MSCs and mesenchymal progenitor cells (MPCs) are studied for their potential in regenerative medicine. MSCs in particular have great potential, because various reports have shown that they can differentiate into many different cell types. However, the difference between mesenchymal stem/progenitor cells and so-called fibroblasts is unclear. In this study, we found that most of the distinct populations of primary fibroblast-like cells derived from various human tissues, including lung, skin, umbilical cord, and amniotic membrane, contained cells that were able to differentiate into at least one mesenchymal lineage, including osteoblasts, chondrocytes, and adipocytes. We therefore propose that primary fibroblast-like cell populations obtained from various human tissues do not comprise solely fibroblasts, but rather that they also include at least MPCs and possibly MSCs, to some extent.
Disclosure of potential conflicts of interest is found at the end of this article.
Fibroblasts are easily obtained from various tissues and organs. Because of this, fibroblasts and fibroblast-derived cell lines have been widely used in biological experiments. The morphology of MSCs and mesenchymal progenitor cells (MPCs) resembles that of fibroblasts . However, unlike fibroblasts, MSCs and MPCs have the potential to differentiate into tissues of at least one mesenchymal lineage, such as bone, cartilage, or fat . Thus, MSCs and MPCs are thought to be promising materials for regenerative medicine and gene therapies. In fact, MSCs have already been used in the clinic to repair or regenerate somatic tissues, such as bone defects  and infarcted heart [3, 4]. In addition, studies in animals have suggested that MSCs might be used to regenerate other cells or tissues, such as retina , neurons , and hepatocytes . Gene transfer into MSCs mediated by adenovirus , retrovirus , or lentivirus vectors  has also been reported.
Human MSCs can be obtained from various tissues, such as bone marrow [1, 11, 12], adipose tissue , umbilical cord blood [14, 15], placenta [16, 17], and skin . In addition, a number of studies have identified MPCs that have limited differentiation potential compared with MSCs [19, , , , , , –26]. MSCs and MPCs are generally isolated by repeated passage of adherent cells derived from a particular tissue of interest, followed by analysis of the ability of such cells to differentiate into mesenchymal lineages. Adherent cells that fail to differentiate into mesenchymal lineages are considered to be fibroblasts. Thus, it is possible that many primary adherent cell cultures considered to be fibroblastic in nature and that have not been tested for their ability to differentiate into mesenchymal lineages might include MSCs or MPCs. We examined this idea directly by evaluating the potential of a number of independent so-called “primary fibroblast cultures” to differentiate into mesenchymal lineages. We evaluated a number of cell populations from our lab and other laboratories. All cell populations from other laboratories were obtained from an archived tissue culture collection, and they all had been deposited and registered as primary fibroblast cultures.
Materials and Methods
All cell materials used in this study were derived from human tissues, were not manipulated with any virus or gene, were not immortalized, and were not cloned cell populations. All populations except those derived from amniotic membranes were purchased from the Cell Engineering Division of RIKEN BioResource Center (Tsukuba, Ibaraki, Japan), an archived tissue culture collection. All purchased cells were cultured following the procedures described by each depositor.
The adherent cells derived from human amniotic membranes were prepared in our laboratory as follows. Amniotic membranes were minced thoroughly with scissors, treated with 1 mg/ml collagenase (Wako Chemical, Osaka, Japan, http://www.wako-chem.co.jp/english) for 60 minutes at 37°C, followed by incubation in 0.25% trypsin/0.02% EDTA solution (Sigma, St. Louis, http://www.sigmaaldrich.com) for 15 minutes, and then filtered through a metal mesh (400 μm). Cells (1 × 107) were cultured in 10-cm plastic dishes in four different media: (a) MSCGM BulletKit (Cambrex, Walkersville, MD, http://www.cambrex.com) for amniotic membrane (AM) no.11-3 cells; (b) α-minimal essential medium (MEM) (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 10% fetal calf serum (FCS; Hyclone, Logan, UT, http://www.hyclone.com), 10 ng/ml human epidermal growth factor (hEGF; R&D Systems, Minneapolis, http://www.rndsystems.com), and 105 unit/ml human leukemia inhibitory factor (Chemicon, Temecula, CA, http://www.chemicon.com) for AM no.12-1 cells; (c) α-MEM supplemented with 10% FCS and 10 ng/ml hEGF for AM no.17-5 cells; and (d) and α-MEM supplemented with 10% FCS for AM no.17-9 and AM no. 19-4 cells. Twenty-four hours later, nonadherent cells were discarded and fresh medium was added to the culture. The cells were trypsinized and replated when they reached 90% confluence. Specifically, the cells were treated with 0.25% trypsin/0.02% EDTA solution, and one-third of the cells were plated on a new 10-cm dish. Cells were subsequently subjected to the following analyses following three or four passages.
