Highlights
- Spontaneous and S-phase-specific chromosome aberrations in differentiated cells
- Higher frequency of residual γ-H2AX foci after exposure to DNA-damaging agents
- Higher frequency of cells with 53BP1 and RIF1 co-localization in differentiated cells
- Higher frequency of cells with a reduced number of RAD51 or BRCA1 foci
Summary
The nitric oxide (NO)-cyclic GMP pathway contributes to human stem cell differentiation, but NO free radical production can also damage DNA, necessitating a robust DNA damage response (DDR) to ensure cell survival. How the DDR is affected by differentiation is unclear. Differentiation of stem cells, either inducible pluripotent or embryonic derived, increased residual DNA damage as determined by γ-H2AX and 53BP1 foci, with increased S-phase-specific chromosomal aberration after exposure to DNA-damaging agents, suggesting reduced homologous recombination (HR) repair as supported by the observation of decreased HR-related repair factor foci formation (RAD51 and BRCA1). Differentiated cells also had relatively increased fork stalling and R-loop formation after DNA replication stress. Treatment with NO donor (NOC-18), which causes stem cell differentiation has no effect on double-strand break (DSB) repair by non-homologous end-joining but reduced DSB repair by HR. Present studies suggest that DNA repair by HR is impaired in differentiated cells.
Introduction
Stem cells have the dual ability to self-renew over the lifetime of an organism and also to differentiate into multiple cell lineages (Weissman et al., 2001, Seita and Weissman, 2010). The majority of mammalian cells in situ originate from a corresponding progenitor daughter cell that is terminally differentiated. Various factors, including reactive oxygen species, that accumulate during differentiation and over the stem cell lifespan, can cause DNA damage (Mikhed et al., 2015). In addition, differentiation-dependent changes in chromatin structure and transcriptional alterations (Nashun et al., 2015, Tran et al., 2015) can also affect genomic integrity by altering the DNA damage response (DDR) and repair facility. Thus, genomic stability is likely to be under increased stress during differentiation. How factors that induce differentiation, such as NO donors, affect stem cell genomic stability is unclear.
Stem cells benefit throughout their lifetime from a robust DNA damage repair activity that enhances resilience toward various environmental factors. Indeed, somatic cells and stem cells differ significantly in their radio-sensitivity (Chlon et al., 2016, Maynard et al., 2008, Lan et al., 2012, Momcilovic et al., 2009, Wilson et al., 2010). However, it is not known how DNA double-strand break (DSB) repair mechanisms are affected during stem cell differentiation. In order to understand whether stem cell differentiation affects DNA damage repair, we compared DDRs and DNA repair in human embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) with their isogenic, differentiated progeny, including neural progenitor cells (neuroectodermal lineage) and their subsequent differentiation products: astrocytes and dopaminergic neurons. DNA damage repair by homologous recombination (HR) was significantly reduced after cell differentiation in all cells examined.
Results
Characterization of Differentiation Markers in iPSCs
Human iPSCs (B12-2) and ESCs (H-9) were used to compare the DDR between undifferentiated and differentiated cell status. The cell lines used were positive for OCT4 or Nanog (Figure 1A) and cell markers (ectoderm β-III tubulin [TUJ1], mesoderm smooth muscle actin [SMA], and endoderm alpha-feto protein [AFP]) and confirmed for embryoid body (EB)-directed differentiation into the three germ layers. During EB-directed differentiation, the first germ layer to be formed is ectoderm, which is identified by the cell marker (TUJ1) in our temporal differentiation (d11). Further, from d14 onward, all three germ layers were observed as indicated (Figure 1B). In other words, on day 11 only TUJ1 stained well; SMA and AFP did not stain, which is reflected in the Figure 1B. Western blot analysis revealed a time-dependent decrease in Nanog, OCT4 (Figure 1C), and hMOF (Figure 1D), while sGCβ1 (Figure 1C) protein levels increased during differentiation. Levels of the hMOF acetylation product H4K16ac were also reduced in differentiated cells (Figure 1D) (Gupta et al., 2008, Kumar et al., 2011, Thomas et al., 2008, Li et al., 2012). During differentiation, levels of H4K20me2 and H3K9ac were not significantly reduced (Figure 1D).
