Facing inevitable PARPis resistance: Mechanisms and therapeutic strategies for breast cancer treatment

BRCA1/2 gene mutations, which result in a dysfunction of homologous recombination repair, have been discovered in at least 5% of breast cancer (BC) patients with the increase in BC incidence in recent years. PARP inhibitors (PARPis), the first drugs with clinical approval based on synthetic lethality, have been approved to treat BRCA1/2‐mutant BC. However, as with other targeted drugs, PARPis drug resistance has become a significant obstacle in the application of PARPis. In this paper, we discuss the mechanism of PARPis, the clinical application of PARPis as monotherapy and the possible induced resistance mechanism. By exploring the resistance mechanism, we aimed to identify appropriate effective therapeutic techniques to overcome PARPis resistance and improve the efficacy of PARPis as well as to provide theoretical and experimental evidence for the clinical use of PARPis in BRCA1/2‐mutant BC.

history, age, lifestyle factors, and hormonal changes. Additionally, there are regional and ethnic variations in the incidence of BC.
Based on the hormone receptor (estrogen receptor [ER] and progesterone receptor [PR]) and HER2 (human epidermal growth factor receptor-2) status, BCs are clinically categorised into three primary subtypes: luminal ER-positive and PR-positive (luminal A and luminal B), HER2-positive, and triple-negative breast cancer (TNBC). Up to 10% of BC patients have inherited (germline) DNA mutations with BRCA1/2 genes containing the most prevalent BC-related germline mutations. According to current world survey statistics, more than 5% of BC patients have mutations in the BC gene (BRCA1/2), and the cumulative risk of BC in carriers of this mutation is 69%-72% by the age of 80. 3 BRCA1/2 germline mutations are the strongest known genetic risk factors for epithelial ovarian cancer and are found in 6%-15% of women with epithelial ovarian cancer. BRCA1/2 carriers with epithelial ovarian cancer respond better than non-carriers to platinum-based chemotherapies. This yields greater survival, even though the disease is generally diagnosed at a later stage and a higher grade. 4 In China, the frequency of total BRCA1/2 pathogenic variants in BC patients is about 5.7%, but the overall incidence of BRCA1/2 pathogenic variants in high-risk BC patients (early-onset, family history, male, and bilateral BC patients) can be up to 14.4%, specifically: early-onset BC patients 7.4%, family history BC patients 15.9%, male BC patients 14.5%, and bilateral BC patients 16.6%. 5 Women with the BRCA1/2 mutation are more likely to develop BC than average women. 6 In addition, research indicates a correlation between BRCA mutations and hormone receptor status: people with gBRCA mutations are more likely to develop TNBC than the general population from 9.3% (Australia) to 15.4% (United States) in unselected TNBC studies. 7 In China, of all molecular subtypes, TNBC demonstrated the highest frequency of BRCA1/2 pathogenic variants (11.2%), while the frequency of HER2-positive BC subtypes was the lowest (1.7%). 5 With the development of various molecular biomarkers and targeted therapy research in BC patients, BC treatment strategies are diversifying. However, adverse treatment events, cancer recurrence and metastasis, and drug resistance continue to plague clinical practice. This review mainly focuses on the clinical use, drug resistance mechanism, and mechanism of action of the targeted medication PARP inhibitors (PARPis). Additionally, several therapeutic strategies to enhance PARPis treatment effects based on the mechanism of drug resistance to PARPis are proposed.

BREAST CANCER
In addition to traditional surgery, radiotherapy, and chemotherapy methods, numerous immunotherapies and targeted therapies have been developed for clinical use in the treatment of BC. 8 However, BC patients with BRCA mutations still have few treatment options. These mutations cause individuals to develop BC at a younger age and at a higher risk of invasive metastasis and recurrence, which result in loss of the chance for surgical treatment at the time of initial diagnosis and a worse prognosis. 9 Currently, chemotherapy is the primary treatment modality for BC, whereas endocrine therapy is the preferred format for HR-positive diseases. 10 Even so, many patients still experience recurrence, some of whom were initially identified as having metastatic lesions, and go on to die from BC. Therefore, the emergence of the biomarkertargeted oral drug PARPi has significant implications for BC treatment, introducing a novel concept for BC treatment: synthetic lethality. PARPis (olaparib or talazoparib) are now the preferred treatment option for advanced TNBC patients with BRCA1/2 mutations, 11 as many experiments have proven that it improves progression-free survival compared to single chemotherapy.

