Repression of BET activity sensitizes oral squamous cell carcinoma to PARP inhibition via inhibiting DNA homologous recombination and immune inhibition

Background: To investigate the root cause of therapeutic resistance underlying oral squamous cell carcinoma (OSCC) to the PARP inhibitor olaparib and explore the combination with the BET inhibitor JQ1 to augment its treatment effects. Material and Methods: Cell viability were detected in the study. We also evaluated the expression of a series of factors through quantitative real-time PCR, immunofluorescence and western blot. The effect of combining JQ1 and olaparib in xenograft OSCC mouse model was also examined. Results: The sensitivity of Cal27 cells to olaparib was better than Scc25. Functional assays demonstrated that olaparib induced HR repair and upregulated PD-L1 expression, which results in drug resistance of OSCC to olaparib. Variations of these factors in the two cell lines may explain different sensitivity to olaparib. Moreover, the JQ1 BET inhibitor and olaparib synergistically exhibited anti-cancer effects in OSCC in vitro and in vivo and inhibited essential HR repair factors RAD51, BRCA1, and TOPBP1 through the ATR/CHK1 pathway and immune suppression mediated by the PD-L1 pathway. Conclusions: Elevated HR and PD-L1 are involved in resistance mechanisms of OSCC to olaparib, attenuating its anti-tumor effects. Our results suggest that the strong synergistic anti-cancer activity of combining olaparib with BET inhibitor in OSCC may be via suppression of the ATR/CHK1-mediated DNA damage response and PD-L1-related immune escape, indicating this combination strategy as a possible therapy in OSCC. These results demonstrated a synergistic inhibitory effect of two drugs on tumor cell viability in OSCC. treatment, the S phase population also decreased as a result of JQ1-induced G1 phase arrest. Taken together, our data suggested that OSCC are sensitive to the BET and PARP inhibitor synergistic treatment both in cells and in mouse model. Our research is the first study to demonstrate the combination effect of BET suppression with PARP inhibition in OSCC.


Background
Oral squamous cell carcinoma (OSCC) is an usual cancer and affects nearly 500,000 patients annually worldwide. [1] Although recent advances in treatment methods of OSCC have generated promising avenues for therapeutic intervention, the average 5-year survival rate is still approximately 50%. [2]Therefore, more effective drugs are needed for OSCC patients.
Homologous recombination (HR) depending on sequence homology/identity, is a kind of the most accurate DNA repair mechanisms. [9,10] Previous studies showed that HR-deficient cancer cells are relatively more sensitive to the treatment of PARP inhibitors. RAD51, BRCA1 and TOPBP1 are three critical tumor suppressors crucial for DNA double strand break (DSB) repair and active the progression of cancers. [11,12] [13,14]Therefore finding strategies which could selectively suppress these factors in tumors, sensitizing HR-proficient tumors to olaparib and benefiting the patients.
Current research has manifested that conventional anticancer therapies influence tumor-targeting immune responses and indicated that PARP inhibitors are associated with immune escape, which has negative effects on durable responses in some tumors. [15][16][17]Thus, describing the crosstalk regarding anticancer treatments and drug-related immunity precisely could result in better combinatorial methods. PD-L1 taking part in inhibiting autoimmunity, and its overexpression is associated with poor prognosis in many kinds of cancers. [18][19][20]Although PD-L1 is particularly important in cancer progression, it is hard to regulate its expression. Previous reports showed that there is an increase of PD-L1 as an reaction to DNA damage in some tumors, [21,22]which suggests that PD-L1 may also regulated by DNA damage drugs including PARP inhibitors, but whether it plays a similar role in OSCC is still unknown.
BET family proteins can associate with mitotic chromosomes and control some significant genes through attracting transcription factors. [23] BET bromodomain inhibitors including JQ1, exhibiting antitumor activities in a range of preclinical models of many tumors.
In our experiments, we detected the mechanisms of PARP inhibitor olaparib in OSCC and its resistance pathways and examined the potential benefit of combination treatment with the BET inhibitor JQ1.

Cell culture
OSCC cells were provided by Shanghai Ninth People's Hospital, culturing in high glucose DMEM (Hyclone, Logan, UT, USA) supplemented with 10% fetal bovine serum at 37°C in a humidified incubator containing 5% CO 2 .

Colony formation assay
OSCC cells were cultured in 6-well plates at a density of 2000 cells/ well, and then treated with 5 µM olaparib ± 0.2 µM JQ1, While treating Scc25 cells with 20 µM olaparib ± 1 µM JQ1, as our results showed that Scc25 cells were less sensitive to olaparib. After 48 h of culture, cells were fixed, with staining, photographing and counting subsequently.

