Abstract
Germline mutations in Fanconi anemia (FA) genes predispose to chromosome instability syndromes, such as FA and cancers. FA gene products have traditionally been studied for their role in interstrand cross link (ICL) repair. A fraction of FA gene products are classical homologous recombination (HR) factors that are involved in repairing DNA double-strand breaks (DSBs) in an error-free manner. Emerging evidence suggests that, independent of ICL and HR repair, FA genes protect DNA replication forks in the presence of replication stress. Therefore, understanding the precise function of FA genes and their role in promoting genome stability in response to DNA replication stress is crucial for diagnosing FA and FA-associated cancers. Moreover, molecular understanding of the FA pathway will greatly help to establish proper functional assays for variants of unknown significance (VUS), often encountered in clinics. In this short review, we discuss the recently uncovered molecular details of FA genes in replication fork protection pathways. Finally, we examine how novel FA variants predispose to FA and cancer, due to defective replication fork protection activity.
Background
Fanconi anemia (FA) is a rare genetic disorder, which is primarily autosomal recessive and affects approximately five in a million births worldwide (1). Clinically, FA is a heterogenous disorder, presented by developmental abnormalities, bone marrow failure, genome instability features, predisposition to solid tumours and leukemia (2). To date, at least 22 genetic subtypes of FA have been identified, whose gene products are crucial for proper ICL repair (1). The 22 FA genes are FANCA, FANCB, FANCC, FANCD1 (BRCA2), FANCD2, FANCE, FANCF, FANCG, FANCI, FANCJ (BRIP1), FANCL, FANCM, FANCN (PALB2), FANCO (RAD51C), FANCP (SLX4), FANCQ (ERCC4), FANCR (Rad51), FANCS (BRCA1), FANCT (UBE2T), FANCU (XRCC2), FANCV (REV7) and FANCW (RFWD3) (1,3). Out of the 22 genetic subtypes, many of the FA genes under HR pathway such as FANCO (RAD51C), FANCR (RAD51), FANCS (BRCA1) and some others are considered as ‘FA-like genes’ due to lack of clear genotype–phenotype correlation with FA (4). Bi-allelic loss of FA genes results in FA, and these patient-derived cells are inefficient at repairing ICLs. When the Watson and Crick strands are covalently linked by an ICL, preventing helicase-mediated unwinding of chromosomal DNA (5), FA gene products are primarily involved in resolving the ICLs (6). When an ICL is detected, each subunit of the FANCD2–FANCI heterodimeric complex is mono-ubiquitinated. This mono-ubiquitination results in FANCD2–FANCI being clamped onto the DNA (7–9). The mono-ubiquitination of FANCD2–FANCI leads to the recruitment of structure-specific endonucleases for the unhooking of the ICL at the site of converged replication forks. The site of the lesion is then bypassed with specific trans-lesion synthesis polymerases. Then the missing information from the opposing strand for which synthesis is not yet complete is replaced with homologous recombination (HR) (6).
Based on the biochemical activities and functions of FA genes, they are placed into three different groups. Group I comprises of FA core complex (FANCA, B, C, E, F, G, L, FAAP20 and FAAP100)—a mega dalton E3 ubiquitin ligase. Group II comprises of FAND2/FANCI heterodimeric complex. Finally, group III comprises of repair factors such as nucleases (FANCP (SLX4) and FANCQ (ERCC4)), trans-lesion synthesis polymerases (FANCV (REV7) and polymerase ζ)) and HR factors ((FANCD1 (BRCA2), FANCJ (BRIP1), FANCN (PALB2), FANCO (RAD51C), FANCR (Rad51), FANCS (BRCA1) and FANCU (XRCC2)) (Fig. 1a) (10). Group I FA core complex mono-ubiquitinates group II FANCD2/FANCI complex (Fig. 1a). Upon mono-ubiquitination, FANCD2/FANCI complex clamps onto the DNA and recruits group III repair factors consisting of nucleases, trans-lesion synthesis polymerases and HR factors for proper ICL repair.