| Functions of PARP
Base excision repair, nucleotide excision repair, mismatch repair, homologous recombination (HR) repair, and nonhomologous end joining (NHEJ) are the five significant ways in which DNA damage is repaired in healthy cells, maintaining genomic integrity. Many proteins are involved in DNA damage repair, including PARP enzymes (PARP1 and PARP2), which are essential for DNA damage repair. Their primary function is to identify single-stranded DNA breaks (SSBs) and double-stranded DNA breaks (DSBs), initiate DNA repair mechanisms, and stabilise replication forks during the repair process. 12 PARP1 is capable of quickly identifying SSBs, binding to DNA at the SSBs site, and then performing PARP poly (ADP-ribos)ylation (PARylation). 13 When PARP PARylates, it recruits repair factors to the lesion and electrostatically breaks the PARP1-DNA connection, thereby releasing PARP1 from the DNA and initiating the DNA repair mechanism ( Figure 1). 14 However, despite the vital role of PARPs in DNA damage repair, PARP1 knockout mouse models did not exhibit developmental defects or early onset of cancer. 14 This may be because when PARP1 is knocked down, sequential SSBs result in more cytotoxic DSBs, which activate the HR repair system and ensure genomic stability and integrity. Furthermore, the alternative role of PARP2 may also account for this experimental result. 14 All of these results suggest that PARP inhibition alone does not lead to significant cell death.

| Mechanism of action of PARPis
In 2005, two critical studies demonstrated synthetic lethal (SL) interactions between PARPis and BRCA1 or BRCA2 mutations ( Figure 2), 15,16 providing theoretical support for PARPis as anticancer drugs in clinical trials. BRCA1/2 protein is essential for DNA homologous recombination repair (HRR). 17 In the process of DNA repair, NHEJ is usually faster than HRR, but DNA end resection is the primary factor that determines whether repair is carried out via NHEJ or HRR 18 During the HRR process, when a DSB occurs, BRCA1 counteracts 53BP1 to promote DNA end resection by inhibiting NHEJ. BRCA1 then interacts with the PALB2 protein, which binds to BRCA2 to recruit RAD51 to the DNA damage site, promote the formation of RAD51 filaments, invade the homologous chromatid, and use it as a template for repair. In addition, BRCA1/2 also protects the stalled replication fork from degradation during replication stress. In summary, BRCA1/2 not only repairs DNA damage but also prevents its occurrence. 19 Therefore, when BRCA1/2 is mutated, the HRR function is impaired, and the cells then adopt low-fidelity repair methods, such as NHEJ, which directly join the two broken DNA ends. These repair methods result in DNA alterations and genetic material loss, 12 which may lead to the development or progression of cancer. In addition, this HR deficiency increases chromosomal instability in tumour cells, resulting in increased sensitivity to DNA-damaging drugs (e.g., platinum compounds and PARPis).
The main events of PARPi-induced BRCA-deficient cell death are the inhibition of poly ADP-ribosylation (PAR) synthesis, formation of double-strand breaks, and collapse of replication forks. When PARPis bind to PARPs and prevent PARylation, the ability of PARP1 to repair SSBs is reduced. Simultaneously, PARP is trapped in the DNA strand complex, leading to the accumulation of unrepaired SSBs. 20,21 During the S phase of the cell cycle, these unrepaired SSBs transform into DSBs, 22 which are lethal to BRCA1/2-mutant cancer cells. In normal cells, the HRR pathway can repair DSBs, whereas in BRCA1/2-mutant cells, the lack of HRR results in the accumulation of DSBs, collapse of the replication fork, and eventually cell death. 23 Thus, PARPis have less impact on healthy cells and are more effective than chemotherapeutic agents. In addition, DSBs can be repaired via the nonhomologous endjoining pathway when PARPase is inhibited in BRCA1/2-mutant cells; however, the low fidelity of DSB repair via this pathway can result in the death of tumour cells. 12 F I G U R E 1 Functions of PARP in SSBs repair. When single-stranded DNA breaks (SSBs) occurs, PARP1 identifies SSBs, binding to DNA at the SSBs site, and then performing PARP poly (ADP-ribos)ylation (PARylation). When PARP PARylates, it recruits target proternsrepair factors to the lesion and electrostatically breaks the PARP1-DNA connection, thereby releasing PARP1 from the DNA and initiating the DNA repair mechanism. LIU ET AL.

