ABSTRACT
DNA repair plays a crucial role in embryonic and somatic stem cell biology and cell reprogramming. The Fanconi anemia (FA) pathway, which promotes error-free repair of DNA double-strand breaks, is required for somatic cell reprogramming to induced pluripotent stem cells (iPSC). Thus, cells from Fanconi anemia patients, which lack this critical pathway, fail to be reprogrammed to iPSC under standard conditions unless the defective FA gene is complemented. In this study, we utilized the oncogenes of high-risk human papillomavirus 16 (HPV16) to overcome the resistance of FA patient cells to reprogramming. We found that E6, but not E7, recovers FA iPSC colony formation and, furthermore, that p53 inhibition is necessary and sufficient for this activity. The iPSC colonies resulting from each of these approaches stained positive for alkaline phosphatase, NANOG, and Tra-1-60, indicating that they were fully reprogrammed into pluripotent cells. However, FA iPSC were incapable of outgrowth into stable iPSC lines regardless of p53 suppression, whereas their FA-complemented counterparts grew efficiently. Thus, we conclude that the FA pathway is required for the growth of iPSC beyond reprogramming and that p53-independent mechanisms are involved.
IMPORTANCE A novel approach is described whereby HPV oncogenes are used as tools to uncover DNA repair-related molecular mechanisms affecting somatic cell reprogramming. The findings indicate that p53-dependent mechanisms block FA cells from reprogramming but also uncover a previously unrecognized defect in FA iPSC proliferation independent of p53.
INTRODUCTION
Human papillomaviruses (HPVs) are pathogens that commonly infect basal stem and progenitor cells in the epidermis and can control keratinocyte proliferation and differentiation as a means to perpetuate the viral life cycle (1, 2). Two viral proteins, E6 and E7, have been extensively characterized for their ability to bind and modulate cellular factors that regulate fundamental processes, including proliferation, survival, transcription, and histone modification (3, 4). In the adult epidermis, E6/E7 proteins support the regenerating stem cell compartment while ensuring retention of a full cellular differentiation capacity. The cellular processes affected by E6/E7 proteins all play key roles during the reprogramming of somatic adult cells into induced pluripotent stem cells (iPSC).
Induced pluripotent stem cells are self-renewing, pluripotent cells derived by reprogramming of somatic cells through exogenous expression of the embryonic stem cell (ESC) transcription factors OCT-3/4, SOX2, KLF4, and c-MYC (OSKM), termed the “Yamanaka factors” (5). The complete conversion of a somatic cell into a pluripotent stem cell requires drastic changes in proliferation rates, cell morphology, metabolism, epigenetic modifications, and gene expression (6, 7). These changes occur over a 10- to 20-day period, during which the success of reprogramming in an individual cell depends stochastically on responses to various impediments (8). One such impediment is DNA damage that occurs during early reprogramming (9). The p53 tumor suppressor responds to this damage and can trigger cell cycle arrest, senescence, or apoptosis, depending on the severity of the damage and the ability of the cell to repair it. Thus, p53 activity represses reprogramming at this early stage (10, 11). Repression of p53 increases reprogramming frequency, and anti-p53 short hairpin RNA (shRNA) is now often introduced alongside the Yamanaka factors to improve efficiency (10–13). The acquisition of the high proliferation rate characteristic of pluripotent cells can also be difficult to achieve in reprogramming somatic cells, and thus, increasing the proliferation rate by targeting cell cycle regulators, such as the retinoblastoma protein (Rb), has been demonstrated to increase reprogramming efficiency (14).
iPSC approximate ESC, a cell type that exists only in the inner cell mass of the blastocyst and ultimately gives rise to the entire embryo proper. These cells possess the unique responsibility to prevent genomic mutations that would be passed on to the cells of the entire organism, including the germ line. It is likely for this reason that ESC have evolved to maintain a significantly lower mutation frequency than somatic cells (15). They accomplish this by both increasing the use of error-free DNA repair pathways at the expense of error-prone pathways and undergoing rapid apoptosis in response to elevated DNA damage levels (16–21).