The Fanconi anemia (FA) pathway. FA pathway comprises of 22 proteins categorised into 3 groups. In the presence of ICL-induced DNA damage and DDR activation, group I FA core complex proteins mono-ubiquitinate heterodimeric group II FANCD2/FANCI complex. FANCD2/FANCI mono-ubiquitination facilitates group III repair protein recruitment which comprises of nucleases, TLS polymerases and HR factors whose function is essential to resolve ICL lesion.
Of all the FANC genes, more than 80% of germline mutations occur in FANCA/FANCC/FANCG and ~8% in FANCE/FANCF genes (11). FA caused by the rest of the genes represents a minority of patient cases; however, with the help of next-generation sequencing, the variants of unknown significance (VUS) of rare FA genes are mounting. Functional characterization of VUS is required to assess their pathogenicity for appropriate treatment and follow-up of patients (12). Likewise, mono-allelic mutations in FA genes, within the HR pathway (from group III repair proteins, mentioned above), predispose to early-onset breast and ovarian cancer. Often, the HR-perturbing mutations in BRCA1/2 (FANDS/D1) are deletion/truncation, insertions and single-nucleotide alterations (12,13). Such germline heterozygous mutations in the essential HR factor BRCA2 (FANCD1) increase breast and ovarian cancer risk of an individual by 50 and 15% (14). Moreover, the majority of patients with bi-allelic mutations in FANCD1 display developmental anomalies, and nearly 100% develop malignancy by the age of 5 (13,15).
Functional analyses on the FA genes have been largely centred on ICL and HR repair given their importance for these processes (16–18). Detailed roles of FA genes in ICL repair were recently reviewed (16–19). However, the latest studies highlight that, besides their classical function in ICL and HR repair, FA genes are involved in DNA replication processes when replication forks are decelerated or stalled (20–23). A number of FA genes such as FANCD1 (BRCA2), FANCR (RAD51), FANCS (BRCA1), FANCU (XRCC2), FANCO (RAD51C) and FANCN (PALB2) are indeed classical HR factors which are now known to have replication fork-associated functions (24).
In this review, we summarize the recent findings and discuss the replication fork protection mechanisms of FA gene products. Interestingly, a proportion of FA genes are involved mainly in counteracting nuclease-mediated degradation of nascent DNA in the presence of replication stress to prevent DNA breaks. Furthermore, we discuss how mono-allelic or bi-allelic loss of these genes contributes to FA and cancer predisposition. Finally, we argue that compromised fork protection/restart pathways, in part, contribute to tumour initiation and progression in patients harbouring pathogenic mutation in FA genes.
FANCD1-, FANCS- and FANCR-mediated replication fork protection: lessons from cultured mammalian cell lines to patient-derived FA mutant cell lines
DNA replication forks are safeguarded by complexes that prevent accumulation of toxic DSBs (25). Replication stress induced by the chemotherapeutic agents such as hydroxyurea (HU), mitomycin C (MMC) and cisplatin has a profound effect on replication fork architecture and the rate of DNA replication, which is evident by the accumulation of ssDNA stretches at and behind replication forks (26,27). Genotoxic stress, induced by sublethal concentrations of HU, MMC and cisplatin, remodels the replicating nascent DNA into a four-way junction called reversed forks (RVFs) (Fig. 2a). RVFs are a controlled fork remodelling event, required for stalled replication fork resolution/restart (28,29). However, the accumulation of RVFs is pathological. Approximately, 20–30% of the global replication forks undergo fork reversal, in the presence of genotoxic stress, and this could potentially originate from repetitive regions that are difficult to replicate (centromeric and telomeric regions), or it may be partially stochastic. The recent understanding on the factors that promote RVFs proposes that the persistent replication fork gaps are remodelled by SWI/SNF chromatin remodelling ‘annealing helicases’ aka translocases (SMARCAL1, ZRANB3 and HLTF) and RAD51 (Fig. 2a) (30–35).
