Author contributions: T.E.: designed all experiments and edited the manuscript; M.H., Y.Y., Y.T., and T.K.: performed the experiments, analyzed data, and edited the manuscript; Y.H.: designed and performed part of the drug candidate experiments and edited the manuscript; N.F.: produced and provided Sendai virus vectors and edited the manuscript; H.H.: designed and performed part of the shRNA experiments and edited the manuscript; Y.N., H.F., and N.H.: provided the patient samples and data and edited the manuscript; T.E and Y.T.: wrote and prepared the manuscript and edited the manuscript.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS August 7, 2012.
Fibrodysplasia ossificans progressiva (FOP) is a rare congenital disorder characterized by progressive ossification of soft tissues. FOP is caused by mutations in activin receptor-like kinase 2 (ALK2) that cause its constitutive activation and result in dysregulation of BMP signaling. Here, we show that generation of induced pluripotent stem cells (iPSCs) from FOP-derived skin fibroblasts is repressed because of incomplete reprogramming and inhibition of iPSC maintenance. This repression was mostly overcome by specific suppression of ALK2 expression and treatment with an ALK2 inhibitor, indicating that the inhibition of iPSC generation and maintenance observed in FOP-derived skin fibroblasts results from constitutive activation of ALK2. Using this system, we identified an ALK2 inhibitor as a potential candidate for future drug development. This study highlights the potential of the inhibited production and maintenance of iPSCs seen in diseases as a useful phenotype not only for studying the molecular mechanisms underlying iPS reprogramming but also for identifying drug candidates for future therapies. STEM CELLS2012; 30:2437–2449
Fibrodysplasia ossificans progressiva (FOP) is a congenital disorder of progressive and widespread postnatal ossification of soft tissues and muscles [1–3]. Severe debilitation, reduced life expectancy due to joint fusion, and restrictive ventilatory impairment with thoracic involvement are major symptoms of this disease. Patients with FOP have gradual worsening of pulmonary function and eventually die by 40 due to respiratory failure if they do not receive the appropriate support. There is no effective therapy for preventing the ectopic ossification associated with FOP. Recent studies have revealed that this disorder is caused by mutations in Activin A receptor type I, the gene encoding the bone morphogenetic protein (BMP) type I receptor activin receptor-like kinase 2 (ALK2) [4–9]. The most common mutation is R206H, which is thought to alter ALK2 and confer constitutive activity to the mutant receptor. Mesenchymal cells derived from primary teeth of FOP patients showed elevated basal expression of RUNX2 and alkaline phosphatase (AP), which are involved in bone formation . These data suggest that the dysregulation of BMP signaling seen in FOP patients results in ectopic expression of osteogenesis-related genes and aberrant ossification. Several other mutations in ALK2, such as G356D, underlie phenotypic variations of FOP and these also alter ALK2 and confer constitutive activity to the mutant receptor . The weaker kinase activity of ALK2 (G356D) compared to that of ALK2 (R206H) suggests that clinical variation is due to differences in the bioactivity of ALK2 mutants .
Induced pluripotent stem cells (iPSCs) derived from patients with incurable diseases represent a powerful tool not only for biomedical research but also for investigating the effects of drugs on patient-derived cells [12–16]. These cells are derived from differentiated somatic cells and functionally resemble embryonic stem cells (ESCs) . This process, known as reprogramming, is triggered by the expression of four transcription factors, Oct3/4, Sox2, Klf4, and c-Myc, which are the same core factors underlying pluripotency in ESCs [17, 18]. This reprogramming process changes a cell's gene-expression profile from that of a somatic cell back to that of a pluripotent state . It is well-known that iPSCs derived from somatic cells harboring pathogenic gene mutations represent the cellular phenotype of the disease [19–21].
Investigations of the process for generating iPSCs are valuable for understanding the molecular mechanisms underlying cellular reprogramming. Recent knockout-mouse studies have identified several genetic mutations that modify the efficiency of iPSC generation. For example, iPSCs can be generated with higher efficiency from p53- and Ink4a/Arf-null fibroblasts than from normal fibroblasts [22, 23]. A mutation in p21, which is a molecule involved in downstream p53 signaling, partially mimics this phenotype, suggesting that activation of p53 and Ink4a/Arf signals can inhibit cellular reprogramming. These results raise the notion that some genetic mutations underlying human diseases also affect the reprogramming processes and eventually abolished iPSC generation. However, it is still unclear how pathogenic gene mutations affect the cellular reprogramming required for the generation and maintenance of human iPSCs.
