210
(1990)

De Mattia, G. et al. Influence of reduced glutathione infusion on glucose
metabolism in patients with non-insulin-dependent diabetes mellitus. Metabolism
47, 993-997 (1998).

Dincer, Y., Akcay, T., Alademir, Z. & Ilkova, H. Effect of oxidative stress on

glutathione pathway in red blood cells from patients with insulin-dependent
diabetes mellitus. Metabolism 51, 1360-1362 (2002).

115

 

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Constantin, A., Constantinescu, E., Durnitrescu, M., Calin, A. & Popov, D.
Effects of ageing on carbonyl stress and antioxidant defense in RBCs of obese
type 2 diabetic patients. J Cell Mol Med 9, 683-691 (2005).

Nwose, E.U., Jelinek, H.F., Richards, R.S. & Kerr, P.G. in Redox Report, Vol. 11
99—104 (Maney Publishing, 2006).

Ghigo, D. et al. Changes in erythrocyte deformability resulting from alterations of
redox metabolism in diabetes mellitus. Ric Clin Lab 13, 375-381 (1983).

Fischer, T.M., Haest, C.W.M., Stoehr, M., Kamp, D. & Deuticke, B. Selective
alteration of erythrocyte deformability by sulfliydryl reagents. Evidence for an

involvement of spectrin in membrane shear elasticity. Biochim Biophys Acta 510,
270-282 (1978).

Schwartz, R.S., Madsen, J .W., Rybicki, A.C. & Nagel, R.L. Oxidation of spectrin
and deformability defects in diabetic erythrocytes. Diabetes 40, 701-708 (1991).

Cicha, I., Tateishi, N., Suzuki, Y. & Maeda, N. Rheological changes in human red
blood cells under oxidative stress: effects of thiol-containing antioxidants.
Pathophysiology 6, 121-128 (1999).

D'Amico Oblak, T., Meyer, J.A. & Spence, D.M. A microfluidic technique for
monitoring bloodstream analytes indicative of C-peptide resistance in type 2
diabetes. Analyst 134, 188-193 (2008).

Bravi, MC. et al. Polyol pathway activation and glutathione redox status in non-
insulin-dependent diabetic patients. Metabolism 46, 1194-1198 (1997).

Gabbay, K.H. Role of sorbitol pathway in neuropathy. Advances in Metabolic
Disorders, Supplement 2, 417-424 (1973).

Chylack, L.T., Jr. & Kinoshita, J .H. A biochemical evaluation of a cataract
induced in a high-glucose medium. Invest Ophthalmol 8, 401-412 (1969).

Ellsworth, M.L., Forrester, T., Ellis, C.G. & Dietrich, H.H. The erythrocyte as a
regulator of vascular tone. Am J Physiol Heart Circ Physiol 269, H2155-216l
(1995)

116

 

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

McCullough, W.T., Collins, D.M. & Ellsworth, M.L. Arteriolar responses to
extracellular ATP in striated muscle. Am J Physiol Heart Circ Physiol 272,
H1886-H1891 (1997).

Collins, D.M., McCullough, W.T. & Ellsworth, M.L. Conducted vascular
responses: communication across the capillary bed. Microvasc Res 56, 43-53
(1998)

Dietrich, H.H., Ellsworth, M.L., Sprague, R.S. & Dacey, R.G., Jr. Red blood cell
regulation of microvascular tone through adenosine triphosphate. Am J Physiol
Heart Circ Physiol 278, H1294-H1298 (2000).

Sprague, R.S., Ellsworth, M.L., Stephenson, A.H. & Lonigro, A.J. ATP: the red
blood cell link to NO and local control of the pulmonary circulation. Am J Physiol
Heart Circ Physiol 271, H2717-H2722 (1996).

Vanna, S.D.K., J. H. Sorbitol pathway in diabetic and galactosemic rat lens.
Biochim Biophys Acta632-640 (1974).

Morrison, A.D., Clements, R.S. & Travis, S.B. Glucose utilization by the polyol
pathway in human erythrocytes. Biochem Biophys Res Commun 199-205 (1970).

Travis, S.F., D., M.A. & Clements, R.S. Metabolic alterations in the human
erythrocytes produced by increases in glucose concentrations. J Clin Invest 2104-
2112 (1971).

Williamson, J .R. et al. Hyperglycernic pseudohypoxia and diabetic complications.
Diabetes 42, 801-813 (1993).

Barnett, P.A.G., R. G.; Chylack L. T. The effects of oxidation on sorbitol pathway
kinetics. Diabetes, 426-432 (1986).

Sprague, R.S. et al. Effect of L-NAME on pressure-flow relationships in isolated
rabbit lungs: role of red blood cells. Am J Physiol Heart Circ Physiol 269, H1941-
H1948 (1995).

Edwards, J., Sprung, R., Spence, D. & Sprague, R. Chemiluminescence detection
of ATP release from red blood cells upon passage through microbore tubing.
Analyst 126, 1257-1260 (2001).

117

44.

45.

46.

47.

48.

49.

50.

51.

52.

Sprung, R., Sprague, R. & Spence, D. Determination of ATP release from
erythrocytes using microbore tubing as a model of resistance vessels in vivo. Anal
Chem 74, 2274-2278 (2002).

Fischer, D.J., Torrence, N.J., Sprung, R.J. & Spence, D.M. Determination of
erythrocyte deformability and its correlation to cellular ATP release using
microbore tubing with diameters that approximate resistance vessels in vivo.
Analyst 128, 1163-1168 (2003).

Oblak, T.D.A., Root, P. & Spence, D.M. Fluorescence Monitoring of ATP-
Stimulated, Endothelium-Derived Nitric Oxide Production in Channels of a
Poly(dirnethylsiloxane)-Based Microfluidic Device. Anal Chem 78, 3193-3197
(2006)

Carroll, J.S., Ku, C.-J., Karunarathne, W. & Spence, D.M. Red Blood Cell
Stimulation of Platelet Nitric Oxide Production Indicated by Quantitative

Monitoring of the Communication between Cells in the Bloodstream. Anal Chem
79, 5133-5138 (2007).

Paris, A. & Spence, D.M. Measuring the simultaneous effects of hypoxia and
deformation on ATP release from erythrocytes. Analyst 133, 678-682 (2008).

Meyer, J .A., Froelich, J .M., Reid, G.E., Karunarathne, W.K.A. & Spence, D.M.
Metal-activated C-peptide facilitates glucose clearance and the release of a nitric
oxide stimulus via the GLUTl transporter. Diabetologia 51, 175-182 (2008).

Kosower, N.S., Kosower, B.M., Wertheim, B. & Correa, W.S. Diamide, a new
reagent for the intracellular oxidation of glutathione to the disulfide. Biochem
Biophys Res Commun 37, 593-596 (1969).

Shantz, L.M., Talalay, P. & Gordon, G.B. Mechanism of inhibition of growth of
3T3-L1 fibroblasts and their differentiation to adipocytes by

dehydroepiandrosterone and related steroids: role of glucose-6—phosphate
dehydrogenase. Proc Natl Acad Sci U S A 86, 3852-3856 ( 1989).

Gupte, S.A. et al. Pentose phosphate pathway coordinates multiple redox-
controlled relaxing mechanisms in bovine coronary arteries. Am J Physiol Heart
Circ Physi01285, H2316—H2326 (2003).

118

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

Leopold, J .A., Cap, A., Scribner, A.W., Stanton, R.C. & Loscalzo, J. Glucose-6-
phosphate dehydrogenase deficiency promotes endothelial oxidant stress and
decreases endothelial nitric oxide bioavailability. FASEB Journal 15, 1771-1773
(2001)

Yang, N.C., Jeng, K.C., Ho, W.M., Chou, S.J. & Hu, M.L. DHEA inhibits cell
growth and induces apoptosis in BV-2 cells and the effects are inversely

associated with glucose concentration in the medium. J Steroid Biochem Mol Biol
75, 159-166 (2000).

Tomlinson, D.R., Willars, G.B. & Canington, A.L. Aldose reductase inhibitors
and diabetic complications. Pharmacol Ther 54, 151-194 (1992).

Nakarnura, J. et al. Polyol pathway, 2,3-diphosphoglycerate in erythrocytes and
diabetic neuropathy in rats. Eur J Pharmacol 294, 207-214 (1995).

Eaton, R.P. et al. A commentary on 10 years of aldose reductase inhibition for
limited joint mobility in diabetes. J Diabetes Complications 12, 34-38 (1998).

Parry, G.J. Management of diabetic neuropathy. Am J Med 107, 27S-33S (1999).

Suzuki, T. et al. Characterization of polyol pathway in Schwann cells isolated
fiom adult rat sciatic nerves. J Neurosci Res 57, 495-503 (1999).

Cameron, N.E., Cotter, M.A., Basso, M. & Hohman, T.C. Comparison of the
effects of inhibitors of aldose reductase and sorbitol dehydrogenase on
neurovascular function, nerve conduction, and tissue polyol pathway metabolites
in streptozotocin-diabetic rats. Diabetologia 40, 271-281 (1997).

