Germline pathogenic variants in metaplastic breast cancer patients and the emerging role of the BRCA1 gene

Metaplastic breast cancer (MpBC) is a rare, aggressive breast cancer (BC) histotype. Scarce information is available about MpBC genetic predisposition. Previous studies, mainly consisting of case reports, retrospective reviews and others on target therapies, pointed to a possible involvement of the BRCA1 gene in increasing MpBC risk, without ever confirming it. In this study, we retrospectively reviewed all BC patients counseled at our Institute for genetic testing of at least BRCA1 or BRCA2 (BRCA) genes and we found that 23 (23/5226 = 0.4%) were affected by MpBC. About 65% (15/23) of MpBC patients harbored a germline pathogenic variant (PV): 13 in BRCA1 (86.7%), including two patients who received genetic testing for known familial PV, one in TP53 (6.7%), and one in MLH1 (6.7%). We observed a statistically different frequency of MpBC in patients who carried a PV in the BRCA genes (13/1114 = 1.2%) vs. all other BC patients (10/4112 = 0.2%) (p = 0.0002). BRCA carriers proved to have an increased risk of developing MpBC compared to all other BC patients who were tested for BRCA genes (OR = 4.47; 95% CI: 1.95–10.23). Notably, MpBCs were diagnosed in 2.1% (13/610) of BRCA1 carriers. No MpBCs were observed in BRCA2 carriers (0/498 = 0%), revealing a statistically significant difference between the prevalence of MpBCs in BRCA1 and BRCA2 carriers (p = 0.0015). Our results confirmed that BRCA1 is involved in MpBC predisposition. Further studies on unselected patients are needed to elucidate the authentic role of BRCA1 and to explore the possible implication of other genes in MpBC predisposition.


INTRODUCTION
Metaplastic breast cancer (MpBC) is an aggressive malignancy characterized by the presence of two or more cell types, most commonly an admixture of epithelial and mesenchymal elements [1].MpBC is a rare condition, accounting for about 0.2-5% of all breast cancers (BCs) [2].However, it carries the worst prognosis in comparison to other BC types and plays a significant role in global BC mortality [3].
Although MpBCs may include all the different clinical BC subtypes, most of them are triple-negative breast cancers (TNBCs), meaning that they are characterized by ≤1% nuclear expression of hormone receptors and HER2 negative status, with or without gene amplification [4,5].Similar to TNBC, MpBC is typically high-grade but has a more aggressive behavior, showing great propensity for recurrence and specific chemoresistance, in particular in neoadjuvant settings [6,7].
It is well-known that the TNBC phenotype is closely associated with a hereditary cause of the disease [8], most frequently the Hereditary Breast and Ovarian Cancer (HBOC) syndrome.HBOC is an autosomal dominant condition caused by germline pathogenic variants (PVs) in BRCA1 and BRCA2 genes.In addition, other Homologous Recombination DNA repair genes, such as PALB2, RAD51D, RAD51D, and BARD1, have been associated with TNBC [9].The latest National Comprehensive Cancer Network (NCCN) clinical practice guidelines recommend at least BRCA (BRCA1 and BRCA2) genetic testing for all TNBC patients [10], although until 2021, it was proposed only for TNBC cases ≤60 years, as to date still suggested by several national recommendations.More specifically, BRCA1 PV carriers are more likely to develop TNBC than BRCA2 carriers [11].
Besides HBOC, early-onset BC is a well-known phenotype in Li Fraumeni syndrome (LFS), a rare hereditary disorder due to germline PVs in TP53 gene and responsible for epithelial and mesenchymal tumors.However, germline mutations in TP53 are not typically associated with TNBC [12].Accordingly, genetic testing of TP53 gene is recommended in all patients with BC diagnosed under 31 years, regardless of family history (FH) [13].
Notably, only a few studies have explored the association specifically between MpBC and germline PVs in cancer predisposition genes.Due to MpBC rarity, histological diversity, and aggressive nature, scarce information is available about genetic predisposition to MpBC.Here, we reported our series of patients affected by MpBC who underwent genetic testing for at least BRCA genes to evaluate the prevalence of germline PVs and assess their possible role in increasing MpBC risk, also by comparison with the literature.

