Mechanobiology June 26th - June 2nd 2016

Mechanobiology: June 26th  - June 2nd 2016

Sox2 is dispensable for the reprogramming of melanocytes and melanoma cells into induced pluripotent stem cells
Jochen Utikal, Nimet Maherali, Warakorn Kulalert, Konrad Hochedlinger

Summary

Induced pluripotent stem cells (iPSCs) have been derived at low frequencies from different cell types through ectopic expression of the transcription factors Oct4 and Sox2, combined with either Klf4 and c-Myc or Lin28 and Nanog. In order to generate iPSCs more effectively, it will be crucial to identify somatic cells that are easily accessible and possibly require fewer factors for conversion into iPSCs. Here, we show that both human and mouse melanocytes give rise to iPSCs at higher efficiencies than fibroblasts. Moreover, we demonstrate that a mouse malignant melanoma cell line, which has previously been reprogrammed into embryonic stem cells by nuclear transfer, remains equally amenable to reprogramming into iPSCs by these transcription factors. In contrast to skin fibroblasts, melanocytes and melanoma cells did not require ectopic Sox2 expression for conversion into iPSCs. iPSC lines from melanocytic cells expressed pluripotency markers, formed teratomas and contributed to viable chimeric mice with germ line transmission. Our results identify skin melanocytes as an alternative source for deriving patient-specific iPSCs at increased efficiency and with fewer genetic elements. In addition, our results suggest that cancer cells remain susceptible to transcription factor-mediated reprogramming, which should facilitate the study of epigenetic changes in human cancer.

Introduction

Induced pluripotent stem cells (iPSCs) have been generated from the fibroblasts of several species, including mouse, rat, rhesus monkey and human, following ectopic expression of the transcription factors Oct4 and Sox2, combined with either Klf4 and c-Myc or Lin28 and Nanog (Hockemeyer et al., 2008; Liao et al., 2008; Liu et al., 2008; Maherali et al., 2007; Okita et al., 2007; Park et al., 2008b; Takahashi and Yamanaka, 2006; Takahashi et al., 2007; Yu et al., 2007; Wernig et al., 2007; Hochedlinger and Plath, 2009; Li et al., 2008).

iPSCs acquire most of the features of embryonic stem cells (ESCs) including immortal growth and pluripotency, as measured by their ability to differentiate into multiple cell types in teratomas and their contribution to germline-competent chimeras in mice (Maherali et al., 2007). Patient-derived iPSCs might be an ideal source for studying complex diseases in vitro and potentially for treating disorders in the clinic. Indeed, iPSCs have been shown to alleviate the symptoms in mouse models of Parkinson's disease and sickle cell anemia (Hanna et al., 2007; Wernig et al., 2008a). Moreover, iPSCs have already been derived from human individuals suffering from diseases such as Parkinson's disease, diabetes or amyotrophic lateral sclerosis (ALS), thus enabling disease modeling and drug screening approaches (Park et al., 2008a; Dimos et al., 2008).

Since the initial report of reprogramming fibroblasts into iPSCs, several additional cell types, including stomach cells (Aoi et al., 2008), liver cells (Aoi et al., 2008; Stadtfeld et al., 2008b), pancreatic β cells (Stadtfeld et al., 2008c), lymphocytes (Hanna et al., 2008) and neural progenitor cells (NPCs) (Eminli et al., 2008; Kim et al., 2008; Silva et al., 2008), have been successfully converted into iPSCs in mouse. By contrast, the derivation of human iPSCs is still limited to fibroblasts and keratinocytes (Park et al., 2008b; Takahashi et al., 2007; Yu et al., 2007; Aasen et al., 2008; Maherali et al., 2008; Lowry et al., 2008). Interestingly, NPCs do not require ectopic Sox2 expression for reprogramming (Eminli et al., 2008; Kim et al., 2008; Silva et al., 2008) due to their high endogenous Sox2 levels (Ellis et al., 2004). Considering future clinical applications of the iPSC technology, however, NPCs are an undesirable cell type due to their difficult accessibility. We therefore evaluated whether melanocytes could be reprogrammed into iPSCs; melanocytes are, like NPCs, of neuroectodermal origin and hence might require fewer factors for conversion into iPSCs, thus potentially facilitating replacement of factors with chemicals or non-integrating approaches (Stadtfeld et al., 2008; Okita et al., 2008). Moreover, melanocytes are easily accessible from the skin and can be marked genetically using Cre-lox technology in mice to trace their origin.

In addition to using nuclear reprogramming as a means of deriving patient-specific iPSCs, induced pluripotency could be a powerful tool for distinguishing between epigenetic and genetic alterations occurring during development and tumorigenesis. We have previously shown that nuclear transfer can reprogram a malignant melanoma genome into a pluripotent state, demonstrating that the epigenetic state of this particular tumor cell was reversible upon exposure of its nucleus to oocyte cytoplasm (Hochedlinger et al., 2004). In order to assess whether transcription-factor-mediated reprogramming is equally suitable for reprogramming tumor cells into a pluripotent state, in this study we attempted to convert the previously used melanoma cell line (Chin et al., 1999) into iPSCs.

