i

PEPTIDE MODIFICATION OF SODIUM ALGINATE TO

INDUCE SELECTIVE CAPTURE OF CARDIAC CELL

POPULATIONS

By

Melissa A. Brown

A thesis submitted in conformity with the requirements

for the degree of Master of Applied Science

Graduate Department of Chemical Engineering and Applied Chemistry

University of Toronto

© Copyright Melissa A. Brown 2009

ii

PEPTIDE MODIFICATION OF SODIUM ALGINATE TO INDUCE SELECTIVE

CAPTURE OF CARDIAC CELL POPULATIONS

Melissa A. Brown

Master of Applied Science, 2009

Graduate Department of Chemical Engineering and Applied Chemistry

Institute of Biomaterials and Biomedical Engineering

University of Toronto

ABSTRACT

Isolation of selected populations from heterogeneous cell mixtures and retrieval of the

captured population of interest for regenerative medicine and diagnostics applications is

one of the challenges that may be addressed by microfluidics. An affinity adhesion

strategy was tested using the tetrapeptides RGDS (arg-gly-asp-ser), REDV (arg-glu-asp-

val) and VAPG (val-ala-pro-gly) to modify an alginate hydrogel surface layer to

selectively adhere fibroblast (FB), endothelial (EC) and smooth muscle cell (SMC)

populations, respectively, of the non-myocyte cardiac cell fraction. Incorporation of

peptides into sodium alginate gel surface coatings demonstrated a preferential, seeding

density-dependent adhesion relationship on alginate-RGDS when tested with a

cardiomyocyte-depleted cell suspension in both static culture and in microfluidic devices.

Seeding density-dependent attachment was seen with close to 100% release of viable cells

from coated surfaces upon application of ethylenediaminetetraacetic acid (EDTA).

Further work will optimize the system with REDV and VAPG to capture ECs and SMCs.

iii

ACKNOWLEDGEMENTS

I would like to thank Dr. Milica Radisic for allowing me to pursue this research in this

highly competitive, innovative and interesting field. I would also like to thank my family

for their unwavering support and my labmates and friends for keeping my spirits high.

Thanks are also due to our collaborators Dr. Shashi Murthy and Brian Plouffe from

Northeastern University and to Rohin Iyer who made time to help with my experiments

and offered advice when I needed it; to Dr. Axel Guenther and Dr. Aaron Wheeler for

generously letting me use their lab equipment; to Dr. Julie Audet for sitting on my

committee and to Dr. Craig Simmons for bringing his expertise (and humour) to my

committee meetings.

This research was funded with grants from NSERC, and NIH, and with an Open

Fellowship from the University of Toronto.

iv

TABLE OF CONTENTS

ABSTRACT ........................................................................................................................ ii

ACKNOWLEDGEMENTS ............................................................................................. iii

TABLE OF CONTENTS ................................................................................................. iv

LIST OF FIGURES .......................................................................................................... vi

LIST OF SYMBOLS AND ABBREVIATIONS ........................................................... vii

1. BACKGROUND ........................................................................................................ 1

1.1 Motivation ............................................................................................................. 1

1.2 Hypothesis ............................................................................................................. 3

1.3 Overall Objective and Specific Aims.................................................................. 3

1.4 Literature Review ................................................................................................ 4

2. Methods ....................................................................................................................... 9

2.1 Cell source ............................................................................................................ 9

2.2 Cell culture and maintenance............................................................................. 9

2.2.1 Primary cardiac fibroblasts ................................................................................ 9

2.2.2 Cell lines ........................................................................................................... 10

2.3 Synthesis and purification of peptide-modified sodium alginate .................. 11

2.4 Fourier Transform Infrared Spectroscopy (FT-IR) quantification of RGDS

content in modified alginate ........................................................................................ 11

2.5 Device fabrication .............................................................................................. 12

2.6 Sample preparation ........................................................................................... 12

2.6.1 Coating glass coverslips for static studies ........................................................ 12

2.6.2 Coating of the microfluidic devices .................................................................. 13

2.7 Experimental procedure ................................................................................... 14

2.7.1 Static Experiments ....................................................................................... 14

2.7.2 Flow Experiments ............................................................................................. 16

2.8 Toxicity testing and viability assessment of EDTA treatment ...................... 17

2.9 Characterization of effects of operation parameters on input cell count and

composition ................................................................................................................... 17

2.10 Sample fixing and immunofluorescent staining .......................................... 19

3. Characterizing the syringe cell output for flow experiments in microdevices ... 20

3.1 Effects of pump angle and flow rate on cell output and composition ............... 20

3.2 Conclusion .............................................................................................................. 21

v

4. Selective adhesion of primary rat cardiac cells to peptide-modified sodium

alginate .............................................................................................................................. 22

4.1 Introduction ....................................................................................................... 22

4.2 EDTA dissolution of alginate ........................................................................... 25

4.3 Adhesion response to peptide modification and seeding density .................. 26

4.4 Overall cell recovery ......................................................................................... 28

4.5 Cell viability in static experiments ................................................................... 30

4.6 Cell capture and release in microfluidic devices ............................................ 31

4.7 Conclusion .......................................................................................................... 32

5. Characterization of alginate functionalized with alternative peptides REDV and

VAPG ................................................................................................................................ 34

5.1 Adhesion response of cell lines to alginate-peptide ........................................ 34

5.2 Characterization of alternate peptides with primary non-myocytes............ 37

5.3 Conclusion .......................................................................................................... 38

6 Summary, Future Work and Recommendations .................................................. 39

6.1 Summary of results ........................................................................................... 39

6.2 Future Work ...................................................................................................... 40

6.3 Recommendations ............................................................................................. 41

7 REFERENCES ......................................................................................................... 44

8. APPENDICES .......................................................................................................... 48

S1.1.2 GMBS functionalization of Hele-Shaw devices with peptides .......................... 48

S1.1.4 Immunofluorescent staining protocol ............................................................. 49

S1.1.5 Results and Discussion – isl-1 staining on adhered cells in Hele-Shaw

devices........................................................................................................................ 50

vi

LIST OF FIGURES

Figure 1: Cell source for experiments................................................................................ 9

Figure 2: Experimental procedure and cell collection stages for static experiments. ...... 15

Figure 3: Test setup for characterizing effects of operation parameters on input cells. .. 18

Figure 4: Effects of pump tilt angle and flow rate on total cell output and cell suspension

composition by staining for TnI and Vim. ........................................................................ 21

Figure 5: FT-IR quantification of peptide content of alginate-RGDS.. ........................... 25

Figure 6: Cell viability and retrieval with EDTA and trypsin treatment.. ....................... 26

Figure 7: Average adhered cells per coverslip, adhesion on glass vs. on alginate-RGDS

and unmodified alginate surfaces...................................................................................... 27

Figure 8: Percentage of cells released from sample surfaces with EDTA treatment based

on image analysis.. ............................................................................................................ 28

Figure 9: Percentage of total cells retrieved from the washing steps over the experiment

as a percentage of the initial seeded cell number. ............................................................. 29

Figure 10: Percentage of total cells retrieved over the experiment as a percentage of the

initial seeded cell number , with control PBS- solution in the release step.. .................... 29

Figure 11: Percentage viability of cells collected at Wash 1, Wash 2 and EDTA-release

time points.. ....................................................................................................................... 31

Figure 12: Cell attachment and detachment of cardiac fibroblast at a shear stress of 1 dyn

cm-2

in microfluidic channels............................................................................................ 33

Figure 13: H5V endothelial cell line adhesion to peptide-modified alginate.. ................ 35

Figure 14: A7r5 smooth muscle cell line adhesion to peptide-modified alginate.. ......... 36

Figure 15: NIH 3T3 fibroblast cell line adhesion to peptide-modified alginate.. ............ 36

Figure 16: Cell adhesion response to peptide-modified alginate. .................................... 37

Supplementary Figures

Figure S 1: Geometry of Hele-Shaw microfluidics device showing shear stress profile

and characteristic equation. ............................................................................................... 50

Figure S 2: Immunofluorescent staining for progenitor cell islet 1 markers in captured

cells in peptide-modified microfluidic devices.. ............................................................... 51

vii

LIST OF SYMBOLS AND ABBREVIATIONS

CM cardiomyocyte

DAPI 4',6-diamidino-2-phenylindole

DMEM Dulbecco’s Modified Eagle Medium

DMSO dimethyl sulfoxide

EC endothelial cell

EDC (1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride)

EDTA ethylenediaminetetraacetic acid

FB fibroblast cell

FBS Fetal Bovine Serum

GMBS N-[g-Maleimidobutyryloxy]-succinimide ester

HBSS Hank’s Balanced Salt Solution

HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid)

PBS- Phosphate Buffered Saline (without calcium or magnesium ions)

PBS+ Phosphate Buffered Saline (with calcium or magnesium ions)

PDMS poly(dimethylsiloxane) (aka silicone)

PFA paraformaldehyde

PP1 cells Preplate 1 cells - collected after one preplating step

Pre-PP1 cells Cells collected prior to the preplating step

Preplate cells Cells left adhered to flasks during the preplating step when PP1

cells are removed

Silane 3-mercaptopropanetrimethoxysilane

SMC smooth muscle cell

TC PS tissue culture treated polystyrene

TnI anti cardiac troponin-I antibody

Vim anti-vimentin antibody

1

1. BACKGROUND

1.1 Motivation

Coronary Heart Disease (CHD) is currently the leading cause of mortality in North

America and was estimated to have a total cost in excess of $156 billion US in medical

costs and lost productivity in 2008. At present over 16 million people in North America

suffer from CHD, with approximately 770 000 new cases being treated annually. Out of

the CHD sufferers the vast majority experience at least one myocardial infarct (MI) event

or heart attack [1]. In this condition blockages in the coronary arteries surrounding the

heart lead to loss of perfusion in regions of the cardiac muscle and subsequent cell

damage or death. This damaged tissue remains functionally dead and usually results in

weakened organ function. For decades scientists and researchers have investigated ways

to mitigate, repair or reverse the damage caused by a MI. Many of these strategies

involve direct injection of various cell types into damaged regions to promote repair and

regeneration [2, 3] while others focus on fabricating engineered tissues and cell sheets for

direct grafting onto the heart muscle [4-6].

A crucial requirement for all these approaches is a cell source which could be expanded

in vitro and be able to successfully re-establish the heart organizational structure and

function. Many groups have used cardiomyocytes (CMs) as the obvious choice for heart

repair [2, 7] while others have explored stem cells and their more differentiated progeny

for implantation [8-11]. The left ventricle of the adult heart is extremely dense,

containing approximately 5 x 109 cells in total and it is estimated that one billion cells

are lost during a MI [12]. Adult cardiomyocytes are considered to be terminally

differentiated with little proliferative capacity thus high cell numbers are required for

regenerative medicine applications as self renewal and propagation are not possible.

Embryonic stem cells are expandable and with directed differentiation can be made to

produce cells in the cardiac lineage [13]. This tactic does have several drawbacks, key

among them being the teratogenic potential of implanted undifferentiated embryonic stem

cells as well as the heterogeneity of the cell mixtures produced through various

2

differentiation techniques [14-16]. Additionally, embryonic stem cells have many ethical

concerns which limit their potential usage in cell therapies. In the search for a less

controversial cell source several groups have recently published their discovery of

resident stem cells in the adult heart [5, 17-19]. This endogenous stem cell population is

attributed with maintaining their regenerative, proliferative and differentiation capacity

over time and can develop into various cardiac cell types. The main differentiated cell

types needed for repair of myocardium are contractile cardiomyocytes (CMs). Fibroblasts

(FBs) are reqired for extracellular matrix production, endothelial cells (ECs) and smooth

muscle cells (SMCs) for vasculature. Finally, as this is an autologous cell source the

reduced teratogenic and immunogeniticity potential compared to embryonic stem cells

make it very attractive for use in cardiac regeneration therapies.

The greatest obstacle to utilizing this seemingly ideal cell source is the relative rarity of

adult stem cells (less than 1% of cells in the adult heart) [5, 18] and in some cases (e.g.

isl-1+ cells) the lack of known characteristic surface receptor groups for this population.

