n vitro bioassembled human extracellular matrix and its application in human embryonic stem cell cultivation 1

1
Chapter 1: Introduction
1.1 Human Embryonic Stem Cells
Stem cells are cells that are at an early stage of development, which can
proliferate by self-renewal and retain the potential to differentiate into two or
more cell types [1, 2, 3, 4].

There are two types of stem cells - adult stem cells and ES cells. Adult stem
cells are found within adult organs whereas ES cells are derived from the inner
cell mass of pre-implantation embryos and have a high nucleus to cytoplasm
ratio, prominent nucleoli and distinct colony morphology.

The first ES cell line was derived from mouse embryos [5, 6], while hESCs
were derived from human blastocysts [7].

hESCs characteristically have high telomerase activity and express known
surface markers - SSEA-3, SSEA-4, Tra-1-60 and Tra-1-81 [7, 8]. The POU
transcription factor, Oct-4 [9, 10] and Nanog [11, 12] are also highly
expressed in hESCs. In contrast to adult stem cells, these cells are able to
differentiate into cells of all three germ layers - the endoderm, mesoderm and
ectoderm; when implanted into SCID mice, they form teratomas and when
allowed to grow to over-confluence, hESCs differentiate to endodermal and
trophoblastic lineages [7]. When cultured as suspended small clumps in
differentiation medium and subsequently plated onto gelatin-coated plates,
they form embryoid bodies with heterogenous morphology in the outgrowths

2
which are positive for differentiation markers β-tubulin III, cardiac troponin I
and α-fetoprotein [4].

Given the hESCs’ ability to proliferate extensively and their capacity to
differentiate into lineages that are otherwise unavailable, hESCs are
considered an attractive therapeutic tool for tissue engineering, cellular
transplant therapy and as drug discovery tools [2, 3, 4, 7, 13, 14]. In tissue
engineering, stem cells can possibly be used to generate healthy tissue to
replace those damaged by trauma or disease. It is believed that diseases that
involve specialized cells known for limited regeneration, such as Parkinson’s
disease, Alzheimer’s disease, heart diseases, stroke, arthritis, diabetes and
spinal cord damage, can potentially be treated by cellular transplant therapy.
To test a large library of drugs in the process of drug discoveries, large
numbers of cells of the known target disease have to be available. In the case
of specialized cells with limited proliferation potential, such large numbers are
difficult to attain. hESCs, with their ability to proliferate and subsequently
differentiate into the required cell type, promise to alleviate this problem in
drug discovery.

Despite the promise of hESCs in tissue engineering and cellular transplant
therapy, the gap between scientific research and clinical applications for safe
and effective stem cell based therapies is still wide. Generally, for hESCs to be
applied for human clinical trials, various bottlenecks have to be cleared before
the process is considered safe. hESCs have to undergo proliferation to
generate a large number of pluripotent cells, which are then induced to

3
undergo differentiation into the desired cell type. Issues that need to be
addressed include limiting the transfer of infectious agents from the donor
cells to the host, reducing host immune responses by limiting immunogen
exposure and retention in donor cells, reducing the risk of tumour generation
upon implantation and attaining the desired efficacy of the differentiated cell
type. All these issues are dependent on the proliferation and differentiation

process that the hESCs will undergo before implantation.

1.2 Conventional Culture of hESCs
Typical culture of hESCs requires the use of mitotically inactivated MEF
feeder layers and fetal bovine serum to maintain their undifferentiated state [3,
4, 7]. The feeder layers are presumed to provide soluble factors and an
anchoring substrate for hESCs, hence providing a substitute growth-conducive
microenvironment that maintains hESCs’ pluripotency.

However, the use of MEF and fetal bovine serum has proven disadvantageous
for the clinical applications of hESCs. Although MEFs are arrested in a post-
mitotic state with mitomycin C treatment prior to hESC co-culture, there is no
way to remove them completely when the cells are brought into suspension.
Typically, MEF constitute 9% to 38% of the confluent co-culture. As a result,
MEFs are unavoidably implanted together with the hESCs into the host, even
though this may only be a small admixture. The presence of MEFs in the
cellular graft could lead to host immune rejection and also the transfer of
infectious agents. MEFs are of animal origin, which exposes animal pathogens
to hESCs. It was found that hESCs exposed to animal-derived products

4
expressed Neu5Gc, an immunogenic non-human sialic acid, thus rendering
them unsuitable for clinical transplantation into humans [15]. MEFs were also
found to contain Neu5Gc [15].