Results

Reprogramming of genetically marked primary mouse melanocytes into iPSCs

To test whether cells of the melanocytic lineage can be reprogrammed into iPSCs, we isolated primary melanocytes from the epidermis of neonatal mice using previously published protocols (Yang et al., 2001). In order to rule out the presence of contaminating fibroblasts in our cultures, we used animals that expressed Cre recombinase from promoter elements of the Wnt1 gene (Danielian et al., 1998), as well as a conditional (enhanced yellow fluorescence protein) EYFP reporter gene driven off the ROSA26 promoter (Srinivas et al., 2001), leading to specific and irreversible labeling of neural-crest-derived cells, including melanocytes. After 5 days in culture in melanocyte growth medium, EYFP-positive cells were sorted by fluorescence activated cell sorting (FACS) to separate them from the remaining keratinocytes, and were cultured in melanocyte growth medium. Resultant melanocyte cultures were homogeneously positive for EYFP and contained melanosomes, which are melanin-containing organelles characteristic for melanocytes.

Fig. 1.

Generation of iPSCs from genetically marked primary mouse melanocytes in the presence or absence of viral Sox2. (A) Experimental scheme for generating iPSCs from primary melanocytes. (B) Derivation of iPSCs from primary melanocytes of the lineage tracing mouse model Wnt1-Cre/ROSA26R-EYFP; shown are primary melanocytes (left: brightfield and fluorescent image) and one picked iPSC colony grown on feeder cells (right: brightfield and fluorescent image) after doxycycline withdrawal. Note that high melanin content of melanocytes results in quenching of the EYFP signal (lower left image). (C) EYFP-positive iPSC colonies develop only from EYFP-positive melanocytes. The reprogramming of EYFP-negative tail fibroblasts from Wnt1-Cre/ROSA26R-EYFP mice does not give rise to EYFP-positive iPSC colonies. (D) Brightfield and fluorescent image of primary melanocyte-derived iPSCs stained for the pluripotency marker Nanog. (E) Primary melanocytes (PM) contain the pigment melanin whereas iPSCs derived from melanocytes lost melanin. (F) Quantitative RT-PCR analysis of marker gene expression in primary melanocytes (PM), two iPSC lines derived with four factors (4F) or three factors (3F) from primary melanocytes, and V6.5 control ESCs. Values represent means ± s.e.m. (G) Bisulfite sequencing of the Oct4 and Nanog promoter regions in primary melanocytes, iPSCs produced with three or four factors from primary melanocytes and ESCs. White circles represent unmethylated CpGs; black circles denote methylated CpGs.

Primary melanocytes were infected on days 1, 2 and 3 with doxycycline-inducible lentiviral vectors expressing Oct4, Sox2, c-Myc and Klf4 (Stadtfeld et al., 2008a) (Fig. 1A), giving rise to ESC-like colonies at day 9. These colonies continued to grow after doxycycline withdrawal on day 19, indicating that colony growth was independent of transgene expression. At day 24, ESC-like colonies were picked and expanded into stable iPSC lines (Fig. 1B). Interestingly, the efficiency of reprogramming melanocytes was more than three times higher than that of fibroblasts (0.19% versus 0.056%, see also Table 1).

View this table:
Table 1.

Efficiency of reprogramming mouse and human melanocytes into iPS cells

To rule out the possibility that the Wnt1-Cre transgene became spuriously activated during the reprogramming process, we reprogrammed tail-tip fibroblasts from Wnt1-Cre/ROSA26-EYFP mice. No EYFP colonies were observed (0/45), confirming specificity of the lineage tracing system (Fig. 1C).

iPSCs derived from melanocytes expressed the pluripotency markers Nanog and Oct4 (Fig. 1D; data not shown), lost expression of the melanocyte markers tyrosinase and dopachrome tautomerase, and attenuated melanin production (Fig. 1E,F; and data not shown). Demethylation of the Oct4 promoter region in iPSCs, which is heavily methylated in primary melanocytes, demonstrated faithful epigenetic reprogramming of iPSCs (Fig. 1G). In contrast to other mouse cell types that have been reprogrammed previously, such as fibroblasts, NPCs and hepatic cells (Maherali et al., 2007; Stadtfeld et al., 2008c; Stadtfeld et al., 2008b; Eminli et al., 2008), primary mouse melanocytes were devoid of methylation at the Nanog promoter. Melanocyte-derived iPSCs differentiated in vitro into embryoid bodies (data not shown) and into mesodermal, ectodermal and endodermal derivatives in the context of teratomas (Fig. 2A). This shows that melanocytes remain amenable to reprogramming into pluripotent cells.

