1. Introduction: the concept of anticipation in the systemic therapy of prostate cancer
For a long time, the natural history of metastatic prostate cancer (mPC) has been simply divided in two separated phases: 1. Hormone sensitive (HSPC) and 2. Hormone refractory (HRPC), subsequently better defined as Castration resistant (CRPC). These two phases were considered as two distinct pathologies with different therapeutic approaches, the first subjected to different lines of manipulation of the classic androgen deprivation (ADT), the second to chemotherapy (CHT) with taxanes. The systemic treatment of mPC has undergone an epochal positive evolution in recent years, in particular with the development of new generation hormone manipulations and androgen receptor signaling inhibitors (ARSI). Most of these new therapies were initially indicated in the late stages of the disease (post CHT-mCRPC) [
1,
2] where options and patient survival were limited. However, several clinical trials have shown that ARSI and CHT are equally effective in earlier stages, such as first-line mHSPC [
3,
4,
5,
6]. The intention of treating more patients for a longer time has led to the concept of
anticipation in the systemic treatment of mPC. With anticipation came several changes in therapeutic strategies, in particular: 1. The use of both docetaxel and ARSI in mHSPC; 2. The stratification of mPC in classes such as high and low volume, high and low risk, or de novo and progressive tumors; 3. The competition in the use of these drugs between the mHSPC and mCRPC phases; 4. The shift from a clear distinction between the HSPC and CRPC phases and their management to a merger of the two. At present, the management of mCRPC represents a second line of therapy, sharing several of the recommended drugs with mHSPC. The concept of anticipation of therapy has merged with that of intensification of care, leading to first-line therapeutic schemes with doublets or triplets of drugs in mHSPC. As a matter of fact, ADT alone no longer represents a therapeutic option in mHSPC and the current recommendation is for an association with ARSI, docetaxel or both [
3,
4,
5,
6]. In patients with high volume/high risk de novo mHSPC, the recent trials PEACE1 [
7] and ARASENS [
8] introduced the concept of triplets, with abiraterone or daralutamide in combination with docetaxel and ADT.
The main criticalities associated with the concepts of anticipation and intensification of care are: 1. An uncontrolled progression of anticipation in the already ongoing non-metastatic PC; 2. The development of early drug resistance with limited choice of treatment for late mCRPC and the need to repeat the same combination of drugs used as first-line therapies in HSPC. The key point of failure in this process, accentuated by the numerous therapeutic options available, is the identification of valid prognostic indicators of treatment response and the development of tailored precision therapies for each patient. The stratification of mPC on the basis of volume, risk and development of the disease in not sufficient for precision medicine. The genetic analysis of pathogenetic variants (PV) of genes involved in the DNA Damage Repair (DDR) mechanisms and the development of Poly ADP ribose polymesare (PARP) inhibitors (PARPi) led to the first valid precision medicine option tailored for mPC. PARPi are following the same evolutionary path traced by ARSI, expanding their indication starting from the late stage of second line mCRPC treatment in a limited population and time frame. Inevitably, also PARP inhibitors will be absorbed by the concept of anticipation and intensification of care. Mounting evidence in favour of the anticipation of PARPi in the first-line mCRPC therapy is already emerging and further one in mHSPC is underway. (
Figure 1).
The purpose of this review is to analyze how the genetic basis of DDR PV and the clinical results with PARPi currently available can guarantee an anticipation process, always in compliance with precision medicine and tailored indications.
2. The rational and the genetic profile that sustain PARP inhibitors in PC
To address the underlying issue of genomic erosion, a complex network of DDR mechanisms has evolved [
9,
10]. Genetic abnormalities in DDR response systems are associated with life-threatening diseases such as immune deficiency, premature aging and cancer susceptibility. There are six major DNA repair pathways in humans: base excision repair (BER), nucleotide excision repair (NER), single strand break repair (SSR), homologous recombination repair (HRR), non-homologous end joining (NHEJ), mismatch repair (MMR). When the lesion is situated in only one strand of DNA, the undamaged complementary strand acts as a template for the NER and BER repair mechanisms. When a damaging agent induces DNA double-strand breaks (DSBs) MMR, HRR and NHEJ come into the picture [
11,
12]
PARP is a multifunctional protein that plays a critical role in SSR and BER mechanisms, primarily by recruiting DNA repair proteins to the sites of damage. So far, eighteen members of the PARP family have been found, among which PARP-1 is the most important [
13]. PARP's main enzymatic function is to add ADP-ribose to substrate proteins by cleaving NAD
+ and releasing nicotinamide, thus activating a complex cascade which ends in the recruitment of different proteins, such as DNA polymerase theta (POLQ), DNA ligase I and XRCC1, to the site of damage. Despite being well-known for its role in SSR and BER, recent evidence suggests that PARP can also affect DSB repair, mainly via controlling the expression of key HR genes BRCA1 and RAD51 [
14,
15,
16].