Fanconi anemia (FA) is a genetic disease characterized by bone marrow failure (BMF) and extreme cancer incidence (22). It is caused by mutations in genes that participate in the FA DNA repair pathway, which is required for error-free repair of DNA interstrand cross-links by homologous recombination (HR) and is also involved in promoting HR at DNA double-strand breaks (DSBs) (23). The FA pathway comprises a core complex of FA proteins, including FANCA, FANCB, FANCC, FANCE, FANCF, FANCG, FANCL, FANCM, and other associated proteins, which is assembled in the presence of interstrand cross-link (ICL) DNA damage and functions to promote the activation of ubiquitination of FANCD2 and FANCI, which then dimerize and localize to the site of damage, where they recruit nucleases, the helicase FANCJ, and the HR machinery to repair the damaged DNA (24). Consistent with the essential role of the FA pathway in the repair of ICLs, cells deficient for the FA pathway are hypersensitive to ICL-inducing agents such as mitomycin C (MMC), which causes G2/M arrest and the formation of radial chromosomes in FA cells. Recent studies have revealed that cells from FA patients are resistant to somatic cell reprogramming and that complementation of the defective FA gene restores normal reprogramming efficiency to these cells (25, 26). Thus, a functional FA pathway is required for efficient somatic cell reprogramming. However, the mechanism by which FA cells fail to reprogram is not yet fully understood. A recent study by Muller et al. showed that mouse FA cells experience even higher levels of reprogramming-induced DNA damage and senescence than control cells and suggested that senescence was caused by an inability to repair damage (26). By culturing the reprogramming cells in hypoxia, which is known to increase reprogramming efficiency in normal cells, those researchers were able to reprogram mouse and human FA cells and grow iPSC lines (26). However, the mechanism by which hypoxia promoted FA reprogramming and iPSC growth remains elusive. Importantly, these human FA iPSC lines were not characterized under normoxic conditions. In a separate study, Gonzalez et al. reported that mouse FA and HR-deficient cells experienced high rates of apoptosis and reduced cell proliferation during reprogramming compared to control cells and also that reprogramming efficiency could be increased by inhibition of p53 (27). However, there was no reported attempt to derive iPSC lines from the reprogrammed colonies in that study. A final study attempted to reprogram FA patient cells and reported the derivation of lines at a low frequency (28). However, these lines failed to produce teratomas, suggesting that they are not fully pluripotent. Thus, it remains unknown whether iPSC lines from FA patients can be grown under normoxic conditions.
In this study, we sought to overcome the resistance to reprogramming in FA patient cells and study associated molecular mechanisms through expression of the oncogenes carried by high-risk human papillomaviruses (HPVs). High-risk HPV E6 and E7 target various cell cycle and genome stability regulators to drive proliferation of their target cell. E6 targets p53 for degradation through interaction with the E6AP ubiquitin ligase (29). It can also activate c-MYC and telomerase (30–32). E7 targets the Rb family pocket proteins for degradation, allowing constitutive activation of the E2F transcription factors (33), and also binds and inhibits p21, which controls the G1/S transition downstream of p53 (34, 35). Thus, these proteins systematically regulate the cell cycle checkpoints and the DNA damage response, two regulatory processes that are critical for somatic cell reprogramming. We tested the effect of these oncogenes by themselves and in combination on reprogramming efficiency in FA patient cells and found that E6, but not E7, overcomes the block to reprogramming. We then utilized mutants of E6 to establish that it requires the ability to degrade p53 in order to promote reprogramming, establishing a role for p53 in failed reprogramming of human FA cells. Finally, we attempted to grow both complemented and noncomplemented FA iPSC lines from iPSC colonies, but only complemented colonies produced lines under normal culture conditions despite the continued repression of p53. Together, our studies reveal that repression of p53 overcomes the barrier to reprogramming in FA patient cells but that the resulting iPSC remain sensitive to the deficiency of the FA pathway and fail to self-renew. This observation suggests a broader role and versatile activities for FA and the HR machinery in ESC and iPSC biology.
MATERIALS AND METHODS
Cell culture and reprogramming.Patient-derived keratinocytes were obtained according to a Cincinnati Children's Hospital Institutional Review Board (IRB)-approved protocol and cultured from fresh skin punch biopsy specimens on irradiated J2-3T3 feeder cells, as described previously (36). After the second passage, keratinocytes were transduced with the empty retroviral LXSN vector or LXSN encoding wild-type or mutant human papillomavirus 16 E6. LXSN-E6 vectors were generous gifts from Elliot Androphy (Indiana University) and Saleem Khan (University of Pittsburgh). Keratinocytes were transduced with anti-TP53 shRNA or the nonspecific shRNA control by retroviral infection with pRS vectors (61). All transductions were performed by using supernatants produced from 293T cells in the presence of 8 μg/ml Polybrene. The transduced keratinocytes were selected with 200 μg/ml G418 or 10 μg/ml hygromycin B for 3 days. The resulting keratinocyte populations were then transduced with either the empty or FANCA-expressing MIEG retrovirus, which also expresses green fluorescent protein (GFP). GFP-positive cells were sorted on a BD FACS Aria cell sorter at 3 days posttransduction and replated. These cells were then transduced with a polycistronic lentivirus expressing OCT-4, SOX2, KLF4, and c-MYC (OSKM) as well as red fluorescent protein (RFP) (37). RFP expression was used to confirm equal transduction. Four days after transduction with OSKM, the cells were trypsinized and plated onto irradiated CF-1 mouse embryonic fibroblasts (MEFs) (GlobalStem) in human ESC (hESC) medium, as described previously (38). The medium was changed daily for 19 days. To establish iPSC lines, colonies with characteristic iPSC morphology were picked and transferred into dishes coated with Matrigel (BD Biosciences) in mTeSR1 medium (Stemcell Technologies). mTeSR1 medium was changed daily. Cells were passaged by treatment with 1 mg/ml Dispase until colony edges were rounded, washed 3 times with a 1:1 mixture of Dulbecco's modified Eagle's medium (DMEM)–F-12 medium (Invitrogen), scraped off the dish with a cell scraper in mTeSR1 medium, gently triturated, and then replated.
