Histochemistry and Cell Biology

, Volume 131, Issue 2, pp 267–282

Isolation and characterization of Oct-4+/HLA-G+ mesenchymal stem cells from human umbilical cord matrix: differentiation potential and detection of new markers

Authors

    • Sezione di Anatomia Umana, Dipartimento di Medicina SperimentaleUniversità degli Studi di Palermo
  • Rita Anzalone
    • Sezione di Anatomia Umana, Dipartimento di Medicina SperimentaleUniversità degli Studi di Palermo
  • Simona Corrao
    • Sezione di Anatomia Umana, Dipartimento di Medicina SperimentaleUniversità degli Studi di Palermo
  • Francesca Magno
    • Sezione di Anatomia Umana, Dipartimento di Medicina SperimentaleUniversità degli Studi di Palermo
  • Tiziana Loria
    • Sezione di Anatomia Umana, Dipartimento di Medicina SperimentaleUniversità degli Studi di Palermo
  • Melania Lo Iacono
    • Sezione di Anatomia Umana, Dipartimento di Medicina SperimentaleUniversità degli Studi di Palermo
  • Antonino Di Stefano
    • Laboratorio di Citoimmunopatologia dell’ Apparato Cardio-Respiratorio, Fondazione S. Maugeri, IRCCSCentro Medico di Veruno
  • Pantaleo Giannuzzi
    • Divisione di Cardiologia, Fondazione S. Maugeri, IRCCSCentro Medico di Veruno
  • Lorenzo Marasà
    • Servizio di Anatomia Patologica, Ospedale Civico M. Ascoli
  • Francesco Cappello
    • Sezione di Anatomia Umana, Dipartimento di Medicina SperimentaleUniversità degli Studi di Palermo
  • Giovanni Zummo
    • Sezione di Anatomia Umana, Dipartimento di Medicina SperimentaleUniversità degli Studi di Palermo
  • Felicia Farina
    • Sezione di Anatomia Umana, Dipartimento di Medicina SperimentaleUniversità degli Studi di Palermo
Original Paper

DOI: 10.1007/s00418-008-0519-3

Cite this article as:
La Rocca, G., Anzalone, R., Corrao, S. et al. Histochem Cell Biol (2009) 131: 267. doi:10.1007/s00418-008-0519-3

Abstract

The presence of multipotent cells in several adult and embryo-related tissues opened new paths for their use in regenerative medicine. Extraembryonic tissues such as umbilical cord are considered a promising source of stem cells, potentially useful in therapy. The characterization of cells from the umbilical cord matrix (Wharton’s Jelly) and amniotic membrane revealed the presence of a population of mesenchymal-like cells, sharing a set of core-markers expressed by “mesenchymal stem cells”. Several reports enlightened the differentiation capabilities of these cells, even if at times the lack of an extensive characterization of surface markers and immune co-stimulators expression revealed hidden pitfalls when in vivo transplantation was performed. The present work describes a novel isolation protocol for obtaining mesenchymal stem cells from the umbilical cord matrix. These cells are clonogenic, retain long telomeres, can undergo several population doublings in vitro, and can be differentiated in mature mesenchymal tissues as bone and adipose. We describe for the first time that these cells, besides expressing all of the core-markers for mesenchymal stem cells, feature also the expression, at both protein and mRNA level, of tolerogenic molecules and markers of all the three main lineages, potentially important for both their differentiative potential as well as immunological features.

Keywords

Mesenchymal stem cellsUmbilical cord matrixDifferentiation protocolsTolerogenic propertiesSelf-renewal markers

Introduction

Several literature reports in recent years have enlightened the presence of undifferentiated cellular populations, collectively named as “adult stem cells”, in various adult tissues; in both human and animal models, (Young et al. 1995; Smith 2006; Tamaki et al. 2007; Zummo et al. 2007). The umbilical cord (UC) is an extraembryonic formation essential to provide feeding to the fetus during the intrauterine development. It has been shown that UC matrix (the Wharton’s Jelly surrounding umbilical vessels) contains a great number of fibroblastoid mesenchymal cells, which have been recently characterized as expressing markers, shared also by bone marrow mesenchymal stem cells (BM-MSC) in both human and animal models (Romanov et al. 2003; Miao et al. 2006; Csaki et al. 2007). The presence of such undifferentiated cells has been considered extremely promising for applications in regenerative medicine (Can and Karahuseyinoglu 2007), also in consideration of ethical and safety issues (embryo manipulation needed for sourcing, likelihood of tumor formation in recipient animals) that rose against the use of embryonic stem cells (ESC) (Baker et al. 2007).

In the literature, the “core” markers that are currently used to characterize mesenchymal stem cells (MSC) from different sources comprise CD44, CD73, CD90, and CD105, molecules for which a large consensus exists (Payushina et al. 2006; Chamberlain et al. 2007; Kolf et al. 2007), together with the absence of expression of endothelial/hematopoietic markers (CD31, CD34, CD45, CD79). Cells extracted from umbilical cord are able to propagate in vitro for several population doublings, therefore constituting a reliable source for applications requiring high-cell numbers. Moreover, several reports suggest that genomic mutations at the chromosomal level, which are commonly observed in ESC lines expanded in vitro (Rebuzzini et al. 2008), are not a feature of these cells, which on the contrary showed a higher chromosomal stability.

On the other hand, recent critical findings showed the limits of the real achievements that can be reached by in vivo applications of such cells. Often, the mere presence of a few “core markers”, together with an assay of the in vitro differentiation capability, has been considered sufficient to translate the application of such cellular populations to in vivo settings, with contrasting and often disappointing results (Chiavegato et al. 2007) with respect to encouraging in vitro experiments (De Coppi et al. 2007). Main issues emerging from in vivo settings regard two key properties of the “differentiated” stem cells implanted to the host: first, stem cells may differentiate towards different lineages, as seen for example in the heart, therefore giving only a limited contribution to myocyte regeneration (Franco et al. 2007); second, in some settings the transplanted cells elicited a strong immune response in the host, similar to the acute rejection occurring after organ transplantation, therefore addressing the problem of a correct characterization of cells in terms of immune molecules expression (Chiavegato et al. 2007).

