ha an .23, 55,.

!bl£.c
a. F:

 

11...}.
.03...

 

 

 

 

iii
if h...

m
hf :dmvﬁ. 3.1%;

1.. ,2 siiﬁvsa).
.3... \- 5: I h.‘

.1.hﬁlu...ss.....ni.: a 1.

513.232 31!?

a... I

x. 91...!
5...! inﬁllmﬂi
. ,. 1 a
.d. .bh...’ _
1.21.31
5...!

3.1.45.1... féﬁqﬁnwiﬁ $.2me .. fﬁﬂﬁ .éjﬁw 2min“. . at";
V 13‘. ‘ u 3 .1 .. .. .. I. :..: .. 2.x. . Li
. .. . . . L 53.5...

I t . . . , I

 

 

mm LIBRARY
2- Michigan State
LOGCI University

This is to certify that the

dissertation entitled

DEVELOPMENT OF A MICROFLUIDIC BASED
MICROVASCULAR MODEL: TOWARDS A COMPLETE
BLOOD BRAIN BARRIER (BBB) MIMIC

presented by
LUIZA I. GENES-HERNANDEZ

has been accepted towards fulﬁllment
of the requirements for the

 

 

 

 

PhD deQree in Chemistry
, ‘ : ' ' g wrProfessor’s Signature
/t‘V"’8
Date

MSU is an Afﬁrmative Action/Equal Opportunity Employer

 

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 K:IProj/Acc&Pres/ClRC/DateDuerindd

DEVELOPMENT OF A MICROFLUIDIC BASED
MICROVASCULAR MODEL: TOWARDS A COMPLETE BLOOD
BRAIN BARRIER (BBB) MIMIC

By

Luiza I. Genes-Hernandez

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Chemistry

2008

ABSTRACT

DEVELOPMENT OF A MICROFLUIDIC BASED MICROVASCULAR MODEL:
TOWARDS A COMPLETE BLOOD BRAIN BARRIER (BBB) MIMIC

By

Luiza I. Genes-Hemandez

Multiple sclerosis (MS) is an inﬂammatory and demyelinating disease of the
central nervous system (CNS) that is the most common neurodegenerative condition in
young adults, especially women, striking people between the ages of 20 and 40. It is
unclear which compounds are most responsible for progressive myelin destruction and
remyelination in MS.

Recently, elevated levels of nitrate and nitrite have been measured in cerebral
spinal ﬂuid (CSF), urine, and serum of MS patients. The origin of this excess nitric oxide
(NO) and its mechanism of action are unknown. Herein, we have constructed a theory
that the overabundance of NO metabolites in the CSF is actually blood-home; in other
words, it is possible that the NO levels are high due to overstimulation by red blood cell
(RBC)-derived adenosine triphosphate (ATP). Our hypothesis is that MS-RBCs are more
deformable, therefore releasing more ATP, which in turn will lead to higher production
of NO. One of the biological fates of NO is the ability to cross the blood brain barrier
(BBB) into the CNS, where it may result in the breakdown of the myelin sheath,
compromise the integrity of the BBB and lead to axonal degeneration. To prove or
disprove this hypothesis, the role of RBC-derived ATP and NO derived from brain
endothelium has been investigated. In an attempt to make a connection between RBC-

derived ATP and the increased levels of NO reported, we were able to demonstrate that

the ATP released from the RBCs obtained from 18 MS patients was more than twice the
amount of ATP released from the RBCs of 11 healthy, non-MS controls.

Moreover, in this work we applied the recent developments in microfabrication
technology to create a BBB mimic with blood components using lithographically-derived
microchips in order to monitor the fate of endothelium-derived NO that is stimulated by
mechanically deformed RBCs. We showed that an array of endothelial cells, addressable
by an underlying microﬂuidic network of channels containing RBCs, can be employed as
an in vitro model of the in vivo circulation to monitor cellular communication between
different cell types. Results obtained from this array suggest that the ability of iloprost, a
stable analogue of prostacyclin, to stimulate NO production in endothelial cells, may be
due to its ability to stimulate ATP release from the red cell. These results provide
evidence that the described device may serve as a controlled, in vitro platform for
performing in viva-type measurements.

Moreover, this device was successfully used in a study that attempts to classify
the possible role of prolactin releasing peptide (PrRP) (a peptide that stimulate the release
of prolactin from the pituitary gland) in regulating the vascular tone by decreasing the

endothelium-derived NO production through a decrease in ATP.

~ To my beautiful angel, Soﬁa ~

iv

ACKNOWLEDGEMENTS

More than anything during this journey, I learned not to take for granted things
like family, friends and God. Words alone cannot express the tremendous love, moral
support and ﬁnancial help that my family, especially my parents, surrounded me with all
these years. Without their help I could not have been at this point in my life and without
my father’s guidance I would have never been chosen the path of science.

I also have to thank my teachers that inspired me and helped me to get here. I am
deeply thankful to my advisor, Dr. Dana Spence, for challenging me with a difficult
project, which if I recall correctly I insisted in taking, and pushing my limits to discover
an inner strength that I did not know I had before. I am grateful to him, not only for being
a good advisor and teacher, but also a really good person and understanding the
challenges I went through outside the lab. When I joined his group in 2004, at Wayne
State University, I was not quite sure about what future might hold for me. I always
wanted to learn more biology and chemistry, but also not be completely detached from
my engineering background. Dr. Spence’s enthusiasm about his research and reassurance
made everything look really bright and gave me the conﬁdence in approaching a totally
different and at the time new for me, ﬁeld of science. I am really grateful to him for
making me take this challenge and I now realize how much I have grown since then, not
only as a ready to be independent scientist, but also as a person. And in all honesty, he is
the reason why, I now know that I want to pursue an academic career and someday have

my own research group.

In these ﬁve years I have gained tremendous experience and knowledge and more
importantly, beautiful and hopefully ever lasting friendships. I met wonderful people and
made amazing friends, and I would like to call them, my Sri Lankan family. I thank you,
Wasanthi, Madushi, Ajith, Wathsala for being there for me at times when I needed help
and support, both physical and spiritual. I realized that, there is a small world after all,
and people are so much alike.

I would also like to acknowledge Nicole, who was like a “ray of sunshine in a
cloudy day”. Nicole and I had very fruitful discussions about research and not only. I
would like to say that we connected at a very special level and she will always be close to
my heart.

I would also take the opportunity to mention Dr. Paul Root, for the coolness that
he always exuded while in the lab, and I would like to thank him for sharing his
knowledge about cell culturing and nitric oxide biochemistry. There are many people that
directly, or indirectly, contributed to me becoming the person and scientist that I am
today and I would like to thank them all.

My tremendous love and all my heart go to my little Soﬁa and my wonderful
loving husband, Frank. Both of you are the reasoning of my life and my work, and both
of you came into my life at a time when I was completely lost. I love you and I am

forever grateful for having you.

vi

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. x
LIST OF FIGURES .......................................................................................................... xi
CHAPTER 1
l .1 INTRODUCTION ........................................................................... 1
1.1.1 Red Blood Cell (RBC) Structure and Properties ............................... 1
1.1.2 RBC-Derived ATP: An Important Factor in Vascular Regulation ........... 5
1.1.3 CFTR-Facilitated ATP Transport/Release out of the Cell ..................... 8
1.1.4 RBC Deformability and Disease Pathogenesis .................................. 9
1.1.5 Nitric Oxide (NO) and the Blood-Brain Barrier (BBB) ...................... 12
1.2 MICROFLUIDIC TECHNOLOGY ..................................................... 17
1.2.1 Short History ....................................................................... 17
1.2.2 Scaling Effects when Miniaturizing the Chemistry ........................... 18
1.2.3 Material for Microsystems ....................................................... 20
1.2.4 Fabrication of Microﬂuidic Systems ........................................... 23
1.2.5 Tools for Microﬂuidic Applications ............................................ 26
1.2.6 Detection Modes in Microﬂuidic Devices ..................................... 31
1.3 THESIS OBJECTIVE ...................................................................... 34
LIST OF REFERENCES ........................................................................... 36
CHAPTER 2
2.1 MULTIPLE SCLEROSIS: AN AUTOIMMUNE DISEASE? .............................. 49
2.2 NITRIC OXIDE AS AN ACTIVITY MARKER IN MS ............................. 51
2.3 RBC DEFORMABILITY IN MS ........................................................ 54
2.4 POSSIBLE ROLES OF THE RBC/RBC-DERIVED ATP IN MS .................. 55
2.5 EXPERIMENTAL ......................................................................... 56
2.5.1 Design and Methods .............................................................. 57
Generation of washed RBCS
Standard A TP assay
Determination of A TP release ﬁom RBCS
2.5.2 Results and Discussion ........................................................... 62

2.6

2.7

Quantitative Evaluation of the MS RBC ATP release
CONSTITUTIVE PRODUCTION OF NITRIC OXIDE IN BRAIN

MICROVASCULAR ENDOTHELIAL CELLS ....................................... 62
EXPERIMENTAL ......................................................................... 64
2.7.1 Design and Methods .............................................................. 65

Preparation of cells
F lorescence determination of NO production
Control experiment

vii

2.7.2 Results and Discussion ............................................................ 66
ATP constitutes an in vivo stimulus of NO production in bBMVECs
AT P-derived NO production with cells passage

2.8 CONCLUSIONS ........................................................................... 70
LIST OF REFERENCES ........................................................................... 71
CHAPTER 3
3.1 THE BLOOD BRAIN BARRIER ....................................................... 76
3.2 BLOOD BRAIN BARRIER MODELS ................................................. 78
3.3 MICROFLUIDIC TECHNOLOGY FOR THE DEVELOPMENT OF A BLOOD
BRAIN BARRIER MIMIC ............................................................... 82
3.4 EXPERIMENTAL ......................................................................... 83
3.4.1 Design and Methods .............................................................. 83

Microchip fabrication
Preparation of N0 standards
Detection of NO
Detection of A TP flux through the polycarbonate membrane
Fabrication of the microchip-based BBB
3.4.2 Results and Discussion ........................................................... 90
N0 ﬂux through the polycarbonate membrane
Cell immobilization on polycarbonate membrane/PDMS channels
Detection of A TP flux through the polycarbonate membrane

3.5 CONCLUSIONS AND OTHER CONSIDERATIONS .............................. 99
LIST OF REFERENCES .......................................................................... 101
CHAPTER 4

4.1 ADDRESSING A VASCULAR ENDOTHELIUM ARRAY WITH BLOOD
COMPONENTS USING UNDERLYING MICROFLUIDIC CHANNELS.....104

4.2 EXPERIMENTAL ........................................................................ 106
4.2.1 Material and methods ............................................................ 106
Preparation of RBCs

Measurement of A T P via chemiluminescence
Measurement of bPAEC—derived N0

Preparation of glybenclamide solution

Preparation of the microﬂuidic devices
Immobilization of the bPAECs on the membrane array

4.2.2 Results and Discussion ......................................................... 111
4.3 CONCLUSIONS ......................................................................... 1 19
LIST OF REFERENCES ......................................................................... 120

viii

CHAPTER 5
5.1 A PROPOSED MECHANISM FOR THE INVOLVEMENT OF PROLACTIN-
RELEASING PEPTIDE (PrRP) IN THE REGULATION OF VASCULAR

TONE ....................................................................................... 122

5.2 EXPERIMENTAL ........................................................................ 125

5.2.1 Materials and Methods .......................................................... 125
Collection of RBCs

Measurement of the ATP release
Mass spectrometry analysis
Fabrication of the microﬂuidic array
Cell immobilization

Measurement of Nitric Oxide

5.2.2 Results and Discussion ......................................................... 130
ATP binding to PrRP
Antagonistic effect of PrRP on the ATP release from C-peptide/iloprost
treated RBCs
LIST OF REFERENCES ........................................................................ 140
CHAPTER 6
6.1 OVERALL CONCLUSIONS ........................................................... 143
6.2 FUTURE DIRECTIONS ................................................................ 148
LIST OF REFERENCES ......................................................................... 151

ix

LIST OF TABLES

Table 1. Classiﬁcation of microvalves ........................................................... 28

LIST OF FIGURES

Figure l. a) Healthy RBC with its usual discocyte shape. b) Typical dimensions
of a healthy RBC.l ........................................................................ 2

Figure 2. a) Composition of the membrane of a RBC. b) Image of the spectrin
cytoskeleton.2 .............................................................................. 4

Figure 3. Proposed mechanism for the effect of RBC-derived ATP on the cells
at the luminal and extraluminal side of a blood vessel .............................. 7

Figure 4. Proposed mechanisms for CFTR-facilitated ATP transport/release out
of cell. For hypothesis A to occur, CFTR must be expressed in concert
with appropriate cofactors or regulators. Hypothesis B and C suggest
positive CF TR regulation of a separate and independent ATP transport
mechanism and CFTR-facilitated fussion of ATP-ﬁlled vesicles,
respectively.3 ............................................................................ 10

Figure 5. Electrons (e-) are donated by NADPH to the reductase domain of the
enzyme (eNOS) and proceed via FAD and F MN redox carriers to
the oxygenase domain. There, they interact with the haem iron and BH4
at the active site to catalyze the reaction of oxygen with L-arginine,
generating citrulline and NO as products. Electron ﬂow through the
reductase domain requires the presence of bound calcium-
calmodulin (Ca2+/CaM).61 ............................................................. 14

Figure 6. A) Representation of the steps in the lithographic process. B) Scheme
of rapid prototyping ..................................................................... 25

xi

Figure 7. Types of micromixing. A. Cyclone micromixer; B. Hydrodynamic
focusing by compressing a central stream by two outer streams at much
larger ﬂow rate; C. Multi-lamination ﬂow patterns of a split-and-
recombine mixer visualized by a dilution-type experiment;
D. Staggered herringbone micromixer for the generation of chaotic ﬂows;
E. Time lapse images of three-electrode mixing at 8 Hz; F. Electrokinetic
instability micromixing; G. Numerical simulation results obtained for two
inlet ﬂows pulsed at 180C] phase difference; H. Photographs showing
acoustic microstreaming in a microchamber.4 ....................................... 30

Figure 8. The two-step bioluminescence reaction between luciferin/luciferase and
ATP ....................................................................................... 59

Figure 9. Instrumental setup employed to measure the RBC derived ATP release ........ 61

Figure 10. ATP released from RBCs subjected to shear-induced stress. The average
ATP release from the RBCs obtained from healthy controls was
138 +/- 21 nM while the release from the RBCs obtained from
people with MS was 375 +/- 51 nM. The error bars are reported as
the SEM for 11 controls and 18 MS samples ...................................... 63

Figure 11. Effects of ATP (100 uM) and L-NAME (100 pM) on NO
production in bBMVEC assessed with DAF-FM DA. L-MAME was
used to demonstrate that the increase in DAF-FM DA ﬂuorescence
upon stimulation with ATP (100 uM) was the result of NO production
(*p 5 0.001 vs. control) ............................................................... 67

Figure 12. ATP-derived NO production is decreasing with increasing the passage
of the cells. For cell division 3 there is a signiﬁcant difference between
the control (black bars) and the bBMVECs stimulated with 100 pM
ATP (grey bars) (*p 5 0.001 vs. control) ............................................ 69

xii

Figure 13.

Figure 14.

Figure 15.

Figure 16.

Figure 17.

The Blood Brain Barrier. a) The BBB is formed by endothelial cells
of the level of the cerebral capillaries. These endothelial cells interact
with perivascular elements such as basal lamina and closely
associated astrocytic end-feet processes, perivascular neurons and
pericytes to form a functional BBB. b) Cerebral endothelial cells are
unique in that they form complex tight junctions (TJ) produced

by the interaction of several transmembrane proteins that effectively
seal the paracellular pathway. These complex molecular junctions
make the brain practically inaccessible for polar molecules, unless
they are transferred by transport pathways of the BBB that regulate
the microenvironment of the brain.4 ................................................ 77

Typical in vitro model of the BBB. Brain endothelial cells are grown

at the bottom of ﬁlter inserts together with glial cells grown at the

bottom of 6-well plate. Soluble factors secreted from the glial cells

induce the BBB phenotype in the brain capillary endothelium.4 ................ 80

Photolithographic procedure for the fabrication of silicon master.

1) Spin coat 4 mL of SU-8 50 photoresist at 1000 rpm for 20 s;

2) Pre-bake for 5 min at 95°C to evaporate solvent and densify ﬁlm;

3) Post-bake for 5 min at 95°C to cross-link the exposed portions

of the photoresist ﬁlm; 4) UV exposure, 350-450 nm, 50W for 60 s

to polymerize and cross-link; 5) Develop with propylene glycol

monomethyl ether acetate (PGMEA); 6) Soft-lithography to create
microﬂuidic chip ....................................................................... 84

Detection of ATP ﬂux through a polycarbonate membrane using a

ﬂow of luciferin/luciferase mixture hydrodynamically pumped into

a spiral-shaped channel of a PDMS device. The set up is enclosed

in a light excluding box over a PMT ................................................ 87

Proposed design for the microﬂuidic-based BBB mimic A. Assembly

of a 3 dimensional ﬂuidic device incorporating microchannels

(100 um diameter) for reagents delivery and a polycarbonate

membrane sandwiched in between the PDMS layers. B. Micrograph

of the channels junction with the polycarbonate membrane (13 mm

diameter and 0.2 pm pore size) ...................................................... 89

xiii

Figure 18. A. Detailed representation of the microﬂuidic channels junction
with polycarbonate membrane in between. 81 represents the channel
under the membrane, where NO (3.8 uM stock solution in PBS) is
being hydrodynamically pumped at a ﬂow rate of luL min". 82 represents
the upper channel that poses as the CNS, where DAF-FM (10 nM) is
being introduced. B. Micrographs of the upper channel before (1),
at (2), and after (3) channels junction showing the formation of
the highly ﬂuorescent DAF-FM/NO adduct ........................................ 92

Figure 19. Detection of NO ﬂux through the polycarbonate membrane with a pore
size of 0.2 pm and at a ﬂow rate of 1 pL min—1 as assessed by DAF-F M
(10 nM). The control sample is PBS buffer. *, **, and *** represent
the statistical signiﬁcant differences versus control, 100, and
200 nM NO, respectively (*p 3 0.001; ** and ***p _<_ 0.05).
The error bars represent the mean values i standard error (n=4) ................. 93

Figure 20. Micrograph of bBMVEC immobilized in a 100 pm PDMS channel.
The channel was pre-coated with ﬁbronectin (100 pg mL") and incubated
for 1 h at 37°C and 5% C02 to ensure cell adhesion .............................. 96

Figure 21. Detection of the ATP ﬂux through a polycarbonate membrane with
pore size of 1.0 pm and 13 mm diameter as assessed by
chemiluminescence in a spiral-shaped channel, as part of a 3D
microﬂuidic device. ATP and luciferin/luciferase ﬂow rates are
3 uL min'1 and lpL min], respectively. Control represented by
RBCs 7% (0 uM ATP). Iloprost is used as an ATP release stimulant
of RBCs. * and ** represent the values statistically different from
control and 0.5 uM ATP, respectively (*p 5 0.001, **p S 0.05) (n=4). ...98

Figure 22. (a) Schematic of experimental set up for chemilmninescence
measurements taken to determine the iloprost-stimulated ATP
release from RBCs in real time shown in (b) ..................................... 112

Figure 23. Comparison to the controls of iloprost (1 nM) and RBCs alone as well
as the inhibition of ATP release when iloprost-stimulated RBCs were
incubated with glybenclamide (250 nM) (*p S 0.005) ........................... 114

xiv

Figure 24. Iloprost-derived ATP release from RBCs stimulates NO production
in a microﬂuidic device-containing an endothelium immobilized
to a polycarbonate membrane. In (a), the microﬂuidic device
contains a polycarbonate membrane sealed between two PDMS
layers that separates a channel containing RBCs (bottom layer) from
a layered endothelium (top layer). In (b), ﬂuorescence microscopy
images are shown for pre- (images on the top) and post-(images
on the bottom) addition of iloprost ................................................ 115

Figure 25. The increase in ﬂuorescence intensity due to endothelium-derived
NO production when a sample of RBCs incubated with iloprost
(23.0 d: 3.5) vs. a sample of RBCs alone (16.5 d: 1.0) is pumped
through the channel. The values reported are the average intensity
and standard deviation (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) .................. 117

Figure 26. Micrographs of a vascular endothelium array with blood components
are shown in (a). The wells in columns A and B contain bPAECs cultured
on a polycarbonate membrane. Column C contains cell free wells.
The rows represent the wells of the array that are addressable
by the underlying microchannel network, each delivering a different
reagent to the wells as described in the text. The quantitative
representation of the data is shown in (b). The difference in
ﬂuorescence intensity between samples containing the RBCs in
the absence (row 3) and presence (row 4) of iloprost are statistically
different from the other rows. There is also a difference in the
intensities of rows 3 and 4 (*p S 0.005) ........................................... 118

Figure 27. Representation of the mechanism of RBC-mediated regulation of
the vascular tone and the possible involvement of prolactin-releasing
peptide (PrRP) in scavenging the RBC-derived ATP. ATP binds to
the P2y purinergic receptors found on the endothelial cells that line up
the luminal side of the blood vessel and through a Ca2+-dependent mechanism
triggers the activation of the eNOS, and therefore production
of NO. NO activates the guanylate cyclase signaling cascade that
ultimately results in vasodilation ................................................... 132

XV

Figure 28.

Figure 29.

Figure 30.

Figure 31.

PrRP binding to ATP. Filled circles (0) represent the ATP

standards (0-1.5 uM) without the peptide. Open triangles (V), squares (c1)

and rhomboids (0) show the decrease in ATP signal after the incubation

with PrRP (10 nM) for 0, 30, and 60 min, respectively (n=3).

The decrease in chemiluminescence signal suggests that PrRP binds

to ATP (p 5 0.005) ................................................................... 133

Effect of PrRP (1 OnM) on the ATP release from RBCs treated with
Zn2+-activated C-peptide. The decrease in chemiluminescence

intensity corresponds to a decrease in measured ATP released from the

RBCs when PrRP was added to the metal activated C-peptide.

Filled bars correspond to the 90 min incubation with PrRP (n=5) and

the open bars correspond to 6 hrs incubation (n=8). The results are

reported as the average :1: standard error of the mean. The differences

in chemiluminescnence intensity between the samples incubated

in the presence and absence of PrRP are statistically signiﬁcant

(*p 5 0.005) ........................................................................... 135

The effect of the PrRP (10 nM) on the measured ATP release from

the RBCs treated with iloprost (10 nM) (n=3). All samples contain

RBC (ﬁnal hematocrit 7%). Control sample is represented by RBC 7%.

The iloprost-treated RBCs samples incubated with or without

PrRP are signiﬁcantly different (**p S 0.05) .................................... 137

Antagonistic effect of PrRP on the endothelium-derived NO production
in the bovine pulmonary artery endothelial cells (bPAECs). Picture I
shows micrographs of vascular endothelium array with blood
components. The rows a to f represent the wells of the array that are
addressable by the underlying microchannel network each delivering

a different reagent to the wells as described in the text. The quantitative
representation of the normalized data from 4 different experiments is
shown in II. The difference in ﬂuorescence intensity between

samples containing RBC + C-pep +Zn(II) incubated in the absence

(bar C) and presence (bar D) of the PrRP is statistically signiﬁcant

(*p < 0.05) There is also signiﬁcant difference in the intensities of rows
(bars E and F), where the RBCs + Iloprost samples were incubated

in the absence and the presence of PrRP ("p < 0.001), respectively.
Errors bars are the mean i standard error A: RBC7%; B: RBC + PrRP;
C: RBC + C-pep + Zn; D: RBC + C-pep + Zn + PrRP; E: RBC + Ilop;
F: RBC+Ilop+PrRP ................................................................ 138

xvi

CHAPTER 1

1.1 INTRODUCTION

1.1.1 Red Blood Cell (RBC) Structure and Properties

Red blood cells (RBCs) or erythrocytes (from Greek erythros for “red”, kytos for
“hollow” and cyte for “cell”) are the most abundant type of cells in the blood (4 to 5
million cells per uL), making up approximately 45% of the total blood volume.
Compared to a typical mammalian cell, the RBC is considered a dead cell because it has
no means of reproducing and repairing itself due to the lack of a nucleus and
mitochondria. The juvenile RBCs start their journey as reticulocytes in bone marrow.
These cells have a lifespan of about 4 days, spending 3 days at the site of their production
(bone marrow) and 1 in the peripheral blood. Once matured, the nucleus is extruded and
their ability to divide is lost. The RBCs short lifespan of 120 days and their primary
function of reversibly transporting oxygen and carbon dioxide by circulating in the
vessels throughout the entire body deﬁne the RBC cell philosophy of “living in a rush”.

I RBCs are smaller than most other human cells and under physiological conditions
they have the shape of biconcave disks of 6-8 pm in diameter, 2 pm in thickness at the
periphery and 1 pm at the center (Figure 1). This shape (as well as the loss of organelles
and nucleus) optimizes the cell for the important function of oxygen exchange with the

surroundings.

a) b)

Volume = 54 p3

 

Surface Area = 135 “z

 

Figure 1. a) Healthy RBC with its usual discocyte shape. b) Typical dimensions of a
healthy RBC.‘

This function is carried out by hemoglobin (Hb), the most abundant protein (> 90%)
present in RBC.As shown in Figure 2, the RBC is composed of a lipid membrane with an
underlying two-dimensional (10 nm in thickness) elastic network of spectrin.5 The
cytoskeleton is anchored to the lipid bilayer through proteins spanning the membrane.
The triangular structure of the spectrin network is presented in Figure 2. This particular
geometry, as well as the intrinsic elastic properties of the spectrin, gives remarkable
deformability to the RBC. This deformability is key property of the RBCs in their motion
through the complex network of the microcirculation. Their deformability is mainly
determined by the membrane surface area to volume ratio, cellular morphology,
mechanical properties of the cell membrane, and the viscosity of the cellular
components.6

The RBCs, with a mean diameter of 5 pm, are often subjected to a randomly
varying shear stress in the cardiac chamber and larger vessels. In addition, they must also
alter their shape to pass through the 3- to 4-um diameter capillaries of the
microcirculation as oxygen carriers. The ability of RBCs to deform and maintain the
oxygen-carrying capability under a randomly varying shear ﬂow is of great interest from
the standpoint of rheological and functional behavior of RBCs in the cardiovascular
system.7 The deformability of the RBC under shear stress has been studied extensively
during the last 30 years. The increased shear stress was suggested to be a major stimulus
for nitric oxide (NO), a small gaseous free radical that is released from the
endothelimnf"9 NO is an important signaling molecule and it is involved in many
physiological and pathological processes within the mammalian body, both beneﬁcial

and detrimental.

a)

Extracellular liquid J’J‘ i .1:

, > . i n ' .
l‘ iv H) Ii. I‘ ,5va ”I. M l !

   

Glycoprotein

 

Cytoskeleton Cholesterdl

Transmenibrane protein

b)

 

Figure 2. a) Composition of the membrane of a RBC. b) Image of the spectrin
cytoskeleton.

1.1.2 RBC-Derived ATP: An Important Factor in Vascular Regulation

In 1991, Ignarro’s group described an in vitro bioassay system to study
endothelium-mediated, shear stress-induced, or ﬂow dependent generation of
endothelium-derived relaxing factor (EDRF), which had just recently been identiﬁed as
No.9 It was generally accepted that NO was produced by the endothelial cells and
relaxed vascular smooth muscles. '0

Later, in 1995, Sprague et al. reported that in order to demonstrate ﬂow-induced
endogenous NO synthesis in the pulmonary circulation of an isolated perfussed rabbit
lung, RBCs obtained from either rabbits or healthy humans were a required component of
the perfusate.11 To strengthen their claim, they showed that in the absence of the RBCs,
ﬂow-induced increases in shear stress did not stimulate NO synthesis. Moreover, they
concluded that an additional key factor must be existing in the rabbit pulmonary
circulation that relates RBCs to the vasorelaxation. Therefore, they proposed that the
property of these RBCs responsible for the stimulation of NO synthesis was their ability
to release adenosine triphosphate (ATP) in response to mechanical deformation.12

This hypothesis was based on multiple ﬁndings, the ﬁrst of which being that ATP
is an in vivo stimulus of endothelial NO production.”’14 Secondly, ATP is of particular
l5,l6

interest because it is present in millimolar concentrations in the RBCs and it is

released in response to physiological stimuli, such as hypoxia and hypercapnian‘l8
Additionally, when RBCs are traversing the resistant vessels in vivo, they become

deformed.19 Another factor they considered was that ATP reduces vascular resistance in

isolated lungs. Moreover, the latter action of ATP was prevented by the NO synthesis

inhibitor NG-nitro-L-arginine methyl ester (L-NAME).20 It has been shown that as their
deformability increases, erythrocytes of rabbits and humans can release increased
amounts of ATP. This RBC-derived ATP can then act on the endothelial cells to
stimulate endogenous NO synthesis and enable RBCs to participate in local regulation of
vascular caliber. I 1,12,13,2r-24

Depending on the type of purinergic receptor to which ATP binds, it can induce
antagonistic effects on the vasculature. For example, ATP binding to the P2X receptor
found on the vascular smooth muscle cells results in contraction of the cells,25 whereas
ATP binding to the P2Y receptor found primarily on the endothelium, results in NO
synthesis and/or vasodilation.“l4 Thus, ATP applied to the luminal side of a vessel, that
released from the RBCs within the circulation, would be expected to produce
endothelium-dependent relaxation through an interaction with the endothelial P2Y
purinergic receptor (Figure 3).

All these ﬁndings suggest that RBCs may be an important determinant in the
stimulation of NO production in the circulatory system. Therefore, as RBCs are
traversing the circulatory system they are increasingly deformed by increments in the
velocity of blood ﬂow through a vessel and/or by reduction in vascular caliber.
Moreover, RBCs release ATP which stimulates endothelial synthesis of NO, resulting in
relaxation of the vascular smooth muscle and thereby, an increase in vascular caliber.26

On the other hand, failure of this mechanism for deforrnation-induced ATP
release has been associated with primary pulmonary hypertension (PPH), cystic ﬁbrosis

(CF) and alveolar hypoxia. It has been shown that the RBCs of CF patients do not release

ATP in response to mechanical deformation and CF patients often develop pulmonary

STRESS

11M.
<1?) ®
® ATPATP 1 “Pl ATP A" No memo“...

