Article Text

Original research
PARP inhibitors enhance antitumor immune responses by triggering pyroptosis via TNF–caspase 8–GSDMD/E axis in ovarian cancer
  1. Yu Xia1,2,
  2. Pu Huang3,
  3. Yi-yu Qian1,2,
  4. Zanhong Wang4,
  5. Ning Jin1,2,
  6. Xin Li1,2,
  7. Wen Pan1,
  8. Si-Yuan Wang5,
  9. Ping Jin6,
  10. Emmanuel Kwateng Drokow7,
  11. Xiong Li8,
  12. Qi Zhang9,
  13. Zhengmao Zhang10,
  14. Pingfei Li11,
  15. Yong Fang1,2,
  16. Xiang-Ping Yang12,
  17. Zhiqiang Han2,4 and
  18. Qing-lei Gao1,2
  1. 1Cancer Biology Research Center (Key Laboratory of the Ministry of Education, Hubei Provincial Key Laboratory of Tumor Invasion and Metastasis), Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology, Wuhan, Hubei, China
  2. 2National Clinical Research Center for Obstetrics and Gynecology, Department of Gynecological Oncology, Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology, Wuhan, Hubei, China
  3. 3Department of Obstetrics and Gynecology, Shandong Provincial Hospital Affiliated to Shandong First Medical University, Jinan, Shandong, China
  4. 4Department of Obstetrics and Gynecology, Shanxi Bethune Hospital, Taiyuan, Shanxi, China
  5. 5Department of Geriatrics, Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology, Wuhan, Hubei, China
  6. 6Department of Pediatrics, Tongji Hospital of Tongji Medical College of Huazhong University of Science and Technology, Wuhan, Hubei, China
  7. 7Hunan Provincial Key Laboratory of Clinical Epidemiology, Central South University, Changsha, Hunan, China
  8. 8Department of Gynecology & Obstetrics, the Central Hospital of Wuhan, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China
  9. 9Xianning Medical College, Hubei University of Science and Technology, Xianning, Hubei, China
  10. 10Department of Obstetrics and Gynecology, The Fourth Affiliated Hospital of Hebei Medical University, Shijiazhuang, Hebei, China
  11. 11Renmin Hospital, Hubei University of Medicine, Shiyan, Hubei, China
  12. 12Department of Immunology, School of Basic Medicine, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China
  1. Correspondence to Dr Qing-lei Gao; qlgao{at}tjh.tjmu.edu.cn; Dr Zhiqiang Han; hanzq2003{at}126.com

Abstract

Background In addition to their established action of synthetic lethality in tumor cells, poly(ADP-ribose) polymerase inhibitors (PARPis) also orchestrate tumor immune microenvironment (TIME) that contributes to suppressing tumor growth. However, it remains not fully understood whether and how PARPis trigger tumor-targeting immune responses.

Methods To decode the immune responses reshaped by PARPis, we conducted T-cell receptor (TCR) sequencing and immunohistochemical (IHC) analyses of paired clinical specimens before and after niraparib monotherapy obtained from a prospective study, as well as ID8 mouse ovarian tumors. To validate the induction of immunogenic cell death (ICD) by PARPis, we performed immunofluorescence/IHC staining with homologous recombination deficiency tumor cells and patient-derived xenograft tumor tissues, respectively. To substantiate that PARPis elicited tumor cell pyroptosis, we undertook comprehensive assessments of the cellular morphological features, cleavage of gasdermin (GSDM) proteins, and activation of TNF-caspase signaling pathways through genetic downregulation/depletion and selective inhibition. We also evaluated the critical role of pyroptosis in tumor suppression and immune activation following niraparib treatment using a syngeneic mouse model with implanting CRISPR/Cas9 edited Gsdme−/ ID8 tumor cells into C57BL/6 mice.

Results Our findings revealed that PARPis augmented the proportion of neoantigen-recognized TCR clones and TCR clonal expansion, and induced an inflamed TIME characterized by increased infiltration of both innate and adaptive immune cells. This PARPis-strengthened immune response was associated with the induction of ICD, specifically identified as pyroptosis, which possessed distinctive morphological features and GSDMD/E cleavage. It was validated that the cleavage of GSDMD/E was due to elevated caspase 8 activity downstream of the TNFR1, rather than FAS and TRAIL-R. On PARP inhibition, the NF-κB signaling pathway was activated, leading to increased secretion of TNF-α and subsequent initiation of the TNFR1–caspase 8 cascade. Impeding pyroptosis through the depletion of Gsdme significantly compromised the tumor-suppressing effects of PARP inhibition and undermined the anti-immune response in the syngeneic ID8 mouse model.

Conclusions PARPis induce a specific type of ICD called pyroptosis via TNF–caspase 8–GSDMD/E axis, resulting in an inflamed TIME and augmentation of tumor-targeting immune responses. These findings deepen our understanding of PARPis activities and point toward a promising avenue for synergizing PARPis with immunotherapeutic interventions.

Trial registration number NCT04507841.

  • Immune modulatory
  • Ovarian Cancer
  • T cell
  • Tumor microenvironment - TME
  • Neoadjuvant

Data availability statement

Data are available on reasonable request. Raw data from bulk TCR-seq of paired clinical specimens has been deposited in the Genome Sequence Archive for Human at the National Genomics Data Center with accession number HRA007230 under project PRJCA016620 (https://ngdc.cncb.ac.cn/bioproject/browse/PRJCA016620). The deidentified bulk TCR-seq profile is available at the NCBI’s Gene Expression Omnibus (GEO) database (accession number: GSE222557). Raw data from bulk TCR-seq of ID8 tumors has been deposited in GSA (https://ngdc.cncb.ac.cn/gsa/s/kCskQmer) with accession number CRA014711. The data from bulk RNA-seq of A2780 cells is available in OMIX (https://ngdc.cncb.ac.cn/omix/preview/ZAxvMET5) with accession number OMIX005758. Graphs of quantitative data and representative images of blot and gel generated in this study are available from the corresponding author upon reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Pyroptosis is a specific type of immunogenic cell death that can initiate a strong inflammatory cascade to heat the tumor immune microenvironment (TIME) and enhance tumor-targeting immune response, greatly contributing to tumor suppression.

  • Poly(ADP-ribose) polymerase inhibitors (PARPis) have the potential to reprogram the TIME by driving the transition from ‘cold’ to ‘hot’; however, it remains elusive whether these therapeutic agents can trigger immune responses that selectively target tumors and whether pyroptosis plays a pivotal role in this process.

WHAT THIS STUDY ADDS

  • PARPi exposure heightens tumor-targeting immune responses, as evidenced by the significantly increased proportion of neoantigen-recognized T-cell receptor (TCR) clones and the expansion of TCR clonality.

  • PARPis induce tumor cell pyroptosis through a GSDMD/E-dependent mechanism, largely contributing to the reprogramming of TIME.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • This study not only uncovers a novel form of cell demise resulting from PARPi administration but also elucidates an unappreciated mechanism underlying the enhancement of antitumor immune responses induced by PARPis, suggesting that the intensification of pyroptosis is a promising approach to amplify the efficacy of PARPis in the treatment of tumors.

Introduction

The advent of poly(ADP-ribose) polymerase inhibitors (PARPis), such as niraparib and olaparib, has ushered in a seminal era in targeted cancer therapeutics.1–5 The clinical triumph of PARPis lies in their ability to induce synthetic lethality in tumors with homologous recombination deficiency (HRD) caused by mutations in BRCA or other genes involved in homologous recombination repair (HRR), particularly exemplified in ovarian malignancies.3 Remarkably, a growing body of evidence highlighted an additional dimension to their therapeutic prowess. Treatment with PARPis showed an enhancement of immune responses in tumors, encompassing both HRD6 and homologous recombination proficiency (HRP).7 This enhancement was characterized by the conspicuous accumulation of peripheral CD8+ T cells within tumor beds, which was associated with increased levels of chemokines, including CCL5 and CXCL10.6 However, it is not yet completely clear if these infiltrated T cells target the tumors effectively and specifically. To fully grasp the immune-modulating role of PARPis, it is imperative to determine whether and how PARP inhibition orchestrates immune responses that selectively target tumors.

Immunogenic cell death (ICD) is known to instigate immune responses targeting neoantigens subsequent to pharmacologic intervention in tumors.8 9 Recently, there has been increasing interest in a type of ICD called pyroptosis, mainly because of its remarkable ability to accelerate a strong inflammatory cascade.4 10–12 The induction of pyroptosis hinges on the cleavage of gasdermin (GSDM) family proteins and caspase activation. Morphologically, it involves initial cell swelling, gradual emergence of bubble-like protrusions on the cell membrane, and eventual rupture of the cell membrane.12 When cells undergo pyroptosis, they release various inflammatory cytokines, damage-associated molecular patterns, and neoantigens into the tumor microenvironment. This, in turn, reinforces antitumor immune responses.4 10 13–15 For example, granzymes secreted from activated T cells or NK cells could penetrate tumor cells, activate GSDMs, and initiate pyroptosis, thereby boosting the cytotoxicity of lymphocytes.16 17 Also, delivering activated GSDMA3 into tumors through silyl-linked gold nanoparticles18 or employing an oncolytic virus armed with GSDME19 significantly enhanced the antitumor immune response. This heightened immune response renders tumors more susceptible to immune checkpoint blockade (ICB). In light of this, our goal is to figure out the potential of PARP inhibition in inducing pyroptosis, thus reshaping the tumor immune microenvironment (TIME) and amplifying tumor-targeting immune responses. Understanding these mechanisms could open new avenues for improving cancer treatments.