In Vitro Differentiation
Induction of osteogenic, adipogenic, and chondrogenic differentiation was performed according to previously reported methods  with some modifications.
For induction of osteogenic differentiation, cells were cultured on type I collagen-coated plastic dishes (BD Biosciences, San Diego, http://www.bdbiosciences.com). When cells reached 100% confluence, they were incubated in differentiation medium for 4 weeks with medium changes occurring every 2 to 3 days. The differentiation medium was as follows: Dulbecco's modified Eagle's medium-high glucose (DMEM-HG; Sigma) supplemented with 10% FCS (Hyclone), 10−7 M dexamethasone (Sigma), 0.5 μM ascorbic acid 2-phosphate (Sigma), and 10 mM β-glycerophosphate (Tokyo Kasei Kogyo, Tokyo, http://www.tokyokasei.co.jp). To detect mineralization (calcium deposits), cells were fixed with ice-cold 70% ethanol and stained with alizarin red S (Sigma). To confirm osteogenic differentiation further, we evaluated the expression of the osteoblast-specific protein osteocalcin (see below).
For chondrogenic differentiation, a cell pellet consisting of 2.5 × 105 cells was incubated at the bottom of a 15-ml conical tube in 0.5 ml of differentiation medium made up of α-MEM supplemented with 3.5 g/ml glucose (Sigma), 1% (v/v) ITS-plus (Sigma), 2 mM l-glutamine (Invitrogen), 100 μg/ml sodium pyruvate (Invitrogen), 0.2 mM ascorbic acid 2-phosphate (Sigma), 10−7 M dexamethasone (Sigma), and 10 ng/ml transforming growth factor (TGF)-β3 (R&D Systems). Twenty-four hours later, the medium was replaced and then subsequently again every 2 days until 28–30 days. Finally, cell pellets were fixed with 4% paraformaldehyde (Wako) for 24 hours at room temperature. The production of mucopolysaccharide, an indicator of chondrogenic differentiation, was measured by staining thin sections of the pellets with toluidine blue. Preparation of thin sections and staining with toluidine blue were performed by SRL Inc. (Tokyo, http://www.srl-group.co.jp/en/).
To perform reverse transcription-polymerase chain reaction (RT-PCR) on the differentiated cells, chondrogenic differentiation was performed essentially according to a previously reported method  using differentiation medium MSCGM BulletKit Chondrogenic (Cambrex) supplemented with 10 ng/ml bone morphogenetic protein (BMP)-2 (R&D Systems), 10 ng/ml BMP-4 (R&D Systems), and 10 ng/ml TGF-β3 (R&D Systems). The medium was changed every 3 days, and cells cultured for 14 days were subjected to semiquantitative RT-PCR (see below).
For analysis of adipogenic differentiation, cells were first grown to 100% confluence. The cells were then incubated for 3 days in induction medium made up of DMEM-HG supplemented with 10% FCS, 0.2 mM indomethacin (Sigma), 1 μM dexamethasone (Sigma), 0.5 mM 3-isobuthyl-1-methylxanthine (Sigma), and 10 μg/ml human insulin (Sigma) (step 1). The cells were next incubated for 3 days in maintenance medium made up of DMEM-HG supplemented with 10% FCS and 10 μg/ml human insulin (step 2). Incubation of the cells in induction medium and then maintenance medium (step 1 and step 2) was repeated four to five times over a period of 24 to 30 days. To detect fat deposition in the cells, cells were fixed with 4% paraformaldehyde (Wako) for 30 minutes at room temperature and stained with oil red O (WALDECK-GmbH & Co. KG, Muenster, Germany, http://www.chroma.de/).