Knowledge of these surface groups would allow for identification and sorting of the cells

using existing standard technologies such as flow cytometry, fluorescence-activated cell

sorting (FACS) and magnetic cell sorting (MACS). As characteristic surface receptor

assemblies are not known, an alternate approach could be to isolate or enrich the

population of interest by negative selection (removal of other cell types).

Several strategies have been attempted in cell separation studies, from sorting of cells by

size using microseives and microchannels, to surfaces altered to recognize cells by

addition of various affinity groups such as antibodies or proteins [20, 21]. One of the

more successful approaches has been used in the isolation of T and B lymphocytes from

model mixtures [22] and whole blood [23] using antibody-mediated selection in

microscale devices. A microfluidics device model has several advantages for use in cell

separation. These devices can have complex designs while maintaining ease of

fabrication and use, cost effectiveness and high efficiency. Additionally control of the

surface chemistry within these microchannels can synergistically aid device design in the

separation process. A hindrance to the success of microfluidic devices in achieving this

3

goal lies in the inaccessibility of cells within microchannels upon successful capture.

Cells can be cultured within microchannels once bound to the channel surface but they

cannot be removed easily through use of simple treatments apart from utilization of

proteolytic enzymes such as trypsin [24, 25].

In order to address all these issues we envision a high throughput microfluidics system

which can be implemented in a clinical setting for isolation of cell populations from

patient cardiac tissue biopsies. CMs can be isolated by physical size as they are the

largest cells. Various non-myocyte populations can then be isolated by adhesion

separation. To do this we must first select an adhesion strategy highly specific to the

various cardiac cells and then characterize the effectiveness of the system. We can then

adapt and optimize the modification technique to allow for successful recovery of the

cells while ensuring that they remain viable. To this end we have selected a peptide

surface modification strategy and in the following demonstrate our efforts towards

isolation of neonatal rat cardiac populations.

1.2 Hypothesis

Non-adhesive sodium alginate hydrogel could be modified with tetrapeptides to render it

selectively adhesive to the major non-cardiomyocyte cell populations comprising the

neonatal rat heart while reducing non-specific cell adhesion. Cells adhered to an alginate-

peptide layer can be recovered without enzymatic treatment, collected and used in further

experiments. These methods extended to microfluidic devices can allow for selective cell

capture and retrieval in flow conditions.

1.3 Overall Objective and Specific Aims

The objective of this study was to characterize the use of three tetrapeptides, RGDS (arg-

gly-asp-ser), REDV (arg-glu-asp-val) and VAPG (val-ala-pro-gly), to modify a sodium

alginate hydrogel for use as a surface layer to selectively adhere primary rat cardiac cell

populations in static culture conditions and in microfluidic devices.

4

Specific Aims:

1. Modification of sodium alginate with the tetrapeptide RGDS and characterization

of the adhesion and EDTA-release characteristics using primary cardiac

fibroblasts in static culture conditions.

2. Characterization of the cell output from syringe pumps to determine the operating

conditions for cell delivery to microfluidic devices.

3. Characterization of the capture and recovery of primary cardiac fibroblasts on

alginate-RGDS surface coatings in microfluidic devices under flow conditions.

4. Modification of sodium alginate with REDV and VAPG peptides to establish the

adhesion response of endothelial and smooth muscle cells

1.4 Literature Review

The recent discovery of resident cardiac stem cells has gained much public interest, with

various groups reporting finding cardiac progenitors in the adult heart. The cells were

identified based on expression of stem cell markers such as isl-1 [17] , c-kit and Sca-1

[18, 26] and on differentiation and renewal characteristics. While there is no general

consensus on the definitive hallmarks for this population the evidence points to the

existence of a rare population of cardiac cells, comprising fewer than 1% of the total cells

[18], which can self renew and appears able to differentiate into all the cells in the cardiac

lineage. These resident progenitors may be involved in repair and maintenance of the

cardiac tissue especially after injury and could play a vital role in treatment strategies for

damaged hearts. The greatest hurdle to be overcome, however, lies in positively

identifying and isolating these cells from the rest of the heart cells; a difficulty with such

low occurrence in the native heart cell tissue. Current approaches in tissue engineering

and regenerative medicine are limited by the deficiencies of using terminally

5

differentiated non-proliferative cardiomyocytes, or from teratoma development and

limited control over directed cell differentiation with embryonic stem cell utilization [27].

For any treatment strategies involving these resident cardiac stem cells a purification step

would be necessary to isolate or enrich the cells before use. The current technological

standards for cell sorting are fluorescence-activated cell sorting (FACS) and magnetic

cell sorting (MACS) [21]. A commonality between these methods is the prerequisite of

knowledge of a cell surface receptor group characteristic to the population. Once this

condition is satisfied, appropriate adhesion molecules on target cells can be selected for

with relative ease. Our knowledge of the phenotype and microstructure of these cells is

however rudimentary at best. The few known identifiers of this progenitor population

include the marker isl-1 and early cardiac transcription factors GATA-4 and nkx2.5 [4,

18]. As these are not surface-expressed, membrane permeabilization is required for the

immunolabelling step prior to sorting by conventional means (a process resulting in cell

death). Additionally, as these are primary cells from a genetically unmodified source,

they cannot be pre-labeled with markers such as green or yellow fluorescent proteins for

identification as is done with many embryonic stem cell lines. As such, an alternative

method must be found by which the cells of interest can be selected for while maintaining

viability and normal cellular development after processing.

In the field of microfluidics various non-invasive methods of cell separation have been

utilized. On the microscale cell sorting has been done using fluid jets, features of the

microchannel design, sieves and electrophysical means to segregate cells based on

physical attributes such as size and electric potential. In the cardiac cell environment the

greatest physical difference in cell populations is that of size; cardiomyocytes are

significantly larger than non-myocytes. As the stem cell population is expected to

comprise a part of the non-myocyte fraction, a first step would logically be the removal

of cardiomyocytes and collection of the non-myocyte portion. Many size-separation

microdevice designs have been tested, with microchannel geometry (varying channel

height), microsieves and complex flow patterns being implemented to evince segregation

[28-31]. Taking apart the non-myocyte fraction however poses a greater stumbling block.

6

Once CMs have been removed, the remaining cells do not have major distinguishable

physical characteristics to tell them apart. For this case ultrastructural differences must be

examined to select for or against a particular cell type.

Similar difficulties have been resolved in the realm of microfluidics in the separation of

leukocytes from whole blood [32] and negative selection of fibroblasts, endothelial cells

and smooth muscle cell lines from heterogeneous mixtures [20, 33]. In each situation

cells were selected for through contact with surfaces modified with either antibodies (for

the blood cells) or with peptide sequences specific to each cell type (for the multiple cell

lines). The latter example can be adapted reasonably well to our situation. As fibroblasts,

smooth muscle cells and endothelial cells comprise the majority of the non-myocyte

fraction (with relative proportions in the neonatal heart of 49%, 3-4% and 2-3%

respectively [34]) this peptide adhesion strategy may prove suitable for separation of the

non-myocyte cell mixture.

Plouffe et al. utilized three tetrapeptides in a sequential negative selection process to

remove ≥ 96% of cells from an initial heterogeneous mixture of all three cell types with

>83% purity of the captured cells [20]. Peptides val-arg-pro-gly (VAPG) and arg-glu-asp-

val (REDV) are selective for integrins β3 and α4β1 respectively, recognizing smooth

muscle and endothelial cells in that order, and arg-gly-asp-ser (RGDS) specific to the

integrin sequences β1 and αvβ3, and binding to all cell types [35-37]. The use of peptides

is an innovative approach for many reasons; peptides are easy to synthesize, are more

stable than proteins and can be incorporated in high concentrations due to their small

physical size. Additionally, careful selection of peptide sequences can result in high

affinity cell recognition.

Finally, a concern not usually addressed in the field of microfluidics today is cell retrieval

after capture. Many groups have worked on the capture of cells within microchannels via

interaction with groups covalently bound to the channel surfaces, making it difficult to

later retrieve cells from the channels if they are required for some alternate purpose. For

tissue engineering research Shimizu et al. have produced removable cardiomyocyte

7

monolayers on the macroscale [38], and Ernst et al. detachable fibroblast layers in

microfluidic devices [39] using thermo-responsive polymers which alter the surface

hydrophobocity with temperature. We envision that cells could perhaps be captured on a

removable layer containing peptide groups on the cell-contacting surface. An attractive

system would be a hydrogel such as sodium alginate which can be converted from a

liquid to semi-solid gel with addition of divalent cations (and revert to liquid form with

ion removal). This material has been studied in many tissue engineering and cell injection

applications where alginate gels were used as a biocompatible scaffold or encapsulator to

sustain cell growth and infiltration over the short term [40, 41]. Moreover, several groups

have successfully modified alginate gels with peptides (including RGD) to improve cell

attachment, proliferation and infarct healing in the heart [27, 42, 43]. An alginate system

thus has several advantages; ease of attachment of peptides with 1-ethyl-3-[3-

dimethylaminopropyl]carbodiimide hydrochloride (EDC) chemistry, adaptability for use

in any microsystem geometry, simplicity of gelling and dissolution procedures and

proven biocompatibility in our model system

Here we have described several techniques which may be used to develop a multistage

system for systematic depletion and cell retrieval from a heterogeneous cardiac cell

mixture. The major step would engender a negative selection process whereby cells are

removed through interaction with peptide groups incorporated onto the contact surface.

This involves removal, in stages, of endothelial cells, smooth muscle cells and

fibroblasts. As a shear-stress dependent optimal adhesion relationship was observed for

each cell line previously [20]. It is also imperative to determine the shear stress behaviour

of primary cells in the system. A previously characterized device used by Murthy et al.

[22] based on the Hele-Shaw principles of shear flow in microdevices and on the design

of Usami et al. [44] provided a microchannel with linear shear stress profile which could

be used to establish the shear stress adhesion behaviour of cells. A system to recover

adhered cells while maintaining viability will be investigated with a peptide-modified

alginate hydrogel coating forming the basis of the study. One or several of the approaches

described above may be combined in our final methodology and applied to our goal of

8

finding a microfluidic solution to the problem of isolating a rare cardiac stem cell

population.

9

2. Methods

2.1 Cell source

Neonatal (1 to 2 days-old) Sprague–Dawley rats were euthanized according to the

procedure approved by the University of Toronto Committee on Animal Care. The hearts

were removed, quartered, and the cells were isolated by an overnight treatment with

trypsin (4 °C, 6 120 U/mL in Hank’s Balanced Salt Solution, HBSS, 50 rpm agitation on

an orbital shaker), followed by a serial collagenase digestion (220 U/mL in HBSS) as

described in previous work [45]. The supernatants from five collagenase digests of the

tissues were collected and centrifuged at 750 rpm for 4 mins, resuspended in culture

medium, and then preplated into T75 flasks (Falcon) for 1hr followed by removal of the

supernatant. To obtain essentially pure cardiac fibroblasts, the cells that remained

attached after one preplating step were expanded for 8 days before being trypsinized.

Figure 1: Cell source for experiments. Diagram depicts steps to obtain pre-preplate cells at physiological

ratios for flow characterization and cardiac fibroblasts without cardiomyocytes for adhesion experiments.

2.2 Cell culture and maintenance

2.2.1 Primary cardiac fibroblasts

The adherent cells remaining in the flask after the preplating step (here called preplate

cells) consisted mainly of cardiac fibroblasts which proliferate rapidly in culture [46].

10

The cells were cultured for eight days in tissue culture flasks and then trypsinized with

0.25% porcine trypsin-EDTA (Gibco, 25200-072) and re-plated. This trypsinizing step

removes any remaining CMs from the culture. Culture medium was refreshed in the

flasks of the passage 2-8 cardiac fibroblasts one day prior to the running of each static

experiment. For trypsinization flasks were rinsed in PBS and incubated with 0.25%

trypsin-EDTA for 4 minutes at 37°C in a humidified incubator (Hera Cell 150, Mandel,

Guelph, Ontario) and quenched with CM medium. The CM medium consisted of

Dulbecco’s Modified Eagle Medium (DMEM, Gibco 11965-092) with 4.5 g/L glucose, 4

mM L-glutamine, 10 % fetal bovine serum (FBS, US Certified, Gibco, 16000-044), 100

U/mL penicillin, 100 µg/mL streptomycin (Pen-Strep, Gibco, 15140-122), and 10 mM 4-

2-hydroxyethyl-1-piperazineethanesulfonic acid buffer (HEPES, Gibco 15630-080). For

static experiments the cell suspension was centrifuged at 1500 rpm for 7 minutes, the

supernatant aspirated and the cells resuspended in serum-free CM medium to the desired

concentration and seeded onto each surface in a 15 µL volume. Cells were isolated and

resuspended in serum free media at a concentration of 100 000

cells/mL for all

subsequent flow experiments.