ATP

ATP ATP ATP ATP ATP
Luminal side PZYI P2Y”

      
    
   
   
 
   
 

ATP 1,. ”0

Endothelial Cells

 

Smooth Muscle Cells

cGMP ‘9 Relaxatio

Extraluminal side

Figure 3. Proposed mechanism for the effect of RBC-derived ATP on the cells at the
luminal and extraluminal side of a blood vessel.

hypertension.26 As a result of these ﬁndings, it has been suggested that the cystic ﬁbrosis
transmembrane regulator (CFTR) is a component of the pathway that links mechanical
deformation of the RBC to the release of ATP.

1.1.3 CFTR-Facilitated ATP Transport/Release out of the Cell

CFTR is CAMP-activated chloride channel expressed in epithelial cells that is
responsible for transepithelial salt and water transport.”29 Mutations in this protein are
responsible for the hallmark defective chloride secretion observed in CF. CFTR regulates

0

other transporters, including Cl'/(HC03') exchangers,3 epithelial sodium channel

(ENaC),3"33 outwardly rectifying chloride channel (ORCC),34'36 aquaporin water
channels,37 and ATP release mechanisms.3’38’39

CFTR belongs to the ABC (ATP binding cassette) transporter family and is
activated both by phosphorylation of its R-domain by PKA (protein kinase A) as well as
ATP binding at its nucleotide binding domains, namely NBDl. CFTR is capable of
hydrolyzing ATP, although this activity is not coupled with the ion transport; rather it
may serve to initiate or modulate conformational changes that underlie channel opening
or closing.

There exists a vast body of work that shows that CAMP-stimulated ATP release is
dependent on CFTR, that expression of CFTR is associated with ATP conductance, and
that there are multiple interpretations of the ATP transport phenomenon.3638’”43 In these
previous studies, the precise route of the ATP through the membrane was not deﬁned;

however, ATP transfer across the membrane occurred only in the presence of CFTR and

mdr (multidrug resistance gene). It is clear that if CFTR can facilitate ATP release, CFTR

itself must be modulated by regulators or stimuli in a concerted manner or CFTR itself

must modulate an ATP release pathway that is expressed universally (Figure 4).3

1.1.4 RBC Deformability and Disease Pathogenesis

Abnormal RBC deformability has been also associated with such diseases as
diabetes and multiple sclerosis.“45 For example, it has been shown that deformability is
decreased in RBCs of patients with type I and type II diabetes. In addition, the whole
blood of patients with MS was found to be more viscoelastic when compared with the
whole blood of healthy patients.”48

Interestingly, it is known that the RBCs of MS patients have abnormally high
levels of glutathione (GSH), the most abundant non-enzymatic antioxidant present in a
cell. In 1979 Szeinberg et al showed that the activity of glutathione peroxidase (GSH-Px)
in erythrocytes of patients with MS was signiﬁcantly lower than in a control group of 30
patients with various neurological disorders.49 Therefore, they were conﬁrming a similar
ﬁnding of a decrease in activity of GSH-Px in erythrocytes of a group of MS patients in
Denmark.50

In the next two decades, an important body of work was published in regard to the
abnormal activity of this enzyme in MS RBCs.49’5 1'5 5 Glutathione peroxidase catalyses
the reaction between the reduced monomeric glutathione (GSH) and hydrogen peroxide
with the formation of the glutathione disulﬁde (GS-SG). Glutathione reductase then

reduces the oxidized glutathione to complete the cycle. If the activity of the GSH-Px is

reduced, an increase in the ratio of GSH/GSSG will result, thus altering the redox status

A

ﬁﬁlﬁﬂ
lllll

Cytoplasmic

      

ﬂ ﬂ 3 Extracellular
Ill

 

 

illli
Milli

Cytoplasmic

Extracellular

ﬁﬁﬂll
lllll

Figure 4. Proposed mechanisms for CFTR-facilitated ATP transport/release out of cell.
For hypothesis A to occur, CFTR must be expressed in concert with
appropriate cofactors or regulators. Hypothesis B and C suggest positive
CFTR regulation of a separate and independent ATP transport mechanism and
CFTR-facilitated fussion of ATP-ﬁlled vesicles, respectively.

10

in the cell. The levels of GSH have been linked with the deformability characteristics of
the RBCs. That is, low levels of GSH are found in diabetic red cells, which are known to
be less deformable than the RBCs from healthy patients.5(”57 Moreover, recent studies
inour group involving diabetic RBCs have shown that a weakened oxidant defense
system in RBCs will lead to lower levels of ATP release.58 We demonstrated that RBCs-
derived ATP is affected by oxidant insults and that the recovery from this insult is
independent of the pentose phosphate pathway, the pathway that maintains the
concentrations of GSH. Moreover, in this study we showed that erythrocytes from
individuals with type II diabetes released 50% less ATP than the erythrocytes from
healthy, non-diabetic patients. As a consequence of decreased levels of ATP release from
RBCs, less endothelium-derived NO may be available for the regulation of the vascular
caliber. Our group also reported that there is a direct relationship between the ability of
the extracellular ATP to stimulate NO in endothelial cells and diabetic complications.58

If, in the case of MS patients, the RBCs are more deformable and the increase
deformability results in an increase of ATP release, then an increase in the NO
production would be expected. Indeed, it is recognized that there exist higher than normal
levels of NO metabolites, nitrate and nitrite, in the cerebral spinal ﬂuid (CSF) and serum
of patients with MS.“‘59 Taken collectively, we have constructed a theory that perhaps
explains the overabundance of NO metabolites in the CSF is actually blood-bome; in
other words, it is possible that the NO levels are increased due to overstimulation by
RBC-derived ATP.

Our hypothesis is that MS-RBCs are more deformable, therefore releasing more

ATP, which in turn will lead to increased production of NO. However, a molecular level

11

understanding of the mechanisms by which RBC deformability properties relate to the
occurrence of MS is lacking. To prove or disprove this hypothesis, the role of RBC-

derived ATP and NO derived from brain endothelium must be investigated.

1.1.5 Nitric Oxide (N O) and the Blood-Brain Barrier (BBB)

During the past decade NO has emerged as an important mediator of
physiological and pathophysiological processes. Elevated NO biosynthesis has been
associated with nonspeciﬁc immune-mediated cellular cytotoxicity and the pathogenesis
of chronic, inﬂammatory autoimmune diseases including rheumatoid arthritis, insulin-
dependent diabetes, inﬂammatory bowel disease, and MS. Recent evidence suggests,
however, that NO is also immunoregulatory and suppresses the function of activated
proinﬂarnmatory macrophages and T lymphocytes involved in this diseases.60

NO is a small gaseous radical that is biosynthesized from the guanidino nitrogen
atom of L-arginine within the active site of NO synthases (NOS), which in mammals are
incoded by three genes corresponding to neuronal (nNOS), inducible (iNOS), and
endothelial (eNOS) isoforms.6| These isoforrns have been classiﬁed on the basis of their
constitutive (eNOS and nNOS) versus inducible (iNOS) expression, and their calcium-
dependence (eNOS and nNOS) or calcium-independence (iNOS). They share an overall
amino acid sequence homology of approximately 55%, and there is strong sequence
conservation in the regions of the proteins that are important in the catalysis.62

The NOS isoforrns carry out a ﬁve-electron reduction process utilizing oxygen,

and cofactors such as reduced NADPH, tetrahydrobiopterin (BH4), ﬂavin adenine

12

dinucleotide (FAD), ﬂavin mononucleotide (FMN) and iron protoporphyrin IX (heme)
participate in the formation of NO (Figure 5). L-Arginine in the presence of NADPH and
02 is oxidized to N-hydroxyarginine, which is reoxidized to citrulline producing NO.63

Although NO has many biological fates, one possible scenario is that it crosses
the blood brain barrier (BBB) into the central nervous system (CNS), where it may result
in a breakdown of the myelin sheath, compromise the integrity of the BBB, and lead to
axonal degeneration.64 The BBB is comprised of the endothelial cells that line the
capillaries of the brain. The distinguishing features of the brain capillary endothelium
include the presence of tight intercellular junctions.65 The existence of these junctions
between the endothelial cells forms a very selective ﬁlter for the blood-home
molecules.66’67 Therefore, the passive diffusion of the compounds across the BBB is
dependent upon their physicochemical properties, such as lipid solubility, molecular
weight, electrical charge, or extent of ionization.

Bradbury et al showed that lipid soluble substances penetrated the cerebral
endothelial plasma membranes readily and also equilibrated easily between blood and
brain tissue.68 However, water soluble molecules, such as amino acids, nucleosides, and
hexoses, and macromolecules are transported in the brain by means of selective carrier
mechanisms or speciﬁc transporters.65

The importance of the barrier arises from the fact that it restricts the movement of
speciﬁc substances, such as infectious organisms from the circulatory system, to the brain
tissue and extracellular ﬂuid. In the CNS, the endothelial cell membrane is further

surrounded by the end feet of the astrocytes outside the capillary wall, which inﬂuences

and conserves the barrier function of the cerebral endothelial cells.

13

    

Citrulllne
NO (or NO')

Figure 5. Electrons (e-) are donated by NADPH to the reductase domain of the enzyme
(eNOS) and proceed via FAD and FMN redox carriers to the oxygenase
domain. There they interact with the haem iron and BH4 at the active site to
catalyze the reaction of oxygen with L-arginine, generating citrulline and NO
as products. Electron ﬂow through the reductase domain requires the presence
of bound calcium-calmodulin (CaZJ'ICaM).6|

It is well established that there are many pathophysiological conditions that
disrupt the integrity of the BBB and can lead to the increased permeation of substances
across the BBB.69 For example bacterial meningitis and MS are two of the conditions that
can alter the BBB and allow normally excluded substances to enter into the cerebrospinal
and interstitial ﬂuids. Moreover, when inﬂammation occurs (as in MS or encephalitis) a
massive leucocyte migration into the brain takes place.”71 Cytokines (signaling proteins
and glycoproteins used in cellular communication), secreted in the early steps of
inﬂammation, are found to increase the permeability of the barrier, therefore permitting
compounds to enter the brain by opening the BBB.72

Other inﬂammatory mediators, such as eicosanoids (e.g, prostaglandins), adhesion
molecules, and free radicals can play an important role in the permeability of the BBB.73
Of the free radicals, reactive oxygen species (ROS) can cause extensive damage to the
membrane lipids in the CNS by peroxidating the fatty acids, thus destroying the myelin
and cell membranes.”75

One of the free radicals that can damage the BBB is NO. Endothelial cells of the
BBB are known to possess the inducible form of NOS. When produced in excessive
amounts, usually as part of a biochemical cascade stimulated by injury or inﬂammatory
conditions, NO production may no longer be beneﬁcial for the cells; instead, the large
quantities of NO become neurotoxic.76

As mentioned above, elevated levels of nitrate and nitrite have been reported in
cerebral spinal ﬂuid (CSF), urine and serum of MS patients.44 Particularly, nitrate and

nitrite levels were signiﬁcantly higher in CSF compared to serum,77 which could be the

consequence of BBB dysfunction mediated by No.78 Although nitrate and nitrite are

15

elevated in CSF, there is no signiﬁcant correlation between their levels and clinical
activity,77 possibly because the origin of this excess NO and its mechanism of action are
unknown.

It is well established that endothelial cells can produce NO in small amounts and,
when released, are able to control the local blood ﬂow through the activation of the
guanylate cyclase signaling cascade that ultimately results in vasodilation. It has also
been previously shown that the mechanism by which NO is released into the circulatory
system involves ATP, and that this ATP is released from erythrocytes that are subjected
to mechanical deforrnation.'2’79 Moreover, it was reported that as the deformability
increases, erythrocytes of rabbits and humans can release higher amounts of ATP.”’80
Collectively, these ﬁndings involving the ability of RBCs to produce ATP and the
increased deformability of RBCs from MS patients suggest a possible mechanism for the
increased levels of NO found in the CSF of MS patients.

However, a molecular level understanding of the mechanisms by which the RBCs
deformability properties relate to the occurrence of MS is lacking. To gain insight into
these mechanisms, it would be ideal to employ an in vitro model that mimics an in vivo
system. A possible approach to developing such a system would be a microfabricated
ﬂuidic approach.

In this work we propose to create a system that would allow mechanically

deformed RBCs to release ATP, stimulate the endothelium—derived NO and that would

allow one to monitor the fate of that NO in a dynamic fashion.

16

1.2 MICROFLUIDIC TECHNOLOGY

1.2.1 Short History

The history of analytical miniaturized devices starts with the remarkable work of
Terry in 1979.8' Even though, at the time he was not acknowledged for his work, the gas
chromatograph fabricated on silicon, he was presenting, was able to separate a simple
mixture of compounds in a matter of seconds. Terry was bold enough to take the existing
microfabrication technology used in microelectronics and successfully apply it to study
basic scientiﬁc phenomena that occur at small dimensions. The device required only
small quantities of reagents, had a relatively short analysis time and showed efficiency in
separation that were better than larger counterparts. One of the other earliest examples of
reported silicon micromachined devices was a coulorimetric acid-base titration system,
which employed a solid state pH-sensitive sensor to determine the acid or base
concentration of a solution.82

In the early 1980’s, Michael Widmer proposed the concept of “miniaturized total
chemical analysis systems” or uTAS, in which silicon chip analyzers incorporated
sample pretreatment, separation, and detection.83 The complete analysis (sampling,
sample transport, any necessary chemical reactions, separations, detection) were
incorporated into a ﬂow system. Due to inclusion of sample pretreatment, the detector no
longer needs to be as selective as a sensor, because interfering compounds are eliminated,

masked or separated from compound of interest.

17

At the beginning, the analytical performance of these devices was more desirable
than their reduced size. However, miniaturization was the route to shorter analysis time
and also versatility. The central technology for a number of miniaturized systems was the
microﬂuidic technology, where liquids and gases were manipulated in channels with

cross sectional dimensions on the order of 10-100 pm.

1.2.2 Scaling Effects when Miniaturizing the Chemistry

As the dimensions of a system change, the relative performance of a system and
the operational success of miniaturized chemical separation, reaction and detection
devices can be predicted by the scaling laws. During the last two decades, microﬂuidics,
micrometer-scale total analysis systems (pTAS) or so called “lab-on-a-chip” devices
have required consideration of scaling laws and dimentionless parameters for
downscaling purposes.83

There exist two main approaches to be considered when miniaturizing a device:
the dimensionless-parameter approach and the similarity approach. In the dimensionless
approach, parameters such as volume, column length, linear ﬂow rate, retention time and
pressure drop can be assumed to be constant over the entire system. The Reynolds
number (Re), used to characterize laminar and turbulent regimes is a particularly useful
dimensionless parameter.

Re = v-p d/r] (l)
where v represents the average velocity , p the ﬂuid density, 77 is the dynamic viscosity

and d is the characteristic length, such as the diameter of the capillary. Because the

18

Reynolds number (Re) is proportional with the characteristic length d, for miniaturized
systems, a small volume will be obtained, indicating laminar ﬂow.

On the other hand, the approach that considers proportionality provides helpful
information related to the behavior of a simple ﬂow system when miniaturized. If
miniaturization is assumed as a downscaling process in three dimensions (represented by
a typical length, d) the behavior of the physical variables of interest can be predicted.84 In
a system, where the timescale is the same for the miniaturized system as for the full-scale
system, the parameters that would change are: linear ﬂow rate that in a tube would
decrease by a factor of d, volumetric ﬂow rate by a factor of d3, and the Reynolds number
by d2.

In miniattuized systems, diffusion has a more dominant role in mass ﬂow and
becomes more important when molecular diffusion, heat diffusion or ﬂow characteristics
dictate the separation efﬁciency of the system. Additionally, in a difﬁrsion-controlled
system, the time is regarded as a surface and is proportional with d2. This system is in
accordance with the band-broadening theory in chromatography and electrophoresis. So,
if a system is downscaled by 1/ 10, the time variables decrease by 1/ 100, pressure
increases by a factor of 100, and voltage requirements remain constant, but the essential
chemistry and separation behavior retain the same quality.85

Miniaturization can exploit the particle Brownian motion to efﬁciently mix
solutions. In small structures, the mixing by diffusion is enhanced greatly because the
time a molecule needs to travel a distance d decreases as I/dz. Because of the low
Reynolds numbers, therefore no turbulence in the micrometer regime, much effort is

focused on improving the mixing capabilities. By using herringbone mixers, Stroock et

19

al.86 demonstrated that the mixing process can be enhanced by chaotic advection. Also,

Song et al showed that by using a two-phase ﬂow, mixing in droplets was possible.87

1.2.3 Materials for Microsystems

Microfabrication revolutionized the modern science and technology. Although

microfabrication has its basis in microelectronics applications in other areas rapidly

88-91 92,93

emerged. These include systems for microalnalysis, micro-volume reactors,

4 95,96

combinatorial synthesis,9 microelectromechanical systems (MEMS), and optical
components.97

One particularly exciting use for microanalytical devices is in the separation and
analysis of chemical and biological substances. In addition, microstructures used in
MEMS have evolved to integrate electronics with monitoring, actuating, and controlling
tools for use in engineering. Most of the devices categorized as MEMS are fabricated
from silicon using standard microlithographic techniques (silicon bulk micromachining
and polysilicon surface micromachining).98

Microsystems can be built on various substrates with a range of materials, from
crystalline silicon to amorphous silicon, glass, quartz, metals, and organic polymers. Each
of these materials posses their advantages and disadvantages depending on the type of
application.

Silicon processing itself can be carried out using thin ﬁlms of organic photoresists

for etch patterning and this technique is very well developed. Moreover, two and three

dimensional shapes can be reproduced with high precision.98 Because similar techniques

20

used in electronic integrated circuits are applied, silicon microdevices can be batch
fabricated. Also, due to the electronics industry, silicon is a readily available material.
Chemically, one of the most attractive features of silicon and silicon dioxide is the
chemical and thermal stability.

However, there are also some disadvantages that limit the spectrum of utilization
of these materials. Crystalline silicon is relatively expensive, brittle, and opaque in the
UV/visible regions. In contrast to glass or plastic, silicon will conduct current limiting its
usefulness for electrokinetic devices. Because of these drawbacks and also complicated
surface chemistry, alternative materials to silicon have been used to fabricate
microsystems. Such materials are glass, quartz and some rigid organic polymers (e.g.,
epoxy, polyurethane, polyimide, polystyrene, and polymethylmethacrylate).99’100

These materials have properties that make them useful for fabrication of
microsystems. More speciﬁcally, they are transparent in the visible and UV regions,
therefore they can be adapted to optical detection in microanalytical systems. Also, the
surface properties of these substrates (e.g., wetting, adhesion, adsorption, reactivity) can
be modiﬁed using a variety of surface chemistries. Microstructures of these materials can
be replicated inexpensively by molding or embossing.”104

Two of the most actively developed polymers for microﬂuidics and other
applications are poly(dimethylsiloxane) (PDMS) and poly(methylmethacrylate)
(PMMA).'°5"07 Elastomeric materials, such as PDMS, gained popularity mostly because

of their optical properties. However aside the fact that PDMS is optically transparent to ~

300 nm,108 it can deform reversibly and repeatedly without permanent distortion,109 it can

21

be molded for feature sizes of 0.1 to 10 pm,”0 it is durable and chemically inert, non-
toxic, commercially available and inexpensive.

When selecting the material for a device, one of the most important features is its
surface chemistry. PDMS, for example, is highly hydrophobic material, comprised of
repeating units of —O-Si(CH3)2-. This hydrophobicity makes the channels less wettable
with aqueous solutions and is prone to adsorption of hydrophobic species. However, by
exposing the PDMS to plasma oxidation, its surface will be rendered more hydrophilic by
the formation of silanol groups (Si-OH) at the expense of methyl groups (Si-CH3).”"”3
These polar groups can then condense with appropriate polar groups on a different
surface, and by bringing the two surfaces in conformal contact will result in the formation
of covalent bonds that represent the basis of irreversible sealing. For PDMS and glass, for
example, these bonds are Si-O-Si after the loss of water.

There are also other parameters of the material that have a critical role in the
success of the application of microﬂuidic devices. Such parameters include
autoﬂuorescence (for devices with optical detection), permeability (when using living
cells), chemical resistance (when using non-aqueous solutions) and analyte adsorption.l '4

Heat dissipation, characterized by the thermal conductivity, in a substrate material
is especially important when considering electroosmotic pumping, which is the most
common method to propagate ﬂow in microﬂuidic systems. In this case, a device or
channels made of/in ﬁrsed quartz will withstand higher localized heating compared to any
plastic, and the efﬁciency of the separation will not be compromised. Moreover, the

electroosmotic ﬂow (EOF) in channels made of different polymer materials is highly

variable due to the wide range of different charge and charge densities. The

22

electroosmotic ﬂow is generated by the surface charge on the microchannel walls in
combination with an electric ﬁeld along the microchannel.

All of these parameters have been extensively evaluated and characterized for the
wide range of applications of the microﬂuidic devices such as: miniaturized analytical

systems, biomedical devices, tools for chemistry, biochemistry and medical applications.

1.2.4 Fabrication of Microfluidic Systems

One of the most commonly used techniques for the fabrication of microﬂuidic
systems is photolithography.”5’”6 Here, a substrate, most often made of silicon, is spin-
coated with a thin layer of photoresist (photosensitive polymer) and exposed to a UV
light source through a photomask. The photoresist, exposed to UV light through a quartz
plate covered with patterned microstructures, becomes either more (positive resist) or less
(negative resist) soluble in a developing solution. Therefore, the pattern of the photomask
is transferred on the ﬁlm of photoresist. Finally, the substrate holding the desired pattern
can serve as a master to make PDMS stamps.

Spin coating involves the acceleration of the coating material, which is deposited
in the center of the substrate that is being rotated. The process of the ﬁlm forming is
driven by two independent parameters, namely viscosity (7]) and rotational speed (w).
Other variable process parameters involved are the solid content and the speed time.
Generally, the rotational speeds for spin coating are reported to be between 1500 and
8000 rpm, this giving a resist ﬁlm uniformity 3c 5 nm from substrate to substrate with a

typical thickness of 0.8-3 pmm

23

The photoresist processing consists of six steps illustrated in Figure 6 A) such as
dehydration or chemical clean and priming, resist coating, soft baking, exposure,

”8 The dehydration/chemical cleaning

development, and post-development inspection.
step is performed to eliminate any moisture or any traces of contaminants (organic, ionic
or metallic) adsorbed by the substrate surface and it is followed by the deposition of a
silicon dioxide (SiOz) which serves as a barrier layer. The next step is the spin coating of
the resist onto the wafer surface. Depending on the type of application and/or the
thickness desired, the speed can be calculated according to the formula:

2 = k Pz/wm (2)
where z is the ﬁlm thickness; k is a constant; P is the percentage of solids; and w is the
rotational speed.

A soft baking step is performed usually between 90-95°C using either a
convection oven, IR source, or hot plate. This step is important for the evaporation of
solvent on the spun-on resist, to improve the adhesion of the resist to the wafer, and
anneal the shear stresses induced during the spin coating. After spin coating and baking,
the wafer must be exposed to UV light that will produce the pattern image on the resist.
Depending on the type of resist used, the irradiated regions on the resist will become
more or less soluble in the developer, therefore, giving a positive image or a negative
image of the mask on the wafer. The development process involves chemical reactions
where the unexposed parts of the resist get dissolved in the developer.1 '8

A widely used technique for microfabrication is soft lithography. This technique
belongs to a class of non-photolithographic techniques that are based on printing of self-

119

assembled monolayers (SAMS) and molding of organic polymers. Soft lithographic

24

A a Silicon wafer
+— Oxidation, SiOz

.. '. _. - . ' photoresist

 

 

 

  

 

 

 

 

   

Photoresist development

 

Master

 

 

 

 

 

 

 

 

 

 

PDMS Peel PDMS from master

 

 

 

Seal against a ﬂat surface

 

PDMS / Microchannel

 

 

 

 

 

Figure 6. A) Representation of the steps in the lithographic process. B) Scheme of rapid
prototyping.

25

techniques include microcontact printing,120 micromolding in capillaries,l2l microtransfer

104,110

molding, replica molding, embossing,'02 and injection molding. Replica molding is

the easiest, more economic, and more amenable technique for analytical or biological
applications. Here, microstructures are directly formed by casting and curing a UV
curable polymer against an elastomeric mold. Feature sizes ranging from millimeters to
approximately 3 nm can be replicated with high ﬁdelity.

Another way of creating pattern structures on the submicrometer scale is the
micro contact printing technique (j.tCP).'22 Using this technique, SEMS can be patterned
in a simple way, allowing the fabrication of multi-array biosensors.

There are several other methods for microfabrication that are gaining more and

102

more popularity, for example: embossing (or imprinting), injection molding,”l laser

123 124

ablation, electrochemical micromachining (EMM), and ultrasonic machining.125

1.2.5 Tools for Microfluidic Applications

Microﬂuidic devices can be used for a wide range of applications, from obtaining

126,127

molecular diffusion coefﬁcients, to measurements of the ﬂuid viscosity,128 to

129,130 131-133

chemical binding coefﬁcients, and enzyme reaction kinetics.

Other applications of microﬂuidic devices include capillary electrophoresis,'34

130,135

isoelectric focusing, immunoassays,'36'I39 ﬂow cytometry,140 sample injection of

proteins for analysis via mass spectrometry,"“"43 PCR ampliﬁcation,”""147 DNA

148-150 153,154

anal sis, cell mani ulation,ISl cell se aration,152 cell attemin , and chemical
y P P P g

gradient formation. 1 55" 56

26

Many of these applications have utility for clinical diagnosticsm’158 Because of
the small size of the channels, from micro- to nanometers, the amount of reagents and
analytes is quite small, hence, microﬂuidic devices present a signiﬁcant advantage to
conduct biomedical research and to create clinically useﬁrl technologies. One of the long
term goals in the ﬁeld of microﬂuidics is to create integrated portable clinical diagnostic
devices for home and bedside use, thereby eliminating time consuming laboratory
analysis procedures.

In an effort to achieve this goal, recent microﬂuidic devices have undergone the
transition from simple to highly integrated systems. Individual microﬂuidic components
such as sample preparation, analysis and detection can be integrated on a single chip.
Higher level of integration has been enabled by components such as microvalves,159
micropumps,160 ﬂow sensors and other physical sensors, ﬂow selectors, micromixers,4
microreactors, extractors, droplet generator, ﬁlters, traps, and sieves.

One of the most important components for the fabrication of a fully integrated

8,83’84’161'163 were considered to be the

microﬂuidic system, such as lab-on-a chip or uTA
microvalves and micropurnps. Valves and pumps provide direct control of the ﬂow and
therefore are essential elements of a system that handles ﬂuids at small scale.
Microvalves are essential components in many devices for portable chemical
analysis, drug delivery, mixing, dosing and they can be roughly classiﬁed in two groups:
passive valves (work without an actuator) and active valves (work with an actuator) that
use mechanical, non-mechanical and external moving parts as shown in Table 1.

Microvalves have a number of advantages over traditional valves. For example, because

of their small size they do not require excessive power consumption, they have no

27

 

Categories

 

Active Mechanical
Non-mechanical
External
Passive .
Mechanical

Non-mechanical

Magnetic

Electric

Piezoelectric
Thermal

Bistable

Electrochemical
Phase change

Rheological

Modular

Pneumatic

Check valve

Capillary

External magnetic ﬁelds
Integrated magnetic inductors
Electrostatic

Electrokinetic

Bimetallic
Thermopneurnatic
Shape memory alloy

Hydrogel
Sol-gel
Paraffin

Electro-rheological
F erroﬂuids
Built-in

Rotary

Membrane

In-line

Flap

Membrane

Ball

In-line mobile structure
Diffuser

Abrupt

Liquid triggered

Burst

Hydrophobic valve

 

Table 1. Classiﬁcation of microvalves.

28

leakage and result in minimal dead volume. Some of their ideal characteristics are their
inﬁnite differential pressure capability, insensitivity to particular contamination, zero
response time, potential for linear operation and ability to operate with liquids and
gases.'64
For achieving ﬂow rates controllable over a certain range (0.25 to 10 uL min")
and with as little pulsation of possible, micropumps were the best choice. Current
micropumps are roughly divided into two groups namely, reciprocating micropumps and
continuous flow micropumps. Reciprocating micropumps use the oscillatory or the
rotational movement of the mechanical parts to displace the ﬂuid, while continuous ﬂow
micropumps rely on the direct transformation of nonmechanical or mechanical energy
into a continuous ﬂuid movement. '60
Beside microvalves and micropumps, micromixers are another important component in
a microﬂuidic system. Rapid mixing is essential in many of the microﬂuidic systems used
in biochemical analysis, drug delivery and sequencing or synthesis of nucleic acids. Due
to low Re numbers, in microﬂuidic devices the mixing of different ﬂowing streams rely
only on molecular diffusion. However, if the application requires fast mixing and
analysis, the mixing can be induced by using input of energy from the exterior, termed
active mixing. Some examples of external energy inputs are ultrasound,165 acoustic,

166,167

bubble-induced vibrations, electrokinetic instabilities,168 periodic variation of ﬂow

169,170 172

rate, electrowetting-merging of droplets}.’1 piezoelectric vibrating membranes,

73

magneto-hydrodynamic action,l small impellers, integrated microvalves/pumpsI74

(Figure 7).

