Development of cell sheet constructs for layer by layer tissue engineering using the blood vessel as an experimental model 3

Figure 4-6: Expression of endothelial markers Immunocytochemical assays demonstrate the expression of surface endothelial markers CD31 PECAM and CD144 VE-Cadherin.. As the cell populatio

4.2.2.2 Matrigel tube formation assay

I then employed the Matrigel tube formation assay to compare the vasculogenic capacity of fetal blood and cord blood derived EPC When the cells were trypsinised and transferred to a Matrigel, substrate, established networks were formed from both groups within 16 hours However, it was noted that FB-RPC was capable of more robust tube formation, resulting in networks visible to the naked eye Microscopic examination of the networks revealed structures indicative of secondary angiogenic sprouting

Figure 4-5: Matrigel vascularisation assay

Tube formation in Matrigel by (a) UCB-EPC and (b) FB-EPC Greatly increased

network formation was found in FB-EPC, with secondary angiogenic sprouting

(a)

A functional assay of acLDL uptake was performed to confirm the endothelial nature

of the cells In contrast to the immunostaining results, acLDL was taken up by more than 90% of all cells in both groups, suggesting their functional endothelial identity

Figure 4-6: Expression of endothelial markers

Immunocytochemical assays demonstrate the expression of surface endothelial

markers CD31 (PECAM) and CD144 (VE-Cadherin) Cells also express von

Willebrand factor in well defined Weider-Palade bodies, albeit at reduced levels in

FB-EPC All cells took up acetylated low density lipoproteins (acLDL), seen as

127

4.2.2.4 Microarray analysis of blood outgrowth cells

To further elucidate differences between UCB-EPC and FB-EPC, I proceeded to compare the differences in expression of the two populations (n=1) through a whole genome expression array approach As the cell population appeared to be contaminated with fibroblast-like cells, I decided to carry out MAC Sorting to enrich the population for CD31 (an endothelial marker) expressing cells prior to RNA extraction for microarray The selected cells were plated overnight to reduce changes

in expression level due to the sorting process The quality of extracted RNA was assessed using spectrophotometry and formaldehyde gel electrophoresis prior to labelling for microarray

Figure 4-6: Assessment of RNA quality

Quantity of RNA was assessed by spectrophotometry prior to labelling for microarray Sufficient, good quality RNA was obtained, as determined by denaturing gel

FB- EPC

UCB-28S 18S

Analysis of the microarray data shows 27,530 out of 55,000 genes to be appreciably expressed in either line The profiles of both groups were largely similar, and thus, of these genes, 434 were found to be differentially expressed by five-fold, with 251 upregulated and 183 downregulated in fetal blood To identify genes of interest, I then proceeded to use the Gene Ontology Browser to identify processes associate with these genes In particular, I chose to look for overlaps with ontologies associated with the development process which may be significant Thus, I identified 33 developmental processes that were upregulated in FB (summarized in Table 4-1) comprising 47 genes (Table 4-3) and 13 processes that were upregulated in UCB-EPC (Table 4-2), comprising 33 genes (Table 4-4)

Studying the processes which have been significantly upregulated in FB-EPC, a correlation with processes involved in blood vessel formation was found, including angiogenesis, blood vessel morphogenesis, vasculature development and blood vessel development, and thus supporting a possible greater vasculogenic potential of FB-EPC Six genes were found to be common to these four processes: Kinase insert domain receptor, Jagged 1, Placental Growth Factor, Interleukin 8, angiopoietic-like 4 and roundabout homolog 4

In contrast, UCB-EPC are largely involved directing cell differentiation, as well as processes involving regulation of cell growth and size The common genes involved

in regulation of cell growth and size are HtrA serine peptidase 1, insulin-like growth factor binding protein 4, suppressor of cytokine signalling 2, insulin-like growth factor binding protein 6, phospholipase C an dubiquitin conjugating enzyme E2E 3

