Globally, breast cancer affects more than one in ten women and has become the most frequently diagnosed cancer worldwide. It is the most prevalent malignancy among women of all ethnicities, accounting for approximately 30% of all newly diagnosed cancers in women. In the United States, an estimated one in eight women (about 13%) will be diagnosed with breast cancer in her lifetime [1,2] and an estimated 42,170 U.S. women will die from breast cancer in 2025 [3].

Significant disparities exist between African American (AA) and European American (EA) women in both incidence and mortality rates [4,5]. Although EA women have higher age-adjusted incidence rates (143 vs. 119 per 100,000), AA women are more often diagnosed at younger ages and with more aggressive tumor subtypes, typically characterized by higher histological grade and negative estrogen/progesterone receptor status. Notably, AA women have a 41% higher mortality rate from breast cancer compared to EA women. Among women under 50, the incidence of aggressive subtypes such as triple-negative breast cancer (TNBC) is nearly twice as high in AA women [5]. Thus, while EA women of European descent are slightly more likely to be diagnosed with breast cancer, women of African descent are more likely to die from this disease.

Hereditary breast cancer is predominantly attributed to mutations in the BRCA1 and BRCA2 genes, which are associated with a markedly increased lifetime risk of breast and ovarian cancers. These mutations account for approximately 5% of all breast cancer cases and over 80% of hereditary cases [4,6,7]. The lifetime risk of breast cancer in BRCA1 and BRCA2 mutation carriers ranges from 35% to 85% respectively [8,9], with BRCA1 mutations conferring an estimated risk of 55%–72% and BRCA2 mutations 45%–69% [1,5]. Despite these well-established risks, most genetic studies on hereditary breast cancer have been conducted in non-Hispanic white populations, with limited representation of AA women [10]. Consequently, the mutation spectrum within AA populations remains underexplored.

Previous research suggests that the frequency and spectrum of germline mutations in BRCA1, BRCA2, and other susceptibility genes such as CHEK2 may differ substantially between AA and EA populations [10–12]. At the Howard University Cancer Center, we have assembled one of the largest cohorts of AA families at substantial risk for hereditary breast cancer. We hypothesize that AA patients possess a distinct and potentially broader spectrum of BRCA1 germline mutations compared to those observed in other populations.

To investigate this hypothesis, we employed denaturing high-performance liquid chromatography (DHPLC) and DNA sequencing to analyze BRCA1 mutations, including missense variants, polymorphisms, haplotypes, and large genomic rearrangements, in 99 AA patients from high-risk hereditary breast cancer families. We used Sorting Intolerant From Tolerant (SIFT) analysis to predict the functional consequences of detected amino acid substitutions. In addition, we assessed potential splicing effects by examining disruptions of conserved exonic splicing enhancer (ESE) motifs and canonical splice site sequences.

Human subjects

Ninety-nine AA patients from families at substantial risk for hereditary breast cancer with histologically confirmed breast cancer were recruited at Howard University Hospital in Washington, D.C. The study was approved by the Howard University Institutional Review Board (IRB-09-MED-86, IRB-04-MED-04, IRB-97-MED-35), and written informed consent was obtained from all participants. Inclusion criteria for the high-risk cohort were based on one or more of the following: 1) two or more first- or second-degree relatives with breast cancer, or with both breast and ovarian cancer; 2) diagnosis of early-onset breast cancer (≤40 years); 3) bilateral breast cancer; 4) diagnosis of both breast and ovarian cancer in the same individual; 5) or a family history of male breast cancer. Among the 99 high-risk patients: 71% reported a family history of multiple breast cancer cases; 14% had families with both breast and ovarian cancer; 8% were diagnosed with early-onset breast cancer (≤40 years); 3% had bilateral breast cancer; 2% had a family history of male breast cancer; and 1% had bilateral breast cancer without any reported family history (Table 1).

A control group of 122 disease-free individuals, with no personal or family history of breast or ovarian cancer, was also recruited through the free screening mammography program at the Howard University Cancer Center (HUCC). Controls were frequency-matched to cases by age.

DNA extraction and polymerase chain reaction (PCR)

To identify genomic rearrangements and germline mutations in the coding regions (exons), intron-flanking sequences, promoter, and untranslated regions of the BRCA1 gene, genomic DNA was extracted from cryopreserved peripheral blood lymphocytes of both cases and controls using the QIAamp DNA Blood Maxi Kit (Qiagen Inc., Valencia, CA) following the manufacturer’s protocol.

