Research ArticleSKIN DISEASE

Induced pluripotent stem cells from human revertant keratinocytes for the treatment of epidermolysis bullosa

See allHide authors and affiliations

Science Translational Medicine  26 Nov 2014:
Vol. 6, Issue 264, pp. 264ra164
DOI: 10.1126/scitranslmed.3009342


Revertant mosaicism is a naturally occurring phenomenon involving spontaneous correction of a pathogenic gene mutation in a somatic cell. It has been observed in several genetic diseases, including epidermolysis bullosa (EB), a group of inherited skin disorders characterized by blistering and scarring. Induced pluripotent stem cells (iPSCs), generated from fibroblasts or keratinocytes, have been proposed as a treatment for EB. However, this requires genome editing to correct the mutations, and, in gene therapy, efficiency of targeted gene correction and deleterious genomic modifications are still limitations of translation. We demonstrate the generation of iPSCs from revertant keratinocytes of a junctional EB patient with compound heterozygous COL17A1 mutations. These revertant iPSCs were then differentiated into naturally genetically corrected keratinocytes that expressed type XVII collagen (Col17). Gene expression profiling showed a strong correlation between gene expression in revertant iPSC–derived keratinocytes and the original revertant keratinocytes, indicating the successful differentiation of iPSCs into the keratinocyte lineage. Revertant-iPSC keratinocytes were then used to create in vitro three-dimensional skin equivalents and reconstitute human skin in vivo in mice, both of which expressed Col17 in the basal layer. Therefore, revertant keratinocytes may be a viable source of spontaneously gene-corrected cells for developing iPSC-based therapeutic approaches in EB.


Revertant mosaicism (RM) refers to the coexistence of cells in one individual carrying germline mutations together with cells in which the inherited, disease-causing mutation is corrected by a spontaneous genetic event, thereby giving rise to a mixture of mutant and corrected cells, also named revertant cells. Several mechanisms can account for RM, including a reverse point mutation, crossing-over, gene conversion, and a second-site mutation (13). Somatic reversion of a mutant phenotype was first identified in Lesch-Nyhan syndrome in 1988 (4). Since then, RM has been identified in several human genetic disorders, including Duchenne muscular dystrophy, adenosine deaminase deficiency, Fanconi anemia, and Wiskott-Aldrich syndrome (1).

Among skin diseases, RM has been described for the genetic blistering disorder epidermolysis bullosa (EB), which is considered a prototype of cutaneous basement membrane diseases. EB is a diverse group of genodermatoses, including EB simplex, junctional EB, dystrophic EB, and Kindler syndrome (5). EB is characterized by the formation of blisters and erosions in the skin and mucous membranes in response to trauma. The standard care for EB is dressing or skin grafting for the blisters and ulcers; however, this is challenging owing to the fragility of the skin. Patients with EB have a markedly shortened life span depending on the clinical and genetic subtypes (6).

Clinical trials are being pursued using gene-, protein-, drug-, and cell-based approaches for the treatment of EB (7). Several approaches of gene therapy including transplantation of genetically modified epidermal keratinocytes using retroviral vector (8, 9) and induced pluripotent stem cell (iPSC)–based gene therapy (10) were previously reported. In recent clinical trials, keratinocytes as well as fibroblasts and bone marrow–derived stem cells, including mesenchymal stem cells, have been evaluated for cell-based therapy and tissue engineering applications. Although several clinical approaches for treatment of EB, such as allogeneic fibroblast injection (11) and bone marrow transplantation (12, 13), are under development, currently there is no cure for any form of EB.

RM was previously reported to occur in the skin of patients with mutations in five different EB genes, including KRT14 encoding keratin 14 (14), LAMB3 encoding the β3 chain of laminin-332 (15), COL17A1 encoding type XVII collagen (16, 17), COL7A1 encoding type VII collagen (18), and FERMT1 encoding kindlin-1 (19). Recent studies suggest that RM is not a rare event (20), and it has further been postulated that perhaps all patients with generalized non-Herlitz junctional EB (JEB-nH) with COL17A1 mutations exhibit some revertant patches (21). Clinically, blistering cannot be evoked in these patches of normal-appearing skin even after mechanical friction. Histologically, normal expression of the revertant gene product is detected, whereas expression is typically lacking in affected areas (3). Progressive improvement and expansion of the RM skin patches have been observed in some cases, suggesting that the RM occurred in epidermal progenitors or stem cell niches. A recent study reported that RM for EB was observed only in keratinocytes and not in blood or fibroblasts (15).

Patient-specific iPSCs are advantageous for cell- and gene-based therapies owing to their unlimited proliferative capability and ability to differentiate into multiple types of cells (22). To enable regenerative approaches to be applicable across a broad range of disorders, the development of safe and efficient genome-editing technologies is essential in iPSC-based approaches. Several technologies have been developed using disease-specific iPSC lines, such as zinc finger nucleases and transcription activator–like effector nucleases (TALENs), to improve the efficiency of the targeted DNA double-strand breaks. However, the low efficiency of gene targeting by spontaneous double-strand break–mediated conventional homologous recombination, and the risks associated with undesired artificial modifications in the host genome (23) currently preclude clinical application.

