SEARCH

SEARCH BY CITATION

Keywords:

  • Human embryonic stem cells;
  • Mesenchymal stem cells;
  • Cell surface markers;
  • Adipogenesis;
  • Chondrogenesis Osteoprogenitor;
  • Selectable marker;
  • Somatic stem cells

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Adult tissue-derived mesenchymal stem cells (MSCs) have demonstrated therapeutic efficacy in treating diseases or repairing damaged tissues through mechanisms thought to be mediated by either cell replacement or secretion of paracrine factors. Characterized, self-renewing human ESCs could potentially be an invariable source of consistently uniform MSCs for therapeutic applications. Here we describe a clinically relevant and reproducible manner of generating identical batches of hESC-derived MSC (hESC-MSC) cultures that circumvents exposure to virus, mouse cells, or serum. Trypsinization and propagation of HuES9 or H1 hESCs in feeder- and serum-free selection media generated three polyclonal, karyotypically stable, and phenotypically MSC-like cultures that do not express pluripotency-associated markers but displayed MSC-like surface antigens and gene expression profile. They differentiate into adipocytes, osteocytes, and chondrocytes in vitro. Gene expression and fluorescence-activated cell sorter analysis identified CD105 and CD24 as highly expressed antigens on hESC-MSCs and hESCs, respectively. CD105+, CD24− monoclonal isolates have a typical MSC gene expression profiles and were identical to each other with a highly correlated gene expression profile (r2 > .90). We have developed a protocol to reproducibly generate clinically compliant and identical hESC-MSC cultures.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Multipotential mesenchymal stem cells (MSCs) are stem cells that have documented evidence of therapeutic efficacy in treating musculoskeletal injuries, improving cardiac function in cardiovascular disease and ameliorating the severity of graft-versus-host disease [1]. Being lineage-restricted, they have limited but robust potential to differentiate into mesenchymal cell types, such as adipocytes, chondrocytes, and osteocytes, and have a negligible risk of teratoma formation. Host immune rejection of transplanted MSCs is routinely circumvented through autologous or immune-compatible allogeneic transplantation. MSCs can be isolated from several adult tissues including bone marrow, adipose tissues, and cord blood and expanded ex vivo. However, the availability of tissues for their isolation remains limiting and requires invasive procedures, and ex vivo expansion of MSCs, although significant, is nonetheless finite.

An alternative source for generating MSCs is the infinitely expandable and pluripotent human ESCs (hESCs) that will also eliminate the need for potentially risky invasive techniques. Host immune rejection of hESC-derived MSCs could potentially be circumvented by using either autologous hESCs generated by nuclear transfer or immune compatible allogeneic hESCs when banks of hESC lines become sufficiently large. In addition, MSCs have been shown to have immunomodulatory effects and could potentially induce immune tolerance or suppression in recipients [1, [2]3]. Therefore, the problem of host immune rejection may be less intractable for the transplantation of allogeneic MSCs. Recent reports have also suggested that some of the reparative effects associated with MSC transplantation are not mediated by differentiation of MSCs but by paracrine factors secreted by MSCs [4]. Paracrine secretion by MSC has been postulated to promote arteriogenesis [5], support the stem cell crypt in the intestine [6], protect against renal ischemic injury [7, 8], support and maintain hematopoiesis [9], ameliorate tissue damage in acute myocardial ischemia [10, [11], [12]13] and limb ischemia [14], and support megakaryocyte and proplatelet formation [15]. Therefore, some future applications of MSCs in regenerative medicine may not be cell-based. Instead, MSC-based therapies may involve biologics secreted by MSCs.

Therefore, the issue of immune rejection may be less intractable in MSC-based therapies using either transplantation of cells or secreted paracrine factors In this regard, the use of well-characterized renewable hESC lines to generate identical batches of clinically relevant MSCs will be additionally useful in developing cost-effective, consistent, and qualified production of therapeutic biologics.

Although MSC or MSC-like cells have been derived from hESCs by either transfection of a human telomerase reverse transcriptase (hTERT) gene into differentiating hESCs [16] or coculture with mouse OP9 cell line [17], the use of exogenous genetic material and mouse cells in these derivation protocols introduces unacceptable risks of tumorigenicity or infection by xenozootic infectious agents.

Here we describe the development of a protocol that could be used for the derivation of highly identical and clinically compliant MSC cultures from hESCs that circumvents the use of animal products, transfection of genetic material, or coculture with a mouse cell culture by gently trypsinizing and culturing hESCs in a feeder-free condition and a medium supplemented with basic fibroblast growth factor (bFGF) and platelet-derived growth factor (PDGF) to encourage proliferation of putative MSCs [18, 19]. Initially, three polyclonal MSC-like cultures, HuES9.E1 and HuES9.E3 and H1.E2 were generated from Hues9 and H1 hESCs, respectively [20, 21]. These hESC-derived MSCs (hESC-MSC) have a bone marrow (BM)-MSC-like surface antigen profile, that is, positive for CD29, CD44, CD49, CD105, and CD166, and negative for CD34 and CD45, a gene expression profile resembling that of BM- and adipose-derived MSCs and a differentiation potential that includes adipogenesis, chondrogenesis, and osteogenesis. Like BM-MSCs, hESC-MSCs have significant proliferative capacity in vitro. By comparing relative gene expression levels between hESC-derived MSCs and hESCs, we identified surface antigens that are either highly expressed in hESC-derived MSCs or their parental hESCs. Using a combination of these markers to sort for putative MSCs and against hESCs, we generated single cell-derived MSC cultures that were highly similar to each other and to the earlier derived hESC-MSCs. Therefore, highly similar batches of MSCs can be reproducibly generated from well-characterized hESC lines.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Derivation of hESC-MSC