Sox2 is dispensable for the reprogramming of murine melanocytes into iPSCs

It has been previously shown that cells with endogenous Sox2 levels can be reprogrammed in the absence of ectopic Sox2 expression (Eminli et al., 2008; Kim et al., 2008; Silva et al., 2008). Given that melanocytes, like NPCs, are of neuroectodermal origin, we assessed Sox2 expression and indeed detected Sox2 transcripts, albeit at lower levels than in NPCs (Fig. 1F; supplementary material Fig. S1A,B). Interestingly, the expression of Sox2 was higher in low passage (passage 1) melanocyte cultures compared with high passage (passage 3+) cells. We therefore reasoned that Oct4, Klf4 and c-Myc alone might be sufficient to induce pluripotency in primary melanocytes. Indeed, Wnt1-EYFP-positive melanocytes that were infected with lentiviruses expressing Oct4, c-Myc and Klf4 gave rise to ESC-like cells that grew into stable iPSC lines upon discontinuation of doxycycline (data not shown), indicating that ectopic Sox2 expression is dispensable for reprogramming melanocytes into iPSCs (supplementary material Fig. 1C). In agreement with the progressive decline in Sox2 expression during serial passaging, we were able to obtain iPSCs only at low passages from Wnt1-EYFP-positive melanocytes using three factors. At higher passages, the cells had to be supplemented by viral Sox2 expression to generate iPSCs.

The efficiency of reprogramming melanocytes in the absence of ectopic Sox2 expression was lower than that with all four factors (0.19% with four factors versus 0.03% with three factors) (Table 1). Three-factor iPSCs expressed pluripotency markers, as demonstrated by quantitative PCR (qPCR) (Fig. 1F). They showed demethylation of the Oct4 and Nanog promoter regions (Fig. 1G), and could differentiate both into embryoid bodies in vitro (data not shown) and into mesodermal, ectodermal and endodermal derivatives in teratomas (Fig. 2A). Further, melanocyte-derived iPSCs gave rise to chimeric mice competent of germline transmission (Fig. 2B-D).

Because fibroblasts can be reprogrammed in the absence of viral c-Myc expression, we infected primary melanocytes with inducible lentiviruses expressing Klf4, Oct4 and Sox2 alone to assess their potential to produce iPSCs. We observed ESC-like colonies after 3 weeks that could be expanded into stable, doxycycline-independent iPSC lines. The efficiency of iPSC formation was 0.02%, which is approximately tenfold lower than with reprogramming melanocytes in the presence of c-Myc (see Table 1) and similar to observations in fibroblasts (Nakagawa et al., 2008; Wernig et al., 2008b). Three-factor iPSCs (without c-Myc) also differentiated in vitro into embryoid bodies and into mesodermal, ectodermal and endodermal derivatives in the context of teratomas (data not shown). Because Sox2 and c-Myc were individually not required for reprogramming melanocytes into iPSCs, we attempted to derive iPSCs with Oct4 and Klf4 expressing doxycycline-inducible viruses alone. Although we observed morphological changes in cells after doxycycline addition, we were unable to derive stable doxycycline-independent iPSC lines, despite repeated attempts.

Fig. 2.

Differentiation potential of iPSCs obtained from primary mouse melanocytes with three (3F) or four (4F) reprogramming factors. (A) Hematoxylin and eosin stainings of teratomas derived from iPSCs show differentiation into cell types from all three germ layers [endoderm: epithelial structures (left images); ectoderm: keratinized epithelium (center images) and mesoderm: muscle fibers (right images)]. (B) Viable newborn chimeras from iPSCs derived from primary melanocytes. Chimeric pup (right) and non-chimeric littermate (left) are shown under regular light (left image) and UV light (right image). (C) Adult chimera derived from three-factor female iPSCs shows obvious coat color chimerism. (D) Germline contribution of three-factor melanocyte iPSCs. Images of pups derived from matings between BDF1 wild-type males and female iPSC chimeras. The presence of agouti coat color indicates germline transmission (arrow).

Reprogramming of human melanocytes into iPSCs

We next tested whether human melanocytes are equally amenable to reprogramming into iPSCs as mouse melanocytes. To this end, we infected primary human melanocytes (obtained from Promocell, Heidelberg, Germany) with lentiviruses containing a reverse tetracycline transactivator (rtTA) and the four reprogramming factors (Oct4, Sox2, c-Myc and Klf4) as previously described (Fig. 3A) (Maherali et al., 2008). After infection of melanocytes with viral supernatant for 3 consecutive days, cells were plated onto mouse embryonic fibroblasts (MEFs) in melanocyte growth conditions and induced with doxycycline. Medium changes with 50% melanocyte and 50% human ESC medium were performed every other day until human iPSCs developed. Human non-ESC-like colonies emerged at day 5 and contained only viral integrations for Oct4 and c-Myc (data not shown), as previously observed during the reprogramming of human fibroblasts (Takahashi et al., 2007). We were unable to expand these colonies after doxycycline withdrawal. After 10 days, however, colonies resembling human ESCs were observed. These colonies showed well-defined phase-bright borders with surrounding sheets of feeder cells and high nucleus-to-cytoplasm ratios with prominent nucleoli (Fig. 3C). The colonies all expressed the pluripotency marker Oct4 (Fig. 3D) and could be expanded in the absence of doxycycline.

Fig. 3.