Defects in the DDR pathways can result in genomic instability, gene mutations, and, eventually, lead to the development of cancer. On the other hand, germline or acquired mutations in DDR also provide cancer-specific vulnerabilities that can be targeted by synthetic lethality-based therapies [
17]. PARPi are the first successful example of a targeted therapy that uses synthetic lethality to kill cancerous cells with DNA-repairing deficiencies (e.g., BRCA1/2 mutation). These drugs act by competing with NAD
+ for the catalytically active site of PARP molecules, interfering with SSB repair by BER. The unrepaired SSB can be converted to DSB, and the primary mechanism to repair such lesions during cell cycle is HRR. While HRR-proficient cells can repair the DSB originated from the SSB, thus ensuring genetic stability and cell survival, HRR-deficient cells are unable to do so, resulting in apoptosis and ultimately cell death [
16,
17,
18]. Olaparib was the first drug to be developed in this group. Initially approved in 2014 for patients with ovarian cancer with germline BRCA PV and subsequently in breast, pancreatic and prostate cancer [
19].
Germline HRR PV increase the risk of developing PC by 8 times at the age of 65 years and are linked to more aggressive PC, higher risk of lymph-node invasion and distant metastasis at the time of diagnosis [
20]. Moreover, germline BRCA 1/2 PV are associated with higher probability of failure in patients with localized disease undergoing active surveillance and higher risk of recurrence in patients who underwent curative treatment [
21,
22]. Several studies have analyzed the prevalence of germline and somatic mutations in HRR genes, both in metastatic and localized PC. A 2019 systematic review by Lang et al. [
23] found a median prevalence of germline HRR PV and specifically of BRCA2 PV in patients with familiar PC of 29.3% (range, 7.3-91.67%) and 3.7% (range, 1.3-7.9%) respectively. In the same study, for unselected patients, the frequency of somatic and germline mutations was, respectively, 3.9% and 1.5% for ATM, 1.1% and 0.6% for BRCA1, 4.9% and 1.1% for BRCA2, 1.3% and 0.5% for PALB2, 1.5% and 0.5% for RAD51C. The overall prevalence of somatic DDR genes PV ranged between 4.9 and 22%, while the rate of germline DDR PV was 17.2-19%.
Regarding metastatic PC, the incidence of germline PV in HRR genes was found between 11% and 33% in a pan-cancer analysis of whole genomes [
23]. Similarly, 8% of germline and 23% of somatic HRR PV were found in 150 metastatic PC by the International Stand Up to Cancer/Prostate Cancer Foundation team (SU2C-PCF) [
24]. In this cohort, BRCA2 was the most prevalent mutation (13%) followed by ATM (7.3%), MSH2 (2%) and BRCA1. More specifically, in mCRPC, the incidence of somatic HRR PV was 24% (BRCA 13%, ATM 7.3%, MSH2 2%, BRCA1 0.3%) in a study by Eeles et al. [
25] and 28% in the Profound study, which analyzed the results of 2792 biopsies of mCRPC patients.
Considering localized PC, the rates of PV in HRR seem to be lower than those seen in mPC. Despite that, a 2019 study by Kim et al. [
26] found an overall incidence of HRR pathways alteration in localized PC of 29.9%, higher than suggested by a previous study by Marshall et al. [
27] (11% in Gleason Grade Group 5 and 15.8% in cT3 patients).
In the Profound study, a total of 4,858 tissue samples were tested and reported centrally [
28]. Considering all samples examined by Next Generation Sequencing (NGS), 83% were primary tumor samples (96% were archival and 4% newly obtained) and 17% were metastatic tumor samples (60% were archival and 33% newly obtained). NGS results were generated more frequently from newly obtained compared with archival samples (63.9% vs. 56.9%) and metastatic compared with primary samples (63.9% vs. 56.2%). Although generation of an NGS result declined with increasing sample age, approximately 50% of samples >10 years generated results.
