Histone deacetylase inhibitors (HDI) have shown promise as candidate radiosensitizers for many types of cancers. However, the mechanisms of action are not well understood, and whether they could sensitize multiple myeloma (MM) to radiation therapy is unclear. In this study, we show that suberoylanilide hydroxamic acid (SAHA) at low concentrations has minimal cytotoxic effects, yet can significantly increase radiosensitivity of MM cells. SAHA seems to block RAD51 protein response to ionizing radiation, consistent with an inhibitory effect on the formation of RAD51 focus in irradiated MM cells. These effects of SAHA on RAD51 focus are independent of cell-cycle distribution changes. Furthermore, we show that SAHA selectively inhibits the homology-directed repair (HDR) pathway. The results of this study suggest that SAHA, a recently approved HDI in clinical trials for malignancies, at lower concentrations may act as a radiosensitizer via disruption of the RAD51-dependent HDR pathway. Mol Cancer Res; 10(8); 1052–64. ©2012 AACR.
Multiple myeloma (MM) is an incurable plasma cell malignancy that accounts for about 1% of human cancers (1). Standard systemic treatment has utilized alkylating agents, anthracyclines, steroids, antiangiogenics agents, proteasome inhibitors, and hematopoietic cell transplantation. MM cells are also very radiosensitive and locally radiocurable. As a result relatively low doses of radiotherapy have proven to be extremely effective in eradicating focal lesions and in palliating symptoms. Although most myeloma patients initially respond to such approaches, median survival is approximately 4 years because of the refractory and/or relapsed disease (2).
Particularly challenging are patients with progressive disease after first-line therapies (3). These patients are often considered for hematopoietic cell transplantation (HCT). Total body irradiation (TBI) containing conditioning HCT regimens are difficult to tolerate given the fact that MM patients are older and the regimens are often combined with chemotherapy agents such as melphalan. Total marrow irradiation (TMI) delivery using large field image guided intensity modulated radiotherapy approaches provides a more targeted form for TBI. This results in a significant reduction in dose to critical normal organs and is better tolerated in HCT patients (4). Clinical trials have clearly showed the feasibility of combining TMI doses of up to 1200 cGy with concomitant melphalan in patients as old as 66 years of age undergoing HCT (5, 6). However, further improvements in the therapeutic index of TMI through dose escalation or the addition of chemotherapy agents that add to the toxicities of the conditioning regimen remain challenging. Whether radiotherapy is used to target local regional disease or to multiple regions in the body, there is a clear need for agents and strategies that will increase the radiosensitivity of these cells to radiation therapy without significant additional toxicity.
Histone deacetylases (HDAC) and its counterpart, histone acetyltransferases (HAT), modify the chromatin conformation by catalyzing the addition or removal of acetyl group from core histones and non-histone proteins, and play important roles in many diverse and essential biologic processes including intracellular signaling (7). HDAC inhibition results in accumulation of acetylated nucleosomal histones and induces differentiation and/or apoptosis in transformed cells. Inhibitors of HDAC, histone deacetylase inhibitors (HDIs), have currently emerged as a new class of therapeutic agents for human malignant diseases (8–11). The potential of HDIs as radiosensitizer has also been showed in many types of solid tumors in vitro. These HDIs, such as FK228, MS-275, and valproic acid can enhance cell-cycle arrest, apoptosis, and increase DNA double-strand breaks (DSB) in irradiated cancer cells (12–15). However, HDIs ability to enhance radiosensitivity of MM cells has not been explored, and a more detailed understanding of mechanisms involved is needed.
Suberoylanilide hydroxamic acid (SAHA), an organic hydroxamic acid, is a nonspecific HDI inhibiting the activity of classes I, II, and IV HDAC enzymes (7). As an orally administered HDI with a manageable profile of side-effects and preliminary evidence of antitumor activity, SAHA was recently approved as single-agent therapy for refractory cutaneous T-cell lymphoma. Phases I and II clinical trials combining SAHA with chemotherapy (including trials for advanced and relapsed and/or refractory MM), and radiotherapy (for solid tumors) are also ongoing (available from: www.clinicaltrials.gov; refs. 16 and 17).
In this study, we show that SAHA at low, minimally cytotoxic doses enhances radiosensitivity of MM cells in vitro, with a novel mechanism involving the inhibition of the homology-directed repair (HDR) pathway of chromosomal breaks. SAHA significantly inhibits the formation of cofoci of RAD51 with phosphorylated H2A.X (γ-H2A.X), a sign of effective DNA damage repair, in irradiated MM cells. The inhibitory effect of SAHA on the focus formation of RAD51 was not because of changes of S and G2/M cell-cycle phase distribution, but is associated with the reduced stabilization of RAD51 protein and its chromatin association in response to irradiation. However, unlike other drugs (18, 19), low-dose SAHA could not increase radiosensitization in nonhomologous end joining (NHEJ)-defective cells. In summary, our data suggest that SAHA may serve as a candidate radiosensitizer, which could have clinically important ramifications for MM patients undergoing radiotherapy and chemotherapy. These strategies may also eventually prove to be applicable to other hematologic malignancies.