Several reports have shown the use of revertant skin cells for EB therapy. Cultured revertant keratinocytes taken from a patient with JEB-nH, caused by mutations in COL17A1, were expanded in culture and grafted back onto the patient (2). However, functional improvement was not observed in this case, and blistering could be provoked owing to the low percentage of revertant cells in the graft (<3%). Biopsy specimens of revertant skin taken from a JEB-nH patient with mutations in LAMB3 were transplanted into blistered regions, resulting in long-term restoration of healthy skin; however, this method of punch grafting covered only small areas and allowed limited expansion of revertant skin (24). COL17A1 revertant keratinocytes display a growth disadvantage compared to mutant keratinocytes taken from the same skin biopsy when expanded in vitro (25). However, the percentage of COL17A1 revertant keratinocytes stabilizes during skin equivalent production and remains stable in vivo, showing long-term survival of revertant cells. These studies encouraged us to use patient-specific iPSCs from revertant keratinocytes for cell therapy, because this would help overcome current limitations of in vitro expansion, such as depletion of stem cells.

We previously reported the generation of iPSCs from patients with EB and efficient derivation of functional keratinocytes from unaffected and EB iPSCs (26). However, this approach required gene correction of the mutant iPSCs. Here, we present a strategy for “natural gene therapy” by generating patient-specific iPSCs from the revertant keratinocytes present in RM patches in JEB-nH patients. These revertant iPSCs were then differentiated into keratinocytes with normal COL17A1 and used to create skin equivalents in vitro and reconstitute skin on mice in vivo. This approach represents an important step forward in generating renewable, spontaneously genetically repaired, patient-specific cells for EB therapy.


Generation and characterization of revertant iPSCs

Revertant skin was identified as clinically unaffected areas in a female patient (EB026-01) with JEB-nH (16). She had the characteristic phenotype of JEB-nH with generalized skin blistering from birth, dental abnormalities, universal alopecia, and nail dystrophy. The proband was compound heterozygous for two COL17A1 mutations, a maternally inherited frameshift mutation c.1601delA and a paternally inherited nonsense mutation c.3676C>T; p.R1226X. Patches of healthy skin were observed on different body sites, including the dorsal part of the left hand and arm (Fig. 1A). In passage 2, about 70% of the keratinocytes cultured from biopsies of revertant skin patches from patient EB026-01 expressed Col17 by immunostaining, whereas the other 30% consisted of mutant keratinocytes, which were negative for Col17 (Fig. 1B) (16).

Fig. 1. Col17 expression in revertant mosaic keratinocytes.

(A) Clinical appearance of RM in patient EB026-01. Patches of the skin that never blister (outlined with a dashed line) show pigmentation. Adapted from (25) with permission. (B) Cultured keratinocytes taken from a revertant skin area expressed Col17 in passage 2. Scale bar, 30 μm. DAPI, 4′,6-diamidino-2-phenylindole. (C) Schematic representation of our study. Strategies for generating iPSCs were combined with the concept of natural gene therapy, which can avoid genetic modifications by starting with revertant keratinocytes. These naturally corrected iPSCs can be differentiated into several types of cells for cell therapy, including keratinocytes (KCs), fibroblasts (FBs), and bone marrow stem cells (BMSCs). (D) Sequence analysis of revertant and mutant iPSCs from two patients showed correction of the exon 18 c.1601delA mutation (blue arrow) in revertant iPSCs, whereas the heterozygous COL17A1 mutation c.3676C>T was still present in exon 51 (red arrow). Mutant iPSCs had a c.2237delG mutation (red arrow) without secondary mutation. (E) Morphology of revertant iPSCs and mutant iPSCs. (F) Immunocytochemical analysis of iPSCs using stem cell markers alkaline phosphatase (AP), OCT4, SOX2, SSEA3, TRA-1-60, and TRA-1-81. Scale bars, 100 μm.

We generated iPSCs from both revertant and mutant keratinocytes by retroviral transduction of the four reprogramming factors c-MYC, SOX2, OCT4, and KLF4 (Fig. 1C). After culturing the transduced cells on feeder layers, we obtained both revertant (EB026-01) and mutant (EB025-01) iPSC lines. Sequence analysis of revertant iPSCs identified the heterozygous c.3676C>T transition in exon 51 (p.R1226X); however, the mutation in exon 18 (c.1601delA) was corrected by mitotic gene conversion (Fig. 1D), as described previously (16). The mutant iPSCs were generated from a second patient who was homozygous for the c.2237delG mutation (EB025-01) (21) (Fig. 1D). The morphology of two iPSC lines exhibited small-cell shape and had a large nuclear-to-cytoplasmic ratio, prominent nucleoli, and tight and flat colonies, which appeared similar to that of human embryonic stem cells (ESCs) (27) (Fig. 1E). The iPSC colonies were identified by positive AP staining, and the expression of several stem cell markers, including OCT4, SOX2, SSEA3, TRA-1-60, and TRA-1-81, was confirmed (Fig. 1F).