Study population
The study included 5226 consecutive BC patients who were referred to genetic counseling at the Division of Cancer Prevention and Genetics of the European Institute of Oncology (IEO) between 2002 and 2022.All the patients received germline genetic testing of at least BRCA1 or BRCA2 genes following the clinical criteria available at the time and according to their personal/family history.Specifically, 4973 (95.2%)BC patients underwent germline testing for unknown mutations (screening of gene exons and splice junctions) in at least BRCA1 and BRCA2 genes, while 253 (4.8%) for a known familial PV in BRCA1 (n = 128) or BRCA2 (n = 119) or both (n = 6).Additionally, 547 and 272 BC patients also underwent germline testing for unknown mutations in TP53 and MLH1, respectively.
We then retrospectively searched for patients affected by MpBC (n = 23) and, among these, for those who were found to be PV carriers (n = 15).Histological diagnosis had been performed or reviewed at IEO by pathologists with extensive experience in breast cancer following the current World Health Organization (WHO) classification system [14].Personal and family history, clinical, histopathological and genetic data were collected and stored in a dedicated institutional database.
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of the European Institute of Oncology (UID 0833, date of approval 23/05/2018).Informed consent was obtained from all patients involved in the study.

Genetic analysis in MpBC patients
Germline genetic analysis on the 23 patients with MpBC was carried out using the molecular methods available at the time of testing, i.e., Denaturing High Performance Liquid Chromatography (DHPLC) and Sanger sequencing and/or Multiplex Ligation Dependent Probe Amplification (MLPA) and/or Next-Generation Sequencing (NGS).Details on molecular methods, genes analyzed and selection criteria for each MpBC patient are indicated in Table S1.

Statistical analyses
Chi-square or Fisher's Exact tests were used to compare the relative frequencies of MpBC in BRCA PV carriers (BRCA-positive) vs. all other BC patients (non BRCA-positive) or among PV carriers in BRCA1, BRCA2 or both genes.
Multivariable logistic regression model was used to calculate the odds ratio (OR) with 95% confidence interval (CI) indicating the association of MpBC with BRCA PV carriers adjusting for age at first BC diagnosis.Two-sided p < 0.05 was considered statistically significant.Statistical analyses were performed using the Statistical Analysis System software, Version 9.2 (SAS Institute, Cary, CA, USA).

Review of the literature
We then compared our series of MpBC PV carriers with those reported in the literature.We performed a PubMed search (up to Dec 31, 2022) of the following keywords in the title/abstract of the articles: [germline variant OR germline mutation] AND [metaplastic breast cancer].Additional papers were identified by a manual search of references from original articles and reviews.Studies not reporting molecular details of the germline PVs identified in MpBCs and/or FH of breast/ovarian cancers and/or not specifying the selection criteria used to address MpBC patients to genetic testing were excluded from this review.

RESULTS
Genetic testing results are outlined in Fig. 1.Out of 5226 BC patients who underwent genetic testing for at least BRCA1 or BRCA2 genes, 23 (0.4%) were diagnosed with a MpBC.Twenty-one (21/23 = 91.3%)patients underwent germline testing for unknown mutations in BRCA1 and BRCA2.Out of these 21 MpBC patients, five underwent genetic testing also for TP53 and four also for MLH1 gene.Two (2/23 = 8.7%) MpBC patients were addressed to genetic testing for a known familial PV in BRCA1.Overall, we identified 15 (15/23: 65.2%) MpBC carriers of a germline PV: 13 in BRCA1 (86.7%), including the two MpBC patients who received genetic testing for the known familial mutation, one in TP53 (6.7%) and one in MLH1 (6.7%).More specifically, thirteen out of 21 (13/ 21 = 61.9%)MpBC patients who underwent germline testing for unknown mutations were found to be carriers of a PV: 11 in BRCA1, 1 in TP53, and 1 in MLH1.