Materials and Methods
Reagents and antibodies
SAHA was obtained from Regulatory Affairs Branch, CTEP, NCI, NIH. Anti-phospho-histone H2A.X (γ-H2A.X, ser-139), anti-RAD54 (4E3/1), anti-DNA-pK (4F10C5), anti-tubulin, and anti-acetyl-histone H4 serum were purchased from Upstate. Polyclonal antibodies against RAD51 (H-92), Mre11 (H-300), Nibrin (H-300) and KU86, and monoclonal antibodies against RAD50 (G-2), KU70 and cyclin A (H-3) were from Santa Cruz Biotech Inc. Pooled siRNA oligos for RAD51, Ku70, and control SiRNA-A were also from Santa Cruz Biotech. N-benzoyloxycarbonyl (Z)-Leu-Leu-leucinal (MG132) was purchased from Calbiochem-Novabiochem.
Human MM cell lines U266B1, RPMI8226, and MM1.s were obtained from the American Type Culture Collection (Manassas, VA). KMS-11 was kindly provided by Dr Yen Yun (City of Hope, Duarte, CA). All MM cell lines were maintained in ATCC-formulated RPMI-1640 medium with 2 mmol/L l-glutamine, 10 mmol/L HEPES, 1.0 mmol/L sodium pyruvate, and 15% nonheat-inactivated FBS (Omega Scientific). A panel of human sarcoma U2OS cells (HTB-96; ATCC) integrated with the DR-GFP, SA-GFP, and EJ5-GFP reporters were described previously (20), and were maintained in RPMI-1640 medium with 2 mmol/L l-glutamine, 4.5 g/L glucose, 100 Units/mL penicillin and 100 μg/mL streptomycin, and 10% heat-inactivated FBS (Omega Scientific). Hamster lung fibroblast cell line V79 and their HR-deficient derivatives VC8 (defective in the BRCA2 gene; ref. 21) were kindly provided by Dr Timothy O'Connor (City of Hope, Duarte, CA). V79, VC8, and Chinese hamster ovary (CHO) cell lines AA8 (wild type) and V3 (defective in DNA-PKcs expression; ref. 22) were maintained in DMEM medium with 2 mmol/L l-glutamine, 100 Units/mL penicillin and 100 μg/mL streptomycin, and 10% heat-inactivated FBS.
Cell viability assay
MM cells were washed once in RPMI-1640 medium. 1 × 105 cells were plated per well in a 12-well format with RPMI-1640 complete medium. The cells were allowed to recover for 24 hours after which they were exposed to the drugs in RPMI-1640 medium for indicated time. Cells were collected by detachment from the plate using a cell scraper for MM1.s cells, and by gently resuspending with pipetting for other cells. Cells were then washed once with PBS, and resuspended in 1 mL PBS. Cell viability was determined by Trypan blue exclusive counting.
MM cells were irradiated using a Mark I Cs-137 Irradiator (J.L. Shepherd Association) at a dose rate between 1.20 and 1.26 Gy/min. Administered doses were validated using commercially available nanodot optically stimulated luminescence dosimeters (Landauer Inc.).
To evaluate radiosensitivity, cells in log phase were seeded at 2 × 105/mL for 24 hours, and treated with SAHA at the indicated concentrations, or dimethyl sulfoxide (DMSO) as control. Irradiation was then delivered 16 hours later. Irradiated cells were maintained in conditioned medium for an additional 3 hours, and then 500 to 5000 cells were suspended in RPMI-1640 medium containing 0.6% SeaPlaque agarose (Lonza) and overlaid onto 60-mm dish containing a solidified bottom of 0.6% agarose in RPMI-1640 medium. Once the top layer solidified, 1 to 2 mL of medium was placed on top to keep the plates moist. Plates were incubated for 2 to 3 weeks until colonies were visible. The plates were then stained with 0.005% crystal violet for 30 minutes and destained with PBS. Colonies with greater than 50 cells were counted as surviving colonies and the number of colonies was normalized to that observed for unirradiated controls. Mean inactivation doses were determined by the method of Fertil and colleagues (23), and the sensitizer enhancement ratio (SER) for HDAC inhibitor treatment was calculated as the ratio of mean inactivation dose control/mean inactivation dose SAHA-treated.