Asano, T., Saito, Y., Kawakarni, M. & Yamada, N. Fidarestat (SNK-860), a
potent aldose reductase inhibitor, normalizes the elevated sorbitol accumulation in
erythrocytes of diabetic patients. J Diabetes Complications 16, 133-138 (2002).

Hamada, Y.N., Jiro Clinical Potential of Aldose Reductase Inhibitors in Diabetic
Neuropathy. Treat Endocrinol 3, 245-255 (2004).

Rakowitz, D., Gmeiner, A. & Matuszczak, B. Discovery of Novel Aldose
Reductase Inhibitors Characterized by an Alkoxy-Substituted Phenylacetic Acid
Core. Archiv der Pharmazie 339, 559-563 (2006).

119

65.

66.

67.

68.

69.

70.

71.

72.

Vinson, J.A., Staretz, M.E., Bose, P., Kassm, H.M. & Basalyga, 8.8. In vitro and
in vivo reduction of erythrocyte sorbitol by ascorbic acid. Diabetes 38, 1036-1041
(1989)

Wang, H., Zhang, Z.-b., Wen, R.-r. & Chen, J .-w. Experimental and clinical
studies on the reduction of erythrocyte sorbitol-glucose ratios by ascorbic acid in
diabetes mellitus. Diabetes Res Clin Pract 28, 1-8 (1995).

Vincent, T.E., Mendiratta, S. & May, J.M. Inhibition of aldose reductase in
human erythrocytes by vitamin C. Diabetes Res Clin Pract 43, 1-8 (1999).

Anaja, H.P. Diagnostic performance of red cell sorbitol assay in a Nigerian
teaching hospital. Clinica Chimica Acta 262, 1-11 (1997).

Shinohara, R., Ohta, Y., Yarnauchi, M. & lshiguro, 1. Improved fluorometric
enzymatic sorbitol assay in human blood. Clinica Chimica Acta 273, 171-184
(1998)

Umeda, M. et al. Determination of D-sorbitol in human erythrocytes by an
enzymatic fluorometric method with an improved deproteinization procedure.
Ann Clin Biochem 38, 701-707 (2001).

Cereser, C. et a1. Quantitation of reduced and total glutathione at the femtomole
level by hi gh-perforrnance liquid chromatography with fluorescence detection:
application to red blood cells and cultured fibroblasts. J Chromatogr B Biomed
Sci Appl 752, 123-132 (2001).

Hood, L. & Perlmutter, RM. The impact of systems approaches on biological
problems in drug discovery. Nat Biotechnol 22, 1215-1217 (2004).

Sprague, R.S., Ellsworth, M.L., Stephenson, A.H., Kleinhenz, M.E. & Lonigro,
A.J. Deformation-induced ATP release from red blood cells requires CFTR
activity. Am J Physiol Heart Circ Physiol 275, Hl726-1732 (1998).

120

CHAPTER 4: ADDRESSING ENDOTHELIUM WITH BLOOD COMPONENTS

4.1 MULTIPLE CELL TYPES WITH THE uFA: CELL-CELL INTERACTIONS

The uFA has been employed as an in vitro model of the in vivo circulation system
by addressing an array of cultured bovine pulmonary artery endothelial cells (bPAECs)
with underlying microfluidic channels containing various RBC samples.1 This device has
been used to demonstrate the ability to monitor cellular communication between multiple
cell types to help better understand the various mechanisms of vasodilation.

Many technologies, such as those designed to perform ADMET screening, are
beginning to emerge as significant tools in drug discovery2 and systems biology3 and
motivation for continued advancements in this area is high.4 As discussed in detail in
chapter 2, the most logical way to improve the drug discovery process at the preclinical
stage is to create technologies that incorporate controlled, in vitro designs that mimic the
in vivo environment, allowing for the determination of molecular-mediated
communication between different cell types. Importantly, there is a necessity to
incorporate components of the circulation, specifically the aspect of flow-induced shear
stress, to accurately mimic the system as it exists in vivo.

There have been many reports involving microfluidics or lab-on-a-chip
technology to create in vivo mimics using controlled, in vitro platforms and many of
these devices incorporate cellular components.5'8 In vitro platforms have integrated
various cell types including a dopaminergic cell model (PC12 cells),9 endothelial cells,10
and isletsll contained within microfluidic devices. Although previous reports have

described the successful direct immobilization of endothelial cells in the channels of a

121

 

microfluidic device,10'12 the uFA first described in chapter 2, incorporates a track-etched
polycarbonate (TEPC) membrane to separate the flow of RBCS from the bPAECs as
illustrated in figure 4.1. The integration of the membrane allows for easily accessible
wells for the culture of an endothelial cell layer with a future prospect of integrating other
cell types (such as PC12 cells) as well. Thus, creating a working mimic of the blood
brain barrier in a three dimensional device, enabling the communication between the
circulation (in particular, RBCS) and the nervous system. In this construct, our device
enables communication among tissues, not just cells.

This uFA serves as a precursor for the implementation of a microfluidic flow-
based system, with already available microtitre plate technologies, to elucidate the
communication between RBCS and bPAECs. The channels were fabricated such that the
internal diameter closely approximates those of resistance vessels in vivo as RBCS were
pumped through these channels. Moreover, this device was used to monitor ATP release
from RBCS that were incubated in the presence or absence of iloprost (a substrate known

to stimulate ATP release via a G-protein coupled receptor mechanism).13

4.2 ILOPROST: PROSTACYCLIN RECEPTOR-MEDIATED ATP RELEASE

The mechanism by which iloprost induces ATP release from RBCS was first
proposed by Olearczyk et al. '4 and has been adapted for figure 4.2. Here, the binding of
iloprost to the prostacyclin (PGIz) receptor (IPR)”18 leads to the cascade of intracellular

l9-21

events within the proposed signal transduction pathway, as also observed in the

mechanical deformation-induced ATP release from RBCS, by way of G5 activation.

122

 

 

 

Blank Master 6

 

 

 

 

 

 

 

— Cultured bPAECs
TEPC Membrane —* :

 

 

 

 

 

 

Channel Master —* - f RBCS

 

 

 

Figure 4.1 Cross-sectional illustration of the uFA used to monitor the cellular
communication between RBCs transversing through the underlying microfluidic
channels with the cultured bPAECs within the array-like wells. Here, RBC-derived
ATP release is able to diffuse to the bPAECs though the TEPC membrane,
stimulating the production of intracellular NO which is imaged using the
fluorescence probe DAF-F M DA.

123

 

Mechanical
Deformation

Iloprost l

1

ATP

    

Figure 4.2 Illustration of the mechanism by which iloprost stimulates the
release of ATP from the RBC through the signal transduction pathway first
proposed by Olearczyk et al. by binding to the prostacyclin receptor (IPR)
which activates the g-protein (Gs) cascading events.

124

 

Iloprost is a therapeutic agent prescribed to increase vasodilation and reduce
hypertension, as a treatment for cardiovascular and circulatory disorders.22 It is a second
generation structural analog of PG122324 that is more chemically stable and demonstrates
oral effectiveness for the decrease of peripheral resistance and pulmonary arterial
pressure. Our group has previously shown that incubation of a RBC sample with a 1 uM
solution of iloprost for approximately 30 min will increase the amount of RBC-derived
ATP release in both a flow and non-flow based system (this significance will be
discussed later in detail).25 Previously, the ATP release from RBCS was measured
utilizing microbore tubing placed over a photomultiplier tube (PMT) housed within a
light excluding box. More recently, as presented here, a microfluidic system has replaced
the microbore tubing,l where it is used to mimic the resistance vessels and allow for on-

chip mixing.

4.3 EXPERIMENTAL

Using the uF A, and incorporating multiple cell types, the additional ATP release
from RBCS in the presence of iloprost and its subsequent effect on cultured bPAECs is
demonstrated, as it pertains to the mechanism of vasodilation that occurs in vivo.
Specifically, RBCS incubated with and without iloprost were pumped through the
underlying microfluidic channels of the uFA in order to allow for the cellular
communication between these RBCS and cultured bPAECs within the array-like wells via
ATP, stimulating the intracellular production of NO. Using fluorescence spectroscopy,
the intracellular NO production within the bPAECs contained in individual wells is

determined.

125

4.3.1 DETERMINING ILOPROST-STIMULATED ATP RELEASE

In order to correctly monitor and determine the cellular communication between
the RBCS and bPAECs within the uF A, and demonstrate that the intracellular
endothelium-derived NO production is, in fact, due to an increased amount of
pharmaceutically-induced ATP release from RBCS, the system must first be used to
quantify ATP release. Illustrated in figure 4.3, a t-channel microfluidic device was used
to validate that an increase in ATP release from RBCs occurs when these cells are
incubated with iloprost.

