Cyclin D2 Promotes the Proliferation of Human Mesenchymal Stem Cells

Background: Human mesenchymal stem cells (hMSCs) hold promise for use in cell-based therapies and tissue engineering. Although hMSCs are thought to be stable ex vivo, it is possible that they undergo an undesirable transformation to a phenotype of unlimited proliferation during ex vivo. In this study, we searched for the factor required for unlimited proliferation of hMSCs. Methods: Changes in gene expression were evaluated between hMSCs and Ewing’s sarcoma cell lines, which may be derived from hMSCs, using GeneChip Human Genome U133 plus 2.0 Array. A gene up-regulated by at least 10-fold in Ewing’s sarcoma cell lines, Cyclin D2, was overexpressed in hMSCs by a lentiviral vector. Results: Overexpression of Cyclin D2 in hMSCs altered cell morphology and promoted cell proliferation. Expression of transforming growth factor-b2 (TGF-b2), which induces senescence in hMSCs, was down-regulated in Cyclin D2overexpressing hMSCs. Furthermore, Gene Ontology analysis revealed that Cyclin D2 overexpression activated expression of genes associated with proliferation and interphase. Conclusions: Cyclin D2 promotes hMSC proliferation and is a candidate biomaker for hMSC transformation.


Introduction
Mesenchymal stem cells (MSCs) self-replicate and differentiate into a variety of cell types such as osteoblasts, chondrocytes, adipocytes, and smooth muscle cells [1][2][3][4][5]. These capacities have made MSCs useful in studies of bone and cartilage regeneration [6][7][8]. One of the sources of human MSCs (hMSCs) is adult bone marrow, although they occur at a rate of one per one-hundred-thousand nucleated cells [6], and the available volume of bone marrow is limited. To secure the numbers of hMSCs required for tissue regeneration, the cell must be expanded ex vivo. Although hMSCs are stable ex vivo, it is possible that they undergo transformation to an unlimited proliferation phenotype during expansion.
Previous studies have demonstrated that Ewing's sarcoma is derived from MSCs [9][10][11][12]. Ewing's sarcoma is a malignancy that primarily affects children and young adults, with a peak incidence between the ages of 14 and 20 years. It arises mainly in bone and less commonly in soft tissues. The t(11;22)(q24;q12) chromosomal translocation generating EWS-FLI-1 fusion gene is found in 85% of cases [13]. EWS-FLI-1 knockdown inhibits cell proliferation in Ewing's sarcoma cells [14,15]. Thus, EWS-FLI-1 expression is believed to play a key role in Ewing's sarcoma development. However, EWS-FLI-1 expression does not transform normal murine and human fibroblasts [16,17], suggesting EWS-FLI-1 promotes malignant transformation in selective cells.
Several reports have demonstrated that EWS-FLI-1 expression transforms murine MSCs; indeed, tumors form when these cells are injected into immunodeficient mice [9,12]. In contrast, EWS-FLI-1 expression in hMSCs does not accelerate cell proliferation and transformation (10). EWS-FLI-1 expression in hMSCs induces a gene expression profile that closely mimics that of Ewing's sarcoma [9][10][11] without affecting proliferation. Therefore, MSCs are thought to be the origin of Ewing's sarcoma, but because EWS-FLI-1 alone cannot transform hMSCs, we believe other factors are required for transformation.
The most important safety concern when using hMSCs in cell-

Viral infection
hMSCs were infected with the lentiviral vector containing Cyclin D2 (hMSCs/CyclinD2) or empty vector (hMSCs/Empty) at 37°C for 24 h. Infected cells were selected with 1 mg/mL puromycin for 14 days and the bulk of the resistant cells was used in subsequent experiments.

Cell proliferation
Proliferation of hMSCs/CyclinD2 and hMSCs/Empty was measured with the TetraColor ONE reagent (Seikagaku Co., Tokyo, Japan). Cultures were incubated for 2 h in medium containing the reagent. Absorbance was read at 450 nm (reference at 600 nm) on a plate reader (SH-9000, Corona Electric Co., Ibaraki, Japan).

Cyclin D2 expression in Ewing's sarcoma cell lines versus hMSCs
To identify the factors required for proliferation of hMSCs, we compared the gene expression profiles of hMSCs and four Ewing's sarcoma cell lines (Hs 822.T, Hs 863.T, RD-ES, and SK-ES-1) ( Figure 1a). Hs 822.T and Hs 863.T had similar expression profiles, as did RD-ES and SK-ES-1. The expression profiles of Hs 822.T and Hs 863.T were more similar to those of hMSCs than to those of the other Ewing's sarcoma (RD-ES and SK-ES-1). Therefore, we first compared the expression profiles of hMSCs, Hs 822.T, and Hs 863.T. We identified 44 genes that differed by at least 10-fold between hMSCs and Ewing's sarcoma cell lines (data not shown). These were narrowed to 9 genes by selecting genes that also differed from hMSCs by more than 10-fold in RD-ES and SK-ES-1 (Table 1). CCND2 (Cyclin D2) stood out in this group of 9 genes, because it represents a family of key cell-cycle regulators. Indeed, aberrant expression of