2.2.2 Cell lines

Cell lines were used to characterize adhesion properties of homogeneous cell

suspensions: rat aortic smooth muscle cells (ATCC, A7r5), mouse endothelial (ATCC,

H5V) and mouse fibroblast cells (ATCC, NIH 3T3).

A7r5 and H5V cells were thawed into tissue culture flasks containing medium composed

of 89% v/v 4.5g/L high glucose Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma,

D5796), 10% v/v Fetal Bovine Serum (FBS, Gibco, US Certified, 16000-044), 100 U/mL

penicillin G, 100 µg/mL streptomycin (Pen-Strep, Gibco, 151401-122). The cells were

passaged at or before confluence and medium changed every 3-4 days. NIH3T3 cells

were thawed into tissue culture flasks containing CM medium (composition described in

Section 2.2.1).

11

2.3 Synthesis and purification of peptide-modified sodium alginate

Low viscosity sodium alginate was obtained (Sigma, A2158) and made up to 2% wt/vol

in 2-(N-morpholino)ethanesulfonic acid (MES) buffer (Sigma, 1M, M1317). EDC (1-

ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride) chemistry was utilized to

covalently attach a tetrapeptide RGDS (arg-gly-asp-ser, American Peptide, 44-0-14),

REDV (arg-glu-asp-val, American Peptide, 44-0-12) or VAPG (val-ala-pro-gly, Sigma,

V-0883) to the alginate. Briefly, for each 800µl batch of alginate-peptide, 15mg alginate,

2.5x10-5

mols EDC (Pierce, 22980) and 6.1x10-5

mols N-hydroxysulfosuccinimide (sulfo-

NHS, Pierce, 24510) were added to 300µl MES buffer, vortexed to mix, and incubated at

room temperature for 2 hours. Five hundred microlitres of 1mg/ml peptide in PBS-

(Gibco, 10010-023) was then added and allowed to react at room temperature for 20

hours. This peptide concentration was selected after consideration of other studies

utilizing RGDS peptide for selective cell adhesion [20, 40]. The alginate-peptide was

then dialysed for 2 days (3500 MWCO) to remove all unreacted reagents. The resulting

solution was flash frozen with liquid nitrogen and lyophilized for 2 days.

For static experiments the alginate-peptide was reconstituted to 2% wt/vol in MES buffer

at room temperature, sonicated for 10 minutes and concentrated sodium hydroxide added

(to a final concentration of 1.67µM NaOH/mg alginate-peptide) to adjust the solution pH

and viscosity. For flow experiments the freeze dried alginate-RGDS was dissolved in

MES buffer to a concentration of 12mg/mL. To assist in dissolving, the solution was

repeatedly vortexed and briefly incubated at 37°C to hydrate. Additionally, 10 N NaOH

was added to the solution to adjust viscosity. The solution was centrifuged very briefly to

remove any residual bubbles.

2.4 Fourier Transform Infrared Spectroscopy (FT-IR) quantification of RGDS

content in modified alginate

RGDS peptide content was quantified using Fourier Transform Infrared Spectroscopy

(FT-IR, Perkin Elmer Spectrum 2000) by Dr. Murthy at Northeastern University. Ten

microliters of an aqueous sample of 100 μg/mL, 50 μg/mL and 25 μg/mL were analyzed

12

on poly(tetrafluoroethylene) IR sample cards (Crystal Labs, Garfield, NJ) to develop a

calibration. This was followed by dissolving 1 gram of alginate-RGDS in 1 milliliter of

an aqueous buffer, which was then analyzed. Curves were compared and the RGDS

content was determined by the amide stretching peak at 638 cm-1

to be 38 µg of RGDS

per milligram of solid modified alginate stock.

2.5 Device fabrication

The design and fabrication of the microfluidic devices followed previously described soft

lithography techniques [47] and was performed entirely at the laboratory of Dr. Murthy at

Northeastern University. The chamber that was utilized was a 1 × 50 × 0.07 mm (W × L

× H) straight channel. A negative master was fabricated and assembled at the George J.

Kostas Nanoscale Technology and Manufacturing Research Center at Northeastern

University using conventional photolithography techniques. To form the polymeric

chambers, poly(dimethylsiloxane) (PDMS, Dow Corning) elastomer was mixed (1:10

ratio), poured onto the negative master wafer, degassed, and allowed to cure overnight.

PDMS replicas were then pulled off the wafers, inlet and outlet holes were punched with

a 19-gauge blunt-nose needle and exposed to oxygen plasma and then immediately

placed in contact with glass cover slides in order to create an irreversible bond between

the PDMS and glass.

2.6 Sample preparation

2.6.1 Coating glass coverslips for static studies

A solution of 100mM CaCl2 was made up in distilled water and filtered with a 0.45µm

filter. Circular glass coverslips 12mm in diameter (VWR, 89015-724) were treated in a

plasma chamber for 30 seconds and alginate or alginate-peptide pipetted onto each

surface. The coated coverslips were immediately placed into a spin-coating machine

(Model WS-400B-6NPP/LITE, Laurell Technologies Corporation, North Wales, PA,

USA) and spun for 4 seconds at 5000rpm to produce a uniform thin coating on the glass

surface. Each coverslip was carefully removed from the device with tweezers and placed

13

into a Petri dish containing CaCl2 to gel the surface coating. This procedure was

performed in small batches to prevent the solutions on the coverslips from drying out

before being placed in the spin coater. Samples were kept overnight in the CaCl2 bath at

4°C. Prior to use, uncoated and coated coverslips were rinsed in sterile PBS+ (PBS

containing both calcium and magnesium ions, Gibco, 14040-133) and then placed into a

sterile well plate. Cell seeding was executed as quickly as possible to prevent drying of

the coated surfaces.

2.6.2 Coating of the microfluidic devices

In the laboratory of Dr. Murthy at Northeastern University four different alginate

monomer solutions in MES buffer (Pierce Biotechnology, Rockford, IL) (2, 6, 12, and 16

mg/mL) were investigated by Mr. Brian Plouffe to qualitatively determine the efficiency

of adsorption. To each solution, 100 µL of fluorescein isothiocyanate (FITC; Vector

Laboratories, Burlingame, CA) was added as a fluorescence indicator. Each solution was

injected into a microfluidic device and allowed to adsorb for 1 hr. The solution was then

rinsed with MES buffer at 10 μL/min for 10min followed by a 100 mM CaCl2 (Sigma

Aldrich) solution at 10 µL/min for 10 min to form a thin layer of alginate gel on the walls

of the microchannels. Each device was imaged under a Nikon Eclipse TE2000 inverted

microscope at 10 magnification using a fluorescein (480 ± 30 nm/535 ± 40 nm)

excitation/emission filter. It was determined that there was no fluorescence at

concentrations of 2 and 6 mg/mL. Additionally, 16 mg/mL concentration was found to

be too viscous a working solution for injection; therefore for all subsequent alginate and

alginate-RGDS experiments in microfluidic devices 12 mg/mL monomer solutions in

MES buffer were utilized. Both alginate and alginate-RGDS adsorptions were allowed to

adsorb for 1 hr. and rinsed with MES buffer at 10 μL/min for 10min followed by a 100

mM CaCl2 (Sigma Aldrich) solution at 10 µL/min for 10 min. to form the hydrogel layer.

14

2.7 Experimental procedure

2.7.1 Static Experiments

For each condition a bare glass positive control and an alginate surface negative control

were compared with alginate-peptide surfaces.

Passage 1 cardiac fibroblasts (or a cell line) were trypsinized and resuspended in serum-

free CM medium (or PBS) such that the required number of cells was seeded onto each

surface in as small a volume as possible (15-45 µL). To determine the relationship

between cell seeding density and the number of cells adhered to the surfaces, experiments

were conducted with cell seeding densities of 10 000, 50 000, 100 000 and 500 000 cells

per coverslip. The cell suspension was carefully spread with a pipette tip to cover the

entire surface of the coverslip (ensuring that the surface coatings were not disturbed) with

the liquid staying on top each coverslip and not spreading to the well surface.

Plates containing coverslips were placed into a 5% CO2 humidified incubator at 37°C for

30 minutes. After the incubation 500µl of CM medium was carefully added to each well

and the plate gently agitated to remove non-adherent cells (Wash 1). The coverslips were

removed to fresh wells with 1mL warm CM medium already in the wells. The CM

medium from the CM medium wash was collected and pooled for each surface treatment

(Wash 1). The empty wells were rinsed with PBS- (PBS devoid of both calcium and

magnesium ions) and this liquid added to Wash 1 tubes. The plates were incubated for 1

hour (at the same conditions as before) before imaging the coverslips in an environmental

chamber at 37°C. Coverslip surfaces were imaged by brightfield microscopy with an

Olympus IX81 microscope with a Qimaging camera (Model Retiga) and the total number

of cells adhered to each surface was then determined by cell counts and the results

graphed.

The CM medium in the wells was then collected (Wash 2) and the coverslips carefully

rinsed with PBS+ to remove any serum (and the PBS+ pooled with Wash 2 liquid).

Coverslips were removed to a fresh plate where 500µl of 50mM EDTA

(ethylenediaminetetraacetic acid, Bioshop, EDT111.500, 0.5M stock diluted in PBS-) at

15

37°C was added to each well to re-solubilize the alginate and alginate-RGDS and remove

cells from glass coverslips. Plates were placed on an orbital shaker for 10 minutes at

70rpm at room temperature before 1ml of warm CM medium was added to quench the

reaction. The liquid was collected and the coverslips rinsed in PBS-. (In some cases an

additional wash was implemented to remove non-adherent or dead cells from the sample

surfaces prior to imaging). Coverslips were then re-examined and samples were imaged

again in case of remaining adhered cells.

Figure 2: Experimental procedure and cell collection stages for static experiments. Diagram plots the

time course of the experiment, showing the main stages of cell seeding, sample incubation, imaging and

treatment with EDTA. Cell collection time points are also highlighted.

Pooled cell suspensions from Wash 1, Wash 2, (Wash 3 if applicable) and the releasing

EDTA step were centrifuged at 1500rpm for 7 minutes to pellet the cells. Supernatant

was aspirated and the cells resuspended in 100-200µl of PBS-. The total cell count and

viability of collected cells was determined via hemacytometer counts with trypan blue

staining. Cells were also manually counted from brightfield images of each coverslip (n >

6 images per sample, n ≥ 3 samples per group) to determine the average number of cells

adhered to each coverslip. A mass balance was performed to determine the percentage

recovery of total cells seeded. Cell counts on images were performed using ImageJ

software (version 1.36b, NIH, Bethesda, MD). Statistical analysis was done using

16

SigmaStat 3.0 software, and groups compared using 1- and 2-way ANOVA and the

paired t-test.

2.7.2 Flow Experiments

Flow experiments were performed by Mr. Brian Plouffe in the laboratory of Dr.

Murthy at Northeastern University. The microfluidic device used in this investigation

was a 1 mm × 50 mm × 0.07 mm (W × L × H) channel, fabricated by standard soft

lithography techniques. Cell adhesion was investigated on two different channel

surfaces: channels coated with unconjugated alginate hydrogel, and channels coated

with peptide-conjugated alginate hydrogel. Cardiac fibroblasts were used for all

experiments and fluid flow rates for all steps (incubation, rinse, cell seeding, and cell

release) were controlled to ensure a shear stress of 1 dyn/cm2.

Microchannel surfaces were coated using a 12 mg/mL solution of native alginate or

peptide-modified alginate. Per the calibration shown in Figure 5, the 12 mg/ml

alginate-RGDS has a peptide content of 456 µg RGDS/mL.