The present paper illustrates the isolation and characterization of mesenchymal stem cells, named here as HEMSCs (human extraembryonic mesoderm stem cells), from the mucous tissue which constitutes Wharton’s Jelly of umbilical cord. The innovative findings of the present work reside in an extended characterization of markers, known to be expressed in mesenchymal stem cells, alongside with markers characteristic of embryonic stem cells (as molecules responsible for the maintenance of an undifferentiated state and molecules responsible for the self-renewal). In particular, Oct-1, Oct-4 and Nanog were for the first time identified in Wharton’s Jelly-derived cells obtained by non-enzymatic isolation procedures. Moreover, since several issues regarding their immunogenicity and the presence of T-cell co-stimulatory molecules have been raised for mesenchymal stem cells, despite their hypoimmunogenicity, the expression of such proteins has been evaluated (Ryan et al. 2005). The presence of normal type I MHC (HLA-A), and the lack of type II MHC (as HLA-DR) enhance the significance of present findings, namely the evidence of expression of HLA-G, demonstrated for the first time in umbilical-cord derived mesenchymal stem cells. These data suggest an immunosuppressive role for these cells mimicking the processes occurring in vivo at the fetus–maternal interface (Rabreau et al. 2000; Moffett and Loke 2004). In addition, we described for the first time the expression of cytokeratins 8, 18, 19 (and the absence of cytokeratin 7) in umbilical cord mesenchymal stem cells, isolated by non-enzymatic methods. This expressional pattern should characterize mesenchymal cells from human cord, with significant differences with the standard trophoblastic lineage (Haigh et al. 1999; Karahuseyinoglu et al. 2007). Moreover, the expression of early endodermal markers such as GATA-4 and HNF4α (hepatocyte nuclear factor 4α) should let hypothesize that these cells should undergo endoderm-specific differentiation. Moreover, we confirmed the expression of neuroectodermal markers as GFAP (glial fibrillar acidic protein) and NSE (neuron-specific enolase) also in undifferentiated cells, as described in earlier reports (Romanov et al. 2003; Mitchell et al. 2003). Collectively, present data indicate that HEMSCs obtained by non-enzymatic isolation from the umbilical cord may constitute a population of cells, which can reliably be expanded in vitro while maintaining their multipotency as well as the expression of immunomodulatory molecules. All of these properties strongly suggest the potential of these cells for their use in regenerative medicine applications.

Materials and methods

Tissue samples and cellular isolation protocol

All samples were obtained after mothers’ informed consent and treated in accordance with the tenets of the Declaration of Helsinki and local ethical regulations.

Umbilical cords (n = 24) were obtained immediately after full-term births, after normal vaginal delivery or cesarean section. Umbilical cords were stored aseptically in cold saline and then cellular isolation started within 6 h from partum. To perform cellular isolation, the cord was rinsed in warm HBSS (Gibco), then cut into small pieces (about 1.5-cm length), which were sectioned longitudinally to expose the Wharton jelly under the amniotic membrane. Vessels were not removed from the matrix, and some incisions were made on the inner matrix with a sterile scalpel to expose a wider area of tissue to the contact with culture medium. Then the cord sections were transferred to 6-well plates (Corning), one cord piece for each well, and covered completely with culture medium. The isolation and subculture of cells were made using DMEM low-glucose (Sigma), supplemented with 10% FCS (fetal calf serum, PAA), 1× NEAA (non-essential aminoacids, Sigma), 1× antibiotics–antimycotics (Gibco), and 2 mM l-glutamine (Sigma). The isolation method made no use of proteases to detach cells from the embedding matrix. Therefore, based on the “mesenchymal” migratory capability of cells, cord fragments were left in the culture medium for 15 days, with medium change every second day. Cellular exit from the cord and attachment to the plastic surface of the tissue culture slide was monitored by phase-contrast microscopy. Finally, after 15 days of culture, the remnants of the cord fragments (usually the intact vessels surrounded by a scarcer matrix) were removed from the wells, and cells attached to wells were cultured until reaching the confluence.

Cell culturing and passaging

After reaching confluence, primary cells were subcultured routinely in culture medium. Cellular detachment from tissue-culture flasks has been performed using Accutase (Invitrogen) instead of standard trypsin solution. Primary populations of HEMSCs were cultivated for up to 15 passages (corresponding to about 60 population doublings). In order to establish freezing conditions for long-term storage of cells, complete medium with 10% DMSO showed the best results as deep-freezing medium in terms of replating efficiency and cellular survival. For characterization experiments described below, cells at different passages were used in order to ensure the maintenance of markers expression along the population doublings of cells. For immunocytochemical analyses, cells were plated in 8-well chamber slides (BD Biosciences) and were subjected to immunochemistry when they reached 90% confluence. For DNA and RNA extraction, cells were cultured either in 6-well tissue culture plates or in 25-cm2 tissue culture flasks (Corning).

Immunocytochemical analysis

After culturing, cells grown in chamber slides were washed with PBS and fixed in methanol for 20 min at −20°C. Air-dried slides were then stored at −20°C until use. For the immunocytochemical procedure, cells were permeabilized with 0.1% TritonX-100 in PBS (Sigma). After a subsequent rinse with PBS, slides were exposed to 0.3% H2O2 in PBS, were then blocked with 1% FBS in PBS, and were incubated for 2 h with the primary antibody. The detection was performed using an avidin–biotin complex kit (LSAB2, DAKO); 3.3′- diaminobenzidine (DAB chromogenic substrate solution, DAKO) was used as developer. Nuclear counterstaining was obtained using hematoxylin (DAKO). The primary antibodies used for this study are listed in Table 1. Immunopositivity was scored using a semiquantitative approach. Three independent observers (FC, LM, GLR) evaluated the immunocytochemical results and semi quantified the percentage of positive cells for each specimen. Ten high-power fields were examined in each culture slide and counting of the cells was performed at 40× magnification.
Table 1

List of antibodies used in the present study

Antigen

Host

Manufacturer

Dilution

CD10

Mouse

Dako

1:50

CD13

Mouse

Chemicon

1:50

CD31

Mouse

Dako

Pre-diluted

CD33

Goat

Santa Cruz

1:50

CD34

Mouse

Dako

Pre-diluted

CD38

Goat

Santa Cruz

1:100

CD45

Mouse

Dako

Pre-diluted

CD79

Mouse

Dako

1:100

CD117

Rabbit

Stressgen

1:200

CEA

Mouse

Dako

Pre-diluted

Cytokeratin 7

Mouse

Dako

1:100

Cytokeratin 8

Mouse

Sigma

1:200

Cytokeratin 18

Mouse

Sigma

1:800

Cytokeratin 19

Mouse

Chemicon

1:100

Connexin-43

Mouse

BD Laboratories

1:50

GATA-4

Mouse

Santa Cruz

1:200

GFAP

Mouse

Dako

1:500

MyoD

Mouse

NeoMarkers

1:100

Myosin (smooth muscle)

Mouse

Sigma

1:500

α-SMA

Mouse

Dako

Pre-diluted

Nestin

Mouse

BD Laboratories

1:50

NSE

Mouse

Dako

Pre-diluted

Prol-4-Hydroxylase (5b5)

Mouse

Dako

1:50

Vimentin

Mouse

Santa Cruz

1:100

Von Willebrand factor

Mouse

Dako

1:50

The antibodies used in the present study, with indications of the working conditions used, are listed in Table 1.

