Dinaciclib potently suppresses MCL-1 and selectively induces the cell death in human iPS cells without affecting the viability of cardiac tissue

Induced pluripotent stem (iPS) cells hold great potential for being a major source of cells for regenerative medicine. One major issue that hinders their advancement to clinic is the persistence of undifferentiated iPS cells in iPS-derived tissue. In this report, we show that the CDKs inhibitor, Dinaciclib, selectively eliminates iPS cells without affecting the viability of cardiac cells. We found that low nanomolar concentration of dinaciclib increased DNA damage and p53 protein levels in iPSCs. This was accompanied by negative regulation of the anti-apoptotic protein MCL-1. Gene knockdown experiments revealed that p53 downregulation only increased the threshold of dinaciclib induced apoptosis in iPS cells. Dinaciclib also inhibited the phosphorylation of Serine 2 of the C-terminal domain of RNA Polyemrase II through CDK9 inhibition. This resulted in the inhibition of transcription of MCL-1 and the pluripotency genes, NANOG and c-MYC. Even though dinaciclib caused a slight downregulation of MCL-1 in iPS-derived cardiac cells, the viability of the cells was not significantly affected, and beating iPS-derived cardiac cell sheet could still be fabricated. These findings suggest a difference in tolerance of MCL-1 downregulation between iPSCs and iPS-derived cardiac cells which could be exploited to eliminate remaining iPS cells in bioengineered cell sheet tissues.


Results
Low concentration dinaciclib treatment induces apoptosis in human iPS cells. Firstly, to ascertain the effect of inhibition of CDKs on the viability of human iPS cells, we treated with different concentrations of dinaciclib and found that 6 nM was sufficient to induce significant cell death within 6 hours of treatment, with more than 90% elimination of iPS cells achieved within 24 hrs (Fig. 1a). To test if the observed effect is due to induction of apoptosis, Annexin V-FITC binding was examined using flow cytometry analysis after 6 hours of dinaciclib treatment. We found that the percentage of Annexin V positive cells was significantly higher in the dinaciclib treated group (49.3% ± 10.7%, n = 3) compared to that in DMSO treated cells (10.3% ± 2.9%, n = 3) (Fig. 1b). Additionally, cleaved active Caspase-9 could be detected within 6 hrs of treatment (Fig. 1c).

Low concentration dinaciclib induced apoptosis in human iPS cells is p53 dependent. Since
CDK1 is involved in DNA repair, we aimed to investigate the effect of dinaciclib on the levels of DNA damage in human iPS cells 22,23 . To do so, whole cell lysates from cells treated with DMSO or varying dinaciclib concentrations for 6 hrs were examined by Western Blot for ɣ -H2AX signal. We observed that the levels of DNA damage following dinaciclib treatment have increased when compared to the DMSO treated group (Fig. 2a). Since the tumor suppressor and transcription factor p53 is vital for the DNA damage response pathway, we examined its levels in iPS cells following dinaciclib treatment. We found that dinaciclib caused a significant increase in p53 protein levels within 6 hrs of treatment (Fig. 2b). To test whether the observed apoptosis was p53 dependent or not, we utilized a p53 specific shRNA and we were able to generate a stable p53 knockdown iPS cell line with 85% downregulation (± 0.9%, n = 3) (Fig. 2c). We found that p53 knockdown inhibited dinaciclib induced apoptosis at 6 nM (Fig. 2d). This suggests that p53 may play a major role in the apoptosis induced by dinaciclib in human iPS cells.
It has been previously reported that chemical inhibition of CDK1 in murine ESC results in p53 mediated increase in Noxa mRNA levels and subsequent proteolytic degradation of the anti-apoptotic protein, MCL-1 17 . We have also observed a decrease in MCL-1 protein levels following dinaciclib treatment in human iPS cells (Fig. 2e). Additionally, we found that the decrease in MCL-1 caused by treatment with 6 nM dinaciclib was p53 dependent, as levels of MCL-1 protein were not decreased in the p53-knockdown cells following treatment (Fig. 2f). This finding suggests that, like in mouse ESCs, human iPS cells seem to be sensitive to decreased MCL-1 levels.

Dinaciclib induces apoptosis in human iPS cells independent of p53 through CDK9 inhibition.