In this study, it was observed that PARP inhibition effectively triggered a type of ICD, pyroptosis, engendering an augmentation of neoantigen-specific T-cell receptor (TCR) clones of tumor-infiltrating T cells and an overall increase in the frequency of intratumoral immune cells engaging in both innate and adaptive immune responses. Further investigations revealed that PARP inhibition activated the NF-κB–TNF–caspase 8 pathway, contributing to the cleavage of GSDMD/E and subsequent pyroptosis. Therefore, our results have identified a novel mechanism by which PARPis exploit the pyroptosis to reprogram the TIME and elicit the tumor-targeting immune responses. These discoveries not only significantly advance the understanding of the pharmacological effect of PARPis but also promote the development of efficacious therapeutic approaches for tumors.

Results

The PARP inhibitor (PARPi) triggers tumor-targeting immune responses and tumor ICD

In a recent endeavor, we launched a phase 2 prospective clinical study (NANT, NCT04507841) to assess the therapeutic efficacy of neoadjuvant niraparib monotherapy in newly diagnosed HRD ovarian cancer (OC).20 Within the framework of this large clinical trial, matched specimens before and after neoadjuvant niraparib treatment were obtained and subjected to a highly informative technique of bulk TCR-sequencing (TCR-seq) to understand the impact of PARP inhibition on neoantigen-specific adaptive immune responses. Our findings revealed a significant increase in the proportions of neoantigen-recognized TCR clones following niraparib exposure (figure 1A). Meanwhile, niraparib treatment enhanced the TCR clonal expansion, as evidenced by a reduction in the top 25% clonotypes of the T-cell repertoire (figure 1B,C). To further validate these findings, we established a syngeneic mouse model by engrafting mouse OC ID8 cells into C57BL/6 mice. An analysis of bulk TCR-seq data from these ID8 tumors also showed an enhancement in TCR clonal expansion (online supplemental figure S1A–C) along with a reduction in TCR diversity (online supplemental figure S1D) in response to niraparib exposure. More importantly, T cells infiltrated in niraparib-challenged tumors exhibited a heightened potential for cytotoxicity against ID8 cells (online supplemental figure S1E). These tumor-killing T cells are often primed by antigen-presenting cells such as dendritic cells (DCs) in peripheral lymphatic organs,21 thus we proceeded to evaluate the activation of T cells in spleens. We found that the activated CD69+CD8+ T cells (online supplemental figure S1F,G) and CD137+CD8+ T cells (online supplemental figure S1F,H) were enriched in niraparib-treated spleens. We further assessed the antigen-presentation capacity of DCs within tumor tissues, draining lymph nodes, and spleens (online supplemental figure S1I). A uniform increase in the proportions of DC differentiation (CD11c+MHCIIHigh) and maturation (CD11c+MHCIIHighCD86High) was observed in niraparib-treated tumors (online supplemental figure S1J), draining lymph nodes (online supplemental figure S1K), and spleens (online supplemental figure S1L), compared with untreated counterparts. Collectively, these observations indicate that PARP inhibition augments tumor-targeting immune responses.

Supplemental material

Figure 1

The PARP inhibitor induces tumor immunogenic cell death. Neoantigen-recognized T-cell receptor (TCR) clones of tumor-infiltrated lymphocytes before and after niraparib treatment from patients enrolled in NANT ((A) n=11). The frequency of clonotypes in the top 25% of TCR repertoires was quantified ((B) n=11) and shown in pie charts (C). (D) OVCAR8 and (E) primary homologous recombination deficiency (HRD) ovarian cancer (OC) cells were treated with or without olaparib or niraparib for about 3 days and then stained with HMGB1 and calreticulin. Scale bars: 20 µm. (F–H) Clinical OC and triple-negative breast cancer (TNBC) tumors were implanted in immunodeficient NOG mice to generate patient-derived xenograft (PDX) models followed with or without niraparib treatment. Representative images of PDX tumors with immunohistochemical staining of HMGB1 and calreticulin (F). The ratios of HMGB1 cytoplasm-positive cells and calreticulin membrane-positive cells were quantified (G,H). Scale bars: 100 µm, n=8. Mean values±SEM. *P<0.05, **p<0.01, and ***p<0.001 by Student’s t-test in (A,B) and (G,H).

To exclude the possibility that the induction of strengthened immune response was attributed to a direct effect of PARPis on immune cells, we evaluated the cytokine production and apoptosis of T cells, as well as the differentiation and apoptosis of DCs, under olaparib and niraparib administration in vitro. In the case of T cells, the expression level of IFN-γ remained consistent in CD4+ T cells, whereas it declined in CD8+ T cells on exposure to the two PARPis (online supplemental figure S2A,B). Following the PARPis treatment, the frequency of CD8+ T cells producing granzyme B decreased (online supplemental figure S2B). Treatment with niraparib resulted in increased production of IL-2 in both CD4+ and CD8+ T cells, whereas olaparib did not affect IL-2 production (online supplemental figure S2A,B). Additionally, both olaparib and niraparib facilitated the apoptosis of CD4+ and CD8+ T cells (online supplemental figure S2C). As for DCs, the PARPis partially disrupted their differentiation and activation (online supplemental figure S2D), yet exerted minimal influence on DC apoptosis (online supplemental figure S2E). Consequently, these intriguing findings demonstrate that PARPis have a partial restraining effect on T-cell and DC functions in vitro, thereby suggesting that PARPis are unlikely to activate the immune response directly.

To delve deeper into whether ICD played a role in the enhanced DC differentiation and maturation, along with augmented T-cell expansion and cytotoxicity, we examined two ICD biomarkers high-mobility group box 1 (HMGB1) and calreticulin in HRD OC cell line OVCAR8 and HRD primary OC cells. The results showed that PARP inhibition significantly boosted the release of HMGB1 from cell nuclei and the translocation of calreticulin to cell membranes (figure 1D,E). These changes in HMGB1 and calreticulin were also duplicated in HRD OC and triple-negative breast cancer (TNBC) patient-derived xenograft (PDX) tumors using immunohistochemical (IHC) staining (figure 1F–H). Therefore, these compelling findings suggest that PARP inhibition triggers ICD, hinting at the possible mechanism to induce tumor-targeting immune responses.

The PARPi induces an inflammatory TIME

To explore whether PARPis could alter the TIME status, we examined the frequencies of diverse innate and adaptive immune cell populations, including CD4+ T cells, CD8+ T cells, CD19+ B cells, CD56+ NK cells, CD14+ monocytes, and GZMB+ cytotoxic cells, in the paired human OC specimens before and after treatment with niraparib. Our findings revealed a significant increase in the frequencies of all the tested immune cells following niraparib treatment, as determined by IHC staining (figure 2A–C). On exposure to olaparib and niraparib, ID8 tumors also exhibited enhanced infiltration of immune cells, including CD45+ leukocytes, CD4+ T cells, CD8+ T cells, CD11C+ DCs, NK1.1+ NK cells, and GZMB+ cytotoxic cells, with statistical significance (figure 2D–F,H). Additionally, there was an increased trend in the infiltration of CD14+ monocytes and CD19+ B cells (figure 2D,G). Collectively, these observations suggest that PARP inhibition produces a systemic inflammatory TIME within both clinical OC tumors and ID8 mouse tumors, thereby supporting the hypothesis that PARPis elicit ICD.

Figure 2

The PARP inhibitor promotes infiltration of both innate and adaptive immune cells. (A–C) Paired clinical ovarian cancer (OC) tumors before and after niraparib monotherapy were subjected to immunohistochemical (IHC) staining of CD4, CD8, CD19, CD56, CD14, and granzyme B (GZMB). Representative images of tumor sections before and after niraparib exposure (A). The numbers of indicated cells were counted in a ×20 field of a microscope ((B–C) n=11). Scale bars: 200 µm. (D–H) Olaparib/niraparib (40 mg/kg) was administered to ID8-bearing immunocompetent CD57/B6 mice for about 4 weeks. ID8 tumors were subjected to IHC staining. Representative images of tumor sections with staining CD45, CD4, CD8, NK1.1, CD19, CD11C, CD14, and GZMB (D). The numbers of indicated cells were counted in a ×20 field of a microscope ((E–H) n=5). Scale bars: 200 µm. Mean values±SEM. *P<0.05, **p<0.01, and ***p<0.001 by Student’s t-test in (B–C) and (E–H).

The PARPi induces caspase-dependent and pyroptosis-like cell death in tumor cells

To further characterize the ICD induced by PARPis, we initially assessed the forward scatter (FSC) values across multiple cancer cell lines, including OC (HOC7, SKOV3, A2780) and breast cancer (MDA-MB-231, MDA-MB-468, HCC1937), using flow cytometry in the presence of niraparib. The BRCA gene is wild type (WT) in HOC7, SKOV3, A2780, and MDA-MB-231 cells, whereas it is mutated in MDA-MB-468 and HCC1937 cells. The results showed that both BRCA-WT and BRCA-mutated tumor cells exhibited a consistent augmentation in FSC values in a dose-dependent and time-dependent pattern, indicating cell swelling after PARP inhibition (online supplemental figure S3A). Furthermore, the inhibition of PARP by olaparib and niraparib resulted in membrane rupture in tumor cells, as demonstrated by increased staining of propidium iodide (PI) and leakage of intracellular calcein in HRD OVCAR8 (figure 3A) and HRD primary OC cells (figure 3B) under microscopy. In line with this, olaparib and niraparib treatment increased the population of PI and annexin V double-positive OVCAR8 and A2780 cells (online supplemental figure S3B,C) and potentiated lactate dehydrogenase (LDH) release of these cells (online supplemental figure S3D,E). ICDs with characteristics such as cell swelling and membrane rupture typically include necroptosis, ferroptosis, and pyroptosis.22 23 Thus, to unravel the specific type of ICD elicited by PARPis, the tumor cells were pretreated with a series of death inhibitors in the presence of PARPis. Notably, the percentage of tumor cells positive for both PI and annexin V was dramatically reduced in the presence of Z-VAD-FMK (pan-caspase inhibitor) but was less affected by the incubation with necroptosis inhibitor Necrostatin-1 and ferroptosis inhibitor Ferrostatin-1 (figure 3C,D). Moreover, PARP inhibition induced the formation of bubble-like protrusions on the surface of tumor cells, including HRD primary OC cells and HRD cell lines (OVCAR8, MDA-MB-436, and HCC1937) (figure 3E). Conversely, treatment with Z-VAD-FMK depleted the appearance of these protrusions on PARP-inhibited tumor cells, such as OVCAR8 and A2780 cells (figure 3F,G). Z-VAD-FMK could also abrogate the elevated LDH release induced by PARPis in the two cell lines (figure 3H,I). These results provide evidence that PARP inhibition orchestrates caspase-dependent cell demise imbued with pyroptosis-like characteristics in tumor cells.