NO Donors Induce Genomic Instability in Stem Cells
We examined whether NO donors induced differentiation by treating stem cells with NOC-18 (5 μM). Differentiation markers such as NKx2.5 (Figure 2A) and myosin light chain 2 (MLC2) protein (Figure 2B) were found to be significantly increased compared with controls. These results are consistent with our earlier report (Mujoo et al., 2008). To determine whether NO also induces DDR, differentiated cells were treated with NOC-18 (0.5 mM) at a sub-toxic dose (95% survival), and response markers were analyzed by western blot and for signaling/repair factor foci formation. Phosphorylation of ATM as well as that of Chk1, Chk2, and H2AX (Figure 2C) was detected in western blots, all indications of DDR activation by NO. Treatment of stem and differentiated cells with NO also resulted in a significant increase in γ-H2AX, 53BP1, and RAD51 foci formation (Figures 2D–2F, S1A, and S1B). There was no difference in the levels of DNA DSB damage induced by ionizing radiation between stem and differentiated cells, indicating that susceptibility to initial damage is independent of differentiation status (Figure S1C). Overall, these results are consistent with previous reports observed in various primary and cancer cells (Fionda et al., 2015, Nagane et al., 2015, Oleson et al., 2014).
DNA DSB and Chromosome Aberration Analysis in Undifferentiated and Differentiated Cells
We next examined DSB rejoining of IR-induced DSBs by the Comet assay and found a significant increase in residual/unrepaired DSBs in differentiated cells compared with undifferentiated cells (Figures 3A and 3B ), suggesting that differentiated cells have a reduced DSB repair capacity. The increased residual DSBs observed in differentiated cells correlated with chromosome aberrations when measured by basal level and IR-induced chromosome aberrations at metaphase. Differentiated cells were found to have a significantly higher frequency of chromosome aberrations (breaks, gaps, radials, dicentrics, micronuclei, aneuploids, and polyploids) compared with undifferentiated cell (Figures 3C and S2A–S2C). About 8%–12% of differentiated cells had chromatin blebbing (Figures 3D and S2D), with a high frequency of micronuclei and endo-reduplicated chromosomes (Figure S2D). Analysis of the cell-cycle specificity of the IR-induced chromosome aberrations indicated no significant differences between undifferentiated and differentiated cells in IR-induced G1-specific aberrations (Figure 3Ei), suggesting that the non-homologous end-joining (NHEJ) repair pathway is largely intact. However, the frequency of S-phase-specific chromosome aberrations was significantly higher in differentiated cells compared with undifferentiated cells (Figure 3Eii), while G2-type chromosomal aberrations did not show any significant differences (Figure 3Eiii). The differentiated cells did not have a significantly decreased percentage of S-phase cells, as there was no significant difference in 5-ethynyl-2′-deoxyuridine labeling between undifferentiated and differentiated cells (Figures S3A and S3B). Since DSB repair by HR occurs largely in S phase, the present data suggest that differentiated cells may have defects in the HR repair pathway.
DDR in Undifferentiated Stem Cells and Differentiated Cells
The initial events of DDR are ATM autophosphorylation and H2AX phosphorylation at Ser139 (γ-H2AX). The frequency of IR-induced γ-H2AX foci induction, which can serve as a surrogate marker for DNA DSBs, was initially identical in undifferentiated and differentiated cells. However, γ-H2AX foci disappearance was delayed in differentiated cells, indicating compromised DSB repair (Figures 4A and S3C). The induction of IR-induced γ-H2AX foci was identical in differentiated ESCs (Figures S4A and S4B) and iPSCs (Figure 4A), suggesting the overall sensing of DNA damage is not affected by differentiation. The higher frequency of residual γ-H2AX foci in differentiated cells indicates a defect in DSB repair that could lie in either the NHEJ and/or HR pathway, although the chromosome aberration data above suggest HR is most likely compromised.