Following gel formation the CaCl2 solution was rinsed from the device at 5 μL/min for

10min. with MES buffer. Suspensions of cardiac fibroblasts (10 × 104

cells/mL) were

flowed through microfluidic devices at a flow rate of 5 μL/min for a period of 20 min,

respectively, using a Harvard Apparatus PHD 2000 syringe pump (Holliston, MA). Cell

adhesion within the devices was measured using a field finder (with 1 mm × 1 mm grids)

placed under the microfluidic chamber. Adhered cells were manually counted at selected

points along the device axis under a Nikon Eclipse TE2000 inverted microscope. Cell

counts were taken between 5 and 25 mm from the device inlet, along the device axis. All

flow experiments were performed at room temperature. Next, a 50 mM solution of

EDTA was injected into the device at 5 μL/min for 10 min followed by another manual

cell count.

17

For flow devices cell viability was examined for cells captured and subsequently released

within the microfluidic devices. Viability was also checked prior to experimentation for

comparison. For viability analysis pre- and post-experimentation, cells were incubated in

a 4 μM EthD-1 (dead cell indicator) and 2 μM calcein (live cell indicator) solution in

media for 1 hr. Live and dead cells were visualized and counted at 10× magnification

using fluorescein and rhodamine filters, respectively, on the Nikon microscope. To

ensure that EDTA would not adversely affect cell viability in post-experimental assays,

cells detached from the device were collected in centrifuge tubes containing 1 mL of CM

medium (20× dilution) to neutralize EDTA.

2.8 Toxicity testing and viability assessment of EDTA treatment

To determine which concentration of EDTA would be least harmful to the cells an initial

toxicity study was conducted. To 15µL of cardiac fibroblasts in suspension (6.7 x 106

cell/mL in serum-free medium were added) 500µL of EDTA was added at concentrations

of 0.5mM, 5mM, 10mM, 25mM and 50mM. The cells and EDTA were incubated for 15

minutes at room temperature on an orbital shaker at 70 rpm. The cells were then

centrifuged and resuspended in 100µL of PBS-. Cell counts were performed using a

hemacytometer and trypan blue staining to assess the number and viability of cells

retrieved after the treatment. Input cells were assessed to be > 98% viable. Trypsin and

PBS- were added to the cells in place of EDTA as positive and negative controls

respectively and CM medium was added to the trypsin group after the incubation to

quench the reaction. To account for potential effects of the CM medium, one vial was

treated with the highest EDTA concentration (50mM) and CM medium added after the

incubation step.

2.9 Characterization of effects of operation parameters on input cell count and

composition

To extend the alginate-peptide system to use in microdevices the effects of experiment

operation parameters on the input cells needed to be characterized. For these experiments

cardiac cells at physiological ratios were obtained prior to the preplating step (pre-PP1) of

18

the cell isolation procedure described in Section 2.1. In normal experiments cell

suspension was loaded into syringes and run with the pump tilted at 10° to the horizontal.

Cells flowed through tubing and entered the devices at a flow rate of 10µL/min over the

course of a 45 minute experiment. The main operation parameters which could affect the

cells included the pump flow rate and tilt angle. Normal and test setups are shown in

Figure 3 below.

Figure 3: Test setup for characterizing effects of operation parameters on input cells. Two parameters

to be varied were pump flow rate (yµL/min) and tilt angle of the pump above horizontal (x°). Tubing 15cm

connected the syringe to the collection tube, which would quantify the cells normally entering the inlet of

the microfluidics device.

For evaluating the contribution of pump tilt angle on the number and composition of cells

entering the devices the pump flow rate was fixed at 10µL/min (the flow rate initially

used in experiments with the devices) and the pump tilt angle set at either 0°, 45° or 90°

(vertical) to the horizontal. Cells were infused with the pump over 45 minutes at room

temperature and collected as shown for analysis. To determine the effects of the pump

flow rate angle on the cells entering the microdevices, the pump tilt angle was fixed at

10° to the horizontal and the syringe pump was run at flow rates of 10µL/min, 20µL/min

and 30µL/min. The total number of cells exiting the tubing at the ‘device inlet’ point was

determined by cell hemacytometer counts. Cells were then cytospun onto slides and

immunostained for cardiac troponin-I and vimentin as described in Section 2.13 to assess

the proportions of CMs and non-myocytes. Cell proportions were compared to the initial

composition of the cell suspension.

19

2.10 Sample fixing and immunofluorescent staining

To fix input and output cells from microdevices or pump parameter testing for

characterization the cells were suspended in CM medium or PBS and 15-30 x104 cells

were cytospun onto glass slides (Thermo, Shandon Cytospin 4) at 1000rpm for 5 minutes.

Slides were incubated with PFA for 20 minutes, rinsed with PBS and stored immersed in

PBS at 4°C for immunostaining.

Immunostaining agents were employed for the detection and identification of cells on the

surfaces. Monoclonal rabbit anti-cardiac troponin I antibody (TnI, Chemicon, AB1627,

1:100 DF) was used to stain cells with cardiomyocyte phenotype. Non-myocytes were

stained with Cy-5 conjugated monoclonal mouse anti-vimentin antibody (Vim, Sigma, C-

9080, 1:75 DF), with both TnI and Vim antibodies being used simultaneously to stain the

cells for assessment of relative quantities of cells. Primary antibodies made in rabbits

were stained with goat anti-rabbit FITC-conjugated secondary antibody (Vector, FI1000,

1:100). Primary antibodies made in mice were stained with goat anti-mouse FITC-

conjugated secondary antibody (Sigma, F9006, 1:64 DF). Cell nuclei were counterstained

using 4’,6-diamidino-2-phenylindole (DAPI, Sigma, D8417, 3mM stock, 1:100 DF).

20

3. Characterizing the syringe cell output for flow experiments in microdevices

3.1 Effects of pump angle and flow rate on cell output and composition

As seen in Figure 4 (B, D) the effects of pump flow rate were much less significant than

the contribution of pump tilt angle (Figure 4A, C) to cell output and composition. At a

fixed tilt angle at maximum only ~50% of the original cells were detected in the cells

exiting the tubing. In comparison, fixing the flow rate and changing the tilt angle had a

more significant effect on the cells exiting the tubing. The greatest number of cells was

detected when the pump was situated vertically (at a 90° angle), with close to 65% of the

input cells exiting. A 45° tilt angle was the worst case, with only 30% of the cells exiting

while a 0° angle produced intermediate results. The cell composition also showed great

consistency with increased non-myocyte and decreased CM percentages at all flow rates

and tilt angles compared to the input cell suspension as determined from immunostaining

for TnI and Vim. CM ratios decreased to ~30% as compared to the input, and the non-

myocytes increasing slightly to ~55% compared to the original ~40%.

It is possible then that one of the major operation parameters for microfluidic devices

may be the pump tilt angle. At angles other than 90° cell depletion via non-specific

adhesion to cell contact surfaces (the syringe and tubing interiors) prior to entering the

inlet drastically reduces the number of cells entering the devices. At all angles significant

changes to the input cell composition were observed. It should be noted however that the

90° angle, 10µL/min condition was common to both the tilt angle and flow rate tests,

with the cell output percentage being significantly different in each case (64% vs. 49%

respectively, Figure 4 A&B). As the output percentages were averaged over tests run at

different times and with different batches of cells, this data reflects the possibility of large

variability in results and indicates that these findings should be interpreted with caution.

In addition to non-specific adhesion, cell settling in the syringes during experimental runs

could have played a significant role in the reduction of the cell output, especially in CM

removal as these are the largest (and thus the heaviest) cells in the suspension. This effect

21

should be accounted for in the setup of flow experiments and steps taken to reduce cell

loss via this route [48].

Figure 4: Effects of pump tilt angle and flow rate on total cell output and cell suspension composition

by staining for TnI and Vim. Percentage of total input cells remaining in output at A) fixed 10µL/min

flow rate and pump tilt angles of 0°, 45° and 90° B) 90° fixed tilt angle and flow rates of 10µL/min,

20µL/min and 30µL/min. Percentage of collected cells that are CMs (TnI+) and non-myocytes (Vim+)

compared to input composition at C) fixed 10µL/min flow rate and pump tilt angles of 0°, 45° and 90° B)

90° fixed tilt angle and flow rates of 10µL/min, 20µL/min and 30µL/min. Proportions characterized by

immunostaining of cells cytospun onto slides and fixed. Averages ± standard error. * significant compared

to input for TnI and Vim (p < 0.001). Solid lines show significant difference between groups (p = 0.004).

3.2 Conclusion

From our tests we can cautiously determine that a more effective setup would set syringes

of cell suspension at a 90° angle during cell injection. Variation of the tilt angle and flow

rate did not improve the cell composition ‘drift’ seen, as the proportion of CMs was

decreased, and the proportion of non-myocytes increased in each condition as compared

to the input cell suspension. Further improvements may incorporate continuous agitation

of the cell suspension to reduce settling of cells during the course of the experiment but

these methods are not explored here.

01020304050607080

0° 45° 90°

% t

ota

l ce

lls o

utp

ut

Tilt angle

% cells output

01020304050607080

10µL/min 20µL/min 30µL/min

% c

ells

ou

tpu

t

Flow rate

% cells outputB A

C D

10µL/min, variable angle 90° angle, variable flow rate

22

4. Selective adhesion of primary rat cardiac cells to peptide-modified sodium

alginate

This chapter is similar in content to a manuscript submitted to Lab on a Chip, authored by

Brian D. Plouffe, Melissa A. Brown, Rohin K. Iyer, Milica Radisic and Shashi K.

Murthy.

4.1 Introduction

Tissue and stem cell engineering are potential methods of tissue regeneration and

repair that have seen major research efforts in recent years, particularly with the

discovery of rare stem and progenitor cell populations in tissue niches [17, 49]. The

isolation of these rare cell subpopulations is a challenge, particularly if the separation

is to be carried out in clinical point-of-care settings. The ability of microfluidic

devices to isolate rare cells was recently demonstrated by Nagrath et al.[50] with

antibody-coated micropillars that captured circulating tumor cells from whole blood.

Adhesion-based cell separation within microfluidic devices can be carried out by

either positive selection, as demonstrated by Nagrath et al, or by negative selection

[51]. Negative selection is an attractive approach when markers for target cells are not

fully known but its implementation poses several challenges. Since this approach

requires 100% removal of the non-target cells with minimal target cell capture,

extensive characterization of the adhesion profiles of all cell types in the

heterogeneous suspension is needed. Positive selection, by contrast, involves direct

capture of the target cell type, and is therefore easier to implement. However, a major

limitation of this approach is the lack of techniques to recover cells with minimal

perturbation from the microfluidic device following isolation. In the aforementioned

study, Nagrath et al. lysed the captured cells on chip in order to extract genomic and

proteomic information. On-chip cell lysis works well for devices designed for

diagnostic applications, but is not relevant for applications requiring recovery of

viable and functional cells.

23

Few studies have focused on the detachment of cells after capture in microfluidic devices.

The detachment of cells adhered to any surface requires the application of a force that is

greater in magnitude to that of adhesion. Fluid shear forces have been shown to be the

simplest method for cell detachment [52-54]. Although this is a local and simple method

of cell release, excessive exposure to fluid shear may result in cell damage and reduction

in viability. An alternative approach is to cleave the protein ligand that is bound to the

capture surface using enzymes, such as trypsin. However enzymatic exposure can cause

morphological changes due to a disruption of the cell membrane and glycocalyx, leading

to losses in cellular activity. Furthermore, enzymatic digestion has been shown to

directly affect both the behavior and chemical makeup of the cells themselves [55, 56].

These limitations illustrate the need to establish a general technique to capture and release

cells in micro-scale devices without extensive physical or chemical perturbations to the

cell environment.

Several recent publications in the literature have described the design of surface

coatings that can facilitate cell detachment when an external stimulus is applied, such

as an electrical potential or a small temperature change [57-60]. An example of the

former is a surface coating that consists of ligands bound to the surface via an

electroactive chemical functional group [58]. The electroactive quinoine ester

undergoes a chemical change to lactone upon applying an electrical potential. This

approach requires electrode incorporation into the capture device and careful

optimization of release parameters. The use of a thermally-responsive polymer, such

as poly(N-isopropylacrylamide), which is hydrophobic at 37°C and hydrophilic below

32°C, is another recently-described approach [39, 57, 60]. The hydrophobic surface is

adhesive to cells and its transformation results in nearly-complete cell release. The

shortcomings of this method are the lack of adhesion specificity in a flow regime and

potential adverse effects of lowering the temperature below the physiological

temperature of 37°C.