Staining and immunofluorescence.Reprogrammed cultures were stained on day19 postplating by using the Alkaline Phosphatase Detection kit (Millipore) according to the manufacturer's instructions. Parallel cultures were stained for Tra-1-60 by using a biotin-conjugated Tra-1-60 antibody (eBiosciences), streptavidin-conjugated horseradish peroxidase (HRP) (BioLegend), and the DAB substrate kit (Vector Laboratories). Briefly, cells were fixed on the culture dish in 4% paraformaldehyde (PFA), blocked with 10% goat serum in 0.5% phosphate-buffered saline (PBS)–Tween for 10 min, incubated with primary antibody in blocking buffer overnight, washed with PBS, incubated with streptavidin HRP in blocking buffer for 1 h, washed with PBS, and then treated with DAB substrate according to the manufacturer's instructions. For OCT-4 and NANOG immunofluorescence (IF), cells were fixed on the culture dish with 4% PFA, washed with PBS, permeabilized with 0.5% PBS–Triton X-100 for 10 min, blocked with 10% donkey serum in 0.5% PBS–Triton X-100 for 30 min, incubated with primary antibodies in blocking buffer overnight, washed with PBS, incubated with appropriate fluorescent secondary antibodies (Jackson Laboratories) with 4′,6-diamidino-2-phenylindole (DAPI), washed and overlaid in PBS, and then imaged. Antibodies included OCT-3/4 (catalog number sc-5279; Santa Cruz Biotechnologies) and NANOG (catalog number ab109250; Abcam) antibodies.
Gene expression analysis.iPSC cultures were treated with 1 mg/ml Dispase until the colony edges were rounded, the plates were then washed with a 1:1 mixture of DMEM–F-12 medium, and the colonies were scraped off the dish. The colonies were pelleted by gentle centrifugation, and the RNA was then extracted from the cell pellet by using the RNeasy minikit (Qiagen) according to the manufacturer's protocol. One microgram of total RNA was used for cDNA synthesis using the QuantiTect reverse transcription kit (Qiagen). The resulting cDNA solution was diluted 1:25 and used for quantitative real-time PCR (qPCR) analysis using Sybr green and gene-specific primers. Relative gene expression values were determined by using the ΔΔCT method with the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene as the housekeeping gene. Error bars indicate standard deviations, and significant differences were determined by t test with a P value of <0.05.
Antibodies.Antibodies for Western blot analysis included p53 (catalog number DO-1; Calbiochem), p21 (catalog number EA10; Calbiochem), actin (catalog number C4; Seven Hills Bioresearch), vinculin (catalog number V9131; Sigma), FANCA (Cascade Biosciences), and FANCD2 (Novus) antibodies.
Teratoma formation.Three confluent 35-mm culture dishes of iPSC were harvested with Dispase followed by scraping, as described above. The cells were pelleted and gently resuspended to retain cell clumps in 70 μl of a cold 1:1 mixture of DMEM–F-12 medium. Immediately before injection, 35 μl of Matrigel was added, and the cells were then loaded into a syringe. Cell clumps were injected into the flanks of non-scid-gamma (NSG) mice with a 30-gauge needle. The mice were monitored for tumor formation for at least 16 weeks and were then sacrificed. The tumors were paraffin embedded, sectioned, and stained by hematoxylin and eosin (H&E). Slides were evaluated for the presence of structures derived from all three germ layers by a clinical pathologist.
Statistics.All pairwise comparisons were examined for statistical significance by Student's t test with a P value of <0.05. Error bars indicate the standard deviations of the means.
RESULTS
HPV E6 oncogene expression rescues reprogramming in FA patient cells.Recent studies have established that cells from both mice and humans with defects in the FA pathway undergo somatic cell reprogramming at dramatically reduced efficiencies (25–28, 39). We previously immortalized skin keratinocytes from individuals with FA using the high-risk HPV E6 and E7 oncogenes (36). Since immortalized somatic cells are known to reprogram with increased efficiency (40), we hypothesized that these cells could be reprogrammed despite being deficient for the FA pathway. Keratinocytes from patients within the FANCA complementation group were immortalized with HPV16 E6 and E7 and then transduced with either the control or a FANCA-expressing retroviral vector. The resulting isogenic FA-deficient and -proficient keratinocytes were then subjected to reprogramming by transduction with an OCT-3/4-SOX2-KLF4-MYC polycistronic lentiviral vector (OSKM). After 19 days under ESC culture conditions, both noncomplemented and complemented (FANCA) cells formed large alkaline phosphatase-positive (AP+) colonies with ESC-like morphology, suggesting that immortalization by HPV overcomes the resistance to reprogramming in FA cells (Fig. 1A). Next, we sought to determine which oncogene was sufficient for reprogramming or if E6 and E7 were required. To do this, cells that had been immortalized with E6 and E7, either individually or together, were compared for their relative reprogramming efficiencies (Fig. 1B and C). Remarkably, E6-immortalized cells produced as many AP-positive colonies as E6 and E7 in combination, while E7-immortalized cells produced no colonies. This result suggested that E6 possesses an activity that overcomes resistance to reprogramming in FA patient cells.