Figure 3

PARP inhibition induces a form of cell death similar to pyroptosis. (A,B) OVCAR8 and primary homologous recombination deficiency (HRD) ovarian cancer (OC) cells were treated with or without PARPis (olaparib and niraparib) for about 3 days and then stained with propidium iodide (PI) and calcein. Representative images and quantification of PI+ calcein OVCAR8 (A) and primary HRD OC cells (B). Scale bars: 100 µm. Quantification of PI+ annexin V+ OVCAR8 (C) and A2780 (D) cells on olaparib and niraparib treatment for about 3 days in the presence and absence of inhibitors, including Necrostatin-1 (Nec-1), Ferrostatin-1 (Fer-1), and Z-VAD-FMK (Z-VAD). (E) Representative images of OVCAR8, primary HRD OC cells, MDA-MB-436, and HCC1937 cells, under different doses of olaparib and niraparib exposure as indicated for 3 days. Representative images of OVCAR8 (F) and A2780 (G) cells under olaparib and niraparib exposure with or without Z-VAD incubation. The lactate dehydrogenase (LDH) levels in the OVCAR8 (H) and A2780 cells (I) culture medium were tested after treatment with olaparib or niraparib in the presence or absence of Z-VAD for 3 days. Mean values±SEM. *P<0.05, **p<0.01, and ***P<0.001 by Student’s t-test in (A–D) and (H,I).

PARP inhibition elicits pyroptosis in tumor cells through the activation of GSDMD and GSDME

Pyroptosis is mediated by the activation of GSDMs, namely GSDMA/B/C/D/E.12 24 To investigate which GSDM member could be activated by PARP inhibition, we initially checked the GSDM mRNA expression profiles in OC cell lines using the Cancer Cell Line Encyclopedia (CCLE) database. We found that the vast majority of these cell lines exhibited higher levels of GSDMD and GSDME than other family members (online supplemental figure S4A). This intriguing finding was further supported by the abundance of GSDMD and GSDME proteins in OVCAR8 and A2780 cells, and both the two GSDMs underwent proteolytic cleavage to their activated forms on treatment with olaparib and niraparib in a dose-dependent manner (figure 4A,B). Consistently, the cleavage of GSDMD and GSDME was also evident in the IHC staining of OC PDX tumors after the niraparib challenge (figure 4C–E). To determine the impact of the activated GSDMs on the antitumor capacity of PARPis, we employed small interfering RNAs (siRNAs) to silence the expression of GSDMA/B/C/D/E (online supplemental figure S5A). As shown in figure 4F,G, only tumor cells with downregulated GSDMD or GSDME displayed a decrease in cell death ratios following exposure to olaparib or niraparib. In addition, the knockdown of GSDMD or GSDME effectively abolished the generation of membrane blebbing (figure 4H) and LDH release (online supplemental figure S5B,C) of OVCAR8 and A2780 cells on the PARPis treatment. These phenomena were also confirmed in MDA-MB-436 TNBC cells (online supplemental figure S5D,E). Thus, the cellular results collectively corroborate that PARPis triggered tumoral pyroptosis.

Figure 4

PARP inhibitors induce pyroptosis by cleaving GSDMD and GSDME. (A,B) Western blotting analyses of GSDMA/B/C/D/E expression and cleavage in OVCAR8 and A2780 cells on olaparib and niraparib treatment at indicated doses. (C–E) Paired clinical ovarian cancer (OC) tumors before and after niraparib monotherapy were subjected to immunohistochemical (IHC) staining of GSDMD-N and GSDME-N. Representation images of IHC staining (C,D). Scale bars: 200 µm. IHC scores of GSDMD-N and GSDME-N were evaluated by two pathologists without patient data ((E) n=5). Quantification of propidium iodide (PI)+ annexin V+ OVCAR8 (F) and A2780 (G) cells on olaparib and niraparib treatment for about 3 days with knocking down GSDMA/B/C/D/E, respectively. (H) Representative images of OVCAR8 and A2780 cells on olaparib and niraparib treatment for 3 days with knocking down GSDMD/E, respectively. (I) The GSDMD/E levels in CRISPR/Cas9-edited ID8 cells were determined by western blotting. (J–L) Wild-type (WT) or Gsdmd/e-deficient ID8 cells were transplanted into immunocompetent C57BL/6 mice with or without niraparib treatment for about 4 weeks (n=6). Volumes of tumors with or without niraparib treatment (J,K). Images of tumors with or without niraparib treatment (L). Mean values±SEM. *P<0.05, **p<0.01, and ***P<0.001 by Student’s t-test in (E–G) and two-way analysis of variance in (J,K).

To test the role of GSDMD and GSDME in tumor suppression by PARPis in vivo, we employed the syngeneic OC mouse model. ID8 cells were genetically engineered using CRISPR/Cas9 to induce knockouts (KOs) of Gsdmd and Gsdme (figure 4I), respectively, and then subcutaneously transplanted into C57BL/6 mice. Once the bearing tumors reached about 100 mm3, oral administration of niraparib was initiated. During 4 weeks of treatment, we monitored the tumor volumes of the WT, Gsdmd KO, and Gsdme KO tumors. The tumor volumes were similar among these three groups without treatment (figure 4J,L); however, the growth of Gsdmd-deficient and Gsdme-deficient tumors was less impacted by niraparib exposure compared with the WT tumors (figure 4K,L). These observations highlight the important role of GSDMD/E and pyroptosis in PARPi-induced tumor suppression.

To understand whether the GSDM protein is responsible for TIME reprogramming during PAPRis treatment, we performed bulk TCR-seq analysis on WT and Gsdme KO ID8 tumors, both with and without niraparib treatment. In comparison with niraparib-treated WT tumors, niraparib-treated Gsdme KO tumors showed a significantly increased frequency of clonotypes in the top 25% of the TCR repertoire (figure 5A,B) and decreased TCR clonality (figure 5C). These findings suggest compromised T-cell clonal expansion within Gsdme-deficient tumors. On the contrary, the TCR diversity was elevated (figure 5D). The deficiency of Gsdme also limited the differentiation and maturation of tumor-infiltrated, lymphatic, and splenic DCs after niraparib treatment (figure 5E–G). Accordingly, the flow cytometry analyses revealed a significant decrease in the proportion of CD3+ T and NK cells in niraparib-treated Gsdme KO tumors compared with niraparib-treated WT tumors (figure 5H). The production of IFN-γ in CD4/8+ T cells and granzyme B in CD8+ T cells was reduced in the absence of Gsdme after niraparib administration (figure 5I,J). However, Gsdme deficiency had a minimal effect on the expression of IFN-γ and granzyme B in NK cells (figure 5I,J). Similarly, Gsdme KO tumors exhibited reduced frequencies of CD4/8+ T cells and GZMB+ cytotoxic cells when compared with WT tumors, as determined by IHC staining (figure 5K,L). Next, we aimed to investigate the critical role of pyroptosis in the established synergistic effect of combining PARPis and ICB for tumor suppression. To this end, we established an orthotopic engraftment model by transplanting WT and Gsdme KO ID8 tumors via intrabursal injection and exposing them to niraparib and/or anti-PD-1 antibodies. As expected, the synergistic effect was attenuated in Gsdme KO ID8 tumors (figure 5M,N). Taken together, these results suggest that PARPis initiate tumor-targeting immune responses by triggering pyroptosis.

Figure 5

Deficiency of GSDME blunts the immune response triggered by PAPR inhibitor in vivo. (A–L) Immunocompetent C57BL/6 mice were transplanted with wild type (WT) or Gsdme-deficient ID8 cells and challenged with niraparib for about 4 weeks. Tumors were harvested and subjected to bulk T-cell receptor (TCR) sequencing and evaluation of immune status. The frequency of T-cell clonotypes in the top 25% of TCR repertoires was shown using pie charts ((A) n=4) and quantified (B). The clonal expansion was evaluated using clonality (C). The TCR diversity was calculated by normalized Shannon diversity entropy (D). The proportion of dendritic cell differentiation (CD11c+MHCIIHigh) and maturation (CD11c+MHCIIHighCD86High) in tumors (E), lymph nodes (F), and spleens (G) was determined by flow cytometry. The proportion of CD3+ T cells and NK cells among CD45+ immune cells in tumors was determined by flow cytometry (H). The production of IFN-γ in tumor-infiltrated CD4/8+ T and NK cells was examined by flow cytometry (I). The expression of granzyme B in tumor-infiltrated CD8+ T and NK cells was evaluated by mean fluorescence intensity (J). Representative images of GSDME, CD4, CD8, and GZMB IHC staining in tumor sections (K) and numbers of indicated immune cells in a ×20 field of a microscope (L). Scale bars: 200 µm. WT or Gsdme-deficient ID8 cells were intrabursally transplanted into C57BL/6 mice and received niraparib and/or anti-PD-1 treatment. Representative bioluminescent images of mice bearing WT and Gsdme-KO ID8 tumors at the endpoint (M). The tumor weight was quantified in each mouse treated with niraparib and/or anti-PD-1 antibodies ((N) n=7). Mean values±SEM. *P<0.05, **p<0.01, and ***p<0.001 by Student’s t-test in (B–J, L, and N).