AS MONOTHERAPY IN BREAST CANCER
Presently, there are five clinically developed PARPis that have proven to be effective in BC clinical research and can be ranked according to their ability to trap PARP (from the most potent to the least potent: talazoparib >> niraparib > olaparib = rucaparib >> veliparib). 22 In the assessment of the molecular mechanism of action, veriparib, the weakest PARP1 trapper, does not induce the same level of synthetic lethality as more potent PARPis (rucaparib, olaparib, talazoparib, and niraparib) at the same dose. 24 However, while a single-agent activity requires adequate PARP capture, it is not directly related to clinical efficacy. Due to the lower maximum tolerated dose achieved, more potent PARPis are typically administered in the clinic at lower doses. 22 At present, the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) have approved olaparib and talazoparib as effective single agents for treating gBRCA1/2 and HER2-negative BC. [25][26][27][28] Additionally, niraparib (NCT00749502), 29 rucaparib (NCT02505048), 30 and veliparib (NCT01149083 and NCT01853306) 31,32 are in clinical trials as monotherapies for BRCA-mutant locally advanced/metastatic BC (Table 1); however, the FDA and EMA have not yet approved them for clinical usage.

| Olaparib
In the phase trial OlympiAD (NCT02000622), compared to standard therapy of the physician's choice (capecitabine, eribulin, or vinorelbine), olaparib monotherapy was associated with a longer median progression-free survival (mPFS) of 2.8 months and a 42% lower risk of disease progression or death in patients with HER2-negative, gBRCA1/2 mutation-associated metastatic BC (7.0 vs. 4.2 months; hazard ratio for disease progression or death, 0.58; 95% confidence interval [CI], 0.43-0.80; p < .001). 35 The response rate in the olaparib group was approximately double that of the standard treatment group (59.9% vs. 28.8%). Moreover, studies have revealed no significant difference in overall survival (OS) when comparing olaparib monotherapy with chemotherapy, but there is a potential OS benefit for subgroup analysis of metastatic BC patients who have not received chemotherapies before. 34 The majority of adverse events (AEs) with olaparib were Grades 1-2 with 38.0% (49.5% in the chemotherapy T A B L E 1 Clinical trials of PARP inhibitors for locally advanced and/or metastatic breast cancer. group) being Grade ≥3 AEs. Nausea (58.0%) and vomiting were the most common adverse effects of olaparib. However, they were all mild to moderate in severity (Grades 1-2), typically occurring early in treatment and diminishing or even disappearing as the drug was administered. Compared with chemotherapy, olaparib was significantly more likely to be associated with anaemia (40.0%) in the first 3 months of treatment with 16.1% of cases being Grade ≥3. The risk of anaemia decreased with supportive therapy and over time. No significant cumulative toxicity was observed during monotherapy with olaparib. 34

| Talazoparib
In , the objective remission rates and clinical benefit rates of talazoparib were superior to those of PCT. 36,37 Anaemia, fatigue, nausea, and neutrophil granulocytopenia were the most common adverse effects of talazoparib. The majority of these nonhaematological AEs were Grade 1 or 2 and occurred primarily within the first 4 weeks of talazoparib treatment. Haematologic AEs (68.2%), including anaemia, neutropenia, and thrombocytopenia, usually occurred within the first 16 weeks of treatment. Anaemia occurred in 54.9% of patients treated with talazoparib (19.0% with chemotherapy), and Grade ≥3 anaemia occurred in 45.0%. However, the rate of permanent talazoparib discontinuation due to haematologic AEs was low (<2%). If patients experience severe adverse reactions, the dose of talazoparib can be adjusted. 41,42 Interestingly, the Asian population seems to have a higher tolerance for talazoparib than the overall EMBRACA study population. 38