129

Figure 4-7: Comparative gene expression of UCB-EPC versus FB-EPC

(a) Microarray analysis shows differentially regulated genes in UCB-EPC compared with FB-EPC (b) 434 genes demonstrate five-fold differences in gene expression

Category

Genes in Category

% of Genes in Category

Genes in List in Category

% of Genes in List in Category p-Value

GO:45596: negative regulation of cell differentiation 27 0.0907 3 1.63 6.09E-04

GO:45671: negative regulation of osteoclast differentiation 7 0.0235 2 1.087 0.000782

GO:45670: regulation of osteoclast differentiation 8 0.0269 2 1.087 0.00104

GO:48534: hemopoietic or lymphoid organ development 217 0.729 6 3.261 0.00237

GO:45656: negative regulation of monocyte differentiation 12 0.0403 2 1.087 0.00241

131

GO:45655: regulation of monocyte differentiation 15 0.0504 2 1.087 0.00378

Table 4-1: List of ontological processes upregulated in FB-EPC compared to UCB-EPC

Category

Genes in Category

% of Genes in Category

Genes in List in Category

% of Genes in List in Category p-Value

GO:45666: positive regulation of neuron differentiation 7 0.0235 2 1.538 0.000392

GO:45664: regulation of neuron differentiation 22 0.0739 2 1.538 0.00413

Table 4-2: List of ontological processes downregulated in FB-EPC compared to UCB-EPC

229800_at Doublecortin and CaM kinase-like 1

203881_s_at dystrophin (muscular dystrophy, Duchenne and Becker types)

208297_s_at ecotropic viral integration site 5

208394_x_at endothelial cell-specific molecule 1

203184_at fibrillin 2 (congenital contractural arachnodactyly)

1555480_a_at filamin binding LIM protein 1

218665_at frizzled homolog 4 (Drosophila)

227405_s_at frizzled homolog 8 (Drosophila)

230559_x_at FYVE, RhoGEF and PH domain containing 4

219901_at FYVE, RhoGEF and PH domain containing 6

209604_s_at GATA binding protein 3

204689_at hematopoietically expressed homeobox

201162_at insulin-like growth factor binding protein 7

211506_s_at interleukin 8

209098_s_at jagged 1 (Alagille syndrome)

201596_x_at keratin 18

203934_at kinase insert domain receptor (a type III receptor tyrosine kinase)

221841_s_at Kruppel-like factor 4 (gut)

204153_s_at manic fringe homolog (Drosophila)

202291_s_at matrix Gla protein

209086_x_at melanoma cell adhesion molecule

218678_at nestin

202149_at neural precursor cell expressed, developmentally down-regulated 9

204105_s_at neuronal cell adhesion molecule

218902_at Notch homolog 1, translocation-associated (Drosophila)

210809_s_at periostin, osteoblast specific factor

209652_s_at placental growth factor, vascular endothelial growth factor-related protein

1559400_s_at pregnancy-associated plasma protein A, pappalysin 1

205128_x_at

prostaglandin-endoperoxide synthase 1 (prostaglandin G/H synthase and cyclooxygenase)

235131_at ras homolog gene family, member J

226028_at roundabout homolog 4, magic roundabout (Drosophila)

217591_at SKI-like

202363_at sparc/osteonectin, cwcv and kazal-like domains proteoglycan (testican)

212558_at sprouty homolog 1, antagonist of FGF signaling (Drosophila)

201998_at ST6 beta-galactosamide alpha-2,6-sialyltranferase 1

206283_s_at T-cell acute lymphocytic leukemia 1

235086_at Thrombospondin 1

232068_s_at toll-like receptor 4

204653_at transcription factor AP-2 alpha (activating enhancer binding protein 2 alpha) 209909_s_at transforming growth factor, beta 2

Table 4-3: List of genes upregulated in FB-EPC compared to UCB-EPC

133

226777_at ADAM metallopeptidase domain 12 (meltrin alpha)

205609_at angiopoietin 1

214297_at Chondroitin sulfate proteoglycan 4 (melanoma-associated)