Although other feeder layers using human cell types have been developed,
they are fibroblasts derived from fetal skin and muscle of aborted embryos, or
from adult tissues such as Fallopian tubes and foreskin [16]. However, at
present, no GMP-grade xeno-free human feeder cells are available [16].

Additionally, the use of feeder layers can confound downstream analysis of
hESCs. For example, during flow cytometry analysis of pluripotency marker
expression by hESCs grown on MEFs, the MEFs have to either be separated
from the hESCs prior to analysis, or labeled with a marker specific for MEFs
and subtracted during the analysis process. In reverse transcriptase-polymerase
chain reaction analysis and other characterization analyses, MEFs can
contribute 9% to 38% of the entire co-culture population, therefore adding
significantly to background noise.

Even the use of human cells as feeders poses challenges for hESC culture, as
the dependence on a co-culture system makes it technically challenging for
large-scale expansion, such as in bioreactor settings [4, 16]. To generate
clinically-relevant numbers of differentiated cells from hESCs, a large number
of undifferentiated hESCs are first required, as only a fraction of the
undifferentiated hESCs will differentiate into the desired lineage. To generate
these undifferentiated hESCs, a large number of MEFs or human feeders are

5
hence required. As MEFs have to be harvested from developing mouse
embryos and have a limited culture lifespan (6-8 passages), a large number of
mouse embryos must be sacrificed making this a technically challenging and
resource-wasting process. Human feeders such as foreskin fibroblasts also
have a limited lifespan, fibroblasts from fallopian tubes are limited in number,
and the acquiring of human feeders from aborted embryos faces the same
problems as the acquisition of MEFs.

Finally, feeder layers have batch-to-batch variability and the undefined quality
of feeders makes reproducibility difficult. One of the major criteria for
developing cellular therapies using hESCs is to provide cells with defined
quality characteristics that are safe for the patient. In order to do so, GMP

need to be employed. Other than preventing microbial contamination in the
product, GMP requires the development of validated standard operating
procedures to ensure that the cells are produced in a reproducible manner [16].
As such, the batch-to-batch variability of feeder layers would present
difficulties in achieving GMP standards.

1.3 Feeder-Free Culture of hESCs
To tackle the limitations of feeder layer culture, feeder-free culture systems
have been developed. Substitutes for soluble growth factors and anchoring
substrates are the underlying principles of these systems, which try to mimic
the microenvironments of the hESCs. Indeed, keeping in mind the origin of
hESCs, the neighbouring cells and their ECM serves as environmental cues
for the developing hESCs. Laminin, a basement membrane protein, and

6
collagen I and III are produced as early as the 8-cell stage mouse embryo,
while other ECM proteins, such as Fn, HSPG and collagen type IV appear
later [17]. In 5 day old mouse blastocysts, laminin, entactin, collagen IV,
HSPGs and Fn were found in the basement membrane between the ectoderm
and the primitive endoderm which later becomes the inner cell mass [17]. As
the embryo develops, new germ layers will emerge, generating new
epithelium and basement membrane, providing anchorage for the hESCs [18].
The presence of ECM proteins in developing embryos indicates that these
proteins play a role in maintaining ES cell. As such, feeder-free culture
systems typically consist of various ECM substrates supplemented with
medium containing various growth factors.

1.3.1 Matrigel
One commonly used feeder-free culture substrate is Matrigel™, a solubilized
basement membrane preparation extracted from EHS mouse sarcoma [19]. Its

major component is laminin, followed by collagen IV, HSPGs, and entactin
[19, 20]. The membrane is harvested using 2M urea and 0.05M Tris-HCL, pH
7.4 [19] and reconstituted as a gel at temperatures above 10ºC, with an
optimum temperature of 35°C [20]. The culture of hESCs on Matrigel requires
the use of conditioned medium from MEF feeder layers or chemically defined
medium.