Repairosome Foci Analysis of HR-Related Protein Factors
The 53BP1 protein is involved in suppression of HR (Morales et al., 2003, Ward et al., 2003, Zimmermann et al., 2013), and the first downstream effector of 53BP1 activity is RIF1 (Chapman et al., 2013, Di Virgilio et al., 2013, Escribano-Diaz and Durocher, 2013, Escribano-Diaz et al., 2013, Feng et al., 2013, Zimmermann et al., 2013). While the initial formation of 53BP1 foci post irradiation was identical; there was a significant delay in 53BP1 foci clearance in differentiated cells (Figure 4B). The frequency of RIF1/53BP1 foci co-localization was also higher in differentiated cells compared with undifferentiated cells (Figures 4C–4E). Accumulation of RIF1 at DSB sites containing phosphorylated 53BP1 (Anbalagan et al., 2011, Bonetti et al., 2010) inhibits the DNA resection step of HR (Bunting et al., 2010), suggesting differentiation-dependent suppression of HR-mediated DSB repair. Cellular levels of RAD51 and Chk2 proteins are similar in differentiated versus undifferentiated cells but Chk2 phosphorylation at Thr68 increases during differentiation (Figures 4Fi, 4Fii, and 4Fiii).
Formation of MDC1 repairosomes was identical between undifferentiated and differentiated cells (Figure 5A), but a significant decrease in BRCA1 and RAD51 foci formation after irradiation was observed in differentiated cells (Figures 5B, 5C, and S4C). The reduced frequency of IR-induced BRCA1 and RAD51 foci could affect DNA resection, an early step in HR-mediated repair, and when we examined resection-related proteins for repairosome formation, lower levels of IR-induced MRE11, RAP80, and FANCD2 foci were detected in differentiated cells (Figures 5D–5F). The data supports the argument that differentiated stem cells have a reduced ability to repair DSBs by HR, probably due to failure to displace 53BP1 and allow subsequent HR proteins to load at the DSB site.
We further examined whether the DDR is similarly altered when primitive cells derived from a developing animal organ are fully differentiated in vitro. Later-stage astrocytes had a higher frequency of cells with delayed disappearance of γ-H2AX foci and a reduced number of RAD51 foci, indicating that differentiated astrocytes showed a similar DDR defect as observed in the stem cell-derived differentiated cells (Figures S5A–S5E), suggesting decreased HR is a general feature of cell differentiation.
In addition to repair of IR-induced DNA damage, HR-mediated repair is applicable to other types of damage, especially that arising from replication fork blockage and stalling (Gupta et al., 2014b, Hunt et al., 2013). Interstand cross-links (ICLs) create obstructions to fundamental DNA processes and are repaired predominantly during S phase when replication forks converge at ICL sites (Raschle et al., 2008). We first examined the repair of DNA damage induced by camptothecin, which binds to topoisomerase 1 and forms a DNA covalent complex that is specifically repaired by HR. After camptothecin treatment, differentiated cells exhibited delayed loss of γ-H2AX foci and a lower frequency of RAP80 foci compared with undifferentiated cells (Figures 5G and 5H). Similarly, treatment with the DNA cross-linking drug cisplatin induced a higher frequency of cells with delayed disappearance of γ-H2AX foci in differentiated cells (Figure 5I). Furthermore, RAD51 foci formation, a marker of HR, after cisplatin or hydroxyurea (HU) treatment was reduced in differentiated cells (Figures 5J and 5K).
DDR in Neural Progenitor Cells and Differentiated Astrocytes or Dopaminergic Neurons
Neural progenitor cells (NProC), prepared from B12-2 iPSCs, are of neuro-ectodermal origin and PAX 6 positive/OCT4 negative (PAX 6+/OCT4–) (Figure S6A). Following NProC cell differentiation, mature astrocytes were detected as glial fibrillary acidic protein (GFAP) positive (Figure S6B) and mature dopaminergic neurons as β III tubulin (TUJ1) positive (Figure S6C). iPSCs (B12-2), progenitors (NOProC), and differentiated cells were exposed to IR and foci formation by the HR markers (RAD51 and BRCA1), NHEJ (53BP1 and RIF1), and γH2AX were measured (Figure 6). There was a significant delay in the disappearance of γ-H2AX foci in neural progenitors, astrocytes, and dopaminergic neurons (Figure 6A) and an associated delay in 53BP1 and RIF1 foci clearance compared with undifferentiated cells (Figures 6B and 6C). These results are consistent with mixed-culture differentiated cells (Figures 3 and 4). Moreover, BRCA1 and RAD51 foci were correspondingly reduced in astrocytes and dopaminergic neurons compared with neural progenitor cells (Figures 6D and 6F). Figure 6E is a representative photomicrograph of DAPI (a) 53BP1 (b), and RAD51 foci (c) in astrocytes. These results provide further support for the model that HR-mediated DDR is significantly downregulated in terminally differentiated cells such as astrocytes and dopaminergic neurons.