BRCA1
Additional genetic testing for: Fig. 1 Flowchart of the study showing the number of patients and genetic testing results.
have an increased risk of developing MpBC compared to all other BC patients who were tested for BRCA1 and BRCA2 genes (OR = 4.47; 95% CI: 1.95-10.23),adjusting for age at first BC diagnosis.We also investigated whether there was any trend in year of cancer diagnosis, but we did not find any indication of differences with time potentially influencing the association between MpBC and BRCA status.Notably, MpBC was diagnosed in 2.1% (13/610) of BRCA1 carriers.No MpBCs were observed in BRCA2 carriers (0/498 = 0%), revealing a statistically significant difference between the prevalence of MpBCs in BRCA1 and BRCA2 carriers (p = 0.0015) (Table 2).Table 3 shows clinical and molecular characteristics of the 15 MpBC patients found to be carriers of a PV.Germline PVs were defined as follows: seven small deletions, two small insertions, two large rearrangements, three nonsense, and one missense.All PVs in BRCA1 and TP53 genes were truncating (i.e., predicted to result in a truncated protein product), while the one in MLH1 gene was a missense substitution.
Median age at MpBC diagnosis was 40 years (range 27-58).MpBCs were histologically heterogeneous, with the majority showing mesenchymal differentiation (47%).Notably, the TP53 PV carrier developed MpBC with sarcomatoid elements.All patients were diagnosed with a TNBC, except for two cases of HER2 overexpressed BC.
Nine out of 15 patients (60%) were also diagnosed with other tumors, before or after MpBC: seven BRCA1 PV carriers and the TP53 PV carrier developed other BCs of ductal histology (either infiltrating or in situ), while the MLH1 PV carrier was also affected by colorectal cancer.
Eight patients showed family history of BC and/or OC (53%), while the remaining seven patients (47%) did not report any relative affected by BC or OC.Pedigrees of the 15 MpBC PV carriers of our series are shown in Fig. S1.
We also performed a detailed literature search to identify previous papers about germline PVs in MpBC, describing also FH of BC/OC and selection criteria used for genetic testing.Eight case reports have been published so far on MpBC patients harboring germline PVs [17][18][19][20][21][22][23][24].All of them were found to carry a C5 variant in BRCA1 gene; notably, one was also carrier of a C4 splicing variant in TP53 gene (Table 4).
The nucleotide changes in BRCA1 gene included five small deletions or insertions leading to frameshift, one nonsense mutation, one large deletion and one missense substitution.Except for this last variant, all PVs were expected to result in a premature termination of the protein.
Median age at MpBC diagnosis of PV carriers was 30 years (range 15-49).All BCs were of TN subtype; however, the histological subtype of MpBC varied greatly among patients.
Five out of eight cases (62.5%) were affected by other BCs in addition to the MpBC: all of them were invasive ductal carcinomas, except for the patient described by Breuer et al. [19], who was diagnosed with a metachronous bilateral MpBC with squamous differentiation.
A positive family history of BC and/or OC was described in five out of eight (62.5%) of the screened MpBC cases.

