PARPs are a family of proteins that are essential for several cellular processes including DNA repair, replication fork stability and genomic stability (Schreiber et al., 2006; Lord and Ashworth, 2017; Forment and O’Connor, 2018). PARP-1 and 2 act as DNA damage sensors, rapidly binding to breaks in DNA strands to hydrolyse NAD⁺ and produce linear and branched PAR chains in a process called poly(ADP-ribosyl)ation (PARylation) (Krishnakumar and Kraus, 2010; Rouleau et al., 2010; Murai et al., 2012). PARylation of chromatin proteins recruits DNA repair proteins to sites of damage causing PARP-1/2 to then dissociate from DNA via auto-PARylation (El‐Khamisy et al., 2003; Schreiber et al., 2006; Murai et al., 2012). The PARylation by PARP is not only important in DNA repair but chromatin modulation, regulation of DNA transcription and replication, protein degradation and cell cycle (Schreiber et al., 2006; Martí et al., 2020).

PARPi bind to the catalytic domains of PARP-1/2 and compete with NAD⁺, inhibiting PARylation, effectively disrupting recruitment of DNA repair proteins and PARP dissociation, thereby “trapping” PARP-1/2 on damaged DNA, and further reviewed in (D’Andrea, 2018; Wakefield et al., 2019). The trapping of PARP proteins on DNA stalls replication forks leading them to become dysregulated and collapse (Murai et al., 2012; Wakefield et al., 2019). Active PARP-1 regulates replication fork progression and when inhibited, replication fork stalling leads to a majority of single-stranded breaks (SSBs) being processed into double-stranded breaks (DSBs) (Berti et al., 2013). In healthy cells, these DSBs are repaired by the high-fidelity homologous recombination (HR) DNA repair pathway to successfully repair the damage (Ter Brugge et al., 2016).The HR DNA repair pathway is pivotal in accurate repair of DSBs and restarting stalled/collapsed replication forks, with BRCA1 and BRCA2 being crucial for the protection of replication forks during replication stress (Chen et al., 2018). In HGSOC cells with mutant BRCA1/2 or other defective HR genes, the inhibition of PARP forces cancer cells to rely on error-prone repair DNA pathways or otherwise unrepaired damage persist into mitosis, leading to the rapid accumulation of mutations, genomic instability, and eventual cell death. The dual loss of the HR pathway and PARP function is synthetically lethal, in that the simultaneous inhibition of the two pathways leads to cell death, whereas loss of only one does not (Ashworth and Lord, 2018). It is within this realm of synthetic lethality that PARPi works best, as seen in the treatment of women with BRCA-mutant HGSOC experiencing sustained and profound responses to PARPi, compared with women with HR proficient carcinomas.

There are currently several PARPi available including olaparib, rucaparib, niraparib, veliparib, pamiparib and talazoparib being tested in phase III trials, with the first three mentioned having both Food and Drug Administration (FDA) and European Medicines Agency (EMA) approval for use in OC in the clinic (Pilié et al., 2019; Lee and Matulonis, 2020). A key feature of all PARPi molecules is a benzamide moiety that binds to the catalytic center of PARP, disrupting enzymatic activity. However, the disruption of catalytic activity alone is not enough to explain the vastly different outcomes in anti-tumour responses and efficacy in the clinic (Sun et al., 2018; Kim et al., 2021). The most effective PARPi trap PARP at sites of DNA damage and this could be due to difference in size and flexibility of each molecule influencing how PARPi bind and effect conformational changes. Allosteric destabilization of a critical helical regulatory domain neighboring the catalytic domain was crucial for cytotoxic and PARP-trapping effects with this being most prominent for rucaparib, niraparib and veliparib compared with olaparib and talazoparib (Zandarashvili et al., 2020). Talazoparib was reported to trap PARP roughly 100-fold more than niraparib, olaparib and rucaparib (Murai et al., 2014). However, the capacity for PARPi trapping does not relate to overall clinical benefit, as talazoparib is also noted for having increased toxicity in the clinic (Murai et al., 2014).

