Human papillomavirus E7 protein induces homologous recombination defects and PARPi sensitivity

Cervical tissues were collected from West China Second University Hospital, Sichuan University. HPV positivity and subtypes were determined by the hybridization capture using HC2 High-Risk HPV DNA Test (QIAGEN, 5199–1220). RLU/CO value > 1.2 were dragonized as high-risk HPV positivity. All sample collections were informed to patients and consented by them, and were carried out in accordance with the Ethics Guidelines and Regulation of the West China Second University Hospital, Sichuan University. Patients signed the informed consent and agreed to use it for scientific research.

HPV-negative cervical cancer cell lines (C33A) and human kidney epithelial cell line (HEK293T) were purchased from National Infrastructure of Cell Line Resource. RPE1-hTERT was gifted by the Genome Damage and Stability Centre, Sussex. Cells were cultured in DMEM (GIBCO) supplemented with 10% Fetal Bovine Serum (FBS, Hyclone) and 1%P/S (Penicillin and streptomycin, GIBCO). For transfection with Lipofectamine 3000® (Invitrogen, L3000015), cells were plated in 24-well plates and propagated to 80% confluency. Transfection was performed according to supplier’s instructions and cells harvested 72-h post-transfection.

The open reading frames of HPV16 E6 and E7 were inserted into the EcoRI and NotI site of pLVX-IRES-ZsGreen1 vector with a N-terminal HA-tag, using NEBuilder® HiFi DNA Assembly Master Mix (NEB, E2621).

Custom-made X-ray machine (Wandong Ltd, Beijing) was used to conduct ionizing radiation at 1 Gray/min and dosages were depicted in respective experiments. Calibration of radiation dosage was carried out by manufacture annually. 2 μM chemotherapy chemical (CPT) was added to cells with confluence of 60%–70%.

Cells grown on cover glasses were fixed with 4% paraformaldehyde (PFA), then used 0.3% Triton X-100 to permeabilize. Or in some experiments cells were fixed and permeabilized with methanol:ethanoic acid (3:1) for 15 min. Cells were blocked for 30 min at 37 °C with 3% BSA in PBS supplemented with 0.2% Triton X-100. Blocked cells were then incubated for 40 min in 37 °C with diluted primary antibodies: Mouse anti-γH2AX (Millipore, 05–636), Mouse anti-ATM-pS1981 (Cell Signaling, 4526S), γH2AX Rabbit anti-53BP1 (Cell Signaling, 3428P), Rabbit anti-BRCA1-pS1524 (Bethyl, A300-001A), Rabbit anti-RAD51 (Santa, sc-8349). After washed by PBS, cover glasses were incubated for 40 min in 37 °C with secondary antibodies: Rabbit IgG F(ab′)₂ fragment-Cy3 (Sigma) and goat anti-Mouse IgG-FITC antibody (Santa). After mounting with DAPI (VECTOR, H-1200), pictures were captured using an Olympus fluorescence microscope (BX 51) and analyzed with Image-Pro Plus. 100–200 cells were counted for quantitative immunofluorescence assay from three independent experiments.

For histological staining, freshly collected cervical tissues were embedded with optimal cutting temperature compound (O.C.T, Sakuraus, 4583) after using PBS clean, immediately frozen with liquid nitrogen and stored in -80 °C. Frozen tissue blocks were sectioned at 5 microns using microtome (LEICA, CM3050).

Cells were resuspended in SDS sample loading buffer (pH8.0) and boiled for 5–8 min after radiotherapy and chemotherapy treatment. Nuclear extracts were separated through 8–15% SDS–PAGE, and then transferred onto PVDF membrane (Roche) which pre-soaked in methanol. Primary antibodies were diluted to the optimum concentration. HRP-conjugated anti-mouse or anti-rabbit IgG were got from DAKO. Blots were analyzed using Chemi docXRS (Bio-Rad). Antibodies used for immunoblotting included: anti-NBS1 phosphor-Serine 343 (Abcam, ab109453), anti-RPA32 phosphor-Serine 33 (Novus, NB100-544), anti-CHK2 phosphor-Threonine 68 (CST, 2348), anti-CHK1 phosphor-Serine 345 (CST, 2348S), anti-p53 phosphor-Serine 15 (CST, 9284S), anti-ubiquityl-histone H2B (Millipore, 05–1312), Tubulin (Sigma, T6074), anti-HA (Roche,11,867,423,001).

Survival after homologous recombination repair defective target drug PARP inhibitor (Olaparib) treatments was assessed by clonogenic assays. Cells were collected and counted in a cell counting board. After plating of 600 cells in 6 cm dish, the plates were treated with Olaparib (0.5, 0.75, 1, and 1.5 μM). The cells were treated for 14 days for colony formation. Colonies were fixed by ice methanol and stained with Giemsa solution.

SPSS 20.0 software was used to statistical analysis. All data came from three parallel assays. Student t test and one-way ANOVA was used to statistical significance analysis and the statistically significant difference was set at P < 0.05. Microsoft Office Excel 2007 was used for plotting.

To investigate if DNA damage responses is aggravated in precancerous lesions of HPV-positive cervical epithelium, cervical tissues were biopsied and cryosectioned, and subjected to immunofluorescent staining. Subject tissues ranged from cervical inflammation, CIN (cervical intraepithelial neoplasia) stages I–III to cancer, all infected with high-risk subtype of HPV16 according to pathological diagnosis (Fig. 1A). Ki67 staining revealed active proliferation in cervical epithelium with inflammation, which expanded in CIN lesions to higher compartments of presumably differentiated cells. Cancerous samples displayed high Ki67 manifestation throughout tissues. Regarding the phosphorylation of H2AX, HPV + epithelium exhibited pan-nuclear or focal signals in the apical side, largely colocalized with Ki67 + compartment, indicative of abnormal activation of checkpoint. γH2AX signals exacerbated in CIN I–III, however, were restricted in cancer samples. This is reminiscent to previous indication that hyperactive checkpoint prevents precancerous cells from cancer initiation (Wang et al. 2015). Moreover, autophosphorylation of ATM kinase (ataxia telangiectasia mutated) that monitors various forms of DNA breaks was manifested in nuclei of CIN epithelial cells (Fig. 1B), suggesting unrepaired DNA lesions in these cells.

