2-hexyl-4-pentynoic acid (HPTA), a novel radiosensitizer to breast cancer cells through increasing the instability of DNA repair proteins

MTT, clonogenic survival assays, comet assays, immunouorescence and western blot analyses were used to detect the effect of VPA / HPTA on DNA damage induced by radiotherapy for breast cancer through a variety of cell models( MCF7, EUFA423, HCC1937, DMBA-induced rat breast cancer-derived primary culture cell and DMBA-induced transformed human normal breast cell line). At the same time, ow cytometry, immunouorescence and western blot analyses were used to investigate the effect of VPA / HPTA on DNA damage repair induced by radiation. In vivo experiment, the effect of HPTA as radiosensitizer was investigated by DMBA-induced breast cancer in rats. Finally, the possible mechanism of HPTA acting on target was by cycloheximide chase experiment. a rational strategy to improve the radiotherapeutic ecacy of breast cancer.


Introduction
Breast cancer is one of the most common malignant tumors in the world, with more than one million new cases every year and is the main cause of cancer death for women [1,2]. Treatment for breast cancer -surgery, chemotherapy, and radiotherapy -have signi cantly improved the prognosis of patients and reduced the mortality of breast cancer [3]. Radiotherapy for breast cancer includes adjuvant radiotherapy after breast conserving surgery or after mastectomy, for local recurrence and distant metastasis (e.g. brain and bone). Radiotherapy has been shown to improved survival rate post-mastectomy, reduction in locoregional recurrent rate and ipsilateral breast cancer recurrences [4]. However, radiation to the breast has the potential to cause severe complications due to incidental dosage to heart, lungs and contralateral breast [5], incidence and severity of adverse events is dose and volume dependent [6]. All of these are considered the causes of treatment failure. Thus, acquiring a deeper understanding of the mechanisms related to radiotherapy and identifying a novel and suitable radiosensitizer are crucial events for improving the survival of breast cancer patients.
Valproic acid (VPA) is used in the treatment of epilepsy since the early 1970s [7][8][9][10]. In addition to its antiepileptic effect, VPA has been found to induce tumor cell death, inhibit tumor invasion and metastasis, and inhibit angiogeneis in various cancer models. More recent studies have also found VPA to increase the radiosensitivity of solid malignant tumor cells, including esophagus cancer [11], lung cancer [12], prostate cancer [13], and colon cancer [14]. Our previous studies have demonstrated that 500 µM VPA increases the sensitivity of osteosarcoma U2OS cells, breast cancer MCF7 cells and breast cancer tissue-derived primary cells to radiation therapy [15,16]. The mechanism involves inhibiting BRCA1-Rad51-mediated homologous recombination (HR) and Ku80-mediated non-homologous end joining (NHEJ) repairs, resulting in DSBs accumulation and tumor cell death. A high dose (500 µM) is required for VPA's radiosensitization effects [16], this increases the likelihood of adverse side effects such as hepatotoxicity, nephrotoxicity, and teratogenicity [17][18][19]. It is therefore imperative to identify alternatives with low toxicity and high potency.
In this study, we discovered for the rst time that 15 µM HPTA has a similar radiosensitization effect on breast tumor cells as 500 µM VPA. The mechanism involves disrupting the DNA repair pathway by reducing the half-life of both DNA repair proteins BRCA1 and Rad51. In conclusion, we reveal that HPTA may be an e cacious sensitizer for breast cancer radiotherapy, indicating that BRCA1 and Rad51 may be the promising targets for breast cancer radiotherapy.

Immunoblotting
The protocols were described in our previous publication [24]. The primary antibodies include: γH 2 AX

DSB assay
For HR assay, the MCF7 cell line expressing a recombination substrate of the pDR-GFP reporter was studied to determine the number of GFP-positive cells in the treated cells for the analysis of the spontaneous HR frequency using ow cytometry as described previously [15,16]. For NHEJ assay, the U2OS cell line expressing an End Joining reporter (EJ5-GFP) was studied to determine the number of GFP-positive cells in the treated cells for the analysis of the spontaneous NHEJ frequency using ow cytometry as described previously [25].
Animal keeping and the establishment of breast cancer model Female Sprague-Dawley (SD) rats were purchased from Pengyue Laboratory Animal Co. Ltd. Jinan, China. The studies of animal tissue were performed in adherence to national standards on animal care, speci cially the Shandong University Human and Animal Ethics Research Committee's requirements (project identi cation code 81472800, approved on 3 March 2014). All rats were housed in a speci cpathogen-free environment, at a temperature of 23 ± 1 °C. The lights were at a daily rhythm of 12 hrs and the rats were fed fresh food and water ad libitum throughout the experiment. The DMBA (Sigma) was dissolved in puri ed corn oil and adjusted to the concentration of 20 mg/ml. Intragastric gavage (i.g.) was performed on 50-day-old SD rats at a single dose of 1 ml DMBA-oil solution [26,27]. At 40-60 days after gavage, primary tumors could be detected through palpation around the breast.