Additionally, another aspect of PARPi to consider is substrate selectivity and specificity. Most PARPi are highly selective toward PARP-1/2 although, computational in-silico analyses have uncovered 58 potential interactions with kinases of which only 10 were previously known (Thorsell et al., 2017; Antolin et al., 2020). Supporting this is evidence of rucaparib inhibiting the activity of kinases CDK16, PIM3 and DYRK1B in catalytic inhibition assays and additionally niraparib inhibiting the activity of two others, DYRK1A and DYRK1B (Antolin et al., 2020). Additional research in deciphering the specific mechanisms unique to each PARPi may elucidate novel pathways for clinical benefit. Investigation into improving PARPi specificity, tolerability and pharmacokinetic properties continues, with several PARPi in phase I/II clinical trials; including senaparib, which is 20-fold more potent than olaparib, and the highly selective, PARP-1 specific, PARP-1-DNA-trapper, AZD5305 (Cao, 2019; Johannes et al., 2021).

The PARPi olaparib and niraparib have been approved by both the EMA and FDA for use as maintenance therapy after response to first-line treatment with chemotherapy for women with germline or somatic BRCA1/2 mutations or platinum-sensitive HGSOC respectively (Veneris et al., 2020). Additionally, olaparib, rucaparib, and niraparib are approved for use as a maintenance treatment for recurrent platinum-sensitive HGSOC patients and in some additional recurrent OC settings.

The phase III SOLO-1 trial evaluated the efficacy of olaparib in women with advanced BRCA-mutant platinum-sensitive HGSOC and demonstrated a 67% decrease in risk of disease progression or death (hazard ratio [HR] 0.33; 95% confidence interval [CI]: 0.25–0.43). Strikingly, at 5-year of follow-up, the PFS for the placebo arm was 13.8 months compared with the olaparib arm, on which women had achieved an unprecedented 56.0 months PFS, a 4-fold improvement, and 48% of women on olaparib remained disease free at this time, compared with only 20.5% of women on the placebo arm (Banerjee et al., 2021). Similarly, the phase III PRIMA/ENGOT-OV26/GOG-3012 trial (NCT026555016) examined responses to niraparib in platinum-sensitive advanced HGSOC and high grade endometrioid OC, regardless of BRCA mutation and/or HRD status (González-Martín et al., 2019). A significant improvement in PFS on niraparib maintenance was observed in the overall population with a median PFS of 13.8 months compared with 8.2 months for the placebo (HR 0.62, CI 0.50–0.76, p < 0.001). Roughly 50% of women were classified as having HGSOC with HRD and the greatest benefit derived from niraparib was seen in the subset of these with BRCA-mutations (median PFS 22.1 versus 10.9 months, HR 0.40, CI 0.27–0.62); followed by that observed in the non-BRCA HRD HGSOC subset (19.6 versus 8.2 months, HR 0.50, CI 0.31–0.83); lastly the remaining ∼50% of women had HGSOC which was HR proficient and responded the least well to PARPi (8.1 versus 5.4 months, HR 0.68, CI 0.49–0.94) (González-Martín et al., 2019). As seen in both clinical trials, response rates in women with HR proficient and platinum-resistant HGSOC were modest in comparison to HRD and platinum-sensitive HGSOC, thus a spectrum to the benefits derived from PARPi was observed. The combination of PARPi with other drugs to induce HRD in HR proficient disease or targeting other pathways that PARP-deficient tumours rely on, may be the answer to improving response further in HGSOC.