An alginate-RGDS complex prepared as described in Section 2.3 was adsorbed on the

surface of the device and converted into a thin hydrogel layer by means of a rinse

24

with Ca2+

ionic solution. Primary rat cardiac fibroblasts were injected into the device

and subsequently captured from the flow stream by the peptide-functionalized gel.

Following capture, the hydrogel was dissolved using EDTA, a strong ion chelator,

thereby releasing the captured cells. This method is simple and amendable to all

micro-scale devices that operate in a low shear stress regime. In addition, the

carbodiimide conjugation chemistry utilized to functionalize alginic acid can be

applied to a wide range of molecules containing primary amines. Furthermore, cell

release requires no external forces such as heating/cooling, electrical potential, or

irradiation of any kind.

Previous studies [33, 61] have illustrated the utility of using surface-immobilized

tetrapeptides to achieve cell capture within microfluidic devices. Specifically, the

arg-gly-asp-ser (RGDS) tetrapeptide has been shown to provide a high degree of

fibroblast adhesion within microchannels [33]. In the present work, this tetrapeptide

was conjugated to sodium alginate using carbodiimide chemistry with a N-

hydroxysulfosucciminide ester (sulfo-NHS) stabilizer [62]. The amount of RGDS

bound to alginic acid using this protocol was then quantified using Fourier Transform

Infrared Spectroscopy (FT-IR) to be 38 µg of RGDS per milligram of solid modified

alginate stock. FT-IR analysis was performed in a collaborative effort at the Murthy

laboratory.

25

Figure 5: FT-IR quantification of peptide content of alginate-RGDS. RGDS content in the alginate-

RGDS was quantified using FT-IR and a three point calibration. Calibration and quantification performed

at the Murthy laboratory in a collaborative effort.

4.2 EDTA dissolution of alginate

Trypsin was chosen as a positive control for cell retrieval as it is one of the standard

enzymatic cell removal treatments used for cell culture. We would like to replicate the

effectiveness of trypsin while avoiding the negative effects of an enzymatic treatment on

the cells previously described.

As seen in Figure 6 below cell retrieval remained constant (≤ 60% within the margin

of error in the controls) at all EDTA concentrations. When EDTA incubation was

followed by addition of CM medium however, cell retrieval increased to 82 ± 8 %,

which was comparable to the trypsin control of 84 ± 5 %. Cell viability remained

high in all treatment conditions, averaging approximately 80%. Addition of CM

medium appeared to have a beneficial effect on the cell retrieval process, likely due to

the quenching of EDTA and introduction of serum proteins to protect the cells.

Hence, an EDTA concentration of 50 mM accompanied by a culture medium rinse

was used for all subsequent cell detachment studies.

0

0.002

0.004

0.006

0.008

0.01

630 635 640 645 650 655

100 mg mL-1

50 mg mL-1

25 mg mL-1

Alginate-RDGS

Ad

sorb

an

ce

[a.u

]

Wavenumber [cm-1

]

26

Figure 6: Cell viability and retrieval with EDTA and trypsin treatment .Percentages of cell viability and total

cells retrieved for preplate cells treated with EDTA concentrations 0.5mM-50mM (unquenched) EDTA and

50mM EDTA (quenched with CM medium), compared to PBS and trypsin-EDTA-treated controls. Results show

that 50mM EDTA in cell media resulting in comparable release to that of the positive control, trypsin. In addition

it should be noted that all experiments maintain a viability above 75 %. Averages ± standard error with n=4. No

statistically significant differences were determined by 1-way ANOVA analysis.

4.3 Adhesion response to peptide modification and seeding density

Prior to the microfluidic cell adhesion experiments, we first conducted testing of cell

attachment and detachment in static culture at the Radisic laboratory. For this study we

needed to demonstrate selective adhesion of cells to the peptide-modified alginate as

compared to the non-adhesive unmodified alginate control as well as determine the

relationship between cell seeding density and cell adherence on each surface type.

For the preplate cell mixture (comprising mainly cardiac fibroblasts), the cell adhesion

response was characterized with alginate-RGDS surfaces. We expected that increasing

the number of cells seeded would correspondingly increase the total number of cells

adhering to each surface. Furthermore, we may postulate that at some point, the ‘dose-

response’ relationship would taper off into a plateau region as the maximum number of

cells able to interact directly with the surface-accessible peptide groups would have been

0

10

20

30

40

50

60

70

80

90

100

PBS 0.5mM EDTA

5mM EDTA 10mM EDTA

25mM EDTA

50mM EDTA

50mM EDTA, CM medium

Trypsin

% retrieval % viability

27

surpassed. Excess cells would remain suspended within the medium and be removed

during the washing steps.

As shown in Figure 7 below, each surface type demonstrated a statistically significant

increase in cell adhesion as the cell seeding density increased. Importantly, alginate-

RGDS surfaces had significantly increased cell adhesion at every concentration when

compared to unmodified alginate (as determined by Tukey’s test with one-way

ANOVA analysis , p<0.001), whereas bare glass supported the highest cell

attachment, in agreement with literature [63].

Figure 7: Average adhered cells per coverslip, adhesion on glass vs. on alginate-RGDS and

unmodified alginate surfaces. Results are shown for three seeding densities of 10 000, 50 000, 100 000

and 500 000 cells per coverslip. Solid lines compare statistically significant differences within each

seeding density. Dashed lines compare statistically significant differences between groups for each surface

treatment. Statistical significance between groups analysed by 1-way and 2-way ANOVA.

A prime objective of this study was to utilize the peptide-modified alginate to first

capture cells of interest on a coated surface and to then retrieve the cells by initiating

dissolution and solubilizing the coating layer. The key requirement for this process would

be maximal removal of cells from the surface layers with the highest possible viability

maintained. Successful cell recovery is pivotal in extending the application of this

28

technique to the clinical setting, especially in the isolation of desirable cells from a tissue

biopsy. To this end we have tracked the cells throughout the process to assess the

efficiency of our process at releasing cells with removal of the surface coating.

From analysis of brightfield images of sample surfaces (taken before and after the EDTA

releasing treatment) the percentage release of adhered cells from coverslips was

determined. The alginate-RGDS group showed close to 100% cell release after EDTA

treatment with similar results found with the unmodified alginate. Bare glass surfaces

showed a decrease in cell release with an increase in cell seeding density. For this group,

the conditions in the releasing step were insufficient to remove all the adhered cells from

the glass coverslips as compared to an enzymatic method such as trypsin.

Figure 8: Percentage of cells released from sample surfaces with EDTA treatment based on image

analysis. The average number of cells adhered to coverslip surface coatings calculated from brightfield

images taken after EDTA treatment was compared to the pre-treatment average and the percentage of the

adhered cells released determined. All samples Averages ± standard error. No statistically significant

differences were determined between groups by 1-way ANOVA analysis.

4.4 Overall cell recovery

Figure 9 below shows the percentage of the total seeded cells that were retrieved over the

course of each experiment (washes and EDTA release) and accounted for in the washing

solutions. The cells in suspension were counted in addition to counting the cells

0

10

20

30

40

50

60

70

80

90

100

alginate, 50,000

alg-RGDS, 50,000

glass, 50,000

alginate, 100,000

alg-RGDS, 100,000

glass, 100,000

alginate, 500,000

alg-RGDS, 500,000

glass, 500,000

% c

ell

rele

ase

fro

m c

ove

rslip

s

50,000 100,000 500,000

29

remaining on the coverslips (Figure 8) in order to close the material balance for the cells.

Overall approximately 60% of the cells were accounted for, suggesting that some cell

loss occurs during processing. This cell loss may arise during the numerous pipetting and

transferring steps, or may be due to dead or damaged cells apoptosing over the course of

Figure 9: Percentage of total cells retrieved from the washing steps over the experiment as a

percentage of the initial seeded cell number. Calculation of total seeded cells retrieved over the course of

the experiment after collection at all stages. Approximately 70% of the initial 50 000 and 100 000 cells

were accounted for, with ~40% lost to non-specific adhesion to surfaces and apoptosis during processing

steps. Greater cell loss was seen in the 500 000 cell group, likely due to cell death from high cell crowding

and resource depletion during the serum-free incubation. Solid line shows statistically significant

differences between groups by 1-way ANOVA (p = 0.038).

Figure 10: Percentage of total cells retrieved over the experiment as a percentage of the initial seeded

cell number , with control PBS- solution in the release step. Calculation of total number of seeded cells

collected over the course of the experiment. 100 000 initial cells seeded in serum-free medium. Averages ±

standard error, n > 4. No statistically significant differences observed between groups by 1-way ANOVA

analysis.

0102030405060708090

100

alginate alg-RGDS glass

% c

ells

re

trie

ved

% PBS retrieval

30

the experiment and being removed with the supernatant after centrifuging. Lower cell

concentrations averaged ~70% retrieval. However high cell loss was seen in the 500 000

cell group (with only ~45% retrieval), and was likely as a result of cell death from high

cell crowding and resource depletion during the serum-free incubation. In order to

discount the possibility that the EDTA may be negatively affecting the cells, a control

study was performed with 100 000 cells on each coverslip and substituting PBS- for

EDTA during the releasing step. Since some of the calcium ions stabilizing the hydrogel

matrix are washed away on addition of PBS, the top surface of the hydrogel is likely to

begin to go into solution, releasing cells from the surface. In comparison to the cell

retrieval when EDTA was used (Figure 9, 100 000 concentration), there appeared to be

no improvement in the number of cells retrieved in the PBS case for the alginate-RGDS

group. Cell retrieval from alginate and glass surfaces was similarly unchanged within the

margin of error. It may thus be safe to say that the cell loss was not due to EDTA-

mediated cell damage but instead likely resulted during processing and incubation.

4.5 Cell viability in static experiments

Input cells were estimated to be > 98% viable. Wash 1 cells (with non-adhered cells

retrieved after seeding and incubation) averaged approximately 70% viability, which was

fairly constant over all seeding densities. This lowered viability was likely due to the lack

of serum during the first 30 minute incubation. The cells collected in the second wash,

however, had lower viability (~50% overall). Any remaining cells which were

unattached, damaged or dead but were not removed in the first washing step likely

remained on the surfaces. These cells did not adhere or recover during the 1 hour

incubation and were removed in the second wash. After the releasing step, however, the

cell viabilities were improved, matching Wash 1 values. Viabilities were approximately

70%, with a few being in excess of 80%. EDTA release step viabilities at the lowest

seeding concentration was highly variable, mainly due to low total cell attachment and

thus release. As a result, any dead cells would represent a large change in viability. These

results are significant as they represent the critical point in the procedure; retrieval of the

31

captured cells. From this we can conclude that we are indeed able to recover our cells of

interest with high viabilities to keep the cells fit for culture.

4.6 Cell capture and release in microfluidic devices

Protein conjugation to alginate has been well established in research involving cell

encapsulation, cell transplantation, and tissue engineering applications [40, 64-66]. By

applying these same concepts to microfluidic devices in conjunction with the ability of

EDTA to chelate Ca2+

ions [67], capture and release of cells can be achieved. Thus far

we have demonstrated cell adhesion to RGDS-functionalized alginate in static conditions.

Of great importance is the testing of this alginate-peptide system in flow conditions as

our targeted application of the alginate-peptide is cell capture in microfluidic devices. To

Figure 11: Percentage viability of cells collected at

Wash 1, Wash 2 and EDTA-release time points.

Cells collected from seeding density groups: a) 50

000 cells per coverslip, b) 100,000 cells per coverslip

and c) 500,000cells per coverslip. Viability

determined via cell counts with trypan blue stain.

Averages ± Std error. Solid lines show statistically

significant differences between groups by 1-way

ANOVA (p < 0.001).