Somatic analysis of HRR mutation in PC should be considered first, compared to germline analysis. When mutations of HRR genes are acquired during the progression of the disease, biopsy of the metastatic tumor represents the ideal approach to identify molecular alterations. However, biopsies of metastatic lesions can be challenging or not feasible, and at the same time, a single biopsy may not reveal tumor heterogeneity among metastases. The analysis of free circulating DNA (cDNA) is a promising approach as it could overcome the difficulty in obtaining tissue in many cases where this is not possible; currently there is no data that allows this test to be used reliably. The first study that analyzed cDNA in this field is GHALAND study [
29] where the treatment efficacy analysis was performed on the amount of circulating tumor cells present from the eighth week of treatment. Best results were obtained in BRCA cohort than non-BRCA cohort, with a 24% of CTC (circulating tumor cell) response. However, incongruences with different commercial tests currently available for cDNA analysis are as high as 40%, with the risk of patients receiving inappropriate or no treatment.
3. Clinical trials and actual recommendations for PARP inhibitors in second line mCRPC
The latest guidelines of the European Urological Association (EAU) recommend PARPi in pre-treated mCRPC patients with relevant DNA repair gene mutations.
So far, only two PARPi, olaparib and rucaparib, are licensed by the FDA (EMA only approved olaparib), but other drugs of the same class are under evaluation (e.g., niraparib, talazoparib). The actual recommendations are supported by prospective trials analyzing a tailored therapy with PARPi for the treatment of mCRPC progressing after first-line treatment with ARSI and taxanes.
In the second line mCRPC setting, PARPi were analyzed as monotherapies in patients with HRR mutations. This type of analysis also determined a ranking in the HRR PV, withBRCA2 mutation associated with the most significant benefits from all types of PARP therapy.
Olaparib was first evaluated in the TOPARP-B trial [
30], a multicenter, open label, randomised phase II trial. Eligible patients had a mCRPC and a putatively pathogenic mutation or homozygous deletion in a DDR gene that could be associated with sensitivity to PARP inhibition as identified by NGS. Patients were required to have previously received at least one but no more than two taxane based chemotherapy regimens (regardless of prior exposure to novel hormonal drugs). The 98-eligible-patients with DDR gene aberrations were divided into two dose cohort, randomly assigned to each one to receive 400mg or 300mg olaparib twice daily. Only the cohort receiving400mg twice daily met the predefined criteria for success, showing better percentages in all primary endpoints: radiological objective response, 50% decrease in PSA from baseline, conversion of circulating tumour cell count. Even though the results observed varied considerably for different HRR gene aberrations, the greatest antitumor activity was seen in the subgroup with BRCA1/2 mutations.
The Profound trial [
31] can be considered the most relevant study in this field. A prospective, randomized, open-label, phase III trial was carried out at 206 sites in 20 countries and evaluated the safety and efficacy of olaparib in men with mCRPC who progressed while receiving a new hormonal agent (enzalutamide or abiraterone).
Patients were divided into two cohorts depending on their qualifying gene alteration selected for their direct or indirect role in HRR: patients with at least one alteration in BRCA1, BRCA2, or ATM were assigned to cohort A, patients with alterations in any of the other 12 genes were assigned to cohort B. In each cohort patients were randomly assigned to receive olaparib (300mg twice daily) or the physician’s choice of either enzalutamide (160mg once daily) or abiraterone (1000mg once daily) plus prednisone (control group).
Overall, 387 patients met all eligibility criteria and thus underwent randomization from April 2017 through November 2018. The primary endpoint was radiological progression-free survival (PFS), which was significantly longer in the olaparib arm (7.4 months vs 3.6 months in the control group) [HR for progression or death 0.34; 95% CI, 0.25–0.47; P < 0.001], especially in cohort A patients (which have a 66% lower risk of disease progression or death).
In cohort A, as secondary endpoints, the confirmed Objective Response Rate (ORR) was (olaparib vs. control group) 33% vs. 2%, a 50% reduction in PSA was seen in 43% vs. 8% and the median Overall Survival (OS) was 19.1 months vs. 14.7 months (HR 0.69, 95% CI 0.50–0.97, P = 0.0175).
The common adverse events of any grade in the olaparib group were anemia (46% vs. 15%) followed by nausea and fatigue and the overall incidence of grade ≥ 3 adverse events was higher in the olaparib group (51% vs 38%).