Impact of Differentiation on DNA Replication Fork Stalling and Resolution
To determine whether reduced RAD51 and BRCA1 foci formation after treatment with agents that induce interstrand cross-links in differentiated cells is due to altered restart of stalled replication forks, we measured the frequency of stalled replication forks and new replication origin firing by using the chromatin fiber assay (Henry-Mowatt et al., 2003). Cells were pulse-labeled with 5-chlorodeoxyuridine (CldU) followed by HU treatment for 2 hr to deplete the nucleotide pool, and subsequently labeled with 5-iododeoxyuridine (IdU) (Petermann et al., 2010, Singh et al., 2013). Contiguous IdU/CldU signals (Figures 7A and 7B ), identifying restarted forks, were significantly lower in differentiated than in undifferentiated iPSCs (B12-2) or ESCs (H9) (Figures 7Ci and 7Di). Further analysis of the DNA fibers indicated the percentage of stalled forks in differentiated cells after 2 hr of HU treatment was higher than in undifferentiated cells, suggesting that differentiated cells resolve stalled replication forks less efficiently in both iPSCs (B12-2) and ESCs (H9) (Figures 7Cii and 7Dii). In addition, differentiated cells had a shorter DNA tract length distribution, indicating reduced replication fork speeds (Figures 7E and 7F).
Replication fork stalling can also arise from RNA:DNA hybrid formation, the latter also referred to as transcriptional R loops. We observed a higher frequency of R-loop formation in differentiated than in undifferentiated cells (Figures S7A and S7B).
Finally, we directly tested whether NO could affect DSB repair by NHEJ or HR or both pathways by using GFP gene reconstitution assays (Pandita et al., 2006, Pierce et al., 1999). Stably transfected cell lines containing the appropriate GFP substrate constructs were transiently transfected with an I-Scel expression vector and simultaneously treated with a sub-toxic dose of the NO donor NOC-18 (0.5 mM; 5% toxicity). Fluorescent GFP-positive cells, indicating repair of the I-Sce1-induced DSB, were detected by fluorescence-activated cell sorting (FACS) 48 hr after transfection and the percentage of cells with repair was calculated. Treatment with the NO donor had no impact on DSB repair through the NHEJ pathway (Figure 7G) however; HR-mediated repair was significantly reduced, suggesting NO specifically affects DSB repair by the HR pathway (Figure 7H).