Here, we studied disease-derived iPSCs to elucidate how pathogenic gene mutations affect cellular reprogramming. We showed that iPSC generation from FOP-derived skin fibroblasts is repressed. The few FOP-derived iPSCs that we managed to isolate could not be maintained because they differentiated spontaneously into mesodermal and endodermal lineages. We showed that repression of iPSC generation results from inefficient reprogramming of FOP-derived fibroblasts and inhibition of iPSC maintenance. This repression was mostly overcome by specific suppression of ALK2 expression and treatment with an ALK2 inhibitor. Using this system in combination with in silico chemical library screening, we identified an ALK2 inhibitor as a potential drug candidate for future therapeutic applications. The inefficient production of iPSCs is a useful disease phenotype not only for understanding the mechanism of reprogramming but also for identifying drug candidates for future therapies.
MATERIALS AND METHODS
Generation and Detection of Sendai Virus Vector
The Sendai virus (SeV) carrying Oct3/4, Sox2, Klf4, and c-Myc were generated as described previously . To detect SeV genome, nested RT-PCR was performed. The sequences of primers and amplification conditions are listed in supporting information Table S1.
Cell Culture and iPSC Generation with SeV Vector
Fibroblasts from FOP patients and healthy volunteers were generated from explants of skin biopsy following informed consent under protocols approved by the ethics committee assigning authors. Skin samples were minced and cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum. After the fibroblast appeared, it was expanded for iPSC induction.
iPSCs were generated from human skin-derived fibroblasts as described previously . Cells (5 × 105) of human fibroblast cells per well of six-well plate were seeded 1 day before infection and then were infected with SeV vectors at a multiplicity of infection of 3. Seven days after infection, fibroblasts infected were harvested by trypsin and replated at 5.4 × 104 cells per 60 mm dish on the mitomysin C (MMC)-treated mouse embryonic fibroblast (MEF) feeder cells. Next day, the medium was replaced in human iPSC medium. Thirty days after infection, colonies were picked up and recultured again in human iPSC medium.
Maintenance of Human iPSCs
Human iPSCs were maintained on MMC-treated MEF feeder cells in human iPS medium containing DMEM/F12 (SIGMA) supplemented with 20% KNOCKOUT serum replacement (Invitrogen), 2 mM L-glutamine, 1 × 10−4 M nonessential amino acids (SIGMA), 1 × 10−4 M 2-mercaptoethanol (SIGMA), 0.5% penicillin and streptomycin (Nacalai Tesque, Kyoto, Japan, http://www.nacalai.co.jp), and 5 ng/ml basic fibroblast growth factor (Wako, Osaka, Japan, http://www.wako-chem.co.jp/). In some experiment, ALK2 inhibitors such as LDN-193189 (STEMGNT, Cambridge, MA, http://www.stemgent.com/) and Dorsomorphin (DM; SIGMA) were added into the human iPS medium. DiPS used as a control iPSC line was kindly gifted by DNAVEC Corporation (Tsukuba, Japan, http://www.dnavec.co.jp/en/Index.html).
Chemical Library Screening
We extracted known inhibitors for homologous kinases including BMP receptor family from the CHEMBL database to determine the queries (https://www.ebi.ac.uk/chembldb/). The CHEMBL database, which contains 1,118,865 compounds and 4,668,202 activity data, is provided by European Bioinformatics Institute. We constructed a search engine to retrieve ALK family inhibitors and BMP inhibitors from the CHEMBL database. Using the search engine, 236 known kinase inhibitors were obtained. Finally, 153 compounds were selected from commercially available databases system and purchased (Namiki Shoji Co., Ltd., http://www.namiki-s.co.jp/english/; Kishida chemical Co., Ltd., http://www.kishida.co.jp/english/index.html). Unavailable seven compounds known as ALK2 inhibitor (Thomson Reuters IntegritySM) were prepared by ourselves. In total, 160 bioactive chemical compounds were evaluated by cell-based inhibition assay of BMP signaling. Other materials and methods are described in supporting information Supplemental Materials and Methods.