Hues9 and H1 hESCs were grown as previously described [20, 21]. To differentiate hESCs, a confluent 6-cm plate of hESCs was trypsinized for 3 minutes at 37°C, neutralized, centrifuged, and resuspended in knockout Dulbecco's modified Eagle's medium (DMEM) (Gibco, Grand Island, NY, http://www.invitrogen.com), supplemented with 10% serum replacement medium (Gibco), 5 ng/ml FGF2 (Gibco), and 5 ng/ml PDGF AB (Peprotech, Rocky Hill, NJ, http://www.peprotech.com) on a gelatinized 10-cm plate. The cells were trypsinized when confluent and split 1:4. Sorting for CD105+ and CD24− was performed 1 week after hESCs had been trypsinized. The differentiating hESCs were trypsinized for 3 minutes, neutralized, centrifuged, resuspended in the culture medium, and then plated on a bacterial culture dish. After 2 hours at 37°C in CO2 incubator, the cells were harvested, washed with phosphate-buffered saline (PBS), and incubated with CD24-phycoerythrin (PE) and CD105-fluorescein isothiocyanate (FITC) (BD PharMingen, San Diego, http://www.bdbiosciences.com/index_us.shtml) for 40 minutes at room temperature. The cells were then washed with PBS and sorted on a FACS Aria using FACS Diva software (BD Biosciences Pharmingen). BM MSCs were prepared as previously described [22]. The cells were cultured in DMEM supplemented with penicillin-streptomycin-glutamine, nonessential amino acids and 10% fetal calf serum (Invitrogen-Gibco).

Differentiation into adipocytes, chondrocytes, and osteocytes was performed as previously described [17]. Oil red, Alcian Blue, and von Kossa staining were performed using standard techniques. Immunoreactivity for collagen type II was performed on paraformaldehyde-fixed, paraffin-embedded sections using a goat anti-collagen α1 type II and donkey anti-goat IgG antibody conjugated with horseradish peroxidase (HRP) (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com).

Karyotyping

Cells were received at approximately 80% confluence in Petri dish. Cells were treated with colcemid for mitotic arrest and harvested by standard hypotonic treatment and methanol:acetic acid (3:1) fixation. Slides were prepared by standard air drying method and hybridized with SKY paint probe (Applied Spectral Imaging, Migclal Ha'Emek, Israel, http://www.spectral-imaging.com). Posthybridization washes were performed according to the protocols provided by the manufacturer and established in our laboratory. Twenty to 30 metaphase cells per culture were analyzed. The karyotype of each culture is representative of >80% metaphase cells.

Transplantation Studies

Two × 106 cells were resuspended in 30 μl of saline and transferred into the renal subcapsular space as previously described [23]. After 4 months, the mice were sacrificed, and the kidneys were removed, fixed in 4% paraformaldehyde, paraffin-embedded, sectioned at 4 μM, and stained with hematoxylin and eosin.

Western Blot Analysis

Standard procedures were used [24]. Briefly, cells were lysed in radioimmune precipitation assay buffer and centrifuged at 14,000 rpm for 15 minutes at 4°C. Twenty μg of supernatant was denatured, separated on 10% SDS-polyacrylamide gel, and electroblotted onto a nitrocellulose membrane. The membrane was incubated sequentially with a primary antibody, then either a HRP-conjugated secondary antibody or a biotinylated secondary antibody followed by neutroavidin-HRP, and finally, a HRP-enhanced chemiluminescent substrate, ECS (Pierce, Rockford, IL, http://www.piercenet.com). Primary antibodies used were 1:200 dilution of anti-OCT4, anti-SOX-2, and β-actin (Santa Cruz Biotechnology). Secondary antibodies were HRP-conjugated goat anti-rabbit, rabbit anti-goat, and rabbit anti-mouse.

Polymerase Chain Reaction

Genomic polymerase chain reaction (PCR) for mouse- and human-specific repeat sequences were performed as previously described [26]. Genomic PCR for mouse- and human-specific repeat sequences were performed as previously described [25]. The PCR primers for mouse c-mos repeat sequences were: 5′-GAATTCAGATTTGTGCATACACAGTGACT-3′ and 5′-AACATTTTTCGGGGAATAAAAGTTGAGT-3′. The PCR primers for human alu repeat sequences were: 5′-GGCGCGGTGGCTCACG-3′ and 5′-TTTTTTGAGACGGAGTCTCGCTC-3′. Real time reverse transcription (RT)-PCR was performed using Taqman Gene expression assays per the manufacturer's instructions (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). One μg of total RNA was reverse transcribed with a random primer using a High Capacity cDNA Archive Kit. The Taqman primer ID for each gene analyzed were: Sox-2, Hs00602736_s1; Utf-1, Hs00747497_g1; Zfp42, Hs00399279_m1; PPARγ, Hs00602622m_1; Aggrecan, Hs00153936_m1; Collagen, Hs00264051_m1; ALP, Hs00758162_m1; and BSP, Hs00173720_m1.