| RESISTANCE MECHANISMS OF PARPis
As the clinical use of PARPis has increased in the treatment of ovarian cancer, PARPi therapy seemed to be less effective in BRCA-mutated BC compared to ovarian cancer. Although the NCT02000622 and NCT01945775 trials have showed that olaparib and talazoparib can significantly increase PFS, the OS benefits for these PARPi in germline BRCA1/2-mutated metastatic BC patients have not achieved. So similarly to other targeted drugs, PARPi also had to face drug resistance problems, which may become a hindrance to the widespread use in cancer therapy. It has been reported that more than 40% of BRCA-mutated ovarian cancer patients had difficulties in benefiting from PARPi therapies. 43 Understanding these resistance mechanisms will therefore allow us to overcome drug resistance more effectively and improve treatment outcomes. According to research, there are four main mechanisms that contribute to the development of PARPis resistance: restoration of homologous recombination repair, stable replication fork, reduced PARP1 trap, and drug efflux.

| Restoration of homologous recombination repair
Resistance to PARPis therapy in BC cells may be caused by multiple mechanisms, the most widespread and common of which is the restoration of the HRR function in clinical practice. 44 Gene reversion mutations can restore HRR genes (BRCA1/2 and RAD51C/D), resulting in the recovery of HRR function and the subsequent development of drug resistance. 45,46 In BC patients, reverse mutations in BRCA1/2 leading to resistance to PARPis have been observed in clinical practice. 47 For example, in BRCA1-C61G-mutant tumours, due to the disruption of the N-terminal loop structural domain, tumour cells respond poorly to PARPis and develop drug resistance. 48 In addition, HSP90 stabilises the mutation-producing protein product in the BRCA Cterminal (BRCT) structural domain of BRCA1 in PARPiactivated cells. The mutated BRCA1 protein stabilised by HSP90 can effectively interact with PALB2-BRCA2-RAD51 to form RAD51 foci, thereby restoring DNA end resection and RAD51 filament formation, which are essential steps in HRR. 49 This resistance mechanism is highly clinically relevant for patients with BRCA-mutant cancers treated with platinum-based therapy. Patients who progress during platinum-based chemotherapy or shortly after its cessation may have a lowered sensitivity to PARPis (talazoparib). 41 In gBRCA-mutant BCs, RAD51 foci can also serve as functional biomarkers of homologous recombination repair and PARPis resistance.
Inactivation of the NHEJ protein is yet another mechanism for restoring HR (Figure 3.①). When DNA damage occurs in normal cells, BRCA1 initiates HR rather than NHEJ to repair DSBs by inhibiting 53BP1. 50 In mouse models, mutations in the BRCA1 gene led to the loss of HR function, which could be partially restored LIU ET AL. by the loss of 53BP1. 51,52 This may be because the loss of 53BP1 in BRCA1-deficient cells promotes HR recovery in an ATM-dependent way, increases genomic stability, and reduces sensitivity to PARPis. 53 Therefore, 53BP1 can also be used as a biomarker to predict the response of BRCA-mutant cancers to PARPis therapy, and 53BP1negative BC patients have a lower response to PARPis. In addition, the downstream factors of 53BP1: RIF1, REV7, and the shieldin complex are also necessary for NHEJ. Therefore, deletion of the 53BP1-RIF1-REV7shieldin axis, that is, depletion or deletion of the shieldin protein complex or REV7/MAD2L2, can also result in the inactivation of NHEJ proteins and the partial recovery of HR in BRCA1-mutant breast tumours. 54,55 HELB is another protein that mediates resection in a 53BP1-independent manner, and its absence also renders BRCA1-deficient tumour cells resistant to PARPis. 56 In conclusion, BRCA1/2-mutant cells can recover HR via gene reversion mutations or inhibition of NHEJ protein activity, resulting in PARPis resistance.
Similarly, the chromatin remodellers SMARCAL1, ZRANB3, and HLTF are required for MRE11-dependent degradation of replication forks, and depletion of these factors can reduce replication stress and genomic instability, inducing PARPis resistance. 59 SLFN11, another protein associated with replication stress, regulates irreversible and prolonged replication fork arrest. Its deletion may result in a shorter replication fork arrest in the S phase, thereby decreasing PARPis cytotoxicity and inducing PARPis resistance. 60,61 Significantly, it was demonstrated that depletion of SLFN11, SMARCAL1, EZH2, or PTIP did not restore the HR function in BRCA-deficient cells. In this case, PARPiresistant tumours could still be HR-deficient, but PARPis would not exhibit cytotoxic effects on these resistant tumours. In conclusion, BRCA1/2-mutant cells are able to stabilise replication forks, protect genomic stability, and generate resistance to PARPis in the approaches described earlier.