Quantitative polymerase chain reaction (PCR) showed that viral transgene silencing was observed (Fig. 2A) and that additional pluripotent markers were expressed in both revertant iPSCs and mutant iPSCs, as well as ESCs demonstrated by reverse transcription PCR (RT-PCR) (Fig. 2B). Bisulfite sequencing revealed that the NANOG and OCT4 promoter regions in revertant iPSCs and mutant iPSCs were demethylated (reactivated) in a pattern similar to human ESCs, demonstrating that our iPSCs were epigenetically reprogrammed (Fig. 2C). Both iPSC lines grew at similar rates to human ESCs and maintained normal karyotypes (46, XX) after culturing for 20 passages (fig. S1A). High-resolution single-nucleotide polymorphism (SNP) oligonucleotide microarray analysis (SOMA) showed no substantial changes in copy number variants in both iPSC lines compared to the original keratinocytes, except one small (65-kb) deletion on chromosome 13 in mutant iPSCs (fig. S1B).

Fig. 2. Transcriptional profile of revertant iPSCs.

The authenticity of both iPSC lines was analyzed at passages 18 to 28. (A) Quantitative PCR analysis of the four viral (v) reprogramming factors, c-MYC, SOX2, OCT4, and KLF4. Transgene silencing was observed in generated revertant iPSCs and mutant iPSCs as well as human ESCs and normal human keratinocyte (NHK). All samples were compared to the transduced fibroblasts, which were transfected with all four viral reprogramming factors and harvested after 7 days. (B) RT-PCR analysis of stem cell markers in revertant iPSCs and mutant iPSCs. RT+, with reverse transcriptase; RT, without reverse transcriptase. (C) Methylation analysis of NANOG and OCT4 promoter regions in revertant iPSCs and mutant iPSCs compared to human ESCs and normal human fibroblasts (nFB). White circles represent unmethylated cytosine-phosphate-guanosine (CpGs); black circles represent methylated CpGs.

To demonstrate the pluripotency of iPSCs, we examined the expression of germ layer–specific markers in embryoid bodies, which promote the spontaneous differentiation of ESCs and iPSCs. Our analysis of embryoid bodies demonstrated that both revertant iPSCs and mutant iPSCs were capable of differentiating into all three germ layers in vitro. Moreover, we also confirmed the in vivo differentiation capacity of our iPSCs by the presence of markers of all three germ layers in teratomas that formed after intradermal injection of iPSCs into nude mice (fig. S2).

Expression of Col17 in keratinocytes derived from revertant iPSCs

To derive keratinocytes from iPSCs, we used a modified protocol using retinoic acid (RA) and bone morphogenetic protein 4 (BMP4) (26). During the expansion periods, keratinocytes derived from both revertant iPSCs (revertant-iPSC KC) and mutant iPSCs (mutant-iPSC KC) as well as NHKs grew in the keratinocyte medium. Revertant-iPSC KC showed strong Col17 protein expression by immunofluorescence staining as early as passage 1 (P1), whereas mutant-iPSC KC did not show Col17 expression in any keratinocytes from P1 to P4. Revertant iPSCs did not express Col17 by immunostaining (Fig. 3A).

Fig. 3. Differentiation of patient-derived revertant iPSCs into keratinocytes.

(A) Col17 protein expression. Negative control was NHKs without primary antibody. Scale bar, 100 μM. (B) Heat map with dendrograms of row z scores calculated from the mean log2(intensity) of the top 25 and bottom 25 differentially expressed genes in the original revertant keratinocytes compared with revertant iPSCs. COL17A1 gene expression is highlighted by a yellow box. (C) Principal components plot of the 23,459 TCIDs that were mapped by IPA, had gene symbols and ensemble transcript (ENST) IDs, and were on Chr 1 to 22, X or Y. (D) Spearman correlation heat map generated using IPA-mapped TCIDs that had abs[FC(intensity)] ≥2 in the comparison between original revertant keratinocytes and revertant iPSCs. Duplicate TCIDs corresponding to the same gene symbol were removed as described in Materials and Methods. (E) Comparison of the global mRNA expression patterns of revertant-iPSC KC, revertant iPSCs, and the original revertant keratinocytes by microarray analysis. Scatterplots were generated for the TCIDs that were used in (C). Points in red had abs[FC(intensity)] >2. Points on the plot correspond to the mean log2 of Affymetrix Transcript Cluster IDs calculated from technical replicates (n = 3 for the revertant-iPSC FB cell line, n = 2 for the nFB, and n = 1 each for the revertant-iPSC KC, revertant iPSCs, and the original revertant keratinocytes, for a total of eight microarrays).

