Dear Editor, Patient-specific induced pluripotent stem (iPS) cells, generated from somatic cells of disease-affected individuals, hold a tremendous potential for studying disease mechanisms and for drug screening approaches using cell types previously not available (Yamanaka, 2008). Fragile-X (FX) syndrome belongs to the autism spectrum disorders, and is the most common cause of inherited mental retardation with the prevalence of 1/3600 (Crawford et al., 2001). It is nearly always caused by silencing of the FMR1 gene due to abnormal CGG repeat expansions in the 5′-UTR of the gene (Verkerk et al., 1991). Abnormal CGG repeat expansion of over 200 repeats leads to transcriptional silencing and CpG methylation of the gene 5′-UTR and the gene promoter (Sutcliffe et al., 1992). The silencing of the FMR1 gene and the heterochromatinization of the gene promoter were shown to be developmentally dependent (Eiges et al., 2007). In undifferentiated human FX embryonic stem cells (FX-ES cells) derived from affected blastocyst-stage embryos, FMR1 is expressed and gene silencing occurs only upon differentiation (Eiges et al., 2007). Recently, we have generated 11 FX-iPS cell lines from three different FX patients (Urbach et al., 2010). In FX-iPS cells, and contrary to FX-ES cells, the gene is transcriptionally silent both in the pluripotent and differentiated states (Urbach et al., 2010). The absence of FMR1 expression in FX-iPS cells is accompanied by DNA methylation and histone modifications indicative of heterochromatin in the gene promoter (Urbach et al., 2010). Although FX-iPS cells cannot recapitulate the developmentally dependent silencing of FMR1, they represent a unique genetic model system that harbors a completely silent FMR1 locus to study disease phenotypes, and for potential drug screening using the most suitable cells in the context of the disease and the affected individual. Here, we wished to evaluate the reactivation of FMR1 in FX-iPS cells and their differentiated neuronal derivatives by epigenetic modulating drugs. We set out to investigate the ability of several chromatin remodeling drugs to reactivate FMR1 gene expression. Histone deacetylase inhibitor trichostatin-A (TSA) had no significant effect on FMR1 reactivation in FX-iPS cells (Supplementary Figure S1). Next, we wished to evaluate the effect of the demethylating agent 5-azacytidine (5-azaC) on FX-iPS cells. Interestingly, 5-azaC was able to robustly reactivate FMR1 gene expression in FX-iPS cells (Figure 1A and Supplementary Figures S1 and S2). Multiple experiments performed on three FX-iPS cell clones at high concentrations have consistently shown that following 5-azaC administration, FMR1 gene expression is restored, with percent levels ranging between 15% and 45% of the WT levels (Figure 1A). A 10-μM 5-azaC treatment resulted in the highest expression levels of FMR1 (Figure 1A and Supplementary Figure S1). A combination of 5-azaC and TSA treatment culminated in a slight increase in FMR1 expression compared with 5-azaC treatment alone (Supplementary Figure S1). When wild type (WT)-ES cells were treated with 1-μM 5-azaC, FMR1 gene expression did not change (Supplementary Figure S3). The 5-azaC concentrations we used are in accordance with physiological levels found in plasma of patients suffering from solid tumors and hematologic malignancies that were treated with 5-azaC (Rudek et al., 2005). Since neurons are the cell type most afflicted in FX syndrome, we further examined the potential of 5-azaC to restore FMR1 expression in FX-neurons. To generate neurons from FX-iPS cells, we used an established protocol (Schuldiner et al., 2001) for neuronal induction from pluripotent cells. We thoroughly characterized the FX-neurons by various means (Supplementary Figures S4–S6). Following 5-azaC treatment, an increase in FMR1 transcript mRNA level could be detected in FX-neurons, albeit with less efficiency than in FX-iPS cells (Figure 1A). Of note, following 5-azaC treatment, some of the undifferentiated FX-iPS cells underwent cell death or differentiation, as observed by the presence of floating cells, or a change in the morphology of some of the undifferentiated colonies. However, many pluripotent colonies were still present after 7 days of treatment. This intriguing phenomenon was not observed with neuronal cells, suggesting a unique effect of 5-azaC on pluripotent stem cells, possibly by inhibiting DNMT3a and DNMT3b which are predominantly expressed in pluripotent stem cells and not in somatic cells. We next assessed whether FMR1 expression persists after withdrawal of 5-azaC treatment. FX-iPS and FX-neuronal cells were treated for 7 days with 5-azaC, and kept in culture for another 7 days without 5-azaC. As expected, the