Tumor observation on rats
For a period of time after DMBA gavage, rats with appropriate tumor size and location were selected for grouping and then treated. The tumor size of each group was recorded and measured with a vernier caliper. Tumor volume was calculated according to the clinical standard formula "Volume V (mm 3 ) = Length (L) * Width (W) 2 * 0.5".

HE and immunohistochemistry (IHC) staining
The rats were given chloral hydrate (Sigma) for the anesthesia and breast tissues and cancer tissues were isolated, then xed overnight in 4% paraformaldehyde solution, embedded in para n and serially sectioned 5 µm thick for HE and IHC staining.
The sections for IHC staining were dewaxed with xylene and hydrated with 100-75% ethanol. And then boiled in a 10 mM sodium citrate solution above 92 °C for 20 minutes for antigen retrieval. The sections were cooled to room temperature and then washed with TBS. After washing with TBS, incubate the sections in 3% hydrogen peroxide for 15 minutes to remove endogenous peroxidase and washed again with TBS. Incubate with 10% goat serum for 1 hour to block non-speci c binding, then dilute γH 2 AX (1:500,05-636, EMD Millipore) and Rad51 (1:500; PC130, Calbiochem)with solution (TBS + 1% BSA) primary antibody, and the sections were placed at 4 °C overnight. Secondary antibody was diluted by TBST (TBS + 0.25% Triton X-100 + 1% BSA): Biotinylated goat anti-mouse IgG (1:300, BA-9200, Vector), Biotinylated goat anti-rat IgG (1:300, BA-9400, Vector) and incubated sections for 1 hour at room temperature. Wash with TBST and immediately incubated in ABC working solution for 30 minutes, then test with DAB working solution. Subsequently, hematoxylin staining was carried out, and it was dehydrated with ethanol after washing, and then xed with xylene for 5 minutes, observed with neutral gum.
The IOD density of γH 2 AX and Rad51 staining by IHC staining was further quanti ed by Image pro plus software (Media Cybernetics).
Chemical carcinogen-induced rat breast cancer-derived primary cells The method was described in our previous publication [28]. After establishment of the breast cancer model, the rats were given chloral hydrate (Sigma) for the anesthesia and breast cancer tissues were isolated in a sterile manner. The breast cancer tissues were further cut into small pieces (2 × 2 mm) and placed on the coated dishes for primary cell culture. The cells from the second and third passage were used for the immuno uorescence analysis of γH2AX and 53BP1.

Soft agar colony formation assay
For the bottom layer, 1.5 ml/well 0.6% low-melting agar (214230, Difco Agar Nobel) was added into a 6well plate for solidi cation. The top layer contained cell suspension (1 × 104/well) and 0.3% agar. The cells were cultured for 3-4 weeks until visible colony formed, these were stained with 0.005% crystal violet.

RI-1 assay
Cells were seeded onto 60 mm dishes with cell culture media supplemented with 10 µM Rad51 inhibitor 1 (RI-1, S8077, Slleck ) pretreated for 24hrs. Cells were harvested at indicated time points, and lysates were subjected to Western blotting experiments to analyze protein expression at indicated time points. Image J software was used to quantify the relative expression level compared with control treatment.

Cycloheximide Chase Experiment
Cells were seeded onto 60 mm dishes with cell culture media supplemented with 40 µg/mL cycloheximide (CHX, HY-12320, MedChemExpress ). Cells were harvested at indicated time points, and lysates were subjected to Western blotting experiments to analyze protein expression at indicated time points. Image J software was used to quantify the relative expression level compared with control treatment.

Statistical analysis
All statistical analyses were performed with Student's t-test on SPSS Statistics for Windows, Version 23.0 (Armonk, NY: IBM Corp; licensed to Shandong University). and represented as mean ± SD. The p values were designated as: *, P < 0.05; **, P < 0.01, indicating a statistically signi cant difference.