DISCUSSION
To the best of our knowledge, we report here the largest series of MpBC patients harboring a germline PV and fully described in terms of clinical and molecular characteristics, as well as family history of BC/OC.Similarly to the frequency previously described in the literature, we observed that about 0.4% of BC patients developed a MpBC.In our series, 65.2% of MpBC patients were carriers of a PV in a cancer predisposition gene.More specifically, 86.7% of MpBC carriers harbored a PV in the BRCA1 gene.
Notably, previous studies pointed to a possible involvement of the BRCA1 gene in MpBC genetic predisposition.However, only eight single case reports have been published so far with detailed information.These eight MpBC cases carried PVs almost exclusively in the BRCA1 gene, except for the unusual case of a woman affected by HBOC and LFS [24].An additional case of a 54year-old woman affected by TN MpBC with osseous differentiation has been reported [25]: a genetic alteration in BRCA1 was identified in the tumor (c.1530del, p.Gly511AlafsTer21), but the patient refused to undergo germline testing, and she did not have any family history of malignancies.
Furthermore, a recent paper on neoadjuvant chemotherapy in MpBC by Wong et al. [26] reported that two out of 31 (6%) MpBC patients who underwent genetic testing were found to carry a BRCA1 germline mutation, showing a lower frequency of BRCA1 carriers compared to our study.However, this discrepancy could be due to different selection criteria used for genetic testing.
To date, only one study compared the frequency of MpBC between BC carriers of a PV in the BRCA1 or BRCA2 genes vs. all other BC patients [38].By this comparison, the authors found a statistically significant difference in frequencies of histological types between BRCA1 carriers vs. non-carriers.In particular, they reported 3.2% MpBCs in 93 BRCA-positive BCs vs. 0.8% MpBCs in 3157 non BRCA-positive BCs.However, they did not observe any statistically significant difference between BRCA1 and BRCA2 carriers.Similarly to Rodríguez-Fernández et al. [38], we found a statistically significant higher frequency of MpBC in BC patients who turned out to be BRCA carriers than in all other BC patients.Moreover, we also observed a statistically significant difference   between PV carriers in BRCA1 and BRCA2 genes.Indeed, we detected an increased risk for BRCA carriers to develop a MpBC when we compared them to non BRCA-positive BC patients.Of note, these results could be ascribed only to BRCA1 carriers as we did not observe MpBC patients carrying a PV in BRCA2 in our series.On the whole, our results showed that the BRCA1 gene, longly regarded as a possible predisposition MpBC gene, has a crucial role in this condition.Additionally, we also identified a TP53 PV carrier without any alteration in the BRCA1 gene (Fig. S1, patient ID 14).Interestingly, no FH suggestive for LFS syndrome was observed, even if family was relatively small.
We also describe for the first time a germline C4 variant of MLH1 gene in a patient who developed both CRC at 51 and MpBC at 58 years of age (Fig. S1, patient ID 15).Constitutive mutations of mismatch repair genes (including MLH1, MSH2, MSH6 and PMS2) are known to cause the Lynch syndrome (LS), an inherited condition characterized by an increased risk of developing many tumor types, mainly colorectal and endometrial cancers [39,40].However, whether BC belongs to the LS spectrum is a long-standing question [41,42].Investigating MLH1 protein expression in a tumor sample would have been useful to confirm or exclude a possible role for this germline variant in MpBC.This patient underwent surgery in another Institution, therefore we could not perform immunohistochemistry for MLH1 protein on BC tissue slides.On the whole, without a clear evidence of MMR deficiency in the tumor, it remains unknown whether this MLH1 variant could be implicated in MpBC predisposition.Moreover, the patient also harbored other two genetic variants (class C3 according to the ACMG classification): the c.5986 G > A (p.Ala1996Thr) variant in BRCA2 gene and the c.3313 G > A (p.Gly1105Arg) in MSH6 gene.We cannot exclude that the missense substitution in BRCA2 could act as a low penetrance variant for BC risk nor that additional genetic/environmental factors could modulate BC risk in this family.
Almost all PVs detected in MpBC patients were expected to result in a truncated protein product, except for the c.181 T > G C5 missense variant of BRCA1 and the c.375+2 T > C C4 splicing variant of TP53 reported in the literature, and the c.244A > G C4 missense variant of MLH1 gene described in our series.The most common mutation types were small deletions/insertions leading to frameshift, followed by nonsense.
The identified germline PVs were different compared with those already reported, except for the BRCA1 c.5266dup variant, described both in our series and in a previous report [23].In addition, we identified the c.5030_5033del BRCA1 variant in two different MpBC patients.These variants are among the most frequent mutations detected in BRCA1 gene, in Europe or worldwide [43].
At any rate, no hotspot mutations specific for MpBC seem to be identifiable in BRCA1 gene.Several PVs clustered in exon 11 of BRCA1, just because it is the largest one of the gene (3426 base pairs) (Fig. 2).On the whole, the germline variants found in MpBC patients fall within the three regions most frequently mutated in cancer patients.
As regards histological classification, MpBCs of PV carriers reported in the literature were extremely heterogeneous, displaying epithelial, mesenchymal and mixed features.Our series showed the same variability: although the most prevalent phenotype was MpBC with mesenchymal differentiation (47%), we did not observe a specific histological subtype in patients with BRCA1 PVs.Notably, the TP53 PV carrier developed MpBC with sarcomatoid elements.All the eight MpBCs described in previous case reports were of TN subtype.In our series, TN phenotype was the most frequent subtype as well (86.7%), even though 13.3% of MpBCs showed HER2 overexpression (one with a BRCA1 germline PV and the other with the MLH1 germline PV).In the eight case reports reviewed, FH of BRCA-positive MpBC cases was ascertained: three of them were apparently sporadic, three presented positive FH of BC, one of OC, and one of both BC and OC.Our case series reflects literature data, since about 50% of MpBC cases seem to be sporadic.However, it should be stressed that these are highly selected cases, addressed to genetic counseling and germline testing due to specific clinical criteria, which include but are not limited to FH (i.e., development of earlyonset BC and/or of TNBC ≤ 60 years and/or of bilateral BC).
Because of its retrospective nature, our study has some limitations.First, we reviewed patients with MpBC who received genetic counseling over a long period of time .Selection criteria for genetic testing have changed over time, as well as molecular methods and genes selected for germline analysis.Only recently, evidence emerged about the involvement in hereditary BC, particularly in TNBCs, of other genes (such as BARD1, PALB2, RAD51D, and, with a minor impact, BRIP1, RAD51C, and TP53) [44].Accordingly, although all patients have been tested for BRCA1 and BRCA2, only a few have been tested for additional genes (e.g., TP53).Moreover, only the most recent cases underwent multigene panel testing that included several cancer-related genes.Consequently, we conducted statistical analyses comparing BRCA carriers with all non BRCA-positive patients, possibly including PV carriers in other genes in this second group.Of note, we verified if a difference by year of cancer diagnosis could exist and represent a possible bias, however we did not find any significant difference nor influence on the association with histological type.
Lastly, even after adjusting our multivariable analysis for age at first BC diagnosis, the association with MpBC was confirmed.However, we may have missed to include some important confounding variables.Additional studies should confirm this result evaluating further sources of bias and confounding.
Analyses of larger series of cases are required to better characterize MpBC genetic predisposition.These studies, if conducted on unselected MpBC patients and through careful data collection (family history, histological classification, etc.), are expected to elucidate the authentic detection rate of PVs in the BRCA1 gene, as well as to explore the possible involvement of additional genes in increasing the risk of developing MpBC [45].Unraveling MpBC genetic predisposition and its specific molecular landscape is needed for the clinical management of affected patients and families and could pave the way for tailoring new therapeutic strategies.