treatment resulted in reactivation of FMR1 expression, which persists at similar levels even after drug withdrawal, as judged by real-time PCR analysis (Figure 1B). Immunostaining for FMR1 protein (FMRP) showed that 5-azaC treated FX-iPS cultures express detectable levels of FMRP, indicating reactivation in the translational level as well (Figure 1C). In order to determine the cell type-reexpressing FMRP, we performed co-immunostaining for FMRP and the pluripotent markers NANOG or SOX2, and found double staining in pluripotent stem cells (Figure 1C). Co-immunostaining for FMRP and the neuronal-specific marker NCAM1 revealed double staining specifically in neuronal cells (Figure 1C) which are Ki67 negative (Supplementary Figure S7). Importantly, 5-azaC treated neurons clustered together with non-treated neurons as revealed by DNA microarray analysis (Figure 1D). Expression levels of typical neuronal genes were not altered following 5-azaC treatment (Supplementary Figure S6B). Lastly, we examined the levels of epigenetic modifications such as DNA methylation, histone acetylation and methylation in the FMR1 locus following 5-azaC treatment. DNA methylation analysis by pyrosequencing of 17 CpG sites in the FMR1 promoter showed a concentration-dependent extensive DNA demethylation effect upon treatment with 5-azaC in FX-iPS cells as well as FX-neurons (Figure 1E and Supplementary Figure S8). While in untreated FX-iPS cells the gene promoter is highly methylated (>85% methylation), treatment with 1 μM 5-azaC results in the reduction of up to 45%, and 10 μM 5-azaC treatment results in reduction of up to 60%, of the average methylation in the FMR1 promoter compared with non-treated FX-iPS cells. These levels correspond to 5% average methylation levels in WT-iPS and WT-neuronal cells (Figure 1E and Supplementary Figure S9). FX-neuronal cells showed a more moderate reduction in the methylation levels following 5-azaC treatment compared with FX-iPS cells, in accordance with the real-time PCR results (Figure 1E). We further investigated histone modifications indicative of transcriptionally active (H3 tail acetylation and H3 lysine K4 methylation) and repressed (H3 lysine K9 methylation) chromatin states associated with FMR1 silencing in somatic cells of FX individuals (Coffee et al., 1999). A 2-μM 5-azaC treatment culminated in relaxation of the FMR1 locus, manifested by near-to WT elevation of histone H3 acetylation and H3K4 methylation compared with non-treated FX-iPS cells; however, H3K9 methylation was not affected by 5-azaC treatment (Figure 1F). Expression and chromatin analysis in FX-iPS and FX-neuronal cells following 5-azaC treatment. (A) Real-time PCR analysis of FMR1 transcript levels in FX-iPS cells and FX-neurons before and after 7 days of 5-azaC treatment using two different concentrations of 5-azaC (1 or 10 μM, respectively). Values are mean ± SE relative to untreated FX-iPS cells (I), or FX-neurons (II) (n = 2–5 biological repeats for each of the three clones). *P < 0.05, **P < 0.001. (B) Real-time PCR analysis of FMR1 transcript levels in FX-iPS cells and FX-neurons after 7 days of 5-azaC treatment, followed by an additional 7 days of culture without treatment. Note that even after 7 days of 5-azaC withdrawal, FMR1 is still expressed in comparable levels to treated cells. Values are mean ± SE compared with non-treated FX-iPS cells (I), or FX-neurons (II) (n = 2–4 biological repeats). *P < 0.05, **P < 0.001. (C) Immunofluorescence analysis of FMRP in FX-iPS cells and FX-neurons before and after 5-azaC treatment. Note the double staining for NANOG (red)/FMRP (green) and SOX2 (red)/FMRP (green) in treated FX-iPS cells and NCAM1 (green)/FMRP (red) in treated FX-neurons, showing FMRP expression specifically in pluripotent and neuronal cells, respectively. Negative NANOG or SOX2 cells are either differentiated FX-iPS cells or mouse embryonic fibroblasts that support pluripotent stem cells growth. Nuclei were stained with Hoechst33258 (blue). (D) Hierarchical clustering of 5-azaC treated neurons. Shown is cluster analysis of fibroblasts, pluripotent stem cells, FX-neurons, FX-neurons treated with 2 and 10 μM of 5-azaC and differentiated ES cell progenies (teratomas and embryoid bodies) by global gene expression arrays. (E) Pyrosequencing analysis of DNA methylation of the FMR1 promoter in 17 CpG positions following 5-azaC treatment (upper panel). Negative and positive controls are untreated FX-iPS #55 and WT-iPS #28, respectively. Note that higher concentration of 5-azaC treatment resulted in more efficient demethylation of the FMR1 promoter. Pyrosequencing analysis of FX-neurons following 5-azaC treatment. Negative and positive controls are untreated FX-neurons #47 and WT-neurons #28, respectively (lower panel). (F) ChIP analysis of histone H3-tail acetylation, H3K4 and H3K9 methylation in the FMR1 promoter before and after 2 μM 5-azaC treatment. Non-treated FX-iPS cell lines as well as WT-iPS #28 and human ES cells served as negative and positive controls. Real-time PCR analysis was performed on bound and input DNA with primers specific for the FMR1 promoter. Adenine phosphoribosyl transferase (APRT) and Crystalline (CRYAA) served as positive and negative controls, respectively. Values were normalized to the appropriate positive control and shown with their respective standard errors. (G) A schematic representation of the reactivation of the FMR1 locus by 5-azaC. The effect of 5-azaC on DNA methylation and chromatin modifications nearly reach WT levels, while RNA and protein expression is only partially restored. This may result from the inability of 5-azaC to clear the repressive methylation mark on H3K9 in the FMR1 locus. Pluripotent stem cells represent an attractive source to model autism spectrum disorders such as Rett (Marchetto et al., 2010) or FX syndrome (Eiges et al., 2007; Urbach et al., 2010). Since the mouse model for FX syndrome does not recapitulate the silencing of the FMR1 gene due to triplet repeats (Brouwer et al., 2007), human FX-iPS cells are a unique cell source to model FX syndrome by their ability to provide the relevant affected cells. Previously it was shown that 5-azaC was able to reactivate FMR1 gene expression in immortalized FX-lymphoblastoid cell lines (Coffee et al., 1999). Here we report that its administration was able to restore the expression of FMR1 in pluripotent stem cells. Furthermore, we show that 5-azaC, an FDA-approved drug, was able to reactivate FMR1 expression in neurons. While the effect of 5-azaC on methylation is global, treated FX-neurons still clustered with mutated as well as WT neuronal cells, and not other cell types, indicating that global gene expression changes following treatment are probably limited. Since FMR1 is aberrantly hyper-methylated in FX-cells, it may be more sensitive to the demethylating agent than other gene promoters. We cannot demonstrate whether 5-azaC reactivates FMR1 expression directly in post-mitotic neurons, or that its expression in post-mitotic neurons reflects an earlier event occurring in neuronal progenitor cells. However, if the latter is indeed true, this still has important implications for FX patients, as administration of the drug in the early stages of development may slow disease progression by clearing the abnormal methylation from FX neuronal progenitor cells. The finding that FMR1 expression is still maintained after 5-azaC treatment withdrawal supports this conclusion and suggests that it may be sufficient to administer the drug for a short period of time and not continuously. However, FX patients who have a point mutation in the coding sequence of the FMR1 gene (De Rubeis and Bagni, 2011) are not expected to benefit from 5-azaC treatment. In addition, we were able to dissect the various chromatin modifications associated with FMR1 reactivation and expression upon 5-azaC treatment. Although 5-azaC dramatically affected the methylation levels of the gene promoter and H3 acetylation and H3K4 methylation, the levels of FMR1 gene expression were still lower than those observed in WT cells. This may result from the high levels of H3K9 methylation that persist after treatment (Figure 1G). This finding suggests for the existence of two different layers of epigenetic states governing the FMR1 locus. Clearing the repressive mark of H3K9 may eventually yield near-to WT levels of RNA and protein in treated FX cells. In summary, our findings demonstrate a successful combined utility of the iPS cell technology to generate the relevant cell type for disease modeling, as well as for drug screening approaches, and may pave the way to affect the progression of the disease in FX patients. [Supplementary material is available at Journal of Molecular Cell Biologyonline. We thank Dr Holger A. Russ (Tel-Aviv University) and Dr Daniel Ronen (The Hebrew University) for critically reading the manuscript and providing insightful comments, Tamar Golan-Lev for helping with the graphical design, Nadav Sharon for assisting with the statistical analysis and Dr Ittai Ben-Porath for the Ki67 antibody. O.B-N is a Clore Fellow. N.B. is the Herbert Cohn Chair in cancer research. This research was supported by the Legacy Heritage Biomedical Science Partnership Program of the Israel Science Foundation (No. 943/09) and the Centers of Excellence Legacy Heritage Biomedical Science Partnership (No. 1801/10).]
No takes yet. Share an insight, caveat, or question.
Bar‐Nur et al. (2012) studied this question.