Mesenchymal stem cells derived from human induced pluripotent stem cells retain adequate osteogenicity and chondrogenicity but less adipogenicity

Introduction Previously, we established a simple method for deriving mesenchymal stem cells (MSCs) from human induced pluripotent stem cells (iPSC-MSCs). These iPSC-MSCs were capable of forming osteogenic structures in scaffolds and nanofibers. The objective of this study is to systematically characterize the mesenchymal characteristics of the iPSC-MSCs by comparing them to bone marrow-derived MSCs (BM-MSCs). Methods Two iPSC-MSC lines (named as mRNA-iPSC-MSC-YL001 and lenti-iPSC-MSC-A001) and one BM-MSC line were used for the study. Cell proliferation, presence of mesenchymal surface markers, tri-lineage differentiation capability (osteogenesis, chondrogenesis, adipogenesis), and expression of “stemness” genes were analyzed in these MSC lines. Results The iPSC-MSCs were similar to BM-MSCs in terms of cell morphology (fibroblast-like) and surface antigen profile: CD29+, CD44+, CD73+, CD90+, CD105+, CD11b–, CD14–, CD31–, CD34–, CD45– and HLA-DR–. A faster proliferative capability was seen in both iPSC-MSCs lines compared to the BM-MSCs. The iPSC-MSCs showed adequate capacity of osteogenesis and chondrogenesis compared to the BM-MSCs, while less adipogenic potential was found in the iPSC-MSCs. The iPSC-MSCs and the tri-lineage differentiated cells (osteoblasts, chondrocytes, adipocytes) all lack expression of “stemness” genes: OCT4, SOX2, GDF3, CRIPTO, UTF1, DPPA4, DNMT3B, LIN28a, and SAL4. Conclusions The MSCs derived from human iPSCs with our method have advanced proliferation capability and adequate osteogenic and chondrogenic properties compared to BM-MSCs. However, the iPSC-MSCs were less efficient in their adipogenicity, suggesting that further modifications should be applied to our method to derive iPSC-MSCs more closely resembling the naïve BM-MSCs if necessary.


Introduction
Mesenchymal stem cells (MSCs) have been widely used for tissue engineering due to their unique tri-lineage differentiation properties: osteogenesis, chondrogenesis, and adipogenesis [1,2]. One attractive application of MSCs for tissue engineering is to functionalize biomaterials scaffolds by MSC co-culturing and differentiation. This can facilitate the replacement and integration of scaffolds into repairing tissues [3][4][5]. Currently, adult bone marrowderived mesenchymal stem cells (BM-MSC) seem to be the most reliable cell source for tissue engineering [6]. However, limited proliferating capability of the BM-MSCs and their replicative senescence after in vitro expansion hamper their applications in a wide range of clinical applications [7].
Stem cell-derived MSCs have emerged as a significant alternative cell source to BM-MSCs. Human embryonic stem cells were mostly used for MSC derivation (ESC-MSCs), but concerns from ethical, pathological, and immunological perspectives have greatly limited the broad use of ESC-MSCs for clinical interventions. Induced pluripotent stem cells (iPSCs) generated from somatic cells represent a potentially inexhaustible cell resource with a pluripotent potential similar to ESCs. These iPSCs have the potential for patient-specific autologous therapies for diseases resulting from injury, genetic predisposition and/or aging. This approach eliminates most problems of immune incompatibility and ethical concerns. Since the first report of iPSC generation from mice and human fibroblasts by retrovirus-mediated reprogramming in 2006 and 2007 [8,9], several methods have been established with improved safety and efficacy, such as protein transduction, mRNA reprogramming, and transgene free methods with a cocktail of small molecules [10][11][12].
To simplify the MSC generation procedure, we have previously established a method for iPSC-MSC derivation, which involves culturing the iPSCs directly in MSC differentiation medium followed by serial trypsinization-based passaging [16]. We have shown that the iPSC-MSCs derived by this method obtained a typical mesenchymal surface marker profile (CD73+, CD90+, CD105+, CD34-, CD45-) and were capable of osteogenesis in tissue culture vessels, synthetic three-dimensional and nanofibers. They also lack tumorigenicity when transplanted into immunodeficient mice [4,16].
Although the iPSC-MSCs derived by our method retain the mesenchymal functions, their multipotent efficacy as compared to naïve BM-MSCs is still unknown. To further characterize the efficacy of iPSC-MSCs derived by our method, we now compare the proliferative rate, mesenchymal marker profile (CD73+, CD90+, CD105+, CD34-, CD45-), tri-lineage differentiation capacity (osteogenesis, chondrogenesis, and adipogenesis), and pluripotency of two iPSC-MSC lines and one BM-MSC line.