Surface Antigen Analysis

Cell surface antigens on hESC-MSCs and hESCs were analyzed using fluorescence-activated cell sorting (FACS). The cells were tryspinized for 5 minutes, centrifuged, resuspended in culture medium, and incubated in a bacterial culture dish for 2–3 hours in a 37°C, 5% CO2 incubator. Then the cells were trypsinized for 1 minute, centrifuged, washed with PBS, fixed in 4% paraformaldehyde for 0.5 hour at room temperature, washed again, and blocked in 2% fetal calf serum for 0.5 hour at room temperature with agitation. One and a half × 105 cells were then incubated with each of the following conjugated monoclonal antibodies: CD24-PE, CD29-PE, CD44-FITC, CD49a-PE, CD49e-PE, CD105-FITC, CD166-PE, CD34-FITC, CD45-FITC (PharMingen) for 90 minutes at room temperature. After incubation, cells were washed and resuspended in PBS. Nonspecific fluorescence was determined by incubation of similar cell aliquots with isotype-matched mouse monoclonal antibodies (PharMingen). Data were analyzed by collecting 20,000 events on a Cyan LX (Dako North America, Inc., Carpinteria, CA, http://www.dakousa.com) instrument using WinMDI software. Nonspecific fluorescence was determined by incubation of similar cell aliquots with isotype-matched mouse monoclonal antibodies (PharMingen) or with secondary antibody alone.

Illumina Gene Chip Analysis

Total RNA (2 μg) from three samples each of primary BM and adipose-derived MSCs, from two biological replicates of HuES9.E1, HuES9.E3, H1.E2, and three undifferentiated hESC lines, H1, Hes3, and HuES9, were converted to biotinylated cRNA using the Illumina RNA Amplification Kit (Ambion, Inc., Austin, TX, http://www.ambion.com) according to the manufacturer's directions. Samples were purified using the RNeasy kit (Qiagen, Valencia, CA, http://www.qiagen.com). Hybridization to the Sentrix HumanRef-8 Expression BeadChip (Illumina, Inc., San Diego, http://www.illumina.com), washing, and scanning were performed according to the Illumina BeadStation 500x manual. The data were extracted, normalized, and analyzed using Illumina BeadStudio provided by the manufacturer. Transcript signals that were below the limit of detection at 99% confidence were eliminated as genes not expressed.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Generating MSC Cultures from Human ES Cell Lines

When hESC colonies were dispersed by trypsin and then passaged on gelatinized tissue culture plates in the absence of feeder and in serum-free medium that was supplemented with serum replacement medium, FGF2 and PDGF AB, a homogenous culture of fibroblast-like cells was generated within 2 weeks. The cultures have a fibroblastic cellular morphology that resembled BM-MSCs (Fig. 1A). Dispersing hESC colonies by collagenase was not efficient in generating these fibroblast-like cells. Two polyclonal cultures, huES9.E1 and huES9.E3, were independently generated from huES9 ESC line, whereas the third, H1.E2, was generated from the H1 ESC line. Expression of several pluripotency-associated genes was generally reduced. For example, transcript levels of HESX1, POUFL5, SOX-2, UTF-1, and ZFP42 were >101–5 fold below that in the hESCs (Fig. 1B). Protein levels of OCT4 and SOX2 were also reduced (Fig. 1C). As typified by huES9.E1 and H1.E2, hESC-MSCs did not have detectable alkaline phosphatase activity (Fig. 1D). Unlike its parental HuES9 cells, renal subcapsular transplantation of 2 × 106 HuES9.E1 cells in immune compromised SCID mice did not induce the formation of a teratoma during a 4-month observation period (Fig. 1E). The transplanted HuES9.E1 cells did not appear to survive or graft onto the recipient kidney. To assess the possibility that these cells were contaminated or fused with mouse feeder cells [25], these cultures were tested and shown to be negative for mouse-specific c-mos repeat sequences but positive for human specific alu repeat sequences (Fig. 1F). The average population doubling time of HuES9.E1, HuES9.E3, and H1.E2 were 72, 72, and 120 hours, respectively. Population doubling time was highly dependent on cell density and was most optimal at between 30% and 80% confluency. HuES9.E1 and HuES9.E3 cultures had been maintained in continuous culture for more than 25 passages at 1:3 split every week while H1.E2 culture had been maintained in continuous culture for more than 15 passages at 1:4 split every week. The karyotype of all three cultures up to 35 population doublings was either 46 XX inv(9)(p11q12) in HuES9.E1 and HuES9.E3 or 46, XY in H1.E1 in 20/20 metaphase nuclei examined. The chromosome nine inversion in HuES9.E1 and HuES9.E3 originated from the parental huES9 hESC line [20]. To assess the karyotype stability of these cells, we monitored the karyotype of HuES9.E1 up to 84 population doublings. At the 68th population doubling, we began to observe random chromosomal aberrations. Two of 20 metaphase nuclei analyzed had chromosomal aberrations. One nucleus lost one chromosome 18 and another gained one chromosome 20. At 72nd population doubling, 4 of 20 metaphase nuclei had chromosomal aberrations; one lost one chromosome 19 and one chromosome 22, two lost one chromosome 22, and one gained one chromosome 18. At the 84th population doubling, two of 20 metaphase nuclei had chromosomal aberrations; one had lost one chromosome 13 and one chromosome 18, and the other lost one chromosome 22. Therefore, the karyotype of these three cell lines is normal and stable up to at least 35 population doublings, and as such we routinely do not use these cells beyond 35 population doublings.