Regardless of the efficacy of PARPi as a monotherapy, a growing concern is the development of resistance with the prolonged use of PARPi. There are five main classes of resistance that have been characterised; drug efflux, changes in PAR metabolism, mutational changes of binding sites or target proteins, rewiring of stalled fork replication and restoration of the HR pathways (Wakefield et al., 2019). Several articles have reviewed these mechanisms in detail (Noordermeer and van Attikum, 2019; Wakefield et al., 2019; Kubalanza and Konecny, 2020), however the relevance of the different resistance mechanisms will need to be studied in large clinical cohorts for a better understanding of the selective pressures from PARPi in tumour evolution. The resistance landscape in patients is likely more diverse than what has been observed in research settings to date, thus developing a better understanding of the diversity could better inform therapeutic strategies moving forward. New technologies, including in proteomics (e.g., mass spectrometry and protein array analysis), that allow for the dissection of underlying molecular signaling events, could reveal clinically relevant biomarkers and new therapeutic choices for HGSOC, especially in the setting of the prediction and analysis of acquired PARPi resistance (Ghose et al., 2022). However, with our current knowledge, instigating early treatment with PARPi, rapid retreatment upon relapse and use of PARPi in combination therapies are important tools in maximizing PARPi efficacy. Treating early in the upfront maintenance setting, having first performed molecular analysis during first-line chemotherapy in order to match the HGSOC to appropriate combination PARPi therapy, may yield the most success in the treatment of highly heterogenous HGSOC.

In a normal state, DNA is wound around histones and non-histone proteins to form highly compact structures known as chromatin. When access to DNA is required, chromatin structures relax, unravelling bound DNA to allow protein complexes to bind and function, this reorganization of bound DNA is called chromatin remodelling (Sinha et al., 2021). Chromatin remodelling is important for maintaining genomic stability and is important in processes like DNA transcription, replication, and repair. PARP-1 plays an important role in these processes, by regulating chromatin remodelling via PARylation of the histones. In its latent state, PARP-1 is found bound to linker DNA and/or histone proteins, resulting in the condensed structure of chromatin called heterochromatin. In this state, no transcription machinery can access the DNA, repressing gene transcription. In the presence of DNA damage however, PARP-1 becomes active (Kim et al., 2004; Muthurajan et al., 2014). Active PARP-1 PARylates itself and histones, promoting the remodelling of chromatin to become euchromatin as the addition of negatively charged PARs on the histones repels DNA. At sites of DNA damage, histone PARylation causes its eviction from DNA strands, facilitating the recruitment of other chromatin remodelers and further loosening of chromatin for subsequent recruitment of DNA repair proteins (Quénet et al., 2009). More specifically, PARP-1 PARylates and then binds to chromatin remodelers at their PAR-binding domain for subsequent alteration of chromatin structures (Andronikou and Rottenberg, 2021). For example, PARylation of the lysine specific demethylase 4D (KDM4D) at its C-terminal promotes demethylation of the methylated forms of H3K9, reducing chromatin compaction, and allowing gene transcription (Khoury-Haddad et al. 2014). However, PARylation of KDM4D at its N-terminal inhibits the action of this enzyme at the promoter of retinoic acid receptor-dependent genes and represses gene transcription (Le May et al., 2012). In this case, the use of PARPi may abolish this specific PARP activity in chromatin remodelling machinery. Particularly, PARPi interferes with the recruitment of KDM4D to double stranded breaks and thus inhibits the repair process (Khoury-Haddad et al. 2014).

There are several natural inhibitors that can counteract PARP-1’s involvement in chromatin remodelling machinery, one of them is poly-ADP-ribose glycohydrolase (PARG). PARG counteracts the action of PARP-1 by cleaving the PAR on PARylated PARP-1, rendering it inactive (Kim et al., 2004). Amplified in liver cancer protein 1 (ALC1) is a chromatin remodeler that is rapidly recruited to DNA-damage and binds to PARylated PARP-1 (Pines et al., 2012; Ahel et al., 2009). When ALC1 binds to PARylated PARP-1, it not only activates the protein but secondarily protects PAR on PARP-1 from PARG hydrolysis (Ahel et al., 2009; Gottschalk et al., 2012; Singh et al., 2017). Loss of ALC1, and subsequent loss of PAR protection by ALC1, was found to enhance PARP-1/2 trapping on DNA by PARPi, effectively sensitising cells to PARPi (Blessing et al., 2020; Juhász et al., 2020).