Lentivirus-mediated reprogramming
NHDFs (1.5 × 10 5 cells/per well) were seeded in a sixwell plate 1 day before transduction. Cells were transduced with reprogramming lentivirus in the presence of polybrene (8 μg/mL) in D10 medium. Cell media were changed every other day. Six days post-transduction, transduced NHDFs (2 × 10 4 cells/per well) were harvested by trypsinization and seeded on irradiationinactivated mouse feeder cells in six-well plates, and cultured in KSR medium. KSR medium was changed daily. Approximately 21 days post-transduction, the iPSC colonies were ready for picking and expansion.

Immunofluorescence staining
For immunofluorescence staining, cells were fixed in 4 % paraformaldehyde for 20 min, followed by PBS wash (three times, 5 min each) and permeabilization with 0.3 % Triton X-100 in PBS for 10 min. The cells were then blocked with blocking solution (5 % donkey serum in PBS) at room temperature for 30 min and incubated with the primary antibodies overnight at 4°C. Goat antihuman OCT3/4 (Abcam, ab27985, 100× diluted) and rabbit antihuman Nanog (Abcam, ab80892, 100× diluted) were used. Cells were then stained with a secondary antibody for 2 h. Alexa 594 donkey anti-goat IgG (H + L) and Alexa Fluor® 488 Donkey Anti-Rabbit IgG (H + L) (Life Technologies) were used for second antibody staining. For live cell staining of TRA-1-60 and CD44, cells were stained using the live cell imaging kit from Life Technologies (Tra-1-60 AF594, CD44 AF488) according to the manufacturer's protocol. All images were taken with a Leica fluorescence microscope.

Derivation of MSCs from iPSCs generated by lentiviral reprogramming
The iPSC-MSC derivation was performed according to our previous protocol. One characterized pluripotent lenti-iPSC line was used for MSC differentiation. Briefly, 3 days after passaging the lenti-iPSCs to feeder cell culture, the KSR medium was replaced with MSC medium. The lenti-iPSCs were maintained in MSC medium for 2 weeks, with medium changed every other day. Subsequently, cells were passaged to gelatin-coated (0.1 % gelatin, room temperature for 2 h) tissue culture vessels by trypsinization (0.25 % trypsin/1 mM EDTA). Cells were defined as passage 1 (P1) after the first passaging. For maintenance of iPSC-MSCs, cells were passaged when 90 % confluent and seeded with a density of 1.6 × 10 4 cells/cm 2 to new tissue culture vessels.

MSC surface marker characterization by flow cytometry
Detail information on antibodies against the human antigens CD11b, CD14, CD29, CD31, CD34, CD44, CD45, CD73, CD90, CD105 and HLA-DR are shown in Table 1. Cells were harvested by trypsinization and washed with 2 % FBS-PBS twice; 2 × 10 5 cells were re-suspended in 100 μl 2 % FBS-PBS and incubated with the conjugated antibody for 30 min at room temperature in the dark. Stained cells were then washed with 2 % FBS-PBS twice and re-suspended in 500 μl 1 % formaldehyde-PBS for flow cytometry analysis (LSRFortessa); 10,000 events were recorded for each sample and data were analyzed with Flowjo.

Number of population doublings
BM-MSCs, mRNA-iPSC-MSC-YL001 and lenti-iPSC-MSC-A001 were serially subcultured in six-well plates for 2 weeks. Briefly, cells were cultured in six-well plates; when cells were 90 % confluent, they were trypsinized and counted with a hemocytometer. The cells were then cultured in a new six-well plate with the same starting cell density. The number of population doublings (NPD) was calculated according to the following equation: NPD = log10 (Nh/Ni) × 3.33, where Nh and Ni are the numbers of harvested and initiating cells, respectively [19]. NPDs were performed in six replicates for each type of MSC.
Adipogenic differentiation was carried out by culturing the cells at a density of 1.2 × 10 4 cells/cm 2 in DMEMhigh glucose (Invitrogen) supplemented with 15 % horse serum (Sigma), and 100 nM dexamethasone (Sigma), 1 % penicillin/streptomycin (MSC growth medium as control). After 3 weeks, cells were stained with Oil-red O (Sigma-Aldrich) solution to evaluate differentiation using standard techniques as previously described [22].

