The Keio Journal of Medicine
Online ISSN : 1880-1293
Print ISSN : 0022-9717
ISSN-L : 0022-9717
ORIGINAL ARTICLE
Differential X Chromosome Inactivation Patterns during the Propagation of Human Induced Pluripotent Stem Cells
Tomoko Andoh-nodaWado AkamatsuKunio MiyakeTetsuro KobayashiManabu OhyamaHiroshi KurosawaTakeo KubotaHideyuki Okano
Author information
JOURNAL FREE ACCESS FULL-TEXT HTML

2016 Volume 66 Issue 1 Pages 1-8

Details
Abstract

Human induced pluripotent stem cells (hiPSCs) represent a potentially useful tool for studying the molecular mechanisms of disease thanks to their ability to generate patient-specific hiPSC clones. However, previous studies have reported that DNA methylation profiles, including those for imprinted genes, may change during passaging of hiPSCs. This is particularly problematic for hiPSC models of X-linked disease, because unstable X chromosome inactivation status may affect the detection of phenotypes. In the present study, we examined the epigenetic status of hiPSCs derived from patients with Rett syndrome, an X-linked disease, during long-term culture. To analyze X chromosome inactivation, we used a methylation-specific polymerase chain reaction (MSP) to assay the human androgen receptor locus (HUMARA). We found that single cell-derived hiPSC clones exhibit various states of X chromosome inactivation immediately after clonal isolation, even when established simultaneously from a single donor. X chromosome inactivation states remain variable in hiPSC clones at early passages, and this variability may affect cellular phenotypes characteristic of X-linked diseases. Careful evaluation of X chromosome inactivation in hiPSC clones, particularly in early passages, by methods such as HUMARA-MSP, is therefore important when using patient-specific hiPSCs to model X-linked disease.

Introduction

The first derivation of human induced pluripotent stem cells (hiPSCs) launched a promising new cellular platform with exciting potential for use in disease modeling and regenerative medicine.1,2,3 One advantage of hiPSCs is their capacity for self-renewal, which enables hiPSCs to undergo extensive clonal expansion in vitro while maintaining a pluripotent state. However, repeated passaging often leads to unfavorable changes, including chromosomal abnormalities,4,5 loss of tumor-suppressor genes,6 and altered copy number variation.6,7 Previous studies have reported that DNA methylation profiles, including those for imprinted genes, may also change during repeated passaging of hiPSCs.8,9,10

X chromosome inactivation is a form of epigenetic silencing that occurs in female mammals; it prevents female somatic cells from expressing X chromosome gene products at higher levels (2×) than male somatic cells.11,12 In mammals, a single X chromosome is randomly inactivated during gastrulation and remains inactive throughout development; this status is stably inherited through subsequent cell divisions. Although X chromosome inactivation status has been used as an index of reprogramming in the generation of mouse iPSCs,13 the stability of X chromosome inactivation in human iPSCs remains in question. Several groups have reported that X chromosome inactivation changes in response to the culture environment and during long-term culture of hiPSCs.10,14,15,16,17,18,19,20