Similar to the fabrication of the uF A, two layers of PDMS were used to fabricate
this t-charmel microfluidic device. One layer, comprised of 5:1 (elastomer base to curing
agent), was obtained using a t-channel master wafer, and the other, comprised of 20:1,
was a blank PDMS slab without any features. Each layer of PDMS was obtained in the
same fashion as for the uFA, mixing PDMS base and curing agent, degassing the
monomer, and partially baking at 75 °C for 20 min. The channel layer had inlet holes
punctured through the chip using a 20 gauge luer stub adapter. A single waste port was
also created using a 1/8” hollow punch. The device was completed by baking the two
layers together at 75 °C for an additional 30 min, resulting in an irreversible seal between
each layer.

To determine the concentration of ATP released from the RBC samples, a syringe
pump is connected to the device via hypodermic stainless steel tubing for the introduction
of flow into each inlet of the t-channel microfluidic device. The syringes deliver ATP

standards ranging from 0.25-1.0 uM as well as various RBCS samples, in the presence

126

 

Dark Box ' """""""" ;

L/L

Standards or
RBC Samples

 

Figure 4.3 Schematic drawing of the experimental setup used, incorporating a t-channel
microfluidic device with the methodology first employed by our group, to determine RBC-
derived ATP release in a continuous flow system. Here the luciferin and luciferase
enzymatic assay was employed, monitoring the chemiluminescence produced that is
directly proportional to the concentration of ATP released from various samples.

127

and absence of an ATP release stimulus (iloprost) or an ATP release inhibitor
(glybenclamide), through the underlying microfluidic channels. Glybenclamide is
specific for inhibiting CFTR, which is involved in the signal transduction pathway
responsible for the release of ATP from RBCS.25 The ATP standard solutions were
prepared from a 100 uM stock solution (27.6 mg in 500 mL DDW) in PSS. The
luciferin/luciferase mixture used to measure chemiluminescence produced as a function
of the ATP within the solutions was prepared by adding 5 mL of DDW to a vial
containing luciferase and luciferin (FLE-SO, Sigma). In order to enhance the sensitivity
of the assay, 2 mg of luciferin were added to the vial.

The ATP standard solutions and the various RBC sample solutions are each
pumped through one inlet, intersecting with the luciferin/luciferase that is pumped
through the second inlet of the t—charmel microfluidic device. The entire device was
placed over a photo multiplier tube (PMT) to detect the chemiluminescence emission26
that is produced downstream from the t-channel intersection. The current collected by
the PMT from the chemiluminescence produced, is converted to voltage, measured in real
time, and plotted at a rate of 10 measurements per second for 30 seconds onto a computer
using LabView software (National Instruments, Austin, TX) with a program that was

written in-house.

4.3.2 ENDOTHELIUM CULTURE ON THE uFA
In order to monitor the cellular communication between RBCS flowing through
the rrricrofluidic channels and the bPAECs cultured onto the membrane within the array

like wells, a bPAEC cell suspension solution must first be prepared. Here, traditional cell

128

 

culture techniques are used to obtain a confluent monolayer within tissue culture flasks
from which cells are harvested to prepare a cell suspension that can be pipetted into each
of the wells within the uFA. As shown in figure 4.4 this cell solution is obtained by first I
washing a T-75 cell culture flask with 7 mL HEPES (4-(2-hydroxyethyl)-1-
piperazineethanesulfonic acid, pH 7.4) to add buffering capacity to protect cell receptors
during harvesting. This solution is aspirated off prior to the addition of 5 mL
trypsin/EDTA solution (containing 0.5 g/L of trypsin (1:250) and 0.2 g/L of EDTA°4Na
in Hanks' Balanced Salt Solution), which will cleave the peptide chains that are involved
in cell attachment to the flask. The trypsin is allowed to sit on the cells for approximately
3 min, followed by the addition of 9 mL of trypsin neutralizing solution (sterile,
phosphate-buffered saline solution (1x) containing calf serum as a trypSin inhibitor) is
added to the flasks, halting the action of the trypsin.

Vigorous pipetting of the solution within the flask (trypsin and the neutralizing
solution) is used to ensure the cells have become detached from the flask and are
suspended in the solution. This suspension is centrifuged at 2200 rpm for 5 min at 25 °C,
the supernatant is aspirated off and the remaining pellet of cells is resuspended in 1 mL
endothelial growth media, (Clonetics® EGM® MV Microvascular Endothelial Cell
Growth Medium with 5% fetal bovine serum) that has been equilibrated at 37°C and 5%
C02, with repeated pipetting to ensure a homogeneous cell solution. This cell solution
has been determined to contain on average approximately 1.0 x 108 cells/mL.

Prior to the introduction of the suspended cell solution into the wells of the uFA,
10 uL of a 100 ug/mL fibronectin solution is pipetted into each well, providing a

fibronectin coating to promote cell adhesion to the membrane. The fibronectin solution is

129

 

828—8 :8 855380: a 88% 8 wfifiommm @8888 :85 838 3:89:83 8 Baggage fl 8on Humans
05 £53883 05 mo 3682 08 com: .00 mm we 88 m .8.“ 89 as we voméfiaoo 3 828—8 mac. @8838 880w?
:85 $25 cows—cm 38:88:08 838 .3 330:8 828—8 <HQm<8m§8 53> 838858 ,3 via: 05 89a @0688
08“.. £3 2:. .2330 :3 wage 5688 watchman 33cm 05 8m :88 08833550088838-793880.83
-mv-$ mmmm: 83> @8298 .8838 £38m 335096 05 go 3688 ”uofiwznwfi 0.8 30% >8— 803 {mi 08
88MB mM,—Dada mo 8315 05 8% 828385 :3 a 8308 9 bozo—Q80 @0808 05 no 808885 o§8onom 1v 2:»:—

 

 

 

 

 

 

 

8528 :8 80328
msoocowo8om wowegcoo mmwnmwmu Home—0:08 802.800
\JW wum<nw£ Mme ) H. m0m<nHD GOO .. , .. int... .. .fifimuifluunuaohuauw
53%.... i... x
1 1 mze \

Ba 58? a MC,

uovcommsmoe £00

130

prepared from a 5 mg vial of lyophilized powder that is diluted to obtain a 1000 ug/mL
stock solution in DDW. This stock solution is stored as individual aliquots prior to
diluting to the working solution in phosphate buffered saline (PBS). The PBS (within the
fibronectin solution that is contained within each well) is evaporated by placing the
device in a standard cell culture hood, providing increase air flow over the device.
Subsequent serilization is performed by exposure to UV light for 15 min within the same
cell culture hood.

The cell suspension described above is then pipetted into each well and allowed to
incubate for one hour in a cell incubator at 37 °C and 5% CO2 to promote the adhesion of
the cells to the surface. At which point, the media is changed by pipetting off the
solution of each well and replacing it with media that has equilibrated to 37°C and 5%
CO2, taking care not to disturb the adhered cells. The cells are returned to the incubator

for an additional two hrs, allowing the cells time to become confluent on the membrane.

4.3.3 FLUORESCENCE DETERMINATION OF INTRACELLULAR NO

Once the cells have become confluent within the wells of the uFA, they are then
ready to be used for the determination of intracellular NO production as various stimuli
are delivered through the underlying microfluidic channels. First, to ensure that the cells
have sufficient substrate for the production of NO by NOS (i.e., that there is no limiting
reactant), an equilibrated solution of 5 mM L-arginine (0.0044 g in 5 mL Hank’s
Balanced Salt Solution (containing in mM, 1.26 CaCl2, 0.493 MgCl2-6H2O, 0.407
MgSO4-7H2O, 5.33 KCl, 0.44] KH2PO4, 4.17 NaHCO3, 137.93 NaCl, 0.338 Na2HPO4,

and 0.556 Dextrose with a final pH of 7.4)) is first incubated within each well at 37°C

131

 

and 5% CO2 for 30 min. The solution is then removed and replaced with 100 uM DAF-
FM DA (4-arnino-5-methylamino-2’,7’-difluorofluoroescein diacetate), an intracellular
fluorescence probe for NO with excitation and emission wavelengths of 495 nm and 515
nm, respectively. DAF-FM DA, as illustrated in figure 4.5 is able to permeate the cell,
after which diacetate groups are cleaved by an esterase, trapping the charged DAF-FM
fluorescence probe within the cell. Upon reaction with NO, the DAF-FM forms the
benzotriazole derivative that is highly fluorescent. The DAF-FM DA was prepared as a 5
mM stock by dissolving 1 mg within 20 uL anhydrous DMSO and making a subsequent
dilution to 100 uM in HBSS, which was equilibrated to 37°C before incubation with the
bPAECs within each well. The uF A was placed in the incubator at 37°C and 5% CO2 for
30 min, at which point the DAF-FM DA is subsequently removed from the wells and
replaced with equilibrated HBSS prior to fluorescence imaging.