We have observed that p53-knockdown cells still entered apoptosis when treated with slightly higher dinaciclib concentrations, suggesting the presence of other mechanisms through which dinaciclib may be inducing apoptosis in human iPS cells independently of p53 (Fig. 3a). Interestingly, we have found that treatment of p53-knockdown iPS cell line with higher concentrations of dinaciclib (20 and 50 nM) still caused a significant reduction in MCL-1 protein levels, while MCL-1 levels were maintained in the 6 nM cells (Fig. 3b). Considering the short half-life of MCL-1 and because dinaciclib is known to inhibit the kinase activity of CDK9 and RNA pol II phosphorylation, we hypothesized that the decreased MCL-1 levels caused by treatment with high dinaciclib concentrations might be due to a decrease in the transcription of MCL-1 as a result of a decrease in the number of phosphorylated RNA pol II molecules. As expected we found that treatment of iPS cells with dinaciclib for 6 hrs, caused a rapid dephosphorylation of the CTD serine 2 of RNA Pol II (Fig. 3c). To confirm that reduced MCL-1 levels following dinaciclib treatment was also due to hampered transcription and not only the activation of a proteolytic pathway or changes in translation rates, the levels of MCL-1 mRNA following dinaciclib treatment were measured by qRT-PCR. Human iPS cells were treated with increasing concentrations of dinaciclib for 6 hrs and then total mRNA was extracted and used for analysis. As expected, MCL-1 mRNA levels were significantly decreased in a dose dependent manner when iPS cells were treated with 8, 20 and 50 nM of dinaciclib (p < 0.01, n = 3) (Fig. 3d).
Interestingly, we observed that even though 6 hrs treatment of dinaciclib significantly reduced the phosphorylation of Ser2 of RNA Pol II, the treatment did not significantly change the levels of OCT4 mRNA in iPSCs (n = 3) (Fig. 3e). This may indicate that the C-terminal domain (CTD) phosphorylation at Ser2 of RNA Pol II may not be required for OCT4 transcription, as previously reported for other genes 24,25 . On the other hand, mRNA levels of NANOG were significantly decreased by all dinaciclib concentration treatments (p < 0.01 vs. DMSO) (Fig. 3e). Additionally, the mRNA levels of c-MYC in dinaciclib treated iPS cells showed a dose dependent downregulation (p < 0.05 6 nM vs. 20 nM and p < 0.01 6 nM vs. 50 nM). Taken together, these findings suggest that dinaciclib treatment has detrimental effects on human iPS cells and may be utilized to purge out contaminating undifferentiated cells in iPS-derived bioengineered tissues.

Dinaciclib selectively eliminates human iPS cells when co-cultured with normal human cardiac fibroblasts (NHCFs).
To test the efficacy of dinaciclib in the selective elimination of iPS cells, we co-cultured human iPS cells with normal human cardiac fibroblasts and then treated with different dinaciclib concentrations. We found that 6 hrs of treatment with all concentrations of dinaciclib resulted in significant reduction in the number of OCT4 positive cells (p < 0.01 vs control) (Fig. 4a,b), while not significantly changing the numbers of OCT4 negative cardiac fibroblasts (Fig. 4c). We could also observe a dose-dependent pattern in the selective elimination of iPS cells with the percentage of OCT4 positive cells being significantly lower in the 50 nM treated wells when  ) and then stained with Calcein AM to detect live cells (green) and Ethidium homodimer 1 (EthD-1) to detect dead cells (red). Scale bar = 100 μ m (b) iPS cells were treated with DMSO or 6 nM dinaciclib for 6 hrs and then incubated briefly with Annexin V-FITC and analyzed for Annexin V binding using flow cytometry (n = 3). (c) iPS cells were treated with DMSO or 6 nM dinaciclib for 6 hrs. Cells were then harvested and whole-cell lysates were used for Western Blot analysis and detection of the active cleaved p10 subunit of Caspase-9.
Scientific RepoRTs | 7:45577 | DOI: 10.1038/srep45577 compared to the ones treated with 6 nM (p < 0.05) (Fig. 4d). These findings suggest that treatment with 50 nM dinaciclib may be a good strategy to purge out remaining iPS cells in iPS-derived cardiac tissue.