thumbnail image

Figure Figure 1.. Characterization of human ESC (hESC)-MSC cultures. (A): Cellular morphology under phase contrast. (B): Expression of pluripotency-associated genes in hESC-MSC. Transcript levels were measured by Taqman-based quantitative reverse transcription-polymerase chain reaction (PCR) and normalized to that of hESC. The transcript level in hESC was derived from the average of HuES9 and H1 hESC lines as the transcript level for each of the genes tested in HuES9 and H1 hESC was similar. The differences in Ct values between the two lines were <1. (C): Western blot analysis for pluripotency-associated genes in HuES9 and H1 hESC lines, HuES9.E1, HuES9.E3, and H1.E2, hESC-MSC cultures and E14 mouse ESC line. (D): Renal subcapsular transplantation of HuES9 and HuES9.E1. Paraffin-embedded, hematoxylin and eosin-stained cross-sections of kidney 4 months after transplantation with either HuES9.E1 (top) or HuES9 (bottom). (E): Alkaline phosphatase activity in human HuES9 ESC line, mouse E14 ESC line, MEF feeder, HuES9.E1, and H1.E2. (F): Genomic DNA analysis by PCR for the presence of human Alu and mouse c-mos repeat sequences. (G): Chromosomal analysis of HuES9.E1 by G-banding (top panel), spectral karyotyping (SKY) (middle panel), and inversion of chromosome 9 shown (bottom panel). Abbreviation: MEF, mouse embryonic fibroblast.

Download figure to PowerPoint

Surface Antigen Profile

Surface antigen profiling of HuES9.E1, HuES9.E3, and H1.E2 by FACS analysis revealed a surface antigen profile that is qualitatively similar to that defined for BM-MSCs, that is, CD29+, CD44+, CD49a and -e+, CD105+, CD166+, and CD34−, CD45− [26, [27]28] (Fig. 2A) The intensity of fluorescent labeling and distribution of labeled cells varied with each of the hESC-MSC cultures (Fig. 2A). To compare the surface antigen profile of these cells to that of BM-MSCs, HuES9.E1, HuES9.E3, and H1.E1 were grown in the same BM-MSC culture medium supplemented with 10% fetal calf serum for two passages. Despite the change in culture condition, HuES9.E1, HuES9.E3, and H1.E1 continued to be CD29+, CD44+, CD49a+, CD105+, CD166+, and CD34−, CD45− (Fig. 2B; data not shown for H1.E1) and were largely similar to that of BM-MSCs. An exception was CD29 and CD49a, which had a much lower expression in BM-MSCs. These data indicated that the hESC-MSCs exhibited characteristic BM-MSC surface antigen profiles that were stable and were not significantly influenced by changes in their microenvironment.

thumbnail image

Figure Figure 2.. Surface antigen profiling by fluorescence-activated cell sorting (FACS) analysis. (A): HuES9.E1, HuES9.E3, and H1.E2 hESC-MSCs, HuES9 hESCs, and murine embryonic fibroblast feeder cells were stained and analyzed on a Cyan LX instrument using WinMDI software. Nonspecific fluorescence was determined by incubation of similar cell aliquots with isotype-matched mouse monoclonal antibodies. (B): HuES9.E1 and HuES9.E3 hESC-MSCs were passaged twice in serum-containing BM-MSC medium before being analyzed in parallel with BM-MSCs by FACS analysis. H1.E2 hESC-MSCs were also similarly passaged twice in serum-containing BM-MSC medium, and due to instrument availability, were analyzed on a FACS Calibur using Cellquest analytical software. Abbreviations: BM-MSC, bone marrow-MSC; MEF, mouse embryonic fibroblast.

Download figure to PowerPoint

Differentiation Potential of hESC-MSC: Adipogenesis, Chondrogenesis, and Osteogenesis

As all of the surface antigens associated with MSCs are also expressed on many other cell types, and the expression of these surface antigens are variable, identification of presumptive MSCs have traditionally relied on functional parameters [26]. It is reported that the default differentiation pathway of MSCs in culture is osteogenesis with varying degrees of adipogenesis and chondrogenesis [26]. Differentiation potential of HuES9.E1 cells was therefore tested using standard differentiation conditions for adipogenesis, chondrogenesis, and osteogenesis [17]. Adipocytic differentiation was highly efficient with oil droplets observed in >99% of the cells (Fig. 3A). Consistent with its role as important transcription factor in adipogenesis [29], PPARγ mRNA in the hESC-derived MSCs, which was approximately 10–100-fold higher than that in their respective parental ESC lines increased by a further twofold (Fig. 3A). Chondrogenesis or the formation of cartilage was also efficient with >90% of cells producing proteoglycans in extracellular matrix as detected by Alcian Blue staining (Fig. 3B) and approxiamtely 20% of the cells being immunoreactive for collagen II. Transcript level of aggrecan, an cartilage-specific extracellular matrix protein was also increased [30] (Fig. 3B). However, transcript level of collagen II, another cartilage-specific extracellular matrix protein was decreased despite the presence of collagen II immunoreactivity in the matrix. The reason is not known but some mRNAs particularly those with AU-rich elements are known to be destabilized when translated [31, 32]. When HuES9.E1 cells were induced to undergo osteogenesis or the formation of bone, expression of bone-specific alkaline phosphatase (ALP) and bone sialoprotein (BSP) [33] was upregulated by 2–3-fold (Fig. 3C). However, mineralization, a more advanced stage of bone formation [34] as determined by von Kossa staining was poor (Fig. 3C). There was <1% positive staining in the differentiated culture.