The other FDA-approved PARPi, rucaparib, was evaluated in the TRITON-2 trial [
32], an open-label, phase II study. The trial enrolled 115 patients who progressed after one or two lines of next-generation AR-directed therapy and one taxane-based chemotherapy for mCRPC and who had a deleterious germline or somatic alteration in one HRR gene that may confer sensitivity to PARP inhibition. Patients received a starting dose of 600mg twice daily. As primary endpoint, ORR was 43.5% (95% CI, 31.0% to 56.7%;) per independent radiology review and 50.8% (95% CI, 38.1% to 63.4%) per investigator assessment. ORRs were similar for patients with a germline or somatic BRCA alteration and for patients with a BRCA1 or BRCA2 alteration, while a higher PSA response rate was observed in patients with a BRCA2 alteration. PSA responses seemed smaller in the
BRCA1 (15.4%) or mono-allelic patients (11.1%) compared to
BRCA2 (59.8%) or biallelic patients (75.0%). The PSA response rate was 4.1%, 6.7% and 16.7% in the
ATM group (49 patients),
CDK12 cohort (15 patients) and
CHEK2 group (12 patients), respectively. The ORR was 10.5% in the
ATM group, 0% in the
CDK12 cohort, and 11.1% in the
CHEK2 group. The most frequent treatment-emergent adverse events (TEAEs) of any grade were asthenia (61.7%), nausea (52.2%), and anemia (43.5%).
The anti-tumour activity and safety of the PARPi niraparib was evaluated in the GALAHAD trial [
29], a multicenter, open-label, single-arm, phase II study. Patients with germline pathogenic or somatic biallelic pathogenic alterations in BRCA1 or BRCA2 (BRCA cohort, n = 142) or biallelic alterations in other prespecified HRR (non-BRCA cohort; n = 81) were included. Enrolled patients received niraparib 300mg orally once daily until treatment discontinuation, death, or study termination. At final analysis, with a median follow-up of 10 months, the ORR in the “
BRCA group” was 41% (95% CI 23.5–61.1%) compared to 9% (95% CI 1.1–29.2%) in the “non-
BRCA group”; and the CRR was 63% (95% CI 47.6–76.8%) compared to 17% (95% CI 6.6–33.7%), respectively. Median PFS and OS in the BRCA cohort were 8.2 months (95% CI 5.2–11.1 months) and 12.6 months (95% CI 9.2–15.7 months), respectively, versus 5.3 months (95% CI 1.9–5.7 months) and 14.0 months (95% CI 5.3–20.1 months) in the non-
BRCA group.
The most common TEAEs of any grade were nausea (58%), anaemia (54%) and vomiting (38%); the most common grade ≥ 3 events were anaemia, thrombocytopenia and neutropenia.
The last PARPi under investigation, talazoparib, was assessed in the TALAPRO-1 trial [
33] (open-label, phase II). The inclusion criteria were the same of the GALAHAD trial and eligible patients were given oral talazoparib (1mg per day or 0.75mg per day in patients with moderate renal impairment). Finally, 127 cases received at least one dose of the study drug (safety population) and 104 had HRR-deficient measurable disease (antitumor activity population)
. The primary endpoint was confirmed ORR, whereas secondary endpoints were time to objective response, duration of objective response, proportion of patients with a decrease in PSA ≥ 50% from baseline, time to PSA progression, radiological PFS, OS, safety. After a median follow-up of 16.4 months, the ORR was 29.8% (95% CI; 21.2–39.6), 46% in cases with BRCA1/2 mutations with a radiological PFS of 11.2 months, 12% in cases with ATM mutations. Talazoparib produced a 50%-PSA response in 66%, 7%, 6%, and a CTC conversion in 81%, 50%, 20%, in cases with BRCA 1/2, ATM and other mutations respectively.
The most common grade ≥ 3 TEAEs in the overall population were: anaemia (31%), thrombocytopenia (9%), and neutropenia (8%).