In addition, we have treated non-purified iPS-derived cardiac cells with DMSO or 50 nM dinaciclib for 24 hrs and compared LIN28 expression using qRT-PCR as an indicator for residual iPS cells. LIN28 mRNA has recently been shown to be the most sensitive marker for detecting residual iPS cells when co cultured with human primary cardiomyocytes 26 . We have found that dinaciclib treatment decreased LIN28 to only 46% to what is seen in the control (Supplementary Figure 1). Taken together, these findings suggest that dinaciclib might be suitable to eliminate residual iPS cells in iPS-derived cardiac cells.
Dinaciclib treatment is not critical for the survival of iPS-derived cardiac cells. The efficacy of dinaciclib treatment as a method for residual iPS elimination depends on its influence on iPS-derived bioengineered cardiac tissue. Therefore, after observing that dinaciclib selectively induced apoptosis in human iPS cells, we aimed to investigate its effect on the viability of iPS-derived cardiomyocytes. Human iPSCs expressing puromycin resistance gene under the control of the mouse α -myosin heavy chain (α -MHC) promoter were differentiated towards the cardiac lineage. Since high cell density is required for the fabrication of cardiac cell-sheet, residual iPS cells survival may be enhanced and therefore treatment with higher concentrations of dinaciclib may be needed. Therefore, we treated purified iPS-derived cardiomyocytes with 50 nM dinaciclib for 24 hrs and did not observe considerable cell death (Fig. 5a). Additionally, DMSO and dinaciclib treated cells continued to exhibit spontaneous beating following treatment (Supplementary Video 1 and 2).
Next, we aimed to examine whether or not treatment with dinaciclib might negatively affect the process of cell-sheet fabrication which is necessary for the use of iPS-derived cardiac cells in regenerative medicine. Following differentiation, human iPS cell-derived cardiac cells including cardiomyocytes and fibroblasts were cultured on temperature-responsive culture dishes and transiently treated with DMSO or 50 nM dinaciclib for 24 hrs. Cell sheets were successfully obtained from both treatments after lowering culture temperature (Fig. 5b), which suggests that dinaciclib might not have detrimental effects on any of the necessary elements required for the fabrication of bioengineered tissue, including cell-cell junctions, basement membrane proteins or extracellular matrix. Additionally, when non-purified iPS-derived cardiac cell sheets treated with DMSO or 50 nM dinaciclib for 24 hrs were subcutaneously transplanted in immunodeficient rats, beating could still be seen 7 weeks after transplantation (Supplementary videos 3 and 4). Moreover, when non-purified cardiac cells were treated with 50 nM dinaciclib, the protein expression of cardiac Troponin T (cTnT) was not affected (Fig. 5c). Furthermore, when purified cardiomyocytes were treated with DMSO or 50 nM dinaciclib, the percentage of cTnT positive cell were not affected even after 1 week of treatment (Supplementary Figure 2). Even though, the number of cTnT positive cells was not different 24 hrs following treatment with dinaciclib compared to DMSO, 1 week following treatment, the number of cTnT positive cells was lower in the dinaciclib group compared to DMSO (p < 0.01) (Supplementary Figure 2). Therefore, we speculated that dinaciclib might have some inhibitory effects on the proliferation of cardiomyocytes. As expected, the number and percentage of Ki67 positive cardiomyocytes were significantly (p < 0.05) lower in the 50 nM dinaciclib treated cells (Supplementary Figure 3).
As shown previously, we observed that treatment of iPS cells with dinaciclib increased the levels of ɣ -H2AX. Since causing a permanent increase in DNA damage in iPS-derived cardiac tissue might not be a desirable effect when the tissue is considered for the use in regenerative medicine, we examined the levels of ɣ -H2AX in iPS-derived cardiac tissues that were treated with varying dinaciclib concentrations for 24 hrs, two days after treatment stoppage. We found that levels of ɣ -H2AX in the tissues were only slightly upregulated 48 hrs following the stoppage of treatment (Fig. 6a). We have also observed an increase in the protein levels of p53 after 24 hrs of treatment, which may be due to an initial increase in DNA damage (Fig. 6b). The protein levels of MCL-1 were also slightly decreased 24 hrs following treatment (Fig. 6c). To ascertain whether the decrease of MCL-1 protein is due to proteolytic degradation as a result of p53 stabilization and not due to blockage of transcription, we examined the mRNA levels of MCL-1 by qRT-PCR. We found that 24 hrs treatment of iPS-derived cardiac tissue with dinaciclib did not cause a significant decrease in the transcription of MCL-1 (Fig. 6d). Although there was a significant dephosphorylation of the CTD serine 2 of RNA Pol II (Fig. 6e), some amount of phosphorylated CTD serine 2 of RNA Pol II might be enough to maintain the transcription of MCL-1 after the treatment with dinaciclib.