For clonogenic assay with SiRNA, U266B1 cells in log phase were transiently transfected with Ku70 SiRNA or control SiRNA-A. After 24 hours, cells were pretreated with 200 nmol/L SAHA for 16 hours, and irradiated with indicated doses. Cells were then plated 3 hours later after irradiation for colony-formation as described by Matsui and colleagues (24). Briefly, 0.5 mL cells (2000 cells/mL) in RPMI-1640 complete medium were added to 2.5 mL 1.2% methylcellulose medium (STEMCELL Technologies Inc.). The mixtures were then plated onto 35-mm Petri dishes and maintained in a 37°C, 5% CO2, fully humidified incubator. After 14 days of incubation, colonies consisting >50 cells were directly scored using an inverted microscope and colony formation for each condition calculated in relation to values obtained for untreated control cells.
For clonogenic assay with hamster lung fibroblast and hamster ovary cell lines, cells in log phase were plated for 8 hours, and treated with 200 nmol/L SAHA or DMSO as a control; ionizing radiation (IR) was then delivered 16 hours later. Irradiated cells were maintained in SAHA-containing medium for 7 to 10 days until colony counting.
MM cells were collected by centrifugation, and washed once with Ca2+/Mg2+-free PBS, and fixed in 2% paraformaldehyde. 2 to 3 × 105 cells were then attached to poly-l-lysine-coated glass slides, permeabilized with 0.5% Triton-X100 in PBS for 10 minutes, and blocked with 0.05% Tween + 2% BSA in PBS for 40 minutes at room temperature. Cells were then incubated with primary antibodies (anti-RAD51 and anti-phospho-histone H2A.X, or anti-cyclin A antibodies) for 2 hours at 37°C followed by 3 washes with PBS and a 60-minute incubation at 37°C with the corresponding secondary antibodies: Alexa Fluor 594 (Red) and 488 (green) at 1:500 (Invitrogen). Cells were then washed 3 times with PBS and mounted with Vectashield mounting medium containing DAPI (Vector Laboratories Inc.). Images were acquired with LSM 510 confocal microscope (Zeiss) with 40× objective and processed by Photoshop (Adobe). For quantitative analysis, nuclei were analyzed by eye. At least 100 cells from each experiment were selected at random and were counted to calculate the percentage of cells as “positive” for both RAD51 and γ-H2A.X if they displayed >5 discrete dots in nuclei. Cells containing discrete merged dots per nuclei in cells that were scored as RAD51 foci–positive were counted as positive for RAD51/γ-H2AX co-foci. The results from at least 3 different experiments were averaged.
Cell-cycle and bromodeoxyuridine incorporation assay
1 × 106 MM cells in log-phase were treated with or without 200 nmol/L SAHA for 16 hours followed by exposure to 6 Gy IR. Ten micromolars of BrdU (BD Biosciences) were added into cell suspension 1 hour before collection. Cells were fixed at indicated time intervals with cold 70% ethanol. Cells were then labeled with fluorescein isothiocyante (FITC)-conjugated anti-BrdU monoclonal antibody (BD Biosciences) as per manufacturer's instruction, and stained with propidium iodide (PI; 10 μg/mL). Detection of bromodeoxyuridine (BrdUrd) incorporation in DNA synthesizing cells was conducted by flow cytometric analysis (FACS) on a Cyan ADP (Beckman Coulter). Up to 5 × 105 cells were analyzed for each sample.
DNA damage repair assay
To measure repair, 1 × 105 U2OS reporter cells were plated onto a 12-well plate and transfected the next day with 0.8 μg of the I-SceI-expression vector, pCBASce (20), with 3.6 μL Lipofectamine 2000 (Invitrogen) in 1 mL of antibiotic-free medium. For the SiRNA conditions, 10 pmol pooled RAD51 SiRNA oligos (Santa Cruz Biotech Inc.) were also cotransfected. Three hours later after the initiation of the transfection, cells were washed once with growth medium, and incubated in fresh medium with 500 nmol/L SAHA or DMSO for 3 days. GFP-positive cells were quantified by flow cytometric analysis. Up to 5 × 104 cells were analyzed for each sample.
Western blot assay
Cells were washed with cold PBS twice, and then lysed in RIPA buffer containing protease and phosphatase inhibitors (Thermo Scientific) with mild sonication. To determine the acetyl-histone H4 and phospho-histone H2A.X, whole cells were lysed in complete RIPA buffer containing 1 mmol/L TSA and 5 mmol/L nicotinamide.