Human embryonic stem cells (hESCs) were first established by Thomson and colleagues in 1998,21 and since that have time been used as a model system for the study of X chromosome inactivation. hESCs have been found to exhibit diverse X inactivation states as a result of variability in the expression level of X-inactive specific transcript (XIST) in different hESC lines or because of the number of times a given line has been subcultured.22,23,24,25,26,27 Silva et al. proposed that hESCs can be categorized into three classes based on X chromosome inactivation status.24 In Class 1 lines, both X chromosomes are activated (XaXa) in the undifferentiated state and lack X chromosome inactivation marks, such as XIST RNA and histone H3 lysine 27 trimethylation (H3K27me3).23,24,25,26 In Class 2 lines, a single X chromosome is targeted by XIST, and X chromosome inactivation proceeds, accompanied by H3K27me3 on Xi [XaXi, XIST(+)].22,23,24 Class 3 lines also retain a single inactive X chromosome, but XIST expression is absent and is not reactivated upon differentiation [XaXi, XIST(−)].22,23,24,25,26 Class 1 lines (with two active X chromosomes) are considered to be in a pre-X chromosome inactivation state, whereas Class 2 and 3 lines are in a post-X chromosome inactivation state, with one active and one inactive X chromosome. Human ESCs and iPSCs are in a primed state of pluripotency, similar to that observed in mouse epiblast stem cells, which are isolated from the post-implantation epiblast,28,29 and nearly all such cells belong to class 2 or 3. Class 1 and 2 hESC lines are epigenetically unstable and readily proceed toward class 3.22,23,24,25,26 However, recent research has shown that it is possible to maintain XaXa cells by the expression of reprogramming factors (OCT4, SOX2, KLF4, c-MYC) in combination with treatment with small molecules, such as ERK1/ERK2 and GSK3 signaling inhibitors, and leukemia inhibitory factor (LIF).30,31 Maintaining a physiological oxygen concentration in the culture environment during the establishment of hESCs has also been shown to stabilize XaXa cells,32 as have several other methods.33,34,35 Similarly, XaXi hiPSC lines show changes in X chromosome inactivation pattern and can be converted into XaXa hiPSC lines after several passages in the presence of feeder cells.16 However, histone deacetylase inhibitors and oxygen concentration have no effect on hiPSCs.17 Other studies have shown that erosion of X chromosome inactivation, which leads to the reactivation of genes in Xi-restricted H3K27me3-enriched domains, occurs during long-term hiPSC culture.18,19,20 These findings suggest that X chromosome inactivation is unstable and exhibits diverse states during the propagation of undifferentiated hESCs and hiPSCs.

Unstable X chromosome inactivation status may affect the detection of phenotypes in hiPSC models of X-linked diseases, such as Rett syndrome. Our group and others have reported that somatic cells from patients with Rett syndrome exhibit cellular mosaicism of wild-type and mutant MECP2-expressing cells.36,37,38,39 Clonally derived hiPSC clones also include both wild-type and mutant MECP2-expressing clones. In the present study, we evaluated the X chromosome inactivation status of hiPSC clones (RS1 and RS2) derived from fibroblasts of two patients with Rett syndrome.40 We also examined the epigenetic status of Rett syndrome and control hiPSCs during long-term culture to assess the epigenetic stability of X chromosome inactivation in these cells using a methylation-specific polymerase chain reaction (MSP) technique to detect changes in the HUMARA locus (HUMARA-MSP).41 We determined that the X chromosome inactivation state varied over the course of repeated passaging in hiPSC clones.

Materials and Methods

Ethics statement

The RS1 and RS2 fibroblast lines were derived from skin biopsies from a single pair of 10-year-old Japanese monozygotic female twins with Rett syndrome. Signed informed consent was obtained from the twins’ parents. The protocols for obtaining fibroblasts and their use in reprogramming were reviewed and approved by the Research Ethics Committees of Keio University (Approval No. 20080016) and the University of Yamanashi (Approval Nos. 523 and 699). This study was conducted in accordance with the principles expressed in the Declaration of Helsinki.

Generation of hiPSCs and cell cultures

Human wild-type hiPSC lines (WT-hiPSCs; WD5 and WD37) and Rett syndrome hiPSC lines (RS1-13, RS1-44, and RS1-76) were newly generated using standard methods in which retroviruses encoding OCT4, KLF4, SOX2, and c-MYC are transduced into skin fibroblasts. The hiPSC lines were maintained on a layer of mitomycin C-treated SNL 76/7 STO feeder cells (Wellcome Trust Sanger Institute, Cambridge, UK) in Dulbecco’s modified Eagle medium nutrient mixture F-12 (Wako, Kyoto, Japan) containing 20% KnockOutTM Serum Replacement (Gibco, Waltham, MA, USA), 4 ng/ml FGF-2 (PEPROTECH, Rocky Hill, NJ, USA), 2 mM glutamine (Nacalai Tesque, Kyoto, Japan), 0.1 mM nonessential amino acids (Sigma-Aldrich, Darmstadt, Germany), 50 units penicillin, 50 mg/ml streptomycin (Nacalai Tesque), and 0.1 mM β-mercaptoethanol (Sigma-Aldrich) at physiological oxygen concentrations.2,40 The hiPSC clones were characterized following a previously published protocol using G-band analysis, immunostaining with surface and nuclear pluripotency markers (TRA-1–81 and Oct4), and teratoma formation. Genomic DNA was collected at each passage for analysis by HUMARA-MSP.