Microarray analysis
Total RNA was extracted from hMSCs and Ewing's sarcoma cell lines with the RNeasy Mini Kit (QIAGEN, Valencia, CA). Total RNA quantity and quality were assessed on an Agilent 2100 Bio-analyzer (Agilent, Santa Clara, CA); 100 ng total RNA was used to generate biotin-modified amplified RNA (aRNA) with GeneChip 3′IVT Express Kit (Affymetrix, Santa Clara, CA). Reverse transcription (RT) of firststrand complementary DNA (cDNA) with the T7 promoter sequence was performed with the T7 oligi(dT) primer. Second-strand cDNA synthesis was used to convert the single-stranded cDNA into a doublestranded DNA template by using DNA polymerase and RNase H to simultaneously degrade the RNA and synthesize second-strand cDNA. In vitro transcription of biotin-modified aRNA with IVT Labeling Master Mix generated multiple copies of biotin-modified aRNA from the double-stranded cDNA templates. The aRNA was purified and quantified; after fragmentation, it was hybridized to GeneChip Human Genome U133 Plus 2.0 Array (Affymetrix). The arrays were stained with phycoerythrin and washed at the GeneChip Fluidics station 450 (Affymetrix). The microarrays were scanned and data extracted using GeneChip scanner 3000 7G (Affymetrix); image analysis was performed using the Affymetrix GeneChip Command Console Software and digitized with the Affymetrix Expression Console.

Data processing and pathway analysis
Data analysis was performed with GeneSpring GX 11.0 software (Agilent Technologies, Santa Clara, CA). Raw data were normalized to the 50 th percentile per chip and the median per gene. Differentially expressed genes were analyzed using Ingenuity Pathway Analysis (IPA) 9.0 (Ingenuity Systems, Redwood City, CA). Fisher's exact test was used to calculate a P-value. Activation z-score was calculated as a measure of functional and translational activation in Networks and Upstream regulators analysis. An absolute z-score >2 was considered significant.

Cloning and expression of Cyclin D2
Cyclin D2 cDNA was amplified by RT-PCR of mRNA extracted from SK-ES-1 using 5´-GAATTCGCCACCATGGAGCTGCTGT-GCCACGAGG-3´ (forward; EcoR I site underlined) and 5´-CTC-GAGTCACAGGTCGATATCCCGCACG-3´ (reverse; Xho I site underlined). The amplified products were cloned into pTA2 (ToYoBo, Osaka, Japan) and verified by sequencing. The verified Cyclin D2 cDNA was cloned into the EcoR I and Xho I sites of pLVSIN-CMV Pur (TaKaRa, Shiga, Japan). Lentiviral vector was prepared with the Lenti-XTM Packaging System (TaKaRa) according to manufacturer protocols. Cyclin D2 has been associated with tumor progression in many tumor types [18][19][20][21]. We measured Cyclin D2 mRNA expression in hMSCs and Ewing's sarcoma cell lines by real-time PCR and confirmed its extreme induction in Ewing's sarcoma (Figure 1b). Hs 822.T and Hs 863.T exhibited similar Cyclin D2 expression levels; expression was higher in RD-ES and SK-ES-1.

Overexpression of Cyclin D2 promoted hMSC proliferation
We transduced the Cyclin D2 gene into hMSCs by using a lentiviral vector; an empty vector served as a negative control. At 14 days after infection when puromycin selection was completed, Cyclin D2 mRNA expression in hMSCs infected with lentiviral vector containing Cyclin D2 (hMSCs/CyclinD2) was about 10,000-fold higher than in hMSCs infected with empty vector (hMSCs/Empty) (Figure 2a). Cyclin D2 was stably expressed in hMSCs at least 56 days after infection ( Figure 2b). Next, we observed the cell morphology of the hMSCs/CyclinD2. Phase contrast microscopy revealed the normal fibroblast-like morphology of hMSCs/Empty ( Figure 2c) and smaller spread areas in hMSCs/ CyclinD2 (Figures 2c and 2d). Furthermore, proliferation of hMSCs/ CyclinD2 was faster than that of hMSCs/Empty, indicating that overexpression of Cyclin D2 promoted hMSC proliferation. However, the proliferation slowed over time and did not result in unlimited proliferation (data not shown).

TGF-b2 expression was down-regulated in hMSCs/CyclinD2, but p16 expression was not
To investigate the effects of overexpression of Cyclin D2 on the cell cycle, we examined the change in cell cycle-associated gene expression over time (p16, p21, Bmi1, TGF-b1, and TGF-b2). We did not detect significant differences in expression of p21, Bmi1, and TGF-b1 between hMSCs/CyclinD2 and hMSCs/Empty (data not shown). However, TGF-b2 expression was lower in hMSCs/CyclinD2 than in hMSCs/ Empty (Figure 3a). In addition, the increase in TGF-b2 expression during culture was suppressed in hMSCs/CyclinD2 compared with hMSCs/Empty. In contrast, the increasing rate of p16 expression in hMSCs/CyclinD2 was higher than in hMSCs/Empty, although expression in both cell types was comparable 14 days after infection ( Figure 3b).