29

A B C
Side J

hdlde‘aq’ E E g E
’ a -

. ,‘ r t—z-l-"‘_ r

W \ - -

 
     

 

’1l2cyclo: "Tm

Figure 7. Types of micromixing. A. Cyclone micromixer ; B. Hydrodynamic focusing by
compressing a central stream by two outer streams at much larger ﬂow rate; C.
Multi-lamination ﬂow patterns of a split-and-recombine mixer visualized by a
dilution-type experiment; D. Staggered herringbone micromixer for the
generation of chaotic ﬂows; E. Time lapse images of three-electrode mixing at
8 Hz; F. Electrokinetic instability micromixing; G. Numerical simulation
results obtained for two inlet ﬂows pulsed at 180° phase difference; H.
Photographs showing acoustic microstreaming in a microcharnber 4

30

1.2.6 Detection Modes in Microfluidic Devices

In order to quantitatively and qualitatively characterize an analyte of interest, the
microsystem must be equipped with a high performance detection scheme. Reduction of
the size of the microchannels makes the difﬁrsion a very fast process and therefore,
separation techniques on microchips, such as chromatography, electrophoresis,
electrochromatography and separations in immunochemical assays are more
advantageous than the conventional systems.‘75 There are three major types of detection
that should be mentioned: optical detection, electrochemical and mass spectrometry.
Optical detection tipically comprises ﬂuorescence, absorbance, chemiluminescence, and
refractive index.

Fluorescence detection is the most common detection mode for microchip devices
because of its versatility and also sensitivity (picomolar detection limits). The most
sensitive < 1013 M) small volume detection method present to date is the laser-induced
ﬂuorescence (LIF).I78 In fact, single molecule detection has been demonstrated in a
capillary using LIF .179 LIF typically requires off-chip derivatization of the sample which
increases the time of the total analysis. However, because the detection of native
ﬂuorescence has disadvantages, such as weak signal and high background ﬂuorescence
(poor sensitivity), LIF has been considered a more attractive choice. Fluorescence
detection is used not only for single point detection but also to image a large area, such as
monitoring of the ﬂuid ﬂow. For applications involving parallel analysis (monitoring

multiple channels), laser scanning techniques can be utilized.'80 Also, organic light

31

emitting diodes (LEDs)18' and ﬁber optics were considered successful ways of integrated
detection methods for on-chip analysis.

For microseparations, and particularly for capillary electrophoresis, the most
common method of detection used is absorbance, and this is because most organic
compounds can be detected at 195-210 nm. However, due to the small internal diameter
of the microchannels, the path length dependency of absorbance detection makes the
detection of solutes with poor absorptivity very difﬁcult]82 Other optical detection modes
include chemiluminescence, where no external light source is needed and background
signal is minimal, and refractive index detection, which is useful for compounds with
poor absorbance/ﬂuorescence (270 nM LCD for unlabelled proteins).183

Another very sensitive and selective method of detection that has been widely
used in microdevices is electrochemical detection (EC). Electrochemical detection has
the advantage of low cost, relatively easy operation and no need for labeling, required in
LIF There are three types of electrochemical detection: potentiometry (ion selective
electrode, measures voltage), amperometry (measures current), and conductivity
(measures voltage/current). Potentiometric detectors are based on classical ion selective
microelectrodes (ISE) and have the ability to detect very small quantities of organic and
inorganic ions in small probe volumes. The ﬁrst example of potentiometric detection on

l.'84 This study described a micromachined ISE chip for

chip was reported by Manz et a
Ba+ sensing.
The mechanism for amperometry relies on the electron transfer to/from

electroactive analyte at solid electrode (DC voltage applied) and the resultant current is

proportional with the analyte concentration. When amperometric detection is applied for

32

conventional CE, there are some issues that have to be overcome. For example, the
electrophoretic current is usually up to 106 larger than the detector current, thus the
detector has to be decoupled from the high voltage. That was accomplished by using a
ﬁne crack in the capillary as a decoupler. Therefore, electroosmotic ﬂow (EOF)
generation stops at crack and the pressure form EOF pushes the analyte to the electrode.
For the on chip integration there have been different attempts to align the electrodes with
the decoupler. Mathies et al.185 reported a device where the decoupling from high
potential was accomplished by placing the working electrode (10 um wide) 30 um
beyond the end of the separation channel in 1 mm wide exit channel. They used the
device for the detection of neurotransmitters with very good limits of detection (LCD);
for example, dopamine (66 atomol in 18 pL), epinephrine (6.5 nM) and catechol (12
nM). However, there are major drawbacks of these amperometric techniques, one of
them being the strong absorption of the analyte reaction intermediates to the surface of
the electrodes (carbon electrodes). Some researchers got around this problem by utilizing
for example, dual-electrode detectors with two gold working electrodes conﬁgured in

86

seriesl or pulsed amperometric detection (PAD) which utilizes a continuous three-step

potential waveform.‘87

181
8,89and

Mass spectrometry has also been used as a detection mode for both glass
PDMS devices.I90 Although mass spectrometry can provide more chemical information
than any other types of detectors, it is less sensitive than LIF, very expensive and cannot
yet be integrated in a portable chip. However, many groups tried and succeeded to

interface microﬂuidic devices with mass spectrometry, especially electrospray ionization

time-of-ﬂight mass spectrometry (ESI-TOF MS)'9H94 for the detection of large

33

biomolecules with high sensitivity. CE-MS detection on chip has been studied
extensively by many researcherslgs‘196
On chip micromachined ionization sources that resemble electrospray emitter tips

have been reported by a number of groups.'97'20' Arscott et al. reported a planar on—chip
nib-like structure using the epoxy-based negative photoresist SU—8 as part of a
microfabricated ionization source that functions in nanoelectrospray conditions.202
With the progress in microfabrication technologies and detection methods,
researchers can basically choose any detection method that is suitable for their
applications. u-TAS and lab on-a-chip systems will continue to develop and possibly

integrate multiple detection schemes for multiple analytes in same device, for different

types of applications, for example rapid, on-site diagnostic.

1.3 THESIS OBJECTIVE

In this work we propose to apply the recent developments in microfabrication
technology to create an in vitro model that will closely mimic an in vivo system. Because
a molecular level understanding of the mechanisms by which RBCs deformability
properties relate to the occurrence of MS is lacking, to gain insight into these
mechanisms would be ideal to employ such a model system.

More speciﬁcally, this work describes the development of a complete blood brain
barrier (BBB) mimic using lithographically-derived microchips in order to monitor the
fate of endothelium-derived NO that is stimulated by mechanically deformed RBCs. The

microchip-based BBB mimic was developed by addressing an array of cell reservoirs,

34

which pose as the central nervous system (CNS), with a network of underlying channels
that represent the circulatory system. The cells were cultured on a polycarbonate
membrane, integrated in the device that separates the “so called” CNS from the “so
called” blood stream. The diameters of the patterned channels in the fabricated chips will
ultimately approximate those of brain capillaries in vivo.

This approach was used to monitor ATP-stimulated NO levels in the cells cultured
on the polycarbonate membrane. Importantly, using chip technology, we were able to
investigate the ﬂux of NO across the BBB brought about by changes in ATP
concentrations, both in the form of ATP standards and RBC-derived ATP. As the
detection scheme, ﬂuorescence microscopy was employed to enable the qualitative and
quantitative identiﬁcation of endothelium derived-NO after it diffuses through the

membrane of the microchip device.

35

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

LIST OF REFERENCES

Evans, E.; Fung, Y. C. Microvasc Res 1972, 4, 335-347.

Hartwig, J. H.; DeSisto, M. J Cell Biol 1991, 112, 407-425.

Schwiebert, E. M.; Benos, D. J.; Egan, M. E.; Stutts, M. J.; Guggino, W. B.
Physiol Rev 1999, 79, 8145-166.

Hessel, V.; Loewe, H.; Schoenfeld, F. Chem. Eng. Sci. 2005, 60, 2479-2501.
Mohandas, N.; Evans, E. Annu. Rev. Biophys. Biomol. Struct. FIELD Full Journal
TitlesAnnual Review of Biophysics and Biomolecular Structure 1994, 23, 787-
818.

Linderkamp, O.; Meiselman, H. J. Blood 1982, 59, 1121-1127.

Watanabe, N.; Kataoka, H.; Yasuda, T.; Takatani, S. Biophys. J. 2006, 91, 1984-
1998.

Rubanyi, G. M.; Romero, J. C.; Vanhoutte, P. M. Am J Physiol 1986, 250,
H1145-1149.

Buga, G. M.; Gold, M. E.; Fukuto, J. M.; Ignarro, L. J. Hypertension 1991, 17,
187-193.

Furchgott, R. F.; Zawadzki, J. V. Nature 1980, 288, 373-376.

Sprague, R. S.; Stephenson, A. H.; Dimmitt, R. A.; Weintraub, N. A.; Branch, C.
A.; McMurdo, L.; Lonigro, A. J. American Journal of Physiology 1995, 269,
H1941-H1948.

Sprague, R. 8.; Ellsworth, M. L.; Stephenson, A. H.; Lonigro, A. J. American
Journal of Physiology 1996, 271, H2717-H2722.

Busse, R.; Ogilvie, A.; Pohl, U. Am J Physiol 1988, 254, H828-832.

Bogle, R. G.; Coade, S. B.; Moncada, 8.; Pearson, J. D.; Mann, G. E. Biochem
Biophys Res Commun 1991, 180, 926-932.

36

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

Dean, B. M.; Perrett, D. Biochim Biophys Acta 1976, 43 7, 1-5.
Miseta, A.; Bogner, P.; Berenyi, E.; Kellermayer, M.; Galambos, C.; Wheatley,
D. N.; Cameron, 1. L. Biochim Biophys Acta 1993, 11 75, 133-139.

Bergfeld, G. R.; Forrester, T. Cardiovasc. Res. 1992, 26, 40-47.

Ellsworth, M. L.; Forrester, T.; Ellis, C. G.; Dietrich, H. H. American Journal of
Physiology 1995, 269, H2155-H2161.

Goldsmith, H. L. J Gen Physiol 1968, 52, SSuppl-28s.

Hassessian, H.; Bumstock, G. Br J Pharmacol 1995, 114, 846-850.

Collins, D. M.; McCullough, W. T.; Ellsworth, M. L. Microvascular research
1998, 56, 43-53.

Dietrich, H. H.; Ellsworth, M. L.; Sprague, R. S.; Dacey, R. 6., Jr. American
Journal of Physiology 2000, 278, H1294-H1298.

Dietrich, H. H.; Kajita, Y.; Dacey, R. 0., Jr. The American journal of physiology
1996, 271, H1109-1116.

McCullough, W. T.; Collins, D. M.; Ellsworth, M. L. American Journal of
Physiology 1997, 272, H1886-H1891.

Houston, D. A.; Bumstock, G.; Vanhoutte, P. M. J Pharmacol Exp Ther 1987,
241, 501-506.

Sprague, R. S.; Stephenson, A. H.; Ellsworth, M. L.; Keller, C.; Lonigro, A. J.
Exp Biol Med (Maywood) 2001, 226, 434-439.

Quinton, P. M. Nature 1983, 301, 421-422.
Anderson, M. P.; Gregory, R. J.; Thompson, S.; Souza, D. W.; Paul, S.; Mulligan,
R. C.; Smith, A. E.; Welsh, M. J. Science 1991, 253, 202-205.

Bear, C. E.; Li, C. H.; Kartner, N.; Bridges, R. J.; Jensen, T. J.; Ramjeesingh, M.;
Riordan, J. R. Cell 1992, 68, 809-818.

Lee, M. G.; Wigley, W. C.; Zeng, W.; Noel, L. E.; Marino, C. R.; Thomas, P. J.;
Muallem, S. J Biol Chem 1999, 274, 3414-3421.

37

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

(43)

(44)

(45)

(46)

Knowles, M. R.; Stutts, M. J .; Spock, A.; Fischer, N.; Gatzy, J. T.; Boucher, R. C.
Science (Washington, D. C., 1883-) 1983, 221, 1067-1070.

Stutts, M. J .; Canessa, C. M.; Olsen, J. C.; Hamrick, M.; Cohn, J. A.; Rossier, B.
C.; Boucher, R. C. Science 1995, 269, 847-850.

Boucher, R. C.; Stutts, M. J.; Knowles, M. R.; Cantley, L.; Gatzy, J. T. J Clin
Invest 1986, 78, 1245-1252.

Gabriel, S. E.; Clarke, L. L.; Boucher, R. C.; Stutts, M. J. Nature 1993, 363, 263-
268.

’ Jovov, B.; Ismailov, I. 1.; Benos, D. J. J Biol Chem 1995, 270, 1521-1528.

Schwiebert, E. M.; Egan, M. E.; Hwang, T. H.; Fulmer, S. 8.; Allen, S. S.;
Cutting, G. R.; Guggino, W. B. Cell 1995, 81, 1063-1073.

Schreiber, R.; Nitschke, R.; Greger, R.; Kunzelmann, K. J Biol Chem 1999, 274,
11811-11816.

Reisin, I. L.; Prat, A. G.; Abraham, E. H.; Amara, J. F .; Gregory, R. J .; Ausiello,
D. A.; Cantiello, H. F. J Biol Chem 1994, 269, 20584-20591.

Sugita, M.; Yue, Y.; Foskett, J. K. Embo J 1998, 1 7, 898-908.

Abraham, E. H.; Okunieff, P.; Scala, S.; Vos, R; Oosterveld, M. J.; Chen, A. Y.;
Shrivastav, B. Science 1997, 275, 1324-1326.

Abraham, E. H.; Prat, A. G.; Gerweck, L.; Seneveratne, T.; Arceci, R. J .; Kramer,
R.; Guidotti, G.; Cantiello, H. F. Proc Natl Acad Sci U S A 1993, 90, 312-316.

Pasyk, E. A.; Foskett, J. K. J Biol Chem 1997, 272, 7746-7751.

Prat, A. G.; Reisin, I. L.; Ausiello, D. A.; Cantiello, H. F. Am J Physiol 1996, 270,
C538-545.

Giovannoni, G.; Heales, S. J. R.; Land, J. M.; Thompson, E. J. Multiple Sclerosis
1998, 4, 212-216.

Mazzanti, L.; Faloia, E.; Rabini, R. A.; Staffolani, R.; Kantar, A.; Fiorini, R.;
Swoboda, B.; De Pirro, R.; Bertoli, E. Clin Biochem 1992, 25, 41-46.

Blair, G. W.; Matchett, R. H. J Neurol Neurosurg Psychiatry 1972, 35, 730-731.

38

(4'7)
(48)

(49)

(50)
(51)
(52)
(53)
(54)
(55)
(56)
(57)
(58)
(59)
(60)
(61)

(62)

Karlov, V. A.; Savin, A. A.; Smertina, L. P.; Redchits, E. G.; Seleznev, A. N.;
Svetailo, L. I.; Margosiuk, N. V.; Stulin, I. D.: USSR, 1990, pp 47-50.

Simpson, L. 0.; Shand, B. I.; Olds, R. J .; Larking, P. W.; Amott, M. J. Pathology
1987, 19, 51-55.

Szeinberg, A.; Golan, R.; Ben Ezzer, J.; Sarova-Pinhas, 1.; Sadeh, M.; Braham, J.
Acta Neurol. Scand. 1979, 60, 265-271.

Shukla, V. K. S.; Jensen, G. E.; Clausen, J. Acta Neurol. Scand. 1977, 56, 542-
550.

Zachara, B.; Gromadzinska, J.; Czemicki, J .; Maciejek, Z.; Chmielewski, H. Klin
Wochenschr 1984, 62, 179-182.

Szeinberg, A.; Golan, R.; Ben-Ezzer, J.; Sarova-Pinhas, I.; Kindler, D. Acta
Neurol Scand 1981, 63, 67-75.

Polidoro, G.; Di Ilio, C.; Arduini, A.; La Rovere, G.; Federici, G. Int. J. Biochem.
1984, 16, 505-509.

Jensen, G. E.; Clausen, J. J. Neurol. Sci. 1984, 63, 45-53.
Golan, R.; Szeinberg, A.; Ezzer, J. B.; Sarova-Pinhas, 1.; Sadde, M.; Braham, J.
Prog. Clin. Biol. Res. 1980, 39, 467.

Ghigo, D.; Bosia, A.; Pagani, A.; Pescarmona, G. P.; Pagano, G.; Lenti, G. Ric
Clin Lab 1983, 13 Suppl 3, 375-381.

Matkovics, B.; Kotorrnan, M.; Varga, I. S.; Hai, D. Q.; Roman, F .; Novak, Z. Acta
Physiol Hung 1997, 85, 99-106.

Carroll, J.; Raththagala, M.; Subasinghe, W.; Baguzis, S.; Oblak, T. D. a.; Root,
P.; Spence, D. Mol. BioSyst. 2006, 2, 305-311.

Giovannoni, G. Multiple Sclerosis 1998, 4, 27-30.

Parkinson, J. F .; Mitrobvic, B.; Merrill, J. B. Journal of Molecular Medicine
(Berlin) 1997, 75, 174-186.

Alderton, W. K.; Cooper, C. E.; Knowles, R. G. Biochemical Journal 2001, 357,
593-615.

Michel, T.; Feron, 0. Journal of Clinical Investigation 1997, 100, 2146-2152.

39

(63)

(64)

(65)

(66)

(67)

(68)

(69)

(70)

(71)

(72)

(73)

(74)

(75)

(76)

(77)

(73)

Guix, F. X.; Uribesalgo, I.; Coma, M.; Munoz, F. J. Progress in Neurobiology
(Amsterdam, Netherlands) 2005, 76, 126-152.

Karg, E.; Klivenyi, P.; Nemeth, 1.; Bencsik, K.; Pinter, S.; Vecsei, L. Journal of
neurology 1999, 246, 533-539.

Audus, K. L.; Borchardt, R. T. Handbook of Experimental Pharmacology 1991,
100, 43-70.

Brightrnan, M. W.; Reese, T. S. J Cell Biol 1969, 40, 648-677.

Reese, T. S.; Karnovsky, M. J. J Cell Biol 1967, 34, 207-217.

Bradbury, M. W. Circ Res 1985, 5 7, 213-222.

Audus, K. L.; Chikhale, P. J.; Miller, D. W.; Thompson, S. E.; Borchardt, R. T.
Adv. Drug Res. 1992, 23, 1-64.

Andersson, P. B.; Perry, V. H.; Gordon, S. Neuroscience 1992, 48, 169-186.
Lassmann, H.; Zimprich, F.; Rossler, K.; Vass, K. Rev Neurol (Paris) 1991, 14 7,
763-781.

de Vries, H. E.; Blom-Roosemalen, M. C. M.; van Oosten, M.; de Boer, A. G.;
van Berkel, T. J. C.; Breimer, D. D.; Kuiper, J. J. Neuroimmunol. 1996, 64, 37-43.

De Vries, H. E.; Kuiper, J.; De Boer, A. G.; Van Berkel, T. J. C.; Breimer, D. D.
Pharmacol. Rev. 1997, 49, 143-155.

Griot, C.; Vandevelde, M.; Richard, A.; Peterhans, E.; Stocker, R. Free Radical
Res. Commun. 1990, 11, 181-193.

Kim, Y. S.; Kim, S. U. J Neurosci Res 1991, 29, 100-106.

Thiel, V. E.; Audus, K. L. Antioxidants & Redox Signaling 2001, 3, 273-278.

Yuceyar, N.; Taskiran, D.; Sagduyu, A. Clinical neurology and neurosurgery
2001, 103, 206-211.

Mayhan, W. G. Brain research 1996, 743, 70-76.

40

(79)

(80)

(81)

(82)

(83)

(84)

(85)

(86)

(87)

(88)

(39)
(90)

(91)
(92)

(93)

(94)

(95)

Sprague, R. S.; Ellsworth, M. L.; Stephenson, A. H.; Kleinhenz, M. E.; Lonigro,
A. J. American Journal of Physiology 1998, 275, H1726-H1732.

Fischer, T. M.; Haest, C. W.; Stohr, M.; Kamp, D.; Deuticke, B. Biochim Biophys
Acta 1978, 510, 270-282.

Terry, S. C.; Angell, J. B. Theory, Des, Biomed. App]. Solid State Chem. Sens,
Workshop 1978, 207-218.

Van der Schoot, B.; Bergveld, P. Sens. Actuators 1985, 8, 11-22.
Manz, A.; Graber, N.; Widmer, H. M. Sens. Actuators, B 1990, B1, 244-248.

Reyes, D. R.; Iossiﬁdis, D.; Auroux, P.-A.; Manz, A. Analytical Chemistry 2002,
74, 2623-2636.

Manz, A.; Harrison, D. J .; Verpoorte, E. M. J .; Fettinger, J. C.; Paulus, A.; Luedi,
H.; Widmer, H. M. J. Chromatogr. 1992, 593, 253-258.

Stroock, A. D.; Dertinger, S. K. W.; Ajdari, A.; Mezit, 1.; Stone, H. A.;
Whitesides, G. M. Science (Washington, DC, U. S.) 2002, 295, 647—651.

Song, H.; Bringer, M. R.; Tice, J. D.; Gerdts, C. J.; Ismagilov, R. F. App]. Phys.
Lett. 2003, 83, 4664-4666.

Jacobson, S. C.; Hergenroder, R.; Koutny, L. B.; Warmack, R. J .; Ramsey, J. M.
Anal. Chem. 1994, 66, 1107-1113.

Manz, A. Chimia 1996, 50, 140-143.
Bratten, C. D.; Cobbold, P. H.; Cooper, J. M. Anal Chem 1998, 70, 1164-1170.
Clark, R. A.; Hietpas, P. B.; Ewing, A. G. Anal. Chem. 1997, 69, 259-263.

Ananchenko, G. S.; Bagryanskaya, E. G.; Tarasov, V. F .; Sagdeev, R. Z.; Paul, H.
Chem. Phys. Lett. 1996, 255, 267-273.

Song, M. I.; Iwata, K.; Yamada, M.; Yokoyama, K.; Takeuchi, T.; Tamiya, E.;
Karube, 1. Anal Chem 1994, 66, 778-781.

Briceno, G.; Chang, H.; Sun, X.; Schultz, P. G.; Xiang, X. D. Science
(Washington, D. C.) 1995, 270, 273-275.

Kovaca, G. T. A.; Petersen, K.; Albin, M. Anal. Chem. 1996, 68, 407A-412A.

41

(96)

(97)

(98)

(99)

(100)

(101)

(102)

(103)

(104)

(105)

(106)
(107)
(108)

(109)

(110)

(111)

(112)

MacDonald, N. C.; Adams, S. G.; Ayon, A. A.; Boehringer, K. F .; Chen, L. Y.;
Das, J. H.; Haronian, D.; Hofrnann, W.; Huang, X. T.; et al. Microelectron. Eng.
1996, 30, 563-564.

Lee, S. S.; Lin, L. Y.; Wu, M. C. Appl. Phys. Lett. 1995, 67, 2135-2137.

Sniegowski, J. J. T ribol. Issues Oppor. MEMS, Proc. NSF/AFOSR/ASME
Workshop 1998, 325-340.

Frazier, A. B.; Ahn, C. H.; Allen, M. G. Sens. Actuators, A 1994, A45, 47-55.

Williams, G.; D'Silva, C. Sens. VI [Proc Conf Sens. Their Appl.], 6th 1993, 161-
166.

Emmelius, M.; Pawlowski, G.; Vollmann, H. W. Angew. Chem. 1989, 101, 1475-
1502.

Chou, S. Y.; Krauss, P. R.; Renstrom, P. J. Science (Washington, D. C.) 1996,
272, 85-87.

Palmer, E. W.; Hutley, M. C.; Franks, A.; Verrill, J. F .; Gale, B. Rep. Prog. Phys.
1975, 38, 975—1048.

Xia, Y.; Kim, B; Zhao, X.-M.; Rogers, J. A.; Prentiss, M.; Whitesides, G. M.
Science (Washington, D. C.) 1996, 2 73, 347-349.

McDonald, J. C.; Duffy, D. C.; Anderson, J. R.; Chiu, D. T.; Wu, H.; Schueller,
O. J. A.; Whitesides, G. M. Electrophoresis 2000, 21, 27-40.

McDonald, J. C.; Whitesides George, M. Acc Chem Res 2002, 35, 491-499.
Becker, H.; Locascio, L. E. Talanta 2002, 56, 267-287.
Clarson, S. J .; Semlyen, J. A. Siloxane Polymers, 1993.

Rogers, J. A.; Qin, D.; Schueller, O. J. A.; Whitesides, G. M. Rev. Sci. Instrum.
1996, 67, 3310-3319.

Xia, Y.; McClelland, J. J.; Gupta, R.; Qin, D.; Zhao, X. M.; Sohn, L. L.; Celotta,
R. J.; Whitesides, G. M. Adv. Mater. (Weinheim, Ger.) 1997, 9, 147-149.

Chaudhury, M. K.; Whitesides, G. M. Langmuir 1991, 7, 1013-1025.

Chaudhury, M. K.; Whitesides, G. M. Science (Washington, D. C., I883-) I992,
255, 1230-1232.

42

(113)

(114)
(115)

(116)

(117)

(118)

(119)

(120)

(121)
(122)

(123)

(124)

(125)

(126)

(127)

(128)

(129)

(130)

Morra, M.; Occhiello, E.; Marola, R.; Garbassi, F .; Humphrey, P.; Johnson, D. J.
Colloid Interface Sci. 1990, 13 7, 11-24.

Becker, H.; Gartner, C. Electrophoresis 2000, 21, 12-26.
Moreau, W. M. Opt. Eng. 1983, 22, 181-184.

Brambley, D.; Martin, B.; Prewett, P. D. Adv. Mater. Opt. Electron. 1994, 4, 55-
74.

Flack, W. W.; Nguyen, H.-A.; Capsuto, E. S. Proc. SPIE-Int. Soc. Opt. Eng.
2003, 5039, 1257-1271.

Wolf, S.; Tauber, R. N. Silicon Processing for the VLSI ERA, Vol. 1: Process
Technology, 1986.

Whitesides, G. M.; Mathias, J. P.; Seto, C. T. Science (Washington, D. C., I883-)
I991, 254, 1312-1319.

Kumar, A.; Abbott, N. L.; Biebuyck, H. A.; Kim, E.; Whitesides, G. M. Acc.
Chem. Res. 1995, 28, 219-226.

Kim, B; Xia, Y.; Whitesides, G. M. Nature (London) 1995, 376, 581-584.
Mrksich, M.; Whitesides, G. M. Trends Biotechnol. 1995, 13, 228-235.

Grzybowski, B. A.; Haag, R.; Bowden, N.; Whitesides, G. M. Anal. Chem. 1998,
70, 4645-4652.

Schuster, R.; Kirchner, V.; Allongue, P.; Ertl, G. Science (Washington, D. C.)
2000, 289, 98-101.

Iwata, F.; Kawaguchi, M.; Aoyama, H.; Sasaki, A. Thin Solid Films 1997, 302,
122-126.

Kamholz, A. E.; Weigl, B. H.; Finlayson, B. A.; Yager, P. Anal Chem 1999, 71,
5340-5347.

Kamholz, A. E.; Yager, P. Biophys J 2001, 80, 155-160.

Forster, F. K.; Galambos, P. C.; Weigl, B. H.; Holl, M. R.; (University of
Washington, USA). Application: US, 1999, pp 23 pp.

Weigl, B. H.; Yager, P. Sens. Actuators, B 1997, B39, 452-457.

Macounova, K.; Cabrera, C. R.; Yager, P. Anal Chem 2001, 73, 1627-1633.

43

(131)

(132)

(133)

(134)

(135)

(136)

(137)

(138)

(139)

(140)

(141)

(142)

(143)

(144)

(145)

(146)

(147)

Hadd, A. G.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1999, 71, 5206-5212.

Hadd, A. G.; Raymond, D. E.; Halliwell, J. W.; Jacobson, S. C.; Ramsey, J. M.
Anal. Chem. 1997, 69, 3407-3412.

Duffy, D. C.; Gillis, H. L.; Lin, J.; Sheppard, N. F., Jr.; Kellogg, G. J. Anal.
Chem. 1999, 71, 4669-4678.

Karneoka, J.; Craighead, H. G.; Zhang, H.; Henion, J. Anal. Chem. 2001, 73,
1935-1941.

Xu, J.; Lee, C. S.; Locascio, L. E. Book of Abstracts, 219th ACS National
Meeting, San Francisco, CA, March 26-30, 2000 2000, ANYL-009.

Hatch, A.; Kamholz, A. E.; Hawkins, K. R.; Munson, M. S.; Schilling, E. A.;
Weigl, B. H.; Yager, P. Nat. Biotechnol. 2001, 19, 461-465.

Eteshola, E.; Leckband, D. Sens. Actuators, B 2001, B72, 129-133.

Cheng, S. B.; Skinner, C. D.; Taylor, J.; Attiya, S.; Lee, W. E.; Picelli, G.;
Harrison, D. J. Anal. Chem. 2001, 73, 1472-1479.

Yang, T.; Jung, S.-y.; Mao, H.; Cremer, P. S. Anal. Chem. 2001, 73, 165-169.

Sohn, L. L.; Saleh, O. A.; Facer, G. R.; Beavis, A. J .; Allan, R. S.; Notterrnan, D.
A. Proc. Natl. Acad. Sci. U. S. A. 2000, 97, 10687-10690.

Jiang, Y.; Wang, P.-C.; Locascio, L. E.; Lee, C. S. Anal. Chem. 2001, 73, 2048-
2053.

Gao, J .; Xu, J .; Locascio, L. E.; Lee, C. S. Anal. Chem. 2001, 73, 2648-2655.

Figeys, D.; Gygi, S. P.; McKinnon, G.; Aebersold, R. Anal. Chem. 1998, 70,
3728-3734.

Northrup, M. A.; White, R. M.; (Regents of the University of California, USA).
Application: WC, 1994, pp 51 pp.

Belgrader, P.; Okuzumi, M.; Pourahmadi, F.; Borkholder, D. A.; Northrup, M. A.
Biosens. Bioelectron. 2000, 14, 849-852.

Khandurina, J.; McKnight, T. E.; Jacobson, S. C.; Waters, L. C.; Foote, R. 8.;
Ramsey, J. M. Anal. Chem. 2000, 72, 2995-3000.