We also examined 53BP1 that mainly represents DSB repair. In inflammatory epithelium 53BP1 foci was low (1–2 per cell), while with the advancement of CIN stages it increased dramatically, especially in CIN II–III tissues (Fig. 1C). These evidences collectively demonstrate that in HPV + precancerous lesions, levels of DNA damage are elevated along with the disease progression of cervical intraepithelial neoplasia.

To evaluate the viral impact of HPV on DNA damage response, we synthesized and cloned HA-tagged E6 and E7 genes of HPV16 subtype, which were transfected into HEK293T cells in parallel with empty vector (Fig. 2A). In ionizing radiation (IR, 10 Gy)-treated cells, E7 expression did not greatly affect the ATM and ATR (ataxia telangiectasia-related)-dependent checkpoint activation regarding CHK1-pS345, CHK2-pT68, and p53-pS15 (Fig. 2C), except ATM-pS1981 that does not play crucial role in checkpoint activation. When cells carrying pLV or pLV-E6 plasmids were challenged with IR or Camptothecin (CPT, 2 μM) for 2 h, DNA repair proteins were phosphorylated as revealed by immunoblotting with α-pS343 for NBS1 and α-pS33 for RPA32 (Fig. 2B). However, these events representing activated homologous recombination were obviously decreased in the presence of E7. Further, histone H2B monoubiquitination (uH2B) is required for HR processing (Zeng et al. 2016). In E7-expressing cells, decreased RPA32 phosphorylation is accompanied by down-regulated uH2B level (Fig. 2B). We conclude that HPV-encoded E7 rather than E6 perturbs the activation of DNA damage repair, which potentially leads to accumulation of DNA lesions in HPV-infected cervical epithelium.

To strengthen the impact of E7 on DNA repair, we used indirect immunofluorescent staining to reveal the competency of DSB repair HPV-negative cervical cancer cells (C33A). Four hours after IR treatment RPA32-pS33 formed foci in C33A cells transfected with pLV vector, whereas the signal significantly decreased upon E7 expression (Fig. 3A). This indicated that E7 attenuated the production of single-stranded DNA at DSBs. Given ssDNA is required for commitment of homologous recombination, RAD51 foci is reduced in irradiated cells expressing E7 in comparison with control cells, corroborating the failure of HR in the presence of E7 (Fig. 3B).

HR is regulated by the pro-NHEJ factors (i.e. 53BP1) and pro-HR factors (i.e. BRCA1). The balanced occupancy of these factors determines the ssDNA generation, and following HR commitment. We examined the DSB recruitment of 53BP1 and phosphorylated BRCA1 in irradiated cells. When C33A cells expressing E7 are subjected to IR treatment, BRCA1-pS1981 formed less foci 4- and 8-h post-IR relative to cells transfected with empty vector (Fig. 3C). Inversely, more 53BP1 foci persisted till 8-h after IR treatment when E7 was expressed, indicating that E7 suppresses the function of BRCA1 but enhances the blocking effect of 53BP1 in ssDNA production (Fig. 3D). These results imply that HPV-E7 perturbs the balanced loading of counteractive 53BP1-BRCA1 at DSBs, thus disrupting the pathway choice of HR and NHEJ. This leads to a HPV-induced deficiency of homologous recombination repair.

The viral HRD in the presence of E7 suggests that HPV-bearing cells could be sensitive to PARP inhibitors due to lethal chromosomal aberrance caused by illegitimate end joining (Patel et al. 2011). To test the synthetic lethality of E7-indued HRD and PARPi, we expressed E7 in C33A cells in the presence of Olaparib or not, and evaluated cell viability by colony formation (Fig. 4A). Clearly, cells transfected with pLV-E7 plasmid failed to sustain the exposure of Olaparib, forming miniscule or dying colonies. In contrast, C33A transfected with E7 without Olaparib treatment survived well and proliferated robustly comparable to pLV empty vector. Quantitatively, C33A cells with E7 expression exhibited 60% less survival relative to those harboring pLV (Fig. 4B).

It is a paradigm that TP53 gene is frequently mutated in cells carrying HRD including BRCA1 mutation (Nik-Zainal et al. 2016). The loss of pro-apoptosis function of p53 can alleviate the cells death caused by DNA breaks in the absence of functional HR, thus circumventing the p53-dependent anti-cancer network. Similarly, p53 degradation induced by E6 may protect E7-induced HRD cells from cell death. To test if E6 can also antagonize PARPi sensitivity caused by E7, we expressed these proteins individually or concomitantly in C33A or RPE1-hTERT cells, the latter of which is a non-cancerous eye epithelial line. Apparently, co-expressing E6 with E7 alleviated PARPi sensitivity, however, failed to reverse it completely (Fig. 4C, D). The mitigating effect of E6 is attributed to the abrogation of pro-apoptotic activity of p53. E6 expression did not impact E7-induced HRD because no statistic difference was observed for chromatin-bound RAD51 between E7 alone and E7/E6 co-expression cells subjected to CPT challenge (Fig. 4E). Taken together, we conclude that HPV-E7 reduces the HR competence and resultant PARPi sensitivity. These defects are largely resistant to E6-dependent p53 destruction.