Gene expression profiling of revertant iPSCs and revertant-iPSC KC

Global gene expression profiling was performed with the Affymetrix Human Transcriptome Array (HTA) 2.0 to compare revertant-iPSC KC (P1) to the original revertant keratinocytes (P3), revertant iPSCs (P21), revertant iPSC–derived fibroblasts (revertant-iPSC FB) (P9), and normal human fibroblasts (nFB) from a different donor (P4). Microarray analysis showed more than 24-fold up-regulation of COL17A1 gene expression in revertant-iPSC KC and the original revertant keratinocytes compared to revertant iPSCs, revertant-iPSC FB, and nFB. As expected, there was no differential expression observed between revertant-iPSC KC and the original revertant keratinocytes (Fig. 3B and table S1). This indicates that Col17 protein expression is induced early in the differentiation of iPSC into iPSC KC.

Global gene expression profiling showed a strong similarity between revertant-iPSC KC and original revertant keratinocytes (Fig. 3, B to E). A plot of the first two principal components (PC1 and PC2) generated for each cell type from the 23,459 Ingenuity Pathway Analysis (IPA)–mapped transcript clusters identifications (TCIDs) demonstrated the similarity between revertant-iPSC KC and the original revertant keratinocytes, showing that these two clustered together and are apart from the revertant-iPSC FB/nFB cluster and revertant iPSCs (Fig. 3C). The correlation coefficient between the original revertant keratinocytes and revertant iPSCs was 0.3 compared with 0.81 between the original revertant keratinocytes and revertant-iPSC KC. This suggests that the revertant-iPSC KC global expression profile was more similar to that of original revertant keratinocytes than to that of the revertant iPSCs (Fig. 3D). Hierarchical clustering grouped the original revertant keratinocytes with the revertant-iPSC KC, the nFB with the revertant-iPSC FB, and the revertant iPSCs separately (Fig. 3C).

Similarity of the expression profiles of the original revertant keratinocytes and revertant-iPSC KC was further substantiated by Gene Ontology (GO) enrichment analysis performed on lists of genes differentially expressed in revertant-iPSC KC compared to the original revertant keratinocytes, which identified no GO terms significantly enriched in down-regulated genes (table S2). The GO terms enriched in up-regulated genes showed differences between revertant-iPSC KC compared to the original revertant keratinocytes owing to an increase in expression of mesoderm, which suggested that inappropriate differentiation may be occurring in the revertant-iPSC KC at low levels (table S2). Conversely, the differences between revertant iPSCs and original revertant keratinocytes identified ectoderm development and epidermis development in the top two down-regulated overrepresented GO biological process terms (table S3), suggesting the successful differentiation of iPSCs into the keratinocyte lineage.

Terminal differentiation of three-dimensional skin equivalents and skin reconstitution from revertant-iPSC KC

We next examined whether revertant-iPSC KC can express Col17 protein in the basement membrane zone. Three-dimensional (3D) skin equivalents were generated using both revertant-iPSC KC (P3) and mutant-iPSC KC (P3) and compared to NHK (P3). In the 3D skin equivalents using revertant-iPSC KC and nFB, Col17 was expressed at the basement membrane (Fig. 4A). In contrast, there was no expression of Col17 in 3D skin equivalents using mutant-iPSC KC. To confirm the functionality of revertant-iPSC KC, we also examined the 3D skin equivalents by immunostaining using several keratinocyte markers. Type VII collagen, keratin 5, and keratin 14 were expressed in the basal layer; desmoplakin and desmoglein 3 were observed in the suprabasal layers; and keratin 1 and loricrin were expressed in the upper layers (Fig. 4B).

Fig. 4. Generation of 3D skin equivalents using revertant-iPSC KC.

(A) Three-dimensional skin equivalents were generated in vitro using revertant-iPSC KC, which expressed Col17 protein in the basal layer (white dotted line), comparable to NHK. Negative control: NHK without primary antibody. (B) Three-dimensional skin equivalent expression of keratinocyte markers (in green). Negative control: NHK without primary antibody. Nuclei were stained with DAPI (blue). Scale bars, 100 μM. H&E, hematoxylin and eosin.

To preliminarily assess in vivo functional capacity of revertant-iPSC KC, we performed a skin reconstitution chamber assay in immunocompromised mice. Revertant-iPSC KC combined with nFB successfully engrafted in the chamber assay and reconstituted human skin on the back of severe combined immunodeficient (SCID) mice. The human skin generated from revertant-iPSC KC and nFB was identified by coimmunostaining of human nuclei, and the reconstituted skin expressed human Col17 at the basement membrane zone, similar to NHK and nFB (Fig. 5). In addition, human type VII collagen and human keratin 14 were expressed in the basal layer, which appeared as “holes” owing to sectioning through a rete ridge. Human desmoglein 3 was observed in the suprabasal layers, and human keratin 1 was expressed in the upper layers (Fig. 5).

Fig. 5. Generation of reconstituted human skin in vivo using revertant-iPSC KC.

Revertant-iPSC KC (15 × 106) combined with human nFB (12 × 106) cells contributed to formation of normal skin in vivo in immunocompromised mice. Skin reconstitution was compared with the human NHK/nFB combination. Mouse skin and a negative control (no antibody) are provided for comparison. Reconstituted skin was stained for human Col17 protein and other keratinocyte markers. Human skin was identified by anti-human nuclei, anti-human desmoglein 3, or anti-human keratin 1, which does not recognize mouse skin. The apparent “holes” in the epidermis represent sections through rete ridges in the reconstituted skin. The white dotted line indicates the basement membrane. Blue, DAPI (both human and mouse nuclei). Scale bars, 100 μM.