Discussion
In this study, we demonstrate that dinaciclib treatment induces programmed cell death in human iPS cells, but not in their derived cardiac cells. This could allow for the fabrication of a iPS-derived cardiac tissue with a lower risk of contamination of undifferentiated cells. Considering the highly proliferative nature of pluripotent stem cells, the exploitation of the differences in cell cycle regulation between iPS cells and their cardiac derivatives might be a good strategy to purge out remaining iPS cells, especially in iPS-derived cardiac tissue in which cells are not actively dividing.
Dinaciclib, inhibits 50% of the kinase activity of the CDKs 2, 5, 1 and 9 with the concentrations 1, 1, 3 and 4 nM 16 . We found that treatment with 2 nM dinaciclib, which is high enough to inhibit CDK2 and 5 but not CDK 1 or 9, did not induce significant apoptosis in iPS cells (data not shown). This suggests that the inhibition of the CDK2 and 5 might not be detrimental for human iPSCs. This is in line with previous findings that showed that CDK1 can fully compensate for the absence of CDK2 27,28 . When we treated human iPS cells with 6 nM dinaciclib, apoptosis was evident in culture within 6 hrs of treatment. More than 90% of the cells became dead within 24 hrs of treatment (Fig. 1a). Previously, it has been shown that specific knockdown of CDK1 in human iPS cells using siRNA did not result in cell death, but caused an increase in differentiation 18 . However, in both cases, using siRNA or dinaciclib to inhibit CDK1 resulted induced an increase in double strand breaks (Fig. 2a) 18 . These findings suggest that dinaciclib treatment at 6 nM is simultaneously inhibiting other CDKs and that the collateral inhibition is necessary for apoptosis to occur in human iPS cells.
We found that dinaciclib treatment caused a significant increase in DNA damage and in p53 protein levels in as early as 6 hrs of treatment (Fig. 2a,b). This was also accompanied by a decrease in the protein levels of the anti-apoptotic protein, MCL-1 (Fig. 2e). Previous reports have shown that when murine and human ESCs were treated with the CDK1 inhibitor, Purvalanol A, a similar response was observed 17 . Furthermore, it was also reported that specific MCL-1 knockdown resulted in cell death in mouse ESCs 17 . Therefore, to confirm whether the observed decrease in MCL-1 following dinaciclib treatment correlated with the occurrence of apoptosis in human iPS cells, and to ascertain if this reduction is caused by proteolytic degradation of the protein mediated by p53 downstream targets (e.g NOXA), we knocked down p53 and treated with dinaciclib. Interestingly, we found that p53 knockdown prevented apoptosis at 6 nM and prevented the observed reduction in MCL-1 protein levels (Fig. 2d,f). These findings suggest that p53 plays a major role in the induced apoptosis following dinaciclib 6 nM treatment in human iPS cells, and that it does so through proteolytic degradation of MCL-1.
Interestingly, we have found that concentrations slightly higher than 6 nM could not be tolerated by p53-knockdown iPS cells (Fig. 3a). Again, the observed apoptosis induced by such concentrations correlated with a decrease in MCL-1 protein levels. This suggested that other mechanisms may come into play and contribute to the reduction of MCL-1 independently of p53. Dinaciclib is also known to inhibit the kinase activity of CDK9 which, within its complex, P-TEFb, functions by phosphorylating Ser2 in the CTD of RNA Pol II, which then promotes the elongation of nascent short transcripts 16,20 . We found that only when treated with concentrations higher than 6 nM, a significant reduction in the transcription of MCL-1 was observed (Fig. 3d). Furthermore, on target inhibition of CDK9 in human iPS cells was confirmed when we examined the levels of phosphorylated RNA Pol II by Western Blot analysis (Fig. 3c).