The cleavage of GSDM proteins induced by PARP inhibition requires the activation of caspase 8

Next, we sought to determine the mechanism underlying the cleavage of GSDM proteins under PARP inhibition. Given that PARPi-induced pyroptosis was shown above to be caspase dependent, we pretreated HRD primary OC, OVCAR8, and A2780 cells with caspase inhibitors. Notably, Z-DEVD-FMK (caspase 3/8 inhibitor), Z-IETD-FMK (caspase 8 inhibitor), and Z-VAD-FMK (pan-caspase inhibitor) significantly weakened the antitumor ability of niraparib, as measured by PI and annexin V staining. However, the caspase 1 inhibitor VX765 did not exhibit a similar effect (figure 6A–C). These results suggest that the activation of caspase 8 is required for the niraparib-induced death. Consistent with this, we found elevated activity of caspase 3 and caspase 8 in PARP-inhibited OC cell lines (OVCAR8 and A2780) (figure 6D) as well as PDX tumors (OC and TNBC) (figure 6E,F). To further investigate whether the cleavage of GSDMD and GSDME is dependent on the activity of caspase 8, we examined the levels of cleaved GSDMD/E in the presence of the caspase inhibitors. Treatment with Z-DEVD-FMK, Z-IETD-FMK, and Z-VAD-FMK sufficiently led to a reduction in the levels of the functional N-terminal domain of GSDMD and GSDME in OVCAR8 and A2780 cells under niraparib exposure (figure 6G). Finally, Z-DEVD-FMK and Z-IETD-FMK, but not VX765, suppressed membrane blebbing (figure 6H,I) and LDH release (online supplemental figure S6A,B) in OVCAR8 and A2780 cells on PARP inhibition by olaparib and niraparib. In summary, these results suggest that caspase 8 mediates GSDMD/E cleavage on PARP inhibition.

Figure 6

Caspase 8 contributes to the pyroptosis induced by PARP inhibition. Ratios of PI+ annexin V+ OVCAR8 (A), A2780 (B), and primary homologous recombination deficiency (HRD) ovarian cancer (OC) tumor cells (C) on niraparib exposure in the presence or absence of caspase inhibitors, including VX765, Z-DEVD-FMK, Z-IETD-FMK, and Z-VAD-FMK. (D) Determination of activated caspase 3 and caspase 8 levels in OVCAR8 and A2780 cells after olaparib and niraparib treatment by flow cytometry. Histological analyses of activated caspase 3 and caspase 8 in OC patient-derived xenograft (PDX) tumors (E) and triple-negative breast cancer (TNBC) PDX tumors (F) with or without niraparib exposure (n=5/6). (G) Western blotting analyses of the cleavage of caspase 8/3 and GSDMD/E in OVCAR8 and A2780 cells on niraparib treatment in the presence or absence of caspase inhibitors, including Z-DEVD-FMK, Z-IETD-FMK, and Z-VAD-FMK. Representative images of OVCAR8 (H) and A2780 cells (I) on niraparib treatment in the presence or absence of caspase inhibitors, including VX765, Z-DEVD-FMK, and Z-IETD-FMK. Mean values±SEM. *P<0.05, **p<0.01, and ***p< 0.001 by Student’s t-test in (A–C) and (E,F).

TNFR signaling contributes to caspase 8-mediated pyroptosis induced by PARP inhibition

Caspase 8, a highly conserved cysteine protease, serves as the principal initiator of the caspase cascade, triggered by the activation of cell surface death receptors, including tumor necrosis factor receptor superfamily member 1 (TNFR1), TNFR superfamily member 10A/B (TRAIL-R1/2), and TNFR superfamily member 6 (FAS).25–28 To identify the stimulus responsible for cell demise and caspase activation, we employed siRNAs to suppress the expression of TNFR1 (genge for TNFR superfamily member 1), DR5 (gene for TRAIL-R2), and FAS (gene for TNFR superfamily member 6) in tumor cells (figure 7A). We discovered that downregulating TNFR1 expression improved the survival of tumor cells on PARP inhibition, whereas reducing DR5 and FAS expression had a negligible effect on tumor cell survival (figure 7B,C). The use of two FDA-approved antibodies, adalimumab and infliximab, blocking the interaction between TNF-α and TNFR1, also effectively restrained the potential of PARPis to induce cell death (figure 7D,E). Moreover, silencing TNFR1 expression led to a significant reduction of caspase 8/3 activity in OVCAR8 (figure 7F,G) and A2780 cells (figure 7H,I). We then investigated the contribution of TNFR signaling to the cleavage of GSDM proteins. It was noted that reducing TNFR1 expression in tumor cells robustly restrained the olaparib-potentiated and niraparib-potentiated activation of GSDMD and GSDME in OVCAR8 and A2780 cells (figure 7J). In line with this, morphological and LDH examination also confirmed the disappearance of membrane blebbing (figure 7K) and a reduction of LDH release (figure 7L,M) in TNFR1-depleted tumor cells in the presence of olaparib and niraparib. Overall, these observations demonstrate the critical role of the TNFR1 signaling pathway in PARP inhibition-induced caspase activation and pyroptosis.

Figure 7

TNFR is located upstream of caspase 8 activation and is responsible for PARP inhibitor-induced pyroptosis. (A) Knocking down TNFR1, DR5, and FAS expression in OVCAR8 cells with siRNAs. Ratios of PI+ annexin V+ OVCAR8 (B) and A2780 cells (C) with silencing TNFR1, DR5, and FAS expression, respectively, on olaparib and niraparib exposure or not. Ratios of PI+ annexin V+ OVCAR8 (D) and A2780 cells (E) on olaparib or niraparib exposure in the presence or absence of TNF-α neutralizing antibodies, including adalimumab and infliximab. Determination of activated caspase 8 and caspase 3 in OVCAR8 (F,G) and A2780 cells (H,I) with downregulating TNFR1 expression on olaparib and niraparib treatment or not by flow cytometry. (J) Western blotting analyses of cleavage of GSDMD/E in OVCAR8 and A2780 cells with downregulating TNFR1 expression on olaparib and niraparib treatment or not. (K) Representative images of OVCAR8 and A2780 cells with downregulating TNFR1 expression on olaparib and niraparib treatment or not. (L,M) Tumor cells with or without TNFR1 knocking down were exposed to olaparib or niraparib for 3 days. The lactate dehydrogenase (LDH) releases in OVCAR8 (L) and A2780 (M) culture medium were tested. Mean values±SEM. *P<0.05, **p<0.01, and ***p<0.001 by Student’s t-test in (B–I) and (L–M).

DNA damage-activated NF-κB contributes to the TNF-α secretion and PARP inhibition-induced pyroptosis

Next, we attempted to unravel the intricate mechanism governing the process of TNFR1/caspase8-associated pyroptosis. To this end, we examined the levels of death-associated ligands upstream of caspase 8, including TNF-α, TRAIL, and FASL, on olaparib or niraparib stimulation in vitro. Of note, TNF-α exhibited an increase at both transcriptional and translational levels in OVCAR8 and A2780 cells, whereas TRAIL and FASL did not (figure 8A,B, online supplemental figure S7A–J). Our RNA-seq analysis also revealed an upregulation of TNF expression in tumor cells treated with niraparib (figure 8C). To confirm this finding, we extended our investigation to OC PDX tumors by IHC staining, where we observed the significant enhancement of TNF-α expression in response to niraparib treatment (figure 8D). On further analysis of the bulk RNA-seq data, we found the enrichment of NF-κB-related genes, which were closely associated with the transcriptional regulation of TNFA, following exposure to niraparib (figure 8E). Consistently, the increased NF-κB transcriptional activity was observed in niraparib-treated OC PDX tumors (figure 8F). To determine the pivotal role of activated NF-κB in mediating TNF-α transcription, we employed siRNA to downregulate RELA (gene for p65) expression in OVACR8 and A2780 cells (online supplemental figure S8A,B) and then evaluated TNF-α expression and secretion in these engineered cells. Knockdown of RELA in the tumor cells elicited a dramatic dampening effect on both the expression (online supplemental figure S8C,D) and secretion (figure 8G,H) of TNF-α. In addition, the reduction of RELA in the tumor cells abolished the appearance of membrane blebbing (figure 8I,J) and depleted LDH release (figure 8K,L) in response to olaparib and niraparib treatment. To establish the critical role of tumor NF-κB in the enhancement of immune responses, Rela-deficient ID8 cells were generated by CRISPR/Cas 9 (online supplemental figure S8E) and transplanted into female mice via intrabursal injection. Following a 6-week exposure to niraparib, the immune status within ID8 tumors and draining lymph nodes was assessed. The results suggested that the deficiency of Rela significantly impaired the ability of niraparib to induce immune activation, including DC maturation (online supplemental figure S8F,G), T-cell cytokine production (online supplemental figure S8H,I), and T-cell cytotoxicity (online supplemental figure S8J,K).

Figure 8

The activation of NF-κB contributes to the secretion of TNF-α and subsequent pyroptosis. Determination of TNF-α secretion in the tumor culture medium of OVCAR8 (A) and A2780 cells (B) after olaparib and niraparib treatment for 24, 48, and 72 hours. (C) The volcano plot shows the upregulated and downregulated genes in A2780 cells after niraparib treatment. (D) Histological analyses of TNF-α expression in ovarian cancer (OC) patient-derived xenograft (PDX) tumors with or without niraparib treatment (n=5/6). (E) Gene Ontology (GO) analysis of the bulk RNA-seq data to identify upregulated genes enriched in the related signal pathways in A2780 cells with niraparib treatment. (F) Histological analyses of NF-κB activity (p-p65 expression levels) in OC PDX tumors with or without niraparib treatment (n=5/6). Determination of TNF-α secretion in the tumor culture medium of OVCAR8 (G) and A2780 cells (H) with knocking down RELA on olaparib and niraparib treatment or not. (I,J) Representative images of OVCAR8 (I) and A2780 (J) cells with downregulating RELA expression on olaparib and niraparib treatment or not. (K,L) Tumor cells with or without RELA knocking down were exposed to olaparib or niraparib for 3 days. The lactate dehydrogenase (LDH) releases in OVCAR8 (K) and A2780 (L) culture medium were tested. (M) Western blotting analyses of PAR, γ-H2AX, p-ATM, p-p65, c-FLIP, cleaved caspase 8, cleaved GSDMD, and cleaved GSDME with olaparib and niraparib treatment for 24, 48, and 72 hours. (N) Schematic illustration of pyroptosis induced by PARP inhibitors. Mean values±SEM. *P<0.05, ***p<0.001 by Student’s t-test in (A, B, D, F–H, K and L).