Another natural inhibitor of PARP-1 activity is macroH2A1.1 which binds to autoPARylated PARP-1 to prevent PAR hydrolysis, which can promote chromatin recondensation to interfere with transcriptional processes (Timinszky et al., 2009). MacroH2A1.1 is a splice variant of macroH2A1, which is recruited to DSBs and is implicated in regulating PAR metabolism and NAD + turnover (Ruiz et al., 2019). The alternative splice variant, macroH2A1.2, interacts with other enzymes to recondense chromatin through the production of H3K9 methylation marks (Khurana et al., 2014; Alagoz et al., 2015). These compact chromatin marks attract BRCA1, promoting the use of the HR pathway to repair DNA damage (Lee et al., 2013; Alagoz et al., 2015). Loss of macroH2A1.1 has been noted in several cancers and, due to its roles in chromatin condensation and BRCA1 recruitment, depletion of this histone may increase PARPi sensitivity (Ruiz et al., 2019).

DNA methylation is another major epigenetic modification which occurs at the fifth carbon of cytosine when followed by guanine (CpG) in eukaryotic genomes. The methylated cytosine (5 mC) is induced and maintained by DNA methyltransferase (DNMT). Promoter hypermethylation commonly promotes gene silencing. This epigenetic silencing has been observed in HR genes, including BRCA1 or RAD51C, occurring as an early clonal event, contributing to the development of OC cases (Alsop et al., 2012) BRCA1 can in fact partially predict BRCAness in OC (Aref-Eshghi et al., 2020). Homozygous methylation (of all copies present) of the BRCA1 promoter can predict sensitivity to PARPi therapy. On the other hand, heterozygous methylation, (loss of methylation of any copy of the gene present in the cancer), correlates with PARPi resistance (Kondrashova et al., 2018). Similarly, for RAD51C, complete gene silencing correlates with PARPi response whilst loss of methylation of even one allele of RAD51C drives resistance to PARPi (Nesic et al., 2021).

PARP-1 interaction with DNMT1 contributes to the regulation of DNA methylation. PARylation has been shown to maintain unmethylated CpG at specific sites of the genome, while blockade of PARylation increases DNA methylation levels in the genome (DE CAPOA et al., 1999; Zampieri et al., 2012). Interestingly, the modulation of DNA methylation by PARP-1 can be counteracted by CCCTC-binding factor (CTCF), that can induce the auto-modification of PARP-1 (Guastafierro et al., 2008; Zampieri et al., 2012). Pharmacological inhibition of PARP-1 has been shown to change the genome-wide DNA methylation profile, confirming PARP involvement in DNA methylation processes (Nalabothula et al., 2015). Besides inducing more DNA methylation, PARPi has been shown to induce the expression of enhancer of zeste homolog 2 (EZH2), a histone methyltransferase that catalyses trimethylation of the lysine residue on histone3 (Martin et al., 2015). This in turns results in a genome-wide increase of H3K27me3 and thus chromatin compaction (Martin et al., 2015). Both increases in DNA methylation and H3K27me3 in the genome result in increased heterochromatin structure and further silencing of various genes.

With important roles in DNA regulation, PARP-1/2 play a role in T-cell development, differentiation, and function. The development of T-cells is a complex and highly regulated process that begins in the thymus with bone marrow-derived lymphoid precursors and through well-characterized maturation steps give rise to mature T-cells (Koch and Radtke, 2011). PARP-1 modulates activity of nuclear factor of activated T-cells (NFAT) which drives CD4⁺ T-cell differentiation (Olabisi et al., 2008; Valdor et al., 2008). A reduction in the expression of NFAT reliant cytokines was observed in PARP-1 deficient T-cells and furthermore PARP-1 deficiency creates a bias for CD4⁺ T-cell differentiation to a Th1 phenotype (Macian, 2005; Olabisi et al., 2008; Sambucci et al., 2013). The Th1 subset of CD4⁺ T-cells is associated with the production of cytokines such as IFN- γ, IL-2 and TNF -β that induce inflammation and cell-mediated immune responses (Constant and Bottomly, 1997). The Th2 subset promotes B-cell proliferation and differentiation through IL-4 and IL-5 cytokine production and is associated with humoral-type immune responses (Constant and Bottomly, 1997). There is conflicting data on PARP-1 deficiency driving Th1 differentiation of CD4⁺ T-cells, with one study in a model of airway inflammation observing olaparib promoting the Th1 phenotype whereas a model of inflammatory arthritis observed PARP inhibition associated with a suppression of Th-1-associated cytokines. Thus, PARP driven Th1 differentiation is likely mediated by other context-specific factors.