Results
HPTA results in the accumulation of more DSBs in response to radiation Our previous research showed that 500 µM VPA inhibits the proliferation of multiple breast cancer cells and increases the sensitivity of cells to radiation [16]. To explore whether HPTA has a similar effect on breast cancer cells with VPA, we performed MTT experiments using HPTA in MCF7 cells. Accordingly, we set three concentrations: 15 µM, 30 µM, and 60 µM HPTA against 500 µM VPA (positive control) and subject to 4 Gy of ionizing radiation (IR). We found that cell proliferation in the VPA-/HPTA-treatment groups was inhibited after IR treatment as anticipated (P < 0.01) (Figure. 1B). There were no statistical differences between the three different concentrations of HPTA as compared to 500 µM VPA, indicating that HPTA at a low concentration (15 µM) possesses similar radiosensitizing effect as 500 µM VPA. Based on the IC 50 of VPA and HPTA, we opted for 15 µM HPTA for further experiments.
We next performed a clonogenic assay on the MCF7 cells ( Figure. 1C). To better compare the effects of HPTA and VPA, we added 15 µM VPA as a negative control. At the same time, to observe cell survival at different IR doses, we also included t 2, 4, and 6 Gy. The results showed that there was no statistical difference between VPA (15 µM or 500 µM) and HPTA treatment (15 µM) in the unirradiated control group ( Figure. 1D upper, P > 0.05). HPTA-treated cells were more sensitive to IR as compared to the untreated control group (Figures. 1C & 1D, P < 0.05). But VPA was more sensitive only at 500 µM, and there was no difference between the untreated control group and the VPA at a low concentration of 15 µM, which also proved that HPTA was more effective. Similar results were observed in another breast cancer cell line, EUFA423 (Figure. S1A&1B). These data indicate that the 15 µM HPTA has similar radiosensitizing effects as 500 µM VPA in vitro.
Our previous research [16] has shown that 500 µM VPA could cause signi cant DSBs following IR treatment, to explore whether the low concentration of HPTA would result in similar DSBs after IR treatment, DSBs were evaluated by two experiments using comet and immuno uorescence assays. To maximize the DSBs in cells, we set the IR dose at 8 Gy. In MCF7 cells, the alkaline comet assay showed no signi cant change in the VPA-or HPTA-treatment alone compared with the untreated control group pre-IR (P > 0.05; Figure. 1E). At 0 min post-IR treatment, the relative DSBs in the 500 µM VPA (70.04) and 15 µM HPTA treatment group (69.06) were signi cantly higher than that of the control group (52.43) (P < 0.01; Fig. 1E), indicating that VPA or HPTA in combination with IR can further increase intracellular DSBs accumulation. At 30 min post-IR treatment, the comet tail length began to shorten, but in the VPA treatment group (48.52) and the HPTA treatment group (49.70), the comet tail length was still signi cantly longer than that of the IR treatment alone (36.79) (P < 0.01; Figure.  DNA DSBs biomarkers γH2AX and 53BP1 were employed to determine whether 15 µM HPTA can enhance IR-induced DSBs. Immuno uorescence assay showed that the proportion of MCF7 cells with γH2AX or 53BP1 foci formation in the VPA-/HPTA-treatment alone was not increased compared with the control group (P > 0.05; Figure. 1G). The proportion of MCF7 cells with γH2AX or 53BP1 foci increased signi cantly 6 hrs post-IR (81.37% and 87.34% respectively; Figure. 1H). The percentage of γH2AX or 53BP1 foci in the VPA-/HPTA-treatment group was further increased as compared to the control: 98.59%/97.50% (γH2AX) and 98.28%/98.21% (53BP1) (P < 0.01; Figure. 1H), indicating that 15 µM HPTA can generate a similar amount of DNA DSBs as 500 µM VPA. Next, we observed the clearance of γH2AX and 53BP1 foci. At 24 hr post-IR treatment, the γH2AX and 53BP1 foci disappeared signi cantly slower in the VPA-/HPTA-treatment group as compared to the control group (P < 0.01; Figure. 1H), indicating that 15 µM HPTA is equivalent to 500 µM VPA in affecting the ability of cells to repair DNA damage. We analyzed the above data by categorizing the cells containing γH2AX or 53BP1 foci into two groups based on the number of foci in each cell: small foci (n < 20) and big foci (n ≥ 20). The results in Figure. S2A showed that the proportion of cells containing big foci in the VPA-/HPTA-treatment group was signi cantly higher than that in the control group. We next detect the protein levels of γH2AX and 53BP1 by an immunoblotting assay ( Figure. 1I). The levels of γH2AX and 53BP1 were increased by IR treatment, and further increased in the VPA-/HPTA treatment group (P < 0.01; Figure. 1L). The results from the above experiments indicate that 15 µM HPTA enhances IR-induced DSBs, similar to 500 µM VPA. We further veri ed the above results in the EUFA423 cells ( Figure. S1E-1H).
In summary, in both MCF7 and EUFA423 cells, our data demonstrate that 15 µM HPTA can lead to a similar extent of DSBs as 500 µM VPA in breast cancer cells after IR treatment.
Low dose HPTA dysregulates DNA repair pathway in response to IR-induced DSBs We next tested the DSB repair process in breast cancer cells to determine the likely mechanism of HPTA's action.
Mammalian cells employ two major DNA repair pathways: homologous recombination (HR) and nonhomologous end-joining (NHEJ), to maintain cell survival [29][30][31]. Error-free HR repair requires a homologous template such as a sister chromatid, whereas NHEJ joins the two ends of a DSB through a process largely independent of homology [32]. Our previous research also showed that 500 µM VPA reduces HR e ciency [16], so we next explore whether 15 µM HPTA has the same effect on the frequency of HR in the above cell models. The MCF7 cells expressing the pDR-GFP recombination reporter for the HR frequency assay were subjected to ow-cytometry after the introduction of I-SceI-induced DSBs ( Figure. 2A). The HR frequency decreased in cells treated with 15 µM HPTA by 42.88% as compared to the cells without HPTA treatment (P < 0.01; Figure. 2B), indicating that low concentration of HPTA is su cient to dysregulate HR repair pathway.