PCR amplification of BRCA1 target regions was performed using sequence-specific primers as recommended by Teresa Wagner et al. [13,14]. Reactions were conducted in an AmpGene 9700 thermal cycler (Perkin-Elmer 600, Foster City, CA).

PCR reactions were performed in a total volume of 50 µL containing 50 ng genomic DNA, 15 mM Tris-Cl (pH 8.0), 50 mM KCl, 1.5 mM MgCl?, 0.2 mM of each deoxynucleotide triphosphate (dNTP), 0.4 µM of each forward and reverse primer, 1 unit of AmpliTaq Gold® DNA polymerase (Applied Biosystems, Foster City, CA), and 1X Gold PCR buffer. The thermal cycling conditions were followed according to the protocol described by Wagner et al. [13,14].

Denaturing high-performance liquid chromatography (DHPLC) and mutation analysis

In a blinded comparison, DHPLC has been reported as one of the most reliable methods for detecting BRCA1 mutations, ranking second only to DNA sequencing [13]. DHPLC was performed using the WAVE® DNA Fragment Analysis System (Transgenomic Inc., Omaha, NE) equipped with a DNASep® column (Transgenomic Inc., San Jose, CA), which contains C18-alkylated, polystyrene-divinylbenzene beads optimized for nucleotide fragment separation.

Optimal temperatures for resolving heteroduplex and homoduplex DNA were determined using the DHPLC melting algorithm [13]. Gradient parameters were set based on PCR product size using WAVEMaker™ software (v1.5.4). DHPLC was conducted under preset conditions with an initial gradient of 45% buffer A (0.1 M TEAA, pH 7.0) and 55% buffer B (0.1 M TEAA + 25% acetonitrile, pH 7.0), followed by a final gradient of 36% buffer A and 64% buffer B. PCR products (25 µL) were injected and eluted at 0.9 mL/min with a total acquisition time of 8.7 minutes.

All putative mutations and sequence variants identified by DHPLC were confirmed by direct DNA sequencing.

DNA sequencing and mutation analysis

Samples exhibiting heteroduplex peaks on DHPLC were further analyzed. PCR products were purified using the QIAquick PCR Purification Kit (Qiagen) and sequenced by ACGT Inc. (Wheeling, IL). Sequence data were analyzed with Sequencher™ v3.1 (Gene Codes, Ann Arbor, MI). Nucleotide positions were assigned according to the BRCA1 reference sequence (GenBank U14680) based on cDNA numbering and compared to the published BRCA1 mRNA sequence (GenBank HSU43746).

SIFT analysis

SIFT is a computational tool used to predict whether an amino acid substitution affects protein function. The BRCA1 protein sequence (GenBank accession no. HSU43746) was submitted to the SIFT server (http://blocks.fhcrc.org/~pauline/SIFT.html). Substitutions with normalized probabilities below 0.05 were classified as intolerant (potentially damaging), while those with probabilities ≥0.05 were considered tolerated, following established thresholds.

BRCA1 ESE analysis

To assess whether BRCA1 variants disrupted ESE motifs, wild-type and mutant sequences were analyzed using ESEfinder (http://rulai.cshl.edu/tools/ESE/). Four position weight matrices (PWMs) for SR proteins were selected for analysis. Scores exceeding the program’s predefined thresholds were considered high and the position of each predicted ESE within exons was also evaluated to estimate its potential functional impact on splicing.

Multiplex Ligation-dependent Probe Amplification (MLPA)

MLPA [15] was performed to detect BRCA1 copy number variants. Twenty-six patients were screened using the BRCA1-MLPA kit P002 (MRC-Holland, Amsterdam, Netherlands) which includes FAM-labeled primers for all BRCA1 exons and nine control probes.

Amplified, fluorescently labeled PCR products were separated by size using an ABI 3100 capillary sequencer (Applied Biosystems), with ROX-500 as the internal size standard. Data analysis was performed with GeneScan and Genotyper software (Applied Biosystems). Peak heights were normalized and compared to controls to calculate dosage quotients (DQs): approximately 1.0 indicates two copies of the target sequence (2n), ~0.5 indicated deletions (n), while ~1.5 indicate duplications (3n).

Preliminary MLPA screening of 40 patient samples was conducted and suspected rearrangements were confirmed by sequencing breakpoints.