During early T-cell development, PARP2 is essential for the development of CD4/CD8 double positive thymocytes and PARP-1 regulates expression of Foxp3 in CD4⁺ T-cells (Zhang et al., 2013; Luo et al., 2015; Navarro et al., 2017). In-vitro studies show that PARP-1 and PARP2 deficient T cells have a decrease in total CD8⁺ and CD4⁺ populations (Navarro et al., 2017). This observation was prevalent in singular deficiencies and with the dual loss of both PARP-1 and PARP2, with the dual loss having a more dramatic reduction suggesting a, and there was prevalent amounts of DNA damage detected suggesting the reduction in these T-cell populations are a result of accumulating DNA damage and genomic instability and not entirely a block in maturation (Navarro et al., 2017). In the circumstance of PARP-1 deficiency, populations of CD4⁺ T-cells expressing Foxp3 increases due to the lack of Foxp3 PARylation for subsequent degradation (Luo et al., 2015). Expression of Foxp3 on CD4⁺ T-cells causes differentiation into Tregs which are immunosuppressive through their production of inhibitory cytokines, mediation of cytolysis and modulation of DC maturation or function (Vignali et al., 2008; Zhang et al., 2013; Luo et al., 2015). In vivo models studying olaparib in BRCA1-deficient ovarian cancer observed significantly increased proportions of CD4⁺ and CD8⁺ effector T-cells infiltrating intratumorally and peripherally and, notably, an increase in intra-tumoral CD4/Foxp3+ Tregs was not seen (Ding et al., 2018). Suggesting that treatment with PARPi in an in vivo setting does not disrupt T-cell development and function to the extent of tumour benefit. Additionally, olaparib-treated CD8⁺ T-cells showed reduced expression of immune receptors, such as PD-L1, that are associated with T-cell inhibition and exhaustion and produced higher levels of TNFα and IFNγ (Ding et al., 2018).

The recruitment of DC to the tumour microenvironment is an important step in the anti-tumour immune response as they play roles in activating and inducing the differentiation of T-cells (Patente et al., 2019). There is evidence PARPi has an indirect effect in activating DC though its DNA-damaging abilities and creation of cytosolic DNA fragments. Cytosolic DNA activates the cGAS/STING pathway within the cell but can also be exocytosed to activate STING pathways in neighbouring DC (Mouw et al., 2017; Pantelidou et al., 2019). The activation of cGAS/STING was noted in a PARPi treated BRCA1-deficient mouse model of TNBC, but not in DC treated with PARPi alone (Pantelidou et al., 2019). This suggests that DC cGAS/STING activation is not induced by PARPi alone. This notion was supported in a BRCA1 deficient model of OC, with activation of cGAS/STING observed upon treatment with olaparib (Ding et al., 2018). To confirm the paracrine effect of PARPi on DC activation, PARPi treated ovarian cells were co-cultured with naïve DC. Increased levels of TBK-1, IRF3, CXCL10 and IFNβ were observed, confirming cGAS/STING activation and expression of downstream genes, further supporting the indirect activation of DC upon treatment with PARPi (Ding et al., 2018). Furthermore, treatment with PARPi increased DC populations with increased antigen presentation machinery, specifically upregulated costimulatory CD80 and CD86 and antigen presenting major histocompatibility complex class II (MHC II) (Ding et al., 2018; Pantelidou et al., 2019).