202436_s_at cytochrome P450, family 1, subfamily B, polypeptide 1

226281_at delta-notch-like EGF repeat-containing transmembrane

227646_at Early B-cell factor

206115_at early growth response 3

225079_at epithelial membrane protein 2

210310_s_at fibroblast growth factor 5

204451_at frizzled homolog 1 (Drosophila)

210220_at frizzled homolog 2 (Drosophila)

218469_at gremlin 1, cysteine knot superfamily, homolog (Xenopus laevis)

235521_at homeo box A3

228904_at Homeo box B3

201185_at HtrA serine peptidase 1

201508_at insulin-like growth factor binding protein 4

203851_at insulin-like growth factor binding protein 6

226534_at KIT ligand

210302_s_at mab-21-like 2 (C elegans)

228425_at Paired box gene 8

238852_at Paired related homeobox 1

205111_s_at phospholipase C, epsilon 1

225975_at protocadherin 18

232231_at runt-related transcription factor 2

202037_s_at secreted frizzled-related protein 1

206056_x_at sialophorin (gpL115, leukosialin, CD43)

228347_at Sine oculis homeobox homolog 1 (Drosophila)

202935_s_at SRY (sex determining region Y)-box 9 (campomelic dysplasia, autosomal sex-reversal) 222557_at stathmin-like 3

225665_at sterile alpha motif and leucine zipper containing kinase AZK

203372_s_at suppressor of cytokine signaling 2

232528_at Ubiquitin-conjugating enzyme E2E 3 (UBC4/5 homolog, yeast)

205990_s_at wingless-type MMTV integration site family, member 5A

Table 4-4: List of genes downregulated in FB-EPC compared to UCB-EPC

4.2.2.5 Flow cytometry Analysis of fetal liver cells

As the adhesion culture failed to generate endothelial progeny from fetal liver cells, I evaluated the liver mononuclear cell population using flow cytometry for markers associated with EPC The profile suggests that liver MNCs contains a significant population of cells expressing endothelial markers (VEGFR2: 0.35%, CD31: 7.64%, CD105: 6.34%) Significantly, a large proportion of cells were found to express haematopoietic stem cell marker, including CD34 (10.7%) and CD133 (2.14%) Similar to cord blood, the pan-haematopoietic marker CD45 was also highly expressed (88.8%) These results suggested that EPC, which stem from haematopoietic sources, do exist in fetal liver, and that enrichment had to be employed prior to culture to provide a selective advantage over highly proliferative fibroblasts

135

Endothelial markers VEGFR2 0.35 ± 0.27 CD31 7.74 ± 7.26 CD105 6.34 ± 2.04 Stem cell markers CD34 10.70 ± 3.26 CD133 2.14 ± 1.26 CD117 15.52 ± 10.84 Haematopoietic markers CD45 88.77 ± 5.93

Figure 4-8: Flow cytometric analysis of surface expression markers on fetal liver

4.2.2.6 Fetal liver selection and expansion

In light of the evidence to suggest the presence of EPC in the fetal liver, I then proceeded to study the possible selection of EPC using immunoselection CD133, a stem/progenitor cell marker, was chosen because of the reported ability to distinguish EPC from mature EC (Peichev 2000) Furthermore, CD133+ cells isolated from cord blood has been shown to capable of differentiation towards endothelial lineage, and for the construction of microvascular networks (Wu 2004)

Using enrichment by Magnetic Affinity Cell Sorting (MACS), purity was consistently above 80% as determined by flow cytometry Recovery rate, however, was highly variable, and the CD133+ cell yield was 1.7±2.2% of total mononuclear cells (n=7) Expansion in a cytokine-rich cocktail (detailed in 2.1.10.1) resulted in expansion of the putative stem cell population, with a doubling time of 96 hours