Autologous (patient-specific) iPSCs have the potential to provide an unlimited source of cells for gene and cell therapies for certain human diseases. Here, we generate iPSC from JEB-nH revertant keratinocytes and differentiate them into functional keratinocytes, in an effort to use cells from EB patients who have already spontaneously corrected their genomes. In such natural gene therapy, we can avoid the step of gene correction, as would be necessary for a starting cell population of mutant keratinocytes (Fig. 1C). RM is not a rare event among different genetic diseases, occurring in the keratinocytes of all patients with generalized JEB-nH (21). Hence, revertant cells could be a viable starting point for generating patient-specific iPSCs and treating a diverse range of diseases.

We demonstrated the feasibility of revertant iPSC–based therapy generating iPSCs from revertant epidermis and differentiating them into keratinocytes that expressed Col17 protein in the early stages of keratinocyte differentiation. Gene expression profiles in revertant-iPSC KC were consistent with that of the original revertant keratinocytes. Furthermore, the functionality of our revertant-iPSC KC was shown by immunostaining of 3D skin equivalents in vitro and skin reconstitution on immunodeficient mice in vivo that expressed human Col17 protein in the basal layer. According to these findings, our differentiation protocol generates populations of iPSC KC that are similar to NHK. We did not test these cells in a mouse model of EB to show therapeutic efficacy. However, a recent study by Wenzel et al. demonstrates that gene-corrected mouse iPSCs can engraft in animal models of EB, forming functional, type VII collagen–producing skin (28).

The efficiency of generating revertant iPSCs from EB patients also depends on the ratio of revertant and mutant cells in the biopsy samples. Therefore, the evaluation of the original skin samples is of great importance. Because all iPSC lines generated from original revertant keratinocytes of patient EB026-01 were revertant iPSCs, we examined mutant iPSCs from patient EB025-01. In EB patients, RM can manifest as patches of seemingly normal skin that are resistant to blister formation using fragility tests like tape stripping. Clinical evidence of RM is verified through skin biopsy analysis, typically using immunohistochemistry and/or transmission electron microscopy, as well as the identification of the germline mutations in DNA from blood, and identification of the reversion mechanism in the keratinocytes whenever possible. In our case, normal-appearing hemidesmosomes were seen adjacent to hypoplastic hemidesmosomes in the samples of clinically unaffected skin by electron microscopy (16), suggesting that normal Col17 protein was produced by the revertant keratinocytes. In patients with JEB-nH caused by COL17A1 mutations, RM is easily identified as clinically healthy skin owing to patches of homogeneously pigmented skin (21) (Fig. 1A).

Our strategy of generating patient-specific iPSCs from RM epidermis could enable gene therapy approaches to be combined with cell-based treatments, providing autologous sources of naturally gene-corrected iPSCs that can be differentiated into a variety of cell types and transplanted for long-term tissue maintenance (29). Here, we demonstrated the functionality of epidermal progenitors using skin reconstitution in vivo. The success of this approach would facilitate both “outside-in” stem cell therapy in the form of grafts and “inside-out” therapy in the form of systemic transplantation in the future (13). In future studies, genomic sequencing could be used to show the genome integrity in iPSCs and iPSC-derived cells for clinical application. Furthermore, long-term assessment of iPSC-derived cells in vivo will be needed to demonstrate long-lasting functionality.

To realize the clinical use of iPSCs, evidence of safety and efficacy is crucial. Here, we presented the strategy of natural gene therapy starting with RM. We would also emphasize that improvement of genome-editing technology will be necessary for EB patients without RM. Recently, several targeted genome-editing strategies mediated by zinc finger nucleases, TALENs, and adenovirus-associated viral (AAV) variant have been applied to human iPSCs in several diseases, including EB. Gene-corrected iPSCs were differentiated into target cells, including neurons (30), erythrocytes (31), hepatocytes (32), and keratinocytes (10). Gene targeting of human iPSCs by homologous recombination is still inefficient, and contamination of the host genome with residual nonhuman sequences might be a risk for clinical use. The study by Sebastiano et al. using AAV genome editing could lead to cell therapy of EB, and progress is already being made toward clinical use by manufacturing iPSC lines from patients with RDEB (recessive dystrophic EB) under good manufacturing practice conditions (33). Going forward, safety and efficiency will be the most important factors for clinical translation of iPSCs.

The use of viral vectors for reprogramming could have insertional mutagenesis or lead to tumorigenesis (34), which is currently a limitation for clinical translation. However, we did not detect any changes of karyotypes or copy numbers in revertant iPSCs over passage 20 compared to original revertant keratinocytes. Several nonviral systems have been explored in inducing reprogramming, including piggyback transposons (35), plasmid-based derivation (36), recombinant proteins (37), integration-free viral vectors (38), and mRNA (39).