HPV-immortalized FA patient keratinocytes yield AP-positive colonies. (A) AP staining of day 19 cultures of FA patient cells immortalized with HPV16 E6 and E7 and either infected with OSKM or mock treated. (B) AP staining of day 16 reprogramming cultures of FA patient cells immortalized with HPV16 E6 and E7, HPV16 E6, or HPV16 E7. (C) Quantification of AP-positive colonies in 2 replicates of the reprogramming experiment shown in panel B.
The above-described immortalized FA keratinocytes were passaged extensively during immortalization, and we therefore sought primary FA patient samples. Transductions with control or HPV oncogene-expressing vectors, the control or FANCA complementation vector, and the OSKM reprogramming vector were carried out consecutively such that primary control cell populations proliferated actively at the point of reprogramming (Fig. 2A). We obtained fresh skin biopsy specimens from 4 patients in the FANCA complementation group (FA-A) and 1 patient in an unknown complementation group (FA-U) and cultured keratinocytes from each specimen. As expected, transduction with a FANCA-expressing retrovirus complemented the FA pathway in FA-A cells, as shown by the ability to monoubiquitinate FANCD2 in response to treatment with hydroxyurea (HU) (Fig. 2B). The ability of E6 and E7 to promote reprogramming was then analyzed in both complemented (FANCA) and control-transduced cells from patients FA-A1 to -3 (Fig. 2C to F). As expected, noncomplemented cells transduced with the empty vector failed to produce AP+ colonies. In contrast, noncomplemented cells expressing E6 alone or E6 and E7 produced many AP+ colonies. Expression of E6 alone or E6 and E7 also increased AP+ colony formation in complemented (FANCA) cells compared to cells not expressing an oncogene. In contrast, E7 alone failed to induce colony formation in noncomplemented cells but did increase colony formation in complemented (FANCA) cells for 2 out of 3 patients. Similarly to the FA-A cells, E6 also efficiently stimulated the reprogramming of FA-U cells, whereas E7 again failed to induce reprogramming (Fig. 2G). We also tested the effect of E6 and E7 on reprogramming efficiency in keratinocytes from a normal donor and found that both oncogenes increased the frequency of AP+ colonies by themselves and when combined (Fig. 2H). However, some of the colonies produced by E7- and E6/E7-transduced cells were large and diffusely stained compared to the small, darkly stained colonies produced by empty vector- and E6-transduced cells (Fig. 2H). Based on their appearance and the high growth rate of these colonies, it is likely that they were transformed. Of note, we observed a similar phenomenon in E6/E7-transduced cells from FA-A patients. In order to confirm that the colonies produced from E6-transduced cells were truly reprogrammed, we repeated the reprogramming experiment on noncomplemented cells from patients FA-A1 and FA-U transduced with control or E6 vectors and stained the resulting colonies for Tra-1-60, an ESC-specific cell surface marker. We found that the colonies produced by E6-transduced cells were Tra-1-60 positive (Fig. 2H and I). Together, these data confirm that E6 expression is sufficient to overcome resistance to reprogramming in FA patient cells, while E7 is not.
HPV E6 oncogene expression rescues reprogramming in FA patient cells. (A) Schematic representation of the sequential viral transduction strategy used to reprogram FA patient-derived cells. (B) Western blot analysis of complementation of the FA pathway in four patient samples using the empty MIEG or MIEG-FANCA vector. IB, immunoblot. (C) AP staining of colonies formed from complemented and noncomplemented cells from patient FA-A1 transduced with HPV16 E6 and E7, HPV16 E6, or HPV16 E7. (D) Quantification of data from 2 replicates of the experiment shown in panel B. (E and F) Quantification of AP-positive colonies formed from complemented and noncomplemented cells from patients FA-A2 and -3 transduced with the empty vector, HPV16 E6 and E7, HPV16 E6, or HPV16 E7. (G) Quantification of AP-positive colonies formed from FA-U patient cells transduced with the empty vector, HPV16 E6, or HPV16 E7. (H) AP staining of colonies formed from normal donor keratinocytes transduced with HPV16 E6 and E7, HPV16 E6, or HPV16 E7. Quantification of 2 replicates is shown at the bottom. (I and J) Tra-1-60 staining of reprogramming cultures of cells from patients FA-A1 and FA-U transduced with the empty vector or HPV16 E6. Quantification of positive colonies is shown at the bottom.