| Reduced PARP1 trap
PARPis target PARP proteins; therefore, when the target protein PARP is altered or depleted, there is less PARP1 trapping on the DNA strand complex, causing F I G U R E 3 Four resistance mechanisms of PARP inhibitors: ① Gene reversion or inactivation of the NHEJ protein can restore the HRR function; ② Depletion of these factors can stable replication fork; ③ Alteration or depletion of the target protein PARP or loss of PARG can reduce PARP-1 trapping on DNA; ④ Drug efflux.
HR-deficient cells to develop PARPis resistance (Figure 3.③). 62 Moreover, a high level of olaparib resistance has been demonstrated in PARP1 knockout cells, 20 suggesting that PARP1 loss may lead to drug resistance. At present, a PARP1 mutation (1771C>T) has been identified in an ovarian cancer patient with resistance to PARPis. 62 Additionally, PAR glycohydrolase (PARG) is responsible for poly(ADP-ribose) (PAR) chain degradation by preventing PAR accumulation. In PARPi-acting cells, the loss of PARG results in the accumulation of PAR, which restores PARP1-dependent DNA damage signalling and decreases PARP1 trapping (Figure 3.③). 63 Additionally, it stabilises replication forks and facilitates the recruitment of DNA repair enzymes to the site of damage. Overall, the decrease in PARP1 trapping leads to a reduction in PARPi-induced DNA damage, rendering cells less sensitive to PARPis.

| Drug efflux
The overexpression of ATP-binding cassette (ABC) drug transport proteins is usually connected to drug resistance. 64 Long-term treatment with PARPis has been shown to increase the expression of P-glycoprotein (P-gp, also known as MDR1, encoded by ABCB1), which reduces the intracellular concentration of PARPis and generates PARP resistance (Figure 3.④). 65 However, the mechanisms underlying the association between increased expression of the P-gp efflux pump and PARPis resistance remain unknown. After cotreatment with tariquidar, an inhibitor of the P-gp efflux pump inhibitor, the tumour regained its sensitivity to PARPis. 66

ADVANCES TO IMPROVE THE EFFICACY OF PARPis IN BREAST CANCER
Currently, monotherapy is the predominant clinical application of PARPis for advanced BC. However, the emergence of drug resistance has significantly diminished the efficacy of PARPis. We analysed the following therapeutic strategies to overcome drug resistance to PARPis and improve the efficacy of PARPis in BC based on its mechanism of drug resistance. The clinical studies about therapeutic strategies to improve the efficacy of PARPi are listed in a Table 2.