We have also found that short treatment with dinaciclib induced a significant reduction in the pluripotency genes, NANOG and c-MYC, while not affecting the transcription of OCT4 (Fig. 3e). This is in line with a previous study that demonstrated the binding of CDK9 to the regulatory sites of NANOG in ESCs 21 . However, it seems that the phosphorylation of the Ser2 of the CTD of RNA Pol II may not be required for OCT4 transcription. Previous reports have showed that some genes may escape the inhibition of transcription when using pTEFb inhibitors 25 . Therefore, dinaciclib treatment of human iPS cells may concomitantly induce differentiation alongside apoptosis, which are both desired effects when we intend to eliminate residual undifferentiated iPS cells in iPS-derived cardiac tissue. Since no change in OCT4 mRNA expression in iPS cells was observed after treatment with dinaciclib (Fig. 3e), OCT4 transcription may be independent of the Ser2 phosphorylation on the CTD of RNA Pol II. This makes OCT4 a good candidate to detect remaining iPS cells in the co-culture experiment performed in Fig. 4. Therefore, the reduction in OCT4 positive cells in Fig. 4 is highly likely to be a result of selective elimination of iPS cells. Since the data in Fig. 3e represent the relative expression of OCT4 against GAPDH, the decrease of iPS cells in feeder-less condition (iPS cell monoculture) might not affect the relative expression of OCT4 against GAPDH.
We additionally measured LIN28 mRNA levels using qRT-PCR in non-purified iPS-derived cardiac cells treated with DMSO or 50 nM dinaciclib. LIN28 mRNA has been shown to be the most sensitive indicator for detecting residual iPS cells when co cultured with human primary cardiomyocytes. This is based on the finding that LIN28 mRNA could not be detected in primary cardiomyocytes and is highly expressed in human iPS cells when measured by qRT-PCR 26 . Since dinaciclib has been shown to cause a reduction of the phosphorylation of serine 2 of the CTD of RNA Pol II, we speculated that it might reduce the transcription of LIN28 and falsely exaggerate the dinaciclib mediated elimination of residual iPS cells. Therefore, the medium was replaced after 24 hrs of treatment and cells were harvested for qRT-PCR analysis the day after. We found that LIN28 expression was decreased to 46% of the levels found in the control (DMSO) (Supplementary Figure 1). Therefore, we believe that dinaciclib treatment might eliminate remaining undifferentiated iPS cells in iPS cell-derived cardiac cells. However, the degree of dinaciclib mediated elimination of residual iPS cells might be difficult to be determined accurately, as it is unclear whether the remaining LIN28 expression following dinaciclib treatment should be attributed to differences in the degree of maturation between iPS-derived and primary cardiomyocytes or due to persisting undiffereniated cells.
Even though low concentration and short duration dinaciclib treatments were enough to induce DNA damage and apoptosis in iPS cells, purified iPS-derived cardiomyocytes that were 97% positive for cTnT (Supplementary Figure 2) have tolerated dinaciclib with no considerable increase in apoptosis when treated for 24 hrs at 50 nM concentration (Fig. 5a). In addition, 50 nM dinaciclib did not have an effect on the numbers of OCT4 negative NHCF when co-cultured with iPS cells (Fig. 4c). This indicates that cardiac fibroblasts may also have higher tolerance to dinaciclib when compared to iPS cells. Furthermore, dinaciclib treated tissue could be successfully fabricated when iPS-derived cardiac cells (cardiomyocytes and fibroblasts) were cultured on a temperature-responsive dish. Furthermore, the treated cell sheets still exhibited beating 7 weeks after transplantation in nude rats (Supplementary videos 3 and 4). It is worth noting that after 7 weeks, the size of the tissue that was treated with 50 nM dinaciclib prior to transplantation was evidently smaller than the control (DMSO treated). This indicates that the iPS-derived cardiomyocytes utilized in our study possess some degree of proliferative potential that may be reduced with dinaciclib treatment. This is in line with our data presented in (supplementary Figure 3), in which the percentage of Ki67 positive cTnT positive cells were significantly (p < 0.05) lower in the dinaciclib treated cells compared to control. Therefore, there seem to be some inhibitory effects of dinaciclib on iPS-derived cardiomyocytes proliferation.