HPV E6 requires the ability to degrade p53 to promote reprogramming.In order to determine the mechanism by which E6 overcomes the block to reprogramming in FA patient cells, we utilized a set of previously characterized E6 mutant proteins that are deficient for specific molecular activities (41, 42) (Fig. 3A).
HPV16 E6 requires the ability to degrade p53 in order to promote reprogramming of FA patient cells. (A) Chart indicating the activities of E6 for which its mutants are deficient. (B) Western blot showing the effects of HPV16 E6 and E7 and mutants of E6 on p53 and p21 levels in response to treatment with mitomycin C (MMC). (C) AP staining of colonies formed from complemented and noncomplemented cells from patient FA-A1 transduced with HPV16 E6 mutants (compare to Fig. 2C). (D) Quantification of data from 2 replicates of the experiment shown in panel C. (E) Quantification of AP-positive colonies formed from complemented and noncomplemented cells from patient FA-A2 and transduced with the empty vector and HPV16 E6 mutant-containing vectors. (F) Quantification of AP-positive of colonies formed from noncomplemented cells from patient FA-A3 transduced with the empty vector and HPV16 E6 mutant-containing vectors. (G) Quantification of AP-positive of colonies formed from FA-U patient cells transduced with the empty vector and HPV16 E6 mutant-containing vectors. (H) Quantification of AP-positive colonies formed from normal donor cells transduced with HPV16 E6 mutant-containing vectors.
Three of the selected mutant proteins (F2V, L50G, and L110Q) have been reported to be deficient for their ability to degrade p53. The F2V and L110Q mutant proteins maintain the ability to stimulate TERT activity, and the F2V mutant protein maintains binding to the E6AP E3 ubiquitin ligase. The L50G mutant is deficient for all three of these activities. Western blot analysis confirmed the effects of wild-type and mutant E6 proteins on p53 levels in response to the DNA-damaging agent mitomycin C (MMC). As expected, wild-type HPV E6 degraded p53, while p53 levels were unaffected by the mutant E6 proteins (Fig. 3B). MMC treatment increased p53 levels in all cells, but MMC-treated wild-type E6-expressing cells maintained lower p53 levels than did the untreated control or E6 mutant-expressing cells. Thus, we conclude that E6 sufficiently ablates the p53 response to DNA damage, whereas the mutants fail in this regard. A fourth mutant (del118-122), which has been reported to maintain the ability to degrade p53 but not activate hTERT, was tested in select patient samples (43, 44). We expressed these E6 mutant proteins in the above-described FA patient cells and tested their ability to promote reprogramming compared to wild-type E6. All three mutants deficient for anti-p53 activity failed to induce AP+ colony formation in noncomplemented patient FA-A1 to -3 and FA-U cells (Fig. 3C to G). These mutants also failed to increase the reprogramming efficiency of complemented (FANCA) FA-A cells and normal donor cells to the same extent as E6 (Fig. 3D, E, and H). The del118-122 mutant induced reprogramming of noncomplemented FA-A cells only slightly less efficiently than did E6, indicating that the activation of hTERT is not essential for the reprogramming-inducing activity of E6 in FA cells (Fig. 3E and F). Together, these results indicate that E6 requires the ability to degrade p53 in order to promote FA cell reprogramming.
Inhibition of p53 is sufficient to circumvent the block to reprogramming in FA patient cells.Having determined that E6 requires the ability to degrade p53 to overcome the block to reprogramming in FA cells, we sought to target p53 directly to verify this finding. Previous studies utilized a TP53-specific shRNA (sh-TP53) and found that it promotes reprogramming of normal somatic cells (10, 11). Thus, we expressed this same shRNA in cells of patients FA-A2 to -4 and FA-U and tested its ability to promote reprogramming. We confirmed that the sh-TP53-transduced cells had drastically reduced p53 levels relative to those of nonspecific shRNA (sh-NS)-transduced cells both with and without MMC treatment (Fig. 4A). We then subjected the cells to reprogramming and found that the expression of p53-specific shRNA resulted in AP+ colony formation in both noncomplemented and complemented cells from three FA-A patient samples and the FA-U patient sample (Fig. 4B to F). As before, the colonies generated were also Tra-1-60 positive (Fig. 4G and H). For further confirmation that the colonies are indeed pluripotent cells, we performed staining for expression of the ESC-specific transcription factor NANOG and found that colonies were positive under all conditions (Fig. 4I). Thus, we concluded that p53 inhibition is sufficient to restore reprogramming of FA patient cells to iPSC.