Amplification Refractory Mutation System (ARMS)

ARMS, a sensitive PCR-based assay, was employed to detect BRCA1 mutations. This method effectively differentiates heterozygous individuals from homozygotes at target loci. Samples exhibiting multiple DHPLC peaks, such as patient BC70, and representative samples harboring the BRCA1 β-promoter 1802 C/C, C/G, or G/G alleles were analyzed. Primers were designed at a sufficient distance from the polymorphic site to prevent interference [16].

BRCA1 pathogenic mutations

DHPLC analysis, followed by DNA sequencing, identified a deleterious protein-truncating BRCA1 mutation, 5296delGAAA, in two unrelated patients (BC69 and BC88) (Table 2). Patient BC69 was diagnosed with bilateral breast cancer at ages 32 and 48 (ER/PR status unknown) and had a family history including 12 cases of breast cancer and 3 cases of ovarian cancer. Patient BC88 was diagnosed at age 30 with ER-negative/PR-negative breast cancer and had a family history of 3 breast cancer cases, 1 ovarian cancer case, and 2 prostate cancer cases.

A second pathogenic BRCA1 mutation, IVS16 + 6T→C, was detected in two unrelated patients (BC86 and BC90) with a family history of multiple breast cancer cases [17]. Patient BC86 was diagnosed with bilateral breast cancer at ages 35 and 48 (ER/PR status unknown) and had a family history of 2 breast cancer cases. Patient BC90 was diagnosed at age 27 with ER-negative/PR-negative breast cancer and had a family history of 4 breast cancer cases.

BRCA1 missense variations and ESE Sequences
Three missense variations, Ile379Met, Glu1210Lys, and Glu1794Asp, were detected in patients with multiple cases of breast cancer and are predicted to affect BRCA1 function (Table 3). The Ile379Met variation, previously reported in both African American and Caucasian populations, was found in patient BC48 who was diagnosed with bilateral ER-negative/PR-negative, p53-positive breast cancer at ages 41 and 46; and her mother was diagnosed with breast cancer at age 49.

The Glu1210Lys variation, not previously reported in the ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/) or The BRCA Exchange (https://brcaexchange.org), is predicted by SIFT analysis to impact protein function. The patient carrying this variant was diagnosed at age 79 with ER-positive/PR-positive breast cancer and has two affected sisters and a brother with prostate cancer.

Glu1794Asp, detected in patient BC40 diagnosed at age 42, is also predicted by SIFT to affect BRCA1 function. Additionally, two BRCA1 exon 22 missense variants, Gln1785His and Glu1794Asp, are predicted by ESE analysis (http://exon.cshl.edu/ESE/) to disrupt binding of serine/arginine-rich (SR) proteins, potentially affecting splicing. The Gln1785His variant, found in BC01 diagnosed at age 33, is similarly predicted by SIFT to impact protein function.

The rare BRCA1 884G>A (Glu255Glu), located within a known alternative exon 11 splicing enhancer (ESE) was identified in patient BC25. The wild-type alternative splice site typically results in the deletion of most of exon 11. This variant is predicted to abolish the alternative splicing of BRCA1, which may have functional consequences due to its tissue specificity and conservation. Patient BC25 was diagnosed at age 56 with ER-positive/PR-positive breast cancer and had a family history of three breast cancer cases. This variant has not been reported in ClinVar or The BRCA Exchange.

BRCA1 polymorphisms and haplotypes

Numerous coding and non-coding polymorphisms, along with unclassified variations, were detected in the BRCA1 gene through DHPLC analysis. Variants were further evaluated using SIFT analysis, evolutionary conservation, ESE scores, and their proximity to introns. Several polymorphisms identified have not been previously documented.

Haplotype analysis was conducted on the 10 most frequent polymorphisms identified. Table 4 presents the BRCA1 haplotypes of common polymorphisms (frequency ≥ 5% of chromosomes) in AA patients. Haplotype I and II frequencies were 0.60 and 0.041, respectively, representing 64% of AA patients in this cohort. Haplotype I includes the more frequent allele at exons 11K, 11N, 11S, 13, and 16; whereas haplotype II is the least frequent allele at these loci.

BRCA1 genomic rearrangements

Twenty-six patients heterozygous for the BRCA1 exon 11 Pro871Leu polymorphism were analyzed using MLPA to assess the presence of large deletions in BRCA1 exons 1a, 1b, and 2. Among these, 16 cases (62%) exhibited linkage disequilibrium, being heterozygous for both Pro871Leu and the BRCA1 β-promoter C/G1802 polymorphism. Five cases (19%) were homozygous for the G allele (G/G), while the remaining five cases (19%), including patient BC70, were homozygous for the C allele (C/C) (Figure 1, Table 5).