It is well established that damaged DNA can activate NF-κB, leading to the secretion of inflammatory cytokines and regulation of cell death.29 30 A classic downstream effector of NF-κB, c-FLIP, is involved in the prosurvival signaling cascade by inactivating caspase 8.27 31 32 Meanwhile, the nuclear kinase ataxia telangiectasia mutated (ATM), known to respond to DNA breaks, can be activated following DNA damage, impeding the activity of c-FLIP.33 34 In this context, to elucidate the underlying mechanisms of how NF-κB elicited pyroptosis via the TNF–caspase 8 axis, we examined various markers, including poly(ADP-ribose) (PAR), DNA damage marker (γ-H2AX), phosphorylated-ATM (p-ATM), p-p65, c-FLIP, cleaved caspase 8, GSDMD, and GSDME on treatment with olaparib or niraparib at different time points. The results hinted at the following dynamics: (1) poly-ADP-ribosylation inhibition and DNA damage coincided with the activation of ATM and NF-κB after approximately 24 hours of PARPis exposure; (2) c-FLIP initially exhibited an increase due to the NF-κB activation but later experienced a decrease because of the pronounced activation of ATM; (3) caspase 8 underwent cleavage following the substantial decrease in c-FLIP levels; (4) cleaved GSDMD-N and GSDME-N were virtually undetectable in the early stages due to the inactivation of caspase 8 but showed a dramatic increase once caspase 8 activation ensued (figure 8M). Figure 8N provides a schematic illustration encapsulating the process of pyroptosis induced by PARPis. Collectively, these results unveil the upregulation of TNF-α expression in tumor cells subjected to PARP inhibition, owing to the activation of NF-κB in response to DNA damage. This, in turn, sets the stage for the TNFR/caspase 8-mediated orchestration of pyroptosis.

Discussion

In this study, leveraging TCR-seq analyses and IHC staining of both clinical specimens and mouse ID8 tumors, we substantiated the ability of PARPis to stimulate tumor cell ICD, fostering an inflamed TIME and enhancing tumor-targeting immune responses. The ICD was identified as an inflammatory pyroptosis characterized by heightened N-terminal GSDMD/E proteins and membrane rupture. The blockade of pyroptosis by genetically depleting Gsdme expression hindered both tumor-suppressive and immune-activating effects of PARPis. We delved into the intricate mechanisms, revealing an NF-κB–TNF–caspase 8–GSDMD/E axis orchestrating the pyroptosis. Our findings shed light on a novel mode of cell death that contributes to the neoantigen-specific immune response, broadening the horizons of the immunoregulatory impact of PARPis.

Both PARP and HRR play pivotal roles in the DNA damage response, serving as evolutionarily conservative machinery dedicated to safeguarding genomic integrity.3 5 On genetic deficiency or pharmacological inhibition of PARP, accumulation of DNA single-strand breaks ensues, inducing replication fork collapse. This triggers the recognition by BRCA1/2, initiating the HRR process to repair the damaged DNA.5 Hence, in the context of BRCA-deficient tumor cells, PARP inhibition synergistically induces chromosomal instability, cell cycle arrest, growth delay, and even cell demise,1 2 marking synthetic lethality as the cornerstone of PARPis in clinical applications.3 As research advances, evidence has emerged regarding the ability of PARPis to modulate antitumor immune responses. Notably, the cGAS-STING pathway, critical for type I immune response, can be activated by PARPis, resulting in increased PD-L1 expression and CD8+ T-cell infiltration.6 35 The elevated expression of CCL5 and CXCL10 contributed to this recruitment of CD8+ T cells.6 However, it remains a deeper investigation to understand whether the observed immune response enhancement effectively targets the tumor, as the majority of tumor-infiltrated T cells are bystanders.36 37 Here, our study unveiled that PARPis had the capacity to initiate ICD. This was supported by an increase in the proportion of tumor-specific TCR clones, TCR clonal expansion, T-cell cytotoxicity, DC maturation, and the translocation of HMGB1 and calreticulin. ICDs represent dying cells with sufficient antigenicity to engage tumor-targeting immune responses and strongly heat TIME.8 22 To validate these findings, we analyzed paired clinical specimens collected before and after niraparib treatment from the prospective clinical study (NANT). Our results confirmed a substantial increase in both innate and adaptive immune cell infiltration. Additionally, this strengthened tumor-targeting immune response has been observed to be devoid of serious immune-related adverse effects both in clinical settings38–40 and in our mouse models (data not shown). Consequently, our results significantly contribute to advancing our understanding regarding the immunoediting effects of PARPis.

The conventional understanding has long attributed the cell death induced by PARPis to apoptosis.2 Nonetheless, recent research has expanded this perspective and suggested that olaparib could also trigger tumor cell ferroptosis in a p53-dependent manner.41 Furthermore, in combination with arsenic trioxide, olaparib has been shown to enhance both apoptosis and ferroptosis in platinum-resistant OC.42 In our study, we unveiled a new dimension to the repertoire of cell death mechanisms triggered by olaparib and niraparib: pyroptosis. This process was predominantly driven by the heightened activation of the TNF/caspase 8 signaling pathway, contingent on NF-κB activation. Traditionally, TNF has long been acknowledged as an inducer of necroptosis when caspase 8 is inactivated by c-FLIP.27 31 However, damaged DNA not only activates NF-κB but also robustly induces ATM phosphorylation which in turn hampers c-FLIP activity,30 33 34 as supported by our western blotting analyses. Consequently, PARPis engendered caspase 8 activation and pyroptosis, closely linked to DNA damage. This activated caspase 8 could then directly cleave GSDMD or indirectly cleave GSDME by the facilitation of active caspase 3, as reported in prior studies.10 43 The NF-κB pathway is widely recognized for its role in promoting tumor progression, including initiation, proliferation, and maintenance of cancerous cells.44 NF-κB can also be activated by TNF-α, thereby contributing to cancer cell migration and invasion.45 Our findings unveiled an additional function of NF-κB and TNF-α in tumor suppression, specifically their involvement in the initiation of pyroptosis. These findings provide evidence that NF-κB and TNF-α have the potential to act as a double-edged sword in tumors, exhibiting both tumor-promoting and tumor-suppressive effects.

In addition, Wang et al documented that elevated levels of GSDMC, either introduced through lentivirus or induced by IFN-γ, rendered tumors susceptible to PARPis treatment.46 However, in our investigation, we assessed the baseline expression of GSDM family members in ovarian tumor cells using the CCLE database. Our findings uncovered a preferential expression of GSDMD and GSDME in multiple OC cell lines, which was confirmed by western blotting. Moreover, it was observed that the high expression of GSDMD and GSDME underwent cleavage in response to treatment with olaparib and niraparib. This discovery introduces a novel intrinsic cell death process, pyroptosis, triggered by olaparib and niraparib, and elucidates the underlying mechanism, profoundly advancing our knowledge of the pharmacological responses of PARPis.

In conclusion, our investigation reveals that PARPis induce a specific type of ICD known as pyroptosis, thereby driving the inflamed TIME and promoting tumor-targeting immune responses. These findings not only establish a novel connection between cellular demise and immune activation but also deepen our understanding of how PARPis reshape the immune landscape. This insight highlights the great potential of synergizing PARPis with immunotherapy.

Methods

Clinical specimens

Detailed information about this NANT trial is available in our recent publication.20 In brief, these advanced OC patients who were newly diagnosed with high-grade plasma OC harboring HRD (genomic instability score (GIS)≥42 or BRCA mutation) would receive the treatment of niraparib (200 or 300 mg for two cycles, each lasting 28 days). These patients participating in this trial and receiving niraparib challenges have achieved their primary objectives so far. The objective response ratio (ORR) was 62.5% according to the Response Evaluation Criteria in Solid Tumors (V.1.1).47 Meanwhile, our recently published study in Cell reported an ORR of 73.6% based on the Gynecologic Cancer InterGroup CA125 response criteria.48 All the enrolled patients signed an informed consent form, and all the tumor samples were collected with no burden to the patients. The pretreated and post-treated tumor samples were surgically resected from diagnostic biopsy and complete surgical staging and maximal resection, respectively. We collected 22 paired samples from 11 OC patients to perform bulk TCR-seq (The baseline information of these patients is summarized in table 1), 22 paired tumor samples from 11 OC patients for IHC analysis (the baseline information of these patients is summarized in table 2), and another one fresh tumor from an OC patient for PDX generation (the baseline information of these patients is summarized in table 3). Additionally, we also collected one fresh tumor from a TNBC patient to generate a PDX model (the baseline information of these patients is summarized in table 4).

Table 1

The baseline information of patients in bulk TCR-seq cohort

Table 2

The baseline information of patients in the IHC analysis cohort

Table 3

The baseline information of patients in ovarian cancer PDX generation cohort

Table 4

The baseline information of patients in the TNBC PDX generation cohort

Animals

Female C57BL/6 and NOD-PrkdcscidIl2rgem1/Smoc mice (6–8 weeks) were obtained from the GemPharmatech and maintained in a 12 hours light/dark cycle with free access to food and water.