Discussion
Continuing stem cell proliferation is necessary to generate cell populations of different lineages during development and for maintenance of tissue homeostasis. The critical role of stem cell expansion necessitates a high-fidelity-based mechanism to repair DNA damage since any mutations will compromise all the derived cell lineages (Adams et al., 2010, Serrano et al., 2011). Whether the high-fidelity-based DNA repair pathway that utilizes HR is maintained in differentiated progeny cells has been of great interest. We compared DSB repair in stem cells at undifferentiated and differentiated stages, and our previous studies have demonstrated that differentiation of stem cells is accompanied by NO production to activate the NO-cyclic GMP pathway (Mujoo et al., 2006, Mujoo et al., 2008, Mujoo et al., 2011). Moreover, modulation of NO levels post IR exposure can occur due to the conversion of oxidative species into nitrosative signals (Pacher et al., 2007, Pham-Huy et al., 2008, van Gent et al., 2001). We report that NO donors activate the ATM kinase pathway as well as induce DNA damage as detected by γ-H2AX and RAD51 foci formation, surrogate markers for DNA DSBs, in both undifferentiated and differentiated cells. When we compared the DDR in stem cells before and after differentiation, we found that differentiated stem cells have the following: (1) higher frequency of spontaneous chromosome aberrations; (2) reduced DNA DSB repair after IR exposure; (3) higher frequency of S-phase-specific IR-induced chromosome aberrations; (4) higher frequency of residual γ-H2AX foci formation after IR exposure or cisplatin treatment; (5) higher frequency of cells with 53BP1 and RIF1 co-localization; and (6) higher frequency of cells with a reduced number of RAD51 or BRCA1 foci after IR exposure or cisplatin treatment compared with undifferentiated stem cells. Furthermore, analysis of lineage-specific differentiation toward the neuronal pathway revealed that mature astrocytes and dopaminergic neurons have impaired DSB repair by HR compared with undifferentiated and neuronal progenitor cells. The higher frequency of chromosome aberrations found in differentiated cells correlated with reduced DSB repair and a higher frequency of S-phase-specific aberrations, suggesting that differentiation affects DSB repair. The higher frequency of chromosome aberrations is not due to an altered cell-cycle distribution as there is no difference in the distribution of cell-cycle phases between undifferentiated and differentiated cells. Since no difference in IR-induced G1- or G2-specific chromosome aberrations was observed between undifferentiated and differentiated cells, this suggests the NHEJ DSB repair pathway is not affected, as it is the dominant mode of DSB repair in G-1 or G-2 phase cells. In contrast, differentiated cells have a higher frequency of S-phase-specific IR-induced chromosome aberrations, suggesting that differentiation impairs HR DSB repair, which primarily occurs in S-phase cells. Thus, our results suggest that while the NHEJ pathway is minimally altered, DSB repair by HR is reduced by differentiation of stem cells.
Only a subset of differentiated cells had a higher frequency of residual IR-induced γ-H2AX foci, suggesting that most of the cells can repair the DNA damage. Further, since differentiated cells showed a higher frequency of cells with co-localization of 53BP1 with RIF1, this suggests that in such cells, the subsequent recruitment of the HR-related proteins is impaired. Consistent with these observations, differentiated cells have a reduced frequency of foci formation by RAD51, BRCA1, and other HR-related protein.
Defects in HR can also decrease resolution of stalled replication forks, and we found a higher frequency of stalled replication forks and lower frequency of new replication origins in differentiated stem cells. The increased level of stalled forks could be due to reduced resolution and repair of the forks by HR. Alternatively, increased interstrand cross-links or transcriptional RNA:DNA hybrids, also called R loops, may contribute to stalling. This latter mechanism seems more likely since differentiated cells exhibited a higher frequency of R loops. Further, the reduced levels of H4K16ac we found in differentiated cells may be important due to its unique ability to control chromatin structure and protein interactions, which may facilitate a more open, repair-conducive chromatin configuration (Pandita, 2013, Horikoshi et al., 2016). Acetylation of H4K16 also limits 53BP1 association with damaged chromatin to promote repair by the HR pathway (Tang et al., 2013). Thus, the reduced levels of H4K16ac in differentiated cells could be one of the factors contributing to aberrant DSB repair.
Mature astrocytes and dopaminergic neurons exhibit significantly higher residual damage, in comparison with their undifferentiated and neuronal progenitor cells, as demonstrated by the delayed disappearance of γ-H2AX foci 8–12 hr post irradiation. Similarly, compared with undifferentiated cells, delayed clearance of 53BP1 and RIF1 foci was observed in neural progenitors, mature astrocytes, and dopaminergic neurons, which is consistent with the correspondingly lower frequency of BRCA1 and RAD51 (HR proteins) foci formation in these cells. Our observations support previous studies indicating that both ESCs and iPSCs repair DNA lesions by HR compared with their differentiated derivatives (Rocha et al., 2013). Some studies have also shown that, compared with neural stem cells, terminally differentiated descendant astrocytes lack functional DDR signaling. However, astrocytes retain the expression of NHEJ genes and are indeed DNA-repair proficient (Schneider et al., 2012). Future studies will focus on elucidating the molecular mechanisms of DNA DSB repair in stem cells and their lineage-directed progeny.
Leave a Reply