Heterozygosity at the exon 11 polymorphism paired with apparent homozygosity at the promoter polymorphism suggests a deletion encompassing the promoter region (Figure 1). Among the patients sequenced for both Pro871Leu (C/T) and the β-promoter 1802 polymorphisms, 10 were in linkage disequilibrium, whereas 5 (33%) were not. The observed hemizygosity at the BRCA1 β-promoter C/G1802 polymorphism supports the presence of a large deletion encompassing exons 1a, 1b, and 2, the first such report in an AA family. DNA sequencing of a subset of samples that displayed multiple DHPLC peaks also confirmed the Pro871Leu polymorphism.

Preliminary MLPA testing was performed on 40 patient samples and two cancer-free subjects. Figure 2 shows the 9 standards (blue) for copy number normalization and the BRCA1 exons (green) from a disease-free control lying in the normal dosage quotient range (DQ, 0.85–1.15 = normal). Figure 3 illustrates the 9 standards (blue) used for copy number normalization alongside the BRCA1 exons (green) for breast cancer patient BC12. Comparison to these standards, indicated by a dosage quotient (DQ) between 1.35 and 1.65, revealed a duplication of BRCA1 exon 17 in patient BC12.

We identified a deleterious BRCA1 mutation (5296delGAAA) in two unrelated patients with breast and ovarian cancer, and a pathogenic BRCA1 mutation (IVS16+6T→C) in two additional unrelated patients. All four probands, each from different families, were diagnosed before the age of 50 and belonged to families with a history of multiple cases of breast cancer and ovarian cancer, or both cancers in the same individual. Two of the BRCA1 mutation carriers had ER-negative/PR-negative tumors, while the hormone receptor status of the remaining two probands is unknown. Both early onset of disease and ER-negative/PR-negative tumor phenotype are characteristic features of BRCA1 mutation carriers [10].

The deleterious mutations (5296delGAAA and IVS16+6T→C) have been previously reported in African American, African, and other diverse populations and may represent founder mutations [18]. Two pathogenic BRCA1 mutations identified earlier in this cohort were found in women diagnosed with ER-negative/PR-negative breast cancer at ages 48 and 50 [11]. The IVS16+6T→C mutation has been shown to create a cryptic splice site, resulting in a truncated protein [19,20].

We also identified rare missense variants (Ile379Met, Glu1210Lys, and Glu1794Asp) that may impact BRCA1 function, as suggested by SIFT analysis, evolutionary conservation, early age of onset, and ER-negative/PR-negative tumor status [6]. Notably, the Ile379Met variant was found in a patient who exhibited all of these characteristics.

ESE sequences can be disrupted by single-base substitutions, including missense, nonsense, silent mutations, or single-nucleotide polymorphisms (SNPs) [21], leading to exon skipping in breast cancer patients with BRCA1 or BRCA2 mutations [20,21]. Recent studies have shown that some exonic substitutions can abolish ESE motifs.

The Ile379Met variant, previously reported in both AA and Caucasian populations, was identified in patient BC48, who was diagnosed with bilateral, ER-negative/PR-negative, p53-positive breast cancer at ages 41 and 46. Her mother was also diagnosed with breast cancer at age 49. Ile379Met variant was found in a critical domain, the BRCT-associated serine-rich domain [22] and the functional prediction using SIFT analysis [21,23], along with evolutionary comparison of human, ancestral eutherian, and marsupial protein sequences [23,24], suggests that the Ile379Met variant may impact BRCA1 function.

We also identified a unique silent variant (884G>A; Glu255Glu) within an alternative exon 11 ESE sequence that has not been reported in ClinVar or The BRCA Exchange (https://brcaexchange.org. Additionally, two missense variants in exon 22 of BRCA1 (Gln1785His and Glu1794Asp) are predicted to affect splicing. Two other patients carry rare variants located in ESE sites near intron-exon boundaries, which are expected to reduce splicing efficiency. However, transformed lymphocyte cell lines were not available to assess the biological impact of these putative missense and ESE variants.