PCR and quantitative PCR
Total RNA was extracted with an RNeasy Plus Mini Kit (Qiagen) according to the manufacturer's instructions, followed by quantification with a Nanodrop spectrophotometer and qualification by 1 % agarose gel electrophoresis. First-strand cDNA was synthesized with approximately 50 ng of total RNA using a cDNA synthesis kit (170-8891, Bio-Rad). PCR primers for stemness genes and osteogenic genes are listed in Table 2.
For quantitative PCR of osteogenic genes (of runtrelated transcription factor 2 (RUNX2), alkaline phosphatase (ALP), collagen type I alpha 1 (COL1A1), and osteocalcin (OC)), 2 uL (five times diluted) cDNA was used. Quantitative PCR was performed using the Light-Cycle 480 SYBR Green I Master kit on LightCycler 480 (Roche) with the following PCR program: 1 hold at 95°C for 5 min; 50 cycles at 95°C for 10 seconds, 60°C for 10 seconds, and 72°C for 15 seconds. To mask the signal from primer dimers, fluorescent signal was measured at 77°C for RUNX2, 80°C for ALP, 82°C for GAPDH and COL1A1, and 84°C for OC. Relative gene expression was calculated using the 2 -ΔΔct method after normalization to the reference gene GAPDH.

Statistics
All values are reported as mean ± SD. Data were checked for normal distribution using QQ-plots. The osteogenic gene expression between growth and osteogenic media were compared by t-test in each cell line. Quantitative calcium deposition was investigated using two-way analysis of variance (media type × cell line). If interactions were observed, the variables were investigated separately by the Scheffe method. P < 0.05 was considered statistically significant.

MSCs from iPSCs induced by lentiviral-mediated reprogramming have the same fibroblast-like morphology as BM-MSCs
The procedure for iPSC-MSC derivation, which we had established previously, was based on an iPSC line established by mRNA reprogramming from human fibroblasts [16]. It is unknown whether this procedure could be applied to derive functional MSCs from iPSC lines generated by other methods. Lentiviral-mediated reprogramming is favored for its high reprogramming efficiency relative to other methods [17]. We firstly generated one iPSC line using lentiviral-mediated reprogramming. The iPSC line exhibited typical stem cell morphology; expressed pluripotent cell cells markers: TRA1-60 (podocalyxin), OCT4 (octamer-binding transcription factor 4, also known as POU domain, class 5, transcription factor 1), and NANOG (Nanog homeobox) (Fig. 1a); lacked expression of CD44 (the absence of which is crucial for iPSC pluripotency (Fig. 1a)) [23]; and retained normal polyploidy (diploid as fibroblasts, Fig. 1b). We differentiated the iPSC colony into MSCs using our preciously established procedure [16]. The MSCs derived from lentivial reprogramming iPSCs (lenti-iPSC-MSCs-A001) had the same morphology (fibroblast-like) as the BM-MSCs and the mRNA-iPSC-MSC-YL001 line reported previously [16] (Fig. 1c), and are karyotypically normal (Fig. 1d).

The iPSC-MSCs obtained similar mesenchymal surface marker profiles to BM-MSCs through in vitro cultivation
To verify whether the iPSC-MSCs harbor mesenchymal characteristics, we analyzed the expression of five MSCpositive surface markers-CD29 (Integrin beta-1), CD44   (Fig. 2). Among the mRNA-iPSC-MSC-YL001 cells, robust CD90 and CD105 expression (89.18 % and 87.47 % positivity at P14, respectively) were observed, which is consistent with our previous observation [16]. Thus, it appears that lenti-iPSC-MSC-A001 (late passage) and  (Fig. 2). Based on the finding of passagedependent mesenchymal enrichment in the lenti-iPSC-MSC-A001, cells with over 6 passages were used for the following functional characterizations.