Cell viability assays
Cell proliferation was detected with CCK-8 (Dojindo, Japan). Cells were cultured and treated with olaparib with or without JQ1 for 1-4 d. Next, adding CCK-8 solution and incubating for o.5-1 h, after which measuring densities, at 450 nm. The half maximal inhibitory concentration (IC50) was calculated by GraphPad Prism software (GraphPad Software, La Jolla, CA, USA) and combination index (CI) was acquired by CalcuSyn software.

Apoptosis analysis
OSCC cells were cultured at 5 × 10 5 cells/well and treated with olaparib, JQ1 or the combination for 2 d. After staining with PE (BD Biosciences, San Jose, CA, USA), the proportion of apoptotic Cal27 and Scc25 cells were examined by FACS (BD Biosciences, Franklin Lakes, NJ, USA).

Cell cycle analysis
OSCC cells were seeded and treated by different inhibitors. After 48 h, staining cells with propidium iodide (PI, BD Biosciences) and measuring cell cycle with a flow cytometer.

Immunofluorescence
OSCC cells were dealt with olaparib and JQ1 alone or in combination, and then cells were fixed with 4% paraformaldehyde and 0.1% Triton X-100. Staining cells with rabbit anti-human RAD51 antibody (1:250) overnight. After that cells were analyzed and images were obtained with a Leica fluorescence microscope.

Quantitative RT-PCR (qRT-PCR)
OSCC cells were cultured and treated with olaparib, JQ1 or the combination for 48 h. Extracting RNA with TRIzol reagent (TaKaRa Biotech, Shiga, Japan) and reversing transcribed into cDNA with Reverse Transcriptase kit (TaKaRa Biotech). qRT-PCR reactions were performed using the SYBR® Premix Ex Taq kit (TaKaRa Biotech) on a Roche 480 LightCycler (Basel, Switzerland). GAPDH mRNA served as the reference.

Western blotting analysis
After treating, cells were lysed in lysis buffer containing 1% PMSF (Beyotime, Shanghai, China) for 30 min. Equal amounts of proteins were separated, with transferring to a 0.45 µm membrane (GE HealthCare Life Science, Germany) for 1 h at 100 V. Membranes were incubated in 3% nonfat milk for 1 h and then were incubated overnight at 4°C with primary antibodies. Membranes were then incubated with secondary antibodies for 1 h. Immunoreactive bands were visualized by ECL detection reagents.

Xenograft OSCC mouse models
Six-week-old immunodeficient mice were purchased from Vital River Laboratories (Beijing, China). All mice were handled according to protocols approved by the ethics committee of Shandong University. For experiments, 1 × 10 7 Cal27 cells were combined with Matrigel (Corning, NY, USA) injecting at the right flanks of mice. Mice were randomized into four groups and treated as follows when volume of tumors reached about 150-250 mm 3 , approximately one week later: control with vehicle (10% captisol), JQ1 (20 mg/kg), olaparib (50 mg/kg) or the combination (n=4 per group). The gap time was two days between every injections and injected for 3 times in total. At the end of the experiments, dissecting tumors and analyzing.

PARP inhibitor olaparib upregulates HR and PD-L1 in OSCC cells
The IC50 values of olaparib in Cal27 and Scc25 cells were 36 µM and 98 µM, respectively, indicating the sensitivity of Cal27 cells to drug stronger than Scc25 cells. (Figure_2a)Then the CCK8 assays also verified that hypothesis. (Figure_1a)To examine the possible role of HR in reaction to olaparib, treating cells with olaparib at various concentrations and evaluated levels of RAD51, BRCA1, TOPBP1, H2AX and phosphorylated γ-H2AX. In Cal27 cells, 5 µM olaparib treatment resulted in increased RAD51, BRCA1, and TOPBP1 and decreased γ-H2AX phosphorylation, H2AX. Similar results were observed in response to 10 µM, although to a lesser extent compared with 5 µM olaparib. In Scc25 cells, olaparib at 10 µM or 20 µM upregulated the expression of RAD51, BRCA1, TOPBP1 and γ-H2AX. (Figure_1b) To further explore the mechanism of olaparib resistance in OSCC, we examined the PD-L1 expression. PD-L1 was consistent with HR-related proteins (RAD51, BRCA1, TOPBP1, γ-H2AX). (Figure_1b) We also examined the mRNA and protein expressions of HR genes RAD51 and CD274. (Figure_1c) Moreover, BRD4 expression was also upregulated by olaparib, which help explain the relationality of using BET inhibitor to augment treatment effect.
Together these results indicated that the PARP inhibitor olaparib induced the HR-mediated DNA damage response as well as high PD-L1 expression.