Generation of iPSCs from FOP-Derived Skin Fibroblasts
We attempted to generate iPSCs from skin fibroblasts from four patients with FOP and three healthy volunteers by the SeV method (Fig. 1A) [24–26]. Three patients had the common R206H mutation of ALK2 and the remaining patient had the G356D mutation (supporting information Fig. S1A) . The SeV method is suitable for establishing disease-specific iPSCs because it is highly efficient, does not involve integration, and the SeV is easy to remove. The frequency of iPSC colony formation from FOP-derived fibroblasts was significantly lower than that from the normal controls (Fig. 1B, 1C). Almost all colonies generated from FOP-derived fibroblasts exhibited atypical morphologies compared to controls (Fig. 1D, 1E). Selected colonies of FOP-derived fibroblasts did not expand after several passages, exhibited a flat morphology, and disappeared. It is noted that these morphological changes are very similar to those observed in the induction of iPSC differentiation.
Treatment with ALK2 Inhibitors Can Restore the Ability to Generate and Maintain FOP-Derived iPSCs
In FOP, the mutant ALK2 results in dysregulated BMP signaling via its constitutive activation [3, 5, 9]. We next examined the effects of ALK2 inhibitors on the generation and maintenance of iPSCs. Treatment with the ALK2 kinase inhibitor, LDN-193189 (LDN) , restores the colony formation capacity of iPSCs in dose-dependent manner (Fig. 2A). DM , another inhibitor for ALK2, also improves the efficiency of colony formation but not to the same extent as that by LDN. LDN enhances the generation of iPSCs from fibroblast F2 cells harboring the R206H mutation as well as from fibroblast F4 cells harboring the G356D, but with lower efficiency (Fig. 2A).
In addition to improving the efficiency of colony formation, treatment with ALK2 inhibitors allows iPSC colonies to be maintained continuously without morphological alteration. The iPS colonies formed in the presence of LDN and DM exhibited a typical morphology (Fig. 2B and supporting information Fig. S1B). Individual colonies were cultured and maintained in the presence of LDN. Absence of the SeV DNA fragment following amplification with nested primers indicated that SeV was completely removed from the iPSC lines (Fig. 2C) . RT-PCR and immunostaining analyses of these FOP-derived iPSC lines revealed that they expressed a set of markers typical of pluripotent cells (Fig. 2C, 2D) [17, 18]. To confirm the pluripotency of FOP-derived iPSC lines, we transplanted them into the subcutaneous tissues of the immunodeficient mice. Eight to twelve weeks after injection, FOP-derived iPSC lines tested formed teratomas that contained derivatives of all three germ layers (Fig. 2E and supporting information Fig. S2). Immunoblot analysis of phosphorylated Smad1/5/8, which are downstream molecules of ALK2 signaling, indicated that ALK2 kinase activity was higher in FOP-derived iPSCs than in controls (supporting information Fig. S3). We also demonstrated Smads dephosphorylation following treatment with ALK2 inhibitors thus confirming the ability of these inhibitors to suppress ALK2 activity (supporting information Fig. S3).
FOP-derived iPSCs Can Spontaneously Differentiate into Both Mesoderm and Endoderm Lineages Under Conditions for Maintaining iPSCs
The removal of ALK2 inhibitors from FOP-derived iPSC cultures caused the colonies to be disrupted and differentiation to be initiated even under conditions for maintaining iPSCs (Fig. 3A and supporting information Fig. S4). To define the differentiated cell type, the expression of a set of differentiation markers in control and FOP-derived iPSC lines was examined (Fig. 3B, 3C). We observed elevated expression of both mesodermal (MESP1, MESOGENIN, and BRACHYURY) and endodermal (SOX17 and FOXA2) markers in iPSC lines cultured in the absence of the ALK2 inhibitor [29, 30]. CDX2, a marker of trophectoderm , was also upregulated in FOP-derived iPSC lines cultured without LDN. In contrast, the expression levels of neuroectodermal markers, such as SOX1, NESTIN, and NEUROD1 , did not change, even in the absence of the inhibitor (Fig. 3B). The iPSC lines cultured in the absence of LDN expressed pluripotency markers such as OCT3/4  and NANOG [34, 35] at a lower level than those in the presence of LDN (Fig. 3B). Immunostaining analysis also confirmed the elevated expression of mesodermal and endodermal markers (Fig. 3C). The expression patterns of these markers indicated that FOP-derived iPSCs tend to spontaneously differentiate into both mesodermal and endodermal cells rather than into ectodermal cells, despite being culturing under iPSC maintaining conditions.