Lagally, E. T.; Medintz, 1.; Mathies, R. A. Anal. Chem. 2001, 73, 565-570.

44

(148)

(149)

(150)

(151)

(152)

(153)

(154)

(155)

(156)

(157)

(158)
(159)
(160)

(161)

(162)
(163)

(164)

Fan, Z. H.; Mangru, S.; Granzow, R.; Heaney, P.; Ho, W.; Dong, Q.; Kumar, R.
Anal. Chem. 1999, 71, 4851-4859.

Koutny, L.; Schmalzing, D.; Salas-Solano, O.; El-Difrawy, S.; Adourian, A.;
Buonocore, 8.; Abbey, K.; McEwan, P.; Matsudaira, P.; Ehrlich, D. Anal. Chem.
2000, 72, 3388-3391.

Woolley, A. T.; Mathies, R. A. Proc Natl Acad Sci U S A 1994, 91, 11348-11352.

Glasgow, 1. K.; Zeringue, H. C.; Beebe, D. J.; Choi, S. J.; Lyman, J. T.; Chan, N.
G.; Wheeler, M. B. IEEE Trans Biomed Eng 2001, 48, 570-578.

Yang, J .; Huang, Y.; Wang, X.-B.; Becker, F. F.; Gascoyne, P. R. C. Anal. Chem.
1999, 71, 911-918.

Chiu, D. T.; Li Jeon, N.; Huang, S.; Kane, R. S.; Wargo, C. J .; Choi, I. S.; Ingber,
D. E.; Whitesides, G. M. Proc. Natl. Acad. Sci. U. S. A. 2000, 97, 2408-2413.

Folch, A.; Jo, B.-H.; Hurtado, 0.; Beebe, D. J .; Toner, M. J. Biomed. Mater. Res.
2000, 52, 346-353.

Dertinger, S. K. W.; Chiu, D. T.; Jeon, N. L.; Whitesides, G. M. Anal. Chem.
2001, 73, 1240-1246.

Jeon, N. L.; Dertinger, S. K. W.; Chiu, D. T.; Choi, I. S.; Stroock, A. D.;
Whitesides, G. M. Langmuir 2000, 16, 8311-8316.

Weigl, B. H.; Darling, R. B.; Yager, P.; Kriebel, J.; Mayes, K. Proc. SPIE-Int.
Soc. Opt. Eng. 1999, 3606, 82-89.

Cunningham, D. D. Anal. Chim. Acta 2001, 429, 1-18.
Hesketh, P. J .; Bintoro, J. S.; Luharuka, R. Sens. Update 2004, 13, 233-302.
Woias, P. Sens. Actuators, B 2005, B105, 28-3 8.

Auroux, P.-A.; Iossiﬁdis, D.; Reyes, D. R.; Manz, A. Analytical Chemistry 2002,
74, 2637-2652.

Dittrich, P. S.; Tachikawa, K.; Manz, A. Anal. Chem. 2006, 78, 3887-3907.
Vilkner, T.; Janasek, D.; Manz, A. Anal. Chem. 2004, 76, 3373-3386.

Manz, A.; Becker, H.; Editors Microsystem Technology in Chemistry and Life
Science. [In.' Top. Curr. Chem, 1998; 194], 1998.

45

(165)
(166)
(167)
(168)
(169)
(170)
(171)
(172)
(173)

(174)

(175)

(176)

(177)

(178)

(179)

(180)

(181)

(182)

(183)

Yang, Z.; Matsumoto, 8.; Goto, H.; Matsumoto, M.; Maeda, R. Sens. Actuators, A
2001, A93, 266-272.

Liu, R. H.; Yang, J .; Pindera, M. Z.; Athavale, M.; Grodzinski, P. Lab Chip 2002,
2, 151-157.

Liu, R. H.; Lenigk, R.; Druyor-Sanchez, R. L.; Yang, J.; Grodzinski, P. Anal.
Chem. 2003, 75, 1911-1917.

Oddy, M. H.; Santiago, J. G.; Mikkelsen, J. C. Anal. Chem. 2001, 73, 5822-5832.
Glasgow, 1.; Aubry, N. Lab Chip 2003, 3, 114-120.

Qian, S.; Bau, H. H. Anal. Chem. 2002, 74, 3616-3625.

Paik, P.; Pamula, V. K.; Fair, R. B. Lab Chip 2003, 3, 253-259.

Woias, P.; Hauser, K.; Yacoub-George, E. Micro T otal Anal. Syst. 2000, Proc.
micro TAS Symp., 4th 2000, 277-282.

West, J.; Karamata, B.; Lillis, B.; Gleeson, J. P.; Alderman, J.; Collins, J. K.;
Lane, W.; Mathewson, A.; Bemey, H. Lab Chip 2002, 2, 224-230.

Voldman, J.; Gray, M. L.; Schmidt, M. A. J. Microelectromech. Syst. 2000, 9,
295-302.

Uchiyama, K.; Nakajima, H.; Hobo, T. Anal. Bioanal. Chem. 2004, 379, 375-382.

Landers, J. P.; Editor Handbook of Capillary Electrophoresis, Second Edition,
1997.

Timperman, A. T.; Khatib, K.; Sweedler, J. V. Anal Chem 1995, 6 7, 139-144.

Gassmann, E.; Kuo, J. E.; Zare, R. N. Science (Washington, D. C., I883-) 1985,
230, 813-814.

Chen, D. Y.; Dovichi, N. J. J. Chromatogr., B: Biomed. App]. 1994, 657, 265-
269.

Kwon, S.; Lee Luke, P. Opt Lett 2004, 29, 706-708.

Gu, G.; Shen, Z.; Burrows, P. E.; Forrest, S. R. Adv. Mater. (Weinheim, Ger.)
1997, 9, 725-728.

Swinney, K.; Bornhop, D. J. Electrophoresis 2000, 21, 1239-1250.

Markov, D.; Begari, D.; Bornhop, D. J. Anal. Chem. 2002, 74, 5438-5441.

46

(184)

(185)

(186)

(187)
(188)

(189)

(190)

(191)
(192)

(193)

(194)

(195)

(196)

(197)

(198)

(199)

(200)

Tantra, R.; Manz, A. Anal. Chem. 2000, 72, 2875-2878.

Woolley, A. T.; Lao, K.; Glazer, A. N.; Mathies, R. A. Anal. Chem. 1998, 70,
684-688.

Martin, R. S.; Gawron, A. J.; Lunte, S. M.; Henry, C. S. Anal. Chem. 2000, 72,
3196-3202.

Ewing, A. G.; Mesaros, J. M.; Gavin, P. F. Anal Chem 1994, 66, 527A-537A.
Dolnik, V.; Liu, S.; Jovanovich, S. Electrophoresis 2000, 21, 41-54.

F igeys, D.; Gygi, S. P.; McKinnon, G.; Aebersold, R. Anal Chem 1998, 70, 3728-
3734.

Chan, J. H.; Timperman, A. T.; Qin, D.; Aebersold, R. Anal Chem 1999, 71,
4437-4444.

Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69, 1174-1178.
Rohner, T. C.; Rossier, J. S.; Girault, H. H. Anal. Chem. 2001, 73, 5353-5357.

Tang, K.; Lin, Y.; Matson, D. W.; Kim, T.; Smith, R. D. Anal Chem 2001, 73,
1658-1663.

Kameoka, J .; Craighead, H. G.; Zhang, H.; Henion, J. Anal Chem 2001, 73, 1935-
1941.

Xue, Q.; Dunayevskiy, Y. M.; Foret, F.; Karger, B. L. Rapid Commun Mass
Spectrom 1997, 11, 1253-1256.

Lazar, I. M.; Ramsey, R. S.; Sundberg, 3.; Ramsey, J. M. Anal. Chem. 1999, 71,
3627-3631.

Desai, A.; Tai, Y.-C.; Davis, M. T.; Lee, T. I). Transducers 97, Int. Conf Solid-
State Sens. Actuators 1997, 2, 927-930.

Lin, L.; Shia, T. K.; Chiu, C. J. J. Micromech. Microeng. 2000, 10, 395-400.

Licklider, L.; Wang, X. Q.; Desai, A.; Tai, Y. C.; Lee, T. D. Anal Chem 2000, 72,
367-375.

Tang, K.; Lin, Y.; Matson, D. W.; Kim, T.; Smith, R. D. Anal. Chem. 2001, 73,
1658-1663.

47

(201) Schultz, G. A.; Corso, T. N.; Prosser, S. J.; Zhang, S. Anal. Chem. 2000, 72,
4058-4063.

(202) Arscott, S.; Le Gac, S.; Druon, C.; Tabourier, P.; Rolando, C. J. Micromech.
Microeng. 2004, 14, 310-316.

48

CHAPTER 2

2.1 MULTIPLE SCLEROSIS: AN AUTOIMMUNE DISEASE?

Multiple sclerosis (MS) is thought to be an inﬂammatory and demyelinating
disease of the central nervous system (CNS) that is the most common neurodegenerative
condition in young adults, especially women, striking people between the ages of 20 and
40 years. Globally, 2 to 3 million individuals have this disease, 400,000 in the US alone.1
It is ﬁve times more prevalent in temperate climates, especially northern Europe or
northern United States, with approximately 1 in 1,000 individuals of northern European
origin developing MS.2

MS is characterized by a chronic inﬂammatory process resulting in damage to
myelin (the multilayered membrane surrounding the axons) and myelin-producing cells
(oligodendrocytes) and eventual neuronal loss. It has been hypothesized that myelin is
damaged due to an immune attack of autoantibodies and autoreactive T cells activated
against myelin antigens such as myelin basic protein (MBP), proteolipid protein (PLP),
and myelin oligodendrocyte glycoprotein (MOG).3 Myelin is very important for nervous
system function and when this tissue is disrupted, the nerve conduction slows and leads
to neuromotor dysfunction. Moreover, in MS plaque tissue oligodendrocytes are
hypertrophied, with swollen nuclei and disrupted plasma and mitochondrial membranes.4

Studies on experimental allergic encephalomyelitis (EAE) models, which is a
widely used animal model for MS, indicated that inﬂammatory molecules (cytokines,

chemokines, and adhesion molecules) induce the recruitment of leucocytes from the

49

periphery to the CNS. Thus, the inﬂammation reaches the CNS through a disrupted blood
brain barrier (BBB). The pathological hallmark of the disease is the formation of plaques,
areas of white-matter demyelination, or eventual axonal degeneration and loss, leading to
long-term disability.5

However, there are contradictory reports in the literature about the autoimmune
hypothesis of this disease, largely because no autoantigen(s) speciﬁc to or causative for
MS has ever been identiﬁed. Some reports indicate that demyelination may precede
inﬂammation,6 while other say that MS is not a autoimmune disease but either a
genetically determined disorder characterized by metabolically dependent
neurodegeneration7 or resulting from a chronic viral infections” Moreover, there are
indications that MS is a disease triggered by an environmental factor in genetically
susceptible individuals. Dyment et al. showed that there is a 70% disconcordance of MS
among monozygotic twins, suggesting that an exogenous factor caused the disease.ll

Although there is controversy about what causes the disease, and an established
mechanism of its occurrence may not yet exist, it is unanimously accepted that axonal
degeneration is the major determinant of irreversible neurological disability in patients
with MS. Also, it was agreed that MS is somewhat heterogeneous and it has been
classiﬁed into several disease types such as relapsing, primary progressive, secondary
progressive and progressive relapsing. At onset, approximately 85% of the MS patients
show a relapsing remitting course, characterized by attacks of neurological deﬁcits,
followed by complete or incomplete recovery. It has also been shown that 40-65% of

patients with relapsing disease will experience slow progression (secondary progressive

course). '2

50

2.2 NITRIC OXIDE AS AN ACTIVITY MARKER IN MS

Although causative factors for MS have not been identiﬁed, there exists a large
body of evidence for the role of the autoimmunity in the pathogenesis of the disease.13 "4
Among other molecules produced during the inﬂammation cascade, nitric oxide (NO) is
one of the major molecules associated with the damage of the myelin and
oligodendroglia.15 "8

NO is involved in important physiological functions of the CNS, including
neurotransmission, learning and memory, and synaptic plasticity. It also has a role in
gastrointestinal motility and neuroendocrine regulation.”23 It is clear that NO can
modulate the induction of the immune response, permeability of the BBB, trafﬁcking of
the cells to the CNS, and local response in the inﬂammatory milieu.4 Therefore, NO can
have both beneﬁcial and deleterious effects. Because of its short life-time in vivo, NO can
be rapidly converted to more reactive intermediates such as nitrite and nitrate.‘8 Recently,
elevated levels of nitrate and nitrite have been measured in cerebral spinal ﬂuid (CSF),
blood, urine, and serum of MS patients in relapse.24 The neurotoxicity of excessive
concentrations of NO in MS has been previously reported,'6"8’25 however the origin of
this excess and its mechanism of action is yet unknown.

NO is a simple gaseous molecule with a complex chemistry in vitro and poorly
understood biochemistry in vivo. At the site of inﬂammation, NO can exist not only in the
form of a free radical, but also can be converted into a variety of other related compounds

formed by the different conditions present in the tissue environment (e.g., redox

conditions, pH, oxygen tension).26 These compounds include the nitroxyl (NOT) ion,

51

nitrous acid (HNOz), the nitrogen dioxide (N02) radical, peroxynitrite (ONOO‘; a
product of the combination of superoxide and NO) and peroxynitrous acid (ONOOH).
All of these compounds have their own chemistries and biochemistries in vivo and they
have also been found at the sites of inﬂammatory lesions in MS.

The high abundance of NO and its metabolites at sites of inﬂammation may be
due to the increased expression of the inducible form of nitric oxide synthase (iNOS).
This form of the enzyme produces NO continuously at high rates and does not depend on
free calcium. In normal conditions, iNOS is absent in the CNS, although iNOS mRNA
and protein become expressed at the inﬂammatory lesions of animals with EAE.27'29
Other ﬁndings also showed that iNOS mRNA was expressed in abundance in biopsy
samples30 or brains15 of patients with MS.

In this construct, inhibition of NO production, especially of iNOS, as a method of
effective therapy for MS has been extensively investigated for the past 15 years.3"38
However, results have turned out to be inconclusive given the complexity of NO
involvement in the immune regulation.

It has been hypothesized that NO has two major effects on cerebral vessels, both
of which may be involved in the pathogenesis of the MS lesions, namely vasodilation and
disturbance of the BBB integrity.26 It is also believed that vasodilation by itself can lead
to an inﬁltration of inﬂammatory cells through the BBB into the CNS. The reasoning
behind this hypothesis is that with the increase in vasodilation, a decrease in the blood
ﬂow will help leucocytes to bind to the vessel wall and migrate through it. In accordance

with this belief, Giovannoni et al. reported that nitrite and nitrate concentrations in the

CSF of patients with MS were correlated with the breakdown of the BBB, as measured

52

by the leakage of albumin.24 This NO present in MS lesions could arise from various
sources such as nerve terminals,”40 induction of iNOS,“ or the release of
neurotransmitters.“’43

In this work, we hypothesize that another source of NO could contribute to the
elevated levels of this radical (and its metabolites) in CSF or other ﬂuids of patients with
MS. Speciﬁcally, the stimulation of endothelial NOS (eNOS) by RBC-derived ATP may
be the source of NO. eNOS produces low (nanomolar) concentrations of NO in a
calcium-dependent manner and this NO is known to contribute to the regulation of blood
ﬂow by inducing smooth muscle relaxation. However, there exists a ﬁne balance between
the effects of the over production of NO and the underproduction of NO in vivo.
Increased concentrations of RBC-derived ATP may lead to increased levels of NO and
ultimately to relaxation of the vessels. This relaxation will slow the blood ﬂow and will
allow the adherence of leucocytes to the vessel wall and eventual penetration of the BBB
and leakage into the CNS, as previously proposed by Smith et 01.26

All previous studies, however, hypothesized that NO sources are within the CNS.
The NO production in the brain microvascular capillaries is a subject of controversy and
no solid evidence about the constitutive production of NO was yet reported. Herein, we

show that brain microvascular endothelial cells constitutively produce NO and, moreover,

ATP represents an in vitro stimulus for eNOS in this cell line.

53

2.3 RBC DEF ORMABILITY IN MS

It has been previously shown that abnormal RBC deformability is associated with
diseases such as diabetes and multiple sclerosis.44"15 While RBC deformability is
decreased in RBCs of patients with type I and type II diabetes, the whole blood of
patients with MS was found to be more viscoelastic when compared with the whole blood
of healthy patients.”48

Interestingly, it is known that the RBCs of MS patients have abnormally high
levels of glutathione (GSH), the most abundant non-enzymatic antioxidant present in a
cell. The levels of GSH have been linked with the deformability characteristics of the
RBCs. That is, low levels of GSH are found in diabetic red cells, which are known to be
less deformable than the RBCs from healthy patients.”50 Moreover, recent studies in our
group involving diabetic RBCs have shown that a weakened oxidant defense system in
RBCs will lead to lower levels of ATP release.5 I

If, in the case of MS patients, the RBCs are more deformable and the increase
deformability results in an increase of ATP release, then an increase in the NO
production would be expected. Indeed, it is recognized that there exist higher than normal
levels of NO metabolites, nitrate and nitrite, in the cerebral spinal ﬂuid (CSF) and serum
of patients with M82144

Taken collectively, we have constructed a theory that explains the overabundance

of NO metabolites in the CSF is actually blood-bome. In other words, it is possible that

the NO levels are increased due to overstimulation by RBC-derived ATP.

54

Our hypothesis is that MS-RBCs are more deformable, therefore releasing more ATP,
which in turn will lead to increased production of NO. However, a molecular level
understanding of the mechanisms by which RBC deformability properties relate to the
occurrence of MS is lacking. To prove or disprove this hypothesis, the role of RBC-

derived ATP and NO derived from brain endothelium must be investigated.

2.4 POSSIBLE ROLES OF THE RBC/RBC-DERIVED ATP IN MS

RBCs, when traversing microvascular beds, are subjected to mechanical
deformation. Previously, it has been reported in the isolated perfused rabbit lung that in
order to demonstrate ﬂow-induced endogenous NO synthesis in the pulmonary
circulation, RBCs obtained from either rabbits or healthy humans were a required
component of the perfusate.52 Importantly, it was reported that the property of these
RBCs responsible for the stimulation of NO synthesis was their ability to release ATP in
response to mechanical deformation.53 These reports suggest a novel mechanism for the
control of endothelium-derived NO production.

It is proposed that as the RBC is increasingly deformed by increments in the
velocity of blood ﬂow through a vessel and/or by reductions in vascular diameter, it
releases ATP that stimulates endothelial synthesis of NO. Indeed, it has been reported
that RBC-derived ATP contributes to the control of vascular resistance in both the

52-54

pulmonary and systemic circulation.”59 The ﬁnding that ATP is released from RBCs

in response to mechanical deformation, and that the levels of ATP released from RBCs

55

increase as the RBC deformability increases, suggest that the RBC may be an important
determinant of NO production as these cells traverse the intact circulation.

If RBC-derived ATP is a determinant of NO production in vivo, and NO plays a
major role in MS, it is important to investigate the relationship between MS RBCs, ATP,
NO and brain microvascular endothelial cells (BMVEC). The ﬁrst step towards this goal
is to establish that RBCs of MS patients release higher levels of ATP compared to

healthy, non-MS controls and that bBMVECs produce NO upon ATP stimulation.

2.5 EXPERIMENTAL

It has been well established that RBCs release ATP in response to mechanical
deformation.53’54 Deformability measurements of the red cells have been previously
performed by different methods based on viscosity,60 centrifugal sedimentation,61
micropipette aspiration,62 rheoscopy,63 ﬁltration or nucleopore ﬁltration, or nuclear track
microﬁlters and laser diﬁraction or ektacytometer.64'66 However, most of these
techniques required manual manipulation and thus, they were inaccurate and clinically
inapplicable. A more practical method to consider is the ﬁltration-based method, where
the RBCs are passed through the pores of a ﬁlter under pressure. The pores have uniform
size and dimensions similar to those of capillaries (5 pm diameter), in an effort to mimic
the in vivo conditions. The deformability of the RBCs was determined by the rate of
ﬁltered cells.67459

Previously, our group developed an alternative method for studying ATP release

from RBCs as they traverse fused silica microbore tubing with internal diameters

56

comparable to those of resistance vessels in the intact circulation.57’70 Here, this
technology will be used to quantitatively determine the ATP release from mechanically
deformed RBCs of patients with MS. Our hypothesis is that RBCs obtained from patients
with MS are more deformable, therefore releasing more ATP, which in turn will lead to
increased levels of NO production.

For a more realistic depiction of the mechanism of endothelium-derived NO
production in vivo, the RBCs obtained ﬁom MS patients were mechanically deformed by
hydrodynamically pumping them through microbore capillary tubing. A continuous ﬂow
system is employed to create stress on the RBCs once they have entered the microbore

tubing.7|

2.5.1 Design and Methods

Generation of washed RBC. MS and control blood samples were kindly provided by Dr.
James Garbem, M.D., associate professor in the Department of Neurology at Wayne
State University. Blood was centrifuged at 500g at 4°C for 10 min. The plasma and buffy
coat (other elements in the blood) were discarded. RBCs were resuspended and washed

three times in a physiological salt solution [PSS; in mM, 4.7 KCl, 2.0 CaClz, 140.5 NaCl
12 MgSO4, 21.0 tris(hydroxymethyl)aminomethane, 11.1 dextrose with 5% bovine serum

albumin (ﬁnal pH 7.4)]. For all studies reported here, a hematocrit of 7% was prepared.
The value of 7% was found to be optimal for our studies, although higher values 10-20%
represent an average over the microcirculation. It is known that when moving from the

macrocirculation (arteries) to microcirculation (arteriols and capillaries), the viscosity of

57

the blood decreases therefore the hematocrit changes with the vessel diameter. Cells were

prepared on the day of use.

Standard ATP assay. A mixture of luciferin/luciferase was used to measure the ATP by
chemiluminescence71 conform to the reaction shown in Figure 8. Chemiluminescence
intensity is proportional to the ATP concentration. The sensitivity of the assay is
enhanced by the addition of 2 mg of D-luciferin (Sigma, St. Louis, M0) to the crude
ﬁreﬂy extract (Sigma, St. Louis, MO). Intensities of ATP standards were taken on the
day of each experiment by adding authentic ATP to a luciferin/luciferase reaction
mixture. The luciferin/luciferase solutions were prepared by diluting the luciferin in 5 mL
of distilled, deionized 18 MO water (DDW) and adding it to a vial containing 100 mg

luciferase. The luciferin/luciferase mixture was prepared on the day of use.

58

”0 s N co,-

/>_</ . m

 

 

 

 

N S
Mg2+
O
HO S N H O Adenine
/ O—P—O
/ I + 1’91
N s O-
( HO OH
_ y _ 4
-0 S N O
/ + co2 + AMP + H
N S
L 4
'O

S N
/> /< + Light
N S

Figure 8. The two-step bioluminescence reaction between luciferin/luciferase and ATP.

59

Determination of A T P release from RBCs. The amount of ATP release from RBCs was
determined, near-real time, by using 50 um inside diameter fused-silica tubing that was
intended to resemble the vessels in vivo. The instrumentation employed consisted of a
syringe pump (Harvard Apparatus, Boston MA) equipped with two 500 uL syringes
(Hamilton, Fisher Scinetiﬁc), two pieces of ~40 cm long microbore tubing with an
internal diameter of 50 um and an outer diameter of 362 um (Polymicro technologies,
Phoenix, AZ). The RBCs or ATP standards loaded in one of the syringes are pumped
through the tubing at a rate of 6.7 1.1L min". As the RBCs traverse the tubing, ATP is
released and it is combined with a luciferin/luciferase mixture that is loaded in the second
syringe, at a tee (Upchurch Scientiﬁc, Oak Harbor, WA) in the system, and the resultant
chemiluminescence is measured by a photomultiplier tube (PMT) built in house (Figure
9). The light intensity of the chemiluminescence is proportional to the amount of ATP
present.71 A working curve was constructed using ﬁve ATP standards between 0 and 1.5
uM resulting in a slope of 0.092 and a y-intercept of 0.0039 (1'2 = 0.9789). The ATP
standards were prepared from a stock 100 11M ATP in a buffered physiological salt

solution (PSS). All solutions were prepared in the day of the experiment.

60

RBC/ATP samoles

 

Figure 9. Instrumental setup employed to measure the RBC derived ATP release.

61

2.5.2 Results and Discussion

Quantitative evaluation of the MS RBC A T P release. Deforrnation-induced ATP
release from RBCs obtained from control and MS patients was examined in order to
determine if the MS RBCs release more ATP than control RBCs as a result of always
being in a more “deformable” state. The data in Figure 10 provides evidence that MS-
RBCs release higher levels of ATP compared to non-MS RBCs. The results demonstrate
that the overall RBC-derived ATP from the patients with MS (375 +/- 51 nM) is 100%
greater than that of healthy controls (138 +/- 21).These ﬁndings suggest that RBCs of MS
patients may indeed be in a continuous “deformable “state that will lead to higher levels
of ATP release and ultimately, to an increase in ATP-stimulated endothelium-derived

NO.

2.6 CONSTITUTIVE PRODUCTION OF NITRIC OXIDE IN BRAIN

MICROVASCULAR ENDOTHELIAL CELLS

It is well established that vascular endothelium constitutively generates NO in
response to ATP stimulation and that endothelium-derived NO induces a relaxation of
vascular smooth muscle cells. However, little is known about the production of NO in
microvessels, where smooth muscle layers are thin or absent. Additionally, the
mechanism by which mechanical stress induces NO synthesis in the microvessel
endothelium is still controversial. There are contradicting reports in the literature

regarding the constitutive production of NO in cerebral microvascular endothelium.

62

 

 

 

 

 

 

 

 

500
1
1 — Control RBCS
400 - 1:] MS RBCS
E 300 -
a) -1
«‘6’
2
£2
a. 200 -
[-5 .
< .
100 4
.1
0
Figure 10. ATP released from RBCS subjected to shear-induced stress. The average ATP

release from the RBCS obtained from healthy controls was 138 +/- 21 nM
while the release ﬁ'om the RBCS obtained from people with MS was 375 +/-
51 nM. The error bars are reported as the SEM for 11 controls and 18 MS
samples.

63

Kimura et a]. reported that ATP and a calcium ionophore (A23187) induced Ca2+
transients in two types of endothelium, bovine brain microvascular endothelial cells
(bBMVECs) and bovine aortic pulmonary endothelial cells (bPAEC), but they failed to
induce NO production in bBMVEC, as measured with an NO-sensitive ﬂuorescent dye,
DAF-2.72 However, an observed increase in ﬂuorescence was reportedly due to an
increase in NO production in bPAECs. In contrast, Janigro et al. showed that ATP
induces NO release in CNS endothelial cells due to an inﬂux of Ca2+ into the cell through
receptor-operated ion channels.73 In this study, we demonstrated that NO is indeed
constitutively generated in brain microvessels, more speciﬁcally in bBMVECs and that

RBC-derived ATP is an in vivo stimulus of NO production in bovine brain endothelium.

2.7 EXPERIMENTAL

It has been previously shown that endothelium-derived NO production can be
stimulated by ATP-binding to the P2, purinergic receptors found on endothelial cells.
Recently, our group has demonstrated the ability of measuring ATP stimulated
endothelium-derived NO via dynamic ﬂuorescence microscopy as an improvement over
previous reports using amperometry as the detection scheme.74 Different methods were
reported for NO detection such as, chemiluminescence assays, bioassays, Griess reaction,
oxy-hemoglobin assays, electron spin resonance, and electrochemical assays, each with
its advantages and disadvantages.75 Here we demonstrate that ATP induced, endothelium-

derived NO production can be detected in real time using a commercially available cell-

64

permeable ﬂuorescence probe, 4-amino-5-methylamino-2,7’-diﬂuoroﬂorescein diacetate

(DAF-FM DA).

2.7.1 Design and Methods

Preparation of cells. bBMVECs primary cultures (Cambrex, East Rutherford, NJ) were
seeded on collagen-coated 25 mL tissue culture ﬂasks at a density of 20,000 cells/cmz.
The cells were kept in humidiﬁed atmosphere, 37°C and 5% C02 and the media for
conﬂuent growth refreshed every 2-3 days. Generally, the cells reach conﬂuence on day 7
or 8 and form a tight monolayer on ~ day 10 or 11. When the cells from primary culture
reached conﬂuence they were sub-cultured at a split ratio 1:2, and the harvested sub-
cultured cells (divisions 3, 5, and 8) were used for the experiments. However, low
passage is desirable for the cells to express cellular tight junctions and transport
characteristics found in BBB.76 In order to conduct the experiment, the cells were seeded
in 2 inches Petri dishes and grown to conﬂuence. All solutions were prepared in Hank’s

balanced salt solution (HBSS) and equilibrated for 20 min at 37°C and 5% C02.

F lorescenee determination of NO production. For the measurement of NC. with DAF-
FM, the conﬂuent cells were incubated with 10 11M of the diacetylated form of DAF-FM
(Molecular Probes, Eugene, OR) for 20 min. Subsequently, the cells were washed twice
with equilibrated HBSS and stimulated with 10 1.1M or 100 11M ATP. DAF-F M ﬂuoresces
(Aex = 494 nm; Aem = 515 nm) upon binding with NO, as measured by ﬂuorescence

microscopy. The ﬂuorescence intensity (pixel intensity) was reported as an average of 4

65

pictures per dish taken every 5 min during a 30 min period. The images were captured by
an Olympus IX71M inverted microscope (Olympus America, Melville, NY) equipped
with an electrotherrnally-cooled CCD (Orca, Harnarnatsu) and MicroSuite software
(Olympus America). To conﬁrm the ATP-stimulated NO production, cells were
incubated with a competitive inhibitor of NO synthesis, L-nitro arginine methyl ester (L-
NAME 100 nM), for 15 min before loading them with the DAF-FM DA. Moreover, to
ensure that the enzyme eNOS has excess substrate for NO production, the cells were

incubated with L-arginine (100 11M).