Immunocytochemistry

Cells were fixed in 4% paraformaldehyde for 30 min at room temperature, washed twice with 1×phosphate-buffered saline (PBS), and blocked with 5% fetal bovine serum in 1×PBS at room temperature for 1 h. Primary antibodies were applied overnight at 4°C. Before use, Tra-1–81 monoclonal antibody (Millipore, Darmstadt, Germany) was diluted 1/1000, and Oct4 monoclonal antibody (Santa Cruz Biotechnology, Dallas, TX, USA) was diluted 1/100. After primary antibody incubation, cells were washed twice with PBS and incubated with Alexa Fluor®-conjugated secondary antibodies (Invitrogen, Waltham, MA, USA) and Hoechst stain.

DNA sampling and HUMARA-MSP

DNA was isolated from each of the hiPSC clones with DNeasy Blood and Tissue kits (QIAGEN, Hilden, Germany) following the manufacturer’s instructions. Bisulfite treatment of genomic DNA was performed using an EZ DNA Methylation-Lightning kit (Zymo Research, Freiburg, Germany). The bisulfite-treated DNA was amplified via polymerase chain reaction (PCR) with two sets of primers. The primer sets were designed within the CpG island (exon 1) of the HUMARA gene, which is located at Xq11-a12. X chromosome inactivation patterns were determined using HUMARA-MSP, as described in a previous report.41 The primer sequences used for HUMARA-MSP were as follows: AR-U forward, 5′-GCG AGC GTA GTA TTT TTC GGC-3′; AR-U reverse, 5′-AAC CAA ATA ACC TAT AAA ACC TCT ACG-3′; AR-M forward, 5′-GTT GTG AGT GTA GTA TTT TTT GGT-3′; AR-M reverse, 5′-CAA ATA ACC TAT AAA ACC TCT ACA-3′. Two PCR products were generated by each primer set, electrophoresed through 2.5% agarose gels, and separated using an ABI 310 DNA Sequencer equipped with GeneScan Software (Applied Biosystems, Waltham, MA, USA).

Results

Generation of hiPSC clones from Rett syndrome fibroblasts

We generated hiPSC clones (RS1 and RS2) from fibroblasts obtained from monozygotic twins with Rett syndrome using retroviral-mediated expression of four transcription factors (OCT4, KLF4, SOX2, c-MYC)40; a schematic of the experimental procedure is shown in Fig. 1A. We analyzed X chromosome inactivation statuses in both hiPSC lines using the HUMARA-MSP assay at an early passage (P6–P8). In this assay, X chromosome inactivation status can be determined by measuring the length of PCR products from the cytosine-adenine-guanine (CAG) repeats in the HUMARA gene: the lengths of these CAG repeats typically differ in the paternal and maternal X chromosomes.42 To distinguish inactive (methylated) and active (unmethylated) HUMARA gene alleles, we used two types of primers to specify unmethylated and methylated regions of the gene, following a previously described protocol.41 We then re-analyzed the X chromosome inactivation status in other hiPSC clones at later passages as shown in Fig. 1A and compared these results to hiPSC clones derived from control fibroblasts.

Fig. 1

Experimental scheme and generation of hiPSCs.

(A) Experimental scheme used to analyze the X chromosome inactivation (XCI) status of hiPSCs. (B) Bright-field image of Rett syndrome hiPSCs (RS1-13). Scale bar =300 µm. (C) Immunostaining for Oct4 (green) and TRA-1-81 (red) along with Hoechst staining (blue) of Rett syndrome hiPSCs (RS1-13). Scale bar =50 µm. (D) Images of a Rett syndrome hiPSC (RS1-13) karyotype. (E-G) In vivo differentiation of Rett syndrome hiPSCs stained with hematoxylin and eosin: (E) neural rosette-like tissue (ectoderm), (F) epithelium-like tissue (endoderm), (G) cartilage (mesoderm). Scale bar =100 µm. (H) Efficiency of Rett syndrome hiPSC colony formation per 90-mm-diameter dish.

The hiPSC clones used in this study exhibited typical hiPSC morphologies (Fig. 1B) and expressed pluripotency markers (Fig. 1C). These clones were confirmed to have normal karyotypes (Fig. 1D) and the capacity for differentiation into lineages representing the three germ layers (Fig. 1E–G). We analyzed hiPSC colony formation approximately 28 days after transduction to count the number of hiPSC or non-hiPSC colonies formed following the reprogramming of fibroblasts from each of the donors. Although RS1 formed approximately twice as many hiPSC colonies as RS2 did, the difference between RS1 and RS2 in colony-forming efficiency was not statistically significant (Fig. 1H).