We have noticed that treatment of iPS-derived cardiac cells with high concentration dinaciclib (50 nM) has induced an increase in p53 protein levels and caused a minute increase in DNA damage. Previous reports have shown that undifferentiated cells may be more prone to undergo apoptosis as a response to DNA damage than differentiated cells 29 . This could be one reason why iPS-derived cardiac cells had higher tolerance to dinaciclib. In non-purified cardiac cells, we observed a modest increase in apoptosis at 24 hrs post the initiation of treatment. This may be due to the presence of undifferentiated or partially differentiated cells that are sensitive to dinaciclib and not due to a loss in the cardiomyocyte constituent of the tissue. This is demonstrated by the persistence of beating of the tissue following treatment (Supplementary video 2) Treatment of iPS-derived cardiac tissue with dinaciclib has also caused a drastic decrease in the phosphorylation of RNA Pol II (Fig. 6e). However, this decrease in the levels of phosphorylated RNA Pol II did not cause a significant decrease in MCL-1 transcription (Fig. 6d). This may be due to a higher MCL-1 activity in iPS cells compared to iPS-derived cardiac cells. It has been previously reported that MCL-1 protein levels decrease significantly following the differentiation of murine and human ESCs 17 . Additionally, MYC is known to direct the transcription of MCL-1 and previous reports have shown that tumors that have high MYC activity or are MYC-driven, are specifically sensitive to MCL-1 downregulation compared to MYC independent cancers or normal tissue [30][31][32] . Therefore, following differentiation of iPS cells to cardiac cells, a reduction in MYC expression is expected, which could possibly lead to a reduction in MCL-1 transcription. Furthermore, CDK9 inhibition by dinaciclib treatment has been shown to induce apoptotic responses in aggressive MYC-driven B-cell lymphoma through transcriptional blockage of MCL-1 33 . Taken together, human iPS cells seem to be dependent on the levels of MCL-1, and its indirect inhibition by dinaciclib induces apoptosis. It is also worth mentioning that Western Blot analysis of iPS-derived cardiac tissue did reveal a small decrease in MCL-1 protein level (Fig. 5c). This might be mediated by the increased p53 protein levels in iPS-derived cardiac cells (Fig. 5b) which could be causing proteolytic degradation of MCL-1.
Many groups including us reported the strategies to eliminate remaining iPS cells based on the differences between undifferentiated iPS cells and their derivatives. Based on the evidence that showed high activity of survivin in undifferentiated PSCs compared to their derivatives, Mitsui et al. have generated conditionally replicating adenoviruses (CRAs) in which the survivin promoter regulated the adenoviral early region 1 A (E1A) and hence the replication of the virus. Upon treatment with the adenovirus undifferentiated human PSCs were selectively eliminated with no cell death in their differentiated counterparts 34 . Tateno and colleagues have generated a recombinant lectin-toxin fusion protein (rBC2LCN-PE23), made by the fusion of the catalytic domain of Pseudomonas aeruginosa exotoxin A, with the lectin rBC2LCN which was previously shown to selectively bind human iPS cells 35 . The negative selection using cell surface markers that specifically express in iPS cells such as CD30 also might be the promising strategy to eliminate iPS cells 36 . The difference in tolerance against methionine free culture condition and TRPV-1 activation between iPS cells and iPS cell-derive cardiac cells has been reported to be capable of eliminating iPS cells [37][38][39] . It remains unclear whether a single strategy from the ones mentioned above is sufficient to prevent tumor formation in the clinical settings. Combining the use of dinaciclib with other methods might prove more effective in iPS cell elimination. However, the most efficient combination of techniques is yet to be determined.
There are some limitations in this study. We have mainly utilized 30 day-old iPS-derived cardiomyocytes (30 days post-induction of differentiation) to elucidate the effect of dinaciclib on iPS-derived cardiomyocytes. It is possible that the effect of dinaciclib treatment may vary depending on the maturity of iPS-derived cardiomyocytes with immature cells being more sensitive to dinaciclib than older more mature cardiomyocytes. As mentioned above, cardiomyocytes used in this study exhibited some degree of proliferative potential and have increased in number to around 2.5 in one week (Supplementary Figure 2). However, since 50 nM dinaciclib did not induce the cell death of cardiomyocytes (Fig. 5), the toxicity of dinaciclib might not be evident in immature cardiomyocytes. Additionally, when non-purified iPS-derived cardiac cells (23-25 days old), were treated with 50 nM dinaciclib, the beating of the cells was not affected in vitro (supplementary video 2) or in vivo (supplementary video 4) and the expression of cTnT was not affected (Fig. 5c), indicating that dinaciclib might not have significant cytotoxicity on immature cardiomyocytes. Although dinaciclib decreased the number of iPS cells when co-cultured with human cardiac fibroblasts, it remains unclear to what extent is dinaciclib treatment capable of eliminating the residual iPS cells due to the lack of spike-in experiments.
