Large-scale meta-analysis and precision functional assays identify FANCM regions in which PTVs confer different risks for ER-negative and triple-negative breast cancer.

1
Introduction
Many genes have been proposed as risk factors for breast cancer, but few have been established with convincing evidence. Protein truncating variants (PTVs) and some missense variants in BRCA1, BRCA2, and PALB2 are known to be high-penetrance risk factors for breast cancer, and there is a growing consensus that i) PTVs and some missense variants in BARD1, RAD51C, and RAD51D are moderate risk factors for estrogen receptor-negative (ER-neg) breast cancer; ii) PTVs in ATM and CHEK2 are moderate risk factors for ER-positive (ER-pos) breast cancer; and iii) both missenses and PTVs in TP53 are associated with breast cancer risk, although the magnitude of the risk remains to be determined [[1], [2], [3], [4]]. Other breast cancer genes have recently emerged and require further genetic and functional validation.
The FANCM gene encodes a 2048 amino acid (AA) sized protein involved in DNA replication stress response and inter-strand cross-link repair mediated by the Fanconi anemia (FA) pathway [5,6]. FANCM was first implicated in breast cancer risk when the c.5791C > T PTV, known as p.Arg1931∗ but re-annotated as p.Gly1906Alafs∗12 due to its spliceogenic effect, was detected in a breast cancer family [7,8]. Subsequently, case-control studies were performed — including one in non-Finnish Caucasians familial cases with no BRCA1 and BRCA2 pathogenic variants [9] — but the association of FANCM PTVs with overall breast cancer risk was not consistently found [10]. Recently, two very large studies were conducted. BRIDGES reported a borderline association between FANCM PTVs and the risk of ER-neg and triple-negative (TN) breast cancer subtypes [1], whereas CARRIERS found no significant association across different subgroups possibly due to its reduced statistical power [2]. We have previously evaluated the three most common PTVs p.Arg658∗, p.Gln1701∗ and p.Gly1906Alafs∗12. We initially observed that the C-terminal p.Gly1906Alafs∗12 was associated with risk for familial breast cancer (OR = 3.67, P = 0.043) [11]. Later, analyzing studies that were predominantly population-based, we found that the N-terminal p.Arg658∗ was associated with a moderate risk of ER-neg (OR = 2.44, P = 0.034) and TN (OR = 3.79, P = 0.009), whereas no association was found for p.Gln1701∗ and an inconsistent association was found between p.Gly1906Alafs∗12 and ER-neg breast cancer [11,12]. In addition, clinical observations suggested that p.Arg658∗ may confer a higher breast cancer risk and more severe cellular phenotypes than the C-terminal p.Gln1701∗ and p.Gly1906Alafs∗12 [13,14].
In the present study, we aimed to clarify the effect of FANCM PTVs on breast cancer risk by conducting single variant and burden analyses. To this end, we performed meta-analyses of seven case-control studies including clinical and gene variant sequencing data from a total of 268 313 women. Additionally, we performed CRISPR-Select assays to evaluate the impact of p.Arg658∗, p.Gln1701∗ and p.Gly1906Alafs∗12 on cell survival, and we agnostically investigated the essentiality of different FANCM regions using a CRISPR-Cas9-based assay.

2
Materials and methods
2.1
Meta-analyses of genetic data from case-control studies
In these analyses, we included data from seven studies [1,2,11,[15], [16], [17], [18]] totaling 144 681 breast cancer cases and 123 632 controls (Table 1). For some of these studies, analysis results of the association between FANCM PTVs and breast cancer risk were previously published. Raw data were obtained from the BRIDGES, OncoArray and GENESIS studies, whereas summary statistics from locally performed analyses were available for the remaining four studies. For each study, odds ratio (OR) with 95 % confidence intervals (CIs) and P-value (P) for PTV carriers compared with non-carriers were estimated using univariable logistic regression. PTVs were analyzed either individually (p.Arg658∗, p.Gln1701∗, and p.Gly1906Alafs∗12) or combined in burden analyses. In these statistical analyses, breast cancer cases were tested combined or based on their classification, when known, into ER-neg, ER-pos and TN breast cancer subgroups.
Meta-analyses were performed by combining the ORs from each study. We applied fixed-effect models when no substantial heterogeneity was observed across studies (I2<30 %), whereas random-effects models with restricted maximum likelihood (REML) estimation were used when heterogeneity was higher (I2≥ 30 %). Statistical analyses were performed using STATA version 15.1 (StataCorp LLC, College Station, Texas, USA) and the R software. All tests were two-tailed, and P < 0.05 was considered statistically significant.

2.2
CRISPR-SelectTIME assay
The CRISPR-SelectTIME assay was used as previously described [19]. For one 9.6 cm2 well, 60 pmol each of crRNA (Supp.Table 1) and tracrRNA were mixed and allowed to complex for 10 min. One hundred and twenty-five μl of OptiMEM were added to the mixture followed by 10 pmol of each of the variants under analysis and of the synonymous internal normalization variant (WT’) single-stranded oligo DNAs (ssODNs) and Lipofectamine (Thermo Fisher Scientific) before dripping on iCas9-MCF10A cells (TP53 WT). On days 2 and 12 post-transfection, cell population aliquots were collected for the variant:WT’ ratio analysis. Genomic DNA was extracted from cell population aliquots using GenElute Mammalian Genomic DNA Miniprep (Sigma). Two rounds of PCR amplifications (Supp.Table 1) were performed to prepare the products for NGS. Finally, an amplicon sequencing library was generated using MiSeq Reagent kit V2 (Illumina) and sequenced on a MiSeq instrument (Illumina). The NGS data were analyzed using the CRISPResso2 online tool (http://crispresso.pinellolab.org). The frequencies of frameshift indels represented knockout efficiency and the frequencies of each edited variant represented knock-in efficiency.