Overexpression of Cyclin D2 altered the expression of genes associated with cell proliferation and interphase
Total RNA was extracted from hMSCs/CyclinD2 and hMSCs/ Empty 14 days after infection and analyzed by DNA microarray, which identified 690 genes that were differentially expressed by at least 2-fold between hMSCs/CyclinD2 and hMSCs/Empty        Figure 4a). Gene ontology (GO) analysis revealed these genes are associated with movement, development, growth and proliferation, cell cycle, and intercellular signaling and interactions (Table 2). Specific predictions indicated that proliferation and interphase are activated in hMSCs/CyclinD2. The induced genes that are associated with proliferation and interphase are listed in tables 3 and 4; in summary, 94 of 186 genes and 19 of 50 genes exhibited expression shifts consistent with increased in proliferation and interphase, respectively.

Discussion
Although EWS-FLI-1 expression transformed murine MSCs, expression in hMSCs did not promote cell proliferation. In this study, we found that Cyclin D2 expression was extremely high in the Ewing's sarcoma cell lines and overexpression of Cyclin D2 in hMSCs promoted cell proliferation. GO analysis also predicted that cell proliferation and interphase were activated by overexpression of Cyclin D2.
Cyclin D2 is a member of the family of D-type cyclins that mediate cell cycle regulation, differentiation, and oncogenic transformation [22,23]. D-type cyclins inactivate retinoblastoma (Rb) by phosphorylation, inducing release of E2F. Free E2F activates genes involved in the activation and maintenance of DNA synthesis. Thus, overexpression of Cyclin D2 generally has growth-promoting effects. Consistent with this notion, overexpression of Cyclin D2 in HeLa      Increased' means the genes up-or down-expression is predicted to promote proliferation of cells. 'Decreased' means the up-or down-expression is predicted to inhibit. 'Affected' means IPA could not predict whether the expression change promote or inhibit.  [24], did not promote cell proliferation (data not shown). On the other hand, increased expression of Cyclin D2 inhibits proliferation of primary human fibroblasts [25], indicating that Cyclin D2 has both positive and negative roles in the cell cycle, depending on cell type. We found that Cyclin D2 in hMSCs has a positive role in the cell cycle ( Figure 2e).
TGF-b2 expression was suppressed in hMSCs/CyclinD2 compared with hMSC/Empty during culture (Figure 3a). We previously demonstrated that hMSC growth is reduced and TGF-b2 expression increases during long-term culture [26]. We also reported that fibroblast growth factor-2 (FGF-2) stimulates hMSC growth by suppressing the up-regulation of TGF-b2 [27]. It is unclear how overexpression of Cyclin D2 suppresses the TGF-b2 increase, but this suppression may be involved in the acceleration of hMSCs/CyclinD2. In contrast, the expression of p16, which is up-regulated with aging [28], was increased in both cell types during culture, indicating that not only hMSCs/Empty but also hMSCs/CyclinD2 were aging normally. The rate of increase in hMSCs/Cyclin D2 was higher than in hMSCs/Empty (Figure 3b), suggesting that the promotion of cell proliferation in hMSCs/CyclinD2 induced cellular senescence and enhanced p16 expression. p16 is a tumor suppressor gene [29]; thus, this increase in p16 expression probably prevented unlimited proliferation. Consistent with this notion, some Ewing's sarcomas contain a homozygous deletion of the p16 locus [16], possibly facilitating subsequent transformation.
In this study, overexpression of Cyclin D2 promoted proliferation of hMSCs but did not lead to unlimited proliferation. Other factors are required for the unlimited proliferation of hMSCs. IGF2BP1 was aberrantly expressed in Ewing's sarcoma (Table 1), consistent with a previous report of an association between increased IGF2BP1 expression and tumor progression in patients with lung cancer [30]. Thus, we attempted to transduce the IGF2BP1 gene into hMSCs, but IGF2BP1 expression was up-regulated by only 2-fold and transduction efficiency was low (data not shown). The cause for this inefficiency is unclear. Because the growth kinetics of IGF2BP1-transformed E. coli is quite slow (data not shown), it is likely that overexpression of IGF2BP1 is deleterious for hMSCs.
We did not tested whether the other genes listed in Table 1 affect proliferation of hMSCs, because these genes were not thought to directly affect the proliferation. Furthermore, not all Ewing's sarcomas express EWS-FLI-1: indeed, EWS-FLI-1 mRNA was not detected in Hs 822.T and Hs 863.T (data not shown). Thus, we did not transduce the EWS-FLI-1 gene into hMSCs. However, it is possible that the cooperation of these proteins is important for the development of Ewing's sarcoma. Thus, it would be interesting to transduce these genes into hMSCs in addition to Cyclin D2.

Conclusion
Cyclin D2 promotes hMSC proliferation and is a candidate biomaker for hMSC transformation. Increased' means the genes up-or down-expression is predicted to increase the function of interphase. 'Decreased' means the up-or down-expression is predicted to decrease. 'Affected' means IPA could not predict whether the expression change increase or decrease Table 4. 'Interphase' genes differentially expressed by at least 2-fold between hMSCs/CyclinD2 and hMSCs/Empty.