| Early treatment
Compared to advanced disease, early-stage tumours should have fewer acquired resistance mechanisms that negatively impact response duration. 23 Therefore, early treatment can enhance the antitumour effect of PARPis, necessitating enhanced genetic testing to identify BRCA mutations and genetic counselling for potential patients and their families, resulting in early diagnosis and treatment of BC.
In the phase III trial OlympiAD, compared to placebo, olaparib significantly improved invasive disease-free survival and distal disease-free survival in the adjuvant treatment of early-stage, high-risk, HER2-negative, gBRCA1/2-mutant BC. 67 This indicates that olaparib can be used both as a single drug and as an adjuvant in the treatment of BC in its early stages. In the 2021 ASCO guidelines, it is recommended that 1 year of adjuvant olaparib therapy be offered to early-stage BRCA1/2mutant BC patients with a high risk of recurrence. 73

| Combination with chemotherapy
Since PARPis in combination with chemotherapy have been confirmed to improve clinical treatment outcomes compared to standard chemotherapy in advanced or metastatic BRCA1/2-mutant BC clinical trials, it may serve as one of the therapeutic strategies to delay drug resistance. However, the combination treatment scheme is often limited by severe drug toxicities. In phase I trials, 74,75 the frequency, severity, and duration increased when olaparib capsules were administered continuously or intermittently with carboplatin and/or paclitaxel. Therefore, researchers have considered reducing the dose of olaparib in combination with carboplatin and paclitaxel for BC. 76 However, veliparib, a weak PARP trapping agent, can reduce the incidence of myelosuppression compared to other PARPis 77 and has shown the antitumour activity and tolerable side effects. 31 In the phase II trial BROCADE2 (NCT01506609), 68 progression-free survival (PFS) and OS were numerically improved in patients with advanced BRCA1/2-mutant BC when veriparib was added to carboplatin/paclitaxel versus placebo with objective remission rates (ORRs) increasing from 61.3% to 77.8%. In the phase III trial BROCADE3 (NCT02163694), veliparib in combination with carboplatin and paclitaxel (C/P) compared to placebo plus C/P in participants with BRCA1/2-mutant BC increased progression-free survival (14.5 vs. 12.6 months; HR 0.71, 95% CI 0.57-0.88). 69 The most common AEs of Grade 3 or worse were neutropenia (81% vs. 84%), anaemia (42% vs. 40%), and thrombocytopenia (40% vs. 28%). 78 Similar to the treatment of BRCA-associated ovarian cancer, after combination therapy with veliparib and carboplatin/paclitaxel, discontinuation of carboplatin/ paclitaxel before disease progression and continuation of LIU ET AL.  maintenance treatment with veliparib can help preserve drug response and improve PFS in patients with advanced germline BRCA-associated BC (25.7 vs. 14.6 months, HR 0.49, 95% CI 0.34-0.73; p < .001). 70 After veliparib monotherapy, the most frequent AE of Grade 3 or higher was nausea (5%). 64,78

| Combination with immunotherapy
More and more research have proven that the immunosuppressive tumour microenvironment of BC characterised by the decreased expression of TILs and the increased expression of immunosuppressive cells, such as Tregs, had negative effects on chemotherapy and targeted therapy. 79,80 The recent studies showed that the immune response initiated by PARPi was also necessary for the elimination of breast tumour. 81 Increased cellular DNA damage induced by PARPis treatment of BRCA-mutant tumours and activation of the type I IFN pathway via cGAS-STING signalling have pleiotropic effects on the immune response, including promotion of dendritic cell (DC) maturation, release of interferons, and initiation of antitumour T-cell immune responses. 82 Qiwei Wang et al. found that PARPi had limited effects on BRCA1-mutated BC because of the tumourassociated macrophages (TAMs), which hampered the PARPi-induced breast cell DNA damage. But the addition of STING agonist could shift the TAMs phenotype to the anti-tumorigenic M1-like state to be benefit of restoring the PARPi's SL response in BRCA1-deficient BC of mouse models. This finding may provide a prospective approach to increase the PARPi response to resistant BRCA1mutated BCs. 83 Some clinical models have also demonstrated that PARP inhibition inactivates GSK3 and upregulates PD-L1 in a dose-dependent manner, which leads to the inhibition of T-cell activation and enhanced apoptosis of cancer cells. 84 In this instance, the combination of an immune checkpoint blocker (such as anti-PD-L1 or anti-CTLA-4) can significantly enhance the therapeutic effect, providing a theoretical basis for the combination of PARPis and ICB. 85,86 Studies have shown that olaparib combined with an immune checkpoint blocker (ICB) increases antitumour efficacy and immunomodulation in treating BRCAdeficient tumours. 82 In a phase II trial (MEDIOLA), the efficacy of the combination of durvalumab (Imfinzi, an anti-PD-L1) and olaparib in metastatic BRCA-mutant HER2-negative BC was evaluated with a disease control rate (DCR) of 80% at 12 weeks (90% CI: 64.3%, 90.9%) and 50% at 28 weeks (90% CI: 33.9%, 66.1%). 71 The most common Grade 3 or 4 AEs were anaemia (12%), neutropenia (9%), and pancreatitis (6%), 72 and there were no significant adverse drug interaction reactions. Additional phase III clinical trials are required to validate the clinical efficacy and safety of ICB in combination with olaparib.
The clinical toxicity of talazoparib is greater than that of olaparib, 87 so the effects and safety of ICB drugs need to be identified in future clinical research.