Recombinase Rad51 plays a central role in the HR mechanism and our previous studies have also shown that the effect of VPA on the HR pathway is mediated by Rad51 [16]. Thus, we investigated whether HPTA in uences the Rad51-mediated HR pathway. The results from the immuno uorescence assay showed that the percentage of the cells with Rad51 foci formation in the VPA-/HPTA-treatment group was almost the same as the control group pre-IR treatment (P > 0.05; Figure. 2C). At 6 hr post-IR treatment, the percentage of cells with Rad51 foci in the IR control group was 77.91%, and were reduced by 13.31% (P < 0.01) and 14.04% (P < 0.01) in the VPA-/HPTA-treatment groups ( Figure. 2C & 2D-upper). At 24 hr post-IR treatment, the Rad51 foci decreased as DNA repairs gradually completed, the VPA-/HPTA-treatment group still showed a signi cant decrease in Rad51 foci formation, a reduction by 10.39% (P < 0.01) and 9.65% (P < 0.01) respectively as compared with the control group (46.51%) ( Figure. 2C & 2D-upper). The results were veri ed by immunoblotting. At 6 hrs post-IR treatment, Rad51 protein level reduced signi cantly in both the VPA and HPTA treatment groups as compared to the IR treatment alone (P < 0.01; Figure. 2E & 2F). This indicated that 15 µM HPTA impaired Rad51 activity after IR, and the HPTA inhibited HR pathway is Rad51-dependent.
BRCA1 is another important protein that regulates HR via its interaction with Rad51 [33], so we next determined whether BRCA1 activity was in uenced by HPTA. Through immuno uorescence and immunoblotting experiments, we observed similar results as per Rad51( Figure.  To next investigate the possibility that HPTA may also in uence the NHEJ repair in our cell models, we used U2OS cells expressing the EJ5-GFP reporter to measure the NHEJ frequency [25,34] using owcytometry after the generation of I-SceI-induced DSBs (Figure. S4A). The NHEJ frequency decreased in cells treated with 15 µM HPTA by 19.48% as compared to the cells without HPTA treatment (P < 0.05; Figure. S4B), indicating that HPTA leads to the disruption of the NHEJ pathway.
A number of proteins are involved in the NHEJ repair pathway, such as DNA-PKcs, Ku70 and Ku80 [16]. Since our data demonstrated that HPTA had a suppressive effect on NHEJ, it would be reasonable to detect whether HPTA in uences the major NHEJ-associated proteins. We next employed immunoblotting assay to detect DNA-PKcs, Ku70 and Ku80 proteins in our cell models. The immunoblotting assay results in MCF7 cells showed that after 8 Gy IR, no signi cant changes in the three proteins were observed (P > 0.05; Figure.S4C & S4D). Similar results were also noted in the EUFA423 cell line (Figure. S4E). However, the expression of Ku70 and Ku80 proteins was decreased in the VPA treatment group as compared with the IR-control group (P < 0.05), but not in the HPTA treatment group (Figure. S4F). These results indicate that HPTA and VPA are different in the role of key proteins in NHEJ pathway. Through ow-cytometry, we found that HPTA could reduce the NHEJ e ciency (19.48%), even if it was not as obvious as the reduction in HR e ciency (42.88%), but the changes of several key proteins in different cells were not consistent, so we next focused on the effect of HPTA on HR.
HPTA exhibits radiosensitizing properties to IR treatment in DMBA-induced breast cancer in rats in vivo To study whether HPTA has a radiosensitization effect in vivo, we rst investigated the appropriate dose of HPTA on rat breast tumor. The primary model of transformed breast tumor cells in rats was induced by the environmental carcinogen 7,12-dimethylbenz[α]anthracene (DMBA). This was previously described and employed in related studies [28]. The experimental scheme is shown in Fig. 3A. In brief, around 40 days after DMBA gavage to female Sprague Dawley rats, lumps in breast sites could be found. The shape of lumps in the location of mammary glands was irregular ( Figure. 3B). By HE staining ( Figure. 3C), compared with normal breast tissue, abnormal hyperplasia, broadenoma and abnormal proliferation of epithelial cells of the breast were found, indicating that breast cancer in rats was successfully induced.
Reported studies of VPA on glioblastoma utilized intraperitoneal injection of VPA in the range from 150 mg/kg to 600 mg/kg [35]. Therefore, we choose 200 mg/kg as the treatment dose of VPA, which was equivalent to its dose of 500 µM used in vitro experiments.
However, the concentration of HPTA in rats has not been reported. In our pilot experiment, four concentrations: 5 mg/kg, 10 mg/kg, 20 mg/kg and 50 mg/kg were used. After intraperitoneal injection, there was no change in the 5 mg/kg dose group, while the 10 mg/kg, 20 mg/kg and 50 mg/kg concentration showed the same performance as that after the 200 mg/kg VPA injection, but only 20 mg/kg dose group was close to 200 mg/kg VPA, and the rest lasted for a shorter or longer time. Therefore, we opted for 20 mg/kg HPTA as the treatment dose for the study, this is equivalent to 15 µM used in the in vitro experiments.
To achieve better therapeutic effect, we next experimented with four fractionated doses of 2 Gy IR which more closely mirrored the clinical IR treatment used in breast cancer [36]. VPA or HPTA were administrated before, during, and after IR treatment (Figure. 3D). All the rats were alive over the 32 days observation period. As shown in Fig. 3E, at the second day post-IR treatment, IR induced a 21% decreased in breast cancer volume, and the addition of VPA or HPTA further reduced the cancer volume by 62% and 54%. During the 32 days of observation, compared with the IR treatment alone, the VPA-/ HTPA-treatment group signi cantly reduced the tumor volume (P < 0.05). The tumors in the IR treatment control group had recovered to the volume before IR, while the VPA-/ HPTA-treatment group had recovered to about half of the tumor volume before IR. On the 10th day after treatment, we excised the tumors under general anesthesia ( Figure. 3E-right), the tumor size in the VPA-/ HTPA-treatment groups were smaller than that in the IR treatment control group.