Total RNA extraction

Total cellular RNA was isolated using the Quick Prep Total RNA Extraction Kit (GE Healthcare) as described previously (La Rocca et al. 2007). RNA yield was evaluated spectrophotometrically (A260/280) and RNA aliquots were stored at −80°C until use. Total RNA fractions were used for subsequent experiments only if the A260/A280 ratio exceeded 1.7.

RT-PCR analysis

Qualitative RT-PCR was performed using the JumpStart RED HT RT-PCR kit (Sigma). RT-PCR was performed by mixing 2 μg of total RNA and 0.5 μg of pd(T)23, with RNAse free water. Tubes were placed in thermal cycler at 70°C for 10 min. The reaction comprised a reverse transcription step of 50 min (42°C), followed by inactivation of the enzyme at 95°C (5 min). Then 10 pM of specific primers were added and the reactions were cycled for 94°C/2 min, then 35 cycles of 94°C/15 s, 60°C/30s, 72°C/60s, with a final extension at 72°C/10 min.
Table 2

List of PCR primers used for the present study

Name

Accession number

TM (°)

Product size

Forward primer

Reverse primer

Actin, beta

NM_001101

60

350

5′-AAACTGGAACGGTGAAGGTG-3′

5′-TCAAGTTGGGGGACAAAAAG-3′

Actin alpha2

NM_001613

59

321

5′-TGATCACCATCGGAAATGAA-3′

5′-GCTGGAAGGTGGAAATGAA-3′

CD29

NM_033666

59

186

5′-CTGATTGGCTGGAGGAATGT-3′

5′-TTTCTGGACAAGGTGAGCAA-3′

CD31

NM_000442

59

309

5′-CCATGCACCCTCATACACAG-3′

5′-CTGTGCTTGTTCCACCTTCA-3′

CD34

NM_001773

59

367

5′-AAAACGTGTTGCCTTGAACC-3′

5′-AAGCCATGGAGATCAGAGGA-3′

CD44

NM_00610

60

282

5′-TCTCAAGGGCGTAACTCTGG-3′

5′-GCCAATTCTACCAGGCTTGA-3′

CD73

NM_002526

59

308

5′-CCTGCTCAGCTCTGCATAAGTA-3′

5′-CCCTATTTTACTGGCCAAGTGT-3′

CD80

NM_005191

59

259

5′-AGGGCCTCCTTAGATCCCTA-3′

5′-TTAGCTGCCATGAGATGTGC-3′

CD86

NM_175862

59

250

5′-TCCTGGCTGAGAGAGGAAGA-3′

5′-AGACTGCCCCATCCCTTAGT-3′

CD90

NM_006288

58

265

5′-TTTGGCCCAAGTTTCTAAGG-3′

5′-AGATGCCATAAGCTGTGGTG-3′

CD105

NM_000118

55

179

5′-TCCAGCACTGGTGAACTGAG-3′

5′-TGTCTCCCCTGCCAGTTAGT-3′

CD106

NM_001844

57

340

5′-TGGAGGAGTTCCTTGATCTG-3′

5′-CTGAAAGTCAACCCAGTGCT-3′

CD117

NM_000222

60

268

5′-ACTTCAGGGGCACTTCATTG-3′

5′-ACGTGGAACACCAACATCCT-3′

CD133

NM_006017

59

255

5′-GCATGCAAAAGCCATCATAG-3′

5′-ATCCATGCTGGACACCAGA-3′

CD166

NM_001627

59

283

5′-TGGTGTGGGAGATCAAAGGT-3′

5′-TGTGGCTGCCATTAAACAAG-3′

ERas

NM_181352

59

315

5′-GCAAGGTCCTGTAGGGAGAA-3′

5′-GCAGCTTTGAAACCCAAAAC-3′

GATA-4

NM_002052

59

270

5′-CCAGAGATTCTGCAACACGA-3′

5′-ATTTTGGAGTGAGGGGTCTG-3′

GATA-5

NM_080473

59

259

5′-GAATGGCCGGTGATGTATGT-3′

5′-TGAAGCTGATGCCAGACAAC-3′

GATA-6

NM_005257

59

259

5′-ACTAACCCACAGGCAGGTTG-3′

5′-GGTACAAAACGGCTCCAAAA-3′

HLA-A

NM_002116

59

262

5′-TGGGACTGAGAGGCAAGAGT-3′

5′-ACAGCTCAGTGCACCATGAA-3′

HLA-DR-B1

NM_002124

59

349

5′-GCACAGAGCAAGATGCTGAG-3′

5′-AGTTGAAGATGAGGCGCTGT-3′

HLA-G

NM_002127

57

358

5′-TTAAAGTGTCACCCCTCACTG-3′

5′-CCCATCAATCTCTCTTGGAA-3′

HNF4α

NM_178850

59

238

5′-CGAGCAGATCCAGTTCATCA-3′

5′-TTCCCATTTTTCTGGTGAGG-3′

Cytokeratin 8

NM_002273

59

216

5′-TCTGGGATGCAGAACATGAG-3′

5′-AGACACCAGCTTCCCATCAC-3′

Cytokeratin 18

NM_000224

59

263

5′-CTGCTGCACCTTGAGTCAGA-3′

5′-GTCCAAGGCATCACCAAGAT-3′

Cytokeratin 19

NM_002276

59

295

5′-ATGAAAGCTGCCTTGGAAGA-3′

5′-CCTCCAAAGGACAGCAGAAG-3′

Leptin

NM_000230

59

193

5′-CCAGATCCTCACAACCACCT-3′

5′-CTCCCAAAGTGCTGGGATTA-3′

Adiponectin

NM_004797

59

164

5′-GCTGGAGTTCAGTGGTGTGA-3′

5′-ACCAACCTGACGAATGTGGT-3′

Nanog

NM_024865

59

209

5′-CTCCATGAACATGCAACCTG-3′

5′-CTCGCTGATTAGGCTCCAAC-3′

Nestin

NM_006617

60

275

5′-TATAACCTCCCACCCTGCAA-3′

5′-AGTGCCGTCACCTCCATTAG-3′

Oct-1

NM_002697

59

297

5′-GCAACCCTGTTAGCTTGGTC-3′

5′-CTCTCCTTTGCCCTCACAAC-3′

Oct-2

NM_002698

59

286

5′-AGGCCTCAGCGTTCTCTTTT-3′

5′-TGCCAGTCCCTTCTCTCTTC-3′

Oct-4 A

NM_002701

60

273

5′-AGTGAGAGGCAACCTGGAGA-3′

5′-GTGAAGTGAGGGCTCCCATA-3′

Oct-4 B

NM_203289

60

194

5′-TATGGGAGCCCTCACTTCAC-3′

5′-CAAAAACCCTGGCACAAACT-3′

Osteonectin

NM_003118

60

296

5′-TGATGATGGTGCAGAGGAAA-3′

5′-GGGGGATGTATTTGCAAGG-3′

Periostin

NM_006475

59

185

5′-TGGAGTTAGCCTCCTGTGGT-3′

5′-ACAAGGCTCGGTCTTTTCAA-3′

Vimentin

NM_003380

60

345

5′-AGATGGCCCTTGACATTGAG-3′

5′-TCTTGCGCTCCTGAAAAACT-3′

vWF

NM_000552

59

317

5′-GGGGTCATCTCTGGATTCAA-3′

5′-CAGGTGCCTGGAATTTCAT-3′

The primer pairs used in this study, together with annealing temperatures and product sizes, are listed in Table 2.