Generation and characterization of human iPSCs derived from primary human melanocytes. (A) Experimental scheme for the generation of human iPSCs. Primary human melanocytes were infected with separate lentiviral vectors expressing a constitutively active rtTA and doxycycline-inducible copies of Oct4, Klf4, c-Myc and, optionally, Sox2. After infection, cells were seeded on mouse embryonic fibroblasts (MEFs), and doxycycline was applied for 9 days. Human iPSC clones were picked on the basis of human ESC-like morphology and doxycycline-independent growth. (B) Quantitative PCR expression analysis of Sox2 in primary human melanocytes (PM), human iPSCs derived from primary human melanocytes produced with (4F) or without (3F) viral Sox2, fibroblasts differentiated from human iPSCs, and resultant secondary human iPSCs (sec. hiPSCs). (C) Morphology of human iPSCs derived from primary human melanocytes in the presence (4F) or absence (3F) of viral Sox2. (D) Doxycycline-independent melanocyte-derived human iPSCs express Oct4 protein. (E) Bisulfite sequencing of the Nanog and Oct4 promoter regions in primary melanocytes, human iPSCs derived from primary melanocytes by three (3F) or four (4F) factors, and WA09 human ESCs. Promoter regions containing differentially methylated CpGs are shown. White circles represent unmethylated CpGs; black circles denote methylated CpGs. (F) Hematoxylin and eosin staining of teratomas generated from primary melanocyte-derived human iPSCs. Differentiated structures from all three germ layers were present [endoderm: epithelial structures (left images); mesoderm: cartilage and muscle fibers (center images) and ectoderm: neural tissue and keratinized epithelium (right images)].

The reprogramming efficiency was calculated as 0.05% (Table 1). All human iPSC lines lacked expression of the melanin-synthesizing enzymes tyrosinase and dopachrome tautomerase, in contrast to melanocytes that expressed these enzymes strongly (Fig. 4B). In addition, all iPSC lines expressed pluripotency genes from the endogenous loci and lacked expression of the viral transgenes (Fig. 4B and supplementary material Fig. S2B). In accordance with these observations, the Nanog and Oct4 promoter regions in iPSCs were demethylated to a similar extent as in human ESCs, whereas melanocytes showed highly methylated promoter regions (Fig. 3E), demonstrating epigenetic reprogramming of key pluripotency genes in human iPSCs.

To test whether melanocyte-derived human iPSCs could form teratomas in vivo, around 5×104 human iPSCs were injected either under one kidney or testis capsule of immunodeficient SCID mice. Teratomas developed after 7-10 weeks and contained structures representative of the three embryonic germ layers, including cartilage, skeletal muscle, keratinized epithelium and mucous epithelium (Fig. 3F). This shows that human iPSCs derived from human primary melanocytes resemble human ESCs and fulfil the criteria for pluripotency (Maherali and Hochedlinger, 2008).

Ectopic Sox2 expression is dispensable for the reprogramming of human melanocytes

Because mouse melanocytes were amenable to reprogramming into iPSCs in the absence of exogenous Sox2 expression and because human melanocyte cultures were also found to express Sox2 (Fig. 3B), we reasoned that Oct4, Klf4 and c-Myc alone might be sufficient to reprogram human melanocytes into iPSCs. Indeed, human melanocytes infected with viruses expressing Oct4, c-Myc and Klf4 alone gave rise to human ESC-like cells and grew into stable human iPSC lines upon discontinuation of doxycycline (Fig. 3C), indicating that human melanocytes can be reprogrammed in the absence of viral Sox2 expression. The efficiency of reprogramming human melanocytes in the absence of Sox2 virus was lower than that using all four factors (0.05% with four factors versus 0.01% with three factors; see Table 1). iPSCs derived in the absence of Sox2 virus expressed pluripotency markers (Fig. 3D and supplementary material Fig. S2), showed demethylation of the Oct4 and Nanog promoter regions (Fig. 3E), differentiated into embryoid bodies (data not shown) in vitro and into mesodermal, ectodermal and endodermal structures in teratomas (Fig. 3F).

To test whether human melanocytes can be reprogrammed into iPSCs in the absence of c-Myc, we infected melanocytes with viral vectors expressing Oct4, Sox2 and Klf4 and detected a delayed onset of colony formation after 4 weeks and obtained doxycycline-independent, stable iPSC lines only after 5 weeks. The efficiency of iPSC derivation without c-Myc was ∼fivefold lower than with c-Myc (0.01% versus 0.05%; Table 1). iPSCs that were derived in the absence of ectopic c-Myc expression also activated pluripotency markers and differentiated in vitro into embryoid bodies (data not shown).

We next asked whether Oct4 and Klf4 alone might be sufficient for reprogramming human melanocytes into iPSCs. Consistent with observations in mouse, infection of melanocytes with viral vectors expressing Oct4 and Klf4 induced morphological changes in cultures, but no stable iPSC lines were recovered after withdrawal of doxycycline. This suggested that Oct4 and Klf4 together are insufficient to induce pluripotency in human melanocytes. Treatment of human fibroblasts with the histone deacetylase inhibitor valproic acid has recently been shown to replace ectopic Klf4 and c-Myc expression during the reprogramming of human fibroblasts into iPSCs (Huangfu et al., 2008; Maherali and Hochedlinger, 2008). However, when applied to human melanocytes infected with Oct4 and Sox2 alone, we did not observe iPSC formation (data not shown).