For safe and efficient clinical translation, our strategy of generating iPSCs from RM would avoid these gene-targeting processes altogether. Furthermore, this methodology would be widely applicable not only to EB but also to several other inherited diseases in which RM has been observed, such as Duchenne muscular dystrophy and Wiskott-Aldrich syndrome. The combination of natural gene therapy with nonviral reprogramming will greatly advance the feasibility of iPSC-based therapy for clinical use in the future.


Study design

We hypothesized that keratinocytes differentiated from revertant iPSCs could show functionality similar to the original revertant keratinocytes in vitro and in vivo. The aim of this study was to present a strategy of natural gene therapy for treatments of EB by generating iPSCs starting from RM epidermis. We generated revertant iPSCs from a JEB-nH patient (EB026-01) and mutant iPSCs from another JEB-nH patient (EB025-01). We checked the authenticity of these iPSCs at more than passage 18 to examine their stability and cultured them up to 30 passages. iPSCs were differentiated into keratinocytes to examine their functionality. Gene expression profiling was performed on one biological sample each of cells extracted from original revertant keratinocytes, revertant iPSCs, revertant-iPSC KC, and revertant-iPSC FB, which were all from patient EB026-01, and normal fibroblasts extracted from a healthy control. Fold change was calculated between each cell type to identify differentially expressed genes and to demonstrate that COL17A1 was expressed by both the original revertant keratinocytes and the revertant-iPSC KC. Principal component analysis, correlation analysis, hierarchical clustering, and GO enrichment analysis were used to confirm that revertant-iPSC KC were similar to the original revertant keratinocytes. There was no blinding or randomization for skin reconstitution chamber assays.

Cell culture

Written informed consent was obtained from both patients for using cells left over after the diagnostic process for this study. iPSCs were generated from human keratinocytes, as described previously, with minor modifications (26). Human keratinocytes were isolated from revertant and affected skin of two patients (EB026-01 and EB025-01) with type XVII collagen–deficient JEB-nH, which were reported previously (16, 21). These cells were transduced by pMXs-based retroviruses with four transcription factors: c-MYC, SOX2, OCT4, and KLF4. After 5 days, the transduced cells were seeded on mitomycin C (MMC)–treated mouse embryonic fibroblasts (MEFs) in human ESC medium (ESM) [knockout (KO) Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 20% KO serum replacement, 1% GlutaMAX-I, 1% nonessential amino acid, 1% penicillin-streptomycin (Invitrogen), and basic fibroblast growth factor (FGF) (4 ng/ml) (R&D Systems)] until colonies appeared. Colonies of iPS cells were mechanically picked up and replaced on MMC-treated MEF for expansion and characterization. NHK was cultured in CnT-57 medium (low-serum condition) (CELLnTEC), and nFB were cultured in DMEM (Invitrogen) supplemented with 10% bovine serum.

Mutation analysis of patient-specific iPSCs

With the genomic DNA extracted from iPSCs, exons 18, 30, and 51 of COL17A1 with adjacent sequences of exon/intron borders were amplified by PCR using gene-specific primers (4). The amplification conditions were 94°C for 2 min, followed by 35 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 50 s, with a final extension at 72°C for 10 min. The amplified PCR products were directly sequenced in an ABI PRISM 310 Automated Sequencer, using the ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems). The amplified PCR products were observed on 8% polyacrylamide gel electrophoresis, and genotypes between the starting keratinocytes and PS-iPSCs were compared.

Polymerase chain reaction

RNA was extracted using the RNeasy Mini Kit (Qiagen), and DNA was removed by DNase treatment (Invitrogen) to avoid genomic DNA amplification. Complementary DNA (cDNA) was synthesized using total 2 μg of RNA by SuperScript III Reverse Transcriptase and Oligo(dT) Primer (Invitrogen) according to the manufacturer’s instructions. PCRs were performed with Platinum PCR SuperMix (Invitrogen). Human β2 microglobulin (β2M) was used to normalize target gene expression. Quantitative PCR was performed on an ABI 7300 machine and analyzed with ABI Relative Quantification Study software (Applied Biosystems). All reactions were performed with Power SYBR Green PCR Master Mix (Applied Biosystems). The following protocol was used: step 1, 50°C for 2 min; step 2, 95°C for 10 min; step 3, 95°C for 15 s; step 4, 60°C for 1 min; repeat steps 3 and 4 for 40 cycles. All samples were run in triplicate for three independent runs and normalized against an endogenous internal control, human GAPDH. All primer sequences are supplied in table S4.

AP and immunofluorescence staining

For AP staining, iPSCs were fixed with 4% paraformaldehyde and washed with AP staining buffer [100 mM tris (pH 9.5), 50 mM MgCl2, 10 mM NaCl, and 1% Tween 20 in water]. iPSC colonies were incubated with 5-bromo-4-chloro-3-indolyl phosphate and 4-nitro blue tetrazolium chloride (Roche Applied Science) in AP staining buffer. Attached embryoid bodies, iPSC-derived keratinocytes, and sections of 3D skin equivalents and reconstitution assay were fixed with 4% paraformaldehyde. After 1-hour blocking using 10% goat or donkey sera in 0.1% Triton X-100/phosphate-buffered saline (PBS), samples were incubated for 1 hour at room temperature with primary antibodies. After three rinses with PBS, incubation with appropriate secondary antibodies was performed for 1 hour at room temperature. Nuclear staining was performed by VECTASHIELD Mounting Medium containing DAPI (Vector Laboratories). Confocal microscopy (Carl Zeiss LSM 5 EXCITER) was used to visualize and capture immunostained cells with good resolution. All antibodies are supplied in table S5.