Knockdown of p53 overcomes resistance to reprogramming in FA patient cells. (A) Western blot depicting efficient knockdown of p53 by sh-TP53 in FA patient keratinocytes. (B) AP staining of colonies derived from noncomplemented and complemented cells from patient FA-A2 transduced with nonspecific shRNA or TP53-specific shRNA. (C) Quantification of data from the experiment shown in panel B. (D) Quantification of AP-positive of colonies formed from noncomplemented cells from patient FA-A3 transduced with nonspecific shRNA or TP53-specific shRNA. (E) AP staining of colonies derived from noncomplemented and complemented cells from patient FA-A4 transduced with nonspecific shRNA or TP53-specific shRNA. (F) AP staining of colonies derived from FA-U patient cells transduced with nonspecific shRNA or TP53-specific shRNA. (G) Tra-1-60 staining of colonies derived from noncomplemented and complemented cells from patient FA-A4 transduced with nonspecific shRNA or TP53-specific shRNA. Quantification of the data is shown at the right. (H) Tra-1-60 staining of colonies derived from FA-U patient cells transduced with nonspecific shRNA or TP53-specific shRNA. Quantification of the data is shown at the right. (I) Immunofluorescence staining for NANOG expression in colonies produced from reprogramming noncomplemented and complemented cells from patient FA-A4.
Reprogrammed FA-deficient colonies fail to generate stable iPSC lines.Despite the expression of several iPSC markers, the colonies generated from noncomplemented FA patient cells tended to be small or to lack the characteristic morphology of a healthy iPSC colony (Fig. 5A and B). Thus, we sought to determine whether these colonies could form stable iPSC lines. To do this, colonies under each condition were picked and transferred to hESC culture conditions on Matrigel. As expected, stable iPSC lines formed frequently from FANCA-complemented colonies (67% in control-transduced cells and 69% in E6- or sh-TP53-transduced cells) (Fig. 5C). However, no cell lines could be generated from 54 noncomplemented colonies across different experiments (Fig. 5C), suggesting an inability of these cells to proliferate or survive once reprogrammed. Most colonies initially adhered to the substrate, but noncomplemented colonies failed to proliferate and gradually disappeared, while the complemented colonies often proliferated robustly and formed lines. No FA-deficient colonies survived beyond 10 days after the initial replating. In contrast, the iPSC lines derived from complemented colonies were maintained in culture for >20 passages (approximately 80 days). We confirmed the expression of the pluripotency genes POU5F1 (OCT-4) and NANOG by both quantitative PCR (qPCR) and IF analysis in cell lines from three different patients (Fig. 5D to H). The OCT-4 protein is likely expressed endogenously since the m-cherry protein that is expressed by the OSKM vector was undetectable, indicating that the vector had been silenced during reprogramming (data not shown). One control and one sh-TP53 line from patient 3 were tested for the ability to form teratomas in immunocompromised mice, and we confirmed the presence of well-differentiated tissues from each of the three embryonic layers (Fig. 5I). Together, these data verify that the complemented cells form bona fide iPSC lines.
Stable iPSC lines are derived and maintained only from cells possessing a functional FA pathway. (A) Images of assorted AP-positive colonies from reprogramming noncomplemented and complemented patient FA-A2 cells. (B) Images of assorted AP-positive colonies from reprogramming noncomplemented and complemented patient FA-A4 cells. (C) Chart indicating the frequency with which stable iPSC lines were derived from reprogramming cultures of control and complemented FA patient cells. (D to F) qPCR analysis for expression of POU5F1 (OCT-4) and NANOG in parental keratinocytes and iPSC lines derived from complemented cells from patients FA-A2 to -4. (G) Immunofluorescence staining for OCT-4 and NANOG in colonies from iPSC lines derived from complemented cells of patient FA-A3. (H) Immunofluorescence staining for OCT-4 and NANOG in colonies from iPSC lines derived from complemented FA-A4 cells. (I) H&E staining of sections from teratomas derived from injection of iPSC from patient FA-A3. (J) Western blot for FANCA and FANCD2 on iPSC lines and parental keratinocytes from patients FA-A2 to -4 treated with 1 mM hydroxyurea (HU).
Since viruses are commonly silenced during reprogramming, we sought to confirm that the FA pathway remained functional in the iPSC lines derived from FANCA-transduced cells. As expected, GFP expression was undetectable in all FANCA-transduced lines, indicating that significant silencing of this vector had occurred (Fig. 5G and H). However, Western blot analysis revealed that exogenous FANCA expression was still detectable in all lines, indicating that silencing was not complete, and all lines maintained a functional FA pathway, as shown by the presence of monoubiquitinated FANCD2 in HU-treated cells (Fig. 5J). Of note, the line that had the lowest levels of FANCA expression and activity (patient FA-A4 sh-NS line 3) grew poorly relative to other lines and experienced high rates of apoptosis (data not shown). Collectively, the sustained functionality of the FA pathway in all lines despite the reprogramming-induced silencing of the FANCA vector suggests selection for iPSC with sufficient FANCA expression to maintain a functional FA pathway, a phenomenon that has been hypothesized in previous studies (25). Together with our inability to derive noncomplemented iPSC lines, these observations suggest that reprogrammed cells continue to require the FA pathway for sustained self-renewal in culture, even when p53 is inhibited.