(a) When FA-proficient cells encounter replication stress or chemotherapeutic agents, the replication forks are remodelled into RVFs by SMARCAL1, RAD51, ZRANB3 and other yet to be identified fork remodelling enzymes. Controlled resection of the RVFs promotes fork recovery and genome stability. (b) FA-deficient or pathogenic patient-derived cells display defective fork protection. Though SMARCAL1, ZRANB3 and residual RAD51 trigger fork reversal, the nascent DNA is resected by MRE11 nuclease and extended by CtIP and EXO1. This degradation of nascent DNA leads to collapsed forks and chromosome breaks. This figure is created with BioRender.com.
Prolonged replication fork stalling by hydroxyurea (HU), an inhibitor of ribonucleotide reductase (RNR), depletes endogenous dNTP pools and induces replication fork collapse and DSBs. In such HU-induced fork-stalling conditions, isolation of proteins enriched on nascent DNA (iPOND) experiments suggests that the HR factor RAD51 is a stalled replication fork-associated protein (24). Further, electron microscopy (EM) visualization of DNA replication fork intermediates, using cell-free frog egg extracts, identified a key function for FANCR (RAD51), in replication fork protection pathway (23). Depleting or preventing the FANCR association onto the ongoing DNA replication forks resulted in impaired nascent DNA protection and accumulation of ssDNA at/behind the DNA replication forks (22–24,30).
Several laboratories further confirmed using EM and single-molecule fibre assays that the loss of RAD51 loader FANCD1/S (BRCA1/2) results in ssDNA accumulation and poor nascent DNA protection (20,21,31,32). This nascent DNA degradation phenomenon was found to be primarily due to aberrant MRE11 exonuclease access and activity in BRCA1/2-deficient background (20,21,31,32). Until recently, the mechanism and the nature of DNA repair substrates, which prompt degradation of nascent DNA in BRCA1/2-deficient cells, remained unclear. However, studies from the cell-free frog extracts and cultured mammalian cells from different groups reached a common conclusion that the RVFs are intermediates for nascent DNA degradation in BRCA1/2-deficient cells (Fig. 2a and b) (30–33). In the absence of BRCA1/2, the replication stress-induced fork reversal is primarily mediated by SMARCAL1, ZRANB3 and residual RAD51. This fork remodelling event acts as an entry point for nucleases such as MRE11, CtIP and EXO1 for the degradation replicating DNA (Fig. 2b). Collectively, data reviewed here so far indicate that replication fork remodelling and the subsequent protection of nascent DNA by FANCD1/S/R are crucial to prevent genome instability.
In recent years, several studies monitored the replication fork stability in patient lymphoblastoid cell lines (LCLs) and skin fibroblasts, those harbouring mutations in key FA or HR genes. Consistently, with regard to FANCR (RAD51) gene, recent identification of de novo heterozygous dominant mutation in RAD51–T131P displayed FA-like phenotype, where patient fibroblasts and lymphoblastoid cell lines (LCLs) are highly sensitive to ICL-inducing agents with moderate chromosome breaks, but proficient in homologous recombination (HR) repair (36). Further, the biochemical reconstitution studies, using in vitro DNA substrates and mammalian cell culture experiments, revealed that RAD51–T131P showed poor replication fork protection function due to compromised nucleoprotein filament-forming ability (Fig. 2b) (30,31,37). These in vitro biochemical and clinical evidence further confirm a role for FANCR (RAD51) in an HR-independent manner to maintain genome stability. Furthermore, investigation from patient LCLs and fibroblasts, from FANCS (BRCA1) somatic and germline heterozygous genetic variants, showed defective replication fork protection/restart and increased collapsed forks in the presence of HU (38). However, HR repair remained intact in these BRCA1 haploinsufficient cells similar to RAD51–T131P (38). Recently, a study provided compelling evidence that BRCA1 patients harbouring germline mutations within or close to 114-serine-proline-115 phosphorylation and BARD1 isomerisation domain displayed compromised fork stability and reduced RAD51 focus formation in the presence of nucleotide depletion (39). Overall, consistent to data from cultured mammalian cells described above, diminished fork protection appears to be a hallmark in FANCD1/S (BRCA1/2) pathogenic variants.
