PARP1 enzyme plays an important role in DNA damage recognition and signalling. PARP inhibitors are approved in breast, ovarian, pancreatic, and prostate cancers harbouring a pathogenic variant in BRCA1 or BRCA2, where PARP1 inhibition results mainly in synthetic lethality in cells with impaired homologous recombination. However, the increasingly wide use of PARP inhibitors in clinical practice has highlighted the problem of resistance to therapy. Several different mechanisms of resistance have been proposed, although only the acquisition of secondary mutations in BRCA1/2 has been clinically proved.
1. Introduction
Poly-(ADP-ribose) polymerase (PARP) enzyme PARP1 plays an important role in DNA damage recognition and signalling, as it binds single-stranded DNA breaks (SSBs) and then organizes their repair by synthesising PAR chains on target proteins (the so-called PARylation)
[1]. In detail PARP1, once bound to SSBs via N-terminal zinc-finger DNA-binding domain, catalyses the polymerization of ADP-ribose moieties from NAD+ on target proteins, mainly PARP1 itself and histones. This process leads to chromatin relaxation and recruitment of other DNA repair enzymes such as XRCC1
[2,3,4,5][2][3][4][5]. The scaffolding protein XRCC1 stimulates the DNA kinase and DNA phosphatase activities of polynucleotide kinase at SSBs, accelerating the base excision repair reaction
[6]. It is also reported that PARP1 promotes recruiting of MRE11, ATM and BRCA1, which are involved in double-stranded DNA break (DSB) repair by homologous recombination (HR)
[7,8,9][7][8][9]. While PARP1 DNA binding is independent of its catalytic activity, dissociation of PARP1 from DNA requires PARylation presumably through a steric mechanism due to highly negatively charged PAR chains
[10].
PARP inhibitors (PARPi) act mainly in a double way. The first proposed mechanism is the inhibition of the catalytic activity of PARP1, which results in synthetic lethality in cells with impaired HR
[11,12,13][11][12][13]. In fact, inhibition of PARP1 promotes SSBs, which, if unrepaired, consequently lead to DSBs by collapse of the stalled replication fork during DNA replication
[14]. In eukaryotic cells, DSBs are mainly resolved by the error-free mechanism of HR, which uses the intact sister chromatid as a template. HR deficiency induces activation of the more error-prone template-independent non-homologous end-joining (NHEJ) pathway, therefore, together with PARPi causing genomic instability followed by cell death
[15,16][15][16]. Subsequent studies have revealed that most PARPi cause cytotoxicity by trapping PARP1 at sites of DNA damage
[17,18,19][17][18][19]. According to the hypothesis proposed by Murai et al.
[18], PARPi binding to the catalytic domain of PARP1 allosterically modifies interactions between DNA and the N-terminal DNA-binding domain of the protein, to the point that PARP1 becomes trapped on DNA. More recently, a third mechanism of PARPi sensitivity has been identified
[20,21][20][21]. In cells with HR deficiency, aside from the NHEJ pathway, DSBs can be repaired by the microhomology-mediated end joining (MMEJ or Alt-EJ) pathway. Similarly to NHEJ, MMEJ is intrinsically error-prone, as the use of regions of microhomology inside DNA leads to deletions of nucleotides from the strand being repaired and to chromosomal translocations. In this pathway, the efficient recruitment of the DNA polymerase POLQ to the DSB requires PARP1. A PARPi will, therefore, block the MMEJ pathway and cause HR-deficient cell death.
PARP inhibitors are actually approved in breast, ovarian, pancreatic and prostate cancers harbouring a pathogenic or likely pathogenic variant in
BRCA1 or
BRCA2 (
BRCA1/2)
[22,23,24][22][23][24].
BRCA1/2 mutation prevalence varied widely from 1.8% in sporadic breast cancer to 36.9% in estrogen receptor/progesterone receptor low HER2 negative breast cancer
[25]. Germline mutations in
BRCA1/
2 have been identified in 13–15% of women diagnosed with ovarian cancer, and somatic mutations are found in an additional 7%
[26,27,28][26][27][28]. Germline
BRCA1/2 mutations can be found in up to 8% of patients with sporadic pancreatic cancer
[29]. In a sample of 692 patients with metastatic prostate cancer, unselected for family history or age at diagnosis, 5.3% carried a
BRCA2 mutation, and 0.9% carried a
BRCA1 mutation
[30]. The increasingly wide use of PARPi in clinical practice is highlighting the problem of resistance to therapy. Considering the complex interaction between PARP1, HR and other DNA damage repair pathways in the setting of
BRCA1/2 mutated cancers, several different mechanisms of resistance have been proposed, although some of them have been only described preclinically. The aim of this
entr
eviewy is to outline the key molecular findings that could explain the mechanisms of primary or secondary resistance to PARPi (summarised in
Table 1).
Table 1.
Proposed mechanisms of PARPi resistance.
Resistance Mechanism |
Evidence |
References |
Primary resistance |
|
|
PI3K/AKT pathway activation |
Cell lines |
Yi et al. [31] |
Wild-type PTEN |
Cell lines |
Dedes et al. [32] |
Loss of NHEJ |
Cell lines |
Balmus et al. [33], Patel et al. [34], Mc Cormick et al. [35] |
ALC1 overexpression |
Cell lines |
Juhász et al. [36] |
Secondary resistance |
|
|
Upregulation of ABC transporters |
Mouse models, cell lines |
Jaspers et al. [37], Vaidyanathan et al. [38] |
Decreased PARP1 trapping |
Mouse models, cell lines |
Pettitt et al. [39], Gogola et al. [40] |
Restoration of HR |
|
|
-BRCA1/2 reversion mutations |
Tumour DNA and ctDNA from cancer patients |
Tobalina et al. [41] |
-Hypomorphic BRCA1 allele |
Cell lines, mouse models, PDXs |
Drost et al. [42], Wang et al. [43], Cruz et al. [44], Wang et al. [45], Castroviejo-Bermejo et al. [46] |
-Loss of BRCA1 promoter methylation |
Cell lines, PDXs |
Ter Brugge et al. [47], Veeck et al. [48], Wang et al. [49] |
-Loss of end resection regulation (53BP1, RIF1, REV7, Sheldin complex or DYNLL1 depletion) |
Cell lines |
Belotserkovskaya et al. [50], Xu et al. [51], Noordermeer et al. [52], Gupta et al. [53], He et al. [54] |
-RAD51 overexpression |
Ovarian cancer samples, cell lines |
Kondrashova et al. [55], Liu et al. [56], Marzio et al. [57] |
Stabilization of stalled fork (FANCD2 overexpression, RADX depletion, SMARCAL1 inactivation,) |
Cell lines |
Michl et al. [58], Chaudhuri et al. [59], Taglialatela et al. [60], Dungrawala et al. [61] |