Control experiment. Basal levels of NO in bBMVECs were determined using DAF-FM
DA as described above. Any increase in ﬂorescence intensity relative to the basal level
was due to the stimulation of the NO production. The DAF—F M ﬂorescence was
subtracted from the ﬁnal value, thus the data represents the net gain in ﬂuorescence due

to NO production.

2.7.2 Results and Discussion

A T P constitutes an in vivo stimulus of NO production in bBMVECs. Figure 11 depicts
the effect of ATP (100 nM) and L-NAME (100 nM) on bBMVEC NO production,
assessed with DAF-FM DA. The results demonstrate the ability of bBMVECs to produce
NO when ATP, a known stimulus of eNOS in vivo, is added to the cells. The bBMVEC
were loaded with DAF-FM DA and the ﬂorescence intensity was assessed after 30 min.

The black bar represents the ﬂuorescence intensity of the adduct DAF -F M /N0.

66

 

60

50‘

40-

30'

 

20‘

10-

Fluorescence intensity (pixel intensity)

 

 

 

 

 

 

 

control ATP ATP+L-NAME

Figure 11. Effects of ATP (100 pM) and L-NAME (100 1.1M) on NO production in
bBMVECs assessed with DAF -F M DA. L-MAME was used to demonstrate
that the increase in DAF-FM DA ﬂuorescence upon stimulation with ATP
(100 nM) was the result of NO production (*p 5 0.001 vs. control).

67

The light gray bar represents the ﬂuorescence intensity after the addition of 100 11M ATP

to the cells loaded with the probe. Darker gray bar shows that L-NAME (100 1.1M)

inhibited ATP-induced DAF-FM ﬂuorescence, proving the endothelium NO production.

A T P-derived NO production with cells passage. It has been shown that primary or low
passage brain capillary endothelial cultures retain many morphological and biochemical
properties similar to the BBB in vivo.77’78 However, due to the high cost of the
commercially available bBMVECs the use of primary cultures exclusively for our models
is not ﬁnancially practical. Therefore, in our protocols the cells were passaged and the
ability of the endothelium to produce NO upon stimulation was investigated. Data in
Figure 12 shows that with increasing the passage number, the ability of the bBMVEC to
produce N0 upon stimulation is signiﬁcantly reduced. Black bars (control) represent the
basal levels of NO production without stimulation. However upon stimulation with 100
uM ATP, a signiﬁcant increase (> 100%) in NO production is observed for division 3

compared with no increase (equal to basal levels) in divisions 5 or 8.

68

 

60

 

- control
"‘ :1 + ATP (1000M)

 

 

50‘

 

40-

30+

20-

10‘

Fluorescence intensity (pixel intensity)

 

 

 

 

 

 

Figure 12. ATP-derived NO production is decreasing with increasing the passage of the
cells. For cell division 3 there is a signiﬁcant difference between the control
(black bars) and the bBMVECs stimulated with 100 uM ATP (grey bars)
(*p 5 0.001 vs. control).

69

2.8 CONCLUSIONS

In this study we hypothesized that RBCs obtained from patients with MS are
more deformable, therefore releasing more ATP, which in turn will lead to increased
levels of NO production. For a more realistic depiction of the mechanism of endothelium-
derived NO production in vivo, the RBCS obtained from MS patients were mechanically
deformed by hydrodynamically pumping them through microbore capillary tubing. A
continuous ﬂow system is employed to create stress on the RBCS once they have entered
the microbore tubing.7| It was found that MS-RBCs release higher levels of ATP
compared to non-MS RBCS. These ﬁndings suggest that RBCS of MS patients may
indeed be in a continuous “deformable “state that will lead to higher levels of ATP
release and ultimately, to an increase in ATP-stimulated endothelium-derived NO.
Additionally, it was shown that ATP acts as an agonist for NO production in bBMVECs
and the increase in NO production is a function of cell division.

Based on these ﬁndings, we have constructed a theory that explains the
overabundance of N0 metabolites in the CSF of MS patients is actually blood-bome. In
other words, it is possible that the NO levels are increased due to overstimulation by

RBC-derived ATP.

70

(1)
(2)
(3)
(4)

(5)

(6)

(7)
<8)
(9)
(10)
(11)
(12)
(13)
(14)

(15)

(16)

LIST OF REFERENCES

Pryce, G.; Baker, D. Trends in Neurosciences 2005, 28, 272-276.
Noseworthy, J. H. Nature 1999, 399, A40-47.
Garren, H.; Steinman, L.; Lock, C. Ann Neurol 1998, 43, 4-6.

Parkinson, J. F.; Mitrobvic, B.; Merrill, J. E. J. Mol. Med. (Berlin) 1997, 75, 174-
186.

Grigoriadis, N.; Tselios, T.; Deraos, S.; 0rologas, A.; Deraos, G.; Matsoukas, J .;
Mavromatis, I.; Milonas, 1. Curr. Med. Chem. 2005, 12, 1513-1519.

Barnett Michael, H.; Prineas John, W.; Institute of Clinical Neurosciences,
Department of Medicine, University of Sydney, Australia: United States, 2004, pp
458-468.

Chaudhuri, A.; Behan Peter, 0. Arch Neurol 2004, 61, 1610-1612.

Gilden, D. H. Lancet Neurol. 2005, 4, 195-202.

Gilden Donald, H. Lancet Neurol 2002, 1, 145.

Roach, E. S. Arch Neurol 2004, 61, 1615-1616.

Dyment, D. A.; Ebers, G. C.; Sadovnick, A. D. Lancet Neurol. 2004, 3, 104-110.
Coyle, P. K. Adv Neuroimmuno] 1996, 6, 143-154.

Raine, C. S. Ann Neurol 1994, 36 Suppl, 861-72.

Thompson, A. J.; Polman, C. H.; Miller, D. H.; McDonald, W. 1.; Brochet, B.;
Filippi, M. M. X.; De Sa, J. Brain 1997, 120 (Pt 6), 1085-1096.

Bagasra, 0.; Michaels, F. H.; Zheng, Y. M.; Bobroski, L. E.; Spitsin, S. V.; Fu, Z.
F.; Tawadros, R.; Koprowski, H. Proc. Natl. Acad. Sci. U. S. A. 1995, 92, 12041-
12045.

Bo, L.; Dawson, T. M.; Wesselingh, S.; Mork, S.; Choi, S.; Kong, P. A.; Hanley,
D.; Trapp, B. D. Ann Neurol 1994, 36, 778-786.

71

(17)

(18)

(19)
(20)

(21)
(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

(31)

(32)

De Groot, C. J.; Ruuls, S. R.; Theeuwes, J. W.; Dijkstra, C. D.; Van der Valk, P. J
Neuropathol Exp Neurol 1997, 56, 10-20.

Molina, J. A.; Jirnenez-Jirnenez, F. J.; Orti-Pareja, M.; Navarro, J. A. Drugs
Aging 1998, 12, 251-259.

Bredt, D. S.; Snyder, S. H. Annu Rev Biochem 1994, 63, 175-195.
Dawson, T. M.; Dawson, V. L.; Snyder, S. H. Ann. Neurol. 1992, 32, 297-311.
Nathan, C.; Xie, Q.-W. Cel] (Cambridge, Mass.) 1994, 78, 915-918.

Oset-Gasque, M. J.; Parramon, M.; Hortelano, S.; Bosca, L.; Gonzalez, M. P. .1.
Neurochem. 1994, 63, 1693-1700.

Stamler, J. S.; Singel, D. J .; Loscalzo, J. Science (Washington, D. C., 1883-) 1992,
258, 1898-1902.

Giovannoni, G.; Heales, S. J. R.; Land, J. M.; Thompson, E. J. Mult. Scler. 1998,
4, 212-216.

Bagasra, 0.; Michaels, F. H.; Zheng, Y. M.; Bobroski, L. E.; Spitsin, S. V.; Fu, Z.
F.; Tawadros, R.; Koprowski, H. Proc Natl Acad Sci U S A 1995, 92, 12041-
12045.

Smith, K. J .; Lassmann, H. Lancet Neurology 2002, 1, 232-241.

Cross, A. H.; Manning, P. T.; Stern, M. K.; Misko, T. P. J. Neuroimmunol. 1997,
80, 121-130.

Koprowski, H.; Zheng, Y. M.; Heber-Katz, E.; Fraser, N.; Rorke, L.; Fu, Z. F.;
Hanlon, C.; Dietzschold, B. Proc. Natl. Acad. Sci. U. S. A. 1993, 90, 5378.

Van Dam, A. M.; Bauer, J.; Man-A-Hing, W. K. H.; Marquette, C.; Tilders, F. J.
H.; Berkenbosch, F. J. Neurosci. Res. 1995, 40, 251-260.

Bitsch, A.; Wegener, 0; da Costa, 0; Bunkowski, S.; Reimers, C. D.; Prange, H.
W.; Bruck, W.; Department of Neurology, Georg-August-University, Gottingen,
Germany: ENGLAND: United Kingdom, 1999, pp 138-146.

Zhao, W.; Tilton, R. G.; Corbett, J. A.; McDaniel, M. L.; Misko, T. P.;
Williamson, J. R.; Cross, A. H.; Hickey, W. F. J. Neuroimmunol. 1996, 64, 123-
133.

Cowden, W. B.; Cullen, F. A.; Staykova, M. A.; Willenborg, D. O. J.
Neuroimmunol. 1998, 88, 1-8.

72

(33)

(34)

(35)

(36)

(3 7)

(38)

(39)

(40)

(41)
(42)

(43)

(44)

(45)

(46)

(47)

(48)

(49)

Okuda, Y.; Sakoda, S.; Fujimura, H.; Yanagihara, T. J Neuroimmunol 1998, 81,
201-210.

Sahrbacher, U. C.; Lechner, F.; Eugster, H. P.; Frei, K.; Lassmann, H.; Fontana,
A. Eur. J Immunol. 1998, 28, 1332-1338.

Thillaye-Goldenberg, B.; Goureau, 0.; Naud, M. C.; de Kozak, Y. J.
Neuroimmunol. 2000, 110, 31-44.

Zielasek, J.; Jung, S.; Gold, R.; Liew, F. Y.; Toyka, K. V.; Hartung, H. P. J.
Neuroimmunol. 1995, 58, 81-88.

Willenborg, D. 0.; Staykova, M. A.; Cowden, W. B. J. Neuroimmunol. 1999,
100, 21-35.

Singh, V. K.; Mehrotra, S.; Narayan, P.; Pandey, C. M.; Agarwal, S. S. Immunol.
Res. 2000, 22, 1-19.

Rand, M. J. Clin. Exp. Pharmacol. Physiol. 1992, 19, 147-169.

Toda, N.; Tanaka, T.; Ayajiki, K.; Okamura, T. Neuroscience (Oxford) 2000, 96,
393-398.

Brian, J. E., Jr.; Faraci, F. M. Stroke 1998, 29, 509-515.
Faraci, F. M.; Breese, K. R. Circ. Res. 1993, 72, 476-480.

Kobari, M.; Fukuuchi, Y.; Tomita, M.; Tanahashi, N.; Yamawaki, T.; Takeda, H.;
Matsuoka, S. Brain Res. Bull. 1993, 31, 443-448.

Giovannoni, G. Multiple Sclerosis 1998, 4, 27-30.

Mazzanti, L.; Faloia, E.; Rabini, R. A.; Staffolani, R.; Kantar, A.; Fiorini, R.;
Swoboda, B.; De Pirro, R.; Bertoli, E. Clin Biochem 1992, 25, 41-46.

Blair, G. W.; Matchett, R. H. J Neurol Neurosurg Psychiatry 1972, 35, 730-731.

Karlov, V. A.; Savin, A. A.; Smertina, L. P.; Redchits, E. G.; Seleznev, A. N.;
Svetailo, L. 1.; Margosiuk, N. V.; Stulin, I. D.: USSR, 1990, pp 47-50.

Simpson, L. 0.; Shand, B. 1.; Olds, R. J .; Larking, P. W.; Amott, M. J. Pathology
1987, 19, 51-55.

Ghigo, D.; Bosia, A.; Pagani, A.; Pescarmona, G. P.; Pagano, G.; Lenti, G. Ric
Clin Lab 1983, 13 Suppl 3, 375-381.

73

(50)

(51)

(52)

(53)

(54)

(55)

(56)

(57)

(58)

(59)

(60)

(61)

(62)

(63)

(64)

(65)

Matkovics, B.; Kotorrnan, M.; Varga, I. S.; Hai, D. 0.; Roman, F.; Novak, Z. Acta
Physiol Hung 1997, 85, 99-106.

Carroll, J.; Raththagala, M.; Subasinghe, W.; Baguzis, S.; Oblak, T. D. a.; Root,
P.; Spence, D. Mo]. BioSyst. 2006, 2, 305-311.

Sprague, R. S.; Stephenson, A. H.; Dimmitt, R. A.; Weintraub, N. A.; Branch, C.
A.; McMurdo, L.; Lonigro, A. J. American Journal of Physiology 1995, 269,
H1941-H1948.

Sprague, R. S.; Ellsworth, M. L.; Stephenson, A. H.; Lonigro, A. J. American
Journal of Physiology 1996, 271, H2717-H2722.

Sprague, R. S.; Ellsworth, M. L.; Stephenson, A. H.; Kleinhenz, M. E.; Lonigro,
A. J. American Journal of Physiology 1998, 275, H1726-H1732.

Collins, D. M.; McCullough, W. T.; Ellsworth, M. L. Microvascular research
1998, 56, 43-53.

Dietrich, H. H.; Ellsworth, M. L.; Sprague, R. S.; Dacey, R. G., Jr. American
Journal of Physiology 2000, 278, H1294-H1298.

Dietrich, H. H.; Kajita, Y.; Dacey, R. 0., Jr. The American journal of physiology
1996, 271, H1109-1116.

Ellsworth, M. L.; Forrester, T.; Ellis, C. G.; Dietrich, H. H. Am J Physiol 1995,
269, H2155-2161.

McCullough, W. T.; Collins, D. M.; Ellsworth, M. L. American Journal of
Physiology 1997, 272, H1886-H1891.

Chien, S. Annu Rev Physiol 1987, 49, 177-192.

Albalak, A.; Streichman, S.; Marmur, A. Chem. Eng. Commun. 1996, 152-153, 5-
15.

Evans, E.; Fung, Y. C. Microvasc Res 1972, 4, 335-347.

Sutera, S. P.; Gardner, R. A.; Boylan, C. W.; Carroll, G. L.; Chang, K. 0;
Marvel, J. S.; Kilo, C.; Gonen, B.; Williamson, J. R. Blood 1985, 65, 275-282.

Bessis, M.; Feo, C. Actual Pharmacol (Paris) 1981, 33, 39-55.

Stozicky, F.; Blazek, V.; Muzik, J. J Biomech 1980, 13, 417-421.

74

(66)

(67)

(68)
(69)

(70)

(71)
(72)

(73)

(74)

(75)

(76)

(77)

(78)

Riquelme, B.; Foresto, P.; D'Arrigo, M.; Filippini, F.; Valverde, J. Clin
Hemorheol Microcirc 2006, 35 , 277-281.

Ogura, E.; Abatti, P. J.; Moriizurni, T. IEEE Trans Biomed Eng 1991, 38, 721-
726.

Koutsouris, D.; Hanss, M.; Skalak, R. Biorheology1983, 20, 779-787.
F lieger, R.; Grebe, R. Biorheology 1997, 34, 223-234.

Edwards, J.; Sprung, R.; Spence, D.; Sprague, R. Analyst (Cambridge, United
Kingdom) 2001, 126, 1257-1260.

Sprung, R.; Sprague, R.; Spence, D. Analytical Chemistry 2002, 74, 2274-2278.

Kimura, C.; Oike, M.; Ohnaka, K.; Nose, Y.; Ito, Y. Journal of Physiology
(Oxford, United Kingdom) 2004, 554, 721-730.

Janigro, D.; Nguyen, T. S.; Gordon, E. L.; Winn, H. R. American Journal of
Physiology 1996, 270, H1423-H1434.

Spence Dana, M.; Torrence Nicholas, J.; Kovarik Michelle, L.; Martin, R. S.
Analyst 2004, 129, 995-1000.

Archer, S. FASEB Journal 1993, 7, 349-360.

Gumbleton, M.; Audus, K. L. Journal of Pharmaceutical Sciences 2001, 90,
1681-1698.

Audus, K. L.; Borchardt, R. T. Annals of the New York Academy of Sciences
1987, 507, 9-18.

Pal, D.; Audus, K. L.; Siahaan, T. J. Brain Research 1997, 747, 103-113.

75

CHAPTER 3

3.1 THE BLOOD BRAIN BARRIER

The existence of a separation between the blood circulation and the central
nervous system (CNS) was ﬁrst acknowledged by Paul Ehrlich in the 1880s in a trivial
experiment using staining with certain dyes. He made the observation that when the dyes
were injected into the circulatory system, all organs in the body were stained with the
exception of brain and spinal cord.1 Later on, Edwin Goldman, a student of Ehrlich
discovered that the same dyes were able to stain the nervous system when injected into
CNS and failed to stain the rest of the body.2 It was not until 1967 when Reese and
Kamovsky showed using electron-microscopic studies that the barrier is localized at the
capillary endothelial cells within the brain.3

The blood brain barrier (BBB) (Figure 13)4 represents the wall of the capillaries
that feed the brain and it is comprised of microvascular endothelial cells that present tight
junctions with very high electrical resistance (~ 2000 Q x cmz).5 The barrier restricts the
passage of molecules between the systemic circulation and the central nervous system,
thus blocking the entrance of unwanted cytotoxic agents or viruses circulating in the
bloodstream. However, at the same time, the barrier prevents the entrance of a variety of
drugs designed to target the CNS. The membranes of microvascular endothelial cells

selectively express a speciﬁc transport system to mediate the direct transport of nutrients

76

   
 

    

Basal
lamina

  
   

Insulin. LDL.

iron Tf. leptin
receptors

 
 

Amino-acid
transporters

 
   
        
   
    

Glucose
transporter

  

Figure 13. The Blood Brain Barrier. a) The BBB is formed by endothelial cells of the

level of the cerebral capillaries. These endothelial cells interact with
perivascular elements such as basal lamina and closely associated astrocytic
end-feet processes, perivascular neurons and pericytes to form a functional
BBB. b) Cerebral endothelial cells are unique in that they form complex tight
junctions (TJ) produced by the interaction of several transmembrane proteins
that effectively seal the paracellular pathway. These complex molecular
junctions make the brain practically inaccessible for polar molecules, unless
they are transferred by transport pathways of the BBB that regulate the
microenvironment of the brain.4

77

into the CNS and of toxic metabolites out of the CNS.6 Currently, about 15 transporters
are known, however it is estimated that at least 50 more exist.7

Recently, the transporter proteins in the endothelial membranes have been exploited as a
real option of delivering drugs across the barrier.

Brain endothelial cells are very distinct from any other endothelial cell in the
body due to their morphology, biochemistry and function. Although they are the main
structural constituents of the barrier, their interaction with other specialized cells such as
pericytes, perivascular macrophages and astrocytic end-feet is a prerequisite for the
proper function of the BBB. In order to study the drug transport across the barrier, a
suitable in vitro model of the BBB has to be developed, where the quality of the barrier

and the differentiated morphology of its endothelium are highly preserved.8

3.2 BLOOD BRAIN BARRIER MODELS

In the last decade, an increasing effort to create in vitro models of the BBB was
observed for studies of drug delivery to the CNS. Many drugs that showed promise in
studies of CNS disorders are too large (tens or hundreds of kDa) for the passive diffusion
through the brain-endothelium barrier. Therefore, a cell-based model that offers the
potential of mimicking in vivo conditions, such as transcellular and paracellular drug
diffusional processes, metabolism, and active transport processes is highly needed.9

Most of the BBB models exempliﬁed in the literature are based on stationary
cultures and co-culture of cells that take into account the complex environment of the

brain. Early in vitro BBB models were described in terms of tight junction parameters

78

that have to be met, selective permeability properties for different standard sugar
solutions and polarity. The ﬁrst three prototypes of in vitro models consisted of either

10-13

suspensions of isolated cerebral microvessels, isolated brain endothelial cells

16,17

”"5 and co-culture of astrocytes and endothelial cell monolayers on plastic or

cultures,
on either side of ﬁlters.”‘22
Polycarbonate membranes with varying pore sizes were also utilized as support to

'9’23 and to allow

grow endothelial cells,8 or co-culture of astrocytes and endothelial cells
the passage of large drug molecules for transport studies. Typically, BBB models
consisted of endothelial cells cultured on top of a polymeric membrane afﬁxed to a
cylindrical plastic insert. This culture insert was placed into a well of a standard culture
plate, dividing the well into the luminal (blood) and abluminal (brain) chambers (Figure
14).4

Recently, Ma et a]. reported an in vitro BBB model that utilizes membranes
nanofabricated from low stress silicon nitride (SiN). They hypothesized that growing
endothelial cells and astrocytes on the opposite sides of the membrane, therefore
increasing the amount of direct contact between the two while still maintaining two
distinct cell layers, will create a more differentiated and appropriate BBB model.24

Some efforts have been made to develop BBB models with a three-dimensional
architecture that will account for the blood flow differentiation of the cerebral
endothelium.25 Thus, endothelial cells have been culture inside hollow ﬁber tubes (the
intraluminal side) and astrocytes were cultured extraluminally on the surface of the tubes.

In this way, the endothelial cells were exposed to the ﬂow conditions and allowed the

investigation of signaling events in the BBB.

79

6-well plate

Endothelial cells

 

  
  

Luminal
. compartment

Coated (blood)

microporous
membrane

  

 

Abluminal
compartment
(brain)

... s ‘NMM.

 

Glial cells

Figure 14. Typical in vitro model of the BBB. Brain endothelial cells are grown at the
bottom of ﬁlter inserts together with glial cells grown at the bottom of 6-well
plate. Soluble factors secreted from the glial cells induce the BBB phenotype
in the brain capillary endothelium.4

80

However, technical demands of this model limit its use for drug screening.Very recently,
Cucullo et a]. utilized the aforementioned technique to develop a humanized dynamic in
vitro model to study the BBB in epilepsy.26 The model was based on the co-culture of
human microvascular endothelial cells (HMVEC) from normal and drug-resistant
epileptic brain tissue with human brain astocytes (HA) from epilepsy patients or controls.
Here, HMVEC and HA were co-cultured in polypropylene capillaries. However, for these
studies, the dynamic in vitro blood brain barrier (DIV-BBB) setup was purchased from
Spectrum (Spectrum Laboratories, Inc., Rancho Dominguez, CA).

These approaches of modeling the BBB for the pharmaceutical drug discovery
and development process had to compromise between capacity, cost, time and how
closely a model needs to resemble in vivo conditions. For successful use, any in vitro
BBB model has to fulﬁll certain criteria such as reproducible permeability of reference
compounds, good screening capacity, the display of complete tight junctions, adequate
expression of BBB phenotypic transporters, to be reasonably robust and to display a
physiologically relevant morphology.4 To date, the existing in vitro BBB models are
labor intensive, expensive and have low throughput that makes them nearly impossible

for testing the plethora of drugs designed. to target the CNS.

81

3.3 MICROFLUIDIC TECHNOLOGY FOR THE DEVELOPMENT OF A

BLOOD BRAIN BARRIER MIMIC

The use of a microfabricated device is a possible way to develop a complete BBB
mimic. More speciﬁcally, using lithographically-derived microchips, it would be possible
to monitor the fate of endothelium-derived N0 that is stimulated by mechanically
deformed RBCS. The concept of a micro total analysis system (uTAS), also called a “lab
on a chip”, miniaturized or microﬂuidic analysis systems, is a rapidly developing ﬁeld.
Microﬂuidic devices have found a broad area of applications, especially in the biological
and life sciences, from cell handling and analysis, biomimetic and biopowered systems,
clinical diagnosis, immunoassays, DNA, proteins and other bioassays to environmental
concerns and gas analysis.27

Typical dimensions of these microﬂuidic devices range from a few micrometers
to several millimeters in length and width and from 100 nm to 100 um depth and height.
These systems can be made in silicon,28 glass,28 and polymers such as polymethyl
methacrylate (PMMA),29 or poly(dimethylsiloxane) (PDMS).30 Flow of sample or
reagents through the chip is accomplished by either electrophoretic or hydrodynamic
pumpingf’l'33 Detecting the analytes directly in the micron-sized channels can be
performed by a variety of mechanisms including optical (ﬂuorescence, absorbance,
chemiluminescence, refractive index), electrochemical (amperometry, potentiometry,
conductivity), and mass spectrometry or NMR detection. The development of systems
that combined surface engineering with layer-by-layer microﬂuidic technology to create
3D tissuelike structures, e.g., blood vessels have been reported.“35 The utilization of an

ultrathin nanofabricated silicon nitride membrane, where endothelial and astrocyte cells

82

were co-cultured on opposite sides of the membrane have also been reported.24 This
approach constitutes a potential starting point for development of a BBB mimic. Even
though models of the BBB already exist, none incorporate actual blood ﬂow. Here we
report the development of a dynamic system that integrates blood ﬂow and a barrier

mimic.

3.4 EXPERIMENTAL

3.4.1 Design and Methods

Microchip fabrication. Microchips were fabricated using standard soft lithographic
technology. Micrometer size channels were patterned in PDMS based on previous
published methods.30 Masters were made on a silicon wafer (Silicon, Inc., Boise, ID)
using SU-8 photoresist (MicroChem Corp., Newton, MA) and photolithographic
procedures employing masks in the form of negative ﬁlms (Figure 15). The photoresist
was prebaked at 95°C for 5 min prior to UV exposure with a near-UV ﬂood source
(Autoﬂood 1000, Optical Associates, Milpitas, CA) through a negative ﬁlm (2400 dpi,
J ostens, Topeka, KS), which contained the desired channel structures drawn in Freehand
(PC version 10.0, Macromedia, Inc., San Francisco, CA). The masters were cast against a
20:1 mixture of Sylgard 184 elastomer and curing agent (Ellsworth Adhesives,

Germantown, WI) and allowed to cure for ~ 2 hours at 75°C.

83

1) Spin Coat

4) l'\" Exposure

 

Figure 15. Photolithographic procedure for the fabrication of silicon master. 1) Spin coat
4 mL of SU-8 50 photoresist at 1000 rpm for 20 s; 2) Pre-bake for 5 min at
95°C to evaporate solvent and densify ﬁlm; 3) Post-bake for 5 min at 95°C to
cross-link the exposed portions of the photoresist ﬁlm; 4) UV exposure, 350-
450 nm, 50W for 60 s to polymerize and cross-link; 5) Develop with
propylene glycol monomethyl ether acetate (PGMEA); 6) Soft-lithography to
create microﬂuidic chip.

84

fabrication of the microﬂuidic device proposed to mimic the BBB. For the determination

After the incubation period, the PDMS was peeled off the master to yield a pattern of
negative relief channels. Chips containing channels of 100 um depth x 100 um width x
2cm length were used for the of NO and ATP ﬂux through the polycarbonate membrane,
a spiral shaped channel with a diameter of 150 um was initially used to deliver the 4-
amino-S-methylamino-Z’7’-diﬂuoroﬂuorescein diacetate (DAF -F M) probe for N0

detection and luciferin/luciferase mixture for the ATP detection, respectively.

Preparation of N0 standards. Nitric oxide was prepared as a 38 mM stock solution
from spermine NONOate (Cayman Chemical, Ann Arbor, MI) by dissolving 10 mg of
spermine NONOate solid in 1 mL of 0.01 M sodium hydroxide (NaOH) solution.
Working solutions were prepared by dilution of the alkaline NONOate solution in 0.1 M
phosphate buffer saline (PBS, pH 7.4). Sperrnine NONOate is reported to have a half-life
of 39 min at 37°C in 0.1 M phosphate buffer and follows ﬁrst order kinetics for its rate of
decay.36 The working solutions were incubated in a water bath at 37°C for 15 min to
ensure NO release from the donor. The NO samples were hydrodynamically pumped into

the microﬂuidic channel at a ﬂow rate of luL min'1 for 10 min.

Detection of N0. Methods based on ﬂuorogenic probes for the measurement of the
production of NO in cellular systems have been gaining increasing popularity due to their
simplicity and sensitivity. The most successful indicator for N0 reported in the literature
has been 4,5-diaminoﬂuorescein diacetate (DAF-2 diacetate) and 4-amino-5-

methylamino-2’7’-diﬂuoroﬂuorescein diacetate (DAF-FM DA). These probes are

85

membrane permeable and they are deacetylated by intracellular esterases to DAF-2 and
DAF-FM, respectively.37 In all imaging studies DAF-FM or the diacetate form for cell
membrane permeation was used for ﬂuorescence determination. The reason for using
DAF-FM versus DAF—2 is that DAF-FM is a more sensitive reagent for N0 (N0
detection limits for DAF-FM ~ 3 nM versus ~5 nM for BAP-238) and DAF-FM is
independent of pH above pH 5.5. Moreover, the N0 adduct of DAF -F M is known to be

signiﬁcantly more stable than that of DAF-2, enabling more time for image capture.