DNA Extraction

DNA extraction was performed using the Nucleon HT kit (GE Healthcare) following the manufacturer’s instructions. RNAse-treated samples were used for subsequent telomere length assay.

Telomere length assay

Length of telomere regions of genomic DNA was assessed on DNA from cells at different passages, using the Telo TAGGG kit (Roche) following the manufacturer’s instructions. Appropriate controls (DNA extracted from cells with long or short telomere regions) were provided with the kit.

Preparation of cells for karyotype analysis

HEMSCs were incubated with Colcemid (100 ng/ml, Roche) for 20 min. Then the cells were detatched from flasks with accutase (Invitrogen) and centrifuged. The cell pellet was gently resuspended in a hypotonic solution (0.075 M KCl), and fixed in fixative solution (3:1 v/v methanol:glacial acetic acid). Fixed cells were dropped on clean wet slides, air-dried and then stained with Giemsa stain (KaryoMax, Gibco).

G-Banded chromosomes analysis

A minimum of 30 metaphases were fully karyotyped at the microscope (Leica DM5000) equipped with 20× and 100× objectives. Images were captured and analysed using the software karyotyping package Leica CW4000 Karyo.

Clonogenicity assays

In order to give a formal demonstration of the self-renewal capability of HEMSC, a limiting dilution method was applied (Weiss et al. 2006). Briefly, cells at different culture passages were plated as single cells into each well of a 96-well plate. The addition of a single cell per well was confirmed by phase contrast microscopy. After 2 weeks in culture, with medium change each second day, the presence of clones was assessed by phase-contrast microscopy.

Osteogenic differentiation

Differentiation of cells was performed by culturing HEMSCs at different passages in osteogenic differentiation medium, as reported previously (Zuk et al. 2002). Briefly, culture medium was supplemented with 50 μM ascorbate-2-phosphate (Sigma), 10 mM β-glycerophosphate (Sigma), 0.1 μM dexamethasone (Sigma), 10% FCS (PAA), and 1% antibiotic/antimycotic (Gibco). Cells were cultured in six-well tissue culture plates for 3 weeks and medium was replaced every second day. The formation of cell clusters resembling intramembranous ossification was monitored by phase-contrast microscopy along culturing. Controls included HEMSCs cultured in normal growth medium for 3 weeks to monitor the eventual spontaneous formation of bone-like nodules.

Adipogenic differentiation

Differentiation of cells was performed by culturing HEMSCs at different passages in adipogenic differentiation medium, as reported previously (Zuk et al. 2002). Culture medium was supplemented with 0.5 mM isobutyl-methylxanthine (Sigma), 1 μM dexamethasone (Sigma), 10 μM insulin (Sigma), 200 μM indomethacin (Sigma), 10% FCS (PAA), and 1% antibiotic-antimycotic (Gibco). Cells were cultured in six-well tissue culture plates for 3 weeks, and medium was replaced every second day. The formation of cytoplasmic lipid vacuoles was monitored by phase-contrast microscopy along culturing. Controls included HEMSCs cultured in standard growth medium for 3 weeks to monitor the spontaneous formation of lipid vacuoles.

Histochemical staining

To demonstrate the acquisition of the osteogenic phenotype, the Alizarin Red S staining was performed. Briefly, cells were fixed in 4% paraformaldehyde and stained with 1% solution of Alizarin Red S (Sigma). Stained cells were rinsed with water for three times to remove excess stain, and then photographed at the photomicroscope.

Alkaline phosphatase staining (Sigma) was performed using the manufacturer instructions. This stain indirectly measures alkaline phosphatase activity, which is enhanced in osteoblast-like cells.

To demonstrate the adipogenic differentiation, cells were stained with Oil Red O (Sigma), and photographed at the photomicroscope. After medium aspiration, a brief wash was performed with PBS. Cells were fixed with 10% formalin (Sigma) for 30 min at room temperature, followed by subsequent washes with distilled water and 60% isopropanol. Oil Red O working solution was added to the cells for 5 min, followed by four washes (5 min each) with distilled water. The wells were viewed and photographed using an inverted phase-contrast microscope. Following a further step of counterstaining (Meyer’s hematoxylin, 1 minute), lipid vacuoles appeared red and nuclei appeared blue.

Results

Cellular isolation and culturing

The cellular isolation protocol described in “Materials and methods” allowed reproducibly the isolation, from each cord specimen, of fibroblastoid cells, which adhered on plastic culture vessels, and were successfully subcultured giving cells at various culture passages. Figure 1 shows a panel of representative phase-contrast micrographs of HEMSCs at different culture passages. Panels are representative of what were obtained with different cell lines from independent cord samples. Primary cultured cells (Fig. 1a; passage 4) at the first passages, show a mesenchymal morphology; with processes extending between adjacent cells and the presence of cell–cell contacts. Confluent cells (Fig. 1b; passage 4) are arranged in parallel arrays. Cells at higher culture passages (Fig. 1c, d; passage 10) retained the standard morphology and steadily grew in the standard culture medium. Freezing, storage in liquid nitrogen and defrost of cultured cells were routinely performed at different passages in order to demonstrate their ability to survive deep freezing and therefore their long-term storability (not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-008-0519-3/MediaObjects/418_2008_519_Fig1_HTML.jpg
Fig. 1

Light microscopic micrographs of HEMSCs in monolayer culture at different culture passages. In monolayer culture, the cells assumed a polymorphic, fibrobroblast-like morphology, which was maintained throughout the passaging process. a Cells at passage 1, prior to reach confluence. b Confluent cells at passage 1. c Confluent monolayer at passage 6. d Confluent cells at passage 10. Magnification ×10