Analysis of X chromosome inactivation status in hiPSC clones derived from patients with Rett syndrome

We sought to determine the influence of X chromosome inactivation in hiPSCs over multiple passages using HUMARA-MSP, which detects the methylation status of CpG sites in exon 1 of the X chromosome gene human androgen receptor (AR). A schematic of the locus amplified by the HUMARA-MSP is shown in Fig. 2A. When CpG dinucleotides are methylated, indicating that the X chromosome is inactive, amplified products are observed with AR-M (methylated) primers. The presence of amplified products when AR-U (unmethylated) primers are used indicates that the X chromosome is active. The lengths of the PCR products from the CAG repeat region in the AR gene, located at Xq11-12, were 177 bp and 210 bp in RS1 and RS2, respectively. Most of the hiPSCs exhibited only a single peak on PCR with AR-M and AR-U primers. Specifically, approximately 90% of the hiPSC lines displayed a nonrandom pattern of X chromosome inactivation in the undifferentiated state. These data suggest that most of the Rett syndrome hiPSC lines used in this study remained in the post-X chromosome inactivation state; however, there were several clones with two activated X chromosomes (Table 1). The lengths of the HUMARA-MSP products were confirmed by electrophoresis to detect inherited CAG repeat numbers for each chromosome. Electrophoresis showed that products amplified with AR-U primers amplified multiple bands in RS1-13 and RS1-44, whereas AR-M primers amplified a single band (Fig. 2B). These observations suggest that these hiPSC clones include cells in both pre-X chromosome inactivation and post-X chromosome inactivation states within individual colonies, indicating partial-X chromosome inactivation within hiPSC clones.

Fig. 2

X chromosome inactivation analysis of Rett syndrome hiPSC clones with early passage numbers.

(A) Diagram of the X chromosomes containing the AR locus. The respective lengths of the PCR products of the HUMARA gene in the patient fibroblasts were 177 and 210 bps. (B) Electrophoretic separation of the HUMARA-MSP products in established Rett syndrome hiPSCs. U, PCR products amplified with AR-U primers; M, PCR products amplified with AR-M primers.

Table 1. Classification of X chromosome inactivation status in Rett syndrome hiPSCs
Patient Total XaXi clones Partial XaXa clones
RS1 44 39 5
RS2 37 33 4

Analysis of X chromosome inactivation status in hiPSC clones after multiple passages

We next examined whether the variable X chromosome inactivation status in RS1-13 and RS1-44 clones changed over multiple passages in long-term cell culture. Peak images were quantified using GeneScan® software. Three Rett syndrome hiPSC clones (RS1-13, RS1-44, and RS1-76) newly generated in this study were examined at passages 11 (or 12), 20, and 25. Interestingly, the extra peaks observed in RS1-13 and RS1-44 were diminished by P25, suggesting that X chromosome inactivation had been completed during the passaging of these lines (Fig. 3A). We did not observe re-activation of inactivated X chromosomes after repeated passaging of hiPSC clones with X chromosome inactivation (RS1-76) (Fig. 3A). We also examined two normal control hiPSC clones, WD5 and WD37, both newly obtained from a healthy donor. Because it was difficult to distinguish these two bands by electrophoresis (data not shown), we used GeneScan® analysis to visualize peaks amplified by HUMARA-MSP. The HUMARA-MSP method is highly sensitive to subtle differences, enabling us to detect differences as small as 6 bp between paternal and maternal CAG repeats. There were 195 bp and 201 bp CAG repeats in the control lines (data not shown). Both WD5 and WD37 had extra peaks, indicating that they were partial-X chromosome inactivation hiPSC clones at an early passage (P5 or P6). The additional peaks gradually diminished with repeated passages, as was the case for RS1-13 and RS1-44 (Fig. 3B). These data suggest that human hiPSCs eventually undergo X chromosome inactivation in the presence of FGF-2 (in the “primed” state of pluripotency), despite possessing two active X chromosomes at early passages.