Matrigel with MEF Conditioned Medium

7
hESCs can be propagated for long periods of time on Matrigel with
conditioned medium (up to 130 population doublings) [4]. Conditioned
medium is obtained by harvesting culture medium from MEFs, which contains
growth factors and cytokines secreted by the MEFs. The need for conditioned
medium still represents a major limitation. Firstly, culture of hESCs still
requires the culture of MEF, albeit in a separate system, for conditioned
medium. As stated earlier, one of the problems with large scale expansion
using feeder systems that require MEFs is the limited culture lifespan of MEF
and hence the continued need for harvesting MEFs from mouse embryos. As
such, the large-scale expansion of hESCs using Matrigel with conditioned
medium still faces the same problem faced by the standard feeder culture
system. Secondly, the use of conditioned medium from non-human MEFs can
still introduce animal pathogens to hESCs, which can cause the expression of
immunogens in these hESCs, rendering them unsuitable for clinical
application. Thirdly, conditioned medium is still undefined, and hence subject
to batch-to-batch variability.

Matrigel with Defined Medium

hESCs express receptors for various growth factors such as FGF-2, stem cell
factor and fetal liver tyrosine kinase-3 ligand [21]. Xu et. al. found that hESCs
could maintain undifferentiated proliferation on Matrigel in serum-substituted
medium supplemented with 500ng/ml noggin and 40ng/ml FGF-2 [22].
Ludwig et. al. also found that 5 factors, FGF-2, TGFβ, LiCl, γ-aminobutyric
acid and pipecolic acid I medium with human serum albumin were critical for
optimal undifferentiated proliferation of hESCs on Matrigel [23]. These

8
studies have shown that hESCs can be cultivated on Matrigel in culture
medium supplemented with defined concentrations of recombinant factors,
hence removing the dependence on MEFs for conditioned medium.

Culturing of hESCs on Matrigel removes the dependence on MEFs, thereby
eliminating the risk of accidental MEF transplantation, and reducing
downstream analysis complexity. However, Matrigel is of mouse origin,
which still poses a xenogenic threat. As described earlier, Neu5Gc is a non-
human sialic acid that can elicit an immune response in humans, and Matrigel
has been found to contain Neu5Gc [24].

In addition, as a product of a tumour cell line, Matrigel’s components, and
their relative proportions, might be atypical of the native microenvironment
for hESCs. Since Matrigel is of mouse origin, the proteins that it presents to
the hESCs are xenogenic. Moreover, the process of harvesting Matrigel
destroys lysine-derived cross-links and disulphide bonds required for the
stability of collagen and other protein assemblies. The components of
reconstituted Matrigel appear to be linked by noncovalent bonds only, and
although the proportions of components are constant, there were no parallel
multilamellar structures typical of native basement membranes [20], indicating
that the exact original morphology of the ECM cannot be regained.

1.3.2 Purified Matrix Components as coating material

9
Various groups have cultured hESCs on single matrix components, such as
laminin and Fn. Xu et. al. found that hESCs can be cultured on commercially
obtained laminin 111 using MEF-conditioned medium for more than 6 months
[4]. However, it was later reported by other groups that there was significant
variability in results, depending on the source and batch of laminin [22]. Amit
et. al. showed that hESCs could be propagated for more than 15 passages on
recombinant Fn in serum-substituted medium supplemented with TGFβ, LIF
and FGF-2, although these hESCs exhibited lower cloning efficiency as
compared to those grown on MEFs [25, 26].

In addition to the variability of substrate and lower cloning efficiency, such
purified matrix components are expensive for large-scale expansion, and are
oversimplified mimics of the complex ECM that hESCs are exposed to in their
natural microenvironments.

Native in vivo ECM is made up of numerous different components, namely
glycoproteins (such as collagens, laminins, nectins, elastin and microfibrillar
components), proteoglycans (such as decorin, biglycan, syndecan etc) and
hyaluronic acid. Not only does the ECM serve as a scaffold for cell attachment
and provide strength for tissue integration, it also serves as a reservoir of
growth factors and cytokines.

Cell attachment to ECM via cell surface receptors such as integrins results in
cell-matrix interactions, which regulate cell survival, proliferation, motility

and differentiation [27].

10

The ECM stores growth factors, such as TGFβ and FGF 2, which have been
demonstrated to have profound effects on ES cell culture.

Fbn and Fn, components of the ECM, binds to LTBP-1 [28, 29], which in turn,
associates with LAP [28, 30]. LAP binds to TGFβ in its latent form, and the
latter becomes activated when released from LAP [31, 32]. TGFβ prevents ES
cell differentiation by maintaining Smad2/3 phosphorylation [33].