The morphological structure of tumors was observed by HE staining (Figure. 3F). VPA or HPTA treatment led to vacuole structures formation in the breast cancer tissue as compared with the untreated group; there were more vacuoles structures and number of necrotic cells after the IR treatment, and large necrotic areas and cells were seen in the tissues in both VPA and HPTA treatment groups. The morphological ndings are consistent with the above ndings. The results demonstrated that 20 mg/kg HPTA or 200 mg/kg VPA can effectively sensitize breast cancer to IR treatment.
Next, we explored whether DNA damage and repair proteins were in uenced by HTPA in vivo using the DSBs marker γH2AX. Whole-cell lysate extracted from the tumor tissue was analyzed for γH2AX using Western blot. The results showed that there were DSBs in both VPA and HPTA treatment groups ( Figure. 3G), and the concentration of γH2AX in both groups was higher than that in the untreated control group. The concentration of γH2AX post-IR treatment was signi cantly increased, signi cantly higher in VPA treatment group (P < 0.05) and no difference between the HPTA treatment group and the IR control group. We speculated that the reason why there was no obvious change in the tumor tissue of HPTA combination group might be related to the cell lysate containing the protein components of non-tumor cells (connective tissue cells, such as broblast, macrophages, etc.), so we further studied the expression of γH2AX by immunohistochemistry to verify this possibility. We found that compared with other groups, the amount of γH2AX in the tumor areas of the VPA and HPTA treatment groups increased signi cantly (P < 0.01; Figure. 3H). The result showed that more DSBs in tumor cells was induced by the combination of HPTA and IR. The level of Rad51 in the VPA-/ HPTA-treatment group was signi cantly lower than that in IR control group (P < 0.01; Figure. 3I). We also stained Rad51 with immunohistochemistry, and the results were consistent with those of Western blot ( Figure.  The role of HPTA on the radiosensitivity of DMBA-induced rat breast cancer-derived primary culture cells Based on the model of breast cancer in rats, we designed the experiment of breast cancer-derived primary culture cells according to Fig. 4A. Immuno uorescence assay showed that there were no signi cant changes in cells with γH2AX or 53BP1 foci formation in the VPA-/HPTA-treatment group compared with the untreated control group ( Figure. 4B & 4C). At 6 hr post-IR treatment, the positive rate of cells with γH2AX or 53BP1 foci formation in the IR control group (90.04%) increased signi cantly from baseline ( Figure. 4C). Importantly, the rate in the VPA-/HPTA-treatment group was further increased to 100% (P < 0.05; Figure. 4C), indicating that 15 µM HPTA induces DSBs in cells similar to 500 µM VPA. Next, we observed the clearance of γH2AX and 53BP1 foci. At 24 hr post-IR treatment, the γH2AX and 53BP1 foci in the IR treatment alone were signi cantly decreased than in the VPA-/HPTA-treatment group (P < 0.01; Figure. 4C), suggesting that HPTA signi cantly affect DNA repair activity in cells and delay DNA repair process. Consistent results were obtained from our further analysis (Figure. S2C). Together, the results supported our hypothesis that IR-HPTA combination promotes more DSBs accumulation in primary tumor cells.
The role of HPTA on the radiosensitivity of DMBA-induced transformed human normal breast cell line Next, we veri ed HPTA's radiosensitizing effect on DMBA-induced transformed MCF10A cells. DSBs in the tumor cells were observed by immuno uorescence assay using γH2AX and 53BP1 biomarkers. The percentage of cells containing γH2AX or 53BP1 foci signi cantly increased 6 hr after IR treatment (P < 0.01; Figure. 5C & 5D), the proportion was further increased by VPA-/ HPTA-treatment (P < 0.01). At 24 hr post IR treatment, the γH2AX and 53BP1 foci resolved signi cantly slower in the VPA-/HPTA-treatment group than in the IR control group (P < 0.01; Figure. 5C & 5D), indicating that 15 µM HPTA affects the ability of transformed cells to repair damaged DNA.
We detected the protein levels of γH2AX and 53BP1 protein by immunoblotting assay. Similarly, the expression of the two proteins in the VPA-/HPTA-treatment group was signi cantly higher than that of the IR control group (P < 0.05; Figure. 5E & 5F), which was consistent with the immuno uorescence results. In the untransformed MCF10A cells, no statistically signi cant differences between treatment groups were observed ( Figure. S5C -S5F).
To investigate whether HPTA-regulated radiosensitizing effect is associated with DNA repair function in the DMBA-transformed MCF10A cells, the activity of HR associated proteins Rad51 and BRCA1 were examined. The immuno uorescence assay showed that the number of cells containing Rad51 or BRCA1 foci was signi cantly decreased in the VPA-/HPTA-treatment group as compared to the IR control group at 6 hr and 24 hr post-IR (P < 0.01; Figure. 5G & 5H). We next detect the protein levels of the Rad51 and BRAC1 in the transformed cells. The protein levels of Rad51 and BRCA1 had decreased in the VPA-/HPTAtreatment group (P < 0.01; Figure. 5I & 5J) as compared to IR control group; though not in untransformed MCF10A cells ( Figure. S5G -S5J). These data further indicate that HR function was signi cantly inhibited in the VPA-/HPTA-treatment group in DMBA-transformed MCF10A cells, consistent with our earlier studies reported above.