Fig. 3

X chromosome inactivation analysis of Rett syndrome and control hiPSCs using the HUMARA-MSP assay.

(A) X chromosome inactivation patterns in Rett syndrome hiPSC lines (RS1-13, RS1-44, and RS1-76), (B) X chromosome inactivation patterns in control hiPSC lines (WD5 and WD37). Red peaks indicate PCR products amplified with AR-M primers (i.e., inactive X chromosome; Xi); blue peaks indicate PCR products amplified with AR-U primers (i.e., active X chromosome; Xa). The dashed lines indicate the same numbers of CpG measured using an ABI 310 DNA Sequencer equipped with GeneScan® Software. Specifically, a left peak indicates fewer and a right peak indicates more CpG repeats. The vertical axis indicates peak height as determined by the area derived from PCR products; the horizontal axis shows numbers of CpG repeats. Xa and Xi peak area ratios represent X chromosome inactivation patterns.

Discussion

We have shown that single cell-derived hiPSC clones exhibit varying X chromosome inactivation states immediately after clonal isolation, even when simultaneously established from a single donor. Approximately 11% of hiPSC clones had two activated X chromosomes at early passages (P < 5), and these hiPSC clones eventually acquired X chromosome inactivation over long-term culture, thus converting to an XaXi state. These results are consistent with previous reports.8,15,22,23,32 In studies of X-linked diseases, such as Rett syndrome, using hiPSCs, the somatic cells used to derive the pluripotent cell lines may be one of two different cell types, expressing either maternal or paternal alleles of the X chromosome. Somatic cells from patients with Rett syndrome include both wild-type MECP2-expressing cells and mutant MECP2-expressing cells.37,38 Therefore, hiPSC lines that are clonally derived from patients with Rett syndrome clonally express either wild-type MECP2 or mutant MECP2.40,43,44,45 Fortunately, such hiPSC clones are isogenic and represent an ideal pair for studying the specific function of MeCP2 while excluding effects from the genetic background. Although several groups have reported that hiPSCs in the primed state of pluripotency have the same X chromosome inactivation status as the original cells,14,15 reactivation of X chromosomes has also been reported in primed hiPSCs.16,17,18,19,20 In the present study, we used a standard hiPSC culture medium with FGF-2 for maintenance and used retroviral vectors for reprogramming, as described in the first report on hiPSCs.2 Although there have been numerous reports of patient-specific hiPSCs generated using these retroviral vectors,46,47,48,49,50 hiPSC clones with two active X chromosomes may have been included in these previous reports if they were analyzed for X inactivation at an early passage.

Our current results indicate that X chromosome inactivation status is variable in hiPSC clones at early passages, and that this variability may affect the phenotypes of X-linked diseases. To enable reliable analysis of patient-specific hiPSC studies of X-linked diseases, the X chromosome inactivation status of the clones must be evaluated carefully, especially in early passages. Our findings suggest that HUMARA-MSP is a useful tool for evaluating X chromosome inactivation.

It would be interesting to investigate the stability of the methylation status of X chromosome of hiPSCs derived from a male Rett syndrome patient (XY) over extended passaging. However, Rett syndrome patients are nearly all female, because of the nonviability of male hemizygous conceptuses, making acquisition of cellular material from male donors exceedingly difficult.51,52 We hope to conduct such studies in the future, pending availability of the requisite cellular resources.

Acknowledgments

We thank all the members of the Okano laboratory for helpful comments and discussions. The following grants and programs partially supported the research described in this study: grants from the New Energy and Industrial Technology Development Organization, the Ministry of Education, Science, Sports and Culture (MEXT), and the Ministry of Health, Labour and Welfare (MHLW) of Japan to H.O. and W.A.; the Program for Intractable Disease Research Utilizing Disease-specific Human iPS Cells funded by the Japan Science and Technology Agency (JST)/Japan Agency for Medical Research and Development (AMED) to H.O. and W.A.; Grants-in-Aid for Scientific Research (KAKENHI); and the Ministry of Economy, Trade and Industry (METI) of Japan for “Development of Core Technologies for Innovative Drug Development Based Upon IT” and for the Project Focused on Developing Key Technologies for Discovering and Manufacturing Drugs for Next-Generation Treatment and Diagnosis (biological-verifying studies) to T.K.

References
 
© 2016 by The Keio Journal of Medicine
feedback
Top