Enzyme-Linked Immunosorbent Assay to Detect Osteocalcin
To evaluate the expression of the osteoblast-specific protein, osteocalcin, we used the Gla-type Osteocalcin EIA KIT (TaKaRa Bio, Otsu, Shiga, Japan, http://www.takarabiousa.com/html/index.html). Four weeks after the initiation of osteogenic differentiation, cells were washed with phosphate-buffered saline and incubated with 10% formic acid (Wako) for 10 minutes at room temperature. After purification of the extracted proteins with a NAP-5 column (GE Little Chalfont, Buckinghamshire, U.K., http://www.gehealthcare.com/), the proteins were subjected to the assay following the manufacturer's procedure.
Fourteen and 28 days following the initiation of chondrogenic and adipogenic differentiation, respectively, total RNA was extracted from cells using ISOGEN (Wako). Reverse-transcription was carried out with 5 μg of total RNA using Ready-To-Go T-primed First-Strand kit (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) in a 33-μl reaction volume. A 167-μl volume of TE (10 mM Tris-Cl and 1 mM EDTA, pH 8.0) was then added to the reaction mixture, and 3 μl of this was used in each polymerase chain reaction (PCR). PCR was carried out with recombinant Taq polymerase (TaKaRa Bio). Cycling parameters were as follows: denaturation at 94°C for 30 seconds, annealing at 57°C for 30 seconds, and extension at 72°C for 30 seconds. Amplification of the gene encoding hypoxanthine phosphoribosyl transferase (HPRT) was used as an internal control in the PCR. The number of PCR cycles performed was as follows: HPRT, 30 cycles; type II collagen, 35 cycles; lipoprotein lipase, 30 cycles; and fatty acid binding protein aP2, 30 cycles. PCR products were separated on a 1.5% agarose gel and visualized by ethidium bromide staining. The relative density of a specific gene signal was determined by dividing the density of the specific gene band by that of the HPRT band using NIH Image analysis software (Bethesda, MD).
The sequences of the PCR primers were as follows: for human HPRT, the sense primer was 5′-TTG CTG ACC TGC TGG ATT AC-3′ and the antisense primer was 5′-TTT CCA GTT TCA CTA ATG ACA C-3′ (414-base pair [bp] PCR product); for human type II collagen, the sense primer was 5′-GCC ATG AAG GTT TTC TGC AAC A-3′ and the antisense primer was 5′-ACA GTC TTG CCC CAC TTA CC-3′ (419-bp PCR product); for human lipoprotein lipase, the sense primer was 5′-AAT CCT CAG CTG ACA C-3′ and the antisense primer was 5′-TCT GGC TCA TTG CAC-3′ (341-bp PCR product); and for human aP2, the sense primer was 5′-GCC AGG AAT TTG ACG AAG-3′ and the antisense primer was 5′-TCC CTT GGC TTA TGC TC-3′ (206-bp PCR product).
Cells were stained with monoclonal antibodies (MoAbs) and analyzed by FACSCalibur (BD Biosciences). The following MoAbs were purchased from BD Biosciences: fluorescein isothiocyanate (FITC)-conjugated CD14 (FITC-CD34), FITC-CD34, FITC-CD45, FITC-CD90, phycoerythrin (PE)-CD31, PE-CD44, PE-CD166, allophycocyanin (APC)-CD29, APC-CD105, and biotin-NGF-R (nerve growth factor-receptor). FITC-mouse IgG1, PE-mouse IgG1, APC-mouse IgG1, biotin-mouse IgG1, FITC-mouse IgG2a, and PE-rat IgG2b were also purchased from BD Biosciences and were used as isotype controls. APC-streptavidin (eBiosciences, San Diego, http://www.ebiosciences.com/) was used to detect biotin-conjugated MoAbs. A MoAb against STRO-1 was purchased from R&D Systems. PE-conjugated rat anti-mouse IgM (eBiosciences) was used to detect anti-STRO-1 MoAb. Cell viability was monitored by propidium iodide (Sigma) staining. Flow cytometry data were analyzed using CellQuest (BD Biosciences) analysis software.