Detection of A T P ﬂux through the polycarbonate membrane. ATP was measured by
chemiluminescence using a luciferin/luciferase mixture as previously described in
chapter 2, section 2.5.1. The reagents were delivered in a continuous ﬂow setup using
Tygon tubing (0.0200 mm ID x 0.0600 mm OD) ﬁtted with 15 mm, 23 gauge
hypodermic steel tubing bent at 90 degrees. This steel tubing was placed in the inlet
access holes created in fabrication of the microﬂuidic device to deliver solutions directly
into the channels. ATP standard solutions (0.0, 0.5, 1.0 and 1.5 uM) prepared in PSS
buffer were added to aliquots of RBCS and delivered in the straight channel on the
bottom of the device shown in Figure 16. The luciferin/luciferase mixture was
hydrodynamically pumped for 10 min (or until the signal is stabilized) at a ﬂow rate of
1.0 or 2.0 uL min'l into the spiral channel placed on the top of the membrane of the
microchip device (Figure 16). The RBC/ATP samples loaded in one of the syringes were
pumped though the tubing at a rate of 2.0 and 3.0 pL min". Additionally, a solution of
iloprost (a prostacyclin analogue known to induce ATP release) at 1 11M in PBS was

incubated with a solution of RBC (ﬁnal hematocrit 7%) for 15 min and hydrodynamically

86

Luciferin/luciferase
sample inlet

 

Top PDMS slab

 

Bottom PDMS

Polycarbonate
membrane

 

.
o
n - .
.I
s l\.' .‘I .1
.- o ‘ . . , .
s - I
,, I
I
<
I
a

l ‘ I
~ . ’ ‘ . .'.
I-.. 22* ‘ 3 ‘-.‘— ..."m' ' _ ‘
. .I \
. "a . - 9 \
, - . . , \
' '-' \
I \
\
\
\
- \

 

\

 

 

20:1

 

 

RBCS + ATP
sample inlet

Detection

7 window

Waste

’ reservoirs

Figure 16. Detection of ATP ﬂux through a polycarbonate membrane using a ﬂow of
luciferin/luciferase mixture hydrodynamically pumped into a spiral-shaped
channel of a PDMS device. The set up is enclosed in a light excluding box
over a PMT.

87

pumped through the bottom channel for 30 min. As the RBC/ATP or RBC/iloprost
solutions traverse the microﬂuidic channel on the bottom of the device, ATP diffuses
through the polycarbonate membrane into the spiral channel, reacts with the
luciferin/luciferase mixture and the resulting chemiluminescence is detected at the end
segment of this channel. The spiral channel intersects the linear channel underneath it in
17 locations, therefore creating a total contact surface area of 17 x 150 um (width of the
spiral channel) x 100 um (width of the linear channel). Chemiluminescence intensity was
measured in a 1 mm long linear segment (x) situated at the end of the spiral-shaped

channel using a photomultiplier tube (PMT) housed in a light-excluding box.

Fabrication of the microchip-based BBB. Figure 17 represents a schematic of the
microﬂuidic device used to investigate the diffusion of N0 through a polycarbonate
membrane with pore size of 0.2 pm and a diameter of 13 mm. The proposed device is
constructed of two PDMS slabs, each having patterned a 2 cm long, 100 um wide channel
and a polycarbonate membrane sandwiched in between the slabs. The membrane is
tightly and reversibly sealed by the two pieces of PDMS, making the device amenable to

cleaning and reuse.

88

 

//
/
/ c/ ./
//

'5 (“gt-Vega

a ”3’5“
0.2 m,13mm 1535'?“ " "4" “1951;;
P. ‘f‘tt-sz' m ;;.-.,. 35

 

DAF-FM

 

 

 

Figure 17. Proposed design for the microﬂuidic-based BBB mimic A. Assembly of a 3
dimensional ﬂuidic device incorporating microchannels (100 um diameter)
for reagents delivery and a polycarbonate membrane sandwiched in between
the PDMS layers. B. Micrograph of the channels junction with the
polycarbonate membrane (13 mm diameter and 0.2 pm pore size).

89

3.4.2 Results and Discussion

In this work we described two approaches for the construction of a microﬂuidic-
based BBB. The ﬁrst approach consists of immobilization of bovine brain nricrovascular
endothelial cells (bBMVECs) in the microﬂuidic channel underneath the membrane and
monitoring of the NO production by the endothelial cells in the upper channel
representing the CNS. A second approach is to culture the brain endothelial cells onto the
polycarbonate membrane that will separate the two channels that pose as the
microvascular system and CNS, respectively.

In the ﬁrst approach, we are interested in optimizing the NO ﬂux though the
porous membrane in order to obtain the appropriate conditions for the detection of the
analyte of interest. Moreover, here we show that endothelial cells can be immobilized on
both PDMS channels and a coated-polycarbonate membrane.

For the second approach, the optimization of the ATP ﬂux through the
polycarbonate membrane is a prerequisite for the stimulation and detection of the NO
production in bBMVECs cultured on the membrane. Ultimately, the ATP that is
measured will be released by the RBCS obtained from patients with multiple sclerosis

(MS).
N0 ﬂux through the polycarbonate membrane. In order to develop a realistic mimic of

the BBB, the ﬂux of N0 across the polycarbonate membrane has to be determined.

Therefore, a 3D microﬂuidic device was designed with one channel that ultimately

90

mimics the brain microvessels (cerebro-microvasculature), and is separated from a
second channel that will pose as the CNS by a polycarbonate membrane.

One important step in the development of the device is the optimization of
parameters such as ﬂow rates, membrane pore sizes, channels lengths, and channel
design in order to achieve high sensitivity and good limits of detection. These ﬁgures of
merit are particularly important when investigating complex biological systems, such as
the BBB and its periphery.

The ﬁrst step in achieving this goal is the direct detection of NO that diffuses
through the membrane with pore sizes 0.2, 0.4, and 0.6 pm. The ﬂuorescent probe DAF-
F M (10 11M) that ﬂows through the second channel, above the membrane was used to
create a highly ﬂuorescent adduct with N0. Figure 18A schematically shows the
diffusion of NO that is ﬂowing in the channel S1, through the polycarbonate membrane
into the upper channel S2 that contains the probe. The highly ﬂorescent DAF-FM/NO
adduct is formed upon reaction between N0 and DAF-FM and detected using a CCD
camera mounted on an inverted microscope. The micrographs in Figure 18B show the
diffusion of N0 through the porous membrane and subsequent reaction with the probe at
the channels junction.

The polycarbonate membrane experiences a uniform pressure drop across the
entire length and width of the membrane at the channels junction (100 um x 100 pm). A
membrane pore size of 0.2 pm and ﬂow rates of 1.0 11L min'1 for both NO and DAF-FM
streams were found to be optimal for the efﬁcient diffusion of NO through the

membrane. The bar graph in Figure 19 shows that with increasing N0 concentration from

91

Microﬂuidic channles

'
’
I
C
'
'
I
a
I
'
ﬂ
.
'
‘
'
I
l
I
l
I
.
a r ,"
..
.. .
. 1
~,‘ 4 , .
' >1 . .
’- 1 '
v . . .
. \’\. . ' ‘ ‘ . v '
“ . , I ‘ . ..o
._ \ . a . - -
- .. o . - _ o
7‘ ' ' v ‘ I
“ ~ , " C -
r. . , ﬂ . o
‘ .
. - ' .
.
. .

S1 7‘

Polycarbonate membrane

 

Figure 18. A. Detailed representation of the microﬂuidic channels junction with
polycarbonate membrane in between. $1 represents the channel under the
membrane, where NO (3.8 11M stock solution in PBS) is being
hydrodynamically pumped at a ﬂow rate of l 11L min". S2 represents the
upper channel that poses as the CNS, where DAF-FM (10 nM) is being
introduced. B. Micrographs of the upper channel before (1), at (2), and
after (3) channels junction showing the formation of the highly ﬂuorescent
DAF-FM/NO adduct.

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Q 1.64
U)
1:
4% *4:
.0)
g 1.4- T
8 f‘ ,1?
g * rill r-l‘l
O 4' M
a T ,‘Iip‘lvr
,c 1.2- ‘1. 15.1“:
B (“I ,r.
= T . ~ .li‘f’
E WIT.“ i (ll.
. I
2 1.0‘ nil-1152511111 EV .l'l',
'l,
I “if“
. i it; .1
, I ”.5”
0.8 . ‘ l " .
0 100 200 500
N0conc.(nM)

Figure 19. Detection of NO ﬂux through the polycarbonate membrane with a pore size of

0.2 pm and at a ﬂow rate of 1 (1L min‘l as assessed by DAF-FM (10 nM).
The control sample is PBS buffer. *, **, and *** represent the statistical
signiﬁcant differences versus control, 100, and 200 nM NO, respectively (*p
5 0.001; ** and ***p S 0.05). The error bars represent the mean values i
standard error (n=4).

93

0 (PBS only) to 500 nM, an increase in the ﬂuorescence intensity is being detected across
the membrane in the upper channel S2.

It was found that concentrations as low as 100 nM of N0 can be detected using
the proposed device. Moreover, signiﬁcant statistical differences between 100 and 200
nM concentrations of N0 were also reported. The data proves that the ﬁrst approach for
the microﬂuidic device described in this work in conjunction with the ﬂuorescence
detection could be utilized to investigate nanomolar concentrations of N0. However, the
NO production in the brain endothelial cells that is stimulated by RBC derived ATP is
one to two orders of magnitude lower. Therefore, the ﬁrst approach for our device is not
sensitive enough for the low concentration of NO produced by the endothelial cells, both
in vivo and in vitro.

Based on these results, efforts have been directed to the immobilization of
endothelial cells directly on the polycarbonate membrane, therefore the detection of
intracellular NO production could be monitored in real time using ﬂuorescence

microscopy.

Cell immobilization on polycarbonate membrane/PDMS channels. Isolated bBMVECs
have been previously successfully cultured as primary cultures on collagen-treated solid

”"5 polycarbonate membranes, or on

plastic surfaces (standard tissue culture ﬂasks),
commercial membrane inserts.8 At date, there are no reports in the literature describing
bBMVEC immobilized onto PDMS substrates or channels as a part of a microﬂuidic

device.

94

Despite the fact that bBMVECs are small (< 10 pm in diameter), and that they
generally need ~ 10-12 days to reach a conﬂuent layer, we show here a successful
attempt to immobilize these cells in a 100 pg mL'l ﬁbronectin-coated 100 um diameter
channel of a PDMS chip (Figure 20). However, the method of immobilizing the cells in
the PDMS channels and the preservation of cell viability and morphology for the
experimental period proved to be highly challenging. The lack of reproducibility and the
high cost of the brain microvascular cell line were taken into consideration and a second
approach was therefore envisioned.

Here, the cells were immobilized on a polycarbonate membrane in order to make
the bBMVEC cell-coated membrane a part of a microﬂuidic device that will closely

mimic the in vivo BBB and its proximal environment.

Detection of ATP ﬂux through the polycarbonate membrane. RBCs 7% and ATP
standard solutions (0, 0.5, 1.0 and 1.5 nM) were prepared as previously described in
chapter 2. RBC solutions were spiked with increasing concentrations of ATP and the ﬂux
of ATP thought the polycarbonate membrane of the PDMS chip was detected as the
chemiluminescence signal observed at the end portion of the spiral channel (distance x in
Figure 15). The chemiluminescence signal is proportional to the amount of ATP reacted
with the luciferin/luciferase mixture. In this case, a spiral-shaped channel (150 um
diameter) was used to increase the time for ATP molecules to diffuse through the pores
of the membrane.

The reason for using ATP-spiked RBC samples instead of ATP standard solutions

in buffer was to account for the complexity of the RBC matrix and the possible

95

 

Figure 20. Micrograph of bBMVEC immobilized in a 100 pm PDMS channel. The
channel was pre-coated with ﬁbronectin (100 pg mL") and incubated for 1 h
at 37°C and 5% C02 to ensure cell adhesion.

96

interferences in the ATP detection that could occur when RBCS are present. Figure 21
shows that low micromolar ATP concentrations added to a continuous ﬂow of RBCS can
be detected on the other side of the polycarbonate membrane. These ﬁndings suggest that
ATP diffusion though the membrane does not constitute a limiting factor in stimulating
the NO production in endothelial cells cultured on the membrane. Polycarbonate
membranes with pores sizes of 0.2, 0.6 and 1.0 pm were used in these studies. The
optimum pore size diameter was found to be 1.0 pm. These data was found to be in good
agreement with the literature. Recently, it has been reported that there is approximately

four-fold smaller permeability of the 0.6 pm pore membranes than the 1.0 pm pore

membranes (kllum/kggﬂm ~ 4) for a standard solution of ADP.39

The ATP ﬂux was manipulated by varying parameters such as relative ﬂow rate
and concentration. The ﬂow rates for both ATP and luciferin/luciferase streams were
optimized to increase the sensitivity of the method. Therefore, we report here an optimum
ﬂow rate of 3.00 pL min'1 for ATP ﬂow, and of 1.00 pL min'l for luciferin/luciferase,
respectively. Using this device we were able to detect as low as 500 nM concentration of
ATP added to RBCS. Additionally, iloprost (a stable analogue of prostacyclin that is
known to stimulate ATP release via a G-protein coupled receptor mechanism”) was used
to provoke ATP release. Data in Figure 20 shows that indeed, pharmacologically induced
ATP release from RBC can diffuse though the polycarbonate membrane and be
efﬁciently detected using chemiluminescence assay and a PMT housed in a light

excluding box.

97

 

 

 

 

Normalized chemiluminescence 1ntensrty
_L
N
I
l-
- _1=———l
5}

 

 

 

 

 

 

 

 

 

 

 

 

 

I ll ' I l *1
. l l ‘ 'A “fight
1-0 - IIVDTIIIIII . ‘ 1.;-.2. (l' “I
. l , “ll. ,.' ’31,, 1'. .l .1111
I " .‘1'1
I , I" ' 1: I.
0-8 .1 l 1 ‘11": .1 11"" ‘
. 1 7 iii
. r . ,‘1‘1‘ .
. .1 .. .1. l
. l r '1' 1 ‘ 1', 1 1“ I I ’1
0.6 WWI..." . L... , | m... I . l . . I . i I
0 0.5 1 1.5 Iloprost
ATP conc. (pM)

Figure 21. Detection of the ATP ﬂux through a polycarbonate membrane with pore size
of 1.0 pm and 13 mm diameter as assessed by chemiluminescence in a
spiral-shaped channel, as part of a 3D microﬂuidic device. ATP and
luciferin/luciferase ﬂow rates are 3 pL min'1 and 1pL min", respectively.
Control represented by RBCS 7% (0 pM ATP). Iloprost is used as an ATP
release stimulant of RBCS. * and ** represent the values statistically
different from control and 0.5 pM ATP, respectively (*p 5 0.001, **p S
0.05) (n=4).

98

3.5 CONCLUSIONS AND OTHER CONSIDERATIONS

In this work we hypothesized that a complete BBB mimic can be developed on a
microﬂuidic device that will monitor the fate of endothelium-derived NO that is
stimulated by mechanically deformed RBCS. Key to investigating the NO production
from the endothelium and its fate through the cell-coated polycarbonate membrane, that
will constitute the BBB, is the ability to actually detect the N0 after it crosses the barrier.

Primarily, we have demonstrated here that the ﬂux of N0 (ﬂow rate 1.0 pL min'l)
through a bare polycarbonate membrane can be detected using a speciﬁc probe for N0,
DAF -F M (IOpM). However, the device lacks the sensitivity required to detect
physiological relevant levels of N0. Secondly, the ATP ﬂux through a polycarbonate
membrane (ﬂow rate 3.0 pL min") in a 3D microﬂuidic device has been investigated.
The molecular transfer through the membrane was maximized by increasing the area of
contact between the two channels by designing the upper channel in a spiral shape that
crosses the underneath linear channel in multiple points (17 x 150 pm x 100 pm). We
showed here that the proposed device is sensitive enough for the detection of as low as
500 nM of ATP spiked into RBC samples and most importantly, is able to detect
concentrations of ATP released by iloprost-treated RBCS.

Parameters such as membrane pore size, ﬂow rates and probe concentrations have
been optimized to increase the efﬁciency of the detection in the device. Although the
results are promising, the inability to culture the cells on the membrane as part of the 3-D
microﬂuidic BBB and the poor sensitivity for the physiological relevant N0

concentrations constitute limiting factors.

99

Attempts have been made to culture the endothelial cells on isolated collagen-
coated polycarbonate membranes and, after the cells have reached a conﬂuent layer to
integrate the membrane in the device. However, the two pieces of PDMS that make up
the device would not seal together as a result of the membrane wetness from the cell
culture media. As an alternative method, an attempt was made to culture the cells directly
on the membrane by introducing the cell suspension in the upper channel, above the
membrane. However, the lack of reproducibility and technical difﬁculties with the cell
introduction made the device unsuitable for the purpose of our applications.

One way of avoiding these complications would be the integration of microvalves
and multiple channels for the introduction of cells and DAF-F M, separately. Moreover,
cell culture reactors on-chip where the viability of the cells can be sustained for days and
the cell-cell signaling can be monitored in real time could also constitute an alternative
option. In terms of using the polycarbonate membranes as support for endothelial cell
culturing and proliferation, integration of transparent membranes for ease of optical cell
imaging would be highly desirable.

The aforementioned microﬂuidic components, such as microvalves, microreactors
or intricate designs are time consuming and prone to problems of engineering nature even
before introducing the biological factor. A new, simpler and more amenable design has
thus been constructed. Here, the high throughput capabilities of a nricrotitre plate
technology was merged with the capabilities of microﬂuidic technology to create a device
that mimics the BBB in a more realistic way by incorporating blood ﬂow and having the

capability of possible multi-drug screening simultaneously and in real time.

100

(1 )
(2)
(3)
(4)

(5)
(6)
(7)
(8)
(9)

(10)
(11)

(12)

(13)

(14)
(15)
(16)
(17)

LIST OF REFERENCES

Ehrlich, P. Ber]. Klin. Wochenschr. 1907, 44, 233-236.
Goldmann, E. E. Lancet 1913, 172, 1183-1188.
Reese, T. S.; Karnovsky, M. J. J Cell Biol 1967, 34, 207-217.

Cecchelli, R.; Berezowski, V.; Lundquist, S.; Culot, M.; Renftel, M.; Dehouck,
M.-P.; Fenart, L. Nat. Rev. Drug Discovery 2007, 6, 650-661.

Crone, C.; Olesen, S. P. Brain Res 1982, 241, 49-55.

Engelhardt, B. Cell Tissue Res. 2003, 314, 119-129.

Miller, G. Science (Washington, DC, United States) 2002, 297, 1116-1118.
Guerin, J. L.; Bobilya, D. J. Neurochem. Res. 1997, 22, 321-326.

Gumbleton, M.; Audus, K. L. Journal of Pharmaceutical Sciences 2001, 90,
1681-1698.

Betz, A. L.; Goldstein, G. W. Science 1978, 202, 225-227.

Bowman, P. D.; Ennis, S. R.; Rarey, K. E.; Betz, A. L.; Goldstein, G. W. Ann
Neurol 1983, 14, 396-402.

Matsumura, K.; Abe, 1.; Fukuhara, M.; Tominaga, M.; Tsuchihashi, T.;
Kobayashi, K.; Fujishima, M. Am J Physiol 1994, 266, R1403-1410.

Abbott, N. J .; Hughes, C. C.; Revest, P. A.; Greenwood, J. J Cell Sci 1992, 103 (
Pt 1), 23-37.

Audus, K. L.; Borchardt, R. T. Ann N Y Acad Sci 1987, 507, 9-18.
100, F. J Neurochem 1992, 58, 1-17.
Arthur, F. E.; Shivers, R. R.; Bowman, P. D. Brain Res 1987, 433, 155-159.

Laterra, J .; Guerin, C.; Goldstein, G. W. J Cell Physiol 1990, 144, 204-215.

101

(18)

(19)

(20)

(21)
(22)

(23)

(24)

(25)

(26)

(27)
(28)
(29)

(30)

(31)

(32)

(33)

(34)

(35)

Beck, D. W.; Vinters, H. V.; Hart, M. N.; Cancilla, P. A. J Neuropathol Exp
Neurol 1984, 43, 219-224.

Dehouck, M. P.; Meresse, S.; Delorme, P.; Fruchart, J. C.; Cecchelli, R. J
Neurochem 1990, 54, 1798-1801.

Raub, T. J .; Kuentzel, S. L.; Sawada, G. A. Exp Cell Res 1992, 199, 330-340.
Biegel, D.; Pachter, J. S. In Vitro Cell Dev Biol Anim 1994, 30A, 581-588.
Deli, M. A.; Joo, F. Keio J Med 1996, 45, 183-198; discussion 198-189.

Wolburg, H.; Neuhaus, J.; Kniesel, U.; Krauss, B.; Schmid, E. M.; Ocalan, M.;
Farrell, C.; Risau, W. J Cell Sci 1994, 107 ( Pt 5), 1347-1357.

Ma, S. H.; Lepak, L. A.; Hussain, R. J.; Shain, W.; Shuler, M. L. Lab on a Chip
2005, 5, 74-85.

Stanness, K. A.; Westrum, L. E.; Fomaciari, E.; Mascagni, R; Nelson, J. A.;
Stenglein, S. G.; Myers, T.; Janigro, D. Brain Res 1997, 771, 329-342.

Cucullo, L.; Hossain, M.; Rapp, E.; Manders, T.; Marchi, N.; Janigro, D.
Epilepsia 2007, 48, 505-516.

Dittrich, P. S.; Tachikawa, K.; Manz, A. Analytical Chemistry, ACS ASAP.
Dolnik, V.; Liu, S.; Jovanovich, S. Electrophoresis 2000, 21, 41-54.
Becker, H.; Gartner, C. Electrophoresis 2000, 21, 12-26.

McDonald, J. 0; Duffy, D. C.; Anderson, J. R.; Chiu, D. T.; Wu, H.; Schueller,
0. J. A.; Whitesides, G. M. Electrophoresis 2000, 21, 27-40.

Reyes, D. R.; Iossiﬁdis, D.; Auroux, P.-A.; Manz, A. Analytical Chemistry 2002,
74, 2623-2636.

Polson, N. A.; Hayes, M. A. Analytical Chemistry 2001, 73, 312A-319A.

Seiler, K.; Fan, Z. H.; Fluri, K.; Harrison, D. J. Analytical Chemistry 1994, 66,
3485-3491.

Tan, W.; Desai, T. A. Journal of Biomedical Materials Research, Part A 2005,
72A, 146-160.

Tan, W.; Desai, T. Encyclopedia of Biomaterials and Biomedical Engineering
2004, 2, 1580-1593.

102

(36)

(37)

(38)

(39)

(40)

Maragos, C. M.; Morley, D.; Wink, D. A.; Dunams, T. M.; Saavedra, J. E.;
Hoffman, A.; Bove, A. A.; Isaac, L.; Hrabie, J. A.; Keefer, L. K. J Med Chem
1991, 34, 3242-3247.

Balcerczyk, A.; Soszynski, M.; Bartosz, G. Free Radical Biology & Medicine
2005, 39, 327-335.

Kojima, H.; Nakatsubo, N.; Kikuchi, K.; Kawahara, S.; Kirino, Y.; Nagoshi, H.;
Hirata, Y.; Nagano, T. Analytical Chemistry 1998, 70, 2446-2453.

Neeves, K. B.; Diamond, S. L. Lab Chip 2008, 8, 701-709.

Olearczyk Jeffrey, J .; Stephenson Alan, H.; Lonigro Andrew, J.; Sprague Randy,
S. Am J Physiol Heart Circ Physiol 2004, 286, H940-945.

103

CHAPTER 4

4.1 ADDRESSING A VASCULAR ENDOTHELIUM ARRAY WITH BLOOD

COMPONENTS USING UNDERLYING MICROFLUIDIC CHANNELS

Many technologies, such as those designed to perform adsorption, distribution,
metabolism, excretion, and toxicity (ADMET) screening, are beginning to emerge as
signiﬁcant tools in drug discovery1 and systems biology2 and there exists a number of
motivations for continued advancements in this area. A logical way to improve the drug
discovery process at the preclinical stage is to create technologies that incorporate
controlled, in vitro environments that closely mimic in vivo conditions and enable
molecular-mediated communication between different cell types. Such technologies have
been previously reported?”4 Unfortunately, most systems fail to incorporate a major
component of many in vivo processes, speciﬁcally, components of the circulation.

A quantitative understanding of molecule-mediated communication between
different cell types in the circulation is challenging due to chemical and physical
constraints. Chemically, the dynamic metabolic status of the cells dictates that
measurements involving multiple cell types are made almost simultaneously as cellular
events occur,2thus enabling the communication between the cell types.5 Physically,
communication between two different cell types are often initiated by some type of stress

10,11

placed upon one of the cell types. Hypoxia,6 shear stress,” and oxidative stress are all

types of stress that affect cell signaling. Measuring such in vivo events in real-time and in

104

a high throughput manner is difﬁcult, especially if one of the cell types is in the complex
matrix of the bloodstream.

There have been many reports involving microﬂuidics or lab-on-a-chip
technology to create in vivo mimics using a controlled, in vitro platform with many of
these devices incorporating cellular components.”"15 Stimulation has also been carried
out on various cell types including the stimulation of a dopaminergic cell model (PC12
cells),16 endothelial cells,17 and islets18 contained within microﬂuidic devices.

Although previous reports have described the successful direct immobilization of
endothelial cells in the channels of a microﬂuidic device, the device described here
utilized a polycarbonate membrane to separate the ﬂow of red blood cells (RBCS) from
the endothelial cells. First, the use of the membrane in the device enables already
available technologies associated with microtitre plate technology. Furthermore, the use
of the membrane will allow other cell types to be immobilized on or near the
endothelium, thus creating a working mimic of the blood brain barrier in a three
dimensional device.

Such a device has the potential to examine communication between the
circulation (in particular, RBCS) and the nervous system. Here, we report a method that
combines microﬂuidic technologies in an array setup to elucidate the communication
skills between RBCs and bovine pulmonary artery endothelial cells (bPAECs). This
array-type device represents an improvement from the previous design with two single
channels patterned onto PDMS chips (representing the circulatory system and the CNS)

and a sandwiched polycarbonate membrane in between.

105

4.2 EXPERIMENTAL

4.2.1 Material and methods

Preparation of RBCS. RBCS were obtained from male New Zealand White rabbits as
previously described. RBC samples were prepared as a 7% hematocrit with and without 1
pM iloprost (Cayman Chemical, Ann Arbor, MI) which, when added, was incubated off-
chip for approximately 15 min. Iloprost was also added to a sample containing 7% RBCS
incubated with 250 nM glybenclamide for approximately 10 min (Sigma Chemical, St.
Louis M0) to demonstrate that the additional ATP release due to iloprost was not a result
of cell lysis. ATP standards ranging from 0 -1.5 pM were prepared in a physiological
saline solution (PSS) by diluting a 100 pM stock solution made up in distilled and

deionized water (DDW) to allow for the quantitation of ATP release.

Measurement of ATP via chemiluminescence. Each sample or standard solution was
individually mixed at the T-junction within the device with a luciferin/luciferase solution
to examine the iloprost stimulated, RBC-derived ATP. The sensitivity of the assay was
increased by adding 2 mg of luciferin (Sigma Chemical, St. Louis, M0) to a vial of
luciferase and luciferin (FLE-SO, Sigma Chemical, St. Louis, MO). The
chemiluminescence for all samples as well as ATP standards were measured in real time
by a photomultiplier tube (PMT) at 10 Hz for 30 seconds as previously described. These
chemiluminescence measurements were performed by placing the downstream portion of

the chip over the PMT using software written in house.

106

Measurement of bPAEC-derived N0. NO production in bBPAECs was determined by
ﬂuorescence imaging using the ﬂuorescent nitric oxide probe DAF-F M (Molecular
Probes, Eugene, OR). Cells were loaded with 10 pM of the membrane permeable
diacetate form of the probe (DAF-F M DA). Alter loading, the cells were incubated for 20
min at 37°C to ensure intracellular probe distribution. The DAF -F M DA and all the other
solutions were prepared in PBS 1X (Phosphate-Buffered Saline, Mediatech Inc.,

Hemdon, VA) which had been previously equilibrated for 20 min at 37°C and 5% C02.

Samples containing RBCS at 7% hematocrit, RBCS incubated with 1 pM iloprost, RBCS
incubated with 1 mM glybenclamide, RBCS incubated with 1 mM glybenclamide and 1
pM iloprost, and two controls (buffer and 1 pM iloprost) were pumped at a flow rate of
1.0 pL min‘1 at room temperature. After 30 min, the ﬂuorescence images corresponding
to each well were taken and the increase in intensity (corresponding to the NO
production by bPAECs) was analyzed. The probe-loaded cells were visualized using an
electrotherrnally-cooled CCD digital camera (Orca; Hamamatsu, Hammarnatsu City,
Japan) attached to an Olympus IX71M inverted microscope (Olympus America,
Melville, NY). Time-lapse images were acquired using MicroSuite software (Olympus

America).

Preparation of glybenclamide solution. Glibenclamide (Sigma; St. Louis, MO) was
prepared as a 0.01M stock solution by adding 49 mg of glibenclarnide to a solution
containing 2 mL NaOH and 7.94 mL of dextrose in distilled water (50 mg mL") and
heating it slowly to 52°C. Washed RBCS were incubated with glibenclarnide at a ﬁnal

concentration of either 250 nM or 1 mM for 15 min.

107

Preparation of the microﬂuidie devices. In all subsequent experiments, an irreversibly
sealed PDMS microﬂuidic device was fabricated from a master that has been previously
described in chapter 3. Brieﬂy, 4 mL of SU-8 50 negative photoresist (MicroChem Corp.,
Newton, MA) was spin coated (Laurell Technologies Corp., North Wales, PA) onto a 4
inch silicon wafer (Silicon, Inc., Boise, ID) for 20 s. at 2000 rpm. The coated substrate
was prebaked at 95°C for 5 min. prior to UV exposure through a negative ﬁlm
(Pageworks, Cambridge, MA) with a near-UV ﬂood source (Spectra-Physics, Stratford,
CT). The negatives were patterned using Freehand software (PC Version 10.0,
Micromedia, Inc. San Francisco, CA) and taped to 2.5 by 4 inch glass slides to ensure
proper alignment and exposure. The exposed wafer was then postbaked again at 95°C for
an additional 5 min before developing with propylene glycol monomethyl ether acetate
(Sigma-Aldrich Inc). The master was then rinsed with acetone and isopropyl alcohol to
dissolve any remaining excess photoresist and dried with compressed nitrogen. The
dimensions of the T-channel, corresponding to the raised photoresist dimensions of the
master, were measured using a proﬁlometer (Alpha Step-200, Tencor Instrucments,
Mountain View, CA) to be 150.0 pm wide by 118.6 pm deep.