thumbnail image

Figure Figure 3.. Differentiation of HuES9.E1. HuES9.E1 cells were induced to undergo adipogenesis, chondrogenesis, and osteogenesis using standard protocols. Transcript levels were measured using Taqman Gene expression assays, and relative transcript levels were normalized to that of the parental HuES9 hESC. (A): Adipogenesis. (i): Day 14 after inducing adipogenesis, cells were stained for oil droplets by oil red. Inset is a typical HuES9.E1-derfived adipocyte; (ii): PPARγ mRNA level before induction of differentiation, and at day 7 and day 14 of differentiation was measured by Taqman quantitative reverse transcription-polymerase chain reaction (RT-PCR). (iii): Relative PPARγ mRNA levels in HuES9 and HI hESCs, their derivative MSC cell cultures (HuES9.E1, HuES9.E3, and H1.E2) and adult tissue-derived MSCs (BM-MSC and ad-MSC). (B): Chondrogenesis. (i): Day 21 after induction of chondrogenesis, cells were stained for proteoglycans by Alcian Blue (left) and immunoreactivity to collagen type II using a horseradish peroxidase-based visualization assay (right). (ii, iii): Aggrecan and collagen II mRNA levels before induction of differentiation, and at day 14 and day 21 of differentiation were measured by Taqman quantitative RT-PCR. (C): Osteogenesis. (i): Day 21 after inducing chondrogenesis, cells were stained for mineralization by von Kossa stain. (ii, iii): Bone-specific ALP and BSP mRNA levels before induction of differentiation and at day 14 and day 21 of differentiation were measured by Taqman quantitative RT-PCR. Abbreviations: ALP, alkaline phosphatase; BSP, bone sialoprotein.

Download figure to PowerPoint

Gene Expression Profile

Gene expression profiling of the hESC-MSCs were performed to (a) assess the relatedness of hESC-MSC cultures with adult tissue-derived MSCs using BM-MSCs and adipose derived (ad)-MSCs from three different individuals, and three human ESC lines; (b) assess the relatedness between each of the three hESC-MSC cultures; (c) compare the similarity and differences between MSCs derived from hESC and those derived from BM.

Labeled cDNA prepared from total RNA RNA were hybridized to Illumina BeadArray containing approximately 24,000 unique features. Hierarchical clustering of expressed genes in three hESC-MSC cultures, that is, HuES9.E1, HuES9.E3, and H1.E2, three BM-MSC samples, and three adipose-derived (ad)-MSC samples revealed that the gene expression profile of hESC-MSCs was more closely related to that of adult tissue-derived MSCs, namely BM-MSC and ad-MSC, than to their parent hESCs (Fig. 4A). Interestingly, MSCs clustered according to their tissue of origin, and this can be further demarcated into adult versus embryonic tissue as suggested by the clustering of ad-MSCs and BM-MSCs as a distinct group from hESC-MSCs. Differences in MSCs derived from tissues of different developmental stages have also been previously reported [35].

thumbnail image

Figure Figure 4.. Gene expression analysis. (A): Hierarchical clustering of expressed genes in three hESC-MSC cultures consisting of HuES9.E1, HuES9.E3, and H1.E2, three BM-MSC samples, three ad-MSC samples, and three hESC lines consisting of HuES9, H1, and Hes3. (B): Pairwise comparison of gene expression between hESC-MSCs and BM-MSCs (left) and between hESC-MSCs and hESCs (right). (C): Analysis of commonly expressed genes (less than twofold difference) in hESC-MSCs and BM-MSCs. The genes are classified into biological processes using the Panther classification system. Each biological process was determined if it was significantly over- or under-represented (p < .01) by comparing the observed frequency of genes to the expected frequency of genes in the NCBI: H. sapiens gene database of 23,481 genes for each biological process. Significantly over- or under-represented processes were grouped and graphically presented (see supplemental data 1). (D): Analysis of differentially expressed genes (more than twofold difference) in hESC-MSCs and BM-MSCs. Biological processes that were significantly over- or under-represented (p < .01) by genes highly expressed in hESC-MSCs or BM-MSCs were grouped and graphically presented (see supplemental data 2).

Download figure to PowerPoint

Pairwise comparison of gene expression between hESC-MSCs and BM-MSCs revealed a correlation coefficient of .72 suggesting that although there was significant conservation of gene expression in both hESC-MSCs and BM-MSCs, there were also significant differences (Fig. 4B). Pairwise comparison between hESC-MSCs and hESCs confirmed the distinction of hESC-MSCs from hESCs with a low correlation coefficient of .65 (Fig. 4C).

To assess the relatedness between each of the three hESC-MSC cultures, HuES9.E1, H1.E2, and HuES9.E3 were each compared to the same reference consisting of HuES9.E1, H1.E2, and HuES9.E3. The correlation coefficients of HuES9.E1, H1.E2, and HuES9.E3 to the same reference were virtually identical, that is, .93, .95, and .93, respectively, suggesting that HuES9.E1, H1.E2, and HuES9.E3 are highly similar (Fig. 4C).

Of 8,699 and 8,505 genes that were expressed above the limit of detection at 99% confidence level in hESC-MSC and BM-MSC, respectively (supplemental data 1: Table 1, 2), 6,376 genes were expressed in both hESC-MSCs and BM-MSCs at <2.0-fold difference (supplemental data 1: Table 3). As these genes are likely to provide insights into the fundamental biology of MSCs, we examine the biological processes that are driven by these genes. Of the 6,376 commonly expressed genes, 4,064 were found in the Panther classified gene list (http://www.pantherdb.org) (supplemental data 1: Table 4). Classification of these genes into different biological processes revealed that the frequency of genes in some of the biological processes were significantly over- or under-represented (p < .01) when compared to the reference list consisting of 23,481 genes in NCBI: Homo sapiens gene database (supplemental data 1: Tables 5, 6). For example, there were an over-representation of genes in metabolic processes that are likely to be important for growth and self-renewal of putative stem cells. These processes include basal metabolic processes for catabolic and anabolic activities, biosynthesis of secretory products that require extensive post-translational modifications, for example, glycosylation and cellular proliferation (Fig. 4D; supplemental data 1: Table 7). Consistent with their mesenchymal potential, there was also an under-representation of genes involved in ectoderm differentiation particularly neural development. The gene expression analysis also suggested that mitogen-activated protein kinase kinase kinase (MAPKKK) signaling is prominent in both BM-MSC and hESC-MSCs. MAPKKK signaling, which consists of at least three subfamilies, namely the classic mitogen-activated protein kinase (also known as ERK), stress-activated protein kinase/c-Jun N-terminal kinase (JNK), and p38 kinase, are associated with proliferation, differentiation, development, regulation of responses to cellular stresses, cell cycle, death, and survival [36, [37], [38]39].