Immunocytochemistry of mesenchymal stem cells markers

The characterization of markers expressed by HEMSCs was initially performed by immunocytochemical analysis. Experiments were repeated in triplicate on cell lines obtained from different cord specimens in order to assess the reproducibility of the results. Figure 2 shows the representative results obtained by probing the cells for the expression at the protein level of a panel of key markers used to characterize cells. Present data show that, besides mesodermal markers as vimentin (Fig. 2a) and alpha-smooth muscle actin (Fig. 2e), HEMSCs expressed c-Kit membrane antigen (CD117, Fig. 2b), which acts as receptor for stem cells factor (SCF) in hematopoietic stem cells (HSC). CD117 has been characterized as a key marker in HSC, but its expression has been reported in several populations of adult- and fetal-derived mesenchymal stem cells. Nestin expression (Fig. 2c) in the vast majority of cells is a particularly significant finding. In fact, this molecule represents an intermediate filament, characterized in the neuroectodermic lineage as precursor of neurofilament proteins. Its expression in mesenchymal stem cells has been recognized by different authors, even if this molecule is not yet a key marker of these cells. Immunocytochemical analysis for expression of GATA-4 (Fig. 2f) showed an intense perinuclear positivity in HEMSCs. This transcription factor is involved in several differentiative pathways, as the cardiomyocyte one, where it is responsible of the transactivation of promoters specific for the myocardial lineage, as indicated for ANF (atrial natriuretic factor), BNP (brain natriuretic peptide), and the cardiac isoforms of troponin (Peterkin et al. 2005). The perinuclear localization of this molecule is consistent with the undifferentiated phenotype of HEMSCs, where, in the absence of a specific differentiation stimulus, the transcription factor is expressed but is maintained in inactive form outside the nucleus. Finally, another marker expressed by the vast majority of cells is connexin-43 (Fig. 2d), one of the isoforms of this protein is known to be expressed at high levels in embryonic cells as well as in myocardial cells. Its expression is indicative of the capability of HEMSC to form gap junctions. This is the first time that a molecule belonging to the connexins family is described in cells derived from Wharton’s jelly. The potential usefulness of some of these markers to predict differentiation along myocytes is an intriguing hypothesis which deserves much work to be demonstrated. Positivity to alpha-smooth muscle actin has been also assessed in HEMSCs (Fig. 2e). The expression of this molecule has previously been reported for various MSC populations (Romanov et al. 2003), even if there is no general consensus to consider it as a MSC-specific marker.
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-008-0519-3/MediaObjects/418_2008_519_Fig2_HTML.jpg
Fig. 2

Representative panels of immunocytochemical detection of mesenchymal markers on HEMSCs. The isolated cells showed a strong positive signal for vimentin (a), CD117 (c-Kit, b), Nestin (c), and Connexin-43 (d). Smooth-muscle specific actin was expressed by most cells (e), while the early endoderm marker GATA4 was amply expressed by cultured cells, with clear perinuclear positivity (f). Magnification ×20

Table 3 shows the global results of the ICC analysis of HEMSCs. The cells did not show expression of endothelial-specific markers (CD31, vWF) or hematopoietic specific ones (CD33, CD34, CD38, CD45, CD79). These data are in good agreement with those reported in literature regarding the expression pattern of mesenchymal stem cells. On the contrary, HEMSC showed the expression of MSC markers as CD10 and CD13, while lacking the expression of the carcino-embryonic antigen (CEA), an adhesion molecule whose expression is normally lost after the embryonic development and should be restored in neoplastic cells. A new and potentially interesting observation is the positivity of more than 50% of cells for the expression of cytokeratins (CK) (namely CK8, CK18, CK19). These epithelial markers neither have been characterized previously in mesenchymal stem cells, nor are normally expressed in mesenchymal-derived cytotypes. Moreover, their presence in undifferentiated MSCs should be taken in account when an epithelial differentiation protocol is being performed. On the contrary, the expression of trophoblast-specific cytokeratin 7 has not been detected in HEMSCs.
Table 3

Immunocytochemistry results of markers expression by HEMSCs

CD10

+

CD13

+

CD31

CD33

CD34

CD38

CD45

CD79

CD117

+++

CEA

Cytokeratin 7

Cytokeratin 8

++

Cytokeratin 18

++

Cytokeratin 19

+

Connexin-43

+++

GATA-4

+++

GFAP

+

MyoD

Myosin (smooth muscle)

α-SMA

+++

Nestin

+++

NSE

+

Prolyl-4-Hydroxylase

+++

Vimentin

+++

Von Willebrand factor

Results of the immunocytochemical analysis are represented semiquantitatively

Undifferentiated cells did not show the expression of mature myocyte markers, namely the MyoD transcription factor and the smooth-muscle specific isoform of myosin. Interestingly, Table 3 shows that besides Nestin, which is a marker of neural precursor cells, undifferentiated HEMSCs showed the expression of neuroectodermal markers NSE and GFAP, confirming precedent results obtained in UC-derived cells in both humans and animals (Romanov et al. 2003; Mitchell et al. 2003).

Since a number of reports described the capability of umbilical cord-derived MSCs to undergo neural differentiation towards both mature neurons and glial cells, present evidence should suggest this capability by HEMSCs. Further experiments, going beyond the scope of the present paper, should be performed to verify this hypothesis.

The expression of another mesenchymal marker, namely prolyl-4-hydroxylase, a fibroblast marker, strongly suggests the presence of an active collagen production by these cells.

Further experiments have shown that alongside with passaging, cultured cells were morphologically and phenotypically similar to the parental cells (not shown).

Qualitative RT-PCR analysis of gene expression

In order to confirm and extend the data obtained by immunocytochemical analysis, we performed total RNA extraction, followed by retrotranscription and specific amplification of gene fragments by qualitative RT-PCR. Figure 3 shows a representative electrophoretic gel of HEMSCs gene expression pattern. As visible, the three GATA (-4,-5,-6) factors are all expressed by the cord cells. The expression of GATA-4 at the RNA level confirms the datum obtained by immunocytochemistry. All the three GATAs are involved in the developmental processes of several organs, as demonstrated for heart (Peterkin et al. 2005). Oct transcription factors bind the octameric regions of DNA and Oct-4 is involved, in ESC, in the maintenance of self-renewal of cells. We demonstrated for the first time in cord-derived mesenchymal stem cells, the presence of both isoforms of Oct-4 (A and B); a biologically relevant datum in the light of recent findings on the different contributions of the two factors to the self-renewal process. In fact, Lee and coauthors recently demonstrated that the isoform A is mainly responsible of self-renewal of human stem cells (Lee et al. 2006). The transcription factor Oct-1, expressed in ESC and fetal liver cells (Katoh and Katoh 2007) was expressed by HEMSCs, while the absence of Oct-2 mRNA was not surprising, since the expression of this factor is restricted to the lymphoid lineage (Pfisterer et al. 1997; Emslie et al. 2008). As shown in Fig. 3, RT-PCR experiments allowed also assessing the expression of other “core markers” of human MSCs: CD29, CD44, CD73, CD90, CD105.
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-008-0519-3/MediaObjects/418_2008_519_Fig3_HTML.gif
Fig. 3