The iPSC-MSCs possess fast proliferative capability
The iPSC-MSCs have become a promising substitute for BM-MSCs or MSCs derived from other adult tissues.
The advantage of using iPSC-MSCs is attributed to their higher self-renewal capability. We compared the proliferation rate of the two iPSC-MSC lines to BM-MSCs using the Click IT-based proliferation assay and flow cytometry. K562 cancer cells were used a positive control for the assay (55.7 % proliferation rate, Fig. 3a). Higher  Fig. 3d). The fast proliferation capacity of iPSC-MSCs was further validated by population doubling array (Fig. 3e).

iPSC-MSCs possess adequate osteogenic and chondrogenic, but less adipogenic capabilities, compared with BM-MSCs
According to the International Society for Cellular Therapy, the capability of tri-lineage differentiation in vitro is one of the major criteria for defining MSCs [24]. We have previously shown that the mRNA-iPSC-MSC-YL001 line is capable of tri-lineage differentiation, with a focus on osteogenesis on synthetic biocompatible scaffolds [16]. However, the efficacy of tri-lineage differentiation as compared to BM-MSCs is yet to be addressed. Thus, we compared the osteogenic, chondrogenic and adipogenic differentiation capabilities of the two iPSC-MSC lines relative to the BM-MSC line. After 3 weeks of osteogenic induction, calcium mineralization was observed in all three MSC lines by Alizarin Red S staining (Fig. 4a). Mineralization was not observed in the MSCs cultured in normal growth medium. A more homogeneous and robust calcium staining was seen in the lenti-iPSC-MSC-A001 than in the mRNA-iPSC-MSC-YL001. This observation was further confirmed by quantification of calcium content (Fig. 4b).
The lenti-iPSC-MSC-A001 line has an efficacy of osteogenesis similar to BM-MSCs. However, approximately twofold less osteogenic capacity was found in the mRNA-iPSC-MSC-YL001 cultures, which is likely due to less positivity of CD90 (Fig. 2) [25]. Increased expression of four osteogenic genes (runt-related transcription factor 2 (RUNX2), alkaline phosphatase (ALP), collagen type I alpha 1 (COL1A1), and osteocalcin (OC)) was detected in all MSCs after 3 weeks of osteogenic differentiation  (Fig. 4c). The induced expression of RUNX2, ALP and OC was more robust in BM-MSCs relative to the iPSC-MSCs, whereas Col1a1 was more strongly expressed in the iPSC-MSC cultures. Taken together, these results suggest that iPSC-MSCs derived with our method can achieve an adequate osteogenic efficacy similar to that of BM-MSCs.
For chondrogenic differentiation, the lenti-iPSC-MSC-A001, mRNA-iPSC-MSC-YL001 and BM-MSC lines were grown in a "V" shape 96-well plate according to a protocol described previously [26]. Toluidine blue staining of pellet sections was used to assess chondrogenic differentiation. Cartilaginous extracellular matrix is stained purple while undifferentiated or fibrous tissue is stained blue. After Fig. 4 Comparative analyses of the osteogenic capacity of the iPSC-MSCs and BM-MSCs. a Alizarin Red staining of calcium in the BM-MSCs, the mRNA-iPSC-MSC-YL001, and the lenti-iPSC-MSC-A001 with growth medium or osteogenic medium for 3 weeks. Images are representative of six repeats. In the left corner there is an image from one well of the 96-well plate. Scale bar for the microscopic images is 300 μm. b Quantitative calcium deposition of the BM-MSC, mRNA-iPSC-MSC-YL001, and lenti-iPSC-MSC-A001 cells after 3 weeks in the MSC growth medium or osteogenic medium (n = 8). c Quantitative PCR analysis of the expression of osteogenic genes RUNX2, ALP, COL1A1, and OC in the BM-MSC, mRNA-iPSC-MSC-YL001, lenti-iPSC-MSC-A001 after 3 weeks in the MSC growth medium or osteogenic medium (n = 6). *P < 0.05. BM bone marrow, iPSC induced pluripotent stem cell, MSC mesenchymal stem cell 3 weeks of chondrogenic differentiation, cartilaginous extracellular matrix was detected in all MSC lines cultured in chondrogenic medium (Fig. 5d-f), but not in growth medium (Fig. 5a-c). The lenti-iPSC-MSC-A001 and mRNA-iPSC-MSC-YL001 had formed similar cartilaginous structures compared with the BM-MSCs.
The adipogenic potential of the iPSC-MSCs was also characterized by in vitro adipogenic differentiation. Lipid droplets formed by adipogenesis were assessed by Oil Red O staining. After 3 weeks of adipogenic differentiation, lipid droplets were formed in both iPSC-MSC lines and the BM-MSCs in the adipogenic differentiation  (g and j), the mRNA-iPSC-MSC-YL001 (h and k), and the lenti-iPSC-MSC-A001 (i and l) after 3 weeks in MSC growth medium (g-i) or adipogenic medium (j-l). Arrow heads indicate the cartilaginous extracellular matrix (d-f) and lipids (j-l). Scale bar: a-f, 150 μm; g-l, 300 μm medium (Fig. 5j-l) but not in the growth medium ( Fig. 5g-i). However, the adipogenic efficacy of both iPSC-MSC lines was much lower than the BM-MSCs. Lipid droplets from iPSC-MSCs were much smaller than those from BM-MSCs.
The iPSC-MSCs and tri-lineage differentiated cells lack expression of stemness genes Loss of "stemness" in iPSC-MSCs and iPSC-MSCderived lineages is one crucial safety criterion when applying iPSC-MSCs to tissue engineering and cell therapy. Lack of tumorigenic potential of the mRNA-iPSC-MSC-YL001 line was validated previously [16]. We further characterized the loss of "stemness" in the iPSC-MSC lines by analyzing the expression of 11 stemness genes (OCT4, KLF4, C-MYC, SOX2, GDF3, CRIPTO, UTF1, DPPA4, DNMT3B, LIN28A and SALL4) in the original iPSCs, iPSC-MSCs, and tri-lineage differentiated cells (osteogenesis, chondrogenesis, and adipogenesis). All stemness genes were expressed in the iPSCs, whereas most stemness genes, except Klf4 and C-Myc, were not expressed in the iPSC-MSCs and the tri-lineage differentiated cells (Fig. 6). We also found that Klf4 and c-Myc were expressed in the NHDF, which is similar to what has been reported previously [8,27].