Role of FANCD2 in DNA replication fork protection and restart pathways
ICL lesion or HU-induced fork stalling triggers ataxia telangiectasia and Rad3-related (ATR) kinase activation. Activation of ATR and its target Chk1 elicits a DNA damage response (DDR), a requirement for FA pathway activation and completion (40). Similar to FANCD1/S/R, independent of ICL damage, FANCD2 (FA pathway) is strongly activated and co-localizes with FANCS (BRCA1) and FANCR (RAD51) in the presence of HU-induced fork stalling (41–43). FANCD2 associates with replication forks and interacts with MCM proteins in the presence of mild replication stress, induced by the nucleotide depletion or by inhibiting polymerase α/δ/ε by aphidicolin (44). Perpetually, in vitro, the binding of FANCD2/I complex onto the DNA substrates, mimicking replication fork structure, promotes efficient mono-ubiquitination of FANCD2 (45). Moreover, the removal of DNA substrates results in rapid de-ubiquitination of FANCD2/I (46). Consistent with FANCD2’s affinity towards replication forks in human primary cells, FANCD2 restricts the DNA synthesis, in order to prevent the ssDNA gaps after HU treatment (44). Indeed, the phenomenon of DNA replication restriction by FANCD2 was also observed in BRCA2-deficient cells that exhibit endogenous replication stress and increased ssDNA accumulation (a condition similar to HU treatment) (30,43).
Single-molecule DNA fibre assays revealed the replication fork protection function of FANCD2. When FANCD2-deficient cells or cells expressing mono-ubiquitination-deficient mutant FANCD2–K561R are exposed to HU treatment for 5 h, the track length of the nascent DNA is much shorter compared to cells expressing WT FANCD2 (43,47). As hypothesized, the shorter replication tracks were due to MRE11 nuclease-mediated degradation of nascent DNA in FANCD2-deficient or mono-ubiquitination-deficient cells (Fig. 2b). Remarkably, even in FANCD2-deficient cells, the replication fork remodelling by SMARCAL1 sets the platform for nascent DNA degradation (Fig. 2b) (20,32). Stalled replication forks are restarted by multiple pathways. In this context, mono-ubiquitinated FANCD2 is essential for stalled fork recovery, upon transient replication fork stalling by high dose of HU (42,43). This replication fork restart defect is more pronounced in cells that are deficient of both BRCA2 and FANCD2, suggesting an independent role for these factors in fork restart mechanisms (42,43). Collectively, these data suggest roles for FANCD2 in replication fork protection and restart mirroring the function of the key HR factors.
Very recently, cryo-EM structure characterization of human FANCD2/I together with the DNA substrate revealed that ubiquitinated FANCD2/I is locked onto DNA as a nucleoprotein filament (7–9). Complementary evidence supporting broad binding of FANCD2 to DNA damage was shown with ChIP-Seq adjacent to Cas9-induced DSBs (48). This might explain why FANCD2 forms a punctuate large focus in the presence of stalled forks, DSBs or ICL lesions (41). This is similar to RAD51, which is known to form nucleoprotein filament at ssDNA regions, forming strong foci in the presence of breaks (49). The demonstration of FANCD2/I filament formation onto dsDNA attracts a new hypothesis that the mono-ubiquitinated FANCD2 may clamp strongly onto the reversed DNA replication forks preventing nuclease (MRE11 or/and Dna2) access, downstream to SMARCAL1-, ZRANB3- and HLTF-mediated replication fork remodelling. However, this hypothesis needs to be tested in future, in the presence of recombinantly purified MRE11, Dna2, RPA and ubiquitinated FANCD2/I complex with specific DNA substrates such as reversed forks. On the contrary, we cannot rule out the possibility of ubiquitinated FANCD2/I complex locking mechanism at dsDNA, which might prevent spontaneous replication fork reversal. Persistent locking of FANCD2/I complex onto reversed forks might not allow prompt replication fork restart leading to replication fork collapse and chromosome breaks. Hence, timely de-ubiquitination of FANCD2 by USP1 might be essential for replication fork recovery, upon BRCA1/2 deficiency (50). Collectively, in the presence of replication stress, a number of molecular mechanisms, such as replication fork stabilization, fork remodelling to RVFs and controlled degradation of nascent DNA (RVFs), are required for prompt restart and genome stability. Therefore, deregulation in these processes triggers aberrant DNA repair activity and increased mutational load, leading to FA-associated cancers.