| Combination with cell cycle checkpoint inhibitors
Cyclin-dependent kinase 4/6(CDK4/6), which regulates the transition of G1 phase to S phase of the cell cycle, plays an important part in the cell cycle renewed process. A recent study showed that CDK4/6i had the ability of preventing tumour cells recovery from DNA damaging agents. 88 Another study showed that the combination of olaparib with palbociclib, which is a CDK4/6i, has significantly increased olaparib-resistant cell damage and successfully decreased the tumour cell growth after activating the Wnt signalling pathway mediated by palbociclib strongly inhibitive effects. 89 It is also found that CDKs interfered with DNA end resection play a positive role in PARPi resistance. 90,91 Carey et al. found that CDKi dinaciclib could resensitized TNBC, which acquired resistance to PARPi niraparib by downregulating MYC oncogene expression and recovering the synthetic lethality with PARPi. 92 So these research studies provide the reference of clinical value for the synergistic action of CDKi with PARPi in BRCAmutated BC with Wnt pathway overactivation and high MYC oncogene expression that may induce a bad response to PARPi-resistant cells.
When DNA damage occurs, the ATR-CHK1-WEE1 axis is activated, resulting in the inactivation of cell cycle protein-dependent kinase 1/2 (CDK1/2) and the arrest of the G1/S and G2/M phases of the cell cycle in normal cells. 93,94 Thus, the combination of cell cycle (ATR, CHK1, and WEE1) inhibitors with PARPis reduces the time required to repair DNA by HR and promotes replication of damaged DNA, leading to cell death. 95 So ATR/ CHK1 and WEE1 inhibitors' treatment can reverse the resistance state of cancer cells in combination with PARPi. Moreover, it has also been demonstrated that cell cycle checkpoint inhibitors can overcome the resistance to PARPis caused by replication fork stabilisation and increase the sensitivity of PARPis. 96 In PARP-resistant BRCA1-deficient cells, ATR regulates BRCA1-independent HR and fork protection by promoting RAD51 recruitment to DSBs and stalled forks. 97 Thus, in a mouse BC model, combining an ATR inhibitor (AZD6738) with olaparib enhances the antitumour effect of olaparib in BRCA-mutant cells. 97 Based on these findings, a phase II clinical trial (NCT04090567) evaluating olaparib in combination with AZD6738 for the treatment of advanced or metastatic gBRCA-mutant BC is presently underway with results expected in 2023.
Likewise, WEE1 is a protein kinase that inhibits cyclin-dependent kinases 1 and 2 (CDK1/2), thereby activating the G2/M cell cycle checkpoint, leading to cell cycle arrest and allowing time for DNA damage repair. Thus, WEE1 inhibitors (adavocetib/AZD1775) can also be used in combination with PARPis for the treatment of BRCA-mutant cancer. Inhibition of WEE1 under PARPis treatment with homologous recombination deficiency blocks the initiation of the G2 checkpoint, forcing cells to enter mitosis without completing DNA synthesis and repair, resulting in the accumulation of DSBs and loss of genomic integrity and hence enhancing PARPis sensitivity. 98 However, due to its poor tolerability, sequential administration of therapy is recommended. 99