HPTA-mediated radiosensitization to breast cancer cells is dependent on BRCA1 and Rad51 proteins
We used multiple breast cancer cell lines (MCF7 and EUFA423, primary cultured rat breast cancer cell lines, and DMBA-induced transformed human normal breast cell line) and breast cancer animal model, to investigate whether the radiosensitizability of lower concentration of HPTA (15 µM) is associated with BRCA1-Rad51-mediated HR activity.
The HCC1937 cell line expressing a defective BRCA1 gene was used to establish an isogenic wild type BRCA1 reconstituted cell line (wtBRCA1) (Figure. 6A). The clonogenic assay was employed to analyze the surviving fraction in HCC1937 cells expressing wtBRCA1 in response to IR. The VPA-/ HPTA-treatment group did not signi cantly decrease survival fraction (plating e ciency) in the paired cells as compared with untreated cells pre-IR ( Figure. 6B & 6C-up). The survival curve showed that HCC1937 cell expressing defective BRCA1 was more sensitive to IR as compared with BRCA1 cells ( Figure. 6B & 6C-down). Furthermore, the survival fraction in VPA-/HPTA-pretreated BRCA1 cells was signi cantly reduced after 4 Gy and 6 Gy IR, similar to BRCA1 defective cells in response to DNA damage ( Figure. 6C-down). The survival fraction in VPA-/HPTA-pretreated BRCA1-de cient cells only moderately decreased compared to wtBRCA1-cells after IR. These results indicated that HPTA treatment led to cell death by inhibiting BRCA1mediated HR repair pathway in response to IR.
To investigate whether Rad51 is directly involved in the observed HPTA's radiosensitizing effect, we used RI-1 [37], an inhibitor of Rad51, in the MCF7 cells. After treatment with 10 µM RI-1 for 24 hrs, the expression of Rad51 protein was signi cantly inhibited ( Figure. 6D). The clonogenic assay was employed to analyze the surviving fraction in MCF7 cells in response to IR. We noted that HPTA treatment alone did not decrease survival fraction (plating e ciency) as compared with untreated cells, irrespective of RI-1 treatment pre-IR ( Figure. 6E & 6F-up). The survival curve showed that MCF7 untreated with RI-1 was more sensitive to IR treatment as compared with cells treated with RI-1 ( Figure. 6E&6F-down). Furthermore, the survival fraction in VPA-/HPTA-pretreated MCF7 cells was signi cantly reduced after IR, similar to cells treated with RI-1 upon DNA damage, especially under 4 and 6 Gy treatment ( Figure. 6F-down). However, in RI-1-treated cells, the survival fraction in VPA-/HPTA-pretreated cells did not decrease compared to MCF7 untreated with RI-1 after IR. The results indicated that HPTA treatment, like and VPA, can lead to cell death via inhibiting Rad51-mediated HR repair pathway in response to IR.
In the established BRCA1 and Rad51 isogenic cell lines, we demonstrated that HPTA-and VPA-induced radiosensitization in breast cancer cells is associated with BRCA1-Rad51 mediated pathway.