From a treatment perspective, restoration of HR and replication fork protection contribute to cancer chemoresistance in FANCD1/S (BRCA1/2)-deficient cells (51). While FANCR (RAD51) genetic loss is lethal in vertebrates, FANCR (RAD51) overexpression is associated with genome instability, cancer development, chemoresistance and poor prognosis (52). The use of RAD51 inhibitors in combination with chemotherapy, for FANCD1/S (BRCA1/2)-deficient cancers, has been proposed (52). Specific to FANCD1/S-deficient breast/ovarian cancer cells, RAD51 inhibitors might be an attractive druggable candidate for PARP1 inhibitor- or cisplatin-resistant breast and ovarian cancers.
Similarly, increased FANCD2 expression has positive correlation with ovarian carcinoma grade (42) and cancer chemoresistance (53). FANCD2 mono-ubiquitination-deficient cells are synthetic lethal with BRCA1/2 loss, suggesting that FANCD2 repair activity acts as a backup mechanism for BRCA-deficient tumour cell survival (42,43). Further analyses revealed that the mono-ubiquitinated FANCD2 did not restore HR but promoted replication fork protection and error-prone alternative end-joining (alt-EJ) in BRCA1-deficient breast cancer cells (42). These results imply the importance of FA core complex-mediated mono-ubiquitination of FANCD2 which is essential for its downstream replication fork protection function. Consistently, FA core component FANCA loss also prevents replication fork protection (20), suggesting a functional requirement for FA core complex in fork protection, especially in BRCA1/2-deficient tumours. Therefore, pharmacological inhibition of the FA core complex may overcome cisplatin- or PARP inhibitor-induced cancer chemoresistance in FANCD1/S (BRCA)-deficient cells. In search for inhibitors for FA core complex, recently, small-molecule inhibition of FANCL E3 ubiquitin ligase strongly inhibited mono-ubiquitination of FANCD2/I complex and sensitized cells for ICL-inducing compounds (54). Future preclinical studies are necessary to test the concept of drugging FA core complex in BRCA-deficient cancer cells that express high levels of FANCD2.
Concluding Remarks
For the past 10 years, studies have provided a weight of evidence that FA genes play key roles in DNA replication beyond ICL repair. With the advent of sequencing technologies, it is imperative to define functionally the VUS that are being encountered in clinical settings. Once VUS or potentially pathogenic variants in FA genes are identified, patient-derived cell lines become a vital resource for functional assays to define whether the variant is pathogenic or not. Since the majority of the FA genes directly or indirectly regulate RAD51 nucleofilament assembly, monitoring the ability of RAD51 focus formation might help to address the efficiency of HR. This assay can complement genomic scar assays that are now being used to detect HR deficiency (55). Further, it is becoming evident that some pathogenic variants are proficient in HR but display compromised fork protection and recovery in the presence of ICL lesions. For such variants precise methods to monitor subtle genome instability features at the level of replication forks are crucial. Recent study used replication fork protection assays successfully in short-term patient-derived ovarian cancer organoid system to predict chemotherapeutic response (56). Of special interest, defective replication fork protection and reduced RAD51 focus formation positively correlated with cancer cell sensitivity when treated with carboplatin in combination with ATR inhibitor (56). The use of replication fork protection assay is still embryonic in clinical settings. To test the reproducibility of fork protection assays, a large number of studies using patient-derived LCLs, fibroblasts and organoid models are necessary. Furthermore, currently, the capability to assess replication fork protection with fibre combing assays was mostly performed in the presence of prolonged high dose of replication fork-stalling agents such as HU and aphidicolin. Such prolonged fork stalling is unlikely under physiological context. Therefore, thorough understanding on the replication fork protection/stability pathways is required to precisely assess the relationship between fork stability, chemoresistance and sensitivity.
Funding
J.N. is a NMRC Singapore Clinician Scientist. W.C. is an NHMRC career development fellow (GNT1129757).
Conflict of Interest statement: No conflict of interest to declare.