| Combination with IR therapy
It has been demonstrated that ionising radiation (IR) can sensitised the PARPi effects in HR-proficient tumours by starting the export signal of BRCA1 from the cell nucleus to cytoplasm. 100,101 In the preclinical studies, the acquired drug resistance associated with BRCA1independent HR restoration could be reversed by the radiotherapy method. 102 More importantly, some clinical research was also carried out to study the probable efficacy of PARPi combination with IR theapy in solid tumours. For example, there are two phase I clinical trials (NCT01589419 and NCT01264432) designed in advanced rectal cancer and peritoneal carcinomatosis, which have indicated good tolerance and responses in the combination treatment. 103,104 Another phase I clinical trial (NCT01618357) was underway to evaluate the initial effects and safety of veliparib with IR therapy in BC, which is worth expecting.

| Combination with proteogenome
Currently, the ability to identify drug resistanceassociated proteins remains a challenge. Proteomics technologies, such as mass spectrometry and protein array analysis, facilitate the anatomical and proteomic characterisation of potential molecular signalling events and underlying diseases. In this context, proteomic analysis, along with their adaptive response to therapy, can identify new therapeutic options, which can reduce the emergence of drug resistance and potentially improve patient outcomes. 105 As mentioned previously, RAD51 foci 19 and 53BP1 49 can be used as a biomarker to predict the response of BRCA-mutant cancers to PARPis therapy. Through the proteomiecs techonology, TPX2 was identified as the posssible target to predict the sensitivity of cancer cells to PARPi capture ability. TPX2 is a direct PARP1-binding protein that regulates the self-ADP-ribosylation activity of PARP1 and promotes homologous targeted repair of DNA double-strand breaks. Thus, the overexpression of TPX2 leads to increased PARP1 activity and thus increases PARP capture potential. 106 In addition to the protein biomarkers, which have being studied, the FDA has approved several gene expression-based prognostic tests: Oncotype DX3, 107 EndoPredict4, 107 and MammaPrint5. 108 These tests can be used to predict BC risk of recurrence and help guide clinical decisions. Recently, the FDA approved the FoundationOne Liquid CDx trial, a circulating cfDNA (cfDNA)-based assay as a companion diagnostic that can be used to determine which patients may benefit from specific FDA-approved targeted therapies. In clinical practice, it can be used to predict the efficacy of the PARPis rucaparib in treating patients with BRCAmutated (germline or somatic) ovarian cancer and apelisib in treating patients with HR+/HER2−, PIK3camutated BC. 109

| OUTLOOK
PARPis therapy provides targeted treatment for BRCAmutant BC and is safer and more effective than chemotherapy. Moreover, AEs observed during treatment are typically manageable, which reduce the discomfort experienced by patients during treatment. It can be argued that the advent of PARPis has significant implications for biomarker-targeted BC therapy. Nonetheless, drug resistance is unavoidable, and the traditional use of single agents is no longer as effective as it once was. Early dosing or combination therapy strategies are available clinically to overcome resistance and improve the efficacy of PARPis.
As PARP research advances, PARPis may be used for a broader range of people in the future. In that case, it must be developed for improved detection of homologous recombination defects beyond the BRCA1/2 mutation and identification of DNA repair-deficient tumours in other cancer types. (Take prostate cancer for example, patients with BRCA2 pathogenic sequence variants have increased levels of serum PSA at diagnosis, an increased proportion of high Gleason tumours, elevated rates of nodal and distant metastases, and high recurrence rates. Currently approved by the FDA PARPis for the treatment of metastatic castration-resistant prostate cancer are olaparib, rucaparib, and niraparib. 110 ) Alternatively, combinations with additional targeted agents may be considered based on drug resistance mechanisms. In addition, additional biomarkers that can predict response are being developed to provide more accurate patient stratification and guide clinical dosing strategies that maximise the efficacy of PARPis. Through future PARPis research, we expect the emergence of potent drugs with new mechanisms to alleviate the problem of PARPis drug resistance.