Characterization of FOP-Derived iPSCs
We used microarray analysis to further characterize the FOP-derived iPSC lines. The patterns of global gene expression of three FOP-derived iPSC lines treated with LDN were different to those of cells not treated with LDN (Fig. 4A). We confirmed that the expression levels of mesodermal and endodermal markers were higher in iPSCs not treated with LDN than in those treated with LDN, whereas expression levels of neuroectodermal markers such as NESTIN and SOX1  were similar (Fig. 4A). In contrast, LDN treatment had no effect on the expression levels of differentiation markers in control iPSC line (Fig. 4A). We also found that the expression levels of markers found to be elevated in FOP-derived iPSC lines not treated with LDN are enhanced by BMP-4 stimulation in control iPSC lines (supporting information Fig. S5). We performed principal component analysis (PCA)  and hierarchical clustering of all genes to determine overall differences in transcription levels between FOP-derived iPSC lines and control iPSC lines (Fig. 4B, 4C). We added the datasets for 201B7, normal iPSC line, and three ESC lines from Gene Expression Omnibus datasets, GSE29115, GSE22167, and GSE37258 into our analyses. The gene-expression profile of FOP-derived iPSC lines not treated with LDN was distinct from that of control iPSC lines and those treated with LDN. In contrast, the gene-expression profile of FOP-derived iPSC lines treated with LDN was grouped closely to control iPSC lines not treated with LDN as did normal iPSC lines (N3-1 and DiPS) treated with LDN.
As shown in Figure 4C, each dataset for 201B7 and three ESC lines are separated from that for normal iPSC lines (N3-1 and DiPS) and FOP-iPSCs. In PCA analysis (Fig. 4B), the datasets for iPSC (red group) are broadly varied in the y-axis (PC3) direction compared with fibroblast datasets (yellow group). Therefore, the separation seen in the cluster analysis may reflect the above-mentioned differences, and such differences are thought to depend on the methods of iPSC induction (between retrovirus and SeV) and/or those between ESCs and iPSCs.
Comparing each gene-expression profile of FOP-derived iPSC lines treated with LDN to those not treated with LDN revealed upregulated and downregulated genes with a fold change of >2. Overall, with these datasets combined, we identified 526 commonly upregulated and 131 commonly downregulated genes (Fig. 4D). In addition, gene ontology analysis revealed that the molecular signature related to development and differentiation was frequently detected in commonly upregulated genes (Fig. 4E). In contrast, no specific signature was found in downregulated genes with a p-value cutoff of .1. Taken together, these results suggest that constitutive activation of ALK2 can promote the differentiation of iPSCs.
Efficient iPSC Generation Requires ALK2 Inhibitor Treatment Within an Optimal Time Frame
To further explore the effect of constitutive activation of ALK2 on iPSC generation, we investigated the time frame of inhibition of iPSC generation by the ALK2 mutant. We treated cultures with the ALK2 inhibitor for various periods of time and then counted the number of AP-positive (AP+) colonies on day 30 after exposure (Fig. 5A). The highest number of AP+ colonies was detected when FOP-derived fibroblasts were treated with the inhibitor from day 8 to day 30 (Fig. 5B, 5C). Treatment from day 1 to day 30 gave a lower level of efficiency. Furthermore, colony-forming efficiency was not restored by treating iPSC-induction cultures from day 1 to day 7 only. To further examine the period from day 8 to day 30, we counted the number of AP+ colonies from day 8 to day 14, from day 15 to day 21, and from day 22 to day 30 (Fig. 5A, 5C). Unexpectedly, we counted few AP+ colonies in every period. Taken together, these results suggested that the formation of iPSC colonies from FOP fibroblasts requires the inhibition of mutant ALK2 activity from day 8 to day 30 of iPSC induction and not during the early phase from day 1 to day 7.