This device is then used to monitor the production of intracellular NO within the
bPAECs cultured in the wells above, as ATP diffuses from various RBC solutions that
are pumped through the underlying microfluidic channels for 30 nrin. An Olympus D(71
inverted fluorescence microscope (Center Valley, PA) fitted with a mercury-arc lamp and
a fluorescein isotlriocyanate (FITC) filter cube (providing excitation and emission
wavelengths of 470 nm and 525 nm, respectively) was used to image each well. The
images were obtained at 20x magnification and shown in figure 4.6. After capturing the
fluorescence images of the array, the pixel intensities were measured for each individual
well, which is directly dependant on the concentration of NO produced within the

bPAECs of each well.

132

 

     

DAF-FM Diacetate
Cell Membrane

Benzotriazole Derivative

Figure 4.5 Mechanism detailing the process by which 4-amino-5-methylarnino-2’7’-
difluorofluoroescein (DAF-FM) diacetate enters the cell, whereby allowing for the
intracellular detection of NO production. Here, the cleavage of the diacetate groups
fiom DAF-FM DA occurs by an esterase as the molecule enters the cell, further
reaction of DAF-FM with NO forms the fluorescent benzotriazole derivate with
excitation and emission wavelengths of 495 nm and 515 nm, respectively.

133

Microfluidic Array

 
   
  

3:41.

Dichroic Camera
‘— Hg-Arc Lamp

Data Analysis Fluorescence Microscope

Figure 4.6 Fluorescence microscopy experimental setup used to monitor the
intracellular production of NO within cultured bPAECs contained within the
wells of the uFA as a result from ATP released from the RBCS transvering
through the underlying microfluidic channels in this flow-based system. Here,
the mercury arc (Hg-Arc) lamp provides the light that is selectively filtered by a
FITC filter cube for the excitation of the benzotrazole derivative of DAF-FM and
NO which is then imaged using a dichroic camera and Microsuite computer
software.

134

To demonstrate the potential of such a multicellular device to mimic in vivo
stimulation of NO by RBC-derived ATP, RBCs were incubated in the presence and
absence of iloprost. A 100 uM iloprost solution was prepared by diluting 36.1 uL of the
13.87 mM stock solution (Cayman Chemical, Ann Arbor, M1) to 5 mL in PBS which is
then diluted to a 1 uM working solution in PBS. Here, as illustrated in figure 4.7, two
RBC samples are prepared with and without iloprost to determine the fluorescence
intensity increase within the cultured bPAECs, as a function of NO production, due to an
increased RBC-derived ATP release. Both RBC samples were prepared fiom the washed
RBCS as 7% hematocrit solutions, in one of these samples, a final concentration of 1 uM
iloprost was present in the PSS prior to the addition of RBCS. These samples were
allowed to incubate at room temperature for 30 min prior to pumping through the
underlying microfluidic channels of the uFA that contained bPAECs.

In order to effectively demonstrate that the increase in intracellular NO is in fact
due to the increased ATP release from the RBCS flowed through the underlying
microfluidic channels, and not due to cell lysis or the action of iloprost on the bPAECs
themselves, various RBC samples and controls were pumped through the underlying
channels. These solutions, as illustrated in figure 4.8 consisted of RBCS at 7%
hematocrit (row 3), RBCS incubated with 1 uM iloprost (row 4), RBCs incubated with 1
mM glybenclamide (row 5), RBCS incubated with 1 mM glybenclamide and 1 uM
iloprost (row 6), and two controls (buffer (row 1) and 1 uM iloprost (row 2)) were
pumped through the underlying microfluidic channels. The 10 mM glybenclamide stock
solution was prepared by combining 0.049 g with 2 mL 0.1 M NaOH, and 7.94 mL

dextrose solution (1 g dextrose in 20 mL DDW) and heating to 52 °C until dissolved with

135

 

TEPC Membrane

RBC Samples

 

    

 

150 1.1M Channel
(bottom layer)
Array—Like Wells

(top layer)

 

Waste

Figure 4.7

(pore d = 1 pm)

 

l
.—

Initial microfluidic array used to monitor the fluorescence intensity

corresponding to the intracellular production of NO as RBCS with and without the
incubation of 1 uM iloprost are pumped through the underlying microfluidic channels.

136

B

 

—L

PBS buffer
1 uMiloprost 2

7% RBCS

0.)

J:-

7°oRBCs 8; 1 uM iloprost

U1

7% RBCS 8; l 1111\lglybenclan1ide

790RBCS & l 1111\Ig1ybenclamide
w.” 1 uM iloprost

0')

 

Figure 4.8 Fluorescence images of the six channel uFA used for the
determination of intracellular NO resulting from the ATP released from various
RBC and control samples. Specifically, columns A and B contain cultured
bPAECs with and without the presence of intracellular DAF—FM DA,
respectively, while column C contains cell-free wells.

137

stirring. This solution was added to the PSS to create the 7% RBC solutions flowed
through channels five and six.

The 18 wells that comprised the device were prepared in the absence (column C)
and presence (columns A and B) of bPAECs, where only the cells in column A were
incubated with the DAF-FM DA probe for the intracellular fluorescence determination of
NO production. After flowing the control and RBC samples through the underlying
microfluidic channels of the uFA for 30 min, the fluorescence images corresponding to
each well were taken and the increase in intensity (corresponding to the NO production
by bPAECs) were analyzed. The wells in column B did not contain the DAF-FM DA
fluorescence probe in order to eliminate the possibility of any type of fluorescence signal
that might be due to the cells alone. Additionally, the wells comprising column C were
absent of cells in order to demonstrate that no fluorescence was observed from the

membrane alone.

4.4 RESULTS AND DISCUSSION: MULTICELLULAR COMMUNICATION
Initially, iloprost was shown to increase the RBC-derived ATP release using a
flow-based microfluidic device and the luciferin/luciferase chemiluminescence detection
method. Figure 4.9 shows the raw data that was collected for the real-time determination
of ATP within a control 1 uM iloprost solution, and 7% RBC solution with and without
the addition of 1 uM iloprost and 250 nM glybenclamide. Iloprost alone flowing through
the microfluidic device resulted in an errrission signal that is equilivent to that of buffer
alone mixing with the luciferin/luciferase mixture. In the presence of a 7% hematocrit of

RBCS, the measured ATP value was 0.27 :1: 0.04 uM. However, as expected, RBCS

138

 

(107-

W 7% RBCS w/ luM Ilop.

(106 a

j 7% RBCs & 250nM Gly.
0.05 W W/ 111M HOP-

0'04 ‘W 1 PM HOP-

Chemiluminescence Intensity

 

 

0.03 . . . - . . . . .8 B. . a B, a
0 10 20 30
Time(sec)

Figure 4.9 Raw data obtained for the real-time analysis of the chemiluminesence
produced as a function of ATP concentration from 1 uM iloprost, 7% RBCS, 7% RBCS
with 250 nM glybenclamide and 1 uM iloprost, and 7% RBCS with 1 uM iloprost by

placing the microfluidic t-channel chip on a PMT housed within a dark box.

139

 

incubated in 1 uM iloprost released ATP at a concentration of 0.95 :t 0.18 uM, a value
that dropped to 0.29 i 0.14 M in the presence of a sample containing 250 nm
glybenclamide, an inhibitor of ATP release from RBCS. As illustrated in figure 4.10, if
the signal was due to cellular lysis upon the addition of iloprost, glybenclamide would
not be able to reduce the RBC-derived ATP chemiluminescence produced.

Figure 4.11 represents the fluorescence intensity resulting from the intracellular
NO production in response to pumping a 7% RBC sample, in the presence and absence of
1 uM iloprost, through the underlying channels for 30 min. The increases in fluorescence
intensity due to endothelium-derived NO production were determined to be 23.0 i 3.5
and 16.5 :1: 1.0 where the values reported are the average intensity and standard deviation
(n = 3 wells for each sample). The difference in fluorescence intensities between samples
is statistically different (p < 0.005).

The data for the final demonstration of iloprost-induced RBC-derived ATP
release resulting in NO production within the cultured bPAECs are shown in figure 4.12.
The results for column A for each of the six channels containing PBS bufier, 1 uM
iloprost, 7% RBCS, RBCs with 1 uM iloprost, RBCS with 1 mM glybenclamide, and
RBCS with both 1 mM glybenclamide and 1 uM iloprost are compiled. The fluorescence
intensities are presented as the average pixel intensity of the five integrated areas of each
acquired image. Note the significant increase in fluorescence intensity due to
endothelium-derived NO in the presence of the RBCS (A3) and the RBCS incubated with
iloprost (A4). The difference in fluorescence intensity between samples containing the
RBCS in the absence (A3) and presence (A4) of iloprost are statistically different from

the other rows. There is also a difference in the intensities of rows 3 and 4 (p = 0.005).

140

 

 

 

  

 

n = 5
* p < 0.001

1.0 -
E
3
g 0.8 -
m
2
6'2
CI. 0.6 -
F.
<
8,
(I! 0.4 '
9
<

0.2 -

0.0 - . ..