B

C

A

32

this end, straight microchannels were prepared with an alginate-RGDS surface coating as

described in Section 2.6.2. Adhesion and EDTA release characteristics of primary cardiac

fibroblasts seeded onto these coated channels were determined and compared to

unmodified alginate control coatings.

As shown in Figure 12(a), conjugation of RGDS to the alginate backbone showed a two-

fold increase in fibroblast adhesion compared to unconjugated alginate; the unconjugated

alginate controls showed 23 ± 1 cells/mm, whereas 57 ± 1 cells/mm adhered for peptide-

conjugated alginate hydrogels. Rinsing with a 50 mM EDTA solution resulted in the

release of appreciable numbers of these adhered cells (Figure 12(a)). This outcome was

reaffirmed by brightfield microscopy as shown in Figure 12(b,c). These micrographs

illustrate that the cells are nearly all removed from the substrate coated with alginate-

RGDS upon rinsing with EDTA. Also of note is the difference in the cell release of

alginate (69 ± 7 %) with respect to alginate-RGDS (97 ± 3%). It is possible that the

chemical conjugation to the carboxylate groups reduces the number of hydrogen bonding

sites within the hydrogel making the peptide-functionalized alginate more easily soluble

by the EDTA chelator, thus releasing a greater number of cells. The cell adhesion values

are slightly lower than that of the static studies, but this may be a result of the short

incubation time of EDTA with alginate hydrogel. Live/dead assays indicated no

significant difference in cell viabilty after release versus pre-injection.

4.7 Conclusion

In summary, this work demonstrates the application of peptide-functionalized alginate

hydrogels as a method of capture and release of cells in both static culture and

microfluidic channels. We have demonstrated specific adhesion, a concentration

dependent adhesion response and maintenance of cell viability throughout the

isolation procedure in the static culture case. When combined with the capability of

microfluidic devices to selectively capture cells of a particular type from

heterogeneous suspensions, the cell release methodology described herein could

provide viable, functional cells for further application (such as culturing on scaffolds

for tissue engineering) or analysis as part of disease diagnostics.

33

Figure 12: Cell attachment and detachment of cardiac fibroblast at a shear stress of 1 dyn/cm2 in

microfluidic channels (a) coated with native alginate gel or RGDS modified alginate gel. Error bars

denote standard errors for five replicates of each experiment. Representative region of an alginate-RGDS

coated channel (b) pre-EDTA rinse and (c) post-EDTA rinse. The channel width (vertical dimension) is 1

mm.

In further work characterization of the effects of flow conditions on the specificity and

capture dynamics in this system would need to be conducted. Optimization of the system

for capture of endothelial and smooth muscle cells (the next largest populations in the

non-myocyte mixture) using REDV and VAPG peptides could prove beneficial,

especially to vascular tissue engineering research.

(a)

(c) (b)

0

10

20

30

40

50

60

Alginate Alginate-RGDS

Captured

Released

Nu

mb

er

of

Cel

ls [

dy

n m

m-2

]

Surface Coating

Remaining

[ m

m -1

]

34

5. Characterization of alginate functionalized with alternative peptides REDV

and VAPG

5.1 Adhesion response of cell lines to alginate-peptide

With the success of alginate-RGDS at capturing primary cardiac fibroblasts it would be

beneficial to explore alternate peptide sequences to allow for selective isolation of

endothelial and smooth muscle cells; two very important cell types for tissue

vascularization contained in the non-myocyte populations. Peptides REDV and VAPG,

selected for their specificity to endothelial cells and smooth muscle cells respectively,

were incorporated into alginate as described previously. Concentrations of 0.6, 1.0 and

0.8 mg/mL for RGDS, REDV and VAPG respectively were used and glass coverslips

coated and gelled with 100mM CaCl2.

In the non-myocyte mixture used to characterize the alginate-RGDS adhesion fibroblasts

dominate the mixture, with endothelial and smooth muscle cells comprising a minority

(making up 2-3% and 3-4% of the native heart respectively) [34]. As such, initial

experiments were performed with cell lines to determine the interaction between the cells

and modified surfaces with more robust and homogeneous cell populations. Mouse

endothelial (H5V), rat aortic smooth muscle (A7r5), and mouse fibroblast (NIH 3T3) cell

lines were used to characterize surface adhesion to glass, alginate and alginate-peptide

surfaces. H5V cells were expected to show preferential adhesion to REDV surfaces and

A7r5 to VAPG due to peptide specificity for receptors characteristic to each cell type [20,

36, 68]. Additionally, the RGDS peptide binds to a highly conserved sequence found in

all cell types.

In total 100 000 cells in serum-free medium were seeded on each alginate-peptide

surface, with uncoated glass and unmodified alginate used as positive and negative

adhesion controls respectively. Figure 13-Figure 15 depict the adhesion response of each

cell type. As expected, in all cases the unmodified glass control demonstrated the greatest

and unmodified alginate the least cell attachment within the margin of error. VAPG

35

surfaces significantly improved attachment of smooth muscle cells compared to the

alginate control. Similarly, fibroblast cell attachment was significantly increased in

alginate-RGDS samples. REDV peptide did not improve adhesion in endothelial cells

with the alginate-peptide concentrations tested. Furthermore, while RGDS and VAPG

peptides demonstrated improvements in specific capture of their targeted cell types, we

also note that non-targeted capture of smooth muscle cells and fibroblasts on REDV

occurred and that alginate-VAPG surfaces appeared to induce the greatest amount of cell

adhesion overall. There may be a high amount of non-targeted binding on the VAPG-

modified surfaces and analysis of the total amount of peptide incorporated into the

alginate (in the REDV and VAPG cases) and the surface characteristics should be

performed. Lastly, as compared to both the bare glass positive control and adhesion of

cardiac fibroblasts to alginate-RGDS, overall adhesion on peptide-modified alginate was

low. These results were therefore unclear as to the relationship between cell type and

specificity of the cell adhesion using the cell lines.

Figure 13: H5V endothelial cell line adhesion to peptide-modified alginate. Total number of cells

adhered per coverslip as determined from cell counts from brightfield images of samples and averaged over

n ≥ 3. 100 000 cells initially seeded per sample in serum-free medium. Alginate-peptides synthesized at

concentrations of 0.6mg/ml (RGDS), 1.0mg/ml (REDV) and 0.8mg/ml (VAPG). Averages ± standard error.

Solid lines show statistically significant differences between groups as determined by 1-way ANOVA

analysis. * Results significant compared to unmodified alginate. • significant compared to bare glass group.

36

Figure 14: A7r5 smooth muscle cell line adhesion to peptide-modified alginate. Total number of cells

adhered per coverslip as determined from cell counts from brightfield images of samples and averaged over

n ≥ 3. 100 000 cells initially seeded per sample in serum-free medium. Alginate-peptides synthesized at

concentrations of 0.6mg/ml (RGDS), 1.0mg/ml (REDV) and 0.8mg/ml (VAPG). Averages ± standard error.

Solid lines show statistically significant differences between groups as determined by 1-way ANOVA

analysis. * Results significant compared to unmodified alginate. • significant compared to bare glass group.

Figure 15: NIH 3T3 fibroblast cell line adhesion to peptide-modified alginate. Total number of cells

adhered per coverslip as determined from cell counts from brightfield images of samples and averaged over

n ≥ 3. 100 000 cells initially seeded per sample in serum-free medium. Alginate-peptides synthesized at

concentrations of 0.6mg/ml (RGDS), 1.0mg/ml (REDV) and 0.8mg/ml (VAPG). Averages ± standard error.

Solid lines show statistically significant differences between groups as determined by 1-way ANOVA

analysis. * Results significant compared to unmodified alginate. • significant compared to bare glass group.

37

A possible explanation for the adhesion results seen in the cell line experiments could

have been abnormal phenotypic expression of receptor groups by these cells. At high

passages cell lines have been known to exhibit aberrant phenotype, receptor expression

and cell responses [69-71], a factor which must be taken into account when viewing these

results as most of these initial experiments were conducted with cell lines at passages

greater than 15. In comparison, the primary preplate cells have only been trypsinized

twice (passage 2) and would be more likely to retain native receptor protein configuration

and expression.

5.2 Characterization of alternate peptides with primary non-myocytes

When non-myocyte cell mixture isolated as described in Section 2.1 was seeded onto

these surfaces, as shown in Figure 16 below a stastically significant increase in cell

adhesion was apparent on RGDS-modified alginate compared to unmodified alginate,

alg-REDV and alg-VAPG surfaces. The level of adhesion was comparable to that of bare

glass positive control samples. As the majority of the preplate cells are cardiac fibroblasts

(~ 49%, [34]) this result is in accordance with the expected adhesion specificity to

RGDS and was seen in the alginate-RGDS surface characterization in the previous

chapter. Capture of smooth muscle cells (the next largest population) by VAPG peptide

was marginally increased (but not to statistically significant levels) as compared to the

alginate negative control.

Figure 16: Cell adhesion response to peptide-modified alginate. Average number of cells adhered per

coverslip after spin-coating with alginate-peptides and incubation for 45 minutes in serum-free CM

medium. Samples compared to bare glass and unmodified alginate surfaces. Alginate-peptides synthesized

at 0.6mg/ml peptide concentration. Averages ± standard error. Solid lines show statistical significance

between groups from 1-way ANOVA analysis.

38

5.3 Conclusion

These preliminary studies describe a system where fibroblasts and smooth muscle cells

were selectively captured by our peptide-functionalized alginate surface coatings. Non-

targeted adhesion of cell lines may be as a result of phenotypic changes in cells at high

passage number and these experiments may be repeated in the future with this in mind.

The system has not yet been optimized, however, with detection of endothelial cell

adhesion remaining a problem and a serial depletion setup may be needed, similar to that

of Plouffe et al. [33]. Definitely, our approach has proved capable of being utilized with

the primary cardiac rat cell model and has the potential to be broadened for use with

human cells if suitable adhesion moieties are utilized. It remains to be seen whether our

eventual hope of stem cell isolation can be realized but the first steps have been taken

with relative success.

39

6 Summary, Future Work and Recommendations

6.1 Summary of results

Thus far our goal of grafting of the peptides into the alginate hydrogel for temporary cell

capture has produced positive results. Concentrating on RGDS-mediated removal of

cardiac fibroblasts (the largest fraction of the non-myocytes) we have demonstrated

increased adhesion of these cells to alginate-RGDS surface coatings. Using these cells we

were able to establish a concentration-dependent adhesion response with increased cell

seeding density, leading to greater cell adhesion at seeding concentrations of 10 000 - 500

000 cells per sample. Utilizing the cation-mediated phase change properties of the sodium

alginate system, the surface coatings were redissolved with EDTA and the cells collected.

From brightfield microscope imaging of the surfaces and image analysis, close to 100%

cell release was found in the coated surfaces with maintenance of cell viability

throughout the procedure. Some discrepancy exists between the number of cells seeded

onto the surfaces and the total cells collected by the end of all the steps in static

experiments. It is possible that cell loss occurs due to non-specific adhesion to plate and

pipette surfaces, or that cells damaged during processing are removed in the final

centrifuging step. This area requires further investigation and fine tuning of the reagents

and reaction conditions to optimize for maximum cell retrieval.

We have also demonstrated the applicability of the alginate-peptide system to

microfluidics by our creation of a removable surface layer inside microfluidic channels

on which selective capture of cardiac fibroblasts under flow conditions was achieved with

RGDS peptide. Removal of the alginate-RGDS layer using 50mM EDTA and a culture

medium quench resulted in collection of the viable adhered cells without the use of

enzymatic agents such as trypsin. It is apparent that under the right conditions this

technique can be used to realize our goal of achieving a capture-release surface coating

for microfluidics and disease diagnostic systems.

For selective capture of endothelial and smooth muscle cells, the peptides REDV and

VAPG were similarly grafted onto sodium alginate and the system characterized with cell

40

lines as well as the non-myocyte cell suspension. The overall inconclusive adhesion

responses seen by NIH 3T3 fibroblasts, A7r5 smooth muscle and H5V endothelial cell

lines may have in part originated with phenotypic changes in cell surface receptor

expression over the course of time spent in in vitro culture. Cells at earlier passages

should therefore be used in any future experiments. Testing of the system with the non-

myocyte cell mixture supported the RGDS-mediated adhesion improvement in fibroblast

adhesion, with slight improvements in smooth muscle cell adhesion. It is also possible the

relative scarcity of endothelial and smooth muscle cells is obscuring any improvements in

cell attachment so a serial depletion method or selective removal of fibroblasts from the

suspension prior to testing the efficacy of REDV and VAPG may be necessary.