BET inhibitor and PARP inhibitor synergistically suppress cell proliferation
We next examined the effects of treating cells with olaparib, the BET inhibitor JQ1, or combining both inhibitors. Both olaparib and JQ1 at low concentrations had moderate impact on decreasing cell proliferation and viability of Cal27 and Scc25 cells when they were used alone. However, olaparib together with JQ1 greatly triggered cell death. The IC50 values of olaparib were significantly reduced after combination treatment (**P < 0.01) (Figure_2a), indicating that OSCC cells were more sensitive to olaparib. CI calculations confirmed that the combination JQ1 and olaparib treatment exhibited a synergistic effect ( Table 1).
The JQ1 and olaparib combination treatment also caused colony formation inhibition compared with other groups. (Figure_2b) The action of the combined treatment on proliferation was tested using CCK8 assays. The results showed that combination of two drugs cause a remarkable decline in cells' proliferation, comparing with individual inhibitors (**P < 0.01, Figure_2c).
These results demonstrated a synergistic inhibitory effect of two drugs on tumor cell viability in OSCC.

BET inhibitor and PARP inhibitor synergistically induce apoptosis, which correlates with higher p53 levels
We next examined whether the combination treatment affected the apoptosis of OSCC cells. Western blot analysis uncovered that the cleaved caspase-3 which is an important apoptosis marker was activated at a dramatically higher level in Cal27 and Scc25 cells after treating with JQ1 and olaparib. (**P < 0.01, Figure_3a, 3b)Our research also evaluated cell apoptosis by Annexin V-PE staining followed by flow cytometry in OSCC cells treated with variable inhibitors for 2 days. Apoptotic cells accounted for higher proportions in response to the combination treatment than treatment by either inhibitor alone. (**P < 0.01, Figure_3c, 3d) P53 signaling is a key pathway that induces apoptosis. We thus next examined whether apoptosis was mediated by the p53 pathway. The combination treatment led to prominently increased p53 and phosphorylated p53 expressions compared with single treatments. (*P < 0.05, **P < 0.01, Figure_3a, 3b) Together our findings suggest that JQ1 and olaparib exhibit synergistic effects on suppressing viability of OSCC cells, which correlated with activation of the p53 pathway.

BET inhibitor and PARP inhibitor co-treatment results in cell cycle arrest
Then we focused on the effects of the combination treatment on cell cycle in OSCC cells. The results are displayed in Figure. 4a and quantitative analysis is shown in Figure. 4b. JQ1 and olaparib alone resulted in G1 or G2/M phase cell cycle arrest, respectively, while the combination treatment caused G1 and G2/M phase arrest as well as a significant decline in S phase cells.

BET inhibitor and PARP inhibitor synergistically suppress the growth of tumors in vivo
Cal27 cells were injected into Six-week-old female mice and after the tumor volume reached approximately 150-250 mm 3 , mice were randomized into groups and treated with different inhibitors. The combination of JQ1 and olaparib potently inhibited tumor growth including tumor volume and tumor weight, comparing with single inhibitors used alone (**P < 0.01, Figure_5a, 5b, 5c). The results demonstrated that combined treatment exhibits better antitumor effects in OSCC in vivo compared with JQ1 or olaparib treatment alone.

BET inhibitor and PARP inhibitor combination suppresses HR mediated by ATR/CHK1 pathway
To explore the potential mechanism of the enhanced efficacy with combined treatment in OSCC, we examined several key DNA damage response factors involved in HR. Western blot and Immunofluorescence assay showed that these HR-related molecules were highly downregulated by the combination treatment compared with olaparib alone, although the expressions were slightly higher than in response to JQ1 alone. (Figure_6a, 6c) We also observed decreased RAD51 mRNA level in Cal27 and Scc25 cells treated with the combination therapy compared with olaparib monotherapy (**P < 0.01, Figure_6b), which verified that JQ1 combined with olaparib can act by suppressing HR.
ATR/CHK1 plays a critical role in the regulation of replication and is essential in HR, and thus we subsequently examined the ATR/CHK1 pathway. Consistent with our above results, the ATR/CHK1 pathway was increased in response to olaparib compared with controls, while the combination treatment markedly blocked this pathway compared with olaparib alone (**P < 0.01, Figure_6a, 6b), further indicating that JQ1 maximized olaparib's effect through impairing HR mediated by ATR/CHK1 pathway inhibition.
These results indicate that the BET inhibitor JQ1 disrupted HR via inducing ATR/CHK1 inhibition, thus sensitizing the OSCC cell response to the PARP inhibitor.