Mouse models

For PDX models, fresh tumor tissues were transported to the laboratory within 4 hours while stored in saline solution at 4°C and chopped into small pieces, after being removed during surgery. Female NOD-PrkdcscidIl2rgem1/Smoc mice aged 6–8 weeks were subcutaneously engrafted with the tissue fragments on the dorsum after being sedated with 1%–1.5% isoflurane (R510-22-10, RWD Life Cycle). These mice were referred to as the first-generation xenografts (P1). Thereafter, tumor volumes were calculated using the modified formula every other day: tumor volume=L×W×W×0.5, where L is the length and W is the width. The animals were euthanized following anesthesia when the tumor volume exceeded 1000 mm3. The fresh tumor tissues were collected, cut into pieces with equal volume (10 mm3), and immediately transplanted into the new mice as the next-generation PDX model. When the tumors developed to the appropriate size, the mice were randomly divided into two groups for treatment with either vehicle (corn oil, 100 µL/mouse, 5 days a week) or niraparib (40 mg/kg/day, 5 days a week) via intragastric gavage for 4 weeks.

For the murine subcutaneous transplantation model, 5×106 ID8-VEGFHi OC cells with genetically depleting Gsdmd, Gsdme or not using CRISPR/Cas 9 were resuspended in 100 µL PBS mixed with Matrigel (Corning, 354234) at a ratio of 1: 1 and transplanted into C57BL/6 mouse via subcutaneous injection. While the tumor volumes reached about 100 mm3, these tumor-bearing mice would be randomly divided into three groups for treatment with niraparib (40 mg/kg/day, 5 days a week), olaparib (40 mg/kg/day, 5 days a week), or vehicle (corn oil, 100 µL/mouse, 5 days a week) for about 4 weeks, and the tumor volume was recorded.

For the murine orthotopic transplantation model, 3×106 ID8 OC cells with genetically depleting Gsdme, Rela or not using CRISPR/Cas 9 were resuspended in 30 µL PBS mixed with Matrigel (1: 1) and transplanted into C57BL/6 mouse via intrabursal injection. After 1 week of transplantation, the Gsdme-deficient or WT ID8 tumors received niraparib (40 mg/kg/day, 5 days a week) and/or anti-PD-1 antibodies (BioXCell, BE0146, 200 µg/mouse, two times per week) challenge for 5 weeks, whereas Rela-deficient or WT ID8 tumors received niraparib (40 mg/kg/day, 5 days a week) treatment for 6 weeks. At the endpoint, the tumor burden was assessed by the bioluminescent intensity and total tumor weight. The immune status of tumors and draining lymph nodes was evaluated using flow cytometry.

Bulk DNA isolation and TCR-seq

For the patients’ samples, the genomic DNA of tissues was extracted using a QIAamp DNA Mini Kit (QIAGEN, 51306) according to the manufacturer’s instructions. DNA concentrations were quantified using a Qubit dsDNA HS Kit (Invitrogen, Q32854), and DNA quality was evaluated using agarose gel electrophoresis. The DNA amplification of the TCRβ CDR3, including 30 forward V primers and 13 reverse J primers, was performed using MPCR primers from a published paper.49 Following multiplexed PCR within a 50 µL reaction mix, containing 1×QIAGEN Multiplex PCR Master Mix (QIAGEN, 206143), 0.5×Q solution, 0.2 µM VβF pool, and JβR pool, single-stranded PCR products were obtained by denaturation and used in the circularization program. Single-stranded circle DNA molecules were replicated via rolling cycle amplification, and a DNA nanoball (DNB) which contained multiple copies of DNA was generated. Finally, sufficient quality DNBs were loaded into patterned nanoarrays using MGISEQ 2000.

For the tumor tissue from the mouse, the genomic DNA was extracted using a QIAamp DNA Mini Kit (QIAGEN, 51304), and the concentration was tested using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA). The DNA was used as a template for PCR amplification, which was performed to generate the library of TCR. The PCR products were purified, and the barcodes were confirmed. The annealed region of these primers was the CDR3 region, as shown by Multiplex PCR Amplifies. The CDR3 regions were sequenced on the Illumina NextSeq500. According to the sequencing depth, the test covered about 90% of the TCR. The CDR3 regions were identified based on the definition established by the International ImMunoGeneTics (http://www.imgt.org/, accessed on 19 March 2021) collaboration, and the V, D, and J segments contributing to each CDR3 region were identified by a standard algorithm.50

Bulk TCR-seq data analysis

For the bulk TCR-seq data of patient samples, raw sequencing data were processed by the tool IMonitor (V.1.4.1).49 Briefly, the raw paired-end (PE) reads were merged into one sequence by the overlapped region. Low-quality sequences were filtered out. The clean sequences were aligned to reference that including V, D, and J germline sequences (www.imgt.org). Originated V, D, and J genes were determined for sequences, and CDR3 regions were identified. Sequencing errors in CDR3 sequences were corrected according to the base sequencing quality and CDR3 frequency. Nucleotide CDR3 sequences were translated to amino acid sequences.

For the bulk TCR-seq data of mouse samples, the original data were converted to raw sequence reads, and low-quality sequences were filtered out. The MiXCR software (V.3.0.12, MiLaboratories, Sunnyvale, California, USA) was used to identify the V, D, and J genes, extract the CDR3 sequence, and correct PCR errors.51 The productive immune repertoire sequence reads were filtered by removing (1) any reads with CDR3s shorter than four amino acids; (2) CDR3 contigs with a length that was not a multiple of 3; and (3) contigs containing stop codons.

The clonal expansion was evaluated using clonality. The diversity of TCR was analyzed by normalized Shannon diversity entropy.52 The TCR repertoire was measured by calculating the number of individual clonotypes represented in the top 25th percentile by ranked molecule count after sorting by abundance.

TCR repertoire annotated by disease-associated TCR database

Disease-associated TCR sequences consisted of a published database, McPAS-TCR. For each TCR sample, CDR3 amino acid sequences were compared with the CDR3s in databases by Levinshtein distance. The CDR3 sequence in the sample was supposed to be related to the disease if the Levinshtein distance ≤1 between CDR3 in the sample and CDR3 associated with a disease in the database. At last, the proportion of tumor-related CDR3s, the number of tumor-related CDR3s divided by the total number of CDR3 in a sample, was calculated.

Cell line and cell culture

SKOV3, HOC7, HEK293T, MDA-MB-231, MDA-MB-468, MDA-MB-436, and HCC1937 cell lines were obtained from the American Type Culture Collection. A2780 and ID8 cell lines were purchased from Sigma-Aldrich. OVCAR8 was obtained from MDACC’s Characterized Cell Line Core. OVCAR8, A2780, HOC7 and HCC1937 cells were grown in RPMI 1640 (Gibco, 31800-022) containing 10% fetal bovine serum (FBS) (Gibco, 16000-044) and 1% penicillin–streptomycin solution (PS) (Servicebio, G4003). SKOV3 cells were cultured in McCoy’s 5a Medium (Sigma, M4892) with 10% FBS and 1% PS. HEK293T and MDA-MB-231 cells were grown in Dulbecco’s modified Eagle medium (DMEM) (Gibco, 31800-017) containing 10% FBS and 1% PS. ID8 mouse OC cells were cultured in DMEM supplemented with 10% FBS and 1% PS. MDA-MB-468 and MDA-MB-436 cells were grown in Leibovitz’s L-15 (Gibco, 21083-027) with 10% FBS and 1% PS. All cell lines were authenticated through STR profiling and tested monthly for Mycoplasma by PCR. Cells were cultured in a thermostatic incubator at 37°C and 5% CO2 except for MDA-MB-468 and MDA-MB-436 cells. A2780, SKOV3, HOC7, and MDA-MB-231 are HRR proficient, while OVCAR8 (BRCA1 methylation), MDA-MB-436 (BRCA1 mutation), MDA-MB-468 (BRCA2 mutation), and HCC1937 (BRCA1 mutation) are HRD according to the CCLE.

Primary tumor cell isolation

Fresh PDX tumors were collected, cut into small pieces, and digested using the tumor dissociation kit (Miltenyi Biotec, Germany) for 60 min at 37°C. RPMI-1640 with 2% FBS was used to terminate the digestion, and the digested mixture was filtered through a 40 µm filter. Red blood cells were lysed with RBC lysis buffer (Servicebio, China) according to the manufacturer’s protocol. After three washes with sterile PBS, cells were cultured in DMEM supplemented with 10% FBS and incubated at 37°C with 5% CO2 in a humidified cell culture incubator.

Reagents

Olaparib (S1060), Z-VAD-FMK (S7023), Z-IETD-FMK(S7314), Z-DEVD-FMK (S7312), Necrostatin-1 (S8037), and Ferrostatin-1 (S7243) were purchased from Selleck. The anti-human TNF-α for the in vitro experiments were from Bioxcell, including Infliximab (SIM0006) and Adalimumab (SIM0001). VX765 (Inh vx765i) was supplied by Invivogen. Niraparib was obtained from Zai Lab.