For subcellular fractionation, 2 to 3 × 106 cells were resuspended in 0.2 mL cytoskeleton buffer (0.5% Triton X-100, 100 mmol/L NaCl, 3 mmol/L MgCl2, 300 mmol/L sucrose, 1 mmol/L EGTA, 10 mmol/L HEPES, pH 6.8) containing protease and phosphatase inhibitors and incubated for 10 minutes on ice. The cells were spun down at 500 × g for 5 minutes at 4°C. The supernatant and pellet were collected and designated as Triton X-100 soluble and Triton X-100 insoluble fractions, respectively. The Triton soluble fraction was clarified by additional centrifugation at 15,000 × g for 10 minutes at 4°C. The Triton insoluble fraction was washed with buffer A (10 mmol/L NaCl, 5 mmol/L MgCl2, 250 mmol/L sucrose, 1 mmol/L EGTA, 10 mmol/L Tris-HCl, pH 7.6) containing protease and phosphatase inhibitors and resuspended in 200 μL of the same buffer. Eighty U/ml of RNase-free DNase I (Roche) were added and the nuclei (Triton insoluble fraction) were incubated for 15 minutes on ice to digest genomic DNA for easy loading (25). When applied for Western blot assay, 6 μg of total cell lysates and Triton-soluble fractions, and 15 μg of Triton-insoluble fractions were loaded onto SDS-PAGE gel.
To verify the changes of protein levels and protein modification, densitometry for western blot signal was conducted using the AlphaEase FC software (AlphaInnotech).
RNase protection assay
Total RNA was extracted using Trizol reagent (Invitrogen). The RNase protection assay was conducted as previously described (15). Briefly, 2 μg of RNA was incubated with α-32P-UTP-labeled single-stranded RNA probes overnight at 56°C and treated with RNase for 45 minutes at 30°C. The RNA–RNA complexes were resolved by electrophoresis in 6% denaturing polyacrylamide gel and analyzed by autoradiography. Template sets for custom-designed template sets for DSB-related genes were purchased from BD Biosciences.
SAHA induces histone acetylation and reduces cell survival after irradiation in MM cells in vitro
We first determined IC50's and the effective doses of SAHA on histone acetylation in MM cells. These experiments enabled us to choose our subsequent experimental concentrations of SAHA that had only minimal cytotoxicity and so negligible effect on cell growth. Our results showed that the IC50's of SAHA on tested MM cells ranged from 0.7 to 1.4 μmol/L, and a minimum of 200 nmol/L SAHA markedly increased histone H4 acetylation in all 4 cell lines (Fig. 1A and Supplementary Data S1). We next examined the effects of SAHA on cellular responses to IR by clonogenic survival assay. Two micromolars of SAHA induced marked radiosensitization in MM cells as expected, which was however accompanied with significant cytotoxicity (Supplementary Data S2). When exposed to lower concentration of SAHA (200 nmol/L), with detectable biologic effect on acetylation of histone proteins, radiosensitizing effects were also detected in MM cells, with observed SER of 1.20 ± 0.02 for RPMI8226, 1.36 ± 0.05 for U266B1, 1.11 ± 0.08 for KMS-1, and 1.32 ± 0.02 for MM1.s, respectively (Fig. 1B).
SAHA induces the persistence of γ-H2AX subnuclear foci in irradiated MM cells
IR produces DNA DSBs in chromosomal DNA that can lead to cell-cycle arrest or cell death. One of the earliest known responses to radiation-induced DSB formation is phosphorylation of the C-terminal tails of variant H2AX (γ-H2A.X) in chromatin, and the formation of γ-H2AX foci at DNA break sites. Studies have supported the hypothesis that the persistence of γ-H2A.X nuclear foci is an indicator of lethal DNA damage with nonrepaired DNA DSBs (26–28). To test whether the exposure to SAHA changes the formation of γ-H2A.X foci and its persistence in response to IR, we determined the kinetics of γ-H2A.X foci after IR exposure in MM cells. As shown in Fig. 2, accumulation of γ-H2A.X foci dramatically increased within 1 hour after 6-Gy irradiation, with values of positive fractions ranging from 80% to 90% in all 4 cell lines examined, and then decreased slowly over time. By 24 hours after IR, the percentage of cells with residual γ-H2A.X foci dropped to 30.4 ± 8.1, 37.8 ± 4.7, 36.5 ± 4.6, and 42.7 ± 6.1 for MM1.s, U266B1, RPMI8226, and KMS-11 cells, respectively. Exposure to SAHA had no obvious effects on the percentage of cells with γ-H2A.X foci within 7 hours after IR; however, the number of cells with residual γ-H2A.X foci after 24 hours in irradiated cells remained higher at 57.1 ± 4.1, 61.2 ± 2.5, 63.0 ± 6.7, and 62.2 ± 3.8 for MM1.s, U266B1, RPMI8226, and KMS-11 cells, respectively, indicating SAHA exposure induced the persistence of γ-H2A.X foci in irradiated MM cells. The effects of SAHA on the persistence of γ-H2A.X foci were also observed in MM cells after 36 hours postirradiation.
