Soluble histone H2AX is induced by DNA replication stress and sensitizes cells to undergo apoptosis

Background Chromatin-associated histone H2AX is a key regulator of the cellular responses to DNA damage. However, non-nucleosomal functions of histone H2AX are poorly characterized. We have recently shown that soluble H2AX can trigger apoptosis but the mechanisms leading to non-chromatin-associated H2AX are unclear. Here, we tested whether stalling of DNA replication, a common event in cancer cells and the underlying mechanism of various chemotherapeutic agents, can trigger increased soluble H2AX. Results Transient overexpression of H2AX was found to lead to a detectable fraction of soluble H2AX and was associated with increased apoptosis. This effect was enhanced by the induction of DNA replication stress using the DNA polymerase α inhibitor aphidicolin. Cells manipulated to stably express H2AX did not contain soluble H2AX, however, short-term treatment with aphidicolin (1 h) resulted in detectable amounts of H2AX in the soluble nuclear fraction and enhanced apoptosis. Similarly, soluble endogenous H2AX was detected under these conditions. We found that excessive soluble H2AX causes chromatin aggregation and inhibition of ongoing gene transcription as evidenced by the redistribution and/or loss of active RNA polymerase II as well as the transcriptional co-activators CBP and p300. Conclusion Taken together, these results show that DNA replication stress rapidly leads to increased soluble H2AX and that non-chromatin-associated H2AX can sensitize cells to undergo apoptosis. Our findings encourage further studies to explore H2AX and the cellular pathways that control its expression as anti-cancer drug targets.


Background
Chromatin-associated histone H2AX is a key regulator of the cellular response to genotoxic stress and DNA doublestrand breaks [1][2][3][4][5][6]. H2AX is a core histone H2A variant that is randomly incorporated into nucleosomes during DNA replication. It differs from histone H2A by a unique C-terminal tail that contains a highly conserved SQE motif with a serine residue at position 139 [7]. Serine 139 becomes rapidly phosphorylated in response to DNA double strand breaks by protein kinases of the phosphatidylinositol 3-OH-kinase-related kinase (PI3KK) family including ATM, ATR or DNA-PK [8][9][10]. The phosphorylated form of H2AX has been referred to as γ-H2AX [2]. Several lines of evidence suggest that the main function of histone H2AX is to orchestrate the DNA damage response through interactions with BRCT repeat-containing proteins that bind phosphorylated H2AX such as MDC1 [11][12][13]. However, H2AX has also been implicated in a number of other functions including inactivation of sex chromosomes and control of sister chromatid recombination [14,15]. We have recently shown that histone H2AX also has non-nucleosomal functions, specifically, proapoptotic activities in gastrointestinal stromal tumor (GIST) cells treated with the small molecule protein kinase inhibitor imatinib mesylate (Gleevec) [16].
Previous studies in yeast have shown that histone levels need to be tightly regulated and that high levels of nonchromatin bound histones are cytotoxic [17]. Soluble histones can accumulate when DNA replication and histone incorporation into the newly synthesized chromatin become uncoupled, for example after stalling of replication forks [17][18][19].
The present study was designed to determine the mechanisms that lead to increased soluble H2AX. We found that even short-term stalling of replication forks is sufficient to cause an increase of soluble H2AX. Furthermore, we show that transient overexpression of H2AX, but not H2A, promotes apoptotic cell death. In addition, we demonstrate that overexpression of H2AX leads to a nuclear redistribution and/or loss of the active form of RNA polymerase II as well as the transcriptional co-activators CBP and p300. Taken together, our results suggest a model in which replication stress leads to increased soluble H2AX, which sensitize cells to undergo apoptosis in a process that involves a shut-down of gene transcription.