Constitutive Activation of ALK2 in FOP Fibroblasts Inhibits Cellular Reprogramming
To elucidate the mechanisms underlying the inhibition of iPSC generation from FOP fibroblasts, we investigated whether the ALK2 mutation in FOP affects cellular reprogramming. We noticed that Tra-1-60, a pluripotency marker expressed on the iPSC surface, appeared during the early phase (∼ day 7) of iPSC induction (Fig. 6A). Expression of Tra-1-60 was reduced on day 14 but then was markedly increased again up to day 21. The proportion of Tra-1-60+ cells was lower in cultures of FOP-derived fibroblasts undergoing iPSC induction than in cultures of control fibroblast (Fig. 6A). However, Tra-1-60 expression was restored on day 21 in cultures treated with LDN. Consistent with this result, BMP-4 treatment inhibited the expression of Tra-1-60 during iPSC generation (Fig. 6A). To confirm that suppression of Tra-1-60 expression is due to the ALK2 mutation, we generated a shRNA expression construct specific for the ALK2 mutant R206H (Fig. 6B). The reporter plasmids are constructed by inserting synthetic oligonucleotides of normal and mutant ALK2 allelic sequence into 3′UTR of each luciferase genes . The activity of luciferase carrying mutant ALK2 allelic sequence was specifically suppressed by shRNA expression (Fig. 6B). This shRNA effectively suppressed expression of the ALK2 mutant in FOP-derived fibroblasts and restored Tra-1-60 expression to control levels (Fig. 6C, 6D). Consistently, the generation of iPSC was partially improved by shRNA expression (supporting information Fig. S6A). Taken together, these results demonstrate that the suppression of Tra-1-60+ cell generation is due to the mutant ALK2 in FOP.
To further evaluate the reprogramming status of the fibroblasts, we purified Tra-1-60+ and Tra-1-60− cells from day 7 cultures by fluorescence-activated cell sorting and investigated the expression of other pluripotent and fibroblastic markers. Fibroblastic markers, such as PDGFRα and VIMENTIN, which are strongly expressed in fibroblasts, should be downregulated during iPSC generation because they are not expressed in iPSCs. The endogenous genes OCT3/4 and NANOG are expressed in day 7 Tra-1-60+ cells but not in day 7 Tra-1-60− cells in both control and FOP-derived cultures (Fig. 6E). The expression levels of these pluripotent markers are significantly lower in FOP-derived Tra-1-60+ cells than in controls. There was no expression of SOX2, REX1, KLF5, or DNMT3b in neither Tra-1-60+ nor Tra-1-60− cells (supporting information Fig. S6B). The expression of fibroblastic markers was downregulated in Tra-1-60+ cells. However, the expression levels of PDGFRα and β and VIMENTIN were significantly higher in FOP-derived Tra-1-60+ cells than in controls (Fig. 6F). These results suggest that the reprogramming status of FOP fibroblasts is incomplete, despite the appearance of pluripotency marker expression in these cells.
The inhibition of ALK2 by LDN during the later phase of iPSC induction can recover the colony-forming capacity of iPSCs. This finding suggests that newly generated iPSCs cannot be maintained due to the constitutive activation of mutant ALK2 in FOP-derived fibroblasts and that as a result iPS colonies cannot form. To prove this hypothesis, we investigated the effect of BMP signaling on the colony-forming capacity of iPSCs. We showed that BMP-4 and BMP-7 treatments reduced the area of AP+ cells in each iPSC colony (Fig. 6G). Morphological analysis revealed that this enforced BMP signaling disrupts the formation of iPSC colonies (Fig. 6G, 6H). Given the fact that BMP signals can induce the differentiation of iPSCs, these results suggest the existence of at least the two distinct mechanisms underlying the inhibition of iPSC generation; the first is incomplete reprogramming of pluripotent and fibroblastic genes, and the second is the forced differentiation of the cells during and after reprogramming.