Natural killer (NK) cells are effector lymphocytes utilized in the innate immune response against “non-self” cells and “self” cells undergoing stress in the form of infections or malignant transformations (Vivier et al., 2004). When activated, NK cells either have direct cytotoxic attacks on targets or produce large arrays of cytokines and chemokines to initiate antigen-specific immune responses. Specifically, NK cells can directly interact with cells through TNF-related apoptosis-inducing ligand (TRAIL) and the Fas ligand to induce apoptosis or indirectly through secretion of IFNγ and TNFα (Barrow and Colonna, 2017; Souza-Fonseca-Guimaraes et al., 2019). In tumour cells, TRAIL stimulates PARP-1 activation and subsequently the PARylation of high-mobility group box protein 1 (HMGB1) which results in HMGB1 localisation to the cytoplasm. This localisation from nucleus to cytoplasm promotes an autophagic response and protects the tumour cells from TRAIL mediated apoptosis (Yang et al., 2015). Treatment with PARPi suppressed the PARP-1/HMGB1 pathway and re-sensitized tumour cells to TRAIL induced cell death suggesting PARPi can sensitize tumours to NK-cell mediated apoptosis (Yang et al., 2015). Treatment with PARPi has also been shown to upregulate death receptors Fas and death receptor 5 in several cancer cell lines (Meng et al., 2014). Upregulation of these receptors sensitized cells to TRAIL induced apoptosis. This was further supported by a study observing NK cell killing in prostate cancer cells with and without PARPi treatment independent of BRCA status (Fenerty et al., 2018). They found treating tumour cells with olaparib upregulated death receptor TRAIL-2 and significantly increase tumour cell sensitivity to NK cell killing in both BRCA-wildtype and BRCA-mutant cells. Additionally, they replicated these results in additional tumour cell lines, including breast, chordoma, non-small cell lung carcinoma (Fenerty et al., 2018). It is important to note that the presence of NK cells has a favourable impact on OS for HGSOC patients and these findings suggest that PARPi can recruit and sensitize tumour cells to NK cells, and that NK cells contribute to PARPi anti-tumour effects (Henriksen et al., 2020).

Tumours with existing DNA repair defects initially stimulate inflammation and T_H1 immune responses, however, maintaining a constant level of DNA damage and subsequent chronic inflammatory responses encourages the infiltration of immune-suppressive cells (Schreiber et al., 2011; Crusz and Balkwill, 2015; Fridman et al., 2017). Chronic inflammation promotes immunosuppression and in cancer this promotes tumour progression. However, studies suggest PARPi has the potential to counteract this and reinvigorate the anti-tumour immune response (Fridman et al., 2017).

Inflammatory responses can be promoted by PARP-1 through its regulation of several transcription factors, cytokines and chemokines (Pazzaglia and Pioli, 2020). Nuclear factor κB (NF-κB) is a transcription factor that is important in the regulation of genes for inflammatory, apoptotic and cell proliferative responses, and for complete NF-κB-dependent gene transcription PARP-1 acetylation is required (Hassa et al., 2005). Additionally, PARP-1 can activate NF-κB through several mechanisms including through the mono-ubiquitination of NF-κB essential modulator (NEMO) for NF-κB nuclear translocation and sustaining toll-like receptor (TLR) induced NF-κB activation (Stilmann et al., 2009; Hinz et al., 2010; Ji et al., 2016). To study whether treatment with PARPi affects inflammatory responses, Alvarado-Cruz et al. (2021) interrogated BRCA1-mutant triple negative breast cancer (TNBC) cell lines and tumours samples after treatment with veliparib for three weeks. Upregulation of hallmark inflammatory TNFα pathways were observed after treatment with PARPi specifically in the BRCA1-deficient cells, and the mechanistic basis for this upregulation was through the cGAS/STING pathways. Separately, a study using genetically engineered mouse models (GEMM) of HGSOC investigated the effects of olaparib in BRCA1 deficient and BRCA1 wildtype settings. Treatment with olaparib elicited an increase in CD4⁺ and CD8⁺ T-cells as well as a pronounced increase in IFNγ and TNFα (Ding et al., 2018). The increase in CD4⁺ and CD8⁺ T-cells was also associated with the increased presence of DC with a potent antigen presenting capacity, and a decrease in MDSCs in the tumour, spleen and blood (Ding et al., 2018). These responses seen were restricted to the BRCA1 deficient GEMM and mechanistically were associated with the stimulation of the STING pathway. It was proposed that PARPi induced DSBs creating cytosolic DNA fragments that are bound by cGAS, activating STING, and subsequently the production of pro-inflammatory cytokines and IFN responses (Pantelidou et al., 2019; Shen et al., 2019). The recent study by Bruand et al.(2021) suggests that BRCA mutant cells are intrinsically programmed for the cGAS/STING/IFN signalling seen in HGSOC and that PARPi enhances this signalling by increasing the amount of ctDNA present. This potentially explains why the immune responses observed after treatment with PARPi were isolated to BRCA1 deficient GEMM models.