Results from this study demonstrate that treatment of iPS-derived cardiac tissue with dinaciclib might be an effective method for eliminating residual iPS cells. Dinaciclib treatment of iPS-derived cardiac cells did not affect the process of cell-sheet fabrication. Additionally, beating in the tissue persisted after treatment and the expression of the cardiac marker cTnT was maintained. Dinaciclib is effective in inducing apoptosis in human iPS cells as it targets both CDK1 and CDK9. The inhibition of CDK1 and CDK9 caused the downregulation of the anti-apoptotic protein MCL-1, by proteolytic degradation and by blockage of transcription respectively. Considering the highly proliferative nature of iPS cells compared to their cardiac derivatives, this study demonstrates the feasibility of CDKs inhibition as a method to produce safer cardiac tissue for regenerative medicine.
hiPS cells (201B7) expressing α -MHC promoter and rex-1 promoter-driven drug-resistance genes were established as previously described 39 . 253G1 iPS cells were maintained on iMatrix511 (Nippi, Tokyo, Japan) in StemFit AK03N (Ajinomoto, Tokyo, Japan). Cells were passaged as single cells every 7-8 days using TrypLE Select (Life Technologies, Carlsbad, CA, USA) as described previously 40 . 201B7 iPS cells were maintained on inactivated mouse embryonic fibroblasts (MEFs, ReproCELL, Yokohama, Japan) and were cultured as described previously. Sample images were obtained by an inverted microscope (Nikon, Tokyo, Japan) using the NIS-Elements software (Nikon).
Cardiac differentiation in the bioreactor and cardiac cell sheet preparation. The cardiac differentiation protocol in the bioreactor system (ABLE Co., Tokyo, Japan) has been described previously 39 . In short, we induced the cardiac differentiation of human iPS cells for 16 days in 3D suspension culture system and then dissociated single cells with trypsin were cultured for 14 days in cell culture plates. In the process of cultivation on cell culture plates, we treated cells with puromycin (Sigma Aldrich) (1.5 μ g/ml) twice to purify cardiomyocytes. Therefore, we used 30 day-old cardiomyocytes for experiments. For cell sheet fabrication, when puromycin purification was not performed, the cardiomyocytes were around 23-25 days old. To make the cell sheet, the surface of temperature-responsive dishes (UpCell; CellSeed, Tokyo, Japan) was coated with FBS for 3 hrs prior to seeding. Single cells were plated onto 12 or 24-well plates at 5 × 10 6 or 1.5 × 10 6 cells per well respectively in DMEM (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% FBS at 37 °C in humid air with 5% CO2. For cell sheet dissociation, the plates were incubated at 20 °C in humid air with 5% CO2.

Cell sheet transplantation. All animal experiments were performed according to the Guidelines of Tokyo
Women's Medical University on Animal Use, and consistent with the Guide for the Care and Use of Laboratory Animals prepared by the Institute of Laboratory Animal Resources (ILAR). All the experimental protocols were approved by the Institutional Animal Care and Use Committee of Tokyo Women's Medical University. Two monolayered non-purified iPS cell-derived cardiac cell sheets that were treated with DMSO or 50 nM dinaciclib for 24 hrs were collected by lowering culture temperature and one cardiac cell sheets from each condition was transplanted onto the subcutaneous tissue of a male fischer 344 athymic nude rat (aged around 6 weeks) (Charles River Japan, Tokyo, Japan) as described previously 39 . Seven weeks after transplantation, rats were anaesthetized by 2% isoflurane inhalation and macroscopic images were recorded using surgical microscope system (Leica M651 Surgical microscope system, Germany).