1 uM Iloprost 7% RBCs 7% RBCS w/ 7% RBCS & 250 nM Gly.
1 uM Iloprost w/ 1 uM Iloprost

Figure 4.10 Data representation for the quantitative determination of ATP
released from 7% RBCS in the presence and absence of 1 uM iloprost and 250
nM glybenclamide using the t-channel microfluidic chip and the luciferin and
luciferase based enzymatic assay. A control sample of 1 uM iloprost alone
without the presence of RBCS is also plotted, demonstrating the background
chemiluminescence. Average concentrations of ATP in micromolar amounts are
reported where the error is represented by the SEM. The reduction in ATP
release due to the addition of glybenclamide (Gly.) is statistically

141

 

Fluorescence intensity [a.u.]

 

 

 

 

 

 

RBCS RBCS + Iloprost

Figure 4.11 Representative data for the initial determination of
intracellular NO production within bPAECs that are cultured
within the wells of a two channel uFA through which a 7% RBC
sample with and without the incubation of a 1 uM iloprost solution
were flowed. The values reported are the average intensity and
SEM (n = 3 wells for each sample). The difference in fluorescence

intensity between samples incubated in the presence and absence
of iloprost are statistically different (p < 0.005)

142

8

‘8‘

   

Fluorescence Intensity
°s°

§

   

0

A1 A2 A3 A4 A5 A6
Figure 4.12 Data representing the intracellular NO production within
bPAECs cultured in the wells of a six channel uFA, here samples were
flowed through the underlying microfluidic channels for 30 min prior to
obtaining fluorescence images. The samples included PBS buffer, 1 uM
iloprost, 7% RBCS, RBCS with 1 uM iloprost, RBCS with 1 mM
glybenclamide, and RBCs with both 1 mM glybenclamide and 1 uM
iloprost (Al-A6 respectively). The difference in fluorescence intensity
between samples containing the RBCS in the absence (A3) and presence
(A4) of iloprost are statistically different from the other rows. There is also
a difference in the intensities of rows 3 and 4 (p = 0.005).

143

These data suggest that the efficacy of iloprost occurs only through the interaction with
the RBCS.

Such data would not have been found in classic in vitro studies containing only

the iloprost and endothelial cells alone, as the mechanism of action is RBC mediated,l4’27

thus verifying the potential usefulness of the device containing underlying microfluidic
channels that incorporate blood flow. Irnportantly, the device wells are addressable by a
continuous flow of RBCS so that assays for drug action, drug inhibition, and control
studies can be performed simultaneously using the same RBC source. In the future, the
throughput of this approach could be increased by employing a device having a larger
array of wells that are measured using a plate reader. As described in chapter 2, it is
anticipated that the advances developed with this three dimensional device will
successfully bridge results obtained with traditional in vitro studies to those in the in vivo
state of drug discovery, thereby reducing the number of trials that fail at the later stage of

clinical trials.

144

 

10.

REFERENCES

Genes, L.I., Tolan, N.V., Hulvey, M.K., Martin, R.S. & Spence, D.M. Addressing
a vascular endothelium array with blood components using underlying
microfluidic channels. Lab Chip 7, 1256-1259 (2007).

Dittrich, P.S. & Manz, A. Lab-on-a-chip: microfluidics in drug discovery. Nat
Rev Drug Discov 5, 210-218 (2006).

Breslauer, D.N.L., P. J.; Lee, L. P. Microfluidics-based systems biology. Mol
BioSyst 2, 97-112 (2006).

Cohen, C.M. A path to improved pharmaceutical productivity. Nat Rev Drug
Discov 2, 751-753 (2003).

Khademhosseirri, A. et al. Cell docking inside microwells within reversibly sealed
microfluidic channels for fabricating multiphenotype cell arrays. Lab Chip 5,
1380-1386 (2005).

Shin, M. et a]. Endothelialized Networks with a Vascular Geometry in
Microfabricated Poly(dimethyl siloxane). Biomed Microdevices 6, 269-278
(2004).

Takayama, S. et al. Laminar flows: Subcellular positioning of small molecules.
Nature 411, 1016-1016 (2001).

Aaron Sin et al. The Design and Fabrication of Three-Chamber Microscale Cell
Culture Analog Devices with Integrated Dissolved Oxygen Sensors. Biotechnol
Prog 20, 338-345 (2004).

Li, M.W., Spence, D.M. & Martin, R.S. A microchip-based system for
immobilizing PC 12 cells and arnperometrically detecting catecholamines
released after stimulation with calcium. Electroanalysis 17, 1171-1180 (2005).

Spence, D.M., Torrence, N.J., Kovarik, M.L. & Martin, R.S. Amperometric
determination of nitric oxide derived from pulmonary artery endothelial cells
immobilized in a microchip channel. Analyst 129, 995-1000 (2004).

145

 

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

Roper, M.G., Shackrnan, J.G., Dahlgren, G.M. & Kennedy, R.T. Microfluidic
Chip for Continuous Monitoring of Hormone Secretion from Live Cells Using an
Electrophoresis-Based Immunoassay. Anal Chem’75, 4711-4717 (2003).

Oblak, T.D.A., Root, P. & Spence, D.M. Fluorescence Monitoring of ATP-
Stimulated, Endothelium-Derived Nitric Oxide Production in Channels of a
Poly(dimethylsiloxane)-Based Microfluidic Device. Anal Chem 78, 3193-3197
(2006)

Olearczyk, J .J ., Ellsworth, M.L., Stephenson, A.H., Lonigro, A.J. & Sprague, R.S.
Nitric Oxide Inhibits ATP Release from Erythrocytes. J Pharmacol Exp Ther 309,
1079-1084 (2004).

Olearczyk, J.J., Stephenson, A.H., Lonigro, A.J. & Sprague, R.S. Receptor-
mediated activation of the heterotrimeric G-protein Gs results in ATP release
fi'om erythrocytes. Med Sci Monit 7, 669-674 (2001).

Charles Willems, Herbert V. Stel, Willem G. van Aken & Jan A. van Mourik
Binding and inactivation of prostacyclin (PGI<sub>2</sub>) by human
erythrocytes. Br J Haematol 54, 43-52 (1983).

Dutta-Roy, A.K. & Sinha, A.K. Binding of prostaglandin E1 to human
erythrocyte membrane. Biochim. Biophys. Acta, Biomembr. 812, 671-678 (1985).

Sprague, R.S. et al. Prostacyclin Analogs Stimulate Receptor-Mediated cAMP
Synthesis and ATP Release from Rabbit and Human Erythrocytes.
Microcirculation 15, 461 - 471 (2008).

Adderley, S.P. et al. Iloprost- and isoproterenol-induced increases in cAMP are
regulated by different phosphodiesterases in erythrocytes of both rabbits and
humans. Am J Physiol Heart Circ Physiol 296, H161 7-1624 (2009).

Davies, P.F. Flow-mediated endothelial mechanotransduction. Physiol Rev 75,
519-560 (1995).

Sprague, R.S., Stephenson, A.H., Bowles, E.A., Stumpf, M.S. & Lonigro, A.J.
Reduced expression of Gi in erythrocytes of humans with type 2 diabetes is
associated with impairment of both CAMP generation and ATP release. Diabetes
55, 3588-3593 (2006).

146

 

21.

22.

23.

24.

25.

26.

27.

Sprague, R.S., Stephenson, A.H. & Ellsworth, M.L. Red not dead: signaling in
and fiom erythrocytes. Trends Endocrinol Metab 18, 350-355 (2007).

Skuballa, W. & Vorbrueggen, H. Synthesis of ciloprost (ZK 36 374): a chemically
stable and biologically potent prostacyclin analog. Adv Prostaglandin
Thromboxane Leukot Res 11, 299-305 (1983).

Stoll, F. et a1. Pharrnacophore definition and three-dimensional quantitative
structure-activity relationship study on structurally diverse prostacyclin receptor
agonists. Mol Pharmacol 62, 1 103-11 11 (2002).

Ruan, K.-H. & Dogrre, J .-M. Implications of the molecular basis of prostacyclin
biosynthesis and signaling in pharmaceutical designs. Curr Pharm Des 12, 925-
941 (2006).

Carroll, J.S., Ku, C.-J., Karunarathne, W. & Spence, D.M. Red Blood Cell
Stimulation of Platelet Nitric Oxide Production Indicated by Quantitative

Monitoring of the Communication between Cells in the Bloodstream. Anal Chem
79, 5133-5138 (2007).

Price, A.K., Fischer, D.J., Martin, R.S. & Spence, D.M. Deformation-Induced
Release of ATP from Erythrocytes in a Poly(dimethylsiloxane)-Based Microchip
with Channels That Mimic Resistance Vessels. Anal Chem 76, 4849-4855 (2004).

Olearczyk, J .J ., Stephenson, A.H., Lonigro, A.J. & Sprague, R.S. Heterotrimeric
G protein Gi is involved in a signal transduction pathway for ATP release from
erythrocytes. Am J Physiol Heart Circ Physiol 286, H940-945 (2004).

