1. Introduction
Breast cancer (BC) is a multifactorial disease and accounts for 30% of cancers in women
[1]. A variety of factors have been associated with its incidence, several of which are genetic. Indeed, approximately 40% of inherited cancers are due to mutations in the breast cancer susceptibility genes 1 and 2 (
BRCA1 and
BRCA2)
[2]. Recent estimates have shown that approximately 70% of patients with mutations in
BRCA1 and 45% in
BRCA2 will develop BC by the age of 70
[1][3].
Triple negative BC (TNBC) is the most aggressive BC subtype, devoid of hormone receptors (estrogen and progesterone) or human epidermal growth factor receptor 2 (HER2). TNBC is distinct and heterogeneous compared to the other subtypes
[4]. Lack of targeted therapies for TNBC patients has led to an unmet need for new biomarkers
[5][6]. TNBC is closely associated with
BRCA mutations and especially germline mutations
[7].
BRCA1 is located on chromosome 17q21 and encodes a tumor suppressor protein of 1863 amino acid residues.
BRCA1 has been reported to have a plethora of roles in tumorigenesis from the start of a tumor, up to the regulation of epithelial-to-mesenchymal transition, cell motility, adhesion, invasion and ultimately metastasis. DNA damage and dysfunctions of repair are also features of cancer etiology
[8].
BRCA1 is involved in the DNA damage repair (DDR) processes, in particular the repair of double strand breaks (DSBs), hence safeguarding genomic stability and integrity
[9][10].
Poly(ADP-ribose) polymerase 1 (PARP-1) enzyme is the most abundant and best characterized of the 17-member family in humans
[11][12]. PARP-1 is a DNA damage sensor, also involved in DDR, primarily the repair of single-strand breaks (SSBs), but also DSBs
[13][14]. Numerous studies have shown up-regulation of PARP-1 expression in cancer cell lines and patients’ tissues
[15]. PARP-1 is mainly localized in the nucleus, but its cytosolic distribution has also been investigated
[16][17][18].
The DDR process, where both
BRCA1 and PARP-1 are crucial, consists of five pathways. Repair of SSBs by PARP-1 is via the base excision repair pathway
[11][12]. In the case of the more deleterious DSBs, repair is mediated by two complementary systems, namely the non-homologous end joining (NHEJ), which is more error-prone and can also depend on PARP activity, and the homologous recombination (HR) repair system, mediated amongst other DNA damage players by
BRCA1 [11][12][13]. Interestingly, more and more evidence suggests that PARP-1 is also essential in the HR process
[19][20], supporting PARP’s complex roles and the interplay of the repair pathways. Administration of PARP inhibitors will result in the production of a SSB, which will be repaired by HR. In the HR-deficient (BRCA-mutated) cells, however, it will result in a DSB, when this SSB reaches the replication fork. Accumulation of DSBs will prove fatal for the cell as they will lead to cell apoptosis or accumulation of mutations with higher likelihood for cancer development
[5][21][22][23][24].
New treatments in BC target the DDR
[25]. In a very recent meta-analysis, PARP inhibitors have been shown to prolong progression free survival (PFS) and overall survival (OS) in patients with BRCA-mutated advanced BC
[26]. Two PARP inhibitors, namely olaparib and talazoparib, have been approved as monotherapies for the management of locally advanced/metastatic HER2-negative BC, in patients with
BRCA1 or
2 germline mutations
[6][22][23][27]. These inhibitors were identified following the OlympiAD
[28][29] and EMBRACA
[30][31] phase III trials, respectively. In the former study, approximately half of the patients had TNBC. The main outcome of both studies was the statistically significant improvement of PFS of BC patients, who were treated with the inhibitors compared to the control groups (7.0 vs. 4.2 months, HR = 0.58 for olaparib and 8.6 vs. 5.6 months, HR = 0.54 for talazoparib), without however a significant benefit on the OS (19.3 vs. 17.1 months for olaparib and 19.3 vs. 19.5 months for talazoparib).
Circulating tumor cells (CTCs) play an important role in the metastatic activity. The prognostic and predictive value of CTCs in BC is being constantly sought. In fact, CTCs’ detection and enumeration has been reported in early as well as metastatic BC and has been associated with poor clinical outcome (decreased PFS and OS)
[32][33]. The epithelial marker, cytokeratin (CK) and particularly A45-B/B3 antibody (CK8, CK18, CK19) has been widely used for the characterization of cells as CTCs. Attention has lately been focused on the heterogeneity of CTCs and the identification of proteins/biomarkers and/or mutations in driver genes with prognostic and/or predictive significance
[34][35][36].
2. PARP-1 Expression and BRCA1 Mutations in Breast Cancer Patients’ CTCs
Based on BC subtype, expression of PARP-1 was identified in 61% of luminal and 41% of TNBC patients. Based on the disease status, 54% and 33% of patients with early and metastatic BC, respectively, expressed PARP-1. Notably, researchers have found that PARP-1 expression is higher in patients with less severe disease status, i.e., early stage and luminal patients. The percentage of PARP-1 expression in tissue microarrays so far is approximately 24–33%
[17][37]. In addition, after immunohistochemistry of biopsies/tissue microarrays, high (cytoplasmic) PARP expression was detected in 36% of TNBC tumors
[17]. Furthermore, PARP-1 mRNA as well as protein expression has been found to be decreased in luminal compared to HER2-enriched and basal BC tumors
[38]. The role of PARP-1 expression in tumor grade is ambiguous; high tumor grade has been reported to correlate with low (
p = 0.003)
[16] or high
[17][37] PARP-1 expression in BC. Furthermore, high PARP-1 expression has been associated with a more advanced clinical stage
[16].