Fundamental Features of Tested Cells
We studied 25 different human primary fibroblast-like cell populations, including the commonly studied WI-38 cells, which had been deposited into an archived tissue culture collection. The cell populations were derived from lung, skin, and umbilical cord. All cell populations were considered primary fibroblast cultures. The number of cell doublings of the WI-38 and NB1RGB cells was 28 and 47, respectively. The number of cell doublings of the other cell populations had not been registered in the archived tissue culture collection. Because most of the cell populations stopped proliferating after 14 to 48 cell doublings in our laboratory (Table 1), we conclude that none of these cell populations was immortalized. The majority of the cells in each cell population exhibited a normal karyotype (Table 1). In addition, we also studied five independent primary fibroblast-like cell populations that we obtained from human amniotic membranes following only three or four cell passages.
Table Table 1.. Karyotype analyses and proliferative capacities
To induce osteogenic differentiation, cells were cultured on type I collagen-coated dishes in differentiation medium for 4 weeks (as described in Materials and Methods) . Deposition of calcium, an indicator of osteogenic differentiation, was determined by alizarin red S staining (Fig. 1). Using this method, we detected osteogenic differentiation in 19 cell populations (supplemental online Fig. 1). We further subjected the cells to analysis of alkaline phosphatase activity, which is an independent indicator of osteogenic differentiation, and the results were consistent with those of the alizarin red S staining assay (data not shown).
To confirm further the presence of osteogenic differentiation, we evaluated the expression of osteocalcin, a protein specifically produced in osteogenic cells, using an enzyme-linked immunosorbent assay (as described in Materials and Methods). Although the expression of osteocalcin was not strongly correlated with the degree of alizarin red S staining in the various populations, the results generally correlated with the results of the alizarin red S staining assay (Fig. 1B, supplemental online Fig. 1). These results strongly suggest that progenitor cells that can differentiate into osteogenic cells are present in various tissues but not in umbilical cord.
To induce chondrogenic differentiation, a cell pellet was incubated in differentiation medium for 4 weeks (as described in Materials and Methods) . The production of mucopolysaccharide, an indicator of chondrogenic differentiation, was measured by staining thin sections of the pellets with toluidine blue (Fig. 2A, 2B). We found that most of the cell populations analyzed in this assay appeared to differentiate into chondrogenic cells (supplemental online Fig. 2).
To confirm chondrogenic differentiation, we evaluated the expression of type II collagen, a gene specifically expressed in chondrogenic cells, using RT-PCR. The expression of type II collagen was not strongly correlated with the degree of toluidine blue staining in the various populations (Fig. 2C, supplemental online Fig. 2). For example, the expression of type II collagen was very low in cells derived from HUC-F cell populations (Fig. 2C), whereas the pellet derived from this cell population was stained clearly by toluidine blue (supplemental online Fig. 2). Although the discrepancy between the results of the two assays remains to be studied further, progenitor cells that can differentiate into chondrogenic cells seem to be present in various tissues.
To induce adipogenic differentiation, cells were treated as described previously (as described in Materials and Methods) . Following treatment for 4 weeks, the cells were stained with oil red O to detect lipid production. Many cells possessing abundant lipid (adipocytes) were observed after adipogenic differentiation of many of the cell populations derived from skin (Fig. 3A). In other cases, exposure of cells to the differentiation protocol did not give rise to significant numbers of adipocytes (Fig. 3B). Although the numbers of adipocytes induced in the cell populations varied, cells derived from skin were prone to differentiate into adipocytes (supplemental online Fig. 3).