Replication stress induces soluble H2AX
To test whether DNA replication stress can induce soluble H2AX, we used the human osteosarcoma cell line U-2 OS that, unlike GIST cells, does not depend on a single oncogenically activated tyrosine kinase such as KIT and has bona fide normal DNA damage response pathways. Cells were transfected with constructs encoding GFP-tagged H2AX, mutant GFP-H2AX-S139A or empty GFP vector [20]. Expression of overexpressed H2AX proteins was distinguishable from endogenous H2AX based on the increased molecular weight (Fig. 1A). After transient transfection (24 h), a considerable amount of GFP-tagged H2AX proteins were detected in the soluble nuclear compartment ( Fig. 1B; lanes S3, untreated, middle and right panels). When these cells were treated with aphidicolin, a DNA polymerase α inhibitor that induces stalling of replication forks, a rapid increase of soluble proteins was detected in the S3 fractions within 1 h (Fig. 1B). The ribonucleotide reductase inhibitor hydroxyurea (HU), which induces replication stress through nucleotide depletion, had a similar effect (data not shown). Treatment with aphidicolin or HU also led to an increase of endogenous H2AX in the soluble fractions (Fig. 1B). These results highlight previous findings in yeast showing excess levels of free histones in cells in which DNA replication and histone deposition are uncoupled [17][18][19]. Transfection with mutant H2AX in which the phosphorylation site at S139 has been abolished (H2AX-S139A) had a similar effect as wild-type H2AX, suggesting that the increase of soluble H2AX is not dependent on this phosphorylation site. To test whether our results allow generalization, we transiently transfected the human lung cancer cell line H1299 with GFP-tagged H2AX, H2AX-S139A and empty vector. Treatment with aphidicolin for 1 h, also led to an increased fraction of soluble H2AX indicating that his response is not cell type specific (Additional File 1).
To further corroborate our results, U-2 OS cell populations stably expressing GFP-tagged H2AX, H2AX-S139A or empty vector were generated. In stable cell populations, H2AX was undetectable in S3 fractions from untreated cells (Fig. 1C). However, treatment with aphidicolin for 1 h rapidly resulted in an increase of soluble H2AX and H2AX-S139A ( Fig. 1C; S3 fractions).
Taken together, these results show that transient stalling of DNA replication forks is associated with an increase of soluble histone H2AX.

Overexpression of H2AX sensitizes cells to undergo apoptosis
We next tested whether an increase of soluble H2AX can lead to enhanced apoptosis. As shown in Fig. 1B, transient overexpression (24 h) of H2AX is associated with a proportion of H2AX remaining in the soluble nuclear fraction. At the same time, we detected a significant 3.3-fold increase of apoptotic cells (5.9%; p ≤ 0.001) in comparison to empty vector controls (1.8%; Fig. 2A,B). Apoptotic cells were identified based on the characteristic nuclear morphology including DNA condensation and fragmentation and only GFP-positive, transfected cells were counted (Fig. 2B). Likewise, overexpression of mutant H2AX-S139A also caused a 3.8-fold increase of apoptotic cells (6.9%; p ≤ 0.01; Fig. 2A,B). Hence, apoptosis induction does not critically depend on serine 139 phosphorylation, which is underscored by the finding that cells transfected with H2AX-S139A lacked immunostaining for γ-H2AX but still underwent apoptosis ( Fig. 2B; bottom panels). DNA replication stress induces soluble H2AX Figure 1 DNA replication stress induces soluble H2AX. (A) Immunoblot analysis for GFP of U-2 OS cells stably transfected with empty GFP vector (control), GFP-H2AX or GFP-H2AX-S139A for 24 h. (B,C) Immunoblot analysis of soluble nuclear (S3) and chromatin bound (P3) protein fractions obtained from U-2 OS cells transiently (B; 24 h) or stably (C) transfected with empty GFP vector, GFP-H2AX or GFP-H2AX-S139A and treated with 10 μM aphidicolin (APH) for 1 h to block DNA replication. Staining for GFP detects expression of tagged proteins, whereas staining for H2AX detects endogenous protein. Immunoblot for ORC2 is shown to rule out carry-over between fractions. Ponceau S staining demonstrates protein loading.
Overexpression of H2AX but not H2A leads to apoptotic cell death and stimulates apoptosis To determine whether replication stress sensitizes H2AXoverexpressing cells to undergo apoptosis, we used U-2 OS cell populations stably expressing either empty GFP vector, GFP-tagged H2AX or GFP-H2AX-S139A. Cells were treated with aphidicolin for 24 h and the proportion of apoptotic cells was assessed (Fig. 2C). A significant 2.2fold increase of apoptotic U-2 OS cells stably expressing GFP-H2AX was detected from 4.5% in DMSO-treated controls to 9.7% in aphidicolin-treated populations (p ≤ 0.05). In addition, a significant 9.2-fold increase of apoptotic cells was detected in U-2 OS cells stably expressing GFP-H2AX-S139A from 0.5% in DMSO-treated controls to 4.6% in aphidicolin-treated cells (p ≤ 0.0001). No apoptotic cells were found in populations expressing empty GFP vector, regardless of the drug treatment (Fig.  2C).
We tested next whether overexpression of core histone H2A was able to stimulate apoptosis. Transiently transfected cell populations were stained with an anti-myc-tag antibody and assessed for apoptotic nuclear morphology. The proportion of transfected cells was similar to H2AXtransfected populations. No enhanced cell death was detected when myc-tagged H2A was overexpressed under transient conditions in comparison to empty myc-tag vector (Fig. 2D,E).
Taken together, these results indicate that overexpression of histone H2AX, but not histone H2A, can sensitize cells to undergo apoptosis.