Our approach has showed relative success in the rat cardiac cell model and has the

potential to be broadened for use with human cells if suitable adhesion moieties can be

found. Importantly, we have characterized these devices for use with primary cells, a

rarity as cell lines are much more robust, have well-defined and predictable

characteristics and are highly consistent. In contrast, primary cells are more

heterogeneous, highly sensitive to culture conditions and handling and can vary widely in

phenotype, morphology and consistency, especially when cultured in vitro. Our research

attempts to bridge this gap, characterizing our system within the rat model from which it

is but a step to human cells. It remains to be seen whether our eventual goal of stem cell

isolation can be realized but the first steps have been taken with relative success and

further breakthroughs anticipated.

6.2 Future Work

The next logical step would be optimizing the alginate system for capture of endothelial

and smooth muscle cells, the next largest populations in the non-myocyte cell mixture.

For this, a reasonable starting point pointed to the use of REDV and VAPG to immobilize

cells on alginate surface coatings. For characterization the actual procedure may have to

be done in series, with ECs first adhering to REDV coatings before proceeding to VAPG

41

or RGDS depletion. This serial depletion method has been established to be effective in

the ideal cell line situation [20], and would be useful in providing a population enriched

in ECs or SMCs for use with the non-RGDS peptides. Retrieved cells must then be

characterized to determine whether the groups exhibit the necessary specificity in

capture. This can be achieved by directly immunostaining them, performing flow

cytometery on output cells or by culturing these cells for several days followed by fixing

and staining for identifying markers.

Further work with this system would extend the technique for use within microfluidic

channels. Characterization of the effects of flow on the specificity and capture dynamics

in this system have already begun in a collaborative effort with some success already

seen in the alginate-RGDS system using CM-depleted cardiac cells. Once static culture

studies have identified optimal synthesis, reagent and reaction conditions for cell release

and recovery, advancement can be made to the microfluidics approach for the REDV and

VAPG alginate-peptide schemes.

With success of the REDV- and VAPG-alginate systems it may be informative to turn to

the glass-functionalized Hele-Shaw devices used by Murthy et al. [29] to characterize

shear stress cell adhesion behaviour, this time using alginate-peptide enriched input cell

populations. This research would be valuable to improve the cell adhesion in alginate-

peptide devices as knowledge of optimal shear stresses for adhesion of the various cells

would aid in adjusting design and operating parameters within the microfluidics devices

to maximize cell capture.

6.3 Recommendations

1. Characterize the alginate-RGDS surface modification and cell retrieval strategy in

the flow condition in straight-channel microfluidic devices. Study the composition

and viability of output cell populations by flow cytometry and immunofluorescent

staining of cells plated and grown for several days. This would inform us as to the

42

separation capability of the devices as well as the fitness of the cells (by studying

their development).

2. Characterize the alginate-REDV and alginate-VAPG systems in static culture and

determine optimal conditions and limitations of the system in terms of total cell

adhesion, effect of seeding density on adhesion, and capture and release

efficiencies. Captured cells are to be typified with flow cytometry or

immunofluorescent staining to determine composition.

3. Continue investigation into the alginate-REDV and alginate-VAPG systems in the

flow case once static culture investigation has been completed. Characterization

of the cells is to be done as for the alginate-RGDS case described in

recommendation #1.

4. Upon completion of functional devices for cell capture and retrieval, return to the

variable shear stress devices. Using the alginate-peptide enriched cell outputs

determine optimal shear stresses for adhesion. Apply this new information to the

straight-channel constant shear stress alginate-peptide microdevices and

determine whether improvements are evident.

5. Utilizing a multistage separation process on the optimized devices serially deplete

the heterogeneous cell mixture using the alginate-peptide systems. Adherent, non-

adherent and final output cells should be sampled and the component cell

populations identified to assess the functionality of the system.

6. Culture the final output from the multistage process. Determine the composition

of the growing cells at various stages. The cells can be assessed for various known

stem cell markers such as isl-1, c-kit or oct4.

43

7. Multistage device output cells can be lysed and gene sequencing performed to

identify any unusual highly expressed protein or cellular marker in this purified

population which may indicate progenitor cell phenotype.

44

7 REFERENCES

1. Rosamond, W., et al., Heart disease and stroke statistics--2008 update: a report

from the American Heart Association Statistics Committee and Stroke Statistics

Subcommittee. Circulation, 2008. 117(4): p. e25-146.

2. Zimmermann, W.H., I. Melnychenko, and T. Eschenhagen, Engineered heart

tissue for regeneration of diseased hearts. Biomaterials, 2004. 25(9): p. 1639-47.

3. Smits, A.M., et al., The role of stem cells in cardiac regeneration. J Cell Mol

Med, 2005. 9(1): p. 25-36.

4. Jawad, H., et al., Myocardial tissue engineering. Br Med Bull, 2008. 87: p. 31-47.

5. Eschenhagen, T. and W.H. Zimmermann, Engineering myocardial tissue. Circ

Res, 2005. 97(12): p. 1220-31.

6. Ott, H.C., et al., Perfusion-decellularized matrix: using nature's platform to

engineer a bioartificial heart. Nat Med, 2008. 14(2): p. 213-21.

7. Radisic, M., et al., Medium perfusion enables engineering of compact and

contractile cardiac tissue. Am J Physiol Heart Circ Physiol, 2004. 286(2): p.

H507-16.

8. Hassink, R.J., et al., Stem cell therapy for ischemic heart disease. Trends Mol

Med, 2003. 9(10): p. 436-41.

9. Kehat, I., et al., High-resolution electrophysiological assessment of human

embryonic stem cell-derived cardiomyocytes: a novel in vitro model for the study

of conduction. Circ Res, 2002. 91(8): p. 659-61.

10. Gepstein, L., Derivation and potential applications of human embryonic stem

cells. Circ Res, 2002. 91(10): p. 866-76.

11. Heng, B.C., et al., Strategies for directing the differentiation of stem cells into the

cardiomyogenic lineage in vitro. Cardiovasc Res, 2004. 62(1): p. 34-42.

12. Beltrami, A.P., et al., Adult cardiac stem cells are multipotent and support

myocardial regeneration. Cell, 2003. 114(6): p. 763-76.

13. Pal, R., Embryonic stem (ES) cell-derived cardiomyocytes: A good candidate for

cell therapy applications. Cell Biol Int, 2008.

14. Dawson, L., et al., Safety issues in cell-based intervention trials. Fertil Steril,

2003. 80(5): p. 1077-85.

15. Garfein, E.S., D.P. Orgill, and J.J. Pribaz, Clinical applications of tissue

engineered constructs. Clin Plast Surg, 2003. 30(4): p. 485-98.

16. Zimmermann, W.H. and T. Eschenhagen, Embryonic stem cells for cardiac

muscle engineering. Trends Cardiovasc Med, 2007. 17(4): p. 134-40.

17. Laugwitz, K.L., et al., Postnatal isl1+ cardioblasts enter fully differentiated

cardiomyocyte lineages. Nature, 2005. 433(7026): p. 647-53.

18. Barile, L., et al., Endogenous cardiac stem cells. Prog Cardiovasc Dis, 2007.

50(1): p. 31-48.

19. Wang, X., et al., The role of the sca-1+/CD31- cardiac progenitor cell population

in postinfarction left ventricular remodeling. Stem Cells, 2006. 24(7): p. 1779-88.

20. Plouffe, B.D., et al., Peptide-mediated selective adhesion of smooth muscle and

endothelial cells in microfluidic shear flow. Langmuir, 2007. 23(9): p. 5050-5.

21. Pappas, D. and K. Wang, Cellular separations: a review of new challenges in

analytical chemistry. Anal Chim Acta, 2007. 601(1): p. 26-35.

45

22. Murthy, S.K., et al., Effect of flow and surface conditions on human lymphocyte

isolation using microfluidic chambers. Langmuir, 2004. 20(26): p. 11649-11655.

23. Wang, K., B. Cometti, and D. Pappas, Isolation and counting of multiple cell

types using an affinity separation device. Anal Chim Acta, 2007. 601(1): p. 1-9.

24. Lu, X., Ultrasound-induced cell detachment and gene transfection in adherent

cells. Acoustic Research Letters Online, 2005. 6(3): p. 195-200.

25. Zhu, H., J. Yan, and A. Revzin, Catch and release cell sorting: electrochemical

desorption of T-cells from antibody-modified microelectrodes. Colloids Surf B

Biointerfaces, 2008. 64(2): p. 260-8.

26. Eschenhagen, T., et al., Cardiac tissue engineering. Transpl Immunol, 2002. 9(2-

4): p. 315-21.

27. Leor, J., et al., Bioengineerred cardiac grafts: A new approach to repair the

infarcted myocardium? Circulation, 2000. 102(suppl III): p. III56-III61.

28. Ji, H.M., et al., Silicon-based microfilters for whole blood cell separation.

Biomed Microdevices, 2008. 10(2): p. 251-7.

29. Murthy, S.K., et al., Size-based microfluidic enrichment of neonatal rat cardiac

cell populations. Biomed Microdevices, 2006. 8(3): p. 231-7.

30. Yamada, M., et al., Microfluidic devices for size-dependent separation of liver

cells. Biomed Microdevices, 2007. 9(5): p. 637-45.

31. Huang, L.R., et al., Continuous particle separation through deterministic lateral

displacement. Science, 2004. 304(5673): p. 987-990.

32. Sethu, P., A. Sin, and M. Toner, Microfluidic Diffusive Filter for Apheresis

(Leukapheresis). submitted.

33. Plouffe, B.D., M. Radisic, and S.K. Murthy, Microfluidic depletion of endothelial

cells, smooth muscle cells, and fibroblasts from heterogeneous suspensions. Lab

Chip, 2008. 8(3): p. 462-72.

34. Naito, H., et al., Optimizing engineered heart tissue for therapeutic applications

as surrogate heart muscle. Circulation, 2006. 114(1 Suppl): p. I72-8.

35. Mann, B.K. and J.L. West, Cell adhesion peptides alter smooth muscle cell

adhesion, proliferation, migration, and matrix protein synthesis on modified

surfaces and in polymer scaffolds. J Biomed Mater Res, 2002. 60(1): p. 86-93.

36. Massia, S.P. and J.A. Hubbell, Vascular endothelial cell adhesion and spreading

promoted by the peptide REDV of the IIICS region of plasma fibronectin is

mediated by integrin alpha 4 beta 1. J Biol Chem, 1992. 267(20): p. 14019-26.

37. Iuliano, D.J., S.S. Saavedra, and G.A. Truskey, Effect of the conformation and

orientation of adsorbed fibronectin on endothelial cell spreading and the strength

of adhesion. J Biomed Mater Res, 1993. 27(8): p. 1103-13.

38. Shimizu, T., et al., Cell sheet engineering for myocardial tissue reconstruction.

Biomaterials, 2003. 24(13): p. 2309-16.

39. Ernst, O., et al., Control of cell detachment in a microfluidic device using a

thermo-responsive copolymer on a gold substrate. Lab Chip, 2007. 7(10): p.

1322-9.

40. Rowley, J.A., G. Madlambayan, and D.J. Mooney, Alginate hydrogels as

synthetic extracellular matrix materials. Biomaterials, 1999. 20(1): p. 45-53.

41. Wang, L., et al., Evaluation of sodium alginate for bone marrow cell tissue

engineering. Biomaterials, 2003. 24(20): p. 3475-81.

46

42. Landa, N., et al., Effect of injectable alginate implant on cardiac remodeling and

function after recent and old infarcts in rat. Circulation, 2008. 117(11): p. 1388-

96.

43. Yu, J., et al., The effect of injected RGD modified alginate on angiogenesis and

left ventricular function in a chronic rat infarct model. Biomaterials, 2009. 30(5):

p. 751-6.

44. Usami, S., et al., Design and construction of a linear shear stress flow chamber.

Ann Biomed Eng, 1993. 21(1): p. 77-83.