Discussion
With technical improvements in generating pluripotent stem cells, differentiation of iPSCs to functionally specific cell types remains a major challenge. MSCs are a promising cell type for personalized therapies such as tissue engineering [28], immunosuppressive biotherapy [29], and regenerative organ repair [30]. We have previously developed a simple protocol to derive iPSC-MSCs and applied them for bone tissue engineering in synthetic three-dimensional scaffolds and nanofibers [4,16]. Here, we compared two iPSC-MSC lines (derived with our protocol) to naïve BM-MSCs. This study demonstrated that the iPSC-MSCs derived by our protocol display many similar characteristics relative to the BM-MSCs, including fibroblast-like morphology, expression of typical mesenchymal markers, the capacity to efficiently differentiate into osteogenic and chondrogenic lineages, and loss of pluripotency. This study further supports our previous finding of efficient osteogenic potential for iPSC-MSCs in three-dimensional scaffolds and nanofibers [4,16].
There are several advantages of using iPSC-MSCs for MSC-based therapies as compared to primary MSCs. The high proliferative capability iPSC-MSCs proven by this study (Fig. 3) and others [31][32][33], as well as the infinite self-renewal capacity of iPSCs [8], increase the potential of MSC-based cell therapies, which commonly requires repetitive administration of a large number of MSCs. Although MSCs can be easily isolated from various sources of adult tissues, such as BM, fat, placenta, blood, spleen, heart, pancreas, joints and so forth [34][35][36], this repetitive MSC isolation procedure would cause pain and risk of infection for patients. Furthermore, for certain patients with inherited genetic problems or age-related disorders, iPSC-MSCs would represent the best solution for gene-and cell-based therapy. One such example is Osteogenesis imperfect (OI), which is caused by dominant mutations of the type one collagen genes. With an adeno-associated virus-mediated gene targeting strategy, Deyle et al. generated iPSC-MSCs derived from OI patients and subsequently inactivated the mutated COL1A1 and COL1A2 genes in the iPSC-MSCs. The targeted iPSC-MSCs were able to restore normal collagen production and bone formation [37].
In terms of osteogenic and chondrogenic capacities of the iPSC-MSCs, we now have further confirmed their adequate efficacy relative to BM-MSCs [4,16]. Although only the osteogenicity was thoroughly investigated in this study due to our research focus, the chondrogenicity of iPSC-MSCs has been systematically confirmed by studies from other research groups [38,39].
However, this study also showed that both iPSC-MSC lines were less efficient with respect to adipogenic differentiation relative to their osteogenicity and chondrogenicity (Fig. 5). Although both iPSC-MSC lines derived by our protocol were capable of adipogenesis [16], the adipocytes and lipid droplets were much smaller in the cultures of iPSC-MSCs compared with those of BM-MSCs. Similarly, iPSC-MSCs derived by other methods have been reported to have less adipogenicity than BM-MSCs [32,40]. A similar observation was reported for MSCs derived from human ESCs [41]. It is still unknown why MSCs derived from iPSCs or ESCs show less commitment towards adipogenicity. Many intrinsic factors could affect the adipogenicity of iPSC-MSCs, such as the iPSC-MSC derivation and cultivation methods, iPSC-MSC purity, and tissue of origin. Hynes et al. recently compared iPSC-MSCs originating from three different tissues: gingiva, periodontal ligament, and lung [42]. Although the efficacy of adipogenesis of those iPSC-MSCs was not assessed in their study, various capacities were observed among the three iPSC-MSCs lines, suggesting the notion that the tissue of origin plays a vital role in commitment differentiation. In our study, the lenti-iPSC-MSC-A001 had a higher adipogenic capacity than the mRNA-iPSC-MSC-YL001. However, both iPSC-MSCs lines were derived from dermal fibroblasts. Thus, the adipogenic difference may be due to clonal variation, which has been reported in other studies [42,43].