137

Figure 4-9: Fetal liver CD133+ cells

Following immunoselection by Magnetic Affinity Cell Sorting (MACS), >80% CD133+ enriched populations can be derived (Red: Isotype control, Blue: Liver MNC, Black CD133 enriched population) These cells expand in culture with a doubling

time of about 96 hours

4.2.2.7 Haematopoietic assays

Following expansion for one week, I transferred a portion of the cells to semi-solid cultures for haematopoietic colony forming assays as an assessment of the retention of haematopoietic character following expansion As seen in Figure 4-10, Colony forming assays of expanded FL133+ cells generated CFU-E (Colony Forming Unit Erythrocyte, 1.33/103 cells), BFU-E (Burst Forming Unit Erythrocyte, 3.00/103 cells), CFU-M (Colony Forming Unit Macrophage, 11.7/103 cells), CFU-G (Colony Forming Unit Granulocyte, 7.7/103 cells) and CFU-GEMM (Colony Forming Unit granulocyte, erythrocyte, macrophage, megakaryocyte, 6.67/103 cells) My results demonstrate that a range of haematopoietic colonies could be generated, suggesting that the cells continued to retain HSC character Of interest, cobblestone area forming colonies (CAFC) were formed, which is taken as an indicator of haematopoietic stem property (van Os 2008)

139

CFU

Count (colonies per 103 cells)

Figure 4-10: Haemopoietic assays on FL-EPC

Colony forming assays of expanded FL133+ cells generated haemopoietic colonies including (a) CFU-E (Colony Forming Unit Erythrocyte) (b) BFU-E (Burst Forming Unit Erythrocyte) (c) CFU-M (Colony Forming Unit Macrophage) and (d) CFU-GEMM (Colony Forming Unit granulocyte, erythrocyte, macrophage, megakaryocyte) (e) Cobblestone Area Forming Colonies (CAFC), associated with long term culture initiating colonies, were also found

4.2.2.8 Endothelial Differentiation

In parallel with the haematopoietic differentiation assay, I subjected a separate portion

of expanded CD133 cells to endothelial differentiation by removing Flt-3 ligand Lightly adherent colonies surrounding loose clusters of cells, reminiscent of blood islands, could be observed over a week Live staining with antibodies against CD31 suggests that the adherent cells are endothelial in nature This was further supported

by the ability of the cells to take up acLDL Such colonies, previously termed Hills, are associated with angioblasts and the adult haemangioblast (Hill 2003; Loges 2004) However, the cells were unable to undergo further proliferation and gradually diminished under extended culture

CFU-141

Figure 4-11: Endothelial differentiation of FL-EPC

Following transfer to endothelial differentiation medium, (a,b) tightly adherent colonies surrounding loose clusters of cells (CFU-Hill) emerged after a week

Adherent cells express (c) CD31 (green) and (d,e) take up acLDL (red)

(a)

4.2.2.9 Immunocytochemistry

To confirm the capability of the putative liver-derived EPC (FL-EPC) to generate endothelial progeny, I assayed for the expression of various endothelial markers, namely CD31, CD144, VEGFR2, as well as vWF The staining profile was compared with FB-EPC, as well as HUVE cells (Figure 4-12)

As seen in Figure 4-12, terminally differentiated EC such as HUVEC exhibit hexagonal morphology, and uniform staining of CD31 and CD144, localised at the gap junctions FB-EPC appeared elongated, and although CD31 and CD144 were similarly found to be expressed at gap junctions, the positive cells appeared in tight clusters, interspersed with negative cells Consistent with previous studies, endogenous VEGFR2 could be found in all three cell types distributed across the cell cytotosol, possibly on endosomes, as well as on the plasma membranes (Ewan 2006) Finally, vWF is found to be highly localised in granular ultrastructures corresponding

to Weibel-Palade bodies In the case of FB-EPC, although similar granular structures could be found, only a fraction of cells tested positive for vWF

In comparison, FL-EPC are rounder and smaller, with similarly well-defined staining

of gap junction proteins (CD31 and CD144) restricted to cell periphery I noted that VEGFR2 expression was largely localised at the cell membrane vWF staining pattern was also found to be different, appearing only in some cells, and lacked organisation within the cytosol