Examination of PARP-1 localization in BC patients showed that PARP-1 was distributed either in the nucleus or in the cytoplasm of CTCs. It has been shown that PARP-1 has a nuclear localization signal in its N-terminal DNA-binding domain
[39]. In fact, the presence of PARP-1 in the nucleus is essential for the maintenance of genomic integrity and cell survival. Hence the importance of the nuclear localization for PARP functions is well established, whereas the cytoplasmic localization has equally gained interest
[18]. Therefore, it was noteworthy that in the study the nuclear localization was more prominent in TNBC patients’ CTCs (61%), whereas cytoplasmic localization was mostly seen in luminal patients (43%,
p = 0.024), showcasing a higher percentage of CTCs with cytosolic PARP-1. These results imply that PARP-1 was potentially more active in TNBC patients. In line to this observation, the data based on disease status, confirmed nuclear expression of PARP-1 in early and metastatic TNBC patients’ CTCs. Similar to the observations, distribution of PARP-1 in the two subcellular compartments has been previously shown in tissues of different subsets of BC patients
[16][17]. PARP-1 expression has been previously determined in the majority (82%) of BRCA1-dependent BC cases, with a predominantly nuclear subcellular localization
[40].
Although this is a pilot study, and the cohort of patients is rather small it was interesting that the CK-positive early BC patients with the CK
+PARP
+ phenotype had longer OS compared to the negative patients (log-rank
p = 0.046). KMplot database (
https://kmplot.com/analysis/, accessed on 27 September 2021), further supported the data, by demonstrating that high PARP-1 expression was correlated with longer DFI (log-rank
p = 0.022, HR = 0.67) and OS (log-rank
p = 0.015, HR = 0.59) in basal-like BC patients
[41]; TNBC is classified as a subtype of basal-like BC. However, a meta-analysis has shown that high PARP expression was associated with poor OS in BC patients (HR = 1.38, 95% CI = 1.28–1.49,
p < 0.001)
[42].
One of the main objectives of the study was also to detect
BRCA1 mutations in CTCs and not in plasma, to directly compare mutations in cancer cells and PARP expression. Furthermore, the origin of these mutations in plasma is not defined (normal or tumor cells). This approach is giving a better profile of CTCs regarding the DNA repair system. In addition, it is well known that some mutations do not exist in the primary tumor and can be detected in CTCs at a distinct time point during cancer evolution
[43]. Furthermore, it is possible that if the primary tumor is removed and the number of CTCs is very low, the “de novo” mutations could not be detected in the plasma. In addition, the approach is an easy assay with one isolation step of CTCs (Ficoll density gradient).
Nine
BRCA1 mutations were detected in BC patients’ CTCs. The majority of mutations was observed in the CTCs of TNBC patients compared to luminal patients, suggesting that mutational burden tends to be higher in TNBC patients. This coincides with the results of a meta-analysis suggesting that patients with
BRCA1 mutations will more probably have more TNBC and higher tumor burden
[44].
Of the identified pathogenic mutations, M7 was exclusively detected in TNBC patients (19%,
p = 0.03), while M8 was mostly found in TNBC patients (double the percentage of that seen in luminal patients). M7 is a splice acceptor mutation (NCBI;
https://www.ncbi.nlm.nih.gov/snp/rs80358173, accessed on 13 September 2021) and is identified as a pathogenic mutation with Combined Annotation-Dependent Depletion (CADD) > A:34, G:34, C:34, belonging to 0.1% of the most harmful SNPs. M8 is a missense variant with CADD > G:26.8, which means that it belongs to the 1% of the most harmful SNPs. Among the TNBC patients, M7 and M8 were detected in more metastatic patients (44% for both,
p = 0.014 and
p = 0.002, respectively) compared to early patients. M4 (A > T mutation) (
https://www.ncbi.nlm.nih.gov/snp/rs80358069, accessed on 13 September 2021) is a splice acceptor variant. In Chevalier
et al., the reported A > G mutation (c.5194-2A > G) caused a change in the splice acceptor region which affects alternative splicing and is thus identified as pathogenic
[45]. TNBC patients bearing the M4 (
p = 0.008), M7 (
p = 0.019) and M8 (
p = 0.019) mutations were also correlated with decreased OS. These observations tend to agree with the characterization of the mutations as pathogenic and suggest that they can indeed be of particular interest. However, the cohort is small, and these results are only indicative of the severity of these mutations.
Olaparib has been approved by the FDA for adjuvant treatment of BRCA-mutated HER2-negative high-risk early-stage BC patients (phase III OlympiAD trial). The approach can give a real time observation of BRCA1 and PARP expression in cancer cells, providing an interesting tool for stratifying patients that could benefit from this target therapy. Furthermore, the analysis revealed that among the BC patients the majority of BRCA1 mutations were observed in TNBC patients, who also expressed nuclear PARP-1. This is rather interesting as it would imply that patients with high mutation burden and PARP-1 nuclear expression would be more likely to benefit from current regimens with PARP inhibitors.