Studies involving the measurement of ATP release from RBCS required the
fabrication of two individual PDMS molds irreversibly sealed together, one absent of any
features and the other a mold of a T-channel master. The T-channel chip was created by
pouring a degassed 5:1 mixture of Sylgard 184 elastomer and curing agent (Ellsworth
Adhesives, Germantown, WI) onto the silicon master and owing at 75°C for ~ 20 min.
This chip was removed from the wafer and inlet holes were punctured through the chip

using a 20 gage luer stub adapter (Becton Dickinson and Co., Sparks, MD) as well as 1/8

108

inch exit holes. This chip was then placed channel-side down onto a blank chip that was
created by pouring a degassed 20:1 PDMS mixture onto a blank wafer and curing at
75°C for ~ 30 min before removing it from the silicon wafer. The device was completed
by irreversibly curing both chips together at 75°C for an additional hour. A displacement
syringe pump (Harvard Apparatus Inc., Holliston, MA) was used to introduce samples
through Tygon tubing (Therrno Fisher Scientiﬁc Inc., Waltham, MA) that was ﬁtted with
15 mm 23 gage hypodermic steel tubing bent at 90 degrees (New England Small Tube,
Linchﬁeld, NH). This steel tubing was simply placed in the inlet access holes created in
fabrication to deliver solutions directly into the channels.

The SU-8 mold masters with the respective designs were obtained as described in
chapter 3 using soft lithography. PDMS prepolymer was mixed with curing agent with a
20:1 mass ratio, degassed in a vacuum chamber for 5 min and spread on the 6 channel
(100 pm diameter) mold master. PDMS was then partially cured by heating for 10 min at
75°C. A second layer of a 5:1 mixture was poured onto the chip and baked in the oven for
additional 20 to 30 min. A PDMS layer with a soft side (20:1) and a more rigid side (5:1)
was obtained by peeling off from the master. Subsequently, holes were punched on the
PDMS layer as the inlets and outlets using a 23 gage RW hypodermic full hard tube,
internal diameter 0.33 mm (New England Small Tube, Litchﬁeld, NH) and a 1/8 inch
puncher respectively.

A non patterned, 20:1 ratio PDMS slab was obtained similarly, by baking for 30
min and cell reservoir holes are punched using 1/8 inch puncher. The two PDMS layers,
one with 6 straight channels and one with 6 x 3 holes are semi-irreversibly sealed

together with the polycarbonate membrane (13 mm diameter, 1.0 pm pore size) in

109

between by baking at 75°C for additional 20 min. The connections between the
microdevice and the external syringe pumps (Hamilton, Fisher Scinetiﬁc) were achieved
using Tygon ﬂexible plastic tubing (Fisher) (0.020" ID x 0.060" OD) through small hard

pins (0.025” 0D x 0.013" ID) (New England Small Tube, Litchﬁeld, NH).

Immobilization of the bPAECs on the membrane array. Bovine pulmonary aortic
endothelial cells (bPAECs) primary cultures were purchased (Cambrex, East Rutherford,
NJ) and they were fed with fresh medium every 2-3 days. After they reached conﬂuence
the cells were enzymatically detached with trypsin (0.25%), pelleted (220 x g for 5 min)

and resuspended in culture medium until ready to use. Cells from divisions 3, 4 and 5
o 2 0
were used to run the expenments and seeded on collagen-coated 25 cm tissue culture

2
ﬂasks (Corning Inc., NY) at a density of 20,000 cells / cm . The cells were kept in a

humidiﬁed atmosphere at 37°C and 5% C02.

After the device was fully assembled, 20 pL of 100 pg mL'l ﬁbronectin was
added to each reservoir, on the membrane side, followed by incubation at 37°C for one
hour. The concentration and incubation time for ﬁbronectin were optimized such as to
insure the best adhesion of the endothelial cells to the polycarbonate membrane. After
ﬁbronection addition, the device was exposed to UV light for 30 min to crosslink the
ﬁbers and sterilize the device. bPAECs division 3 were harvested from the T-25 ﬂasks
and pelleted by centrifugation at 220 x g for 5 min and 37°C. Then, they were

resuspended in 5 mM L-Arginine (Sigma) (the substrate for eNOS) at a density of 3.1 x

6
10 cells mL". The suspension was homogenized by vortex for a couple of seconds.

Subsequently, 20 pL of the cell suspension was added to each well and incubated at 37°C

110

for one hour. One hour was found to be the optimum time for the cells to adhere to the

membrane in the wells.

4.2.2 Results and Discussion

Here, we report a method that combines microﬂuidic technologies in an array
setup to elucidate the communication skills between RBCS and endothelial cells obtained
from bovine puhnonary artery (bPAECs). RBCS, obtained from healthy rabbits, were
pumped through the channels of a microﬂuidic device. These channels were fabricated
such that the internal diameter closely approximates those of resistance vessels in vivo.
Moreover, the RBCS that were hydrodynamically pushed through the channels were
incubated in the presence or absence of iloprost, a stable analogue of prostacyclin that is
known to stimulate ATP release via a G-protein coupled receptor mechanism.

Figure 22 provides information on the experimental setup (a) as well as the
obtained data (b) .The summary of results from the RBCS of n = 5 rabbits are shown in
Figure 23. Figure 22a describes the system used to obtain the data; brieﬂy, a syringe
pump is connected to the device via hypodermic stainless steel tubing. The syringes
deliver RBCS (in the presence or absence of ATP release stimulant, such as iloprost, or
inhibitors such as glybenclamide) to the channels of the microﬂuidic device. The device,

made from PDMS is then placed over a photomultiplier tube (PMT) contained in a light

111

Dark Box

 
 

Hypodennic Tubing

  
 

 

 

Luciferin/Luciferase

   

 

 

 

 

 

 

 

 

 

 

 

 

l

Irreversibly sealed PDMS layers Waste

   
    
 

1))

0.07 '

W 7 % RBCS w/l pM Ilop.

7 % RBCS & 250 nM Gly.
/ M 11 .
0.05 ‘W W 1 11 Op

004 W 1 11M Ilop.

0.03 ' V v

.9
o
as

Chemiluminescence intensity

 

 

Time (see)

Figure 22. (a) Schematic of experimental set up for chemiluminescence measurements
taken to determine the iloprost-stimulated ATP release from RBCS in real
time shown in (b)

112

excluding box. The ATP released from the RBCS is then combined at an intersection of
channels on the microﬂuidic device with a solution of luciferase containing luciferin
(which is pushed through via the syringe pump). The chemiluminescence product that is
formed results in the emission of light that is measured by the PMT. The amount of
chemiluminescence is proportional with the ATP present.

A representation of the data is shown in Figure 22b and, as shown, the amount of
ATP released from the RBCS increases in the presence of the iloprost. To verify that the
increased chemiluminescence is not due to cellular lysis, the RBCS were also incubated
in the presence of iloprost and glybenclamide, an inhibitor of ATP release from the
RBC.19 As shown, incubation of the RBCS with glybenclamide present reduces the ATP
release from these cells, thus providing evidence the measured signal is due to ATP and,
importantly, that the signal is not due to cellular lysis but rather a release from intact
cells. If the signal was due to lysis, glybenclamide would have no affect on the signal.
Figure 23 summarizes the results from the RBCS obtained from n = 5 rabbits. The data in
Figure 24 veriﬁes that iloprost is capable of inducing ATP release in the channels of the
microﬂuidic device and that this ATP release can be controlled using the appropriate
inhibitors.

In an attempt to classify a possible role for this ATP release in vivo, a device was
prepared that contained a polycarbonate membrane that separated the channels
containing RBCS from a layered endothelium. The schematic of such a device is shown
in Figure 23a. The top of the membrane, the side away from the RBC stream, was pre-
treated with ﬁbronectin prior to introducing bPAECs to the membrane. After the bPAECs

were immobilized to the membrane (as veriﬁed through microscopy), the RBCS were.

113

 

 

 

 

*

0.95 i018

I

 

 

 

 

 

 

 

A 1.0 .
E
.3
“a? 0.8 .
2
2
f: 0.6 .
<1:
53;; 0.29 i 0.14
H
s 0-4 ‘ 0.27 i 0.04
<1

0.2 2 0.10 a 0.03

0.0 1 . 1 . , .

1 11M 1109- 7 % RBCS 7% RBCS w/ 7% RBCS & 250 nM

1 pM Ilop. Gly.w/l pM Ilop.

Figure 23. Comparison to the controls of iloprost (1 pM) and RBCS alone as well as the
inhibition of ATP release when iloprost-stimulated RBCS were incubated with
glybenclamide (250 nM) (*p 5 0.005).

114

a) Polycarbonate membrane
(1 pm pore size) sealed
between two PDMS layers

  
 

RBC samples

\

 
   
 

150 pm Channel
(bottom layer)

Blank PDMS with cell
reservoirs (top layer)

b)

7%RBCS

1 pM Ilop.

 

Figure 24. Iloprost-derived ATP release from RBCS stimulates NO production in a
microﬂuidic device-containing an endothelium immobilized

polycarbonate membrane. In (a), the microﬂuidic device contains a
polycarbonate membrane sealed between two PDMS layers that separates a
channel containing RBCS (bottom layer) from a layered endothelium (top
layer). In (b), ﬂuorescence microscopy images are shown for pre- (images on
the top) and post-(images on the bottom) addition of iloprost.

115

introduced to the microﬂuidic device. First, an aliquot of RBCS in the absence of iloprost
were introduced to the system and, as shown in images 1-3 in Figure 24b, the level of
ﬂuorescence intensity is minimal. However, upon addition of an aliquot of RBCs that had
been incubated in the presence of iloprost, the ﬂuorescence intensity of the bPAECs
increased by 67% (Figure 25). This increase in ﬂuorescence intensity can be attributed to
an increase in NO production by the bPAECs in response to the increase in ATP derived
from those RBCS that were incubated in the iloprost. These data suggest that the efﬁcacy
of iloprost in improving blood ﬂow in the intact circulation may be due, in part, to its
ability to stimulate the release of a known stimulus of nitric oxide synthase, namely,
ATP. In addition to creating a controlled in vitro platform that simulates in vivo
conditions, it is also necessary to have a system that will enable high throughput
measurements that include calibration, drug-dose response data, and proper controls for
veriﬁcation of measurements. In an extension of the data shown in Figures 24 and 25,
data from a device that incorporates multiple wells, all addressable by a microﬂuidic
channel, is shown in Figure 26. Here, 18 wells were prepared in the absence (column C)
and presence (columns A and B) of bPAECs; all of the wells were addressable with the
RBC ﬂow stream. The bPAECs in column A were also loaded with DAF-FM DA, an
intracellular ﬂuorescence probe for NO.

In this system, the wells were subjected to buffer alone (row 1), iloprost (row 2),
RBCS in the absence (row 3) and presence (row 4) of iloprost, RBCS with glibenclamide
(row 5) or RBCS and iloprost combined in the presence of glibenclamide (row 6). The
data shown is summarized in Figure 25b. Note the signiﬁcant increase in ﬂuorescence

intensity due to endothelium-derived NO in the presence of the RBCS and the RBCS.

116

 

40

(.0
O
_l

 

 

 

 

 

 

 

 

A
3.
(U V" 'I
v 1-..,“
g l’.'r'
2 ll ,4:-.'“
0) {WM
4—0 l l,
.5 ,lil" »
m 20- 0" 1;“!
2 “a“... ‘ .
a
a) .4 :9». ..fl‘tijlll
8 Jil I”? l . '1,“ l
. r -‘1‘
2 ..;'I" M I! I
' , i‘l‘l
g 10. ”IN”. ‘ Hit-'lln 1". .
L]. V .II ' pgsrrli’
H" v'
,; 1 ..:i U
l:l
t '.
0 , lr '- ,r'.. ‘6
control Ilop. 1 uM

Figure 25. The increase in ﬂuorescence intensity due to endothelium-derived NO
production when a sample of RBCS incubated with iloprost (23.0 d: 3.5) vs. a
sample of RBCS alone (16.5 :t 1.0) is pumped through the channel. The
values reported are the average intensity and standard deviation (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
S 0.005).

117

a) b)

 

 

 

 

 

 

A B C
*
1600‘
'7:
s: 1200»
3 8
8
a 800‘
4 3 . T
3 600- r
‘5
Lu
2001
o + .

Figure 26. Micrographs of a vascular endothelium array with blood components are
shown in (a). The wells in columns A and B contain bPAECs cultured on a
polycarbonate membrane. Column C contains cell free wells. The rows
represent the wells of the array that are addressable by the underlying
microchannel network, each delivering a different reagent to the wells as
described in the text. The quantitative representation of the data is shown in
(b). The difference in ﬂuorescence intensity between samples containing the
RBCS in the absence (row 3) and presence (row 4) of iloprost are statistically
different from the other rows. There is also a difference in the intensities of
rows 3 and 4 (*p 5 0.005).

118

4.3 CONCLUSIONS

In addition to interacting with an appropriate molecule, there are many
characteristics required of a drug candidate throughout the testing process. Drug
adsorption properties, its ability to be distributed in vivo, its effects on metabolism,
excretion characteristics and its toxicity are often monitored.

Here, we showed that an array of endothelial cells, addressable by an underlying
microﬂuidic network of channels containing RBCS, can be employed as an in vitro model
of the in vivo circulation to monitor cellular communication between different cell types.
Results obtained from this array suggest that the ability of iloprost, a stable analogue of
prostacyclin, to stimulate nitric oxide production in endothelial cells, may be due to its
ability to stimulate ATP release from the red cell. These results provide evidence that the
described device may serve as a controlled, in vitro platform for performing in viva-type

measurements.

119

(1)
(2)
(3)

(4)
(5)

(6)

(7)

(8)

(9)
(10)

(11)

(12)

(13)

(14)

(15)

LIST OF REFERENCES

Dittrich Petra, S.; Manz, A. Nat Rev Drug Discov 2006, 5, 210-218.
Breslauer David, N.; Lee Philip, J .; Lee Luke, P. M01 Biosyst 2006, 2, 97-112.

Pardridge, W. M. Introduction to the Blood-Brain Barrier: Methodology, Biology
and Pathology, 1998.

Shah, M. V.; Audus, K. L.; Borchardt, R. T. Pharm Res 1989, 6, 624-627.

Carroll, J. S.; Ku, C.-J.; Karunarathne, W.; Spence, D. M. Anal. Chem.
(Washington, DC, U. S.) 2007, 79, 5133-5138.

Ellsworth, M. L.; Forrester, T.; Ellis, C. G.; Dietrich, H. H. American Journal of
Physiology 1995, 269, H2155-H2161.

Sprague, R. S.; Ellsworth, M. L.; Stephenson, A. H.; Kleinhenz, M. E.; Lonigro,
A. J. American Journal of Physiology 1998, 275, H1726-H1732.

Sprague, R. S.; Stephenson, A. H.; Ellsworth, M. L.; Keller, C.; Lonigro, A. J.
Exp Biol Med (Maywood) 2001, 226, 434-439.

Sprung, R.; Sprague, R.; Spence, D. Analytical Chemistry 2002, 74, 2274-2278.

Carroll, J .; Raththagala, M.; Subasinghe, W.; Baguzis, S.; Oblak, T. D. a.; Root,
P.; Spence, D. M01. BioSyst. 2006, 2, 305-311.

Evans Joseph, L.; Goldﬁne Ira, D.; Maddux Betty, A.; Grodsky Gerold, M.
Diabetes 2003, 52, 1-8.

Khademhosseini, A.; Langer, R.; Borenstein, J .; Vacanti, J. P. Proc. Natl. Acad.
Sci. U. S. A. 2006, 103, 2480-2487.

Shin, M.; Matsuda, K.; Ishii, 0.; Terai, H.; Kaazempur-Mofrad, M.; Borenstein,
J .; Detrnar, M.; Vacanti, J. P. Biomed. Microdevices 2004, 6, 269-278.

Takayama, S.; Ostuni, E.; LeDuc, P.; Naruse, K.; Ingber, D. E.; Whitesides, G.
M. Nature 2001, 411, 1016.

Sin, A. Chin Katherine, C.; Jamil Muhammad, F.; Kostov, Y.; Rao, 0.; Shuler
Michael, L. Biotechnol Prog 2004, 20, 338- 345.

120

(16)
(17)

(18)

(19)

Li, M. W.; Spence, D. M.; Martin, R. S. Electroanalysis 2005, 17, 1171-1180.

Spence Dana, M.; Torrence Nicholas, J.; Kovarik Michelle, L.; Martin, R. S.
Analyst 2004, 129, 995-1000.

Roper, M. G.; Shackman, J. G.; Dahlgren, G. M.; Kennedy, R. T. Anal. Chem.
2003, 75, 4711-4717.

Sprague, R. S.; Olearczyk, J. J.; Spence, D. M.; Stephenson, A. H.; Sprung, R.
W.; Lonigro, A. J. Am. J. Physiol. 2003, 285, H693-H700.

121

CHAPTER 5

5.1. A PROPOSED MECHANISM FOR THE INVOLVEMENT OF
PROLACTIN-RELEASING PEPTIDE (PrRP) IN THE REGULATION OF
VASCULAR TONE

In 1998 Hinuma et al.1 reported the identiﬁcation of two novel peptides, PrRP-20
and PrRP-31 that stimulated the release of prolactin from the pituitary gland. The
peptides were identiﬁed in the bovine hypothalamus as the endogenous ligands of the
orphan seven-transmembrane-domain receptor (GPRlO/hGPR3). Initial studies revealed
that the role of PrRP was associated with weak prolactin release in vivo1 and in vitro.2‘3
Other roles of this peptide have been reported, for example regulation of energy balance,4
suggesting that PrRP is involved in food regulation and body weight. Lawrence et al.
have shown that intracerebralventricular (i.c.v.) administration of PrRP reduced the food
intake signiﬁcantly in male rats, but did not affect the water intake.5 However, PrRP not
only regulates the energy homeostasis, but also is a potent mediator of stress responses in
the brain.2‘3’6"5

Subsequent work by Samson et al.16 and Jarry et al.17questioned the capacity of
the PrRP to serve as physiological regulator of prolactin (PRL), and other functions for
this peptide have been explored. Samson et al. reported that microinjections of PrRP in
the ventrolateral medulla (V M) resulted in a speciﬁc dose dependent increase in mean

arterial pressure, heart rate, and sympathetic activity. It was shown that PrRP acts as a

modulator of blood pressure in areas with little or no receptor expression and contrary,

122

has no effect in areas of high receptor expression, such as the area postrema (AP).18 It
became clear that although PrRP and its receptor may be acting to release prolactin in
some tissues, it may have some independent functions of its own.8 To date, the
physiological function of PrRP is not yet completely understood.

Prolactin-releasing peptide has two molecular forms, one with a C-terminal 20-
residue peptide (PrRP20) and one with a 31-residue peptide (PrRP3 1), both known to be
derived from a single precursor molecule. Their amino acid sequences,
TPDINPAWYAGRGIRPVGRF-amide and SRAHQHSMEIRTPDINPAWYAGRGIRP-
VGRF-amide respectively, are highly conserved among several species, such as human,
bovine, and rat.3

Roland et al.8 reported that the fragment PrRP(12-31) was the minimum agonist
fragment, with a receptor afﬁnity of Ki = 1 nM. Moreover, an alanine scan of the segment

I.19 found

PrRP(25-31) showed the importance of the three arginine residues. Boyle et a
that the functionally important residues are located within the carboxyl-terminal
heptapeptide segment Ilez5-Arg26-Pr027-Va128-Gly29-Arg3O-Phe3l-NHz. He also
acknowledged that the only essential amino acid residue in PrRP is the Arg,30 which
donates both a critical basic side chain in the L-conﬁguration and is also reported to
contribute an essential backbone NH. The ﬁve Arg and two His residues of PrRP31
impart a very basic character to the peptide at physiological pH, and therefore high
tendency to bind negatively charged molecules. Because of its basic character and
because only the carboxyl-terminal heptapeptide segment is involved in the receptor

binding with the Arg30 residue being the essential amino acid, PrRP could adapt its “body

structure” to other functions of physiological importance.

123

In this study we propose a mechanism for the previously reported
cardiovascular/blood pressure regulatory properties of the PrRP, that is PrRP binding to
the red blood cell (RBC)-derived adenosine triphosphate (ATP) that is released under

various stimuli. It has been previously reported that RBC—derived ATP contributes to the

0-22 3

control of vascular resistance in both the pulmon and the systemic circulation.2
RBCS, when traversing microvascular beds, are subjected to mechanical deformation and
as a response they release ATP. Moreover, it is well established that vascular
endothelium constitutively generates nitric oxide (NO) in response to ATP stimulation
and that endothelium-derived NO induces a relaxation of vascular smooth muscle cells. If
PrRP binds ATP, this will lower the extracellular ATP pool available for the stimulation
of the endothelial NO synthesis and subsequently, a decrease in the vascular
relaxation/increase in the blood pressure might be possible.

This proposed mechanism seems very plausible in the construct of Samson’s
ﬁndings, that signiﬁcant elevations in the mean arterial pressure were observed in
response to i.c.v. administration of low nanomolar concentrations of PrRP31. Also, he
reported that the mean blood pressures returned to pre-injection control levels in
approximately 15 to 25 min from the peptide administration. Moreover, he showed that
effective doses of PrRP failed to alter the thirst and also, affect the salt appetite in rat
models, therefore attesting the speciﬁcity of the blood pressure elevated actions of the
PrRP. All these ﬁndings led us to investigating in more depth the alternative

physiological function proposed for the PrRP as well as the mechanism by which the

peptide exerts this function.

124

In this work, different techniques, such as mass spectrometry, chemiluminescence
and ﬂuorescence microscopy were utilized to show, directly or indirectly, the PrRP
binding to ATP. Moreover, by using an in vitro microﬂuidic-based mimic of an in vivo
microcirculation that we previously reported,24 we constructed a system where the direct
involvement of the PrRP in the regulation of the blood pressure and/or vascular tone

could be monitored in real time.

5.2 EXPERIMENTAL

5.2.1 Materials and Methods

Collection of RBCS. All procedures involving the collection of blood samples from
animals were approved by the appropriate Animal Investigation Committee. For studies
involving rabbit RBCS, male New Zealand white rabbits (2.0—2.5 kg) were anesthetized
with ketamine (8 mL kg", intramuscular injection) and xylazine (1 mg kg", i.m.)
followed by pentobarbital sodium (15 mg kg], intravenous injection). A cannula was
placed in the trachea and the animals were ventilated with room air at 20 breaths min-l
and a tidal volume of 20 mL kg‘l. A catheter was placed into a carotid artery for
administration of heparin and for phlebotomy. After heparin (500 units, i.v.) animals
were exsanguinated. Blood was centrifuged at 500 X g at 4°C for 10 min. The plasma and
buffy coat were removed for other experiments. RBCs were then resuspended and
washed three times in a physiological salt solution (PSS). The PSS was made by

combining 25 mL of TRIS buffer [prepared by mixing 50.9 g of TRIS in 1 L of distilled

125

and deionized water (DDW)] and 25 mL of Ringer’s solution (164.2 g NaCl, 7.0 g KCl,
5.9 g CaClz -2H20, and 2.83 g MgSO4 in 1 L of DDW). After the addition of 0.50 g of
dextrose and 2.50 g of albumin bovine fraction V (fatty acid-free) to the TRIS—Ringer’s
mixture, the entire solution was diluted to 500 mL with DDW and the pH was adjusted to
7.35—7.45. The PSS was then ﬁltered three times using a 0.45 um ﬁlter (Corning, Fisher

Scientiﬁc).

Measurement of the ATP release. All reagents were from Sigma Chemical (St. Louis,
MO) unless otherwise noted. A 100 uM stock solution of ATP was prepared by adding
0.0551 g of ATP to 1000 mL of DDW. ATP standards with concentrations ranging
between 0 and 1.5 uM were then prepared in PSS from the stock. 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. For ATP
measurements involving RBCS, all RBC samples were diluted to a 7% hematocrit. To
stimulate the ATP release from the RBCS, cells were treated with iloprost (Cayman
Chemicals, Ann Arbor, MI), and Zn(II) activated-C-peptide (American Peptide,
Sunnyvale, CA, USA). Detection of ATP release from the RBCS was performed using the
luciferase assay and chemiluminescence measurement by a photomultiplier tube. The
resultant current from the photomultiplier tube, which is proportional to the ATP induced
chemiluminescence, was measured as a potential by a data acquisition board controlled

by a program written with LabView (National Instruments, Austin, TX, USA).

126

Mass spectrometry analysis. All experiments were performed using a Therrno model
LTQ linear ion trap mass spectrometer (Therrno, San Jose, CA, USA). Samples were
prepared by dissolving synthetic human PrRP in puriﬁed water at a concentration of 6.6
umol L". For ATP binding studies, a 20 mo] L'l solution of ATP prepared in puriﬁed
water was added to a PrRP solution then introduced to the mass spectrometer at a ﬂow
rate of 0.5 pL min'1 by nanoelectrospray ionization. The spray voltage was maintained at

2.0 kV. The heated capillary temperature was 250°C.

Fabrication of the microﬂuidic array. Fabrication of the poly(dimethylsiloxane)
(PDMS)-based microchips involves well established soft lithography methods.25 The
fabrication relied on two individual PDMS molds irreversibly sealed together, one absent
of any features and the other a mold of a 6 channels master. The 6 channels chip was
created by pouring a degassed 20: 1 mixture of Sylgard 184 elastomer (Ellsworth
Adhesives, Germantown, WI, USA) and curing agent onto a 4 inch silicon master
(Silicon Inc., Boise, ID, USA) and curing at 75°C for 30 min. An additional layer of a 5:1
mixture was poured on top of the 20:1 mixture to give the chip enough surface brittleness
for ease of puncturing the inlet holes. After additional 30 min bake at 75°C, the chip was
removed from the wafer and inlet holes were punctured through the chip (using a 20
gauge luer stub adapter) as well as 1/8 inch exit holes. A non-pattemed, 20:1 ratio PDMS
slab was obtained as before, by curing for 30 min followed by the creation of cell
reservoir holes using a 1/8 inch hole punch. The two PDMS layers, one with 6 straight
channels and one with 18 holes (6 rows x 3 columns) were irreversibly sealed together

with the transparent polycarbonate membrane (13 mm diameter, 0.03 pm pore size) in

127

between by curing at 75°C for an additional 30 min. The connections between the
microdevice and the external syringe pumps were achieved using the Tygon plastic
tubing (0.02" ID x 0.06" OD) that were ﬁtted with 15 mm, 23 gauge hypodermic steel
tubing bent at 90 degrees. This steel tubing was simply placed in the inlet access holes
created in fabrication to deliver solutions directly into the channels. All channels used in
the studies reported here had a width of 200 um and a depth of 100 pm. After the
bPAECs were immobilized to the membrane and loaded with the NO probe, the RBCS
were introduced to the microﬂuidic device through the underlying channels on the bottom
layer of PDMS. These 18 wells in the presence of bPAECs were addressed with the RBC

ﬂow stream.

Cell immobilization. The design of the microﬂuidic device allows the easy introduction
of the cells to the wells. The top of the membrane, the side away from the RBCS, serves
as a support for the bPAECs. The membrane was pre-treated with 100 pg mL'l
ﬁbronectin (Sigma Chemical Co., St. Louis, MO) and incubated for 1 h at 37°C. The
ﬁbronectin treatment ensures the adhesion of the cells to the membrane. bPAECs, low
passage, were harvested by trypsinization and introduced to the 18 wells of the array and
incubated at 37°C and 5% C02 for 3 h or until the cells had attached to the ﬁbronectin
layer. Once the endothelial cells had adhered to the membrane, ﬂuorescence microscopy
was used to monitor the NO production brought about the RBC-derived ATP in the

underlying microﬂuidic channels.

128

Measurement of Nitric Oxide. NO derived from bPAECS cultured on the membrane of
the microﬂuidic array was monitored using the ﬂuorescence probe 4-amino-5-
methylamino-Z’,7’-diﬂuoroﬂuorescein diacetate (DAF-F M DA) (Molecular Probes,
Eugene, OR).26’ 27 Next 10 uL of a 10 uM solution of DAF-FM in Hank’s Balanced Salt
Solution (HBSS, Sigma) were added to the cells in each well of the microﬂuidic array
and incubated for 30 min at 37°C and 5% C02 to ensure a sufﬁcient amount of dye had
accmnulated in the endothelial cells. After probe incubation, the wells were rinsed twice
with equilibrated HBSS to eliminate any extracellular DAF-FM that might result in an
increased background ﬂorescence.

To measure the NO production in the cells, RBC samples of 7% hematocrit in
PSS were incubated in the presence or absence of iloprost (a stable analogue of
prostacyclin), or metal-activated C-peptide,28 both known to stimulate/increase the ATP
release from RBCS. Contrary, for the samples incubated with the aforementioned
solutions followed by incubation with PrRP, a decrease in ATP release, as result of PrRP
binding to ATP, would translate into a decrease in NO production. All solutions are
pumped for 30 min at a ﬂow rate of 1 uL min'l in the microﬂuidic channels underlying
the array of wells. The ATP within each sample diffuses throught the polycarbonate
membrane into the wells and stimulates the NO production in bPAECs.

Fluorescence images for each well were taken using an Olympus IX71M
microscope (Olympus America, Melville, NY) with an electrotherrnally cooled CCD
(Orca, Hamamatsu) and MicroSuite software (Olympus America). The microscope
incorporated a F ITC ﬁlter cube (Chroma Technology Corp) containing the excitation

(460-500 nm) and emission (505-560 nm) ﬁlters. The results were reported as gray value

129

mean per well, where the gray-level intensity refers to the amount of light energy actually

reﬂected from or transmitted through the physical specimen.