Secondary cells derived from primary human iPSCs

Primary iPSCs produced with doxycycline-inducible lentiviruses can be differentiated in vitro into fibroblast-like cells, generating so-called `secondary cells', which express the four factors more homogeneously and thus reprogram into secondary iPSCs at higher efficiency upon re-exposure to doxycycline (Hockemeyer et al., 2008; Maherali et al., 2008; Wernig et al., 2008c). To test whether human melanocyte iPSCs can be converted into secondary cells, we differentiated several melanocyte-derived human iPSC lines in vitro into fibroblast-like cells (Fig. 4A). Promoter methylation analysis of the Oct4 and Nanog genes revealed a methylation pattern akin to fibroblasts (Fig. 4A), confirming that the cells had acquired a somatic epigenetic pattern. Further, quantitative RT-PCR analysis confirmed the lack of pluripotency gene expression in this cell population (Fig. 4B).

For the generation of secondary human iPSCs, fibroblasts were plated on mouse embryonal feeder layers and cultured under human ESC conditions in the presence of doxycycline. Fibroblast-derived secondary human iPSCs indeed gave rise to iPSCs in the presence of doxycycline, whereas no colonies developed in the absence of doxycycline (data not shown). The frequency of conversion into human iPSCs was 0.24% and thus nearly five times higher than reprogramming using direct viral infection. However, iPSCs took longer to develop than melanocytes (21 versus 10 days). Moreover, we were only able to establish secondary iPSCs from primary iPSC clones produced with four factors, not from those produced with three factors lacking Sox2, consistent with the notion that fibroblasts require ectopic Sox2 expression for reprogramming into iPSCs. Secondary human iPSCs were molecularly and functionally similar to primary human iPSCs. They showed a similar methylation pattern of the Oct4 and Nanog promoter regions to ESCs (Fig. 4A), expressed the pluripotency markers Oct4 and Nanog (Fig. 4B), and formed teratomas after injection under the kidney capsule of SCID mice (data not shown).

Reprogramming of a malignant melanoma cell line into iPSCs in the absence of exogenous Sox2 expression

We wanted to assess whether the malignant transformation of melanocytes would make them refractory to reprogramming by transcription factors. To test this, we took advantage of the malignant R545 melanoma cell line (Chin et al., 1999), which we have previously shown to support the development of cloned embryos and the derivation of pluripotent ESCs (Hochedlinger et al., 2004). R545 cells carry a deletion of the ink4a/arf locus, and conditionally express H-Ras from a melanocyte-specific promoter. Moreover, these cells were found to be trisomic for chromosomes 8 and 11 (Hochedlinger et al., 2004).

Because R545 cells were of melanocyte origin, we first attempted infecting them with retroviruses expressing Oct4 and Klf4 alone, which also conferred resistance to puromycin and blasticidine, respectively (Fig. 5A). In the presence of both drugs, however, no ESC-like colonies emerged despite the appearance of double-resistant cells (data not shown). We therefore superinfected resistant cells with doxycycline-dependent lentiviruses expressing Oct4, Klf4 and c-Myc (Fig. 5A). ESC-like colonies appeared after 2 weeks in culture and were picked for further analysis. Melanoma-derived iPSC lines showed, like melanocyte-derived iPSCs, silencing of viral gene expression (supplementary material Fig. S3C) and demethylation of the Oct4 and Nanog promoters. They also formed teratomas and gave rise to chimeras after labeling with a lentivirus constitutively expressing GFP and subsequent blastocyst injection (Fig. 5C-F). At five months of age and in the absence of doxycycline treatment, these mice were free of obvious tumors. Virus-specific PCR analysis confirmed that iPSCs were devoid of retroviral and lentiviral Sox2 expression (supplementary material Fig. S3A,B), demonstrating that melanoma cells, like primary melanocytes, do not require ectopic Sox2 expression for conversion into iPSCs. Together, these results indicate that the ectopic expression of transcription factors can reverse the epigenetic state of primary melanocytes and certain malignant cells of the melanocyte lineage.

Fig. 4.

Generation of secondary human iPSCs. (A) Experimental scheme depicting the generation of secondary human iPSCs. Primary human iPSCs derived from melanocytes were differentiated in vitro as embryoid bodies for 7 days, then plated under adherent conditions. Fibroblast-like colonies were picked and expanded for at least two further passages prior to re-inducing with doxycycline, resulting in secondary human iPSCs. Shown below are the methylation states of the Nanog and Oct4 promoter regions in primary human melanocytes, human iPSCs derived from melanocytes with four factors, fibroblasts differentiated from human iPSCs and secondary human iPSCs derived from fibroblasts. Promoter regions containing differentially methylated CpGs are shown. White circles represent unmethylated CpGs; black circles denote methylated CpGs. (B) Quantitative PCR analysis of Nanog and Oct4 pluripotency marker expression as well as of the melanin-synthesizing enzymes tyrosinase (TYR) and dopachrome tautomerase (DCT) in primary human melanocytes (PM), human iPSCs derived from primary human melanocytes generated in the presence (4F) or absence (3F) of viral Sox2, differentiated fibroblasts produced from four-factor human iPSCs and iPSCs obtained from secondary fibroblast-like cells (sec. hiPSCs). Values represent means ± s.e.m.