According to ClinicalTrail.gov to date, there are 452 studies registered for "mesenchymal stem cells", 32 studies registered for "induced pluripotent stem cells", and 27 studies registered for both "mesenchymal stem cells" AND "induced pluripotent stem cells". One key concern when using iPSCs and MSCs derived from iPSCs for cell therapy is their potential for teratoma formation. We have shown that the iPSC-MSCs generated by our method do not form teratoma after transplanting into immune deficient mice [16]. In this study, we further characterized a few "stemness" genes in the iPSC-MSCs. Except for KLF4 and C-MYC, the rest of the "stemness" genes (OCT4, SOX2, GDF3, CRIPTO, UTF1, DPPA4, DNMT3B, LIN28A, SALL4) were not expressed in the iPSC-MSC lines, further implicating the safety aspect of the iPSC-MSCs. KLF4 and C-MYC were also expressed in the BM-MSCs and in primary fibroblasts. Similarly, KLF4 and C-MYC were reported to be expressed in human fibroblasts by a previous study [8], suggesting that the presence of KLF4 and C-MYC should not contribute to teratoma formation by the iPSC-MSCs. With more studies conducted to characterize the safety and efficacy aspects of iPSC-MSCs, increasing numbers of clinical trials using iPSC-MSCs are expected in the future. We also observed an expression difference for KLF4 and C-MYC between the mRNA-iPSC-MSC-YL001 and lenti-iPSC-MSC-A001 after osteogenic and adipogenic differentiation. Since KLF4 and C-MYC are expressed in mRNA-iPSC-MSC-YL001 and lenti-iPSC-MSC-A001 cells, the expression of KLF4 and C-MYC in osteogenic and adipogenic samples might have resulted from the proportion of parental iPSC-MSCs that failed to differentiate. The difference between the two iPSC-MSC lines could have resulted from the different mesenchymal heterogeneity of these two iPSC-MSCs as shown in Fig. 2.
Another important challenge for generating clinically compliant iPSC-MSCs is the time-effective generation of a large amount of functional MSCs. Several methods have been described to derive functional MSCs from human iPSCs. These methods vary from each other in terms of ingtermediate embryoid bodies, surface coating of tissue culture vessels, differentiation medium, and supplement of pathway inhibitors (e.g., the p38 MAPK inhibitor, SB203580 and the transforming growth factor-β pathway inhibitor SB431542) [32,44]. The use of molecule SB431542 did not improve the adipogenic capacity of iPSC-MSCs [32]. However, robust adipogenesis of iPSC-MSCs has been achieved by supplementing the SB203580 inhibitor [44]. Our method for iPSC-MSC derivation involves neither intermediate embryoid bodies nor the use of pathway inhibitors. The derivation of iPSC-MSCs is achieved by MSC medium-mediated pre-differentiation, followed by MSC enrichment by serial passaging in MSC culture condition, which take approximately 1 month [4,16]. A similar method has been applied to derive functional MSCs by other groups [42,45]. Compared to the embryoid body-based differentiation method, this method is more time-and costeffective, and compliant with large-scale production. However, as proved in this study, further improvements are still required for our iPSC-MSC method to generate functional MSCs that can resemble the naïve BM-MSC functions.

Conclusions
The iPSC-MSCs derived by our method display most mesenchymal characteristics similar to adult BM-MSCs. The iPSC-MSCs achieve robust osteogenicity and chondrogenicity, but have very limited adipogenicity. Our results suggest that these iPSC-MSCs present a promising alternative cell source for MSC-based therapeutic applications.