Representative panel of RT-PCR analysis of HEMSCs. Cultured cells showed the expression of markers of mesenchymal stem cells (CD29, CD44, CD73, CD90, CD105), markers of embryonic stem cells (Oct-1, Oct-4, Nanog), and transcription factors involved in the development of several mesoderm- and endoderm-derived organs (GATA-4, -5, -6). Oct-2 expression, restricted to the lymphoid lineage, was not detected. M 50 bp ladder

Table 4 summarizes the results of the RT-PCR analysis of HEMSCs. As shown, we first extended the analysis for MSC markers, assessing the expression of CD106, CD117, CD133, CD166. Moreover, we investigated the presence of molecules involved in the immune recognition of MSCs by the immune system of the potential host. In fact, several authors reported that in “in vivo” settings the host response towards differentiated stem cells was related to the expression of B7 costimulatory molecules as CD80 (B7-1) and CD86 (B7-2). We determined by RT-PCR (Table 2) that HEMSC did not express CD86, while expressing CD80. This characteristic should be of key importance in the instauration of a tolerogenic response of the host to avoid transplant rejection (Sansom et al. 2003; Zhu et al. 2005). Moreover, further extending the importance of the data, we determined that HEMSCs are negative for type II MHC (HLA-DR), harbor normal class I MHC (HLA-A) and express also the non-canonical class I MHC HLA-G (Table 4). This is the first time that HLA-G expression is demonstrated in a MSC cell line, both for adult- and fetal-derived cells. This particular form of class I MHC is expressed by trophoblast and placental cells when the embryo establishes its contact with the uterus, and it is thought to induce an immune tolerance versus the embryo implantation process (Rouas-Freiss et al. 1997; Le Bouteiller 2000; Blaschitz et al. 2001). In addition, Table 4 shows that HEMSCs did not express ERas, an embryonic form of the Ras oncogene which codes for a constitutively active Ras protein. The expression of this factor has been correlated to tumorigenicity of embryonic stem cells, and it has been found to be expressed in several tumor cell lines (Takahashi et al. 2003; Yasuda et al. 2007).
Table 4

Expression of markers of different lineages by HEMSCs assessed by RT-PCR

Name

Accession number

Expression in HEMSC

Actin, beta

NM_001101

+

Actin alpha2

NM_001613

+ (*)

CD29

NM_033666

+

CD31

NM_000442

− (*)

CD34

NM_001773

− (*)

CD44

NM_00610

+

CD73

NM_002526

+

CD80

NM_005191

+

CD86

NM_175862

CD90

NM_006288

+

CD105

NM_000118

+

CD106

NM_001844

+

CD117

NM_000222

+ (*)

CD133

NM_006017

+

CD166

NM_001627

+

ERas

NM_181352

GATA-4

NM_002052

+ (*)

GATA-5

NM_080473

+

GATA-6

NM_005257

+

HLA-A

NM_002116

+

HLA-DR-B1

NM_002124

HLA-G

NM_002127

+

HNF 4α

NM_178850

+

Cytokeratin 8

NM_002273

+ (*)

Cytokeratin 18

NM_000224

+ (*)

Cytokeratin 19

NM_002276

+ (*)

Nanog

NM_024865

+

Nestin

NM_006617

+ (*)

Oct-1

NM_002697

+

Oct-2

NM_002698

Oct-4 A

NM_002701

+

Oct-4 B

NM_203289

+

Vimentin

NM_003380

+ (*)

vWF

NM_000552

− (*)

Qualitative RT-PCR analyses allowed determining the expression of key markers of mesenchymal stem cells together with molecules responsible of the immunologic features of these cells

Asterisks (*) indicate results cross-confirmed at the protein level by immunolocalization analyses

RT-PCR experiments allowed confirming the expression of Nestin also at the mRNA level, and allowed detecting of the expression of Nanog, another “embryonic” marker which is involved in the maintenance of an undifferentiated phenotype in ESC cultures.

The expression of the endoderm-restricted marker HNF4α suggests that HEMSCs obtained by non-enzymatic isolation protocol should undergo hepatocyte differentiation, as demonstrated for other MSC populations. This hypothesis is currently being investigated in our laboratories.

Telomere length assay

In order to demonstrate the ability of HEMSCs to undergo a high number of cellular divisions, making them promising candidates for cell therapy in regenerative medicine (where billions of cells should be needed for appropriate regeneration), we analyzed the mean length of telomeric ends in genomic DNA extracted from HEMSCs at different culture passages. As shown in Fig. 4, DNA extracted from cells at both passage 6 (corresponding to about 24 population doublings) and passage 15 (corresponding to about 60 population doublings) showed the presence of long telomeric ends, consistent with the band observed for the positive control (lane H, control DNA provided with the kit).
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-008-0519-3/MediaObjects/418_2008_519_Fig4_HTML.gif
Fig. 4

Telomere length assay on HEMSCs at different culture passages. The maintenance of long telomeres is a key feature of stem cells, ensuring the capability to undergo several cell cycles of replication. Genomic DNA processed with appropriate restriction enzymes was electrophoresed on a 0.8% agarose gel and hybridized with a specific probe for telomeric ends (digoxigenin-conjugated). The relative length of telomeric ends was revealed using an anti-digoxigenin antibody revealed by a chemiluminescence system. M Molecular weight marker, L control DNA with short telomeres, H control DNA with long telomeres, p6 DNA extracted from HEMSCs at passage 6 (corresponding to 24 population doublings), showing long telomeric ends and p15 DNA extracted from HEMSCs at passage 15 (corresponding to 60 population doublings) showing long telomeric ends

Karyotype analysis

The acquisition of chromosomal mutations, ranging from translocation of an arm of a chromosome to another, to frank aneuploidy phenotypes, is a concern which has been raised in different studies on embryonic stem cells (Rao 2006; Mantel et al. 2007; Rebuzzini et al. 2008). Therefore, to ensure the maintenance of a regular karyotype along the culture passages, we performed karyotype analysis using G-banding staining and software-assisted process. Briefly, metaphases not less than 30 were karyotyped for each cell line, at a number of passages ranging between 6 and 13. Figure 5 shows the results of a typical experiment of karyotyping. The presence of a normal karyotype suggests that HEMSCs, alongside with the passaging steps of cell culture, do not acquire chromosomal mutations.
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-008-0519-3/MediaObjects/418_2008_519_Fig5_HTML.gif
Fig. 5

Representative karyogram of HEMSCs showing a normal karyotype

Clonogenicity assays

The ability to generate clones, e.g., a cellular colony derived from a single cell, is a formal demonstration of the self-renewal ability, a characteristic of stem cell populations. Our gene expression data, suggesting expression of key molecules involved in self-renewal were thus confirmed by performing a clonogenicity assay. Briefly, cells at different culture passages were seeded in 96-well plates with an appropriate dilution to deliver single cell per well, as stated in “Materials and methods”. After 15 days of culture, colonies were counted, resulting in a cloning efficiency of 10–12% (not shown), a result which is in good agreement with previous observations made by other groups on umbilical cord stem cells (Weiss et al. 2006). Clonal lines were not further characterized.