Further analysis of the gene expression profiles of hESC-MSC and BM-MSC revealed that 1,142 and 1,134 genes were expressed at >2.0-fold in hESC-MSC and BM-MSC, respectively (Supplementary data 2: table 1 and 2). Of these, 738 and 880 genes, respectively, were located in Panther classified gene list (http://www.pantherdb.org) (supplemental data 2: Tables 3 and 4) and classified into biological processes (supplemental data 2: Table 5 and 6). Biological processes that were significantly over- or under-represented (p < .01) when compared to the reference list of 23,481 genes in NCBI: H. sapiens gene database were selected (supplementary data 2: Table 7). Genes that were preferentially expressed in BM-MSCs were clustered in biological processes that are involved in metabolic processes, cell structure, differentiation, and signaling, whereas preferentially expressed hESC-MSC genes were clustered in those processes involved in proliferation, differentiation, immunity, and signal transduction (Fig. 4E; supplemental data 2: Table 7). The over-representation of genes in biological processes associated with proliferation was consistent with the higher proliferative capacity of hESC-MSC over BM-MSC. Although highly expressed genes in either hESC-MSC or BM-MSC were over-represented in the general categories of differentiation and signaling, the specific biological processes within each category were differently represented in hESC-MSC and BM-MSC. For example, differentiation processes that are associated with early embryonic development such as embryogenesis and segmentation were over-represented in hESC-MSC, whereas those associated with late embryonic development, for example, skeletal development and muscle development were overrepresented in BM-MSC. Similarly, extracellular matrix protein-mediated signaling and MAPKKK cascade were overrepresented in BM-MSC. Together, these observations suggest that differentiation potential and signaling pathway utilization in hESC-MSC and BM-MSC may not be identical.

Distinguishing Surface Markers for hESCs and hESC-Derived MSCs for Isolating of Single Cell-Derived MSC Population

The genome-wide gene expression was queried for highly expressed genes in either hESC-MSC or hESC that encode for membrane proteins to facilitate the isolation of MSCs from differentiating hESCs. From a list of top 20 highly expressed genes encoding for putative membrane proteins in either hESC-MSCs or hESCs, candidate genes were selected for which antibodies against their gene product was commercially available (Table 1). Among those candidate genes that were highly expressed in hESC-derived MSCs are ENG (CD105), ITGA4 (CD49d), PDGFRA, NT5E (CD73) that are characteristic surface markers of MSCs derived from adult tissues [26, [27]28] and among those candidate genes that were highly expressed in hESC are previously identified as highly expressed hESC-specific genes, ITGB1BP3 and PODXL [40], and CD24, whose expression has not been associated with hESCs. We confirmed that CD24 was highly expressed in hESC versus hESC-MSC (Fig. 5A).

Table Table 1.. Surface antigen gene expression
Thumbnail image of
thumbnail image

Figure Figure 5.. Positive and negative sorting for generation of human ESC (hESC)-MSC. (A): FACS analysis HuES9.E1 HuES9.E3 and H1.E2 hESC-MSCs, HuES9 hESCs, and murine embryonic fibroblast feeder cells were stained and analyzed for the presence of CD24 on a Cyan LX instrument using WinMDI software. Nonspecific fluorescence was determined by incubation of similar cell aliquots with isotype-matched mouse monoclonal antibodies. (B): Sorting for CD105+, CD24− cells from HuES9 cells that have been trypsinized and propagated without feeder in media supplemented with platelet-derived growth factor and fibroblast growth factor (FGF)2 for 1 week. CD105+, CD24− cells represented in Q4 were selected for culture. (C): Pairwise comparison of gene expression between Q4.1 and each of the other Q4 cultures, namely Q4.2–Q4.5. (D): Pairwise comparison of gene expression between all Q4 cultures and hESC-MSCs consisting of HuES9.E1, HuES9.E3, and H1.E2, and between all Q4 cultures and BM-MSCs. (E): SKY analysis of Q4.3. Abbreviations: FITC, fluorescein isothiocyanate; MEF, mouse embryonic fibroblast; PE, phycoerythrin.