2.3
Sensitivity of FANCM to genome editing
We used the haploid HAP1LIG4− cell line as previously described (LIG4 KO clone#5 in Ref. [20]) in which FANCM is essential [21]. Therefore, if essential functions of the gene are disrupted, these cells are expected to show a loss of fitness or die. Twelve crRNA (Integrated DNA Technologies) were designed to target six FANCM regions (Supp.Table 1). Each of the crRNAs was annealed with the common tracrRNA (Integrated DNA Technologies) at a final concentration of 1 μM and complexed with 6 pmol of Cas9 Nuclease (Integrated DNA Technologies) in OptiMEM. Each crRNA was named according to the predicted AA position of the Cas9-mediated double-stranded break (DSB). Pre-assembled ribonucleoprotein complexes were transfected into HAP1LIG4− cells using Lipofectamine CRISPRMAX Cas9 Transfection Reagent (Thermo Fisher Scientific). Cells were cultured for 7 days before limited dilution. Approximately 30 clones were amplified for each of the fourteen transfections. DNA was extracted from each clone using Gitschier buffer and 0.5 mg/mL Proteinase K and subjected to PCR amplification (Supp.Table 1). Amplified DNAs were subjected to Sanger sequencing and analyzed using Geneious, TIDE software [22] and InDelphi algorithm [23]. Each of the 12 transfections was repeated three times and a total of 1045 clones were annotated.

3
Results
3.1
Meta-analyses of case-control studies
We conducted meta-analyses combining data from seven case-control studies (Table 1). First, we assessed the effect of the three most common PTVs — p.Arg658∗, p.Gln1701∗ and p.Gly1906Alafs∗12 — on breast cancer risk. Carriers of the p.Arg658∗ PTV showed a significantly increased risk of ER-neg (OR = 2.08, 95 % CI 1.07−4.02, P = 0.030) and TN (OR = 3.26, 95 % CI 1.48−7.20, P = 0.0034) breast cancer subtypes. Triple-negative breast cancer cases are included in the ER-neg group; hence, it should be noted that the risk effect observed in the ER-neg group could be seen as a diluted effect of the risk observed in the TN breast cancer group. In contrast, the C-terminal p.Gln1701∗ and p.Gly1906Alafs∗12 PTVs showed no association with overall breast cancer risk or with the risks of ER-neg, ER-pos and TN (Fig. 1 and Supp.Table 2).
Then, we performed a burden analysis to assess the effect of any FANCM PTV on overall breast cancer risk, and on ER-pos, ER-neg, and TN subtypes. Carriers of FANCM PTVs had an increased risk of developing ER-neg breast cancer (OR = 1.38, 95 % CI 1.10−1.72, P = 0.0047). However, no association was observed between all PTVs and overall breast cancer risk, or the ER-pos and TN subtypes (Fig. 2 and Sup.Table 2).
We repeated the burden analysis excluding p.Gln1701∗ and p.Gly190Alafs∗12. In this analysis, carriers of any of the remaining FANCM PTVs showed an increased risk of developing ER-neg (OR = 1.41, 95 % CI 1.05−1.90, P = 0.023) and TN (OR = 1.64, 95 % CI 1.07−2.51, P = 0.0023, Fig. 2 and Supp.Table 2), but not ER-pos or breast cancer overall. Finally, we performed an additional burden analysis excluding p.Arg658∗ and found that carriers of any of the remaining FANCM PTVs are at increased risk of developing ER-neg breast cancer (OR = 1.43, 95 % CI 1.05−1.93, P = 0.022, Supp.Table 2).

3.2
CRISPR-SelectTIME functional assay on the three most common PTVs
We performed the CRISPR-SelectTIME assay to investigate the effect of the p.Arg658∗, p.Gln1701∗ and p.Gly1906Alafs∗12 PTVs on iCas9-MCF10A cells (Fig. 3). This quantitative assay tracks the frequency of the tested variants over time relative to a neutral variant (WT’) that serves as an internal control [19]. First, we tested the common variant p.Asn655Ser (c.1964A > G) [24] as a negative control and the p.Arg200∗ (c.597A > G) PTV as a positive control. As expected, there was no significant change in the variant/WT’ ratio for p.Asn655Ser and a decrease in the variant/WT’ ratio for p.Arg200∗ (Fig. 3 and Supp.Figure 1). When we tested the p.Arg658∗, p.Gln1701∗ and p.Gly1906Alafs∗12 PTVs, we observed a decrease in the variant/WT’ ratio for all of the PTVs, but with a statistically significant reduction for p.Arg658∗ and p.Gly1906Alafs∗12 (48 % and 33 % respectively) (Fig. 3). We obtained knock-in efficiencies of 1.84 % and 3.13 % for p.Arg658∗ and Gly1906Alafs∗12, respectively (Supp.Figure 1). The lower knock-in efficiency of 0.56 % for p.Gln1701∗ was compensated by the reproducibility of the replicate results. Taken together, these data suggest that p.Arg658∗ is functionally more deleterious than the p.Gln1701∗ and possibly of p.Gly1906Alafs∗12.