5.2.2 Results and Discussion

It has been well established that RBCS participate in vascular regulation and the
property of these RBCS responsible for the stimulation of NO synthesis was their ability

to release ATP in response to mechanical deformation.21 ATP is an in vivo stimulus of

29,30 31,32

endothelial NO production, it is present in millimolar concentrations in the RBCS,

and it is released in response to physiological stimuli, such as hypoxia and

3334 When RBCS are traversing the resistant vessels in vivo, they are

hypercapnia.
increasingly deformed by increments in the velocity of the blood ﬂow through a vessel
and/or by reduction in vascular caliber,35 and as a result they release ATP through a
mechanism involving the cystic ﬁbrosis transmembrane regulator (CFTR). This RBC-
derived ATP can then act on the endothelial cells to stimulate endogenous NO synthesis
and enable RBCs to participate in local regulation of vascular caliber.2"23’ 34’36'38

However, depending on the type of purinergic receptor ATP binds, it can induce
antagonistic effects on the vasculature. Thus, ATP binding to the P2x receptor, which is
primarily found on the vascular smooth muscle cells, results in contraction of the cells,39
whereas ATP binding to the P2y receptor found primarily on the endothelium, results in

NO synthesis and/or vasodilatation.29”3O Thus, ATP applied to the luminal side of a vessel,

that released from the RBCS within the circulation, would be expected to produce

130

endothelium-dependent relaxation through an interaction with the endothelial P2y

purinergic receptor (Figure 27.)

ATP binding to PrRP. In this work we show the possible role that PrRP could play in
the regulation of the mean arterial pressure through a mechanism involving intraluminal
ATP depletion. Mass spectrometry data, obtained in collaboration with Dr. Gavin Reid at
Michigan State University, showed the ability of the peptide to bind ATP. Based on these
ﬁndings, additional experiments involving the luciferin/luciferase assay and
chemiluminescence detection were performed. Standard ATP samples were prepared as
described in the methods section. Data in Figure 28 provides further evidence that PrRP
binds to ATP. Here, upon the addition of 10 nM PrRP, the chemiluminescence signal
from the samples containing ATP standard solutions decreased in a time dependent
fashion. After the immediate introduction of the peptide, a decrease of approximately
30% of the chemiluminescence signal was observed for the ATP concentrations of 0.75,
1, and 1.5 uM. Moreover, about 50% of the initial ATP signal was inhibited by the 60
min incubation with PrRP. This signiﬁes that free ATP in solution decreases either by
binding to another species present and/or by degradation. However, in this case, based on

the mass spectral data and the structural information, PrRP is found to bind ATP.

131

STRESS

ﬂﬂﬂ
Q 65???)

TP AP 5P
® 911’er PrRP NO metabolites
PrR ATP ATP ATP Tl,
ATP ATPP
Luminalside a l P2Y! P2Y” pry NO
PrR . ,

 
    
 

   
 
  

PrR

 

Endothelial Cells

  

     
   
   
    
 

  

Caz"-CaM L- -Arginin'é

mg;

 

Smooth Muscle Cells

Extraluminal side

Figure 27. Representation of the mechanism of RBC-mediated regulation of the vascular
tone and the possible involvement of prolactin-releasing peptide (PrRP) in
scavenging the RBC-derived ATP. ATP binds to the P2y purinergic receptors
found on the endothelial cells that line up the luminal side of the blood vessel
and through a Ca2+-dependent mechanism triggers the activation of the eNOS,
and therefore production of NO. NO activates the guanylate cyclase signaling
cascade that ultimately results in vasodilation.

132

 

0.18

 

 

 

0.16 r
:i
3; 0.14 r .
.E‘
8
.E
a) 0.10 - at
‘e’
3,,” 0.08 ~ 0: f i
Q)
C‘.
.._ g a
g 0.06~ O 9
E 3 3
0 0.04 - O
6 g 5

8
0.02 -
0.00 I I I I I I I I I

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
ATP conc. (uM)

Figure 28. PrRP binding to ATP. Filled circles (0) represent the ATP standards (O-l.5
uM) without the peptide. Open triangles (V), squares (El) and rhomboids (0)
show the decrease in ATP signal after the incubation with PrRP (10 nM) for
0, 30, and 60 min, respectively (n=3). The decrease in chemiluninescence
signal suggests that PrRP binds to ATP (p 3 0.005).

133

Antagonistic effect of PrRP on the A T P release from C-peptide/iloprost treated RBCS.
We hypothesized that PrRP binds to physiological ATP and reduces the pool of ATP
needed for stimulation of the endothelium-derived NO. In order to prove this hypothesis,
RBCS were treated with different agents known to have the ability to increase the ATP
release, speciﬁcally C-peptide and iloprost. We have previously demonstrated that the
ability of the C-peptide to increase the RBC-derived ATP depends on the activation with
metal ions, such as Fey and Cr”.28 Moreover, the ATP release was dependent upon the
ability of the C-peptide to increase glycolysis within the RBCS via increased cellular
glucose transport through GLUTl. However, in this work we used Zn2+ to stimulate the
activity of C-peptide before incubation with RBCS. Figure 29 shows a signiﬁcant ATP
release from the RBCS treated with the Zn2+ activated C-peptide compared to control
(RBC 7%), which is in very good agreement with our previously published data.28
Moreover, the data in Figure 28 demonstrates that PrRP (10 nM) induces a
signiﬁcant decrease in the RBC-derived ATP release and the decrease is higher after 90
min (n=5) incubation compared to 6 h (n=8). This is because the PrRP is being consumed
and ATP release is being partially restored after the 6 h incubation period. We have to

note here that all reagents have been added at the same time and have been incubated for

134

 

3.0

 

- 90min
2.5_ $611

2.0 - l

 

 

 

1.5 -

1.0 - ﬂ T f

0.5 -*

Normalized chemiluminescence intensity

 

 

 

 

 

 

 

 

 

 

 

 

 

0.0 u u
f‘0/0 0,96?

 

Figure 29. Effect of PrRP (10 nM) on the ATP release from RBCS treated with Zn”-
activated C-peptide The decrease in chemiluminescence intensity
corresponds to a decrease in the measured ATP released from the RBCS
when PrRP was added to the metal activated C-peptide. Filled bars
correspond to the 90 min incubation with PrRP (n=5) and the open bars
correspond to 6 hrs incubation (n=8). The results are reported as the average
i standard error of the mean. The differences in chemiluminescnence
intensity between the samples incubated in the presence and absence of PrRP

are statistically signiﬁcant (*p S 0.005).

135

90 min or 6 h. However, for the C-peptide samples, pre-activation of the peptide with the
metal in water followed by the addition of this cocktail in PSS to the RBCS was required.
In order to prove whether the PrRP is actively binding to the ATP released from the
RBCS or it is involved in some mechanism of C-peptide-dependent membrane
deformability, we incubated the RBCS with iloprost. Iloprost is a stable analogue of
prostacyclin that is known to stimulate ATP release via a G-protein coupled receptor
mechanism. Data in Figure 30 provides further evidence that PrRP decreases the pool of
that ATP released from RBCS rather than having a direct effect on the RBCS membrane.

In an attempt to classify the possible role of PrRP in regulating the vascular tone
by decreasing the endothelium-derived NO production through a decrease in ATP, a
device was constructed that closely mimics in vivo microcirculation by using a
transparent polycarbonate membrane to separate a ﬂow of RBCS from a layered
endothelium. This device was previously described and successfully used as a controlled,
in vitro platform for performing in vivo-type measurements.24 This device has the
potential of performing high throughput measurements that include calibration, drug-dose
response data and proper controls.

In this study, all 18 wells contained a layer of conﬂuent bPAECs and were
addressable by a ﬂowing stream of RBCS. The bPAECS were loaded with DAF -F M DA
and the NO production was monitored as ATP diffused from different samples containing
RBC, (row a), Zn(II)-activated C-peptide-treated RBCS incubated with and without PrRP
(rows d and c, respectively), and iloprost treated RBCS incubated with and without PrRP
(rows f and e, respectively). The data shown is summarized in Figure 311. The

quantitative representation of the normalized data from 4 different experiments is shown

136

3.0

 

2.5 ~

2.0- I. I

 

 

0.5 -

Normalized chemiluminescence intensity
_I
U!
l
l.—

 

 

 

 

 

 

 

 

 

00 I i l I A 7 VIII 8 :.
RBC 7% RBC+Ilop RBC+Ilop+PrRP

Figure 30. The effect of the PrRP (10 nM) on the measured ATP release from the RBCs
treated with iloprost (10 nM) (n=3). All samples contain RBC (ﬁnal
hematocrit 7%). Control sample is represented by RBC 7%. The iloprost-
treated RBCs samples incubated with or without PrRP are signiﬁcantly
different (*p < 0.005; **p S 0.05).

137

2.0 7

'oo

Normalized ﬂuorescence intensity
I

 

0.8 -

   

Figure 31. Antagonistic effect of PrRP on the endothelium-derived NO production in the
bovine pulmonary artery endothelial cells (bPAECs) Picture I shows
micrographs of vascular endothelium array with blood components. The rows
a-f represent the wells of the array that are addressable by the underlying
microchannel network each delivering a different reagent to the wells as
described in the text. The quantitative representation of the normalized data
from 4 different experiments is shown in II. The difference in ﬂuorescence
intensity between samples containing RBC + C-pep + Zn(II) incubated in the
absence (bar C) and presence (bar D) of the PrRP is statistically signiﬁcant
(*p < 0.05) There is also signiﬁcant difference in the intensities of rows (bars
E and F), where the RBCS + Iloprost samples were incubated in the absence
and the presence of PrRP (**p < 0.001), respectively. Errors bars are the mean
3: standard error A: RBC7%; B: RBC + PrRP; C: RBC + C-pep + Zn; D: RBC
+ C-pep + Zn + PrRP; E: RBC + Ilop; F: RBC + Ilop + PrRP.

138

in Figure 3111. A signiﬁcant decrease in ﬂuorescence intensity between samples
containing activated C-peptide-treated RBCS or iloprost-treated RBCS and the same
samples incubated with PrRP was observed. This decrease in ﬂuorescence translates to a
decrease in NO production as result of depletion of ATP pools available for endothelium
stimulation. Herein we showed that PrRP has the ability to decrease the NO production
and therefore, constitute a possible alternative mechanism for regulation of the blood

ﬂow by exerting vascular smooth muscle constriction.

139

(1)

(2)

(3)

(4)
(5)

(6)
(7)
(8)

(9)
(10)

(11)

(12)

(13)
(14)

LIST OF REFERENCES

Hinuma, S.; Habata, Y.; Fujii, R.; Kawamata, Y.; Hosoya, M.; Fukusumi, S.;
Kitada, C.; Masuo, Y.; Asano, T.; Matsumoto, H.; Sekiguchi, M.; Kurokawa, T.;
Nishimura, 0.; Onda, H.; Fujino, M. Nature 1998, 393, 272-276.

Minami, S.; Nakata, T.; Tokita, R.; Onodera, H.; Imaki, J. Neurosci. Lett. 1999,
266, 73-75.

Maruyama, M.; Matsumoto, H.; Fujiwara, K.; Kitada, C.; Hinuma, S.; Onda, H.;
Fujino, M.; Inoue, K. Endocrinology 1999, 140, 2326-2333.

Bray, G. A. Int J Obes 1984, 8 Suppl 1, 119-137.

Lawrence, C. B.; Celsi, F.; Brennand, J .; Luckman, S. M. Nat. Neurosci. 2000, 3,
645-646.

Zhu, L. L.; Onaka, T. Neuroscience (Oxford, U. K.) 2003, 118, 1045-1053.
Yamakawa, K.; Kudo, K.; Kanba, S.; Arita, J. Neurosci. Lett. 1999, 26 7, 113-116.

Roland, B. L.; Sutton, S. W.; Wilson, S. J .; Luo, L.; Pyati, J .; Huvar, R.; Erlander,
M. G.; Lovenberg, T. W. Endocrinology 1999, 140, 5736-5745.

Morales, T.; Sawchenko, P. E. Neuroscience (Oxford, U. K.) 2003, 121, 771-778.

Morales, T.; Hinuma, S.; Sawchenko, P. E. J. Neuroendocrinol. 2000, 12, 131-
140.

Mera, T.; Fujihara, H.; Kawasaki, M.; Hashimoto, H.; Saito, T.; Shibata, M.;
Saito, J .; Oka, T.; Tsuji, S.; Onaka, T.; Ueta, Y. Neuroscience (San Diego, CA, U.
S.) 2006, 141, 1069-1086.

Maruyama, M.; Matsumoto, H.; Fujiwara, K.; Noguchi, J .; Kitada, C.; Fujino, M.;
Inoue, K. Endocrinology 2001, 142, 2032-203 8.

Lee, Y.; Yang, S. P.; Soares, M. J.; Voogt, J. L. Brain Res. Bull. 2000, 51, 171-
176.

Iijima, N.; Kataoka, Y.; Kakihara, K.; Bamba, H.; Tamada, Y.; Hayashi, S.;
Matsuda, T.; Tanaka, M.; Honjyo, H.; Hosoya, M.; Hinuma, S.; Ibata, Y.
NeuroReport 1999, 10, 1713-1716.

140

(15)
(16)
(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

Chen, C. T.; Dun, S. L.; Dun, N. J .; Chang, J. K. Brain Res. 1999, 822, 276—279.
Samson, W. K.; Resch, Z. T.; Murphy, T. C. Brain Res. 2000, 858, 19-25.

Jarry, H.; Heuer, H.; Schomburg, L.; Bauer, K. Neuroendocrinology 2000, 71,
262-267.

Horiuchi, J.; Saigusa, T.; Sugiyarna, N.; Kanba, S.; Nishida, Y.; Sato, Y.;
Hinuma, S.; Arita, J. Brain Res. 2002, 958, 201-209.

Boyle, R. G.; Downham, R.; Ganguly, T.; Humphries, J.; Smith, J .; Travers, S. J.
Pept. Sci. 2005, 11, 161-165.

Sprague, R. S.; Ellsworth, M. L.; Stephenson, A. H.; Kleinhenz, M. E.; Lonigro,
A. J. American Journal of Physiology 1998, 275, H1726-H1732.

Sprague, R. S.; Ellsworth, M. L.; Stephenson, A. H.; Lonigro, A. J. American
Journal of Physiology 1996, 271 , H2717-H2722.

Sprague, R. S.; Stephenson, A. H.; Dimmitt, R. A.; Weintraub, N. A.; Branch, C.
A.; McMurdo, L.; Lonigro, A. J. American Journal of Physiology 1995, 269,
H1941-H1948.

Collins, D. M.; McCullough, W. T.; Ellsworth, M. L. Microvascular research
1998, 56, 43-53.

Genes, L. I.; Tolan, N. V.; Hulvey, M. K.; Martin, R. S.; Spence, D. M. Lab Chip
FIELD Full Journal Title:Lab on a Chip 2007, 7, 1256-1259.

McDonald, J. C.; Duffy, D. C.; Anderson, J. R.; Chiu, D. T.; Wu, H.; Schueller,
O. J .; Whitesides, G. M. Electrophoresis 2000, 21 , 27-40.

Kojima, H.; Nakatsubo, N.; Kikuchi, K.; Kawahara, S.; Kirino, Y.; Nagoshi, H.;
Hirata, Y.; Nagano, T. Analytical Chemistry 1998, 70, 2446-2453.

Kojima, H.; Urano, Y.; Kikuchi, K.; Higuchi, T.; Hirata, Y.; Nagano, T. Angew.
Chem, Int. Ed. 1999, 38, 3209-3212.

Meyer, J. A.; Froelich, J. M.; Reid, G. E.; Karunarathne, W. K. A. ,,Spence D. M.
Diabetologia 2008, 51, 175- 182.

Busse, R.; Ogilvie, A.; Pohl, U. Am J Physiol 1988, 254, H828-832.

Bogle, R. G.; Coade, S. B.; Moncada, S.; Pearson, J. D.; Mann, G. E. Biochem
Biophys Res Commun 1991, 180, 926-932.

141

(31)
(32)

(33)
(34)

(35)

(36)

(37)

(38)

(39)

Dean, B. M.; Perrett, D. Biochim Biophys Acta 1976, 43 7, 1-5.

Miseta, A.; Bogner, P.; Berenyi, E.; Kellermayer, M.; Galambos, C.; Wheatley,
D. N.; Cameron, 1. L. Biochim Biophys Acta 1993, 1175, 133-139.

Bergfeld, G. R.; Forrester, T. Cardiovasc. Res. 1992, 26, 40-47.

Ellsworth, M. L.; Forrester, T.; Ellis, C. G.; Dietrich, H. H. American Journal of
Physiology 1995, 269, H2155-H2161.

Goldsmith, H. L. J Gen Physiol 1968, 52, SSuppl-28s.

Dietrich, H. H.; Ellsworth, M. L.; Sprague, R. S.; Dacey, R. G., Jr. American
Journal of Physiology 2000, 2 78, H1294-H1298.

Dietrich, H. H.; Kajita, Y.; Dacey, R. G., Jr. The American journal of physiology
1996, 271, H1109-1116.

McCullough, W. T.; Collins, D. M.; Ellsworth, M. L. American Journal of
Physiology 1997, 2 72, H1886-H1891.

Houston, D. A.; Bumstock, G.; Vanhoutte, P. M. J Pharmacol Exp T her 1987,
241, 501-506.

142

CHAPTER 6

6.1 OVERALL CONCLUSIONS

At the dawn of the 21St century, with the accelerated pace of the pharmaceutical
and chemical industries, microﬂuidic technologies are emerging as powerful tools in
many stages of the drug discovery process. Microﬂuidic technologies have the potential
to radically transform the medicinal chemistry and biology interface.

In this work, we applied the recent developments in microfabrication technology
to create an in vitro model that will closely mimic an in vivo system. More speciﬁcally,
this work describes the development of a blood brain barrier (BBB) mimic with blood
components using lithographicalIy-derived microchips in order to monitor the fate of
endothelium-derived nitric oxide (NO) that is stimulated by mechanically deformed red
blood cells (RBCS). The microchip-based BBB mimic was developed by addressing an
array of cell reservoirs, which pose as the central nervous system (CNS), with a network
of underlying channels that represent the circulatory system. The cells were cultured on a
polycarbonate membrane, integrated in the device that separates the “so called” CNS
from the “so called” blood stream. The diameters of the patterned channels in the
fabricated chips will ultimately approximate those of brain capillaries in vivo.

Because a molecular level understanding of the mechanisms by which RBCS
deformability properties relate to the occurrence of multiple sclerosis (MS) is lacking, to

gain insight into these mechanisms would be ideal to employ such a model system. In this

143

study we hypothesized that RBCS obtained from patients with MS are more deformable,
therefore releasing more ATP, which in turn will lead to increased levels of NO
production. For a more realistic depiction of the mechanism of endothelium-derived NO
production in vivo, the RBCS obtained from MS patients were mechanically deformed by
hydrodynamically pumping them through microbore capillary tubing. A continuous ﬂow
system is employed to create stress on the RBCS once they have entered the microbore
tubing. It was found that MS-RBCs release higher levels of ATP compared to non-MS
RBCS. These ﬁndings suggest that RBCS of MS patients may indeed be in a continuous
“deformable “state that will lead to higher levels of ATP release and ultimately, to an
increase in ATP-stimulated endothelium-derived NO. Additionally, it was shown that
ATP acts as an agonist for NO production in bovine brain microvascular endothelial cells
(bBMVECs) and the increase in NO production is a function of cell division.

Based on these ﬁndings, we have constructed a theory that the overabundance of
NO metabolites in the CSF of MS patients is actually blood—borne. In other words, it is
possible that the NO levels are increased due to overstimulation by RBC-derived ATP.
Using chip technology, we were able to investigate the ﬂux of NO across the BBB
brought about by changes in ATP concentrations, both in the form of ATP standards and
RBC-derived ATP. As the detection scheme, ﬂuorescence microscopy was employed to
enable the qualitative and quantitative identiﬁcation of endothelium derived-NO after it
diffuses through the membrane of the microchip device.

Key to investigating the NO production from the endothelium and its fate through
the cell-coated polycarbonate membrane, that will constitute the BBB, is the ability to

actually detect the NO after it crosses the barrier. Primarily, we have demonstrated here

144

that the ﬂux of NO through a bare polycarbonate membrane can be detected using a
speciﬁc probe for NO, DAF-FM. However, the device lacks the sensitivity required to
detect physiological relevant levels of NO. Secondly, the ATP ﬂux through a
polycarbonate membrane in a 3D microﬂuidic device has been investigated. The
molecular transfer through the membrane was maximized by increasing the area of
contact between the two channels by designing the upper channel in a spiral shape that
crosses the underneath linear channel in multiple points. We showed here that the
proposed device is sensitive enough for the detection of as low as 500 nM of ATP spiked
into RBC samples and most importantly, is able to detect concentrations of ATP released
by iloprost-treated RBCS.

Parameters such as membrane pore size, ﬂow rates and probe concentrations have
been optimized to increase the efﬁciency of the detection in the device. Although the
results are promising, the inability to culture the cells on the membrane as part of the 3-D
microﬂuidic BBB and the poor sensitivity for the physiological relevant NO
concentrations constitute limiting factors.

Attempts have been made to culture the endothelial cells on isolated collagen-
coated polycarbonate membranes and, after the cells have reached a conﬂuent layer, to
integrate the membrane in the device. However, the two pieces of PDMS that make up
the device would not seal together as a result of the membrane wetness from the cell
culture media. As an alternative method, an attempt was made to culture the cells directly
on the membrane by introducing the cell suspension in the upper channel, above the
membrane. However, the lack of reproducibility and technical difﬁculties with the cell

introduction made the device unsuitable for the purpose of our applications.

145

One way of avoiding these complications would be the integration of microvalves
and multiple channels for the introduction of cells and DAF-F M, separately. Moreover,
cell culture reactors on—chip, where the viability of the cells can be sustained for days and
the cell—cell signaling can be monitored in real time, could also constitute an alternative
option. In terms of using the polycarbonate membranes as support for endothelial cell
culturing and proliferation, integration of transparent membranes for ease of optical cell
imaging would be highly desirable.

The aforementioned microﬂuidic components, such as microvalves, microreactors
or intricate designs are time consuming and prone to problems of engineering nature even
before introducing the biological factor. A new, simpler and more amenable design has
thus been constructed. Here, the high throughput capabilities. of a microtitre plate
technology was merged with the capabilities of microﬂuidic technology to create a device
that mimics the BBB in a more realistic way by incorporating blood ﬂow and having the
capability of possible multi-drug screening simultaneously and in real time.

In addition to interacting with an appropriate molecule, there are many
characteristics required of a drug candidate throughout the testing process. Drug
adsorption properties, its ability to be distributed in vivo, its effects on metabolism,
excretion characteristics and its toxicity are often monitored. Here, we showed that an
array of endothelial cells, addressable by an underlying microﬂuidic network of channels
containing RBCS, can be employed as an in vitro model of the in vivo circulation to
monitor cellular communication between different cell types. Results obtained from this
array suggest that the ability of iloprost, a stable analogue of prostacyclin, to stimulate

NO production in endothelial cells, may be due to its ability to stimulate ATP release

146

from the red cell. These results provide evidence that the described device may serve as a
controlled, in vitro platform for performing in viva-type measurements.

Moreover, this device was successfully used in a study that attempts to classify
the possible role of prolactin releasing peptide (PrRP) (a peptide that stimulate the release
of prolactin from the pituitary gland) in regulating the vascular tone by decreasing the
endothelium-derived NO production through a decrease in ATP.

In this study we propose a mechanism for the previously reported
cardiovascular/blood pressure regulatory properties of the PrRP, that is PrRP binding to
RBC-derived ATP that is released under various stimuli. If PrRP binds ATP, this will
lower the extracellular ATP pool available for the stimulation of the endothelial NO
synthesis and subsequently, a decrease in the vascular relaxation/increase in the blood
pressure might be possible. Herein, different techniques, such as mass spectrometry,
chemiluminescence and ﬂuorescence microscopy were utilized to show, directly or
indirectly, the PrRP binding to ATP. Moreover, by using an in vitro microﬂuidic-based
mimic of an in vivo microcirculation we constructed a system where the direct
involvement of the PrRP in the regulation of the blood pressure and/or vascular tone

could be monitored in real time.

147

6.2 FUTURE DIRECTIONS

The ultimate goal of the present work is to determine endothelium-derived NO,
using the microﬂuidic array device and high throughput ﬂuorescence detection, where the
NO is stimulated by ATP from deformed RBCS obtained from people with MS. In this
sense, the microﬂuidic device presented here will enable a comparison to be made
between the ability of RBCS from different patient groups to stimulate endothelium-
derived NO. This endothelium-derived NO and its fate through the BBB is important in
establishing the source of the elevated levels of nitrites and nitrates in the CSF of patients
with MS.

Based on our previous results involving the determination of RBC-derived ATP
from people with MS (Figure 10, chapter 2), it is anticipated that a higher level of ATP
release from those RBCs obtained from the MS patients will be measured in comparison
to the RBCS obtained from healthy controls. In our previous studies, the MS patient pool
was not subdivided into stage of disease; therefore, it will interesting to determine if there
is a further correlation of ATP release with stage of disease.

Another important direction of the project would be to investigate more closely
the relationship between the RBC-derived ATP release from MS patients and factors,
such as cholesterol and zinc, that have been reported to affect the erythrocytes membrane
ﬂuidity. One such report of lower cholesterol levels in the membranes of erythrocytes
obtained from people with MS suggests that the RBCS of patients with MS may be more
“deformable” because, typically, a decrease in cholesterol leads to an increase in

deformability.

148

 

In a series of papers nearly two decades ago, an analysis of erythrocyte membranes
obtained from controls and people with MS was performed. The data in these
manuscripts provided evidence that the levels of zinc in the membranes from the RBCS of
people with MS were signiﬁcantly higher than the RBCs obtained from controls." 2
Moreover, in the work by Dore-Duffy, the zinc levels found to be signiﬁcantly different
were compartmentalized to the red cell membrane. No signiﬁcant differences between
MS patients and controls were found when examining the cell as a whole, serum, or
plasma.2

It is known that zinc exerts physiological and biochemical roles as a component
of the known zinc metalloenzymes and also, plays an important role in the maintenance
of membrane structure and function.3 Therefore, it would be very interesting to try to
correlate the ATP release from RBCS of patients with different stages of MS and the
RBC membrane zinc content. This information would be critical in establishing the
relationship between RBC deformability, ATP release, endothelium-derived NO
production and elevated levels of NO metabolites found in CSF and other ﬂuids of MS
patients.

Another approach in trying to correlate ATP and MS has the basis in recent
reports in the literature, that inhibition of the P2X7 receptors, which are receptors for
ATP, found on oligodendrocytes prevented ATP triggering of these cells excitotoxicity.4
Moreover, a recent report by Feinstein5 has shown that P2X7 deﬁcient mice showed a
reduction in onset of experimental autoimmune encephalomyelitis (EAE), an
autoimmune mouse model of MS. Similar to the work by Feinstein, Voskuhl work

showed a reduction in EAE onset when the mice were administered estrogen analogues.6

149

Voskuhl’s results are key to the work that will be proposed here because preliminary
studies (not shown here) have found that estradiol has the ability to signiﬁcantly reduce
ATP release from RBCS. Thus, one future aim would be to determine the ability of
estradiol to attenuate shear-induced release of ATP from the erythrocytes of people with
MS.

Collectively, the preliminary data involving ATP release from the RBCS of MS
patients and reports in the literature indicating novel roles for ATP in a mouse model of
MS (EAE), suggest that events in the bloodstream may play a role in MS. As shown in
Figure 10 chapter 2, we propose that abnormally high levels of ATP release from the
RBC may be a determinant in the high levels of NO metabolites that have been
reportedly found in people with MS. High concentrations of NO production have been
reported to inhibit collagen production, the substrate for the endothelium, in vivo. Upon
inhibition of collagen production, the BBB is compromised and constituents in the
bloodstream (leukocytes, erythrocytes, bacteria, etc.) are able to penetrate through to the
CNS resulting in the hallmark features of MS (demyelination, oligodendrocyte toxicity,
iron deposits, bacterial infections, etc.). Unfortunately, investigating such a complex
event would be difﬁcult with current technology. Therefore, herein, we propose to
expand our current microchip based blood/endothelium model to include a collagen-
coated membrane and to determine the extent of NO mediated damage of the collagen
base and breakdown of the endothelium barrier. Subsequently, it would be worth
examining the integrity of the collagen-supported microchip-based BBB through a
separation and detection of cells and molecules breaking through the BBB when using

whole blood obtained from people with MS in the presence and absence of estradiol.

150

(1)

(2)

(3)

(4)

(5)

(6)

LIST OF REFERENCES

Cunnane, S. C.; Ho, S. Y.; Dore-Duffy, P.; Ells, K. R.; Horrobin, D. F. Am J Clin
Nutr 1989, 50, 801-806.

Dore-Duffy, P.; Catalanotto, F.; Donaldson, J. 0.; Ostrom, K. M.; Testa, M. A.
Ann Neurol 1983, 14, 450-454.

Bettger, W. J .; O'Dell, B. L. Life Sci. 1981, 28, 1425-1438.

Matute, C.; Torre, 1.; Perez-Cerda, F .; Perez-Samartin, A.; Alberdi, E.; Etxebarria,
E.; Arranz, A. M.; Ravid, R.; Rodriguez-Antiguedad, A.; Sanchez-Gomez, M. V.;
Domercq, M. J Neurosci. 2007, 27, 9525-9533.

Sharp, A. J .; Polak, P. E.; Simonini, V.; Lin, S. X.; Richardson, J. C.;
Bongarzone, E. R.; F einstein, D. L. J. Neuroinﬂammation 2008, 5, No pp given.

Kim, S.; Liva, S. M.; Dalal, M. A.; Verity, M. A.; Voskuhl, R. R. Neurology
1999, 52, 1230-1238.

151

"lllIlllIlllllllll