In parallel experiments for the examination of γ-H2A.X protein levels, we also found that SAHA exposure did not cause obvious changes of baseline γ-H2A.X and IR-induced γ-H2AX within 7 hours after IR. However, it prolonged postirradiation of γ-H2AX until 24 hours in all tested MM cells (Supplementary Data S3).
SAHA modulates RAD51 protein levels and its nuclear cofoci with γ-H2A.X in irradiated MM cells
To define possible roles of DNA repair mechanisms in SAHA-modulated radiosensitization, we examined the effects of SAHA on levels of several proteins known to be involved in the repair of IR-induced DSBs. Western blotting results showed that SAHA exposure had no notable effects on the protein level of KU70, KU86, DNA-pK, RAD50, Mre11, or Nibrin in MM cells after 6 Gy IR (Supplementary Data S4). Of note, preexposure to 200 nmol/L SAHA for 16 hours produced inconsistent changes of baseline RAD51 protein: the protein level slightly increased in RPMI8226 cells, but decreased in MM1.s and U266B1 cells, whereas no changes were observed in KMS-11 cells. However, SAHA exposure dramatically blocked the IR-induced increase in RAD51 levels shortly after the delivery of irradiation in all tested MM cell lines (Fig. 3). SAHA also blocked the increase of RAD54 in irradiated RPMI8226, U266B1, and MM1.s cells.
Upon the generation of γ-H2A.X foci in response to IR, DNA repair factors, such as RAD51, will be recruited to sites of DNA damage (29, 30). To test whether the observed effects of SAHA on RAD51 protein would affect the IR-induced recruitment of RAD51 to DSBs site, we evaluated the formation of nuclear cofoci of RAD51/γ-H2A.X by immunofluorescence staining in irradiated MM cells.
In exponentially growing populations of unirradiated MM cells, only a small fraction (3%–8%) showed subnuclear RAD51 foci (≥5 foci per nucleus), and no RAD51/γ-H2A.X cofoci were detected. After IR treatment, the percentage of cells with RAD51/γ-H2A.X cofoci achieved maximum levels at 24 hours in U266B1, RPMI8226, and KMS-11 cells, and then decreased at 36 hours. In MM1.s cells, the percentage for cells with nuclear RAD51/γ-H2A.X cofoci achieved maximum levels at an earlier time point (7 hours), showing decreases at both 24 and 36 hours. Notably, IR-induced formation of RAD51 foci in these cells was most apparent in ones that were γ-H2A.X focus–positive, with only <2% of cells showing RAD51 foci in the absence of γ-H2A.X foci. When counting RAD51 foci alone, we found that exposure to 200 nmol/L SAHA did not result in obvious changes in spontaneous RAD51 foci in growing MM cells. However, pretreatment with SAHA reduced the proportion of irradiated cells with RAD51 foci at 7 hours and beyond. Furthermore, SAHA exposure also significantly decreased the ability of cells to form nuclear RAD51/γ-H2A.X cofoci in response to IR (Fig. 4 and Supplementary Data S5). For example, at 24 hours after IR, SAHA treatment caused 3.75 ± 0.36, 2.52 ± 0.50, 2.05 ± 0.15, and 2.82 ± 0.89 fold decreases in the frequency of cells with RAD51/γ-H2A.X cofoci in U266B1, RPMI8226, KMS-11, and MM1.s cells, respectively.
SAHA inhibition of RAD51 nuclear focus formation in response to irradiation is not associated with cell-cycle S–G2 phase changes
Previous studies showed that IR-induced RAD51 nuclear focus formation was cell-cycle regulated (31, 32). To address whether the inhibitory effects of SAHA on the formation of RAD51/γ-H2A.X cofoci was because of possible cell-cycle redistribution, we determined the kinetics of nuclear accumulation of cyclin A, a marker for cells in S to G2/M phase (33), and its correlation with nuclear RAD51 foci in response to SAHA combined with IR in U266B1 cells.