147

 

CHAPTER 5: FUTURE DIRECTION AND APPLICATIONS

5.1 uFA: FUTURE APPLICATIONS

Another objective of this work was to demonstrate that hypoxia-induced release
of ATP was capable of stimulating endothelium-derived NO using this uFA. Recent
studies in our group have demonstrated that hypoxia results in, what presents itself as,
mechanical deformation-induced ATP release from RBCS.l Hemoglobin, a
metalloprotein, is attached to the cellular membrane of the RBC and upon conformational
change from oxygenated (R-state) to dexoygenated (T-state) will elicit ATP release in
both a flow and non-flow system. Hypoxia-induced ATP release from the RBC can be
inhibited with prior stiffening of the cell membrane by incubation with diamide, a known
cell stiffener2 as well as with the incubation of glybenclamide, which has been shown to
inhibit CFTR.3 This suggests that the mechanism by which hypoxia induces RBC-
derived ATP release must be of a similar mechanism involving the mechanical
deformation-induced signal transduction pathway previously discussed.4’12

Separately, previous studies in our group have found that there is an increase in
the concentration of ATP release from RBCS of both diabetics and, to a greater extent,
healthy controls, when incubated with metal activated C-peptide. 13 For years C-peptide
was presumed to have little to no biological activity14 which has been thought to be due
to the lack of metal activation. Thus, the ability of metal-activated C-peptide to restore
the ATP release fi'om RBCS obtained from type 2 diabetics, to the normal levels

associated with RBC from healthy controls, demonstrates the physiologically-relevant

prospect of improving blood flow and reducing the complications associated with

148

 

 

 

diabetes. Here, it was also demonstrated that C-peptide will also increase the glucose
uptake into RBCS through GLUTl, the primary glucose transporter in the cell, and is
thought to enhance glycolysis which is the primary route of ATP production.

One of the main advantages to incorporating flow into microtitre plate technology
is the ability to more realistically mimic the circulation system found in vivo and simulate
the shear-forces that RBCS are subjected to. Demonstration of the ability of hypoxia- and
C-peptide-induced RBC-derived ATP release to stimulate the intracellular production of
NO within the bPAECs cultured within the array-like wells of the uFA is presented here.
As previously discussed, this NO production plays a physiologically-relevant role in the
process of vasodilation of the smooth muscle surrounding the vascular system,4'l°’”’15'22
and ultimately, the ability of the circulation system to deliver 02 to deprived tissue.

Another application for this uFA is the ability to monitor the mechanism by
which pentoxyfilline or trental, in contrast to iloprost, requires the shear aspect of flow to
elicit RBC-derived ATP release.3 Trental has been demonstrated to reduce the
complications associated with nephropathy in diabetic patients and is thought to serve as
an anti-inflammatory agent23'26 by either suppressing oxygen radial formation, preventing
lipid peroxidation, or scavenging radicals that are formed. The combination of trental
and iloprost on the effect of RBC-derived ATP release in a non-flow system is

demonstrated, lending to future work possible in the evaluation of the multicellular effect

of stimulating endothelium NO and ultimately vasodilation.

5.2 EXPERIMENTAL

149

 

Three different applications of the uFA, specifically the application for cellular
communication between RBCs and bPAECs cultured within the array-like wells by
monitoring ATP-derived intracellular production as it pertains to vasodilation is
presented. Also, preliminary work illustrating the ability to determine the mechanisms of
action of two therapeutics, trental and iloprost are also discussed demonstrating that
without employing accurate in vitro technologies trental could be mistakenly determined

ineffective.

5.2.1 HYPOXIA-INDUCED ATP RELEASE: EFFECT ON NO

Employing the uF A, the effect of hypoxia-induced RBC-derived ATP release on
the intracellular production of NO within the bPAECs cultured in the array-like wells is
determined using fluorescence microscopy. The samples investigated include 3.5%
RBCs, RBC incubated in hypoxic buffer, and RBCS initially incubated with 100 and 200
uM diamide and 20 and 40 uM diamide separately, prior to incubation in the hypoxic
buffer. In order to demonstrate that the RBC-derived ATP release in response to hypoxia
is dependent upon the signal transduction pathway previously detailed here, RBCS are
incubated with dianride, a known cell stiffener, and niflurrric acid, previously
demonstrated to inhibit CFTR.5 These solutions are increased from 100 to 200 uM for
diamide and fi'om 20 to 40 uM for niflumic acid to further illustrate the ability to inhibit
the RBC-derived ATP release that is responsible for the intracellular NO production
within the bPAECs of each well.

The RBCS were first prepared in PSS that contained the various concentrations of

either diamide or niflumic acid. The niflumic acid was allowed to incubate with the cells

150

 

 

for 30 min, where as the diamide was only incubated for 10 min prior to centrifuging, at
500 x g at 4 °C for 3 min, for the following resuspension in hypoxic buffer. These times
were chosen to maximize the effect of ATP inhibition, as previously reported, the RBC is
able to begin to recover from the stiffening of the cell due to diamide within
approximately 30 min at which point the concentration of ATP release begins to return to
normal.10 In order to introduce the RBCS in a hypoxic environment Oxyrase® was used
to enzymatically remove the 02 from the PSS and here 5 mL of PSS with 1.5 mL of
Oxyrase® was added to each of the centrifuged RBCS.

The uF A was prepared as previously discussed in chapter 4, briefly, bPAECs
were cultured within the array of wells which were then subsequently incubated with a 5
mM L-arginine solution prior to the incubation with 100 M DAF-FM DA, a
fluorescence probe for intracellular NO production, which was removed and replaced
with HBSS prior to fluorescence imaging. Figure 5.1 illustrates the fluorescence image
taken of the entire six channel device as various samples are flowed through the
underlying microfluidic channels for 30 min at a flow rate of 1.0 uL/min. The Olympus
MVX stereomicroscope (Center Valley, PA) fitted with a mercury-arc lamp and FITC
filter cube, allowing for excitation and emission wavelengths of 470 nm and 525 nm was
employed to obtain fluorescence image of the array of 18 wells. As before, the
fluorescence intensities were imaged using Microsuite software and the pixel intensities

were averaged for each of the three wells addressed by a single channel.

151

 

bPAECs + DAF-F I\/1 DA
1

3.5%RBCS

3.5% 13on RBCS

3.5% DEOX RBCS

+ 100 ul\l DIAIVLLDE

35% DEOX RBCS
+ 200 ul\r1 DIAIMIDE

3.5% DEOX RBCS
+ 20 ulvl NIFLUIVIIC

3.5% DEOX RBCS
+ 40 uIVI NIFLUIVIIC

 

Figure 5.1 Fluorescence image of the six-channel uFA for the
determination of intracellular NO production within the bPACEs as a
result of hypoxia. Here, the RBC-derived ATP release from each
sample stimulates the production of NO which is detected using DAF-
FM DA and the Olympus MVX stereomicroscope fitted with a
mercury-arc lamp for excitation and a FITC filter cube with excitation
and emission wavelengths of 470 nm and 525 nm respectively.

152

5.2.2 C-PEPTIDE: RBC-MEDIATED NO PRODUCTION

The same uF A for the determination of intracellular NO production within
bPAECs cultured within the array-like wells is used to demonstrate that an additional
RBC-derived ATP release due to metal-activated C-peptide may play a vital role in
increasing vasodilation in vivo. The samples examined were comprised of 7% RBCS,
RBCS incubated with metal-activated C-peptide, with C-peptide alone and metal alone,
with 1 mM glybenclamide, and finally, with 1 mM glybenclamide and metal-activated C-
peptide. The solution containing 1 mM glybenclamide was prepared first, providing
inhibition of CFTR prior to the incubation with metal-activated C-peptide. As discussed
in chapter 4, the 10 mM glybenclamide stock solution was prepared by combining 0.049
g with 2 mL 0.1 M NaOH, and 7.94 mL dextrose solution (1 g dextrose in 20 mL DDW)
and heating to 52 °C until dissolved with stirring. This solution was added to the PSS
that was used to create the 7% RBC solution for subsequent incubation with metal-
activated C-peptide.

Recent work in our lab has shown that the relevant metal for the activation of C-
peptide resulting in an increased ATP release, is Zn2+ and its preparation allowing for
non-competitive interaction with the peptide is necessary. An 80 uM stock solution of
Zn2+ is prepared by dissolving 0.0055 g in 500 mL DDW. In order to obtain an 83 uM
stock solution of C-peptide, 0.25 g of human C-peptide (American Peptide, Sunnyvale,
CA, USA) is dissolved in 100 mL DDW which is then used to make subsequent dilutions
within RBCs. To prevent competitive binding events between the peptide and other
metals or ions, the Zn2+ and C-peptide are first combined in DDW before combining with

PSS and RBCS. The final concentrations of each Zn2+ and C-peptide are 10 nM in PSS

153

which is incubated with 7% RBCS for 6 hrs prior to pumping through the underlying
microfluidic channels of the uF A.

The fluorescence image, figure 5.2, was obtained by imaging the entire six
channel device after each of the solutions were flowed through the underlying
microfluidic channels for 30 min at a flow rate of 1.0 pL/min. The same Olympus MVX
stereorrricroscope (Center Valley, PA) fitted with a mercury-arc lamp and FITC filter
cube, allowing for excitation and emission wavelengths of 470 nm and 525 nm was
employed. As before, the fluorescence intensities were imaged using Microsuite
software and the pixel intensities were averaged for each of the three wells addressed by

a single channel.