FGF-2 interacts with heparin sulfate proteoglycans in the ECM [34] and
maintains undifferentiated proliferation of hESCs [22]. The biological action
of FGF-2 is prolonged when it is bound to the matrix [35]. Furthermore, the
binding of FGF-2 to ECM protects it from proteolytic degradation [34].

A pure single-component substrate such as laminin or Fn, would lack be
lacking in the complex array of ECM proteins and other components, whereas
in vivo ECM is bio-assembled as a complex mixture of components.
Therefore, such single-component substrates can hardly be termed a mimic of
the complex microenvironment for hESCs, leading recent research to focus on
the use of animals cells-derived ECM, which contain the complex array of
proteins in their in situ architecture.
1.3.3 MEF-Lysed ECM
Klimanskaya et. al. derived and cultured hESCs on MEF-lysed ECM with
medium supplemented with 20ng/mL LIF and 16ng/ml FGF-2 [36]. The
MEFs were lysed with 0.5% DOC in 10mmol/L Tris-HCL, pH 8.0 on ice for

11

30 minutes and remaining matrices were rinsed 5-6 times with PBS. Again,
the use of MEF potentially exposes the hES to MEF-derived pathogens, hence
the authors proposed sterilizing the MEF-lysed ECM before use. Preliminary
studies conducted by the authors found that the MEF-lysed ECM could be
dehydrated, treated with paraformaldehyde, heat-pasteurized (60°C for over
20 hours) or gamma-irradiated without substantial changes to hESC
behaviour. However, as the authors acknowledged, the complex array of
proteins, growth factors and hormones in the ECM would be denatured and
destroyed by the sterilization process.

Also, it has been found that Neu5Gc represent 20% of total sialic acid in
MEFs [15], and whether Neu5Gc is removed by the detergent lysis process
was not yet proven by Klimanskaya et. al., hence MEF-lysed ECM might still
expose hESCs to this immunogen. Moreover, this system does not eliminate
the need for continuous harvesting of MEFs from mouse embryos.

1.3.4 Commercially Available Matrix Substrates
Besides Matrigel, other matrix substrates commercially available for hESC
proliferation include Geltrex (Gibco), Cultrex (R&D Systems), ECM gel
(Sigma-Aldrich) and PuraMatrix (BD). Of these, Geltrex, Cultrex and ECM
gel are all derived from mouse EHS sarcoma, like Matrigel. Hence, like
Matrigel, it is to be expected that being of mouse-origin, they will also contain
Neu5Gc and possibly animal pathogens. On the other hand, PuraMatrix is
made of synthetic amphiphilic oligopeptides that self-assemble in
physiological salt solutions [37]. Since PuraMatrix does not contain any

12
animal-derived materials, it poses no xenogenic threat. However, according to
the manufacturer, PuraMatrix forms a relatively soft fibrous matrix, and
cannot withstand vacuum aspiration around the hydrogel during medium

removal. Thus, PuraMatrix is more fragile than Matrigel in terms of handling,
and there are currently no data supporting hESC proliferation on this material.

1.4 Ideal hESC Culture System
The various culture systems described above have sought to improve the
culture of hESCs, but there are still open issues. The use of mouse-origin
Matrigel, MEF-conditioned medium, expensive purified matrix components,
and MEF-lysed matrix need to be replaced by transplant-friendly and cost-
effective culture systems. In view of the imperfections of existing culture
systems, an optimal system for the culture and derivation of hESCs would
require feeder-free culture on a human ECM, maintained on chemically
defined culture medium containing only human substances and proteins to
produce clinical-grade hESCs [16]. And such a human ECM can made by
human fibroblasts, the professional ECM producer.

Advances in ECM deposition in our laboratory have shown that the use of
MMC in the culture of human fibroblasts can significantly enhance the
production of a collagenous matrix [38, 39]. Such a matrix could then be
treated with detergents for fibroblast removal, rendering the ECM cell-free. As
such, this complex ECM, which does not require the use of mouse-origin
Matrigel nor dependence on MEFs, would be suitable for hES culture

13
application. Moreover, such a cell-free ECM would serve as a study on the
effects of ECM on hESC behaviour.