In unirradiated control and SAHA treated cells, about 8% of cells were positive for RAD51 nuclear foci, and all these cells were also cyclin A nuclear stain positive; an additional ∼12% of the cells were cyclin A–positive alone. After 6 Gy IR, the numbers of cells with cyclin A nuclear staining started increasing within 1 hour, and continued at very similar rates in the control and SAHA-exposed cell populations, reaching 57.7 ± 4.2% for control and 57.1 ± 3.1% for SAHA at 24 hours after IR (Fig. 5A and Supplementary Data S6A). Consistent with the above kinetic results for RAD51/γH2A.X cofoci, the percentage of cells positive for RAD51 foci also increased gradually in cells treated with IR alone, with the maximum of 42.2 ± 4.2% at 24 hours postirradiation. When cells were exposed to 200 nmol/L SAHA, IR also increased the percentage of RAD51 focus–positive cells, but with a significantly reduced frequency compared with IR without SAHA exposure. The maximum percentage of cells with RAD51 foci was only 19.1 ± 1.2 at 7 hours after IR, and then persisted at closely the same level until 36 hours (Fig. 5B). As expected, the formation of nuclear RAD51 foci in response to IR was predominantly in nuclear cyclin A–positive cells (Fig. 5C). Thus, the relative loss of RAD51 focus–positive cells after IR, in SAHA-treated versus control cells, cannot be explained by a smaller population of S-phase cells with the former experimental condition.
As an alternative approach, we also analyzed changes in BrdUrd incorporation (DNA synthesizing cells) and cell-cycle distributions in MM cells after combined IR plus SAHA treatment versus IR plus vehicle. 6 Gy IR induced G2 arrest in U266B1 cells. IR also first slightly increased the percentage of cells entering S-phase shortly after the delivery of IR, and then significantly decreased it. Exposure to 200 nmol/L SAHA did not cause notable changes in either cell-cycle distribution or cell fractions with BrdUrd incorporation in response to IR. Similar results were also observed in RPMI8226 cells (Fig. 5D and Supplementary Data S6B).
SAHA inhibits homology-directed repair
Because SAHA treatment caused a reduction in RAD51 levels and focus formation after IR, we hypothesize that SAHA exposure would specifically inhibit the RAD51-dependent, HDR repair pathway for DSBs. We utilized a set of chromosomally integrated reporter assays in sarcoma U2OS cells in which a site-specific DSB is introduced by expression of the I-SceI endonuclease (20). Subsequently, a GFP + cassette is restored in each reporter cell line as follows: HDR using a downstream homologous template (DR-GFP), single-strand annealing (SSA) between two 266 nt repeats that flank the DSB to cause a 2.7 kb deletion (SA-GFP), or end joining (EJ) between distal ends of 2 tandem DSBs (EJ5-GFP). A previous study with these reporters showed that disruption of RAD51 caused a decrease in HDR and an increase in SSA, and had no effect on EJ (34). In our experiments, the repair frequencies (percentages) in U2OS cells after transient expression of I-SceI were 2.32 ± 0.37, 3.57 ± 0.18, and 4.78 ± 0.80 for HDR, SSA, and EJ, respectively. As expected, when cells were exposed to 200 nmol/L of SAHA, HDR was significantly reduced to 1.5 ± 0.27 and SSA was increased to 8.2 ± 0.50, whereas there was no obvious change to EJ (5.1 ± 0.97; ref. Fig. 6A). Furthermore, SiRNA depletion of RAD51 also led to significantly inhibited HDR, and an increase in SSA, with no obvious change to EJ.
In these reporter experiments, SAHA affected HDR in cells not exposed to IR, whereas the effects of SAHA on RAD51 levels in MM cells were only observed after IR treatment. However, the reporter experiments involved an extended exposure to SAHA. We therefore examined whether SAHA treatment for an extended time would affect RAD51 levels without IR. We found that SAHA treatment at 200 nmol/L for 40 to 72 hours led to a significant decrease in RAD51 protein levels in both U2OS cells and the MM cell line U266B1 (Fig. 6B). Consistent with the above experiments in MM cells (Fig. 3), SAHA exposure for less than 40 hours had no clear effect on RAD51 levels without IR treatment. Thus, although only a short SAHA treatment is necessary to disrupt the accumulation of RAD51 in response to IR, extended SAHA treatment seems to be required to observe decreased levels of RAD51 without IR.
We next employed a pharmacological-genetic epistasis experimental approach to confirm whether the observed radiosensitizing effects of SAHA stem from impaired processing of IR-induced DNA damage by the HR pathway. If, as we suggest, SAHA exposure phenocopies a defect in the HR pathway, then we expect this treatment to have no radiosensitizing effect in cells that are HR-deficient by virtue of biallelic inactivating mutations in some HR pathway gene. Using a matched pair of Chinese Hamster cells having HR-deficiency or not, we compared clonogenic survivals in response to IR with prior SAHA or DMSO vehicle treatment (Fig. 6C). HR-proficient V79 parental cells displayed significant radiosensitization by SAHA: the SER with respect to mean inactivation was 1.26 ± 0.10. Congenic BRCA2-deficient VC8 cells, in contrast, exhibited little if any radiosensitization by SAHA (SER 1.09 ± 0.05). We note that control VC8 cells are more radiosensitive than SAHA-treated V79 cells, indicating that drug-induce HR inhibition is incomplete.