BET inhibitor and PARP inhibitor combination attenuated expression levels of BRD4 and PD-L1
We continued to explore the mechanism behind the combinatorial synergism of JQ1 and olaparib in OSCC cells. The combination of JQ1 and olaparib reduced BRD4 expression to a greater extent compared with olaparib or JQ1 monotherapy in Cal27 and Scc25 cells, which verified that the sensitivity of OSCC to the BET inhibitor was also enhanced by the combination treatment. (*P < 0.05, **P < 0.01) (Figure_7a, 7b, 7c) We next determined whether BRD4 correlates with PD-L1 expression in OSCC by examining a panel of OSCC cell lines. We observed a trend towards a positive relationship between BRD4 and PD-L1 expression. (Figure_7a, 7b) We then examined whether BRD4 and CD274 expression was also positively correlated in OSCC cells. Indeed, there was a significant positive mRNA correlation between BRD4 and CD274 in Cal27 and Scc25 cells treated with JQ1, olaparib or the combination. (Figure_7c) Notably, decreasing PD-L1 expression may be a strategy to optimize the antitumor effects of PARP inhibitors by adding BET inhibition.
Our results showed that the PARP inhibitor upregulated PD-L1 and BRD4 in OSCC cells, while BET blockade attenuated PD-L1-mediated immunosuppression activity, resulting in better treatment effect in OSCC cells. These results verified that the sensitivity of OSCC to the BET inhibitor was enhanced by the combination treatment with the PARP inhibitor.

Discussion
Combination of targeted therapies to augment treatment results is an effective therapeutic strategy for various cancers, including OSCC. PARP inhibitors have been the focus of intense investigation in cancer treatment, particularly as monotherapies or combining with conventional chemotherapeutics regarding DNA damage. [24,25]Our results showed that the effects of PARP inhibitor olaparib in OSCC was moderate both in vitro and in vivo and olaparib reduced levels of the DNA damage marker γ-H2AX in Cal27 cells. Identification of the mechanisms underlying resistance to olaparib and developing combination treatments that can produce potent responses in a preclinical setting of OSCC is required.
The DNA damage response exists in resistance mechanisms to anticancer treatment in many tumors. DNA damage response can be divided into two major pathways: the HR pathway for DSB repair and the nucleotide excision repair pathway for single-strand break repair. [26,27]The ATR-CHK1 pathway is a central signal transduction processes that facilitates DNA repair and promotes cell proliferation. Activation of the ATR-CHK1 pathway can cause cell cycle arrest to allow cellular repair or induce cell death when the damage is irreparable. [28]In our study, the PARP inhibitor olaparib induced levels of several important HR-related factors in Cal27 and Scc25 cells, indicating that HR is induced in response to olaparib. We further found that olaparib upregulates the expression of the ATR/CHK1 pathway, which may contribute to HR repair in OSCC cells. Thus, we propose that olaparib activates the ATR/CHK1 pathway to induce HR in OSCC.
In recent years, accumulating evidence has highlighted the role of immunity in patient response to anticancer agents. [29,30]Our results demonstrated that the PARP inhibitor exhibited low levels of tumor inhibition effects on OSCC, which may be due to the induction of HR-mediated DNA damage response. However, the role of PARP inhibition in cancer-associated immunity has been largely unknown. Our findings showed that treatment with the PARP inhibitor olaparib alone upregulated PD-L1 expression in OSCC cells, which may contribute to the reduced efficacy of PARP inhibitor through inducing immune escape. Thus, strategies that can block the PD-L1 pathway to restore anti-tumor immunity may help potentiate the effects of PARP inhibitors in tumor suppression. We also found that Cal27 OSCC cells were more sensitive to olaparib than Scc25 cells, which may be partly due to different levels of HR repair and PD-L1 expressions, as olaparib induced higher activation of these factors in Scc25 cells.
A number of BET inhibitors have been developed and are entering clinical trials. After wide range of drug screening in preliminary analyses, selecting the JQ1 small molecule inhibitor of BRD4 because JQ1 treatment alone can decrease the DNA damage response and PD-L1 expression. Tumor cell viability fell down dramatically via the addition of JQ1 to olaparib, manifesting that JQ1 play a role in sensitizing Cal27 and Scc25 cells to the PARP inhibitor. We also verified that the synergistic effects of the combination treatment are partly attributable to the promotion of apoptosis mediated by p38 pathway. Notably, while the olaparib-induced G2/M arrest was reduced in Scc25 cells upon co-treatment with JQ1 treatment, the S phase population also decreased as a result of JQ1-induced G1 phase arrest. Taken together, our data suggested that OSCC are sensitive to the BET and PARP inhibitor synergistic treatment both in cells and in mouse model. Our research is the first study to demonstrate the combination effect of BET suppression with PARP inhibition in OSCC.
We explored the mechanism underlying the synergistic treatment in OSCC. We found that JQ1 and olaparib exhibited anti-cancer activities in OSCC cells by inhibiting HR repair induced by the ATR/CHK1 pathway and immune escape mediated by the PD-L1 pathway. Moreover, olaparib alone upregulated BRD4 expression, which may help clarify the synergism of BET and PARP inhibitors. Further studies of the molecular consequences of the combination treatment of BET and PARP inhibitors in OSCC will be necessary to develop effective combinations with optimal therapeutic potential.