Download figure to PowerPoint

Reproducible Derivation of hESC-MSC Cultures

We next tested the utility of these markers to enhance reproducible derivation of identical hESC-MSCs. One week after trypsinization and culture in medium supplemented with serum replacement medium, FGF2 and PDGF AB, the cells in the culture were sorted by FACS for CD105 and against CD24. CD105+ and CD24− cells constituted approximately 5% of the culture (Fig. 5B). Sorted cells were plated onto 10 × 96-well plates at 1 cell per well, 1 × 24-well plate at 100 cells per well, and 3 × 6-well plates at 1,000 cells per well. Of these, only 5 of the 18 1,000 cells per wells have surviving cells that proliferated to beyond 107 cells to generate MSC-like cultures, suggesting that these cultures were likely to be generated from a single cell. Genome-wide gene expression profiling of these five cultures, Q4.1–Q4.5 using the Illumina BeadArray containing about 24,000 unique features revealed a high degree of similarity among the five cultures with four of the lines having a correlation coefficient of .96 and the remaining one with .90 (Fig. 5C). In our hands, the correlation coefficient between technical replicates performed at least 1 month apart using the same RNA sample is routinely in the range of .97–.98. Q4.1–Q4.5 were also highly similar to the hESC-MSCs consisting of huES9.E1, H1.E2, and huES9.E3, and BM-MSCs with a correlation coefficient of .87 and .81, respectively (Fig. 5D). In contrast, the correlation coefficient of Q4.1–Q4.5 to their parental HuES9.E1 hESC line was a low .55 (Fig. 5D). Chromosomal analysis using G banding and SKY was performed on randomly selected Q4.3 culture. Q4.3 has a normal karyotype with a chromosome nine inversion that originated from its parental HuES9 hESC line in 21 of 21 metaphase cells (Fig. 5E) [20]. Together these observations suggested that highly similar MSC cultures can be reproducibly generated by sorting for CD105+ and CD24− cells from trypsinized hESC culture after propagation in media supplemented with bFGF2 and PDGF BB for 1 week.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

This report describes a protocol that could be used to reproducibly generate highly similar and clinically compliant MSC populations from hESCs by trypsinizing and propagating hESCs without feeder support in medium supplemented with FGF2 and PDGF AB before sorting for CD105+, CD24− cells. The isolation of MSC or MSC-like cells from hESC has been previously described. For example, Barberi et al. [17] cocultured hESCs with mouse OP9 cells in the presence of serum for 40 days before sorting for CD73+ cells that constitute approximately 5% of the total cell population, whereas Xu et al. [16] infected hESC-derived embryoid bodies with a retrovirus expressing hTERT. However, the critical components of these protocols, that is, viral infection of exogenous DNA, exposure to mouse cells, and use of serum introduce unacceptable risks of tumorigenicity and xenozootic infection and preclude the use of these MSCs for clinical applications. In contrast, our protocol does not require serum, use of mouse cells, or genetic manipulations and requires less manipulations [35, 41] and time and is therefore highly scalable. The robustness and reproducibility of this protocol was evidenced by the isolation of highly similar MSC cultures from two different hESC lines, HuES9 and H-1, and also a third one, Hes-3 [42] and data not shown).

hESC-MSCs were remarkably similar to BM-MSCs. These presumptive MSCs satisfied the morphologic, phenotypic, and functional criteria commonly used to identify MSCs [26], that is, adherent monolayer with a fibroblastic phenotype, a surface antigen profile that is CD29+, CD44+, CD49a+ CD49e+, CD105+, CD166+, CD34−, and CD45− [26, [27]28], and a differentiation potential that includes adipogenesis, chondrogenesis, and osteogenesis [26]. We further demonstrated that this differentiation potential is lineage-restricted, and unlike that of their parental hESCs, is not pluripotent. Consistent with the restriction in their differentiation potential, hESC-MSCs do not express any of the pluripotency-associated markers, for example, Sox-2, Oct4, or alkaline phosphatase activity. The loss of pluripotency and reduced differentiation potential possibly promotes the robust and efficient differentiation of hESC-MSCs into adipocytes, chondrocytes, and osteocytes and highlights a distinct advantage of lineage-restricted ESC-derived stem cell lines over their parental ESC lines. In addition, the risk of teratoma formation is minimal as nonpluripotent hESC-MSCs cannot induce formation in SCID mice. In fact, hESC-MSCs failed to survive or differentiate into other cell types at the site of the graft, consistent with previous reports of poor survival and differentiation of transplanted immortalized human MSCs in nonobese diabetic (NOD)/SCID mice [43] possibly through the intact innate immune response of these mice [44]. As an additional evaluation, global gene expression was compared to that of the more traditional BM-MSCs. Despite the genetic variations within and between the different hESC-MSC and BM-MSC samples, pairwise comparison of gene expression between three independently derived hESC-MSC populations, and three individual BM-MSC samples were found to be similar with a correlation coefficient of .72.

hESC-MSCs have a substantial proliferative capacity in vitro and could undergo at least 35 population doublings while maintaining a normal diploid karyotype, and a stable gene expression and surface antigen profile. Random nonclonal chromosomal aberrations and alterations in gene expression become manifested only after 35 population doublings. Although hESC-MSC and BM-MSCs shared many distinctive hallmarks of MSCs, genome-wide gene expression analysis suggest that there were not only substantial similarities but also important differences. Not unexpectedly for stem cells with self-renewal and differentiation potential, the commonly expressed genes in both hESC-MSC and BM-MSCs were over-represented in growth, proliferation, and differentiation and under-represented in nonmesenchymal differentiation processes such as ectoderm differentiation particularly neural development. Of note, MAPKKK signaling, which is associated with proliferation, differentiation, development, regulation of responses to cellular stresses, cell cycle, death, and survival [36, [37], [38]39], was upregulated in both hESC-MSC and BM-MSCs and more so in BM-MSCs than in hESCs. The latter was suggested by the over-representation of MAPKKK-associated genes that were preferentially expressed in BM-MSCs. Therefore, the MAPKKK signaling pathway may be a target for manipulating proliferation and differentiation of MSCs. Consistent with previous reports that MSCs isolated from fetal and adult tissues have different immunomodulatory effects, and fetal MSCs have higher proliferative capacity and are less lineage committed than adult MSCs [35, 41], we also observed that the differences between hESC-MSC and BM-MSCs appear to be reflective of their tissue of origin. Genes that were preferentially expressed in hESC-MSC appeared to be associated with embryonic processes such as proliferation and early developmental processes of embryogenesis and segmentation, whereas those in BM-MSCs were over-represented in biological processes associated with more mature cell types, such as metabolic processes, cell structure, and late developmental processes of skeletal development and muscle development. Differences in gene expression may explain the higher proliferative capacity in hESC-MSCs and their relatively poor osteogenic differentiation in vitro.