To confirm adipogenic differentiation, we used RT-PCR to evaluate the expression of lipoprotein lipase (LPL) and fatty acid binding protein aP2 (aP2), which are specifically expressed in adipogenic cells. The expression of both LPL and aP2 tended to correlate with the degree of oil red O staining (Fig. 3C, 3D, supplemental online Fig. 3). Progenitor cells that can differentiate into adipocytes appear to be present in various tissues, but not in amniotic membrane.
Effect of Culture Conditions on Differentiation Potential
We next compared the effect of four different culture media on differentiation of cells derived from the amniotic membrane (as described in Materials and Methods). We observed no difference in the ability of cells cultured in the four different media to generate mesenchymal progenitors (Figs. 1B, 2C, 3C, 3D). Culture of cells in α-MEM supplemented with 10% FCS was sufficient to obtain mesenchymal progenitors.
Next, we evaluated the effect of long-term culture on the differentiation potential of several cell populations. We observed that six cell populations exhibited the potential to differentiate into the three lineages in the differentiation assays described above: HFL-AE-III, HFSKF-AE-V, NB1RGB, PWS-Yamaguchi, SF8536, and SF8543. We cultured HFSKF-AE-V, NB1RGB, PWS-Yamaguchi, SF8536, and SF8543 cell populations long-term. We also cultured long-term the SF8429 cell population, which exhibited the potential to differentiate into osteoblasts and adipocytes. Cells obtained from the archived tissue culture collections were therefore cultured long-term and then subjected to the differentiation assays. NB1RGB cells seemed to lose the potential to differentiate into osteoblasts (Fig. 4A) following 14 cell doublings. On the other hand, SF8536 and SF8543 cells seemed to lose the potential to differentiate into adipocytes (Fig. 4C, 4D) following 32 and 38 cell doublings, respectively. SF8429 cells also lost the potential to differentiate into adipocytes (Fig. 4C, 4D) following 36 cell doublings. These results strongly suggest that the culture period is a critical factor affecting the potential of primary cultured cells to differentiate.
Analysis of Expression of Cell Surface Molecules
To identify markers that could predict the potential of each cell population to differentiate, we analyzed the expression of cell surface molecules by flow cytometry using specific monoclonal antibodies against CD14, CD31, CD34, CD45, CD29, CD44, CD90, CD105, CD166, STRO-1, and NGF-R. None of the cell populations was found to express CD14, CD31, CD34, or CD45 (data not shown), whereas all expressed CD29 (supplemental online Fig. 4), CD44 (supplemental online Fig. 5), CD90 (supplemental online Fig. 6), CD105 (data not shown), and CD166 (supplemental online Fig. 7).
The expression of STRO-1 varied among the cell populations (supplemental online Fig. 8). Some populations did not contain any STRO-1-positive cells, whereas others did express STRO-1. However, the expression pattern of STRO-1 did not correlate with the potential of the cell populations to differentiate. For example, whereas both NB1RGB and HFSKF-AE-V cells could differentiate into all three lineages, NB1RGB cells expressed STRO-1, whereas HFSKF-AE-V cells did not. Likewise, although the expression pattern of NGF-R varied among the cell populations, it did not correlate with the differentiation potential (supplemental online Fig. 9).
We analyzed the potential of a number of primary fibroblast-like cell populations to differentiate. Most of the analyzed cell populations could differentiate into at least one mesenchymal lineage to generate osteoblasts, chondrocytes, or adipocytes. This observation is consistent with previous reports that indicated that MSCs and MPCs were present in various tissues [1, 11, , , , , , –18]. Because six cell populations (HFL-AE-III, HFSKF-AE-V, NB1RGB, PWS-Yamaguchi, SF8536, and SF8543) could differentiate into all three lineages, these cell populations might comprise MSCs. On the other hand, cell populations that showed the potential to differentiate into one or two lineages may represent MPCs.