45. Radisic, M., et al., Functional assembly of engineered myocardium by electrical

stimulation of cardiac myocytes cultured on scaffolds. Proceedings of the National

Academy of Sciences of the United States of America, 2004. 101(52): p. 18129-

18134.

46. Radisic, M., et al., Biomimetic approach to cardiac tissue engineering: oxygen

carriers and channeled scaffolds. Tissue Eng, 2006. 12(8): p. 2077-91.

47. Xia, Y.N. and G.M. Whitesides, Soft lithography. Angewandte Chemie-

International Edition, 1998. 37(5): p. 551-575.

48. Cooper, R. Chips & Tips: Preventing suspension settling during injection. Lab on

a Chip - Chips & tips 2007 21 August 2007 [cited 2009 February 3]; 1st:[Forum

entry]. Available from:

http://www.rsc.org/Publishing/Journals/lc/Chips_and_Tips/suspension_injection.asp.

49. Oshima, H., et al., Morphogenesis and renewal of hair follicles from adult

multipotent stem cells. Cell, 2001. 104(2): p. 233-245.

50. Nagrath, S., et al., Isolation of rare circulating tumour cells in cancer patients by

microchip technology. Nature, 2007. 450(7173): p. 1235-1239.

51. Plouffe, B.D., M. Radisic, and S.K. Murthy, Microfluidic depletion of endothelial

cells, smooth muscle cells, and fibroblasts from heterogeneous suspensions. Lab

on a Chip, 2008. 8(3): p. 462-472.

52. Lu, H., et al., Microfluidic shear devices for quantitative analysis of cell adhesion.

Analytical Chemistry, 2004. 76(18): p. 5257-5264.

53. Wankhede, S.P., et al., Cell detachment model for an antibody-based microfluidic

cancer screening system. Biotechnology Progress, 2006. 22(5): p. 1426-1433.

54. Zhang, X., P. Jones, and S.J. Haswell, Attachment and detachment of living cells

on modified microchannel surfaces in a microfluidic-based lab-on-a-chip system.

Chemical Engineering Journal, 2008. 135(Supplement 1): p. S82-S88.

55. Fujioka, N., et al., Difference in infrared spectra from cultured cells dependent on

cell-harvesting method. Applied Spectroscopy, 2003. 57(2): p. 241-243.

56. Jung, K., et al., Culture of Human Kidney Proximal Tubular Cells- the Effect of

Various Detachment Prodecures on Viability and Degree of Cell Detachment.

Cellular Physiology and Biochemistry, 1995. 5(5): p. 353-360.

57. Yamato, M., et al., Thermally responsive polymer-grafted surfaces facilitate

patterned cell seeding and co-culture. Biomaterials, 2002. 23(2): p. 561-567.

58. Yeo, W.S., C.D. Hodneland, and M. Mrksich, Electroactive monolayer substrates

that selectively release adherent cells. Chembiochem, 2001. 2(7-8): p. 590-593.

59. Zhu, H., J. Yan, and A. Revzin, Catch and release cell sorting: Electrochemical

desorption of T-cells from antibody-modified microelectrodes. Colloids and

Surfaces B-Biointerfaces, 2008. 64(2): p. 260-268.

47

60. Yamato, M., et al., Thermo-responsive culture dishes allow the intact harvest of

multilayered keratinocyte sheets without dispase by reducing temperature. Tissue

Engineering, 2001. 7(4): p. 473-480.

61. Plouffe, B.D., et al., Peptide-mediated selective adhesion of smooth muscle and

endothelial cells in microfluidic shear flow. Langmuir, 2007. 23(9): p. 5050-5055.

62. Hermanson, G.T., Bioconjugate Techniques. 1996, Boston, MA: Academic Press.

63. Morra, M. and C. Cassinelli, Surface studies on a model cell-resistant system.

Langmuir, 1999. 15(13): p. 4658-4663.

64. Alsberg, E., et al., Engineering growing tissues. Proceedings of the National

Academy of Sciences of the United States of America, 2002. 99(19): p. 12025-

12030.

65. Cheetham, P.S.J., K.W. Blunt, and C. Bocke, Physical Studies on Cell

Immobilization Using Calcium Alginate Gels. Biotechnology and Bioengineering,

1979. 21(12): p. 2155-2168.

66. Drury, J.L., T. Boontheeku, and D.J. Mooney, Cellular cross-linking of peptide

modified hydrogels. Journal of Biomechanical Engineering-Transactions of the

ASME, 2005. 127(2): p. 220-228.

67. Couperwhite, I. and M.F. McCallum, Influence of EDTA on Composition of

Alginate Synthesized by Azotobacter-Vinelandii. Archives of Microbiology, 1974.

97(1): p. 73-80.

68. Gobin, A.S. and J.L. West, Val-ala-pro-gly, an elastin-derived non-integrin

ligand: smooth muscle cell adhesion and specificity. J Biomed Mater Res A,

2003. 67(1): p. 255-9.

69. Ellison, B.J. and H. Rubin, Individual transforming events in long-term cell

culture of NIH 3T3 cells as products of epigenetic induction. Cancer Res, 1992.

52(3): p. 667-73.

70. Makino, I., et al., Phenotypic changes of adrenomedullin receptor components,

RAMP2, and CRLR mRNA expression in cultured rat vascular smooth muscle

cells. Biochem Biophys Res Commun, 2001. 288(3): p. 515-20.

71. Kato, S., et al., Characterization and phenotypic variation with passage number

of cultured human endometrial adenocarcinoma cells. Tissue Cell, 2008. 40(2): p.

95-102.

48

8. APPENDICES

S1. Characterization of isl-1+ cardiac progenitor cell capture in peptide-modified

microfluidics devices

S1.1.1 Hele-Shaw device theory

Pre-fabricated Hele-Shaw devices were shipped from the Murthy laboratory at

Northeastern University in Boston, MA, USA and stored at 4°C until used. These devices

were based on the linear shear stress equation derived by Usami et al. [44] (Equation 1)

and device design of Murthy et al. [22]. The construction is such that a linear gradient of

shear stress is present along the channel axis, characterized by the shear stress profile

seen in Figure S 1 below. The main purpose of this device is to determine a range of

shear stress range with maximal cell attachment to the various peptide-functionalized

surfaces and is a prelude to fabrication of straight-channel fixed shear stress peptide-

modified devices.

S1.1.2 GMBS functionalization of Hele-Shaw devices with peptides

The following details the surface modification of glass coverslips and fabricated

microfluidic devices for covalent attachment of peptides.

Prior to carrying out any reactions GMBS (N-[g-malemidobutyryloxy]-succinimide ester,

Pierce 22309) was dissolved to 100mg/ml in DMSO (Sigma, D2650). The peptides arg-

gly-asp-ser (RGDS, American Peptide 44-0-14), arg- glu-asp-val (REDV, American

Peptide 44-0-12) and val-ala-pro-gly (VAPG, Sigma V0883) were dissolved to 5mg/ml in

PBS. The rest of the reactions were carried out inside a chemical fume hood. A glove bag

(Sigma, Z530212) was filled with nitrogen gas and used to pipette 400μl of 3-

mercaptopropyltrimethoxysilane directly into 10ml of 200 proof ethanol, ensuring that

the silane did not contact air. Glass slides were exposed to oxygen plasma for 30 seconds

in a petri dish (Harrick Plasma, Plasmaflo, Ithaca, NY, USA), then removed from the

machine and immediately submerged in the silane/ethanol solution. The dishes were

covered and the surfaces were allowed to react for 30 minutes at room temperature. The

49

solution was removed and the surfaces rinsed in 200 proof ethanol. All silane wastes were

placed into specially labeled containers for proper chemical waste disposal. 28μl of the

GMBS solution was then added to 10ml of 200-proof ethanol and mixed. The mixture

was then pipetted to cover the glass surfaces and reacted for 15 minutes at room

temperature. Once the GMBS solution was removed the surfaces were washed, first with

200-proof ethanol and then PBS, and peptide solution at concentration 1mg/ml (in PBS)

was added. The surfaces were reacted for a final 30 minutes at room temperature and then

washed and stored in PBS at 4°C.

S1.1.3 Experimental procedure

Prior to use devices were flushed, injecting PBS into the microchannel using large bore

needles and 1ml syringes. Pre-PP1 cells were obtained as described in Section 2.1,

centrifuged and resuspended to 500 000 cells/ml in PBS. For each device 500μl of cell

suspension was drawn into a 1ml syringe and the syringe attached to a 30G needle (Small

Parts Inc., NE-301PL-C). The needle was inserted into a 15cm long piece of tygon tubing

(Small Parts Inc, TY-010-C). The syringe was gently depressed to fill the tubing with cell

suspension, and flicked to remove air bubbles. The free end of the tubing inserted into the

input port of the microfluidics device, being careful not to introduce any air bubbles into

the fluid stream. The syringe was then placed into a syringe pump. The output port of the

device was connected via a 2.5cm piece of Tygon tubing (again avoiding air bubbles) to a

tube for collection. Cells were injected at a flow rate of 10μl/min for 45 minutes at room

temperature.

S1.1.4 Immunofluorescent staining protocol

To identify cardiac progenitors by expression of the marker islet-1 (isl-1)

immunofluorescent staining was performed. The isl-1 monoclonal antibody was obtained

from the Developmental Studies Hybridoma Bank, University of Iowa (clone 39.4D5).

Cells were fixed for 20 minutes in 4% PFA, washed with PBS and then permeabilized

50

Figure S 1: Geometry of Hele-Shaw microfluidics device showing shear stress profile and

characteristic equation. Equation 1 and graph showing the linear relationship between shear stress Ʈ and

axial distance from inlet z (in mm). H is the height and w the width of the microchannel, Q the volumetric

flow rate and L the total length of the channel in mm. Reproduced with permission from Plouffe et al. [20]

with 0.1% Triton X-100 (Alfa Aesar, P1379). Cells were incubated with isl-1 primary

antibody at a working dilution of at least 2 µg/mL (1:200 stock) [17]. For detection cells

were incubated with goat anti-mouse FITC-conjugated secondary antibody (Sigma

F9006, 1:64). RIN-m cells (a rat pancreatic beta-islet cell line, ATCC, CRL-2056) were

used as a positive immunostaining control and H5V (mouse aortic endothelial cells,

ATCC CRL-2299) used as a negative staining control for isl-1. For staining devices

PDMS above the microchannel was excised with a scalpel to access the cells for fixing

and staining.

S1.1.5 Results and Discussion – isl-1 staining on adhered cells in Hele-Shaw devices

An initial investigation was performed to ensure that cardiac progenitors in the cardiac

cell mixture would not interact with and adhere on the peptide-modified glass surfaces

intended to remove EC, SMC and FB cell types. Pre-PP1 cardiac cells were run through

bare glass, RGDS, REDV and VAPG devices as described in Section S1.1.3. The PDMS

Equation 1

51

above the channel was excised and the cells immunofluorescently labelled to detect

progenitors by isl-1 expression. For comparison, RIN-m cells, an isl-1+ rat pancreatic

beta-islet cell line served as a positive staining control and H5V mouse endothelial cells

as a negative control. As shown in the representative images in Figure S 2 below, none of

the microchannel surfaces exhibited positively staining cells for isl-1. It may be safe to

say that our choice of peptides does not appear to capture the progenitor population and

they are likely flow through the devices and are collected at the output. It is possible

however that the lack of attachment to the modified surfaces was partially due to the

scarcity of isl1+ cells in the initial cell suspension and further investigation is required.

We may tentatively say then that our approach of using peptides for a negative selection

process holds merit and could eventually be used to realize the isolation of progenitors.

Figure S 2: Immunofluorescent staining for progenitor cell islet 1 markers in captured cells in

peptide-modified microfluidic devices. Representative images of immunostaining for cardiac progenitor

cell islet 1 (isl-1) marker (green) in cells captured on peptide-functionalized glass surfaces of Hele-shaw

microfluidics devices. Nuclei were counterstained with DAPI (blue). Controls included peptide- bare glass

microchannel, positive staining control of isl-1+ RIN-m pancreatic cells and negative staining control of

H5V isl-1- endothelial cells.