Discussion

Our data allow four major conclusions: First, we demonstrate that mouse and human melanocytes can be reprogrammed into iPSCs by the same combination of transcription factors previously used to reprogram fibroblasts and keratinocytes. These melanocyte-derived iPSCs expressed markers similar to those expressed by ESCs and gave rise to teratomas. In addition, mouse iPSCs contributed to adult chimeras that showed germline transmission and hence fulfill all criteria of pluripotency.

Second, mouse melanocytes undergo reprogramming at roughly fourfold higher efficiency than mouse fibroblasts (∼0.19% versus ∼0.05%). Interestingly, primary mouse melanocytes were devoid of methylation at the Nanog promoter region, which is in contrast to human melanocytes and other cultured primary mouse cell types that have been recently reprogrammed (Eminli et al., 2008; Maherali et al., 2007; Stadtfeld et al., 2008c) and might explain why mouse melanocytes are reprogrammed more efficiently. Consistent with this notion is the previous observation that hypomethylated fibroblasts, which lack methylation at the Nanog promoter, are more efficiently reprogrammed to pluripotency by nuclear transfer (Blelloch et al., 2006). It remains unclear, however, whether demethylation of the Nanog promoter in murine melanocytes was pre-existing in vivo or arose upon explantation of cells in vitro. In addition to mouse melanocytes, human melanocytes gave rise to iPSCs faster than did fibroblasts, showing colony formation after 10 days of factor expression. These kinetics are similar to keratinocytes infected with four factors, which yield iPSCs faster and at higher efficiency than fibroblasts (Aasen et al., 2008; Maherali et al., 2008). Keratinocytes and melanocytes are both of ectodermal origin whereas fibroblasts are of mesodermal origin. Because ESCs are embryonic ectodermal cells, it might be faster and more efficient to convert an ectodermal rather than a mesodermal somatic cell type into iPSCs.

Third, we provide evidence that an aneuploid melanoma cell line remains amenable to reprogramming into iPSCs that supported the development of chimeras, suggesting that certain cancer cells are not refractory to reprogramming by transcription factors. This observation should be useful for studying the relative contribution of reversible epigenetic and irreversible genetic changes to cancer. In contrast to nuclear transfer, however, iPSC formation does not require the generation of cloned embryos, which might be incompatible with some genetically aberrant cancer cells (Hochedlinger et al., 2004). Moreover, it should be possible to reprogram human cancer cells into iPSCs to assess their epigenetic state.

Fourth, we show that fewer factors are required for the generation of iPSCs from mouse and human melanocytes as well as from melanoma cells than previously used for reprogramming mouse and human fibroblasts and keratinocytes. To date, melanocytes and NPCs are the only cell types that can be reprogrammed into pluripotent cells in the absence of ectopic Sox2 expression. iPSCs produced with three factors were molecularly and functionally similar to iPSCs produced with four factors, including their ability to generate high-degree chimeric mice competent of germline transmission.

Fig. 5.

Reprogramming of a malignant melanoma cell into iPSCs. (A) Experimental outline for generating iPSCs from R545 melanoma cells. (B) Shown are R545 cells (left image) and R545-derived iPSC colonies on feeder cells at day 50 (right image). (C) Brightfield and fluorescent image of R545-derived iPSCs stained for the pluripotency marker Nanog. (D) Bisulfite sequencing of the Oct4 and Nanog promotor regions in R545 cells, R545 derived iPSCs and ESCs. White circles represent unmethylated CpGs; black circles denote methylated CpGs. (E) Hematoxylin and eosin staining of a teratoma derived from a R545-iPSC shows differentiation into cell types from all three germ layers: keratinized epithelium (left image), epithelium and glandular structures (center image) and cartilage (right image). (F) Viable newborn chimera from iPSCs derived from R545 cells. Chimeric pup (lower image) and non-chimeric littermate (upper image) are shown under UV light (left image). Adult chimera derived from three-factor R545-iPSCs shows coat color chimerism (right image).

The efficiency of reprogramming mouse and human primary melanocytes in the absence of ectopic Sox2 expression is lower than that using all four factors. This is in contrast to NPCs, whose reprogramming efficiency with four factors was in our hands lower than that of three factors in the absence of Sox2 (Eminli et al., 2008). This suggests that cells with abundant levels of endogenous Sox2, such as NPCs, are more sensitive to ectopic Sox2 expression, resulting in toxicicity or the induction of differentiation as seen previously for ESCs (Kopp et al., 2008). In mouse and human early passage melanocyte cultures, however, which show low total levels of Sox2 expression, forced expression of viral Sox2 does not appear to have an adverse effect on reprogramming, but rather enhances efficiency.