Teratoma formation

Colonies of iPSCs were collected by collagenase VI (1 ml/ml) (Invitrogen) treatment and subcutaneously injected into nude mice (Taconic) with Matrigel. Palpable tumors appeared 2 months after injection. Tumors were collected and processed for H&E staining according to standard procedures to check their differentiation capacity into all three germ layers in vivo.

Methylation assay

To identify the methylation status of CpG islands in the NANOG and OCT4 gene promoter, published methods were followed (26, 36). Briefly, a total of 0.4 μg of genomic DNA extracted from various cells was treated with EZ DNA Methylation-Gold Kit (Zymo Research) for bisulfate reaction. The promoter regions of the NANOG and OCT4 genes were amplified by PCR using gene-specific primers (table S4). PCR products were subcloned into pCRII vector (Invitrogen). For defining methylation status of the NANOG and OCT4 promoter regions, 10 clones of each sample were sequenced.

Karyotype analysis

A minimum of 20 metaphases were examined by experienced and certified cytogeneticists in the Microarray and Molecular Cytogenetics shared resource of the Herbert Irving Comprehensive Cancer Center at Columbia University.

High-resolution SOMA

DNA was prepared and hybridized to a CytoScan HD Array Kit (Affymetrix) according to the manufacturer’s instructions. This array contains 2.69 million copy number markers, including about 750,000 SNPs. The resulting data were analyzed with the Affymetrix Chromosome Analysis Suite version 2.0.0 software using human genome build hg19. Analysis was done to reveal changes of any size that contained a minimum of 25 consecutive probes.

Gene expression microarrays

Microarray preprocessing and normalization. For transcriptional profiling of revertant-iPSC FB, revertant-iPSC KC, revertant iPSCs, original revertant keratinocytes, and normal fibroblasts, total RNA was extracted from each sample and reverse-transcribed and amplified using the GeneChip WT PLUS Reagent Kit (WT PLUS Kit) (Affymetrix). Amplified cDNA was biotinylated and then hybridized to HTA 2.0 (Affymetrix), subsequently washed, stained with streptavidin-phycoerythrin, and scanned on an HP GeneArray Scanner (Hewlett-Packard Company). Three technical replicates were prepared for the revertant-iPSC FB cell line, two for the nFB, and one each for the revertant-iPSC KC, revertant iPSCs, and the original revertant keratinocytes, for a total of eight microarrays. Quality control and a gene level robust multiarray average normalization were performed with the Affymetrix Expression Console software (version 1.3).

Microarray differential expression analysis. Differential gene expression analysis was performed with the Affymetrix Transcriptome Analysis Console (TAC) software. The HTA 2.0 array contained 70,523 TCIDs, 67,528 of which were mapped to genes by the TAC software. For each cell line comparison, linear fold change values were calculated using the mean expression levels from the technical replicates. IPA was used to identify and annotate the 24,662 mappable TCIDs. Filtering was performed, and only those Affymetrix TCIDs that were mappable by IPA, that were on Chr 1 to 22, X or Y, and that had both official gene symbols were retained, resulting in 23,459 TCIDs that were used in subsequent downstream analysis and visualization of the data.

Within IPA, a threshold of absolute 2.0 fold change was used to identify genes/TCIDs that were differentially expressed between cell lines. Linear fold change values were calculated using the mean intensity levels from the technical replicates. Duplicate TCIDs corresponding to the same gene symbol from IPA annotation were resolved by taking the maximum fold change value for that gene symbol.

Microarray GO enrichment analysis using Database for Annotation, Visualization and Integrated Discovery. Database for Annotation, Visualization and Integrated Discovery (DAVID) (40) was used to perform GO enrichment analysis. A fold change greater than 5 was used as the threshold for identification of differentially expressed genes from our cell type comparisons. For each comparison, differentially expressed genes were separated into lists that were identified as up and down. These lists were uploaded into DAVID separately to identify GO biological processes whose gene sets were significant enriched in our differentially expressed genes.

iPSC differentiation into keratinocytes

iPSC-derived keratinocytes (iPSC KC) were generated using our previously published protocol with minor modification (26). Briefly, iPSCs (P18 to P30) were directly induced into iPSC KC using either CnT-07 (CELLnTEC) or defined keratinocyte serum-free medium (Invitrogen) supplemented with 1 μM all-trans RA (Sigma) and BMP4 (10 ng/ml) (R&D Systems) on days 0 and 2. Cells were passaged at days 21 to 28, after which time they were switched to CnT-57 medium.

iPSC differentiation into fibroblasts

iPSC-derived fibroblasts (iPSC-FB) were generated using our previously published protocol (41). Briefly, to directly differentiate iPSCs (P21) into fibroblasts, we first generated embryoid bodies in ESM without basic FGF and supplemented with 0.3 mM ascorbic acid (Sigma-Aldrich), transforming growth factor–β2 (10 ng/ml) (R&D systems), and ITS-A Supplement (Invitrogen) on a low-binding dish. For inducing cell outgrowth, embryoid bodies were attached to a gelatin-coated dish and cultured in DMEM (with high glucose) (Invitrogen) supplemented with ascorbic acid and 20% bovine serum for 10 days. Cells that grew out from embryoid bodies were cultured in DMEM supplemented with 10% bovine serum and passaged every week to obtain consistent spindle-shaped cells.