Previous studies demonstrated that TGF-β/Activin signals affect iPSC generation . Smad2 and 3, the downstream molecules of TGF-β/Activin receptors, are phosphorylated by the activation of these signals. However, we could not detect any change in Smad2 and 3 phosphorylations between normal and FOP-derived fibroblasts, suggesting that ALK2 mutation observed in FOP does not affect TGF-β/Activin signals during iPSC generation (supporting information Fig. S6C–S6F).
Inhibited Production and Maintenance of FOP-derived iPSC is a Useful Phenotype for Identifying Drug Candidates for Future Therapies
Our observation that treatment of FOP-derived fibroblasts with an ALK2 inhibitor allows iPSCs to be generated and maintained with pluripotency prompted us to use this system to identify candidate ALK2 inhibitors. We performed in silico screening by the multidirectional similarity search system and collected 160 bioactive compounds (see Materials and Methods). Our screen of the chemical library resulted in the identification of new candidate ALK2 inhibitors. One of them, RK-0071142 suppressed Smad1/5/8 phosphorylation (Fig. 7A, 7B). We used a cell-based assay of BMP signaling to show that the half-maximal inhibitory concentration (IC50) of RK-0071142 was 1.44 μM for BMP-6 and 4.08 μM for BMP-4 (Fig. 7C and supporting information Fig. S7). Treatment of FOP-derived fibroblasts with RK-0071142 restored their capacity to generate iPSCs but with efficiency lower than that of LDN (Fig. 7D). The spontaneous differentiation of FOP-derived iPSCs was suppressed by RK-0071142 at a similar level to that of LDN (Fig. 7E). These results, together with LDN analyses, indicate that the generation and maintenance of FOP-derived iPSCs is a useful system for evaluating the bioactivity of new candidates as ALK2 inhibitors.
In this study, we demonstrated that constitutive activation of ALK2 inhibits the generation and maintenance of human iPSCs. We also show that two distinct mechanisms underlie the inhibition of iPSC generation, the incomplete reprogramming of pluripotent and fibroblastic genes, and the forced differentiation of the cells during and after reprogramming.
Mutations in ALK2 underlie phenotypic variations of FOP and confer constitutive activity to the mutant receptor [3–9]. ALK2 acts as a type I receptor of BMPs, which are members of the TGF-β family . TGF-βs and their family members are implicated in the generation and maintenance of iPSCs. Activation of TGF-β signaling blocks the cellular reprogramming required for the differentiation of a somatic cell into a pluripotent one and thus results in a reduction in the efficiency of iPSC generation . Consistently, the blocking of TGF-β signaling by inhibiting ALK5 kinase, which is a receptor of TGF-β, augments the formation of iPSC colonies . In contrast, a recent study using MEFs demonstrated that BMP signaling during the early-stage of iPSC induction (∼ day 8) can enhance the generation of iPSCs. This signal induces a set of miRNAs associated with the mesenchymal-to-epithelial transition (MET), which can accelerate the generation of iPSCs . BMP signaling occurs via a tetrameric complex of two out of three type II receptors, namely, BMP receptor II (BMPRII), activin receptor type II (ActRII), and ActRIIB; and two out of three BMP type I receptors, namely, ALK2, 3, and 6 . Suppression of BMPRII and ALK3 inhibits the generation of iPSCs, suggesting that enhancement of iPSC generation by BMP signaling is mediated by the receptor consisting of BMPRII and ALK3 . Although these findings indicate that BMP signaling together with the receptors BMPRII and ALK3 are implicated in the generation of iPSCs, the role of other receptors, such as ALK2 and ALK6, in cellular reprogramming remains unknown. We show here that during the early phase of iPSC induction, from day 1 to day 7, treatment with an ALK2 inhibitor can suppress the generation of iPSCs from normal fibroblasts. Thus, BMP signaling mediated by ALK2 is necessary for reprogramming during the early phase of iPSC generation. In contrast to the previous findings for BMPRII and ALK3, we show here that constitutive activation of ALK2 affects both the upregulation of pluripotent markers and the downregulation of fibroblastic markers during the early phase of iPSC generation resulting in incomplete reprogramming. Signaling by BMP type I receptors appears to regulate iPSC generation through different mechanism; ALK3 signaling stimulated by BMPs can enhance the iPSC generation by promoting MET, whereas constitutive activation of ALK2 can suppress iPSC generation by reprogramming the pluripotent and fibroblastic markers insufficiently. The downstream molecules mediated by ALK2 are known to be similar but not identical to those mediated by ALK3. The BMP type I receptors, ALK2, 3, and 6, act as downstream components of type II receptors and phosphorylate Smad proteins [42, 43]. Whereas ALK3 and ALK6 phosphorylates Smad1, 5, and 8, ALK2 only phosphorylates Smad1 and 5 under physiological conditions . In addition, ALK3 signaling has a functionally different effect from that of ALK2 and 6 on cellular apoptosis of hippocampal progenitors . These studies can support our results that the effect of ALK2 on iPSC generation is different from that of ALK3.