Muraglia et al. conducted a detailed analysis of human MSC clones derived from bone marrow , which revealed that lineage commitment was hierarchical in nature; e.g., trilineage cells gave rise to osteochondro and then osteoprogenitors but did not yield adipochondro, adipo-osteo, or chondroprogenitors. Consistent with this observation, most of the bipotential cell populations differentiated into osteogenic and chondrogenic cells, including HFL-II, HFL-III, SF8428, AM no. 11-3, 12-1, 17-5, 17-9, and 19-4. In contrast, however, we observed that SF8403 and SF8433 cells could differentiate into adipogenic and chondrogenic cells and that SF8429 cells could differentiate into adipogenic and osteogenic cells. Because Muraglia et al. analyzed only MSCs derived from bone marrow, the hierarchy of lineage commitment of MSCs may differ among MSC populations derived from different tissues . Another possibility is that SF8403, SF8433, and SF8429 cells comprise two different populations of monopotent progenitor cells. Likewise, it is possible that the cells that possessed bi- or trilineage differentiation potential comprised two or three distinct populations of monopotent progenitor cells. This possibility cannot be ruled out without clonal expansion analysis of each cell population. However, because many reports have shown the presence of MSCs in various tissues [1, 11, , , , , , –18], we believe that most if not all of the populations that have shown trilineage differentiation potential included MSCs.
The ability to predict the potential of a cell population to differentiate into a particular lineage should facilitate clinical application of the cells. However, the cell surface markers we analyzed in this study were unable to predict the potential of the cell populations to differentiate into specific lineages. Gene expression profiling may be useful to correlate the pattern of gene expression with the differentiation profile of each cell population [28, 29]. Although it is still controversial, it has been suggested that MSCs could differentiate into endodermal lineages such as hepatocytes [7, 30]. Gene expression profiling may also be useful to identify cell populations that can differentiate into endodermal cells. Such gene expression profiling should be performed using a number of clonal MSC and MPC populations derived from various human tissues. Many cell populations that have been deposited and registered as fibroblasts in archived tissue culture collections may include abundant MSCs and MPCs and thus may be useful for these investigations.
The ability of different cell populations to differentiate into particular lineages appeared to depend on the source tissue. Cells derived from lung could differentiate efficiently into osteoblasts and chondrocytes but only rarely into adipocytes, whereas cells derived from skin could differentiate into osteoblasts, chondrocytes, and adipocytes. Although cells derived from umbilical cord could differentiate into chondrocytes and adipocytes, the degree of differentiation appeared to be lower than that of cells derived from other tissues. Finally, cells derived from amniotic membrane could differentiate into osteoblasts and chondrocytes but only rarely into adipocytes. Bone marrow is often a source of MSCs and MPCs, although it is usually very difficult to obtain sufficient numbers of cells from the bone marrow of aged individuals. As shown in this study, cells derived from the skin of adults aged 26–72 years of age could differentiate into osteoblasts, chondrocytes, and/or adipocytes. Thus, skin may serve as an additional source of MSCs and MPCs in aged individuals for use in therapies based on regenerative medicine.
As shown in the study of cells derived from amniotic membrane, the ability of cells to differentiate into the various mesenchymal lineages seems not to depend on the nature of the medium (Figs. 1B, 2C, 3C, 3D), although further study is necessary. On the other hand, other culture conditions can have a strong effect on the differentiation potential of cells. Inappropriate culture practices, such as infrequent change of media or overgrowth of cultures, often result in the loss of the ability of cells to differentiate. Prolonged culture of cells can also result in loss of differentiation potential. In fact, long-term culture of cell populations characterized by a trilineage differentiation potential reduced that potential (Fig. 4). Taken together, it is possible that although many or all primary fibroblast-like cell populations may initially have exhibited a capacity to differentiate into mesenchymal lineages, and this potential may have been lost as a result of subsequent culture conditions. In other words, many so-called fibroblast populations may in fact be derived from MSCs or MPCs. Thus, the pluripotent stem cells (embryonic stem cell-like cells) induced in “fibroblast cultures” by defined factors  might have been derived from cells possessing some degree of pluripotency, such as MSCs.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflict of interest.
We thank all members in the Cell Engineering Division for their help, discussion, and secretarial assistance. This work was supported by grants from the Ministry of Education, Culture, Sports, Science, and Technology in Japan. K.M. was supported by the Junior Research Associate grant from RIKEN.