CONCLUSIONS
Taken together, our findings and literature data point to the following conclusions: (1) in MpBC patients described so far, who are highly selected for germline testing of specific genes, BRCA1 seems to play a crucial role in increasing the risk of MpBC development; (2) our results confirmed that BRCA1 is involved in MpBC predisposition; (3) additional studies on larger, unselected series of patients are necessary to elucidate the authentic role of germline BRCA1 PVs in MpBCs and to explore the possible implication of other genes in MpBC predisposition; (4) almost all PVs detected in MpBC carriers were predicted to result in a truncated protein product; (5) no hotspot mutations specific for MpBC seem to be detectable in BRCA1 gene, since the identified PVs are frequently found in HBOC patients or affect the most commonly mutated regions of the protein in cancer patients; (6) BRCA1 PVs do not seem to be associated with a specific histological subtype of MpBC; (7) almost all MpBCs in PV germline carriers are TN, with a few exceptions of HER2 overexpressed cases.Alternate exons are shown as rectangles, with the corresponding number underneath.After exon 3, subsequent exon numbers are increased by one, due to historical mis-annotation of an additional "exon 4".Exon 11 is not to scale, since it covers >65% of BRCA1 sequence.PVs described in our series are indicated in bold; PVs reported in previous case reports are in normal type.a reported twice in our series.b reported both in our series and in ref. [23].The three regions of BRCA1 protein that are mutated in cancer patients with a higher frequency are indicated above the graph.
BC patients tested for at least BRCA1 or BRCA2 n

Fig. 2
Fig. 2 Distribution of unique germline BRCA1 PVs identified in MpBC patients (modified from BRCA Exchange https://brcaexchange.org/).Alternate exons are shown as rectangles, with the corresponding number underneath.After exon 3, subsequent exon numbers are increased by one, due to historical mis-annotation of an additional "exon 4".Exon 11 is not to scale, since it covers >65% of BRCA1 sequence.PVs described in our series are indicated in bold; PVs reported in previous case reports are in normal type.a reported twice in our series.b reported both in our series and in ref.[23].The three regions of BRCA1 protein that are mutated in cancer patients with a higher frequency are indicated above the graph.

Table 1 .
Comparison of MpBC frequency in BRCA-positive vs. all other a BRCA-positive: carriers of a PV in BRCA1 and/or BRCA2 genes.b P-value from Chi-square test.

Table 2 .
Comparison of MpBC frequency between carriers of PVs in BRCA1 or BRCA2 or both genes.
a P-value from Fisher's Exact test.

Table 3 .
Germline PVs, family history and clinical, histological characteristics of our series of MpBC PV carriers.
a cDNA and protein changes are named according to HGVS nomenclature.Reference sequences: b After exon 3, subsequent exon numbers of BRCA1 gene are increased by one, due to historical mis-annotation of an additional "exon 4".c Histopathological classification according to the current World Health Organization (WHO) classification system [14].d Family history of BC

Table 4 .
Germline PVs, family history, and clinical, histological characteristics of the MpBC PV carriers reported in the literature.
a cDNA and protein changes are named according to HGVS nomenclature.Reference sequences: b After exon 3, subsequent exon numbers of BRCA1 gene are increased by one, due to historical mis-annotation of an additional "exon 4".c Pathology-confirmed or self-reported cancer diagnoses.G. Corso et al.