Lentivirus generation and p53 knockdown in 253G1 feeder-free iPS cells. p53 specific sequence and control shRNAs were obtained from Addgene (Cambridge, MA, USA). Lentiviral particles were produced using the second-generation packaging system obtained from Addgene as reported previously 41 . In short, 10 μ g of p53 or control shRNA plasmid, 7.5 μ gs of packaging plasmid (psPAX2 plasmid) and 2.5 μ g of envelope plasmid (pMD2.G plasmid) were transfected in HEK293t cells using FUGENE 6 (Promega, Fitchburg, WI, USA) transfection reagent according to manufacturer's instructions. Medium was replaced 24 hrs following transfection. Two days later, lentivirus containing medium was collected and concentrated using Lenti-x concentrator (Takara Bio, Mountain View, CA, USA). Lentivirus particles were re-suspended in OPTI-MEM (Life Technologies), aliquoted and stored in − 80 °C for later use. 253G1 human iPS cells were transduced with lentiviral particles using 6 μ g/ ml polybrene (San Cruz Biotech). When confluent, transduced cells were passaged and expanded. Successfully transduced cells were selected using 2 μ g/ml puromycin and were used for experiments.
RNA extraction and quantitative RT-PCR. Total RNA was extracted using RNeasy kit (QIAGEN, Hilden, Germany). RNA concentration and purity were determined using NanoDrop 2000. Equal amounts of RNA were then used from all samples to construct complementary DNA (cDNA) strand using the High-Capacity cDNA Reverse Transcription Kit (Life Technologies) according to manufacturer instructions. Quantitative PCR was performed with a 7300 Real Time PCR System (Applied Biosystems (ABI), Foster City, CA, USA). Relative mRNA expression levels were calculated using the Δ CT quantification method and endogenous GAPDH expression was used for normalization. All primers were obtained from ABI, MCL-1: Hs01050896_m1, POU5F1: Hs00999632_g1, NANOG: Hs02387400_g1 and c-MYC: Hs00153408_m1. LIN28: Hs00702808_s1.
Protein extraction and Western Blot. TrypLE Select (0.5x) (Life Technologies) and 0.05% trypsin were used to harvest iPS cells and iPS-derived cardiomyocytes respectively. Pelleted cells were treated with 5% protease inhibitor (Cell Signaling) in RIPA buffer (Life Technologies) vortexing occasionally and incubating on ice for 15 mins. Protein concentration was then determined using BCA Protein Assay Kit (Life Technologies). For shearing DNA in cells, protein was sonicated for 5 minutes. Protein extracts were centrifuged at 12,000 rpm at 4 °C for 5 minutes and supernatant was collected for Western Blot analysis. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and western blotting analysis were carried out as previously described 41 . The membrane was photographed by the chemiluminescence imager (LAS 4000; GE Healthcare).
Immunocytochemistry. Cells were fixed in 4% paraformaldehyde (PFA) for 12 mins at room temperature and then washed 3 times with PBS and then staining was performed as previously described 35 . Nuclei were stained using Hoechst 33258 (Sigma-Aldrich). Samples were imaged (49 images) by ImageXpress (Molecular Device, Sunnyvale, CA, USA) with MetaXpress and AcuityXpress software (Molecular Device).
Live/Dead staining. For iPS cells, cells were treated with DMSO or 6 nM dinaciclib and Live/Dead staining was performed at 7 and 24 hrs post treatment using (LIVE/DEAD Viability/Cytotoxicity Kit, Life Technologies) by following manufacturer's instructions. For purified iPS-derived cardiac tissue, cells were treated with DMSO or 50 nM dinaciclib for 24 hrs. The next day treatments were removed and the cells were stained 24 hrs later. Sample images were obtained by an inverted microscope (Nikon) using the NIS-Elements software (Nikon).
Annexin-V and flow cytometry analysis. Semi-confluent feeder free 253G1 hiPS cells were treated with DMSO or 6 nM dinaciclib for 6 hrs and then harvested and incubated with Annexin V using the Annexin V-FITC Apoptosis Detection Kit obtained from Abcam according to the manufacturer's protocol. Cells were then analyzed by flow cytometry using Gallios (Beckman Coulter, Brea, CA, USA).
Statistical analysis. Data are presented as mean ± standard deviation. Statistical analyses were performed with the Student's t-test for comparison between two groups. Multiple group comparisons were performed by one-way analysis of variance followed by Tukey-Kramer procedure for comparison of means. A value of p < 0.05 was considered statistically significant.