RNA interference

The cells were cultured until they reached approximately 60% confluence before transfection. siRNA were transfected into these cells using Lipofectamine 3000 (Invitrogen, Carlsbad, California, USA) according to the manufacturer’s protocol. The transfected cells were grown in an antibiotic-free medium for 24 hours and then treated with olaparib or niraparib. The knockdown efficiency was examined by qRT-PCR analysis. The siRNA sequences were obtained from RiboBio (Guangzhou, China). The siRNA sequences used were as follows (5′−3′),

  • Si-GSDMA 1, GAAGGAGATGCAAGATCAA

  • Si-GSDMA 2, GCGCCATCCTCTATTTCGT

  • Si-GSDMA 3, CCTCTGTGCTCTTTATGCA

  • Si-GSDMB 1, GGATATGATTGCCGTTAGA

  • Si-GSDMB 2, GGACAAGTGGTTAGATGAA

  • Si-GSDMB 3, CCTTGTTGATGCTGATAGA

  • Si-GSDMC 1, GTGCCATCCTAAAGAAACT

  • Si-GSDMC 2, GCTGTACGATAGCAGTAGT

  • Si-GSDMC 3, GGGTGTGAATGTTGGTATA

  • Si-GSDMD 1, GCAGGAGCTTCCACTTCTA

  • Si-GSDMD 2, GGAACTCGCTATCCCTGTT

  • Si-GSDMD 3, GTGTCAACCTGTCTATCAA

  • Si-GSDME 1, CCCACTGCTTCTTTGTATA

  • Si-GSDME 2, CAAGCAGCTGTTTATGACA

  • Si-GSDME 3, GGTCCTATTTGATGATGAA

  • Si-TNFR1A 1, CTCCATTGTTTGTGGGAAA

  • Si-TNFR1A 2, CCGTTGCGCTGGAAGGAAT

  • Si-TNFR1A 3, TGGGCTGCCTGGAGGACAT

  • Si-FAS 1, GGATTGGAATTGAGGAAGA

  • Si-FAS 2, GGGAAGGAGTACACAGACA

  • Si-FAS 3, CACCAAGTGCAAAGAGGAA

  • Si-DR5 1, GCAAATATGGACAGGACTA

  • Si-DR5 2, TCAATGAGATCGTGAGTAT

  • Si-DR5 3, TCATGTATCTAGAAGGTAA

  • Si-RELA 1, GATTGAGGAGAAACGTAAA

  • Si-control, TTCTCCGAACGTGTCACGT

CRISPR/Cas9

The guide RNAs (gRNA) targeting mouse Gsdme were cloned into the LentiCRISPR-v2 plasmid (Addgene, 52961) according to a prior publication.53 54 For the CRISPR virus package, the lentiviral transfer plasmid was transfected into HEK293T cells with pSPAX2 (Addgene, 12260) and pMD2G (Addgene, 12259) at a ratio of 2: 1: 1. Supernatants of HEK293T cells containing lentivirus were collected 2 days later and used to infect mouse ID8-VEGFHi cell lines. These modified ID8 cells were cultured in a medium with 8 µg/mL puromycin for 1 week, and sorted into single-cell wells for monocolonies expansion. The expression of GSDMD, GSDME, and p65 (protein for Rela) was determined by western blotting.

The following oligonucleotide sequences for the sgRNAs were used (5′−3′):

  • sgRNA-Gsdmd 1, CACCGAGGTTGACACATGAATAACG

  • sgRNA-Gsdmd 2, CACCGTGCGTGTGACTCAGAAGACC

  • sgRNA-Gsdmd 3, CACCGCTGCAACAGCTTCGGAGTCG

  • sgRNA-Gsdmd 4, CACCGAAGCTCCAGGCAGCGTAAAC

  • sgRNA-Gsdme 1, CACCGCACGGACACCAATGTAGTGC

  • sgRNA-Gsdme 2, CACCGCCCATAGGTGTCAGCAACGG

  • sgRNA-Gsdme 3, CACCGAGAAATCACGCACTTCTGCG

  • sgRNA-Gsdme 4, CACCGCTTACCTCTTGACGGCATCC

  • sgRNA-Rela 1, CACCGCGATTCCGCTATAAATGCG

  • sgRNA-Rela 2, CACCGAGCTGATGTGCATCGGCAAG

  • sgRNA-Rela 3, CACCGGTCGGCGTACGGAGGAGTC

RNA isolation and qRT-PCR

Total RNA was extracted using the FastPure Cell/Tissue Total RNA Isolation Kit (Vazyme, China) according to the manufacturer’s instructions and then reverse-transcribed into complementary DNA with HiScript SuperMix (Vazyme, China). The qRT-PCR analysis was performed using ChamQ Universal SYBR qPCR Master Mix (Vazyme, China) on LightCycler 480 (Roche, Basel, Switzerland), and the level of GAPDH was used as a normalization control. The mRNA levels of gene expression were analyzed by the 2-ΔΔCt method. The primers used are as follows (5′−3′),

  • Human GSDMA forward: AGAACAGCACTCTGGAGGTCCA

  • Human GSDMA reverse: CCATCACCACATACAGGTTCTCC

  • Human GSDMB forward: CCAGCAGTATCTGGCTACCCTT

  • Human GSDMB reverse: CCTCCTTTACCGTCTCCAGAGT

  • Human GSDMC forward: CCCATCACCAAACCTGGAAGAC

  • Human GSDMC reverse: TCAACAGCCTCTGTCACCACGT

  • Human GSDMD forward: ATGAGGTGCCTCCACAACTTCC

  • Human GSDMD reverse: CCAGTTCCTTGGAGATGGTCTC

  • Human DFNA5 (GSDME) forward: GATCTCTGAGCACATGCAGGTC

  • Human DFNA5 (GSDME) reverse: GTTGGAGTCCTTGGTGACATTCC

  • Human TNFRSF10B (DR5) forward: AGCACTCACTGGAATGACCTCC

  • Human TNFRSF10B (DR5) reverse: GTGCCTTCTTCGCACTGACACA

  • Human TNFRSF1A (TNFR1) forward: CCGCTTCAGAAAACCACCTCAG

  • Human TNFRSF1A (TNFR1) reverse: ATGCCGGTACTGGTTCTTCCTG

  • Human TNFSF10 (TRAIL) forward: TGGCAACTCCGTCAGCTCGTTA

  • Human TNFSF10 (TRAIL) reverse: AGCTGCTACTCTCTGAGGACCT

  • Human FASL forward: GGTTCTGGTTGCCTTGGTAGGA

  • Human FASL reverse: CTGTGTGCATCTGGCTGGTAGA

  • Human CD95 (FAS) forward: GGACCCAGAATACCAAGTGCAG

  • Human CD95 (FAS) reverse: GTTGCTGGTGAGTGTGCATTCC

  • Human TNF alpha forward: CTCTTCTGCCTGCTGCACTTTG

  • Human TNF alpha reverse: ATGGGCTACAGGCTTGTCACTC

  • Human RELA forward: TGAACCGAAACTCTGGCAGCTG

  • Human RELA reverse: CATCAGCTTGCGAAAAGGAGCC

  • Human GAPDH forward: GTCTCCTCTGACTTCAACAGCG

  • Human GAPDH reverse: ACCACCCTGTTGCTGTAGCCAA

Immunochemistry staining (IHC)

After deparaffinization in xylene and rehydration through graded alcohol solutions, the tissue slides underwent antigen retrieval for 15 min. To block non-specific binding, 5% BSA was applied for 30 min at 37°C. The slides were then incubated with anti-CD4 (Abcam, ab133616), anti-CD8 (Abcam, ab101500), anti-CD19 (Cell Signaling Technology, 90176), anti-CD56 (NCAM1) (Abcam, ab75813), anti-CD14 (Abcam, ab133335), anti-granzyme B (Abcam, ab255598), anti-CD45 (Cell Signaling Technology, 70257), anti-CD4 (Abcam, ab183685), anti-CD8 (Abcam, ab209775), anti-CD11C (Cell Signaling Technology, 97585), anti-CD14 (Abcam, ab182032), anti-Klrb1c/CD161c (NK1.1) (Cell Signaling Technology, 39197), anti-cleaved N-terminal GSDMD (Abcam, ab215203), anti-cleaved N-terminal GSDME (Abcam, ab215191), anti-cleaved caspase 3 (Cell Signaling Technology, 9664), anti-cleaved caspase 8 (Immunoway, YC0011), anti-phospho-NF-κB p65 (Cell Signaling Technology, 3033), and anti-TNF-alpha antibody (Abcam, ab215188) primary antibodies overnight at 4°C. After three washes with tris-buffered saline containing 0.1% Tween ® 20 detergent (T-BST), the slides were incubated with HRP-conjugated secondary antibodies for 1 hour. The HRP activity of the secondary antibodies was determined using the DAB kit (ZLI-9017, ZSGB-BIO, Beijing, China).

Immunofluorescence staining

Cells were seeded in 12-well plates and treated with DMSO, olaparib, or niraparib at different time points. After being washed with PBS, the cells were fixed in 4% paraformaldehyde or absolute ethanol and permeabilized with 0.2% Triton-X in PBS. The anti-HMGB1 (Abcam, ab79823) and anticalreticulin (Abcam, ab92516) primary antibodies were incubated with the cells overnight at 4°C, followed by washing with PBS and incubation with Alexa 594-conjugated secondary antibodies for another 1 hour at 37°C in the dark. Nuclei were counterstained with 4,6-diamidino-2-phenylindole, and images were captured using an OLYMPUS BX41 fluorescent microscope.

Calcein/PI staining

The cells were treated with olaparib or niraparib for different time points, rinsed two times with PBS, and incubated with a Calcein/PI working solution (Beyotime, China) for 15 min at 37°C in the dark. Images were captured using an OLYMPUS BX41 fluorescent microscope. Ten high-power fields (×20) of each slide were randomly selected, and the average number of cells with positive PI staining and negative calcein staining was calculated.