3.3
FANCM sensitivity assay in HAP1LIG4− cells
We then tested the gene agnostically by knocking out different regions, using haploid HAP1LIG4− cells in which FANCM is essential (Fig. 4A) [21]. In this assay, for each of the 12 crRNAs used (Fig. 4B), we compared the observed “clone genotype patterns” with the DNA repair outcomes of Cas9-induced DSBs predicted by the computational model “InDelphi” [23,25] (Fig. 4C and Supp.Figure 2). We noted that all the observed recurrent in-frame indels were among the most frequently predicted by InDelphi and that frameshift events were predicted at all the 12 targeted positions.
We observed that in the three N-terminal regions AA 171–199, AA 516–522, and AA 631–658, only one KO clone was detected out of 515 annotated clones. In contrast, in the three C-terminal regions AA 725–746, AA 1257–1260 and AA 1675–1701, 83 KO clones were detected out of 530 annotated clones (P < 0.0001, Fig. 4C and Supp.Table 3). Furthermore, we noted that short in-frame indels (<21bp) were present in all targeted regions with 123 annotated clones upstream and 124 downstream of AA position 659, respectively. On the contrary, as for the KO clone distribution, long (>54bp) in-frame indels were predominantly observed downstream of AA position 658 with 37 annotated clones and only one upstream of this limit (P < 0.0001, Fig. 4C and Supp.Table 3). These findings suggest that FANCM N-terminal region upstream of the AA position 659 is essential, whereas the C-terminal region downstream of AA position 724 is not.

3.4
Meta-analyses of case-control studies assessing the effect of N-terminal and C-terminal FANCM PTVs
Based on the results of our functional analyses, we wanted to test whether or not N-terminal and C-terminal PTVs have different effects on breast cancer risk. We identified a boundary between essential and non-essential regions of FANCM located between AA position 658 and 725, which could be compatible with the existence of a shorter FANCM mRNA isoform (NM_001308134.2; ENST00000556036.5) encoding a 669 AA long protein. Therefore, N-terminal and C-terminal PTVs were sub-grouped according to this cut-off.
We observed a significant association between only N-terminal PTVs and the risk of TN breast cancer (PTVs located upstream the AA position 670: OR = 2.05, 95 % CI 1.04−4.03; P = 0.0037), whereas no significant associations were found for C-terminal PTVs (PTVs located downstream the AA position 671, Supp.Table 2). Furthermore, in the analysis of upstream PTVs excluding p.Arg658∗, we confirmed the significant association with TN breast cancer risk (OR = 2.87, 95 % CI 1.36−6.06; P = 0.006), further supporting the importance of the N-terminal FANCM region.

4
Discussion
In this study, we conducted meta-analyses of case-control studies to clarify the association between FANCM PTVs and breast cancer risk. Our findings indicate that, among the common FANCM PTVs, p.Arg658∗, p.Gln1701∗ and p.Gly1906Alafs∗12, only p.Arg658∗ resulted associated with a moderate risk of developing ER-neg and TN breast cancer subtypes (Fig. 1). It should be noted that we previously tested these three PTVs in the OncoArray study [11]. In the present single-PTV meta-analyses, additional studies were combined with OncoArray; however, their combined weight was generally limited varying between 27.7 and 47.3 % (Supp. Table 2).
In the CRISPR-SelectTIME assay, we consistently observed that the effect of the p.Arg658∗ on loss of cellular fitness was significantly stronger than that of p.Gln1701∗ and p.Gly1906Alafs∗12 (Fig. 3). Overall, these results corroborate our previously published data showing that p.Arg658∗ is a moderate risk factor for ER-neg and TN and is associated with reduced cell survival and increased number of cell chromatid breaks compared to p.Gln1701∗ and p.Gly1906Alafs∗12 [11]. As further evidence of this reduced effect of p.Gln1701∗ and p.Gly1906Alafs∗12 on breast cancer risk, more convincing associations for both ER-neg and TN were observed when these two variants were excluded from the analysis of all PTVs combined (Fig. 2). As these two variants also account for approximately 50 % of all FANCM PTVs, their inclusion is likely to have diluted the overall association between FANCM PTVs and breast cancer risk, which may explain the BRIDGES and CARRIERS results.
Furthermore, the sensitivity to genome editing indicated that two different FANCM regions exist (Fig. 4C). The first region, from AA 171 to 658, contains the entire DEAH domain and does not tolerate KO, suggesting that it harbors essential functions. The second region, from AA position 725, tolerates KO suggesting that it contains non-essential functions. This may be explained by the existence of a shorter FANCM mRNA isoform (NM_001308134.2; ENST00000556036.5) encoding a 669 AA sized protein resulting from alternative RNA processing that causes the retention of a portion of the intron 11 containing a stop codon. Notably, short FANCM fragments containing only the N-terminal translocase domain have been shown to retain affinity for branched DNA, as well as ATPase activity and R- and D-loop resolution [[26], [27], [28], [29]]. It is known that mRNA transcripts carrying a premature termination codon (PTC) are expected to be completely or partially degraded by nonsense-mediated decay (NMD) unless the PTCs are located in the last exon or in the 3′-most 50 bp of the penultimate exon [30]. Accordingly, we argue that FANCM PTVs may induce different NMD effects on the full-length and short mRNA isoforms, depending on their location. In particular, we hypothesize that p.Arg658∗ would affect both the short and full-length isoforms. The p.Gln1701∗ would not affect the expression of the short isoform but would cause the full-length isoform to be degraded, at least partially, by NMD. The p.Gly1906Alafs∗12 alters normal splicing by introducing a PTC which, due to its position, is expected to escape NMD [8]. This PTV would therefore result in the expression of a normal short isoform and of a protein truncated in the FAAP24 interacting domain. This might explain the different levels of breast cancer risk observed. Moreover, the N-terminal domain of FANCM was recently found to have a dual role [31]. In this context, the essentiality of ATP-dependent branch migration activity for DNA damage survival, but not its involvement with the FA core complex, could be link to our observation since we observed that only the DEAH domain of FANCM was essential but not the entire MM1 domain.