Generation of 3D skin equivalents

Three-dimensional skin equivalents were generated according to the protocol described previously (42). Briefly, type I collagen matrix containing nFB isolated from human foreskin (P4 to P6) was added and polymerized on polyethylene terephthalate membranes (BD Biosciences). After the polymerized matrix was incubated for 5 to 7 days, 106 NHKs or iPSC KC (P2 to P3) were reseeded on the matrix and incubated for 6 days. The culture composite was raised to the air interface and fed by feeding from below to induce epidermal cornification, and 3D skin equivalents were harvested 7 days later.

Skin reconstitution chamber assay

Experiments were performed in compliance with institutional guidelines as approved by the Institutional Animal Care and Use Committee of Columbia University Medical Center.

A sterile silicone chamber (Renner GmbH) was implanted into the back of female SCID mice (Taconic) (n = 2) under anesthesia to prevent inward migration of mouse cells from the adjacent skin. Ten to 15 million cells were suspended in 10 ml of PBS for both the epidermal component (NHK or iPSC KC) and the dermal component (nFB) and mixed together in a 15-ml Falcon tube and centrifuged at 1500 rpm for 5 min. PBS was removed by aspiration, and the resultant cell pellet, composed of epidermal and dermal cells, was deposited into the chamber using a 200-μl pipette. One week after placing the cells, the silicone chamber was removed. Skin was harvested from the chamber sites 4 weeks later to observe reconstituted human skin.

Statistical analysis

Spearman rank correlation coefficients, r, were calculated for each pair of cell types using the mean of the log2(intensity) of the 2739 IPA-mapped TCIDs that were identified as differentially expressed between original revertant keratinocytes and revertant iPSCs using an absolute fold change ≥2 cutoff. Duplicate TCIDs corresponding to the same gene symbol were removed as described above. Correlation coefficients and figures were generated using R (43). Hierarchical clustering of the correlation coefficients was performed with the similarity measure d = 1 – r (Fig. 3D), whereas hierarchical clustering of the top 50/bottom 50 differentially expressed genes was performed with the Euclidean distance calculated from the z scores of each gene.


Fig. S1. Genomic characterization of revertant iPSCs.

Fig. S2. Differentiation analyses of iPSCs.

Table S1. Top and bottom 50 transcripts differentially expressed between revertant iPSCs and original revertant keratinocytes.

Table S2. Top 50 GO terms overrepresented by genes differentially expressed in original revertant keratinocytes compared to revertant-iPSC KC.

Table S3. Top 50 GO terms overrepresented by genes differentially expressed in original revertant keratinocytes compared to revertant iPSCs.

Table S4. Primers used in this study.

Table S5. Antibodies used in this study.


  1. Acknowledgments: We are grateful to the patients for their participation in this study. We thank S. Harel and C. A. Higgins for stimulating discussions. Funding: The Dutch Butterfly Child Foundation (Vlinderkind), the Dr. Ines Mandl Research Fellowship (Columbia University), and grants from DEBRA International (ROOP 2), New York State Office of Science, Technology and Academic Research, NYSTEM (New York State Stem Cell Science), the NIH/National Center for Advancing Translational Sciences Microphysiological Systems Program (grant U18TR000561-01), the Helmsley Stem Cell Starter grants (Columbia University), and the Core facilities of the Skin Disease Research Center at Columbia University (grant P30AR044535 from NIH/National Institute of Arthritis and Musculoskeletal and Skin Diseases). Author contributions: N.U.-A. contributed all experiments. N.U.-A., A.M.G.P., M.I., A.G., M.F.J., and A.M.C. developed the concept and managed the project. N.U.-A., M.I., Z.G., B.L., L.R.R., and A.M.C. generated and characterized the iPSCs. N.U.-A., J.E.C., and A.M.C. developed microarray analyses. N.U.-A. and Z.G. developed 3D skin equivalents and the skin reconstitution chamber assay. N.U.-A. performed PCR and immunohistochemistry. J.E.C. performed statistical analysis. N.U.-A., A.M.G.P., M.I., J.E.C., B.L., A.G., M.F.J., and A.M.C. wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The data discussed in this publication have been deposited in the National Center for Biotechnology Information’s Gene Expression Omnibus (GEO) ( and are accessible through GEO Series accession no. GSE60991.
View Abstract

Navigate This Article