Overexpression of H2AX causes chromatin aggregation
In order to determine the mechanisms that underlie H2AX induced cell death, we performed a morphological analysis of U-2 OS cells transiently transfected with empty GFP vector, GFP-H2AX or GFP-H2AX-S139A (Fig. 3A). Excessive amounts of histones have previously been shown to result in non-specific binding to DNA and/or chromatin leading to structural alterations [16,21]. In transient transfection experiments, we detected an increase of cells with aberrant chromatin aggregation from 4% in empty vector controls to 17.2% in H2AXtransfected cells (4.3-fold; p ≤ 0.001) and 14.8% H2AX-S139A-transfected cells (3.7-fold; p ≤ 0.001; Fig. 3A,B). U-2 OS cells transiently transfected with core histone H2A showed a moderate increase of chromatin aggregation from 5.5% in empty vector controls to 10.4% in H2Atransfected cells that was not statistically significant (Fig.  3C). Chromatin aggregation induced by H2AX was not equivalent to apoptosis since cells with chromatin aggregation did not show TUNEL positivity as exemplified in Fig. 3D (bottom panels).

Overexpression of histone H2AX is associated with a redistribution of components of the transcriptional machinery
We lastly asked whether chromatin aggregation induced by overexpression of histone H2AX can lead to a downregulation of active gene transcription [16]. The majority of cellular mRNA transcription is carried out by RNA pol II. The C-terminal domain (CTD) of RNA pol II consists of tandem heptad repeats with the consensus sequence Y 1 S 2 P 3 T 4 S 5 P 6 S 7 that become hyperphosphorylated and hypophosphorylated during the transcription cycle. Phosphorylation at serine 5 is associated with transcription initiation whereas serine 2 phosphorylation is predominantly found during transcription elongation [22]. Active gene transcription can be visualized by immunofluorescence microscopy against RNA pol II S2 and typically results in a fine granular nuclear staining pattern ( Fig. 4A; top panels) [23,24]. The transcriptional co-activators CREB-binding protein (CBP) and p300 regulate RNA pol II-mediated gene transcription by connecting sequence-specific transcription factors to the basal transcription machinery [25]. These proteins display a similar staining pattern during ongoing transcription as RNA pol II. Transcriptional inhibition has been shown to induce a redistribution of the S2 phosphorylated elongating form of RNA pol II (pRNA pol II S2) into speckled nuclear domains and to lead to RNA pol II degradation [26,27]. Likewise, CBP and p300 are redistributed into nuclear speckles upon transcriptional inhibition [27].
We detected a significant increase of cells with redistribution of pRNA pol II S2 into nuclear speckles or loss of pRNA pol II S2 expression below detection level by immunofluorescence microscopy in cells ectopically expressing H2AX (2.9-fold; 36.5%; p ≤ 0.05) or H2AX-S139A (2.7-fold; 34.2%; p ≤ 0.005) when compared to controls (12.5%; Fig. 4A,B). In the vast majority of these cells, nuclear redistribution of RNA pol II S2 coincided with chromatin aggregation (Fig. 4A; middle panels). No significant effect on pRNA pol II S2 distribution was detected when H2A-transfected cells were analyzed (Fig.  4A,B). In addition to pRNA pol II S2, we found the transcriptional co-activator CBP to be displaced from chromatin in cells transfected either with H2AX (3.2-fold increase, 36%; p ≤ 0.05) or H2AX-S139A (2.1-fold increase, 24%; p ≤ 0.01) in comparison to empty vector controls (11.4%; Fig. 4D,E). Similar results were found when cells were analyzed for expression of p300 (Fig. 4F). Furthermore, pRNA pol II S2 was displaced from chromatin after treatment with aphidicolin as shown by immunoblotting further corroborating the results obtained by immunofluorescence staining (Fig. 4G). Since transcription is required for cell viability, our results suggest that the ability of H2AX to abrogate ongoing transcription may sensitize cells to undergo apoptosis.