5.2.3 REQUIRED ASPECT OF SHEAR: TRENTAL VS. ILOPROST
Our group has previously examined the necessity of a flow system as it applies to
drug efficacy and mechanisms of action; specifically demonstrating the required aspect of

3 Four

shear-induced deformation for trental to increase RBC-derived ATP release.
samples consisting of 3.5% hematocrit RBCS alone, RBCS incubated with 5 uM trental, 1
uM iloprost and the combined incubation with both 5 uM trental and 1 uM iloprost. The
trental stock solution is prepared as a 100 uM solution where 0.0028 g are dissolved in 10
mL PSS which is subsequently diluted to a 10 uM solution prior to the addition to the
washed RBCs with a final concentration of 5 uM in PSS which is allowed to incubate for
30 min. The iloprost solution is prepared as previously described in chapter 4, where a

100 uM stock solution in PBS is diluted to 1 uM with 3.5% RBCs in PSS which is

allowed to incubate for 30 min. For the solution containing both trental and iloprost, a

154

bPAECs I DAIi-FM DA

 

7% RBCS

+CM

+C

+th

+GLY

+GLY +CM

 

Figure 5.2 Fluorescence image of the six-channel uFA used to
monitor the intracellular production of NO within the bPACEs
cultured within the wells of the device due to the incubation of
RBCS with C-peptide. Here, the RBC-derived ATP release from
each sample stimulates the production of NO which is detected
using DAF-FM DA and the Olympus MVX stereomicroscope fitted
with a mercury-arc lamp for excitation and a F ITC filter cube with
excitation and emission wavelengths of 470 nm and 525 nm
respectively.

155

solution containing 5 uM trental and 3.5% RBCS are incubated for 30 min to which the
volume required for a final concentration of 1 uM iloprost solution is incubated for an
additional 30 min, thus, allowing trental to act on the RBCs prior to the addition of
iloprost.

Each of these solutions (100 uL) is combined with the luciferin/luciferase solution
(100 uL) in a cuvette and placed over a PMT housed within a dark box, as illustrated in
figure 5.3. A 100 uM stock solution of ATP was prepared by adding 0.0551 g of ATP to
1000 mL of DDW, subsequently, ATP standards with concentrations ranging between 0
and 1.5 uM were then prepared in PSS from the stock and used to obtain a standard
working curve. To prepare the luciferin/luciferase mixture used to measure the ATP by
chemiluminescence, 5 mL of DDW were added to a vial containing luciferase and
luciferin. In order to enhance the sensitivity of the assay, 2 mg of luciferin were added to
the vial. The chemiluminescence is measured as described in chapter 4, where the
current detected by the PMT is converted to voltage which is measured in real time and
plotted at a rate of 10 measurements per second for 30 seconds. A period of 15 s was
allowed from the time of mixture of the two solutions in the cuvette to the time the
measurement was taken by the PMT. This allowed for improved reproducibility and

helped to ensure that all time intervals were identical for each solution.

5.5 RESULTS AND DISCUSSION
The normalized fluorescence intensity in figure 5.4 indicates that endotlrelium-
derived NO is increased in the presence of hypoxic RBCS (2.13 :l: 0.07) when compared

to RBCS alone (1.00 i 0.21). Specifically, a 3.5% RBC sample incubated in hypoxic

156

-------------- Dark Box

\

I
I

 

 

100 pL L/L ,,
100 ”L RBC or . |
Control Samples ----- Gir-
W

 

Figure 5.3 Schematic drawing of the experimental set up for determining the RBC-
derived ATP release from samples incubated with trental and or iloprost to
demonstrate their opposing mechanisms of action. Here, 100 uL of the luciferin and
luciferase (L/L) solution is combined in a cuvette with 100 uL of each sample. The
chemiluminescence is measured as before where the voltage is measured in real time
and plotted at a rate of 10 measurements per second for 30 seconds onto a computer
using software that was written in—house.

157

2.5

 

n=3 *1 *p<0.001
i p<0.05

Normalized Fluorescence Intensity

 

 

 

3.5% RBCS DEOX RBCS +100 uM +200 uM +20 uM +40 uM
DIAMIDE DIAMIDE NIFLUMIC NIFLUMIC

Figure 5.4 Normalized fluorescence intensity obtained from the fluorescence
image of the six-channel uF A used to determine the NO produced within bPAECs
cultured within the array of wells upon stimulation of various RBC samples
flowing through the underlying microfluidic channels. Here, these samples
include 3.5% RBCS (1.00 :l: 0.21), RBCS in deoxygenated buffer (2.13 i 0.07),
and four deoxygenated samples with prior treatment of 100 uM (1.52 :h 0.33) and
200 uM (0.91 :t 0.07) diamide and 20 [TM (1.27 :t 0.13) and 40 uM (1.10 :l: 0.11)
niflumic acid. Here the intensities are reported as averages of each well,
providing triplicate measurements, and the error is reported as the standard error
of the mean.

158

buffer stimulated more than a 2-fold increase in fluorescence intensity when compared to
RBCs alone, which is directly proportional to the intracellular NO production within the
bPAECs. The fluorescence intensity resulting from the RBCS hypoxic environment is
decreased by 28.6% and 40.4% when RBCS are first incubated with 100 uM diamide or
20 uM niflumic acid prior to introduction into the hypoxic buffer. Demonstrating the
ability to reduce this NO production even further, incubation with 200 uM diamide and
40 uM niflumic acid correspond to a decrease of 57.3% and 48.4% respectively, when
compared to hypoxic RBCS.

Using the same methodology as for the hypoxic study, figure 5.5 represents the
preliminary data obtained to determine the increased NO production within bPAECs due
to an increase in ATP release from the RBCs incubated with metal-activated C-peptide.
The normalized fluorescence intensity reported for RBCs (1.00 :t 0.05) incubated with
the metal-activated C-peptide (1.31 :1: 0.05) was increased by 31.0%. This intensity was
consequently reduced (1.10 i 0.20) upon prior treatment of 1 mM glybenclamide,
demonstrating that the increase in fluorescence intensity was not due to cell lysis upon
the incubation with metal-activated C-peptide.

Finally, the non-flow system and the luciferin/luciferase ATP assay was used to
show that without subjection to mechanical deformation, the RBCS treated with trental
does not elicit an increased ATP release. Figure 5.6 shows that the combination of both
trental and iloprost will result in an increased RBC-derived ATP release even in a non-
flow system. Here, the increase from RBCS (0.098 :t 0.006 M) in response to iloprost

(0.327 i 0.070 uM) was three-fold, and the combination of both trental and iloprost

159

 

 

Normalized Fluorescence Intensitiy

 

 

 

RBCS +CM +GLY +GLY +CM

Figure 5.5 Initial results of the fluorescence intensity observed when
addressing bPAECs within the uFA with RBC samples alone (1.00 d: 0.05) and
with the incubation of metal-activated C-peptide (1.31 i 0.05), 1 mM
glybenclamide (0.94 i 0.11), and 1 mM glybenclamide and metal-activated C-
peptide sequentially (1.10 i 0.20). Here the intensities are reported as averages
of each well, providing triplicate measurements, and the error is reported as the
SEM where the increased fluorescence intensity of RBCS incubated with metal
activated C-peptide is statistically different from RBCS alone (p < 0.05).

160

 

.0 .0 _O
-# U1 0"
I I

[ATP] Release in uM
.0
(A)

  

 

 

3.5% RBCs + Trent + Ilop + Trent/llop

Figure 5.6 Summarized data for the measurement of RBC-derived ATP release
in a non-flow system employing the luciferin and luciferase assay. Here, ATP
concentrations in micromolar are reported for 3.5% RBCS (0.0981 :t 0.0063
uM), with the incubation of each 5 uM trental (0.1126 2t 0.0218 uM), 1 uM
iloprost separately (0.3265 :t 0.0702 M) and combined with 5 uM trental
(0.4628 3: 0.0935 M). The average concentration of ATP released from each
RBC sample is reported where the error is represented by the SEM.

161

(0.463 d: 0.094) was more than four-fold, while no statistical difference in RBC-derived

ATP release was observed for trental (0.113 i 0.094 uM) when compared to RBCs alone.

5.5.1 VASODILATION MIMIC FOR MECHANISMS OF ACTION

Employing the uFA allows for the ability to better mimic the in vivo circulation
system, specifically for the determination of hypoxia- and C-peptide-induced RBC-
derived ATP release resulting in increased NO production in bPAECs cultured in the
uFA. While research has previously determined that both hypoxia1 and C-peptidel3 will
increase the RBC-derived ATP release, this is the first demonstration that this will induce
an increased NO production within bPAECs when compared to RBCS alone. Thus,
demonstrating the multicellular mechanism of action for each hypoxia and metal
activated C-peptide to elicit the production of endothelium-derived NO production.