Osteogenic differentiation of HEMSC cells

The definitive demonstration of “stemness” for a cell line is the ability to differentiate towards more than one mature cell type.

In order to demonstrate the capability to differentiate towards osteoblasts, HEMSCs were cultured for 3 weeks in osteogenic medium. Following culture, fixed cells underwent alkaline phosphatase staining.

As visible in Fig. 6a, untreated cells, cultured in standard medium for 3 weeks, retained the normal fibroblastoid morphology, and were negative for alkaline phosphatase activity. On the contrary (Fig. 6b), osteogenic-induced HEMSCs formed bone nodules, which resulted positive to the specific staining indicating alkaline phosphatase activity.
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-008-0519-3/MediaObjects/418_2008_519_Fig6_HTML.jpg
Fig. 6

Light microscopic demonstration of bone tissue formation by HEMSCs following osteogenic differentiation. When cultured in osteogenic medium, cellular morphology changed from a fibroblastic appearance (a, c) to a more cuboidal shape (b, d), in addition cells were surrounded with an abundant matrix and formed mineralized nodules. Alkaline phosphatase activity (demonstrated by a red staining) was amply positive in cells subjected to osteogenic differentiation (b) with respect to control cells, cultured for the same time in standard growth medium (a). The formation and deposit of a mineralized matrix in differentiated osteoblast-like cells (d) has been assessed by Alizarin Red S staining. Control cells, cultured in standard medium did not show any staining (c). Magnification ×20

The formation of extracellular calcium deposits has been assessed also by alizarin red S staining: differentiated HEMSCs were extensively stained red (Fig. 6d), while control cells were not stained (Fig. 6c).

Moreover, the acquisition of an osteoblast-like phenotype was assessed by RT-PCR. Figure 7 shows that while in control cells, cultured for 3 weeks in standard medium, the osteoblast-specific genes (periostin and osteonectin) were not expressed, when cells underwent differentiation, the two RNAs become detectable, thus confirming data from histochemical staining.
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-008-0519-3/MediaObjects/418_2008_519_Fig7_HTML.gif
Fig. 7

RT-PCR analysis of the expression of specific markers following osteogenic differentiation of HEMSCs. Following the osteogenic differentiation protocol, cells became positive for the expression of specific markers as osteonectin and periostin (lanesO), which were undetectable in control cells (lanesC). Beta actin was used as housekeeping control gene

Adipogenic differentiation of HEMSC cells

HEMSCs were cultured for 3 weeks either in adipogenic medium or in standard culture medium (control cells) prior to Oil Red Ostaining, a histochemical staining specific for lipids. As visible in Fig. 8, control cells do not show any staining for lipid vacuoles (Fig. 8a, e). On the contrary, cells cultured for 3 weeks in adipogenic medium show the presence of multiple red vacuoles, resembling multivacuolar adipocytes (Fig. 8b–d, f).
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-008-0519-3/MediaObjects/418_2008_519_Fig8_HTML.jpg
Fig. 8

Light microscopic demonstration of adipose tissue formation with Oil Red O staining. HEMSCs cultured for 3 weeks in adipogenic medium, showed variations in cellular morphology (b, d, f) and accumulation of neutral lipid vacuoles (demonstrated by Oil Red O staining) with respect to control cells. The latter (a, c, e) were cultured for the same time in standard culture medium, and retained the normal fibroblast-like morphology, without any positivity for the lipid-specific staining procedure. Magnification a, b: ×10; c, d ×40; e, f ×100

Moreover, adipogenic differentiation resulted in the de novo expression of molecules, characteristic of the adipocyte lineage. In fact, Fig. 9 shows that the expression of leptin and adiponectin, markers of mature adipocytes, is restricted to cells that underwent adipogenic differentiation. Therefore, this differentiation protocol resulted in the acquisition of a morphologic and functional phenotype adherent to that of adipocytes in vivo.
https://static-content.springer.com/image/art%3A10.1007%2Fs00418-008-0519-3/MediaObjects/418_2008_519_Fig9_HTML.gif
Fig. 9

RT-PCR analysis of the expression of specific markers following adipogenic differentiation of HEMSCs. Following the adipogenic differentiation protocol (lanes A), cells became positive for the expression of specific markers as adiponectin and leptin, which were undetectable in control cells (lanesC). Beta actin was used as housekeeping control gene

Discussion

Several reports indicated that extraembryonic tissues may be considered a potential source for populations of cells phenotypically resembling BM-MSC (Marcus and Woodbury 2008). This has been demonstrated for placenta (Miao et al. 2006), amniotic fluid (In ‘tAnker et al. 2003; De Coppi et al. 2007), amniotic membrane (Alviano et al. 2007; Magatti et al. 2008), umbilical cord blood (Kang et al. 2006; Neumüller et al. 2006), and umbilical cord matrix (Romanov et al. 2003; Wang et al. 2004). The lack of a standardized system to isolate Wharton jelly’s cells is reflected in the differences between the markers characterized for such mesenchymal cells. Therefore, only a few “core markers” (as CD73, CD90, CD105) are expressed in almost all cellular preparations. This issue is of central importance, since for BM-MSC it has been reported that differences in the isolation and subculture protocols may lead to different cell lines both in terms of gene expression and “stemness” potencies (Wagner et al. 2006). In the present work, we describe for the first time an extended characterization of mesenchymal stem cells, obtained from the umbilical cord matrix by non-enzymatic isolation procedure, named here as HEMSCs (human extraembryonic mesoderm stem cells).

The protocol described has been carried out without the use of proteases to achieve cellular dissociation from tissues. Therefore, even if it should be regarded as more time-consuming than other protocols described in literature, using the protease-free system we obtained cells with the old technique of “plate and wait” without long incubation times in collagenase. In fact, overdigestion of tissue may result in diminished cellular viability, and altered cellular function (e.g., for degradation of cell surface receptors) (Hung and Mauck 2004), as demonstrated for pancreatic islets isolation (Balamurugan et al. 2005). Therefore the initial selection of cells was based on the migratory capability from the original matrix to the plastic surface of the tissue culture flask (migration is one key feature of mesenchymal cells). As shown in “Results”, the isolation protocol allowed to reproducibly isolate adherent, spindle-shaped fibroblastoid cells, which retained their morphological and phenotypical features throughout the culture passages.

We performed characterization of cells to detect the expression of validated MSC markers, together with new ones, both at the protein and RNA level. Typical MSC markers (CD10, CD13, CD29, CD44, CD73, CD90, CD105, CD106, CD166) were detected in HEMSCs by Immunocytochemistry and/or RT-PCR. Moreover, these cells were negative for the expression of hematopoietic/endothelial markers (CD31, CD33, CD34, CD38, CD45, CD79, vWF), another feature of several MSC populations.