To further investigate the differential essentiality of FANCM functions, we assessed the effect of PTVs located upstream or downstream of AA position 670 on breast cancer risk (Supp.Table 2). We only identified a significant association between upstream PTVs and the risk of TN breast cancer. Notably, the analysis of upstream PTVs excluding p.Arg658∗ revealed a particularly strong association with TN breast cancer (OR = 2.87), suggesting that even relatively rarer N-terminal truncations may confer substantial risk for this aggressive breast cancer subtype. The lack of association with the risk of ER-neg breast cancer was however unexpected, as p.Arg658∗ alone was associated with an increased risk of both ER-neg and TN breast cancer (Fig. 1). However, the analysis of PTVs upstream of AA position 670 was based on a smaller sample size compared to the analysis of the p.Arg658∗, which may explain this discrepancy. Therefore, the analysis of PTVs upstream of AA position 670 may not have sufficient statistical power to detect the true effect of these PTVs on ER-neg breast cancer risk. Accordingly, it is noteworthy that the ORs found for PTVs upstream of AA 670 in overall breast cancer, ER-neg and TN cases showed an increasing trend, suggesting a potential true association that requires further validation.
In this study, we found that FANCM N-terminal PTVs are associated with risk of developing ER-neg and TN breast cancer subtypes. While pathogenic variants in BRCA1 predominantly predispose to the same subtypes, germline defect in BRCA2 more often lead to ER-pos tumors [1,2]. All these three genes play key function in the FA/BRCA molecular pathway that is responsible for DNA repair by homologous recombination (HR). In the pathway, FANCM and BRCA1 are early players that sense replication stress and coordinate DNA HR repair initiation, whereas BRCA2 functions later, in RAD51 filament stabilization [31,32]. Hence, it could be speculated that this mechanistic difference might explain the effects of FANCM, BRCA1 and BRCA2 on the risks of developing specific breast cancer subtypes. Additionally, cells deficient in BRCA1, BRCA2 or some other genes involved in the FA/BRCA pathway are sensitive to the poly(ADP-ribose) polymerase (PARP) inhibitors (PARPi), which are now routinely used in personalized cancer therapies. Consistently, we and other have shown that FANCM-depleted cell lines are also hypersensitive to PARPi [11,33,34]. Therefore, it is expected that cancer patients with pathogenic variants in FANCM may also benefit from PARPi-based treatments.
Our study has several strengths. First, it represents the largest meta-analysis to date assessing the association between FANCM PTVs and breast cancer risk, pooling data from more than 270 000 women across multiple large-scale cohort studies, most of which had previously published high-quality data. This unprecedented sample size substantially increased statistical power and enabled more precise risk estimates across breast cancer subtypes. Second, we complemented the genetic evidence with two different robust functional assays based on CRISPR/Cas9 technologies. These experiments highlighted the differential roles of N-terminal and C-terminal regions, consistently with the genetic associations results. Third, by grouping PTVs according to their genomic position, we proposed a biologically plausible mechanistic model linking FANCM functional domains with differential risks for ER-neg and TN breast cancer. Finally, the integration of multiple analytical approaches (single PTVs, burden, and positional analyses) strengthens the overall conclusions and provides a comprehensive view of FANCM's role in breast cancer predisposition. Some limitations should also be acknowledged. Because FANCM PTVs are rare, the number of carriers in specific subgroups was small, which reduced the statistical power of some subgroup analyses. The positional boundary, which is located between AA 670 and AA 725, was defined using functional data and is subject to refinement. For some studies, only summary-level genetic data were available rather than individual-level data, preventing more detailed analyses such as stratification by age groups or for additional covariates. Similarly, certain clinico-pathological tumor characteristics, such as HER2 status, were not consistently available across all contributing studies, limiting our ability to perform additional stratified analyses by tumor characteristics. The lack of coherent clinical information across contributing studies introduces the risk of misclassification of breast cancer subtypes, particularly for ER status, when clinical data were incomplete, inconsistently reported, or assessed with different laboratory methods. Some studies may also have used different inclusion criteria (for example, oversampling patients with strong family history of breast cancer), which could lead to overestimation of risk associations. Furthermore, the majority of the studies included women of European ancestry, which restricts the generalizability of our conclusions to other populations. Finally, this was a meta-analysis of case–control studies, and while this design is particularly suitable for studying rare exposures such as FANCM PTVs, residual confounding cannot be fully excluded.

5
Conclusion
Our functional results suggest that the N-terminal FANCM region upstream of AA 725 harbors essential functions, whereas downstream regions appear dispensable. Genetic data indicate that all FANCM PTVs, excluding the two C-terminal p.Gln1701∗ and p.Gly1906Alafs∗12, are associated with an increased risk of developing ER-neg and TN breast cancer subtypes. Specifically, PTVs upstream of AA position 670 are associated with a moderate risk of developing TN breast cancer. Importantly, our genetic and functional results support the knowledge that carriers of p.Arg658∗ have a moderate risk of developing ER-neg and TN subtypes, whereas carriers of p.Gln1701∗ and p.Gly1906Alafs∗12 appear not to be at risk for breast cancer or any disease subtype. While our data are informative for more efficient counseling of carriers of FANCM individuals PTVs, additional more extensive studies are warranted to better define the risk magnitude of FANCM PTVs.