Discussion
Histone variant H2AX plays a pivotal role in the maintenance of genome stability, in particular in the absence of p53 [1,6,28]. We have previously shown that soluble histone H2AX is involved in apoptosis induction in GIST cells treated with the small molecule inhibitor imatinib mesylate [16]. Here, we extend these results by showing DNA replication stress as one mechanism that can lead to an increase of the pool of soluble H2AX. We show that overexpression of H2AX sensitizes cells to apoptosis in a process that is associated with a disruption of ongoing gene transcription as evidenced by a redistribution of active RNA pol II as well as CBP and p300.
Our report raises a number of questions that warrant further investigation. First, it is unclear why the overall level of apoptosis in U-2 OS after transient overexpression of H2AX is comparatively low (in the range of 5%) in comparison to GIST cells (approximately 40%) [16]. One possibility is that GIST cells are particularly sensitive to increased non-chromatin associated H2AX because they express oncogenic KIT. We have previously speculated that KIT may constitutively downregulate H2AX, which Overexpression of H2AX leads to aberrant chromatin aggregation and apoptosis may render GIST cells particularly sensitive to its reexpression (see also below). Second, despite our current results the precise mechanisms that lead to soluble H2AX remain unclear. Since the reagents used here induce stalling of DNA replication forks, it is possible that the uncoupling of histone production and histone incorporation into chromatin leads to excessive levels of free histones as it has previously been suggested [17][18][19]. However, prolonged replication arrest can also cause DNA breakage and hence the possibility that increased soluble H2AX is a result of accelerated histone eviction during DNA damage-associated chromatin reorganization cannot be dismissed [29]. Third, the precise mechanisms of apoptosis induction by histone H2AX await further clarification. Our results suggest that chromatin aggregation and transcriptional shut down, which cannot be sustained for a prolonged period of time without affecting cell viability, are important. However, this is a relatively non-specific mechanism and other, more direct, pro-apoptotic mechanisms of H2AX may also be involved [30]. Although we cannot rule out the possibility that release of histones from cells undergoing apoptosis contributes to a certain degree to the increase of soluble H2AX [31], the present report provides additional evidence that soluble histone H2AX is a cause for increased cell death and not simply a consequence.
It is conceivable that replication stress is a frequent event in tumor cells [32,33]. Given the potentially deleterious consequences of excessive H2AX [16,17], one prediction would be that mechanisms exist that regulate the pool of non-chromatin-associated H2AX. Such mechanisms are likely to involve the ubiquitin/proteasome machinery since we have previously shown that H2AX can become polyubiquitinated and that proteasome inhibition can stabilize H2AX [16]. Recent findings show that the ubiquitin-conjugating enzyme Ubc13 and the RING-finger ubiquitin ligase RNF8 are involved in DNA damageinduced H2AX ubiquitination [34][35][36][37][38]. Whether these proteins are also involved in the modulation of H2AX levels in tumor cells remains to be tested.
The fact that soluble H2AX induces apoptosis in tumor cells opens new avenues for cancer therapy. Identification of drugs that induce soluble H2AX, e.g. by inhibiting a potential suppressor of H2AX, would be promising. Given the role of the ubiquitin-proteasome system in the regulation of H2AX levels in GISTs [16], it is also conceivable that targeting the protein degradation machinery could lead to increased soluble H2AX levels. Furthermore, the GFP-tagged H2AX vectors could be used in a compound screen with chromatin aggregation as the readout.
In summary, we report here that DNA replication stress rapidly leads to an increase of soluble H2AX. Overexpres-sion of H2AX was found to sensitize cells to undergo apoptosis, which very likely involves the disruption of ongoing gene transcription. Our findings encourage further studies to exploit H2AX and the cellular mechanisms that regulate its expression levels as novel anti-cancer drug targets.

Conclusion
This report shows that stalling of DNA replication leads to an increase of soluble H2AX in human tumor cells. Soluble H2AX was found to sensitize cells to undergo apoptosis and this activity was associated with a transcriptional shut-down. These results encourage further studies to exploit histone H2AX and the mechanisms that regulate its expression as anti-cancer drug targets.

Materials and methods
Cell culture and transfections U-2 OS osteosarcoma cells and H1299 lung cancer cells were obtained from ATCC (Manassas, VA) and maintained in Dulbecco's modified Eagle medium (DMEM; Cambrex, Baltimore, MD) supplemented with 10% fetal bovine serum (FBS; Mediatech, Herndon, VA), 50 U/ml penicillin and 50 μg/ml streptomycin (both BioWhittaker, Baltimore, MD). Cells were transfected with an empty GFP vector or N-terminal GFP fusion plasmids encoding wild-type H2AX and H2AX in which serine 139 has been substituted by alanine (H2AX-S139A) [20] using Fugene 6 (Invitrogen, Carlsbad, CA). In addition, empty myc-tag vector or myc-H2A was used for transient overexpression experiments. Stable cell populations were generated by selecting cells in G418-containing media for 48 h.

TUNEL staining
Apoptotic cells were visualized using the In Situ Cell Death Detection Kit (Roche Applied Sciences, Indianapolis, IN) according to manufacturer's instructions.

Statistical analysis
Statistical significance was assessed using Student's t test for independent samples. P values ≤ 0.05 were considered significant.