HR is thought to afford some backup function for repair of DSB in mammalian cells when the major pathway, NHEJ, is inactivated (35). It follows that SAHA would be expected to be a more potent radiosensitizer in an NHEJ-deficient cell than in a congenic repair proficient one. Contrary to this expectation, however, SAHA at 200 nmol/L radiosensitized parental AA8 and congenic DNA PKcs-defective V3 cells almost equally (SER 1.27 ± 0.07 and 1.31 ± 0.05, respectively). Control and KU knockdown cells yielded very similar results (SER 1.18 ± 0.03 and 1.20 ± 0.03, respectively; Fig. 6C and D). These results indicate that the IR-induced DNA damage associated with radiosensitization by SAHA, or SAHA at low concentrations, must be repaired via the HR pathway almost exclusively.
SAHA modulates RAD51 protein stabilization and reduces the activity of RAD51 in chromatin association in irradiated MM cells
In an attempt to understand in more detail the mechanisms involved in the SAHA-mediated modulation of RAD51 protein levels after IR, we evaluated the effects of SAHA exposure on IR-induced redistribution of RAD51 protein to sites of DNA damage. Cells treated with IR and SAHA were fractionated into a Triton X-100 soluble fraction (TSF) containing free nuclear proteins, and a Triton X-100 insoluble fraction (TISF) containing chromatin-bound proteins (25, 36). As shown in Fig. 7A, 6 hours after irradiation, the amounts of RAD51 in the TSF and TISF increased, concomitant with the increase of total RAD51 protein. However, when cells were exposed to SAHA, the RAD51 in the TISF did not show any obvious increase. Instead, the slightly increased total RAD51 in cells exposed to IR combined with SAHA were confined to the TSF. These results indicate SAHA exposure reduces the chromatin-associated RAD51 specifically in irradiated MM cells.
A growing list of drugs, including HDI, has been recently reported to reduce HR through modification of gene expression at the transcriptional level and, consequently, decreases of baseline RAD51 protein level and its nuclear focus formation in irradiated cancer cells (18, 19, 37–39). However, our RPA assays (Supplementary Fig. S7) showed that exposure to 200 nmol/L SAHA did not change RAD51 mRNA levels within 40 hours in MM cells in response to IR. Interestingly, although RAD51 protein level dramatically decreased in cells exposed to 200 nmol/L SAHA for longer than 40 hours in U266B1 (Fig. 6B), no changes in RAD51 mRNA level were observed in cells after extended exposure to SAHA. However, exposure to higher concentrations of SAHA, such as 2 μmol/L, significantly decreased both mRNA and protein level of RAD51 in U266B1 cells. Of note, 2 μmol/L SAHA exposure also inhibits the transcription of other DNA damage repair factors.
We then tested whether SAHA could modulate protein stabilization of RAD51 in irradiated MM cells. As shown in Fig. 7B, when cells were cotreated with 500 nmol/L MG132, a 26S proteasome inhibitor, the repressed RAD51 levels seen previously with SAHA exposure in irradiated cells was restored by ∼50% in all 3 cell line. A parallel immunofluorescence assay also showed that MG132 treatment restored the cell's ability to form RAD51/γ-H2A.X cofoci at 7 hours after 6 Gy IR in U266B1 cells preexposed to SAHA (Fig. 7C). Taken together, these data suggest that SAHA does not influence the gene expression of RAD51, but, at low concentrations, SAHA may reduce protein stability of RAD51 and its redistribution into the nucleus, leading to the inhibition of the chromatin association in response to irradiation.
Radiation therapy is a key modality in the treatment of cancer. Although the molecular basis of radiation response is complex and multifactorial, the predominant mechanism by which therapeutic irradiation kills most tumor cells is through clonogenic death. DSBs are regarded as the specific lesions that initiate this lethal response (40, 41), and the repair of DSBs is then critical in determining radiosensitivity (42, 43). In mammalian cells, radiation-induced DSBs are repaired by a complex mechanism involving several principle pathways: NHEJ, HDR, and SSA. In HDR, RAD51 serves as a key recombination protein, which mediates the search for a homologous duplex template and the formation of joint molecules between the damaged DNA and the undamaged repair template. Previous studies have found cancer cells with elevated levels of RAD51 protein as compared with normal cells (44), and targeting RAD51 and its dependent HDR might have potential impact on sensitizing cancer cells to radiation therapy and DNA-active chemotherapeutic drugs (39, 45–47).