CRediT authorship contribution statement
Amandine Billaud: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Gisella Figlioli: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Clémence Mooser: Writing – review & editing, Writing – original draft, Visualization, Investigation, Formal analysis, Data curation. Irene Casamassima: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis. Violette Azzoni: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis. Jahnavi Srivatsa: Writing – review & editing, Investigation, Formal analysis. Mara Colombo: Writing – review & editing, Writing – original draft, Formal analysis, Data curation. Laura Caleca: Writing – review & editing, Writing – original draft, Formal analysis, Data curation. Thomas U. Ahearn: Writing – review & editing, Data curation. Irene L. Andrulis: Writing – review & editing, Data curation. Antonis C. Antoniou: Writing – review & editing, Data curation. Matthias W. Beckmann: Writing – review & editing, Data curation. Sabine Behrens: Writing – review & editing, Data curation. Marina Bermisheva: Writing – review & editing, Data curation. Natalia V. Bogdanova: Writing – review & editing, Data curation. Manjeet K. Bolla: Writing – review & editing, Data curation. Bernardo Bonanni: Writing – review & editing, Data curation. Thomas Brüning: Writing – review & editing, Data curation. Nicola J. Camp: Writing – review & editing, Data curation. Archie Campbell: Writing – review & editing, Data curation. Jose E. Castelao: Writing – review & editing, Data curation. Melissa H. Cessna: Writing – review & editing, Data curation. Jenny Chang-Claude: Writing – review & editing, Data curation. Kamila Czene: Writing – review & editing, Data curation. Joe Dennis: Writing – review & editing, Data curation. Peter Devilee: Writing – review & editing, Data curation. Thilo Dörk: Writing – review & editing, Data curation. Alison M. Dunning: Writing – review & editing, Data curation. Mikael Eriksson: Writing – review & editing, Data curation. D. Gareth Evans: Writing – review & editing, Data curation. Peter A. Fasching: Data curation, Writing – review & editing. Jonine D. Figueroa: Writing – review & editing, Data curation. Marike Gabrielson: Writing – review & editing, Data curation. Manuela Gago-Dominguez: Writing – review & editing, Data curation. Anna González-Neira: Writing – review & editing, Data curation. Pascal Guénel: Writing – review & editing, Data curation. Andreas Hadjisavvas: Writing – review & editing, Data curation. Eric Hahnen: Writing – review & editing, Data curation. Ute Hamann: Writing – review & editing, Data curation. Peter Hillemanns: Writing – review & editing, Data curation. Antoinette Hollestelle: Writing – review & editing, Data curation. Maartje J. Hooning: Writing – review & editing, Data curation. Reiner Hoppe: Writing – review & editing, Data curation. Anthony Howell: Writing – review & editing, Data curation. Anna Jakubowska: Writing – review & editing, Data curation. Vessela N. Kristensen: Writing – review & editing, Data curation. Jan Lubiński: Writing – review & editing, Data curation. Michael Lush: Writing – review & editing, Data curation. Siranoush Manoukian: Writing – review & editing, Data curation. Dimitrios Mavroudis: Writing – review & editing, Data curation. Roger L. Milne: Writing – review & editing, Data curation. Anna Marie Mulligan: Writing – review & editing, Data curation. William G. Newman: Writing – review & editing, Data curation. Nadia Obi: Writing – review & editing, Data curation. Mihalis I. Panayiotidis: Writing – review & editing, Data curation. Guillermo Pita: Writing – review & editing, Data curation. Muhammad U. Rashid: Writing – review & editing, Data curation. Valerie Rhenius: Writing – review & editing, Data curation. Emmanouil Saloustros: Writing – review & editing, Data curation. Elinor J. Sawyer: Writing – review & editing, Data curation. Rita K. Schmutzler: Writing – review & editing, Data curation. Mitul Shah: Writing – review & editing, Data curation. Melissa C. Southey: Writing – review & editing, Data curation. Amanda B. Spurdle: Writing – review & editing, Data curation. Thérèse Truong: Writing – review & editing, Data curation. Qin Wang: Writing – review & editing, Data curation. Camilla Wendt: Writing – review & editing, Data curation. Paul L. Auer: Writing – review & editing, Data curation. Nicholas J. Boddicker: Writing – review & editing, Data curation. Clara Bodelon: Writing – review & editing, Data curation. Elizabeth S. Burnside: Writing – review & editing, Data curation. Fei Chen: Writing – review & editing, Data curation. Fergus J. Couch: Writing – review & editing, Data curation. Susan M. Domchek: Writing – review & editing, Data curation. Heather A. Eliassen: Writing – review & editing, Data curation. Christopher Haiman: Writing – review & editing, Data curation. James M. Hodge: Writing – review & editing, Data curation. Chunling Hu: Writing – review & editing, Data curation. Hongyan Huang: Writing – review & editing, Data curation. Sara Lindstrom: Writing – review & editing, Data curation. Maria Elena Martinez: Writing – review & editing, Data curation. Katherine L. Nathanson: Writing – review & editing, Data curation. Susan L. Neuhausen: Writing – review & editing, Data curation. Katie M. O'Brien: Writing – review & editing, Data curation. Janet E. Olson: Writing – review & editing, Data curation. Alpa V. Patel: Writing – review & editing, Data curation. Kathryn J. Ruddy: Writing – review & editing, Data curation. Dale P. Sandler: Writing – review & editing, Data curation. Lauren R. Teras: Writing – review & editing, Data curation. Clarice R. Weinberg: Writing – review & editing, Data curation. Jeffrey N. Weitzel: Writing – review & editing, Data curation. Stacey J. Winham: Writing – review & editing, Data curation. Siddhartha Yadav: Writing – review & editing, Data curation. Song Yao: Writing – review & editing, Data curation. Gary Zirpoli: Writing – review & editing, Data curation. Marketa Janatova: Writing – review & editing, Data curation. Zdenek Kleibl: Writing – review & editing, Data curation. Petra Kleiblova: Writing – review & editing, Data curation. Jana Soukupova: Writing – review & editing, Data curation. Qihong Zhao: Writing – review & editing, Data curation. Lisa Devereux: Writing – review & editing, Data curation. Paul A. James: Writing – review & editing, Data curation. Ian G. Campbell: Writing – review & editing, Data curation. Tu Nguyen-Dumont: Writing – review & editing, Data curation. James G. Dowty: Writing – review & editing, Data curation. Nadine Andrieu: Writing – review & editing, Data curation. Fabienne Lesueur: Writing – review & editing, Data curation. Dominique Stoppa-Lyonnet: Writing – review & editing, Data curation. Miguel de la Hoya: Writing – review & editing, Data curation, Conceptualization. Paolo Radice: Writing – review & editing, Data curation. Claus Storgaard Sørensen: Writing – review & editing, Data curation. Paolo Peterlongo: Writing – review & editing, Data curation.

Funding
This study was supported by a grant from the Fondazione AIRC per la Ricerca sul Cancro (AIRC; IG22860) to Paolo Peterlongo and by fellowships from the Fondazione Umberto Veronesi to Gisella Figlioli and Amandine Billaud. Paolo Peterlongo was also supported in part by the Italian Ministry of Health (Ricerca Corrente and 5 per mille).
This work was supported by Cancer Research UK grant: PPRPGM-Nov20\100002 and by core funding from the NIHR Cambridge Biomedical Research Centre (NIHR203312) [∗]. ∗The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care. Additional funding for BCAC is provided by the Confluence project which is funded with intramural funds from the National Cancer Institute
Intramural Research Program, National Institutes of Health, the European Union's Horizon 2020 Research and Innovation Programme (grant numbers 634935 and 633784 for BRIDGES and B-CAST respectively), and the PERSPECTIVE I&I project, funded by the Government of Canada through Genome Canada and the Canadian Institutes of Health Research, the Ministère de l’Économie et de l'Innovation du Québec through Genome Québec, the Quebec Breast Cancer Foundation. The EU
Horizon 2020 Research and Innovation Programme funding source had no role in study design, data collection, data analysis, data interpretation or writing of the report.
Genotyping of the OncoArray was funded by the NIH Grant U19 CA148065, and Cancer Research UK Grant C1287/A16563 and the PERSPECTIVE project supported by the Government of Canada through Genome Canada and the Canadian Institutes of Health Research (grant GPH-129344) and, the Ministère de l’Économie, Science et Innovation du Québec through Genome Québec and the PSRSIIRI-701 grant, and the Quebec Breast Cancer Foundation.
The BRIDGES panel sequencing was supported by the European Union
Horizon 2020 research and innovation program BRIDGES (grant number, 634935) and the Wellcome Trust (v203477/Z/16/Z).
The ABCS and ABCS-F studies were supported by the Dutch Cancer Society [grants NKI 2007-3839; 2009 4363] and an institutional grant of the Dutch Cancer Society and of the Dutch Ministry of Health, Welfare and Sport. The work of the BBCC was partly funded by ELAN-Fond of the University Hospital of Erlangen. For BIGGS, ES is supported by NIHR Comprehensive Biomedical Research Centre, Guy's & St. Thomas' NHS Foundation Trust in partnership with King's College London, United Kingdom. IT is supported by the Oxford Biomedical Research Centre. The BREast Oncology GAlician Network (BREOGAN) is funded by Acción Estratégica de Salud del Instituto de Salud Carlos III FIS PI12/02125/Cofinanciado and FEDER PI17/00918/Cofinanciado FEDER; Acción Estratégica de Salud del Instituto de Salud Carlos III FIS Intrasalud (PI13/01136); Programa Grupos Emergentes, Cancer Genetics Unit, Instituto de Investigacion Biomedica Galicia Sur. Xerencia de Xestion Integrada de Vigo-SERGAS, Instituto de Salud Carlos III, Spain; Grant 10CSA012E, Consellería de Industria Programa Sectorial de Investigación Aplicada, PEME I + D e I + D Suma del Plan Gallego de Investigación, Desarrollo e Innovación Tecnológica de la Consellería de Industria de la Xunta de Galicia, Spain; Grant EC11-192. Fomento de la Investigación Clínica Independiente, Ministerio de Sanidad, Servicios Sociales e Igualdad, Spain; and Grant FEDER-Innterconecta. Ministerio de Economia y Competitividad, Xunta de Galicia, Spain. The BSUCH study was supported by the Dietmar-Hopp Foundation, the Helmholtz Society and the German Cancer Research Center (DKFZ). CCGP is supported by funding from the University of Crete. The CECILE study was supported by Fondation de France, Institut National du Cancer (INCa), Ligue Nationale contre le Cancer, Agence Nationale de Sécurité Sanitaire, de l'Alimentation, de l'Environnement et du Travail (ANSES), Agence Nationale de la Recherche (ANR). The CNIO-BCS was supported by the Instituto de Salud Carlos III, the Red Temática de Investigación Cooperativa en Cáncer and grants from the Asociación Española Contra el Cáncer and the Fondo de Investigación Sanitario (PI11/00923 and PI12/00070). FHRISK and PROCAS are funded from NIHR grant PGfAR 0707-10031. DGE, AH and WGN are supported by the NIHR Manchester Biomedical Research Centre (NIHR203308). The GC-HBOC (German Consortium of Hereditary Breast and Ovarian Cancer) is supported by the German Cancer Aid (grant no 110837 and 70114178, coordinator: Rita K. Schmutzler, Cologne) and the Federal Ministry of Education and Research, Germany (grant no 01GY1901). This work was also funded by the European Regional Development Fund and Free State of Saxony, Germany (LIFE - Leipzig Research Centre for Civilization Diseases, project numbers 713–241202, 713–241202, 14505/2470, 14575/2470). The GENICA was funded by the Federal Ministry of Education and Research (BMBF) Germany grants 01KW9975/5, 01KW9976/8, 01KW9977/0 and 01KW0114, the Robert Bosch Foundation, Stuttgart, Deutsches Krebsforschungszentrum (DKFZ), Heidelberg, the Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr University Bochum (IPA), Bochum, as well as the Department of Internal Medicine, Johanniter GmbH Bonn, Johanniter Krankenhaus, Bonn, Germany. Generation Scotland (GENSCOT) received core support from the Chief Scientist Office of the Scottish Government Health Directorates [CZD/16/6] and the Scottish Funding Council [HR03006]. The GESBC was supported by the Deutsche Krebshilfe e. V. [70492] and the German Cancer Research Center (DKFZ). The HABCS study was supported by German Research Foundation (DFG Do761/15-1), the Claudia von Schilling Foundation for Breast Cancer Research, by the Lower Saxonian Cancer Society, and by the Rudolf Bartling Foundation. The HEBON study is supported by the Dutch Cancer Society grants NKI1998-1854, NKI2004-3088, NKI2007-3756, NKI 12535, the Netherlands Organisation of Scientific Research grant NWO 91109024, the Pink Ribbon grants 110005 and 2014-187.WO76, the BBMRI grant NWO 184.021.007/CP46, and the Transcan grant JTC 2012 Cancer 12-054. The HMBCS was supported by the German Research Foundation (DFG Do761/15-1), a grant from the Friends of Hannover Medical School, and by the Rudolf Bartling Foundation. The HUBCS was supported by German Research Foundation (DFG Do761/15-1), a grant from the German Federal Ministry of Research and Education (RUS08/017). The study was performed as part of the assignment of the Ministry of Science and Higher Education of the Russian Federation (№1022040500074-9). Financial support for KARBAC was provided through the regional agreement on medical training and clinical research (ALF) betweenStockholm County Council and Karolinska Institutet, the Swedish Cancer Society, The Gustav V Jubilee foundation and Bert von Kantzows foundation. The KARMA study was supported by Märit and Hans Rausings Initiative Against Breast Cancer. The KBCP was financially supported by the special Government Funding (VTR) of Kuopio University Hospital grants, Cancer Fund of North Savo, the Finnish Cancer Organizations, and by the strategic funding of the University of Eastern Finland. kConFab is supported by a grant from the National Breast Cancer Foundation, and previously by the National Health and Medical Research Council (NHMRC), the Queensland Cancer Fund, the Cancer Councils of New South Wales, Victoria, Tasmania and South Australia, and the Cancer Foundation of Western Australia. Amanda B. Spurdle was supported by an NHMRC Investigator Fellowship (APP177524). Financial support for the AOCS was provided by the United States Army
Medical Research and Materiel Command [DAMD17-01-1-0729], Cancer Council Victoria, Queensland Cancer Fund, Cancer Council New South Wales, Cancer Council South Australia, The Cancer Foundation of Western Australia, Cancer Council Tasmania and the National Health and Medical Research Council of Australia (NHMRC; 400413, 400281, 199600). G.C.T. and P.W. are supported by the NHMRC. RB was a Cancer Institute NSW Clinical Research Fellow. The MARIE study was supported by the Deutsche Krebshilfe e.V. [70-2892-BR I, 106332, 108253, 108419, 110826, 110828], the Hamburg Cancer Society, the German Cancer Research Center (DKFZ) and the Federal Ministry of Education and Research (BMBF) Germany [01KH0402]. The MASTOS study was supported by “Cyprus Research Promotion Foundation” grants 0104/13 and 0104/17, and the Cyprus Institute of Neurology and Genetics. MBCSG is supported by grants from the Italian Association for Cancer Research (AIRC). TheMelbourne Collaborative Cohort Study (MCCS) cohort recruitment was funded by VicHealth and Cancer Council Victoria. The MCCS was further augmented by Australian National Health and Medical Research Council grants 209057, 396414 and 1074383 and by infrastructure provided by Cancer Council Victoria. Cases and their vital status were ascertained through the Victorian Cancer Registry. The NBCS has received funding from the K.G. Jebsen Centre for Breast Cancer Research; The Norwegian Health authorities 2014, 2018; the Norwegian Cancer Society, 2015, 2019, 2021: The genetic „Make up“ and metabolic profile of breast cancer patients; relation to clinical course and treatment response“ (V.N.K. 223313).The Ontario Familial Breast Cancer Registry (OFBCR) was supported by grant U01CA164920 from the USA