Substrate
Culture
Medium
Advantages

Disadvantages
MEF Feeder
Layer
Human fetal
fibroblasts,
human adult
fibroblasts

• Soluble factors
• Anchoring substrate
• Unsuitable for
clinical
transplantation
• Confounds
downstream analysis
• Technical challenge
• Dependence on
limited sources of
fibroblasts
• Variability of
feeders
• Xenogenic threat
Conditioned
Medium
• Separate culture of
MEFs
• Dependence on
MEFs
• Xenogenic threat
• Variability of

conditioned medium
Matrigel
Defined
Medium
• Non-dependence on
MEFs
• Xenogenic threat
Purified
Matrix
Component
- Laminin
Conditioned
Medium
• Defined substrate
• Dependence on
MEFs
• Xenogenic threat
• Variability of both
substrate and

14
conditioned medium
• Oversimplified
model of ECM
Purified
Matrix
Component
- Fn
Defined
Medium

• Defined substrate
and medium
• Expensive
• Oversimplified
model of ECM
MEF-Lysed
ECM
Defined
Medium
• Suitable model of
ECM
• Dependence on
MEFs
• Xenogenic threat
Ideal hESC
Culture
System –
Human
fibroblast-
lysed ECM
Defined
Medium
• Reduce xenogenic
threat
• Non-dependence on
MEFs
• Suitable model of
ECM

Table 1: Summary of hESC culture systems

15

1.5 Hypothesis
A human ECM can be bio-assembled in vitro, decellularized and
subsequently applied for the sustained growth and maintenance of
pluripotency of hESCs.

1.6 Objectives
1.6.1 Enhance in vitro Bio-assembly of Matrix Proteins by Fibroblasts
To manipulate fibroblasts to deposit ECM which is rich in proteins
such as collagen, Fn, ligands and growth factors

1.6.2 Complete Removal of Fibroblasts to obtain Cell-Free Matrices
To completely remove fibroblasts while retaining acceptable ECM
protein and ECM structure levels

1.6.3 Characterization of Cell-Free Matrices
To characterize cell-free matrices and identify components of the
matrix

1.6.4 Propagation of hESCs on Cell-Free Matrices
To cultivate hESCs on cell-free matrices with characterization of the
hESCs and to assess the maintenance of their undifferentiated state and
pluripotency


16
Chapter 2: Materials and Methods
2.1 Cell Culture

2.1.1 Matrices
Wi38 fetal lung fibroblasts (ATCC, CCL-75) at passage 20-22 were seeded at
25,000 cells/cm
2
in either 24 well or 6 well culture plates (CelStar, Greiner
Bio-one, 662160) in high glucose DMEM (Gibco-BRL, 10569-010)
supplemented with 10% FBS (Gibco, 10270-106), which was replaced with
0.5% FBS culture medium supplemented with 30µg/ml AcA (Wako, 019-
12063) with 100µg/ml DxS (USB corporation, 70796-50G) on the following
day. After 3 days of incubation at 37°C and 5% carbon dioxide, the fibroblast
cell layers were washed with PBS twice and lysed with 0.5% DOC (Prodotti
Chimici E Alimentari, S.P.A. 2003030085) and 0.5X Complete PrI (Roche
Diagnostics Asia Pacific, 11836145001) 3 times 10mins on ice to obtain cell-
free DxSDOC matrices while a similar protocol of 6 times of 0.5% DOC
treatment was applied to get DxSDOCDOC matrices. DxSNP40/DNase
matrices were made by incubating fibroblasts with 30µg/ml AcA and
100µg/ml DxS for 3 days, washed with PBS twice and lysed with 1% NP40
(Fluka BioChemika, 74385) and 0.5X PrI 4 times 10mins on ice and treated
with DNase I (US Biologicals, D3200) twice for 1 hour each at 37°C. DNase
was used at dilution 1:50 in a buffer made of 10mM Tris pH 7.5 (J.T.Baker,
4109-02), 2.5mM MgCl
2
(Fluka BioChemika, 63068) and 500µM CaCl
2

(Merck, 1.02382.1000). FcNP40/DNase was obtained by incubating
fibroblasts with 30µg/ml Aca and 37.5mg/ml Fc 70 (GE Healthcare, 17-0310-
50) and 25mg/ml Fc 400 (GE Healthcare, 17-0300-50) for 6 days and lysed in

17

a similar manner as DxSNP40/DNase matrices. All matrices were washed
with PBS 3 times to remove all traces of the lysis solutions.