We demonstrated that constitutive activation of ALK2 repressed the formation of iPSC colonies due to the forced differentiation of the cells during and after reprogramming. Consistently, LDN treatment at later stages of iPSC generation (from day 8 to day 30) restored the colony-forming capacity of FOP-derived iPSCs. Therefore, the molecular basis of the inhibitory effect on colony formation is the same as that causing repression of iPSC generation from FOP fibroblasts. BMP signaling is known to be important for maintaining the pluripotency of mouse ESCs. Mouse ESCs can be continuously cultivated with BMP-4 in the presence of leukemia inhibiting factor under serum-free conditions . In contrast, in human ESCs and mouse epiblast stem cell studies, BMP-4 has been shown to induce trophoblastic lineage  as well as germ cell lineage differentiations . In human ESCs, BMP-4 together with FGF2 can switch the cell lineage outcome to mesendoderm . We demonstrated here that constitutive activation of ALK2 in the presence of FGF2 is forced to induce differentiation into both mesoderm and endoderm. This result suggests that BMP signaling mediated by ALK2 can synergy with FGF2 to direct iPSCs into both mesodermal and endodermal cells.
Several recent studies have reported that iPSCs established from patients can not only recapitulate some aspects of diseases but also can be used to better design and anticipate results from translational medicine [12–16]. Our study demonstrates that the aberrant molecular events occurring in diseases can impair the generation of iPSCs and that by restoring these events the efficiency of iPSC generation can also be restored. Thus, investigations into why iPSC production in diseases is inhibited can help unravel the underlying molecular and pathogenic events of these diseases.
This study also highlights the inefficient production of iPSCs as a useful disease phenotype not only for studying the molecular mechanisms underlying iPS reprogramming but also for identifying drug candidates for future therapies. In addition to the other ways to select drug candidates such as the measurement of kinase activity, the reprogramming process serves well as a validation tool for specific application. Although the system used here did not appear to be relevant to the symptoms of FOP, we showed that LDN, a known inhibitor of ALK2, could improve the efficiency of both the generation and the maintenance of iPSCs derived from FOP patients. In combination with in silico chemical library screening, we screened a chemical library to identify candidates with the ability to restore iPSC induction in FOP-derived fibroblasts. Using this approach, we identified a new ALK2 inhibitor candidate, RK-0071142, with potential for future therapeutic applications in the treatment of FOP. Since constitutive activation of ALK2 plays a pivotal role in the pathogenesis of FOP, screening for and evaluating new compounds as ALK2 inhibitors is expected to be the first step toward developing new drugs for the treatment of FOP. The patient-derived cellular model presented here has the potential to not only lead to the discovery of new compounds to treat FOP but also to be applied as a strategy to develop future therapies for other human diseases.
We would like to thank Masaki Takahashi for technical support, and Dr. Yuki Yanagihara and Dr. Kohei Miyazono for providing materials and helpful discussion. We also thank RIKEN Program for Drug Discovery and Medical Technology Platforms for providing the chemical compounds. This study was supported in part by grants from the Ministry of Health, Labor, and Welfare of Japan and Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.