National Cancer Institute of the National Institutes of Health. The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centers in the Breast Cancer Family Registry (BCFR), nor does mention of trade names, commercial products, or organizations imply endorsement by the USA Government or the BCFR. The ORIGO study was supported by the Dutch Cancer Society (RUL 1997-1505) and the Biobanking and Biomolecular Resources Research Infrastructure (BBMRI-NL CP16). The PBCS was funded by Intramural Research Funds of the National Cancer Institute, Department of Health and Human Services, USA. Genotyping for PLCO was supported by the Intramural Research Program of the National Institutes of Health, NCI, Division of Cancer Epidemiology and Genetics. The PLCO is supported by the Intramural Research Program of the Division of Cancer Epidemiology and Genetics and supported by contracts from the Division of Cancer Prevention, National Cancer Institute, National Institutes of Health. The RBCS was funded by the Dutch Cancer Society (DDHK 2004–3124, DDHK 2009–4318). The SASBAC study was supported by funding from the Agency for Science, Technology and Research of Singapore (A∗STAR), the US National Institute of Health (NIH) and the Susan G. Komen Breast Cancer Foundation. SEARCH is funded by Cancer Research UK [C490/A10124, C490/A16561] and supported by the UK
National Institute for Health Research Biomedical Research Centre at the University of Cambridge. The University of Cambridge has received salary support for PDPP from the NHS in theEast of England through the Clinical Academic Reserve. SKKDKFZS is supported by the DKFZ. The SZBCS was supported by Grant PBZ_KBN_122/P05/2004 and the program of the Minister of Science and Higher Education under the name "Regional Initiative of Excellence" in 2019–2022 project number 002/RID/2018/19 amount of financing 12 000 000 PLN. UBCS was supported by funding from National Cancer Institute (NCI) grant R01 CA163353 (to N.J. Camp) and the Women's Cancer Center at Huntsman Cancer Institute (HCI). Data collection for UBCS was supported by the Utah Population Database (UPDB) and Utah Cancer Registry (UCR). The UPDB is supported by HCI, the University of Utah, and NCI grant P30 CA42014. The UCR is additionally funded by the NCI's SEER Program, HHSN261201800016I, and the US Center for Disease Control and Prevention's National Program of Cancer Registries (NU58DP007131).
CARRIERS was supported by National Institutes of Health grants R01CA192393, R01CA225662 and R35CA253187; an NIH Specialized Program of Research Excellence (SPORE) in Breast Cancer P50CA116201 to Mayo Clinic; and the Breast Cancer Research Foundation. CARRIERS members were supported by the Basser Center for BRCA; Susan G. Komen Foundation; the Paul Calabresi Program in Clinical/Translational Research at Mayo Clinic (2K12CA090628-21); and the Penn Center for Global Genomics and Health Equity (M.P.A). BWHS was supported by U01CA164974, R01CA098663 and Susan G. Komen Foundation SAC180086; MEC was supported by UM1 CA164973; NHS was supported by P01 CA87969, UM1 CA186107 and U19 CA148065; NHS2 was supported by UM1 CA176726 and U19 CA148065; SISTER and Two SISTER were supported by Z01-ES044005, Z01-ES102245, Z01-ES049033 and Susan G. Komen Foundation FAS703856 to Two SISTER; WWHS was supported by P30CA014520, University of Wisconsin-Madison Office of the Vice Chancellor for Research and Graduate Education; The American Cancer Society funds the creation, maintenance, and updating of the Cancer Prevention Study-II cohort and Cancer Prevention Study-3. The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through 75N92021D00001, 75N92021D00002, 75N92021D00003, 75N92021D00004, 75N92021D00005.
The work of CZECANCA has been supported by the Ministry of Health of the Czech Republic: DRO-VFN 64165; Charles University: COOPERATIO; and the Ministry of Education Youth and Sports of the Czech Republic: MULTIOMICS_CZ Project CZ.02.01.01/00/23_020/0008540 - Co-funded by the European Union.
Financial support for GENESIS, including sequencing of the FANCM gene, was provided by Ligue Nationale contre le Cancer (grants PRE05/DSL, PRE07/DSL, PRE11/NA), INCa (grant No b2008-029/LL-LC), the comprehensive cancer center SiRIC (Site de Recherche Intégrée sur le Cancer, grant INCa-DGOS-4654), France Génomique Nationale infrastructure, funded as part of the « Investissements d’Avenir » program managed by the Agence Nationale pour la Recherche (ANR-10-INBS-09), the Centre National de Recherche en Génomique Humaine (CNRGH), CEA, the Ligue Comité deParis (grant RS15/75-78/FL) and the Fondation ARC pour la recherche sur le cancer (grant PJA 20151203365/FL).
IdISSC contribution was supported by Instituto de Salud Carlos III grant PI24/00267 co-funded by European Union (ERDF/ESF, “A way to make Europe”/”Investing in your future”) to Miguel de la Hoya.

Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:Matthias W. Beckmann conducts research funded by Amgen, Novartis and Pfizer. Peter A. Fasching conducts research funded by Amgen, Novartis and Pfizer. He received Honoraria from Roche, Novartis and Pfizer.