2.1.2 hESCs
H9 (WiCell, WA09) were passaged using collagenase IV (Gibco, 17104-019)
or dispase (StemCell Technologies, 07913) equally onto Matrigel (BD, hESC-
qualified, 354277), DxSDOC and DxSDOCDOC matrices and cultured using
defined medium mTeSR-1 (StemCell Technologies, 05850). Culture medium
was changed daily and subculture was done every 5-7 days.

2.1.3 iPSCs
DL1 cells were generously supplied by Dr Jeremy Crook of Singapore Stem
Cell Bank, Singapore Stem Cell Consortium, Institute of Medical Biology.
The iPSCs were derived by transfecting IMR-90 fetal fibroblasts with
lentiviral vector expressing Oct-4, Sox2, Nanog, Lin28 and Large T. The
iPSCs were cultured on MEFs prior to seeding onto matrices, and subcultured
every 5-7 days using dispase.

2.2 Characterizations of Matrices
2.2.1 SDS-PAGE
Collagen is synthesized by fibroblasts and secreted in soluble form as
procollagen into the extracellular space. Under in vivo conditions, procollagen
is extracellularly cleaved by C- and N-proteinase enzymes to form insoluble
collagen. The insoluble collagen then deposits in a quarter-staggered array to
form insoluble collagen fibrils. Unlike the in vivo situation, collagen

18
synthesized by fibroblasts in cell culture remains mostly in the soluble form,
which accumulates in the culture medium instead of depositing at the cell
layer. To correctly access the amount of insoluble collagen rather than soluble

procollagen in fibroblast cultures, procollagen was first removed by aspiration
of culture medium, followed by 2 washes of PBS. In this way, only insoluble
collagen found at the cell layer would remain. To further isolate collagen from
other proteins found also at the cell layer, the samples in 24-wells plate format
were digested with 100µl of 0.25mg/ml of pepsin (Roche Diagnostics Gmbh,
10108057) supplemented with 0.5% Triton-X 100 (Bio-Rad, 161-0407),
which degrades all proteins into peptides except collagen. Pepsin-digested
samples were separated under non-reducing conditions using 5% resolving/
3% stacking polyacrylamide gels as outlined in [38, 39]. Protein standards
used were the Precision Plus Dual Color (Bio-Rad,161-0374 ) and collagen
type I (Koken Co., IAC-13). Protein bands were visualized using a Silver
Quest Silver Staining Kit (Invitrogen Life Technologies, LC 6070) according
to manufacturer’s instructions. The amount of collagen present in the samples
was quantified by densitometry of the α-bands of collagen using a GS-800TM
Calibrated Densitometer (BioRad, 170-7980).

2.2.2 Immunofluorescence and Mass Spectrometry
The following antibodies were used to immunolabel DxSDOC,
DxSDOCDOC, DxSNP40/DNase, FcNP40/DNase matrices: mouse anti-PDI
(Invitrogen, S34200), Alexa-Fluor 594 Phal (Invitrogen, A12381), mouse anti-
Col I (Sigma, C2456), rabbit anti-Fn (Dako, A0245), mouse anti-HS
(Seikagaku, 370255), mouse anti-Fbn (Chemicon, MAB 2499), rabbit anti-

19
DCN (gift from Dr Larry Fisher, LF-136), rabbit anti-BGN (gift from Dr
Larry Fisher, LF-112), rabbit anti-LTBP-1 (gift from Dr Carl Heldin, AB39),
Alexa Fluor 594 goat anti-mouse (Invitrogen, A11020), Alexa Fluor 488
chicken anti-rabbit (Invitrogen, A21441) and DAPI (Invitrogen). Mass
spectrometry identifications were done by Christopher Stephen Hughes and
Prof Gilles Lajoie, using Mascot, OMSSA and X!Tandem after analysis on the

mass spectrometer. Scaffold was used to validate MS/MS based peptide and
protein identifications. Protein probabilities were assigned by Protein Prophet
algorithm. Peptide and protein identifications were accepted if they could be
established at greater than 95% and 99% probabilities respectively. Proteins
that contained similar peptides and could not be differentiated based on
MS/MS analysis alone were grouped to satisfy the principles of parsimony.