In this study, we showed that treatment with SAHA, at low concentrations that produce minimal cytotoxicity, can radiosensitize MM cells. Mechanistic investigations revealed that SAHA exposure disrupts the normal RAD51 responses to IR exposure, and causes inhibition of HDR. We found that the formation of RAD51 foci, and their colocalization with γ-H2A.X foci in irradiated MM cells, were significantly inhibited by SAHA. Cells that are HDR-deficient often show defects in RAD51 focus formation, which likely reflect a decreased recruitment of this key HDR factor to sites of DNA damage (48–50). The inhibition of RAD51/γ-H2A.X cofocus formation observed in irradiated MM cells after SAHA pretreatment likely indicates that SAHA inhibits RAD51 recruitment to DSBs during repair. Consistent with causing a defect in DSB repair, SAHA pretreatment resulted in the persistence of γ-H2A.X foci at late time points in irradiated MM cells; the latter finding is thought to reflect lethal DNA damage (26, 27).
Like other HDIs that have been reported to decrease RAD51 protein and inhibit HDR in response to IR treatment (18, 19, 37), SAHA exposure lead to the reduction of baseline RAD51 protein levels. However, we present evidence here that SAHA at low concentration can also block the IR-induced increase of RAD51 protein, and furthermore reduce the formation of nuclear RAD51 foci shortly after IR. These effects of low dose SAHA on RAD51 dynamics occur before SAHA changes the baseline RAD51 protein levels, indicating SAHA can directly disrupt RAD51 responses to IR. In addition, our data also suggest that SAHA at high concentrations may affect other DNA repair pathways through transcription; however, the potential associated cytotoxicity may limit its clinical applicability.
A pharmacological-genetic epistasis analysis confirmed an inhibitory effect of SAHA on the HR pathway, as the SER for the drug was smaller in an HR-deficient line than for its congenic HR-proficient parental line. A remarkable finding of this work was the absence of increased radiosensitization in NHEJ-deficient cells. This result indicates that the cognate DNA lesions in irradiated cells are somehow not substrates for that repair process, despite its dominant role in DSB rejoining in mammalian cells throughout the cell cycle when SAHA is absent. There is currently little understanding of the manner by which cells “decide” whether a given DSB will repaired via NHEJ versus HR. Perhaps, as a consequence of SAHA acting on the cellular “decision machinery,” a subset of DSB becomes inaccessible to repair via NHEJ. Because the alternative HR pathway is inhibited as an independent effect of the drug, however, cell survival is compromised. We are currently investigating this and other possibilities.
We also found that SAHA might enhance the proteasome-mediated degradation of RAD51 protein in irradiated MM cells. Of note, a previous study showed that MG132, a proteosome inhibitor, at 10 μmol/L concentration suppressed HR and RAD51 foci formation, without changes of RAD51 protein level in irradiated murine embryonic fibroblast cells (51). However, we found that MG132 at 500 nmol/L could restore SAHA-inhibited RAD51 protein level and its foci formation after IR. The differences in the effects observed for RAD51 response to IR may be dependent on concentrations of MG132. These data predict that combining SAHA with the proteasome inhibitor bortezomib, may reverse the inhibitory effect of SAHA on HR in response to radiotherapy, but more work needs to be done to confirm this prediction.
The potential of using SAHA as a radiosensitizer in MM treatment through inhibition of HR repair is intriguing, particularly because MM is a radioresponsive disease, radiotherapy plays an important role, and with current advanced technologies in radiotherapy, delivery of dose through allows for the delivery of precisely focused radiation to the major marrow sites where the MM cancer cells most reside. SAHA may also have potential in sensitizing cells to certain chemotherapy agents used in MM, such as Melphalan. Melphalan is an alkylator which is known to be preferentially toxic in cells harboring impaired HR (37).
We and others have been actively evaluating TMI in MM. Using a TMI approach, radiation sensitizers take on greater importance, because dose deposition can now be controlled and redistributed preferentially to bone and marrow and away from normal organs, resulting in selective radiosensitization of bone and marrow compared with normal tissues. The benefits of a radiosensitization strategy with even modest effects will be further amplified because of selective targeting of dose and radiosensitization to marrow and other user-specified target structures. Low dose of SAHA may serve as an ideal radiosensitizer in TMI for MM. SAHA has also showed promise in acute leukemias, where TBI, and potentially TMI, have an important role in patients undergoing HCT. Future plans are to evaluate the effects of SAHA on radioresponse in acute leukemias.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: X. Chen, E.H. Radany, J.M. Stark, J.Y.C. Wong
Development of methodology: X. Chen, P. Wong, C. Laulier
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): X. Chen, P. Wong, J.Y.C. Wong
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Chen, P. Wong, E.H. Radany, C. Laulier, J.Y.C. Wong
Writing, review, and/or revision of the manuscript: X. Chen, P. Wong, E.H. Radany, J.M. Stark, J.Y.C. Wong
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. Chen, P. Wong, J.Y.C. Wong
Study supervision: J.Y.C. Wong
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.