Flow cytometry

Fresh ID8 mouse tumors were collected, cut into small pieces, and dissociated into single-cell suspensions using the tumor dissociation kit (Miltenyi Biotec, Germany) with agitation for 60 min at 37°C. The spleens and draining lymph nodes were mechanically dissociated into single-cell suspensions. The suspensions were filtered through 70 µm strainers, lysed red blood cells, and washed with PBS containing 0.5% BSA. After staining the single-cell suspensions with Fixable Viability Dye (Biolegend, 423105), TruStain FcX PLUS antibodies (Biolegend, 156604) were used to block the Fc receptors. Next, the surface markers were stained with APC anti-mouse CD8a (Biolegend, 100711), FITC anti-mouse CD45 (Biolegend, 103108), PE/Cyanine7 anti-mouse CD11c (Biolegend, 117318), PE anti-mouse NK-1.1 (Biolegend, 108708), PerCP anti-mouse/human CD11b (Biolegend, 101230), BV510 anti-mouse F4/80 (Biolegend, 123135), APC anti-mouse CD86 (Biolegend, 105011), PE anti-mouse I-A/I-E (Biolegend, 107607), BV421 rat anti-mouse F4/80 (BD, 565411), BV421 rat anti-mouse CD86 (BD, 564198), PE-Cy7 anti-mouse CD45 (BD, 552848), BV610 anti-mouse CD4 (Biolegend, 100548), BV510 anti-mouse CD8a (BD, 563068), PE anti-mouse CD107a (Biolegend, 121612) and PE-Cy7 Hamster anti-mouse CD69 (BD, 552879) diluted in PBS containing 0.5% BSA for 30 min on ice. For intracellular staining, T cells were fixed using a Cytofix/Cytoperm kit (BD Biosciences, 554715) for 20 min. After three washes of Perm buffer, the T cells were stained with PerCP/Cyanine5.5 anti-mouse IL-2 (Biolegend, 503822), BV650 anti-mouse lFN-γ (Biolegend, 505832) and FITC anti-human/mouse granzyme B (Biolegend, 372206), BV785 anti-mouse TNF-α (Biolegend, 506341) for 60 min on ice. After three washes of FACS buffer, the fluorescence intensity of these tested markers was determined using CytoFLEX (Beckman Coulter, USA) and analyzed by Flowjo software.

For assessing the cell death ratios, the FITC Annexin V Detection Kit I (BD Biosciences, 556547) was used according to the manufacturer’s instructions. The intensity was determined using CytoFLEX (Beckman Coulter, USA) and analyzed by Flowjo software.

Microscopy

To examine the morphology of the pyroptotic cells, the cells were seeded in six-well plates at 30%–40% confluency and then treated with different drugs. Phase contrast cell images were captured under the OLYMPUS BX41 microscope.

Western blotting

The tumor cells were harvested and lysed using RIPA lysis buffer (Servicebio, Wuhan, China). The protein concentration was measured using the bicinchoninic acid protein assay kit (Beyotime, Shanghai, China). The protein lysate was mixed with five times loading buffer and boiled at 100°C for 10 min. The prepared samples were then loaded into sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) with constant voltage and transferred to polyvinylidene fluoride membranes with a constant current. After blocking with 5% skim milk (Servicebio, Wuhan, China) for approximately 1 hour at room temperature, the membranes were incubated with anti-GSDMA (Abcam, ab230768), anti-GSDMB (Abcam, ab215729), anti-GSDMC (Proteintech, 27 630-1-AP), anti-GSDMD (Abcam, ab209845), anti-DFNA5/GSDME (Abcam, ab215191), anti-cleaved caspase 3 (Abcam, ab32042), anti-cleaved caspase 8 (Cell Signaling Technology, 9746), anti-TNFRSF1A (TNFR1) (ABclonal, A1540), anti-poly/mono-ADP ribose (Cell Signaling Technology, 83732), anti-phospho-histone H2A.X (Cell Signaling Technology, 9718), anti-phospho-ATM (Cell Signaling Technology, 4526), anti-phospho-NF-κB p65 (Cell Signaling Technology, 3033), anti-p65 (Cell Signaling Technology, 8242), anti-FLIP (D5J1E) (Cell Signaling Technology, 56343), and anti-beta-tubulin (ABclonal, AC015) antibodies, and anti-beta-actin (ABclonal, AC026) antibodies overnight at 4°C. After three washes with T-BST, the membranes were then incubated with HRP-conjugated secondary antibodies (1:2000) for 1 hour at room temperature. The blots were visualized using ECL chemiluminescence technology (Bio-Rad).

Caspase 3/8 activation assay

Activated caspase 3 and 8 were detected using FAM-FLICA Caspase 3 Assay Kit (Immunochemistry Technologies, #93) and FAM-FLICA Caspase 8 Assay Kit (Immunochemistry Technologies, #99), respectively, following the instructions. The intensity was determined using a Beckman CytoFLEX Flow cytometer, and the data were analyzed using Flowjo software.

Enzyme-linked immunosorbent assay (ELISA)

The levels of death receptors-related ligands in tumor culture medium were assessed using the TNF-α Human ELISA Kit (DAKEWE, China), FASL Human ELISA Kit (MULTI science, China), and TRAIL Human ELISA Kit (MULTI science, China) following the manufacturer’s protocols. The absorbance was measured at 450 nm using a SpectraMax Paradigm plate reader (Molecular Devices, USA). All samples were measured in duplicate and fitted to a relative standard curve of recombinant protein.

T cells separation and stimulation in vitro

Primary CD4+ and CD8+ T cells were isolated from C57BL/6 mouse spleens using a mouse T-cell enrichment kit (Invitrogen, 8804-6820) and resuspended with 1640 medium supplemented with 10% FBS and 2-mercaptoethanol. Subsequently, these isolated T cells were stimulated with anti-CD3 (BioXCell, BE0002) and anti-CD28 antibodies (BioXCell, BE0328) in the presence or absence of olaparib/niraparib. After 72 hours, the proportion of apoptosis and the expression of functional cytokines were assessed throughflow cytometry.

Bone marrow-derived DCs (BMDC) differentiation and stimulation

The primary BMDCs were differentiated using mouse bone marrow cells with stimulation of 10 ng/mL interleukin 4 (Peprotech, 214-14) and 20 ng/mL granulocyte-macrophage colony-stimulating factor (Peprotech, 315-03). After 5 days of differentiation, 10 ng/mL LPS (Sigma-Aldrich, L7895) with or without 10 µM olaparib/niraparib were added into the culture medium for the other 3 days. Then, the differentiation and maturation were evaluated by flow cytometry.

LDH release assay

LDH release assay was performed using Cytotoxicity Detection Kit (LDH) (Roche, 11644793001) according to the manufacturer’s instructions.

Tumor-infiltrated T-cell killing assay

Tumor-infiltrated T cells were isolated from ID8 tumors with or without niraparib treatment for about 4 weeks by a mouse T-cell enrichment kit (Invitrogen, 8804-6820). To assess the tumor-killing potentials, the tumor-infiltrated T cells were cocultured with ID8 cells at an effector-to-target ratio of 2:1. After 12 hours, the culture supernatants were obtained and LDH levels were measured using a Cytotoxicity Detection Kit (LDH) (Roche, 11644793001) according to the manufacturer’s instructions.

Bulk RNA isolation and sequencing

RNA from A2780 cells treated with DMSO or niraparib was extracted using the Rneasy Mini Kit (QIAGEN, 74104). The concentration of RNA was measured using the NanoDrop instrument (Thermo), and RNA quality was evaluated using the fragment analyzer (AATI). Libraries were constructed using the NEBNext Poly(A) mRNA Magnetic Isolation Module Kit (NEB, E7490L) and the NEBNext Ultra RNA Library Prep Kit (NEB, E7530L) for Illumina PE multiplexed sequencing. Samples were sequenced on the Illumina HiSeq 4000 sequencer with 150 bp PE reads.

Gene set enrichment analysis (GSEA)

The DESeq2 package was used to perform differential expression analyses. The GSEA was performed using the sorted gene list based on log fold change, with the clusterProfiler (V.3.18.0) package.

CCLE analysis

The CCLE was used to interrogate the transcriptomics data of the GSDM family signature in a panel of OC cell lines (https://portals.broadinstitute.org/ccle).

Statistical analysis

Statistical analyses and graphing of the data in this study were performed using GraphPad Prism V.8. The comparison between the two groups was analyzed for statistical significance using a two-tailed Student’s t-test. Time course data in our mouse models were analyzed by two-way ANOVA. All results are represented as the values±SEM of at least three independent experiments. P<0.05 was considered statistically significant and denoted as follows: *p<0.05, **p<0.01, and ***p<0.001.

Data availability statement

Data are available on reasonable request. Raw data from bulk TCR-seq of paired clinical specimens has been deposited in the Genome Sequence Archive for Human at the National Genomics Data Center with accession number HRA007230 under project PRJCA016620 (https://ngdc.cncb.ac.cn/bioproject/browse/PRJCA016620). The deidentified bulk TCR-seq profile is available at the NCBI’s Gene Expression Omnibus (GEO) database (accession number: GSE222557). Raw data from bulk TCR-seq of ID8 tumors has been deposited in GSA (https://ngdc.cncb.ac.cn/gsa/s/kCskQmer) with accession number CRA014711. The data from bulk RNA-seq of A2780 cells is available in OMIX (https://ngdc.cncb.ac.cn/omix/preview/ZAxvMET5) with accession number OMIX005758. Graphs of quantitative data and representative images of blot and gel generated in this study are available from the corresponding author upon reasonable request.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants. The Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (approval number: TJH2020S122-5). Participants gave informed consent to participate in the study before taking part.

Acknowledgments

We thank Zai Lab for providing niraparib.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • YX, PH, Y-yQ and ZW contributed equally.

  • Contributors YX and PH designed the research study. YX, PH, Y-yQ, ZW, NJ, XinL, S-yW, PJ, ZZ, and PL conducted experiments and analyzed data. Y-yQ, WP, and XiongL performed the bioinformatic analyses. YX, EKD, QZ, YF, and X-pY wrote and revised the manuscript. Q-lG and ZH supervised the entire project. Q-lG acts as the guarantor.

  • Funding This study was supported by the National Key R&D Program of China (2022YFC2704200), the National Science Foundation of China (#81902933, #82372928, and #82072889), and the Natural Science Foundation of Hubei Province (#2023AFB760 and #2023AFB1017).

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.