HPTA increases the sensitivity of breast tumor cells to radiotherapy by inhibiting BRCA1 and Rad51
protein synthesis and shortening their half-life We next investigate how does HPTA affects sensitize tumor cells to IR. Cycloheximide (CHX) which hinders the translation process by interfering with shifting steps in protein synthesis, is often used to inhibit protein synthesis in eukaryotic cells [38][39][40][41]. We designed CHX-chase experiment to measure the half-life of the Rad51 and BRCA1 proteins in MCF7 cells. Immunoblotting was employed to analyze the expression of BRCA1 and Rad51 protein in the treated cells. Prior to IR, the HPTA treatment group and the negative control group showed an inverse association with the duration of CHX treatment and there was no signi cant difference between them (P > 0.05; Figure.  Collectively, HPTA inhibits the synthesis of both BRCA1 and Rad51, shortens their half-life and sensitizes tumor cells to IR.

Conclusion
In summary, we discovered for the rst time in this study that 15 µM HPTA has a similar radiosensitization effect on breast tumor cells as 500 µM VPA. The mechanism involves disrupting the HR repair pathway by inhititing both DNA repair proteins BRCA1 and Rad51, as proposed in Fig. 8. Given that HPTA can reduce the half-life of BRCA1 and Rad51 proteins, targeting BRCA1 and Rad51 may be a desirable therapeutic strategy for HPTA to enhance breast cancer radiotherapy.