Furthermore, numerous coding and non-coding polymorphisms were observed in BRCA1, several of which were previously undocumented. In many cases, the same combination of polymorphisms (haplotypes) appeared in different patients. Many patients shared specific combinations of polymorphisms (haplotypes), like those reported in U.S., European, and Asian populations, reflecting linkage disequilibrium [25–27]. For example, in European and Asian control populations, two major haplotypes, haplotype I (frequency: 0.65) and haplotype II (frequency: 0.33), account for approximately 98% of chromosomes [27]. In contrast, our AA cohort exhibited greater haplotype diversity than the European or Asian populations, accounting for 64% of the chromosomes, with haplotype I at a frequency of 0.60 and haplotype II at a frequency of 0.041. Haplotype I corresponds to the more frequent alleles at positions 11K, 11N, Ex13, and Ex16, while haplotype II carries the less frequent alleles at these same sites. Similarly, Dunning et al., reported that two major haplotypes (haplotype I at 0.59 and haplotype II at 0.34) accounted for 93% of chromosomes in Caucasian breast cancer cases [28]. In our AA patients, the same two haplotypes accounted for only 66% of the cases, with haplotype I at 0.60 and haplotype II at 0.06. Notably, the frequency of haplotype II in AA is 6–8 times lower than in Caucasians and Asians, further underscoring the increased haplotype diversity in this population. This greater diversity among AA is consistent with human evolutionary history. As modern humans originated in Africa, there has been more time for genetic variation to accumulate within African and African American populations. Furthermore, the AA population has ancestral roots in diverse regions across Africa and has experienced admixture with EA and Native Americans, contributing to the broader range of haplotype combinations observed [18].

We also identified a large deletion spanning BRCA1 exons 1a, 1b, and 2, extending into the promoter region in one of the 55 high-risk AA families. This deletion was detected in patients heterozygous at the BRCA1 exon 11 Pro871Leu polymorphism but apparently homozygous at the β-promoter C/G1802 polymorphism, consistent with hemizygosity from a large deletion. Linkage disequilibrium between these loci was present in some patients but absent in others, supporting the deletion hypothesis. No such deletion was observed in disease-free controls. This is the first report of such a deletion in an AA family.

Similar deletions have been reported in European populations, where homologous recombination between BRCA1 intron 2 and the highly similar pseudo-BRCA1 intron 2 generates an approximately 37 kb deletion encompassing the promoter and exons 1a, 1b, and 2, resulting in complete loss of BRCA1 transcript expression [29–34]. Our findings are consistent with this mechanism and confirm that the deletion extends into the promoter region, as suggested by the absence of linkage disequilibrium between BRCA1 exon 11 and the β-promoter polymorphisms in several patients.

Confirmation of the BRCA1 Pro871Leu polymorphism was obtained by sequencing samples with multiple DHPLC peaks. ARMS testing of patient BC70 and others confirmed that the deletion extended into the promoter region. In our cohort, AA displayed less linkage disequilibrium than Caucasians which is consistent with prior studies and the increased haplotype diversity observed. Patient BC70, carrying the deletion, was diagnosed with ER-positive/PR-negative breast cancer at age 63, and her sister was diagnosed at age 42.

ARMS testing confirmed deletion-associated alleles and DHPLC showed characteristic multiple peaks. Although MLPA confirmed the deletion of exons 1a, 1b, and 2, its exact extension into the promoter region requires further investigation.

This study reveals a rare BRCA1 genetic alteration, a large deletion encompassing exons 1a, 1b, and 2 and extending into the promoter in AA breast cancer patients. This finding suggests that large genomic rearrangements may contribute substantially to hereditary breast cancer risk in this population. Moreover, the diversity of rare missense and ESE-disrupting variants underscores the need for comprehensive genetic screening in AA. These findings should be validated in larger cohorts, and meta-analyses may further clarify the significance of these rare variants.

Not applicable.

This project was supported by grants DAMD17-01-1-0266 from the DOD Breast Cancer Research Program.

DB, LJR, and RLC conceived and designed the experiments. DB and YMK performed experiments, analyzed and interpreted the data. DB, YMK, and DAS wrote the manuscript.

MA, AHD, AL, OOK, RW, SN, and BK reviewed and corrected the manuscript. All author(s) reviewed and approved the final manuscript.

The study was approved by the Howard University Institutional Review Board (IRB-09-MED-86, IRB-04-MED-04, IRB-97-MED-35), and written informed consent was obtained from all participants.

Not applicable.

The authors declare no competing interests.