Another study interrogated the role of STING in relation to PARPi in both in vitro and in vivo experiments regardless of BRCA1/2 status (Shen et al., 2019). Treating OC cells with talazoparib markedly elevated the phosphorylation of two key components along the STING pathway, IRF3 and TBK1. An increase of total IRF3 and TBK1 translocation from the cytoplasm to the nucleus was observed, suggesting functional signaling of STING (Shen et al., 2019). Additionally, CCL5 and CXCL10, which are two major chemokines activated by STING that positively correlate with the presence of CD8⁺ T-cells, were seen to be upregulated post PARPi treatment. The knockdown of STING, TBK1, IRF3 or cGAS significantly reduced the upregulation of CCL5 and CXCL10 in PARPi treatment in OC cell lines. Further work in mouse models validated these findings showing treatment with PARPi elicited the expression of CCL5 and CXCL10 and induced higher percentages of CD8⁺ T-cells and PD-L1+ cells infiltrating the TME. Treatment with PARPi had no therapeutic effects in immunodeficient mice but prolonged survival and limited tumour growth in immune competent mice (Shen et al., 2019). Additionally, knockout of STING abolished the anti-tumour effects of PARPi establishing PARPi efficacy is based in an immunogenic response. These results do not correlate with the previous studies mentioned thus further work to determine the role of BRCA-loss in STING and IFN responses in HGSOC is needed, which will hopefully further elucidate mechanisms by which PARPi invigorates immune responses.

Both studies also noted the increased expression of programmed cell death-ligand 1 (PD-L1) in cells when treated with PARPi (Ding et al., 2018; Shen et al., 2019). Programmed death-ligand 1 is the ligand of PD-1, which is an immune receptor expressed on CD4⁺/CD8+ T-cells and B cells, and mediates the inhibition of T-cell proliferation and IFNγ production (Iwai et al., 2017). The role of PD-1/PD-L1 is to mediate autoimmune responses, however in cancer, the upregulation of immune checkpoints can be used to suppress the anti-tumour immune response. Upregulation of PD-L1 expression can occur through several mechanisms including PD-L1 promotor binding by NF-kB, JAK1/2 activation and IFNy secretion following type I IFN response (Bellucci et al., 2015; Chabanon et al., 2019; Cai et al., 2020). All of these effects can be stimulated through the cGAS/STING pathway that is activated upon treatment with PARPi. Shen et al., found treatment with PARPi increased percentages of PD-L1+ cells and explored the effects of combining talazoparib and an anti-PD-L1 antibody. Tumours treated with the combination therapy significantly reduced tumour burden compared to either monotherapy, and had the most significant increase in CD8⁺ cell recruitment (Shen et al., 2019). Additionally Jiao et al. (2017) demonstrated that PARPi induced PD-L1 expression regardless of BRCA-status, and effectively reduced the efficacy of active cytotoxic T-cells. Combination of olaparib and an anti-PD-L1 antibody desensitised PARPi treated cells and found the combination more effective than either agent alone (Jiao et al., 2017). Due to the encouraging results of PARPi and checkpoint inhibitor combinations in research, this combination is being explored in several clinical trials.