Preliminary results demonstrate the need to examine the mechanisms of action
within an appropriate system, as it is shown here, trental does not elicit a response in
RBC-derived ATP release in a non-flow system. Similarly, the combination of each
trental and iloprost induces an even greater ATP release from RBCS than with iloprost
alone. This indicates that while iloprost will stimulate RBC-derived ATP release through
IPR and the activation of the Gs-coupled signal transduction pathway within the cell,
trental could possibly increase membrane flexibility, and has been demonstrated to
require the shear-aspect of flow to induce ATP release.3 Hence, the combination of these
two pharmaceuticals, acting each in their own way, elicit an even greater increase in ATP

release.

162

 

It has been shown that RBCS obtained from diabetic patients are stiffened2728 due

29'” as well as a

to a reduced antioxidant ability and intracellular GSH concentrations,
reduction in the expressed G, that is involved in the signal transduction pathway for ATP
release.9 With this in mind, the ultimate objective of this research is to provide a more
appropriate analytical tool to provide individualized medicine, effectively putting the
right drug into the right patient. This uFA allows for a more systems biology approach to
drug discovery and clinical testing by mimicking the resistance vessels found in vivo,
providing a means to better understand the multicellular mechanism of action as it
pertains to vasodilation. Specifically, iloprost may be limited in alleviating the
complications associated with diabetes for a patient with a reduced G, expression,

however addition of trental as well, may allow for increased vasodilation by increasing

the RBC deformability as well.

163

 

REFERENCES

Paris, A. & Spence, D.M. Measuring the simultaneous effects of hypoxia and
deformation on ATP release from erythrocytes. Analyst 133, 678-682 (2008).

Kosower, N.S., Kosower, B.M., Wertheim, B. & Correa, W.S. Diamide, a new
reagent for the intracellular oxidation of glutathione to the disulfide. Biochem
Biophys Res Commun 37, 593-596 (1969).

Carroll, J.S., Ku, C.-J., Karunarathne, W. & Spence, D.M. Red Blood Cell
Stimulation of Platelet Nitric Oxide Production Indicated by Quantitative

Monitoring of the Communication between Cells in the Bloodstream. Anal Chem
79, 5133-5138 (2007).

Sprague, R.S., Ellsworth, M.L., Stephenson, A.H. & Lonigro, A.J. ATP: the red
blood cell link to NO and local control of the pulmonary circulation. Am J Physiol
271, H2717-H2722 (1996).

Sprague, R.S., Ellsworth, M.L., Stephenson, A.H., Kleinhenz, M.E. & Lonigro,
A.J. Deformation-induced ATP release fi'om red blood cells requires CFTR
activity. Am J Physiol Heart Circ Physiol 275, H] 726-1 732 (1998).

Olearczyk, J .J ., Stephenson, A.H., Lonigro, A.J. & Sprague, R.S. Receptor-
mediated activation of the heterotrimeric G-protein Gs results in ATP release
fi'om erythrocytes. Med Sci Monit 7, 669-674 (2001).

Sprung, R., Sprague, R. & Spence, D. Determination of ATP release fiom
erythrocytes using microbore tubing as a model of resistance vessels in vivo. Anal
Chem 74, 2274-2278 (2002).

Fischer, D.J., Torrence, N.J., Sprung, R.J. & Spence, D.M. Determination of
erythrocyte deformability and its correlation to cellular ATP release using
microbore tubing with diameters that approximate resistance vessels in vivo.
Analyst 128, 1163-1168 (2003).

Sprague, R.S., Stephenson, A.H., Bowles, E.A., Stumpf, M.S. & Lonigro, A.J.
Reduced expression of Gi in erythrocytes of humans with type 2 diabetes is

associated with impairment of both cAMP generation and ATP release. Diabetes
55, 3588-3593 (2006).

164

 

10.

ll.

12.

l3.

14.

15.

16.

17.

l8.

19.

20.

Carroll, J. et al. An altered oxidant defense system in red blood cells affects their
ability to release nitric oxide-stimulating ATP. M01 Biosyst 2, 305-311 (2006).

Genes, L.I., Tolan, N.V., Hulvey, M.K., Martin, R.S. & Spence, D.M. Addressing
a vascular endothelium array with blood components using underlying
microfluidic channels. Lab Chip 7, 1256-1259 (2007).

Sprague, R.S., Stephenson, A.H. & Ellsworth, M.L. Red not dead: signaling in
and from erythrocytes. Trends Endocrinol Metab 18, 350-355 (2007).

Meyer, J .A., Froelich, J .M., Reid, G.E., Karunarathne, W.K.A. & Spence, D.M.
Metal-activated C-peptide facilitates glucose clearance and the release of a nitric
oxide stimulus via the GLUTl transporter. Diabetologia 51, 175-182 (2008).

Kunt, T., Forst, T., Pfiitzner, A., Beyer, J. & Wahren, J. The physiological impact
of proinsulin C-peptide. Pathophysiology 5, 257-262 (1999).

Furchgott, R.F. & Zawadzki, J .V. The obligatory role of endothelial cells in the
relaxation of arterial smooth muscle by acetylcholine. Nature 288, 373-376
(1980).

Ignarro, L.J., Buga, G.M. & Chaudhuri, G. EDRF generation and release from
perfused bovine pulmonary artery and vein. Eur J Clin Pharmacol 149, 79—88
(1988)

Moncada, S., Palmer, R. & Higgs, E. The discovery of nitric oxide as the
endogenous nitrovasodilator. Hypertension 12, 365-372 (1988).

Palmer, R.M.J., Ashton, D.S. & Moncada, S. Vascular endothelial cells synthesize
nitric oxide from L-arginine. Nature 333, 664-666 (1988).

Moncada, S., Palmer, R.M.J. & Higgs, E.A. Nitric oxide: physiology,
pathophysiology, and pharmacology. Pharmacol Rev 43, 109-142 (1991).

Sobey, C.G. Potassium channel function in vascular disease. Arteriosclerosis,
Thrombosis, and Vascular Biology 21, 28-38 (2001).

165

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

Webb, R.C. Smooth Muscle Contraction and Relaxation. Adv Physiol Educ 27,
201-206 (2003).

Tuteja, N., Chandra, M., Tuteja, R. & Misra, M.K. Nitric oxide as a unique
bioactive signaling messenger in physiology and pathophysiology. J Biomed
Biotechnol 227-237 (2004).

Martinez-Morales, M.E.D.-E.a.F. Pentoxifylline Diminishes the Oxidative
Damage to Renal Tissue Induced by Streptozotocin in the Rat. Exp Diabesity Res
5, 245-251 (2004).

Gunduz, Z. et al. The Effects of Pentoxifylline on Diabetic Renal Changes in
Streptozotocin-Induced Diabetes Mellitus. Ren Fail 26, 597 - 605 (2004).

Navarro, J.F., Mora, C., Muros, M., Macia, M. & Garcia, J. Effects of
pentoxifylline administration on urinary N-acetyl-beta-glucosarnirridase excretion

in type 2 diabetic patients: a short-term, prospective, randomized study Am J
Kidney Dis 42, 264-270 (2003).

Radfar, M. et al. Effects of pentoxifylline on oxidative stress and levels of EGF
and NO in blood of diabetic type-2 patients; a randomized, double-blind placebo-
controlled clinical trial. Biomed Pharmacother 59, 302-306 (2005).

MacRury, S.M., Small, M., Anderson, J., MacCuish, A.C. & Lowe, G.D.
Evaluation of red cell deformability by a filtration method in type 1 and type 2

diabetes mellitus with and without vascular complications. Diabetes Res 13, 61-
65(1990)

Schwartz, R.S., Madsen, J .W., Rybicki, A.C. & Nagel, R.L. Oxidation of spectrin
and deformability defects in diabetic erythrocytes. Diabetes 40, 701-708 (1991).

Costagliola, C. Oxidative state of glutathione in red blood cells and plasma of
diabetic patients: in vivo and in vitro study. Clin Physiol Biochem 8, 204-210
(1990)

De Mattia, G. et al. Influence of reduced glutathione infusion on glucose
metabolism in patients with non-insulin-dependent diabetes mellitus. Metabolism
47, 993-997 (1998).

166

 

31.

32.

33.

Dincer, Y., Akcay, T., Alademir, Z. & Ilkova, H. Effect of oxidative stress on

glutathione pathway in red blood cells from patients with insulin-dependent
diabetes mellitus. Metabolism 51, 1360-1362 (2002).

Constantin, A., Constantinescu, E., Dumitrescu, M., Calin, A. & Popov, D.
Effects of ageing on carbonyl stress and antioxidant defense in RBCs of obese
type 2 diabetic patients. J Cell Mol Med 9, 683-691 (2005).

Nwose, E.U., J elinek, H.F., Richards, R.S. & Kerr, P.G. in Redox Report, Vol. 11
99-104 (Maney Publishing, 2006).

167

Mlclllllllllllllll lllllllllllllllll lllllllll
3 1293 03063 1901