2.3 Analysis of hESCs
2.3.1 Preparation of Matrigel Coating
175µl of BD Matrigel hESC-qualified Matrix (BD, 354277) was diluted in
12.5ml of DMEM-F12 medium (Invitrogen, 11330032) overnight at 4°C.
1000µl or 200µl of diluted Matrigel was coated onto each well of 6-well
culture plates or 24-well culture plates respectively and incubated for 1 hour at
37°C. Wells were used immediately for hESC culture.

2.3.2 Population Doubling Curve
Representative wells were sacrificed for cell counting at each passage to
obtain the number of cells and from there calculate the population doublings.
The wells were harvested into single cell suspensions using TrypLE Express

20
(Gibco, 12604-021) and neutralized with 10% FBS culture medium. All
counting was done using Trypan Blue and a haemacytometer.

2.3.3 Flow Cytometry
The hESCs were harvested into single cell suspension using TrypLE and fixed
with 4% formaldehyde for 10 minutes at 37°C. The following antibodies were
used to immunolabel the hESCs: goat anti-Oct-4 (Santa Cruz, sc-8628), mouse
anti-TRA-1-60 (Chemicon, MAB 4360), mouse anti-TRA-1-81 (Chemicon,
MAB 4381), rat anti-SSEA-3 (Chemicon, MAB 4303), mouse anti-SSEA-4

(Chemicon, MAB 4304) and Alexa Fluor 488 donkey anti-goat (Invitrogen,
A11055), Alexa Fluor 488 chicken anti-mouse (Invitrogen, A21200) and
Alexa Fluor 488 goat anti-rat (Invitrogen, A11006). Flow cytometry was
performed using CyAn.

2.3.4 Adherent Immunofluorescence
The following antibodies were used to immunolabel the hESCs that were fixed
with formaldehyde: goat anti-Oct-4 (Santa Cruz, sc-8628), mouse anti-SSEA-
4 (Chemicon, MAB 4304), Alexa Fluor 488 donkey anti-goat (Invitrogen,
A11055), Alexa Fluor 594 goat anti-mouse (Invitrogen, A11018) and DAPI.

2.3.5 In Vivo Differentiation
hESCs cultured on Matrigel, DxSDOC or DxSDOCDOC for 18 passages were
harvested using collagenase IV, resuspended in HBSS, and injected into the
right hind limb of each SCID mouse with the generous help rendered by Toh
Wei Seong and Prof Cao Tong.

21

2.3.6 In Vitro Differentiation
The protocol for neural differentiation was adapted from WiCell Research
Institute, National Stem Cell Bank (SOP-CH-207A). hESCs that were cultured
on Matrigel, DxSDOC or DxSDOCDOC were harvested using collagenase
and seeded onto CoStar non-adherent plates (Corning, 3471) to form embryoid
bodies using 20% Knockout Serum (Invitrogen, A1099202) in Knockout
DMEM (Invitrogen, 10829-018). After 4 days, culture medium was changed
to DMEM/F12 (Invitrogen, 10565-042) supplemented with Non-essential
amino acids (Invitrogen, 111140050), N2 (Invitrogen, 17502-048), heparin
(Sigma, H3149) and FGF-2 (ProSpec, CYT-557_10µg) and embryoid bodies
were seeded onto new standard 6 well plates for 3 days. Finally, embryoid

bodies were seeded onto wells coated with 20µg/ml laminin (Sigma, L6274)
according to manufacturer’s instructions and incubated for 7 days. Samples
were fixed with formaldehyde and immunolabeled with rabbit anti-β III
tubulin (Abcam, ab18207) and Alexa Fluor 488 chicken anti-rabbit
(Invitrogen) and DAPI.

2.3.7 Karyotype
Matrices were prepared in 25cm
2
culture flasks (CelStar, Greiner Bio-one,
690-175) in a similar manner as stated in section 2.1.1. 125,000 Wi38 fetal
lung fibroblasts were seeded into each culture flask using 10% FBS in DMEM
and the culture medium was changed to 0.5% FBS in DMEM supplemented
with 30µg/ml AcA and 100µg/ml DxS. After 3 days of incubation at 37°C and
5% carbon dioxide, the fibroblast cell layers were washed with PBS twice and

22
lysed with 0.5% DOC and 0.5X Complete PrI 3 times for 10mins each on ice
to obtain cell-free DxSDOC matrices while a similar protocol of 6 times of
0.5% DOC treatment was applied to get DxSDOCDOC matrices. Matrigel
was prepared in 25cm
2
culture flasks in a similar manner as stated in section
2.3.1. 2.5ml of diluted Matrigel was used for each culture flask and incubated
for 1 hour at 37°C.
1.5 million hESCs were seeded into each culture flask and incubated in 5ml of
mTeSR-1 for 3 days with daily changes of culture medium. Karyotyping
services were obtained from KK Women’s and Children’s Hospital.
Harvesting was done using colcemid and staining done with Trypsin and
Wright’s-Giemsa stain.