4. The prognostic role of HRR PV in non-metastatic and mHSPC as indicators of anticipation tailored treatment
It is widely accepted that BRCA2 gene PV increase the chance of developing PC and are linked to an earlier onset, higher rates of lymph node involvement and distant metastasis at the time of diagnosis [
34,
35]. In this context, the IMPACT study analyzed the role of PSA screening in BRCA PV carriers compared to controls after 3 years of follow-up [
36]. A total of 2932 patients were recruited, 919 were BRCA1 carriers, 709 BRCA1 noncarriers, 902 BRCA2 carriers and 497 BRCA2 noncarriers. The incidence of PC at biopsy was 5.2% in BRCA2 carriers and 3.0% in non-carriers. Moreover, BRCA2 gene PV were associated with a younger age at diagnosis and a higher incidence of intermediate-risk or high-risk disease (77% in BRCA2 carriers vs 40% in non-carriers). These data suggested the possibility of a tailored PV screening strategy for this more vulnerable population with a higher chance of high-risk disease.
Active surveillance (AS) is a well-established option for men diagnosed with favorable-risk and intermediate-risk PC. A study by Carter H.B. et al. [
21] evaluated whether BRCA 1/2 and ATM PV were associated with grade reclassification (GR) in patients undergoing AS and thus with the need of a more aggressive choice of treatment from diagnosis. Considering all 1211 participants, the rate of PV carriers was significantly higher in those reclassified both for the three-gene panel and for BRCA2 alone (3.8% and 2.1%, respectively) than those not reclassified (1.6% and 0.5%, respectively). However, the retrospective nature of this study and the absence of multiparametric magnetic resonance targeted biopsies demand for further evaluation of the role of AS in DDR PV carriers through prospective studies [
37].
Castro E. et al. [
22] evaluated the effect of BRCA PV on metastatic relapse and cause-specific survival after radical treatment (surgery and radiotherapy) for localized disease. A total of 1302 patients were included in the study, of which 67 carried a germline BRCA PV (18 BRCA1 and 49 BRCA2). At 3, 5, and 10 years after treatment, 97%, 94%, and 84% of noncarriers vs 90%, 72%, and 50% of carriers were metastasis-free. The 3-, 5-, and 10-year cancer specific survival (CSS) rates from radical treatment were significantly better in the noncarriers (99%, 97%, and 85%, respectively) than in the carrier cohort (96%, 76%, and 61%, respectively; p < 0.001). In line with these results, Martinez Chanza M. et al. [
38] concluded that BRCA PV were associated with a greater relapse risk in a retrospective series of 380 patients with localized and metastatic HSPC. Data from these studies supports closer follow-up of these patients and the need for randomized prospective clinical trials to standardize the most appropriate management strategies of localized PC in HRR PV carriers.
Different studies have analyzed the impact of BRCA2 and other HRR PV in metastatic PC. In a recent observational study, Antonarakis et al. [
39] investigated the clinical impact of germline HRR PV on the efficacy of first-line ARSI among 172 mCRPC patients. Clinical/radiologic PFS, the primary endpoint, was longer both in patients with versus without any germline DNA-repair PV (median 13.3 vs 10.3 months) and in those with versus without BRCA/ATM PV specifically (median 15.2 vs 10.8 months). These results appear to be in conflict with a retrospective study of 319 patients with mCRPC by Annala M. et al. [
40], which showed that HRR PV carriers had a significant shorter PFS rate compared to non-carriers (3.3 vs. 6.2 months, p = 0.01). Despite that, another recent study [
41], in accordance with Antonarakis et al. [
39], suggested a better response to first-line abiraterone treatment in germline or somatic HRR PV carriers compared to non-carriers, supporting the idea of a “synthetic lethality” of treatment with more effective AR-targeted therapies in patients harboring germline HRR PV.
Prorepair-B was the first prospective trial to analyze the prognostic impact of BRCA1–2 and other HRR genes on CSS in mCRPC patients [
42]. There was no significant difference in CSS rates when considering all HRR mutations together between carriers and non-carriers (HRR PV carriers, 23.3 months vs. non-carriers 33.2 months; p = 0.264). Surprisingly, when only BRCA2 was taken into account, the authors found statistical difference in CSS (BRCA2 PV carriers, 17.4 months vs non-carriers, 33.2 months; p = 0.027). Moreover, a subgroup analyses to determine whether the treatment sequence adopted had any effect on the carrier status' impact on CSS was carried out. BRCA2 carriers were found to have worse outcomes compared to non-carriers when treated with the sequence docetaxel-ARSI (median, 10.7 vs 28.4 months; p < 0.001) but not when treated with the sequence ARSI-docetaxel (median, 24.0 vs 31.2, p = 0.901). Despite the need for validation in larger series, these results suggest that the prognostic impact of BRCA2 mutations may be influenced by the choice of first-line therapy.