Our characterization aimed also to the extension of the “core” signature of MSCs, by investigating the presence of additional markers. Therefore, we demonstrated that these cells express mesenchymal markers (vimentin, α-SMA), neuroectodermal markers (Nestin, NSE, GFAP), and early endoderml markers (GATA-4,-5,-6, HNF4α, cytokeratin-8,-18,-19). The expression of such molecules should indicate the capability of such cells to undergo differentiation not only towards the mesoderm-derived cell types (as we demonstrated for bone and adipose tissue in the present paper), but also towards cell types characteristic of other germ layers, therefore extending the potential applications of such cells in regenerative medicine approaches. In particular, regarding NSE and GFAP expression we confirmed the data obtained by others (Romanov et al. 2003; Mitchell et al. 2003), who demonstrated that induction of a neuronal differentiation protocol should upregulate these proteins which are already present in undifferentiated cells.

One of the key properties of stem cells is self-renewal, i.e., the capability to undergo cellular replication forming a progeny of undifferentiated cells. We demonstrated that HEMSCs are clonogenic, therefore providing a formal demonstration of self-renewal. Moreover, we characterized the expression of Oct-4 isoforms in these cells. The oct-4 gene, also referred to as oct-3, encodes a member of the family of transcription factors containing the POU DNA binding domain, which binds to the octamer sequence motif 5′-ATGCAAAT-3′ (Takeda et al. 1992). Oct-4 has been proposed as a master regulator of the pluripotence of cells (Campbell et al. 2007). We demonstrated for the first time in cord-derived stem cells the presence of both isoforms of Oct-4 (A and B, derived from alternative splicing of the same transcript), a biologically relevant datum in the light of recent findings on the different contributions of the two molecules to the self-renewal process. In fact, Lee and coauthors recently demonstrated that the isoform A (capable of nuclear translocation) is mainly responsible for self-renewal of human stem cells (Lee et al. 2006). Oct-4 expression has been demonstrated previously in ESC cell lines and in limited populations of amniotic fluid stem cells (Marcus and Woodbury 2008), but not in umbilical cord-derived mesenchymal stem cells. Moreover, HEMSCs express Oct-1 and Nanog, other transcription factors characterized in ESC lines, which are involved in the maintenance of the undifferentiated phenotype of these cells.

The in vitro expandability of stem cells depends on the maintenance of a high-telomerase activity, to escape senescence processes. We demonstrated that HEMSCs maintain long telomeres at lower and higher culture passages, therefore ensuring the capability to undergo further cell cycles. The other side of the coin of fast cellular replication may reside in the acquisition of spontaneous mutations which can alter the phenotype of stem cells. Embryonic stem cell lines have been characterized not only for their capability to undergo tenths culture passages in vitro, but also for the acquisition, alongside with the culture process, of chromosomal mutations, as aneuploidies (Rao 2006). Therefore, we performed karyotype analysis which showed that HEMSCs maintained a normal karyotype.

A key property of MSC is the capability to down-regulate host immune responses, thus favoring the engraftment of the implanted cells. This regulation may take place at several steps of the immune recognition process, like DC migration and maturation (English et al. 2008), T-cell recognition (Ryan et al. 2005), and tolerance promotion by binding specific receptors on NK cells (Blaschitz et al. 2001). We demonstrated that HEMSCs harbor a number of characteristics that indicate their potential features leading to immunomodulation processes.

First of all, we showed that HEMSCs express normal MHC class I molecules, confirming the previous results by other groups. This expression is important, since it protects cells from some NK killing processes, like those occurring in killing of tumor cells with downregulation of class I molecules. Moreover, again in agreement with past reports, these cells do not express MHC class II molecules (namely HLA-DR). Class II MHC molecules are powerful alloantigens which, if expressed, mediate recognition by alloreactive CD4+ T-cells.

Finally, and most importantly, we demonstrated for the first time that HEMSCs express HLA-G for umbilical-cord derived MSCs. This particular type of MHC class I, whose expression is normally restricted to the trophoblastic lineage, leads to the development of the maternal tolerance of the mother towards the fetus (Rouas-Freiss et al. 1997; Le Bouteiller 2000). In fact, HLA-G has been shown to bind the two major inhibitory NK receptors, KIR-1 and -2 (members of the Killer-cell immunoglobin-like receptor family), thus stopping NK killing process (Moffett and Loke 2004).

Moreover, these cells featured the expression of CD80 and lacked CD86. This pattern of expression of B7 co-stimulatory molecules is viewed as a favorable combination which can lead to immune tolerance development. In fact, as reviewed by Sansom et al. (2003) CD80 appears to have an inhibitory role through CD152 binding. Moreover, CD80–CD152 interactions appear to occur more effectively in the absence of CD86 expression. Therefore, CD80 and CD86 should have opposite roles in regulating the suppressive activity of regulatory T cells, as suggested more recently by Zhu et al. (2005) who demonstrated that CD86 blockade, leaving intact CD80, induced an alloantigen-specific tolerance.

Differentiation is another key property of stem cells, and we demonstrated that HEMSCs can be successfully differentiated towards osteoblasts and adipocytes by incubation in appropriate induction media. Differentiation has been assessed both by histochemical staining (Alizarin Red S, Oil Red O) and by the expression, in differentiated cells with respect to control ones, of markers characteristic of osteoblasts and adipocytes.

Moreover, chondrogenic differentiation for these cells has also been demonstrated in our laboratory (Anzalone et al. manuscript in preparation), both as pellet-cultured and bead-cultured chondrocytes.

In conclusion, we demonstrated that umbilical cord matrix contains a cellular population of mesenchymal stem cells that can be isolated by a non-enzymatic protocol and successfully expanded for at least 60 population doublings. We demonstrated that these cells express the key markers of MSCs, being also negative for hematopoietic/endothelial markers. Moreover, we characterized new markers as Oct-4, Oct-1, HLA-G, which confer to Oct-4+/HLA-G+ HEMSCs a number of properties useful for regenerative medicine applications. In fact, we demonstrated that these cells are clonogenic, retain long telomeres, and maintain a stable phenotype and a normal karyotype throughout passages. HEMSCs can undergo differentiation towards connective tissues (bone, adipose, cartilage). Moreover, HEMSCs bear a number of immunosuppressive features, between which the expression of class I MHC molecule HLA-G has been described for the first time. The presence of markers of the neuroectodermal and endodermal lineages open further perspectives for the potential use of HEMSCs in regenerative medicine applications targeting different organs and tissues of the body.

Acknowledgments

The work was supported by University of Palermo grants (ex 60%) to FF, FC, GZ.

Copyright information

© Springer-Verlag 2008