2.3.8 DNA Methylation
hESCs propagated on Matrigel, DxSDOC and DxSDOCDOC were washed
with HBSS twice before the cells were harvested in 500µl HBSS by scraping.
The cells were pelleted by centrifuging at 1100rpm for 3 minutes and the
supernatant was removed. The pellet was then transferred into a cryovial and
snap frozen in liquid nitrogen. The vials were then stored in -80°C before
shipping under dry ice conditions to Prof Frank Lyko and Michael Bocker of
University of Heidelberg, German Cancer Research Center for DNA
methylation studies using the Illumina Methylation Assay – Infinium II.
Analysis was done using Illumina Beadstudio Software to generate the
genome-wide correlation data. The methylation states of samples obtained
from propagation trial 3.5 and trial 3.6 was inter-chip normalized before
generating the heatmaps.

23
Chapter 3: Results and Discussion
3.1 MMC enhances deposition of Col I in ECM
3.1.1 Negatively-charged macromolecules
Standard culture of fibroblasts typically requires 10%FBS in DMEM. To
obtain an enhanced collagen content laid down by fibroblasts in ECM, AcA is
added to the culture medium [40]. However, AcA is not directly responsible
for matrix deposition but instead is a co-factor for procollagen
prolylhydroxylase, the enzyme responsible for posttranslational modification
of procollagen, a crucial process that renders it thermostable for secretion.
AcA basically boosts the secretion of procollagen. However, even under AcA
supplementation, collagen deposition is tardy due to a slow enzymatic
conversion of procollagen to collagen. Work by our laboratory has shown that
the addition of negatively charged macromolecules is able to enhance collagen
deposition by fibroblasts [38, 39, 41, 42, 43]. Based on these data we used

ECM deposited under MMC with DxS. Fibroblasts were first seeded for 1 day
using 10%FBS in DMEM, and the culture medium was exchanged the next
day for 0.5% FBS/DMEM supplemented with AcA and the negatively charged
macromolecule DxS. Serum contains up to 80mg/ml of macromolecules in
protein form [39], and the use of low serum culture medium would reduce the
potential of interference on collagenous deposition by the fibroblasts. After 3
days of incubation, the culture medium fraction was removed and the cell
layer was harvested by pepsin digestion. Control samples were fibroblasts
incubated for the same period in low serum culture medium, and AcA control
samples were fibroblasts incubated in low serum culture medium
supplemented with AcA. SDS-PAGE of the samples separated the proteins to

24
reveal the Col I bands, α
1
(I), α
2
(I), β
11
and β
12
(Figure 1A). The optical
density of the alpha bands were measured by densitometry, normalized to the
AcA control samples, and plotted in a graph (Figure 1B). In contrast to AcA-
enriched culture medium, fibroblasts grown in media without AcA only
produced half that amount of collagen in the ECM. Interestingly, the addition
of DxS to the culture condition resulted in an almost 10-fold increase in Col I
deposition compared to the AcA control, showing that the addition of DxS can
be used to increase Col I deposition in fibroblast cultures.

The morphologies of the collagen and Fn laid down under the above
conditions were investigated using immunofluorescence labeling of Col I and
Fn (Figure 1C). When fibroblasts were incubated using low serum culture
medium, the ECM laid down contained some Fn and minimal Col I. When the
culture medium was supplemented with AcA, the resultant ECM still
contained some Fn; however, an increased amount of Col I was laid down in
wispy thin strands, which were co-localized with the Fn. When DxS was
added into the AcA-enriched culture medium, the resultant ECM contained
more Fn, which was noted to be in a granular form. There was also more Col I
detected; and similar to Fn morphology, this Col I also showed a granular
morphology. The greatly increased amount of deposited Col I was clearly
visible.

25