Conclusions
Above all, our research verified that HR repair and upregulated PD-L1 expression in Cal27 and Scc25 cells resulted in adaptive cellular response to the PARP inhibitor olaparib. Adding BET inhibitor effectively attenuates the ATR/CHK1-mediated DNA damage response and PD-L1-related immune escape, increasing sensitivity of OSCC to PARP inhibitors in vitro and in vivo. Our results provide evidence for using BET inhibitors to enhance therapeutic efficacy of PARP inhibitors to HR-proficient tumor cells. (Figure_8) The relationships between PARP inhibitors, HR and PD-L1 is timely and provides strong rational to explore more effective method consisting of PARP and BET inhibitors.
Together, this work provides insights on the mechanisms of acquired resistance to olaparib and explores methods in order to augment the therapeutic strategies of PARP inhibitors. These results suggest a potential clinical benefit of the combination strategy for OSCC patients and may have far-reaching implications for OSCC patients in the future.

Ethics approval and consent to participate
Not applicable

Conflict of interest
The authors have no potential conflicts of interest in the work.

Authors' contributions
Xiaojing Liu participated in the conception and design of the study. Xin Zhang contributed the cell culture and treatment for the study. Fenghe Zhang performed the data collection and analysis. All authors read and approved the final manuscript.

Funding
This research was supported by natural science foundation of Shandong province (ZR2018MH023).

Data Availability Statement
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Abbreviations
OSCC oral squamous cell carcinoma HR Homologous recombination DSB DNA double strand break ATR Ataxia telangiectasia and Rad3-related PARP poly ADP-ribose polymerase Table   Table 1: Synergistic induction of cell death by olaparib and JQ1.   BET inhibitor and PARP inhibitor co-treatment causes cell cycle arrest Cal27 and SCC25 cells were cultured and then treated with JQ1, olaparib or the combination for 2 d. (a), Cells stained with PI and examined with flow cytometry assays to analyze cell cycle distribution. (b), Bar graphs showed the percentage of Cal27 and SCC25 cells in each cell cycle phase after treatment with JQ1 and/or olaparib. JQ1 and olaparib alone caused G1 and G2/M phase cell cycle arrest, respectively, while the combination treatment resulted in G1 and G2/M phase arrest and a profound reduction in S phase cells. Bars represent the mean ± SD of each group (n=6).  BET inhibitor and PARP inhibitor combination suppresses HR mediated by ATR/CHK1 pathway Cal27 and SCC25 cells were cultured and then treated with JQ1, olaparib or the combination for 2 d. (a), Relative protein expressions of RAD51, BRCA1, TOPBP1, ATR, CHK1 and p-CHK1 in Cal27 and SCC25 cells. GAPDH was used as a loading control. (b), Relative mRNA expressions of RAD51 and ATR in Cal27 and Scc25 cells. (c), Immunofluorescent staining showed that combination treatment suppresses the expression of RAD51, compared with JQ1 or olaparib were used alone. Bars represent the mean ± SD of each group (n=6). Significant differences between groups are marked with asterisks. (*P<0.05 ** P<0.01) The mechanism figure of BET inhibitor augments treatment effects of PARP inhibitor in OSCC. Though PARP inhibitor has positive effect on p53-medicated cell death or cell cycle arrest, it also upregulates HR-related DNA damage response and PD-L1-medicated immune escape which weakens treatment results in OSCC. Otherwise, BET inhibitor not only increases cell cycle arrest and induces cell death through p53 pathway, it also suppresses the expressions of HR and PD-L1, and so it provides scientific evidence to enhance the therapeutic potential of PARP inhibitor.