Next, we sought to verify that the failure of noncomplemented colonies to grow as cell lines was not caused by silencing of the E6 or TP53 shRNA constructs, which would allow for reactivation of p53. Since it was not possible to collect enough cells for analyses of the noncomplemented iPSC colonies, we examined mRNA expression levels of E6 and TP53 in the iPSC lines formed from the complemented cells. If reprogramming had silenced these vectors in the noncomplemented colonies, they would be expected to be silenced in the complemented lines as well. Semiquantitative PCR analysis revealed that the E6 vector was completely silenced in two independently derived iPSC lines from patients FA-A2 and -3 each (Fig. 6A), indicating that the inability of the E6 colonies to grow may have been caused by reactivation of p53. However, in stark contrast, the TP53 shRNA still appeared to be expressed and robustly functional in 2 iPSC lines from patient FA-A3 and in 3 lines from patient FA-A4, since TP53 transcript levels were reduced similarly in the sh-TP53-expressing parental cells and iPSC compared to empty vector control cells (Fig. 6B). The expression of CDKN1A (p21) is dependent upon p53, and thus, we used its expression as a readout of p53 transcriptional activity. We found that the relative reduction of the expression level of CDKN1A in sh-TP53 parental cells compared to control cells was maintained in the iPSC lines, even though total CDKN1A levels were dramatically reduced in iPSC. Thus, we conclude that the TP53 shRNA remains functional throughout reprogramming and in iPSC lines. Combined with our inability to grow the noncomplemented sh-TP53 iPSC colonies as stable lines, the continued functionality of the TP53 shRNA in these lines suggests that p53 inhibition is not sufficient to promote the growth of noncomplemented FA pluripotent cells.
The TP53 shRNA remains functional in iPSC lines. (A) Semiquantitative PCR analysis for expression of HPV16 E6 on parental keratinocytes and iPSC lines derived from cells transduced with the empty vector or the HPV16 E6 vector. (B) qPCR analysis of TP53 and CDKN1A (p21) expression in parental keratinocytes and iPSC lines transduced with control vectors, HPV16 E6, or sh-TP53.
DISCUSSION
In this study, we have demonstrated that human FA cells can be reprogrammed to iPSC colonies by the repression of p53 through either expression of high-risk HPV E6 or p53 RNA interference. Inhibition of p53 is known to promote reprogramming by preventing cell death and cell cycle arrest in response to DNA damage during early reprogramming. Thus, we conclude that FA cells fail to repair reprogramming-induced DNA damage adequately and succumb to p53-induced cell death and cell cycle arrest (Fig. 7). Despite the efficient production of iPSC colonies from FA cells by repression of p53, we were unable to grow noncomplemented FA patient-derived iPSC lines, even from cells where p53 remained repressed. This observation suggests that iPSC continue to require the FA pathway for efficient self-renewal even after reprogramming is complete and that p53-independent mechanisms underlie this growth defect (Fig. 7). Since ESC are known to undergo cell cycle arrest and apoptosis at much lower levels of DNA damage than somatic cells, a likely explanation for the apparent self-renewal defect in human FA iPSC is that the cells respond dramatically to the inability to repair endogenous DNA damage and arrest or die.
Inhibition of p53 rescues reprogramming of FA patient cells but cannot support maintenance of FA iPSC lines. The model schematic indicates the effect of FA at stages of somatic reprogramming and iPSC growth.
The most intensely studied function of the FA core complex and its components, such as FANCA, is in the activation of FANCD2 and FANCI proteins to promote the repair of interstrand cross-link DNA damage. Since repair is thought to be the primary cellular FA function in the cell, it is assumed that the defect in reprogramming is mediated by failed repair of reprogramming-induced DNA damage. Inhibition of p53 results in iPSC colony formation, and increased rates of DNA damage have been observed in FA cells during reprogramming. It is therefore likely that p53-induced apoptosis represents one mechanism inhibitory to reprogramming in FA cells. However, the inability to derive iPSC lines from colonies produced by this approach indicates that p53-independent mechanisms are also involved. Additional functions have been ascribed to the FA pathway besides its primary role in DNA repair, and it is possible that these functions may also contribute to reprogramming. Many recent studies have characterized mitochondrial defects in FA cells, which may affect reprogramming cells in several ways (45–48). First, FA mitochondrial defects cause elevated levels of cellular reactive oxygen species (ROS), which cause DNA and protein damage and are increased during reprogramming. It is likely that elevated ROS levels in reprogramming FA cells contribute to increased apoptosis, as was indicated previously (26). Second, the damaged mitochondria of FA patient cells are more likely to rupture, inducing apoptosis at lower levels of stress than normal cells. Finally, the mitochondrial defects in FA cells may also impede the conversion from the oxidative metabolism of a somatic cell to the glycolytic metabolism of a pluripotent cell. This would explain the observed rescue of FA cell reprogramming in hypoxia, which is known to increase glycolytic metabolism (26, 49). Another change that somatic cells must undergo during reprogramming is altered cell structure, including a mesenchymal-to-epithelial transition (if the starting cells are mesenchymal) and increased nuclear-to-cytoplasmic ratio. In this regard, FA has been shown to affect the structure of cultured cells through deregulation of cytoskeletal proteins, including the mesenchymal protein vimentin, and may thus inhibit reprogramming through the deregulation of cell structure and organization (50). Finally, activated FANCD2 has recently been shown to function as a transcription factor that can suppress oncogene expression (51). Since FANCD2 is highly expressed and activated in ESC/iPSC even in the absence of exogenous DNA-damaging agents (Fig. 5J), it is possible that FANCD2 functions as a transcription factor therein to promote the maintenance of pluripotency. Taken together, a number of DNA repair-related and -unrelated mechanisms, not necessarily mutually exclusive, may underlie the observed FA-dependent defects in reprogramming and pluripotent cell self-renewal.