Discussion
The overexpression of HDACs was extensively investigated in many cancers [42,43]. HDACis inhibit histone acetylation and result in transcriptionally inactive chromatin [26,44,45]. Increasing evidence show that many HDACis not only have potential anticancer effects at the cellular and animal level but also are radiosensitizers for some tumors, such as glioblastoma [14] and esophageal squamous cell carcinoma(ESCC) [46]. VPA, a HDACi, increases histone acetylation level, affects chromatin structure and gene regulation by inhibiting HDACs activity [42,44,47,48]. Our previous studies have shown that 500 µM VPA inhibits the proliferation of a variety of breast cancer cells and induce radiosensitivity, suggesting that VPA may be a promising anti-cancer drug and radiosensitizer [15,16]. However, a large dose (500 µM) of VPA is needed to render such anti-cancer and radiosensitization activities, other studies reported that high dose VPA can induce adverse events, such as liver, kidney toxicity and teratogenicity [17][18][19], therefore there is clinical interest to locate a suitable alternative.
Some scholars have studied and compared the effect of VPA and its derivatives on the survival of HepG2 hepatoma cells [49]. It was found that VPA derivatives inhibit the growth of tumor cells at 40 µM. HPTA is a VPA derivative and an effective HDAC inhibitor. Studies have shown that in cerebellar granule cells, HPTA at 5 µM was able to signi cantly increased acetylated histone H3 (ac-H3) level in a dose-dependent manner [6]. In comparison to VPA which at 100 µM increases the level of ac-H3 by 200%, HPTA at 50-100 µM increases the level of ac-H3 by 600% − 700% [6]. In this study,we demonstrated that HPTA had both anti-cancer and radiosensitizing effects on breast cancer, and that HTPA is more e cacious than VPA since a lower concentration of HPTA is needed to achieve effects as high concentration of VPA. The ndings of HPTA's radiosensitizing effect is novel.
A series of our previous studies other group studies have shown that the mechanism of VPA-induced radiosensitization involves the DNA damage repair mechanisms by inducing more DSBs damage to accumulate in the tumor cells, the broken DSBs are unable to be repaired effectively thus lead to the lethality of tumor cells from IR [15,16]. Our current ndings that HPTA acts in a similar mechanism of action to VPA by disrupting BRCA1-Rad51 mediated HR repair pathway ( Figure. 8).
For the NHEJ repair, whilst VPA can inhibit Ku70 and Ku80 proteins in EUFA423 cells similar to our previous report [16], HPTA affects the NHEJ independent of the major repair proteins such as Ku70, Ku80 and DNA PKcs in both MCF7 and EUFA423 cells. The data indicated that there is a different mechanism of action between HPTA and VPA in regulating DNA NHEJ repair function. It may be possible that HPTA in uences other NHEJ-related proteins and should be subject to further exploration.
We found that HPTA inhibits the synthesis of BRCA1 and Rad51 protein and shortens their half-life indicating that BRCA1 and Rad51 are the key target proteins in mediating HPTA radiosensitization. HPTA can induce the activation of 53BP1, which forms a barrier that inhibits DNA-end resection [50]. Although BRCA1 can promote DNA-end resection by relieving the 53BP1-dependent barrier [51], HPTA inhibits BRCA1 protein synthesis and promotes degradation, which is undoubtedly bad for HR repair. We therefore hypothesize that HPTA may lead to the failure of forming Rad51 nucleoprotein laments on damaged single-stranded DNA ends in time, since this process requires the presence of a large number of Rad51 proteins [52], and the homology search and strand exchange mediated by Rad51 nucleoprotein laments are key steps of the homologous recombination process [53]. The inhibition of Rad51 synthesis and the shortening of its half life cycle are fatal to the HR repair of tumor cells, and should be further investigated.
In summary, our work has bridged laboratory research and clinical practice: HPTA is a more e cacious radiosensitizer than VPA, and its clinical application warrants further investigation and promotion.     (upper) and 53BP1 (down) foci formation was evaluated. Each data point in the graph was from three independent experiments (mean SD); P-Values were calculated by t-test (*P<0.05, ** P<0.01).