We were able to detect higher Sox2 expression levels at early passages of mouse and human melanocyte cultures than at late passages. In agreement, we show that primary melanocytes can be reprogrammed in the absence of exogenous Sox2 expression only at lower passages. This suggests that cultured cells might change qualitatively over time or as they reach senescence, thus resulting in a gradual loss of Sox2 expression. Alternatively, there may be a rare Sox2-positive subpopulation in the culture, possibly progenitors, that is lost during passaging through differentiation. We also cannot rule out the possibility that other Sox family members expressed in melanocytes functionally replace the role of Sox2 during reprogramming.

Our findings imply that melanocyte cultures that express one of the reprogramming factors endogenously are more amenable to reprogramming than fibroblasts or keratinocytes and thus might be an appropriate source of cells for attempts to replace viral gene delivery systems with transient expression approaches such as plasmids (Okita et al., 2008), adenoviruses (Stadtfeld et al., 2008b) or small compounds (Huangfu et al., 2008; Shi et al., 2008). Lastly, melanocytes can be easily obtained by skin punch biopsies, which make them an accessible cell type for clinical applications.

Materials and Methods

Lentiviral production

The generation and structure of replication-defective doxycycline-inducible lentiviral vectors, a lentiviral vector constitutively expressing rtTA, and retroviral vectors have been described in detail elsewhere (Stadtfeld et al., 2008a). To produce infectious lentiviral particles, 293T cells cultured on 10-cm dishes were transfected with the LV-tetO vectors (11 μg) together with the packaging plasmids VSV-G (5.5 μg) and Δ8.9 (8.25 μg) using Fugene (Roche). Viral supernatants were harvested on three consecutive days starting 24 hours after transfection, yielding a total of ∼30 ml of supernatant per viral vector. Viral supernatant was concentrated approximately 100-fold by ultracentrifugation at 50,000 g for 1.5 hours at 4°C and resuspended in 300 μl phosphate-buffered saline (PBS), and stored at –80°C. Infections were carried out in 1 ml medium using 5 μl of each viral concentrate per 35-mm plate. Comparisons of different cell types were performed with the same batch and concentration of viral particles.

Mouse cell culture and generation of iPSCs

Newborn pups of the lineage tracing mouse model Wnt1-Cre × Rosa26R-EYFP (mixed background of C57Bl/6 and 129SvJae) (Stock no. 003829, The Jackson Laboratory, Bar Harbor, ME) were sacrificed by CO2 asphyxiation. The mice were sterilized by soaking several times in 1:200 diluted Wescodyne Solution (Steris Corporation) and washed three times in 1× PBS containing penicillin and streptomycin. The skin was carefully peeled off and placed dermis side down in a Petri dish with 5 ml of 0.2% trypsin in Hank's buffered salt solution (HBSS) containing penicillin and streptomycin. The Petri dish was placed in the refrigerator at 4°C overnight. The next day, the trypsinized skin parts were carefully pulled away from the thin sheet of epidermis leaving the dermis. The epidermis sheets were rinsed in PBS and cut into small pieces. The tissue pieces were transferred into a 75-ml flask with 6 ml of melanocyte growth medium (Cascade Biologicals), and individual cells were released by pipetting up and down (Yang et al., 2001). After 3-4 days of culture, cells were treated with 0.25% trypsin-EDTA for 1-2 minutes and EYFP-positive cells were isolated by FACS and further cultured in melanocyte growth medium.

Lentiviral vector infections were carried out in mouse melanocytes in six-well plates at a density of 100,000 cells/well on 3 subsequent days. Medium changes were performed 12-24 hours after infection. One day after the last infection, ESC medium containing 1ug/ml doxycycline was added. Fresh ESC medium with doxycycline was added every other day until iPSC colonies developed. Five days later, cell culture conditions were switched to ESC medium without doxycycline. iPSC colonies were picked into 96-well plates containing PBS without magnesium and calcium using a 10-ul pipette. Trypsin was added to each well and incubated for 5 minutes. Single-cell suspensions were transferred into 24-well dishes containing MEFs. Picked iPSCs were grown on MEFs under ESC conditions. Prior to blastocyst injections, iPSCs were marked with a FUGW lentiviral vector constitutively expressing GFP.

The infection efficiency of primary melanocytes and tail-tip fibroblasts was calculated after three subsequent infections with tetO-GFP lentiviral vector in the presence of the rtTA-expressing lentiviral vector (24% and 87%, respectively). The fraction of cells receiving all viral vectors in the various combinations was determined by multiplying the infection rate of GFP-positive cells, as described previously (Stadfeld et al., 2008a). The final percentage of efficiency was determined by dividing the number of GFP positive colonies by the fraction of cells calculated to express either three or four viral transgenes.