In searching for cell surface markers that will facilitate the preparation of a homogenous hESC-MSC population, gene expression analysis was performed to identify candidate genes that encoded for surface antigens and that were preferentially expressed on hESC-MSC and not hESCs. Not unexpectedly, it revealed many candidate surface antigens known to be associated with MSCs, for example, CD105, CD73, ANPEP, ITGA4 (CD49d), and PDGFRA. However, we noted that some of the MSC-associated surface antigens, for example, CD29 and CD49e, were also highly expressed in hESCs, and their expression was verified by FACS analysis. Therefore, the association of a surface antigen with MSCs may not be sufficient to qualify the antigen as a selectable marker for isolating MSCs from hESC. Although CD73 has been used to successfully isolate putative MSCs from hESCs on the basis that it was highly expressed on MSCs [17], it was also fortuitously more highly expressed on hESC-MSCs than on their parental hESCs (Table 1). As both CD73 and CD105 are highly expressed surface antigens in MSCs and are among the top 20 highly expressed surface antigens in hESC-MSCs relative to hESC (Table 1), the use of either CD73 or CD105 in our case as selectable marker for putative MSCs will be equally effective in sorting for putative MSCs generated by differentiating hESCs. To further enhance the stringency of sorting and to reduce possible hESC contamination, we also identified surface antigens that are highly expressed on hESC and not hESC-MSC by comparing the gene expression profiles of hESCs and hESC-MSCs. Some of the top 20 candidates included the previously identified hESC-specific surface antigens MIBP, ITGB1BP3, and PODXL [40] and CD24. CD24 was not previously identified as an hESC-specific surface antigen where gene expression in hESC was compared with that in hESC-derived embryoid bodies [40], suggesting that CD24 may also be expressed in other differentiated hESC-derived cell types. Indeed, CD24 has been reported to be expressed on many different cell types [45]. FACS analysis confirmed the presence of CD24 expression on hESC and its absence on hESC-MSCs. Therefore, CD24 was an ideal negative selection marker such that when used in conjunction with CD105 as a positive selectable marker for isolating putative MSCs from differentiating hESC cultures, it should enhance the selection specificity for MSCs and reduce contamination by hESC and other hESC-derivatives. This would then reduce the risk of teratoma formation and increase the clinical relevance of this protocol.

The incorporation of positive and negative selectable markers into the derivation protocol resulted in the derivation of five monoclonal isolates with a genome-wide expression profile that was almost identical to each other confirmed the specificity of the selection criteria. Global pairwise gene expression comparison between the five isolates reveal a near identical gene expression profile that is comparable to that observed for technical replicates using the same RNA samples. This suggests that sorting for CD105+, CD24− cells from trypsinized hESCs 1 week after feeder-free propagation in a medium supplemented with FGF2 and PDGF AB will generate consistent batches of hESC-MSC cell culture.

In conclusion, this protocol could be used to reproducibly generate clinically compliant and highly similar hESC-MSC cultures and greatly facilitate the production of MSC-based therapeutic biologics that are of uniform and consistent quality. However, until the issue of immune incompatibility between hESC-MSCs and potential recipients is resolved, cell-based therapy using hESC-MSC will not likely be practical. Nonetheless, hESC-derived MSCs will be superior to those derived from BM in developing therapeutics that are based on paracrine factors secreted from MSCs, for example, the use of MSC-conditioned medium to ameliorate tissue damage in acute myocardial ischemia [10, [11], [12]13] and limb ischemia [14]. Specifically, Dzau and his co-workers [12] had demonstrated that the inhibition of ventricular remodeling and restoration of cardiac function in a rodent model of myocardial ischemia by intramyocardial injection of BM-MSCs could be effected through the administration of cultured medium conditioned by BM-MSCs. They further demonstrated that genes, coding for factors (vascular endothelial growth factor, FGF-2, HGF, IGF-I, and TB4) that are potential mediators of the effects exerted by the conditioned medium are significantly up-regulated [11]. Together, their observations support the hypothesis that BM-MSCs promote myocardial protection and functional improvement through paracrine secretions of cytokines or growth factors. Therefore, the MSC-based therapy needs not be cell-based. Instead, cell secretion, which is less immunogenic would be as therapeutically efficacious. As hESC-MSCs are highly similar to BM-MSCs, they are likely to secrete most of the cytokines and growth factors known to be secreted by BM-MSCs. In addition, they can be reproducibly generated from the same renewable hESCs. This reproducible generation of identical MSC cultures will ensure that conditioned media of consistent and uniform quality will be produced on a large scale. In contrast, similar large scale production of media conditioned by BM-MSCs will necessitate the use of multiple BM donors and will inherently be variable from batch to batch. Furthermore, the robust derivation of highly identical MSCs from a defined cell type such as hESC provides a useful model to study and better understand the ontology and biology of MSC that has remain an enigma despite much intense study [26].

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

This work was funded by a BMRC grant (BMRC Project number 01/1/21/17/045) to S.K.L.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
Supplementary_Data_1.pdf1455KSupplemental Data 1
Supplementary_Data_2.pdf642KSupplemental Data 2

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.