In DNA repair, the FA pathway functions specifically in the recognition and repair of interstrand cross-links but also functions broadly to promote error-free repair of DNA DSBs and repress error-prone repair by nonhomologous end joining (NHEJ). In a recent study, NHEJ-deficient patient cells were reprogrammed, although at decreased efficiency, and iPSC lines were efficiently derived and maintained (52). In comparison with our observations, it appears that iPSC are much more sensitive to the loss of the FA pathway than NHEJ, suggesting that pluripotent cells depend heavily on FA-mediated repair as opposed to NHEJ. Since FA functions to promote HR-mediated repair, NHEJ presumably dominates the repair of DSBs in FA-deficient iPSC. It will be important in future studies to determine the effect of increased NHEJ-associated signaling on iPSC self-renewal and proliferation.
The HPV E6 and E7 oncogenes were utilized in this study because of their well-characterized ability to prevent cell cycle arrest, apoptosis, and senescence, which are the ultimate fates of reprogramming FA cells. There is also considerable evidence that HPV oncogenes collaborate with FA deficiency to drive even greater cell proliferation. We have previously reported that HPV-immortalized FA keratinocytes form hyperplastic epithelium in an organotypic raft culture system compared to complemented cells (36, 53). Furthermore, several studies have implicated FA in the increased incidence and aggressiveness of HPV-induced mouse models of cancer (54–56), and there is controversial evidence that HPV plays a role in the increased incidence of head, neck, and anogenital tumors in FA patients (57–60). The effects of E6 and E7 on somatic reprogramming were characterized in a previous study, which found that both E6 and E7, alone and in combination, promoted reprogramming by preventing senescence (9). Our results indicate that the ability to degrade p53 is required for E6 to promote reprogramming, while the ability to activate telomerase is not. However, our studies cannot rule out the contribution of other E6 functions to reprogramming, such as the interaction with PDZ domain-containing proteins. It is surprising that E7 did not promote reprogramming of FA cells here, indicating that E7 does not rescue the FA-specific defect that prevents reprogramming in this context. Since E7 binds and inhibits p21, a downstream effector of p53 for G1/S arrest, these data suggest that p53 may not function through p21 to block reprogramming in FA cells. Alternatively, it is also possible that E7 does not sufficiently repress p21 signaling for this purpose. Finally, E7 may drive rapid proliferation through Rb/p130/p107 deregulation such that E7-expressing cells are actually more sensitive to DNA repair defects and undergo apoptosis.
Due to their unique position at the top of the cellular development hierarchy, pluripotent cells are biologically distinct from all other cell types in a multitude of ways, including their well-documented sensitivity to DNA damage and altered DNA repair signaling. The molecular mechanisms that establish these unique features are poorly understood and should be studied further. The results of this study, which utilizes HPV oncoproteins to understand pluripotency and self-renewal, indicate an essential role for the FA pathway not only in reprogramming but also at early stages in the life of a pluripotent cell.
ACKNOWLEDGMENTS
We acknowledge the CCHMC/UC Pluripotent Stem Cell Facility for assistance with hPSC culture and analysis.
This work was supported by NIH grants RO1 CA102357, RO1DK080823, and R01DK092456 and by a Pelotonia Postdoctoral Fellowship grant from The Ohio State University Comprehensive Cancer Center. We also acknowledge core support from a Cincinnati Digestive Disease Center Award (P30 DK0789392) and a Clinical Translational Science Award (U54 RR025216).
FOOTNOTES
- Received 2 June 2014.
- Accepted 14 July 2014.
- Accepted manuscript posted online 16 July 2014.
- Address correspondence to Susanne I. Wells, Susanne.Wells{at}cchmc.org.
REFERENCES
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