R545 mouse melanoma cells were cultured in RPMI supplemented with 10% FCS as recently reported (Hochedlinger et al., 2004) and viral infections were performed as described in detail elsewhere (Stadtfeld et al., 2008a).

Human cell culture and generation of iPSCs

Human primary melanocytes were obtained from Promocell, Heidelberg, Germany (NHEM.f-c M2, Lot 6022001 and Lot 7062002.1) and cultured in human melanocyte growth medium (M2, Promocell). Lentiviral vector infections were carried out with human melanocytes in six-well plates at a density of 100,000 cells/well on 3 subsequent days. The infection efficiency of primary human melanocytes after three subsequent infections with tetO-GFP lentiviral vector in the presence of the rtTA-expressing lentiviral vector was 30.1%. Medium changes were performed 12-24 hours after infection. One day after the last infection, cells were transferred on MEFs. Media containing 50% melanocyte M2 medium and 50% human ESC medium containing 0.5 ug/ml doxycycline was added. Medium changes were performed every other day in the presence of 0.5 ug/ml doxycycline until iPSC colonies developed. Upon appearance of human ESC-like colonies, medium was switched to human ESC culture conditions as previously described (Cowan et al., 2004).

Immunofluorescence

Mouse and human iPSCs were cultured on MEFs that were seeded the day before on pretreated coverslips. Cells were fixed with 4% paraformaldehyde, and permeabilized with 0.5% Triton X-100. The cells were then stained with primary antibodies against mOct4 (sc-8628, Santa Cruz Biotechnology) and mNanog (ab21603, Abcam). Respective secondary antibodies were conjugated to Alexa Fluor 546 (Invitrogen). Nuclei were counterstained with DAPI (Invitrogen). Cells were imaged with a Leica DMI4000B inverted fluorescence microscope equipped with a Leica DFC350FX camera. Images were processed and analyzed with Adobe Photoshop software.

Quantitative PCR

RNA was isolated from cells with the TRIzol reagent (Invitrogen). For melanocytes, an additional phenol-chloroform step was performed before RNA clean up with the RNeasy Minikit (Qiagen) to remove melanin. cDNA was produced with the Superscript III First-Strand synthesis system (Invitrogen) using oligo-dT primers. Real-time quantitative PCR (qPCR) reactions were set up in triplicates with the Brilliant II SYBR Green qPCR Master Mix (Stratagene) and run on a Mx3000P qPCR System (Stratagene). Primer sequences are listed in supplementary material Table S1.

Bisulfite sequencing

Bisulfite treatment of DNA was performed with the EpiTect Bisulfite Kit (Qiagen) according to manufacturer's instructions. Oct4 and Nanog promoter sequences were as previously described for mouse (Eminli et al., 2008) and human (Maherali et al., 2008). Amplified products were purified using gel filtration columns, then cloned into the pCR4-TOPO vector (Invitrogen) and sequenced with M13 forward and reverse primers. Analysis shows the promoter regions containing the methylated and unmethylated dinucleotide CpG.

Generation of mouse teratomas and chimeras

For teratoma induction, 2×106 cells of each iPSC line were injected subcutaneously into the dorsal flank of isoflurane-anesthetized SCID mice. Teratomas were recovered 3-5 weeks post-injection, fixed overnight in 10% formalin, paraffin embedded and processed with hematoxylin and eosin.

For chimera production, female BDF1 mice were superovulated with PMS (pregnant mare serum) and hCG (human chorion gonadotropin) and mated to BDF1 stud males. Zygotes were isolated from plugged females 24 hours after hCG injection. After 3 days of in vitro culture in potassium simplex optimized medium (KSOM), blastocysts were injected with iPSCs and transferred into d2.5 pseudopregnant recipient females. C-sections were performed 17 days later and the pups fostered with lactating Swiss females.

Differentiation of human iPSCs

For in vitro differentiation, human iPSC colonies were picked and placed in suspension culture with fibroblast growth media. After 10 days, embryoid bodies were plated under adherent conditions on gelatin-coated plates and further cultured with fibroblast growth media. For teratoma formation around 50,000 human iPSCs were pelleted and injected as a suspension into SCID mice underneath one kidney or both testis capsules of anaesthetized mice. Tumors developed after 8-12 weeks and were processed for histological analysis.

Footnotes

  • Supplementary material available online at http://jcs.biologists.org/cgi/content/full/122/19/3502/DC1

  • We thank Matthias Stadtfeld for critical review of the manuscript. We also thank Laura Prickett, Kathryn E. Folz-Donahue for help with FACS analysis and Patricia Follett for performing blastocyst injections. J.U. was supported by the Dr Mildred Scheel Foundation for Cancer Research. N.M. was supported by a graduate scholarship from the Natural Sciences and Engineering Research Council of Canada and a Sir James Lougheed Award from Alberta Scholarships. K.H. was supported by the NIH Director's Innovator Award, the Harvard Stem Cell Institute, the V Foundation and the Kimmel Foundation. This study has been approved by the MGH animal and biosafety committees. Deposited in PMC for release after 12 months.

  • Accepted July 16, 2009.

References

View Abstract