PCSK9 rs562556 variant and survival in BRCA1 mutation carriers and mutation-negative familial breast cancer

Abstract

Breast cancer etiology involves a complex interplay of various risk factors, with genetic alterations playing a significant role in susceptibility to the disease. While mutations in genes such as BRCA1 and BRCA2 are well established in elevating breast cancer risk, their potential role in predicting metastatic progression remains unclear. A recent study by Mei et al. suggested that a common germline variant in the PCSK9 gene (rs562556) may increase metastasis risk and lower survival rates in breast cancer patients. To investigate this association, we evaluated the relationship between the PCSK9 rs562556 variant genotypes and survival outcomes in two cohorts: patients with BRCA1 mutations and those with familial breast cancer lacking mutations in known susceptibility genes. Our study included 1052 BRCA1 mutation carriers, 467 familial breast cancer patients of unknown genetic etiology, and 1404 controls. Genotyping was performed using TaqMan genotyping and whole-exome sequencing. Survival analyses were conducted using Kaplan–Meier survival curves and Cox proportional hazards models. We found no significant association between PCSK9 rs562556 variant and breast cancer mortality or time to death following distant recurrence in either cohort. This finding does not support PCSK9 prognostic impact in breast cancer that may be modified by additional genetic or environmental factors.

Introduction

The genetic basis of breast cancer has been extensively explored over several decades, leading to the identification of both inherited and somatic mutations in several susceptibility genes. The burden of inherited breast cancer is due to high-risk pathogenic variants in genes such as BRCA1, BRCA2, and PALB2^1,2. In addition, several moderate- and low-penetrance genes contribute to cancer risk, albeit to a lesser extent³. Numerous single nucleotide polymorphisms (SNPs) have also been implicated in modulating risk. While each SNP individually confers only a small effect, their combined impact can be captured through a polygenic risk score (PRS), which may offer utility in personalized risk prediction⁴. Although these inherited genetic alterations are well-established determinants of breast cancer susceptibility, they have generally not been found to predict metastatic progression. Additionally, genome-wide association studies of patients with metastatic outcomes have tested numerous common SNPs, but none have reached genome-wide significance^5,6.

A study by Mei et al. identified a common germline variant in PCSK9 (NM_174936.4:c.1420G>A, p.Val474Ile; rs562556) as a potential driver of breast cancer metastasis⁷. The researchers reported that the AA genotype of rs562556 was associated with reduced survival across several independent breast-cancer patient cohorts and promoted metastatic progression in mouse models. Most strikingly, in a Swedish cohort of early-stage breast cancer patients, homozygosity for the A allele conferred a 22% risk of distant metastatic relapse at 15 years, compared with only 2% in non-homozygous individuals⁷. Mechanistically, PCSK9 influences metastasis by targeting low-density lipoprotein receptor-related protein 1 (LRP1) on tumor cells, leading to the suppression of metastasis-suppressing genes XAF1 and USP18. The study also reported that antibody-mediated inhibition of PCSK9 reduces breast cancer metastasis in multiple preclinical models⁷. These findings suggest that the common PCSK9 rs562556 variant may play a critical role in modulating metastatic potential and survival outcomes in breast cancer patients. Mei et al. further proposed PCSK9 inhibition as a potential strategy for metastasis prevention, while emphasizing the need for additional validation before implementation as a predictive biomarker or therapeutic target in breast cancer management⁷.

It is of clinical interest to determine whether the PCSK9 AA genotype is also associated with survival outcomes in women carrying a pathogenic variant in the BRCA1 gene. If such an association exists, it could inform genetic counseling and decision-making regarding preventive mastectomy. Those with the low-risk genotypes (AG or GG) may be more suited for annual surveillance with MRI, whereas those with the higher-risk AA genotype might consider prophylactic surgery. With this rationale, we sought to determine whether the PCSK9 rs562556 polymorphism is associated with survival in a cohort of Polish women with breast cancer carrying a pathogenic BRCA1 variant, as well as in a smaller cohort of Polish familial breast cancer cases without mutations in known susceptibility genes.

Results

PCSK9 c.1420G>A genotype distributions across the three cohorts are presented in Table 1. No significant differences in genotype frequencies were observed between either case group and the control group. Genotype distributions were in Hardy-Weinberg Equilibrium (HWE) in both the BRCA1-positive and control cohorts. However, the genotype data for the familial breast cancer cohort deviated from HWE, which may be attributed to the highly selective nature of this group. The odds ratio (OR) for developing breast cancer associated with the AA genotype of the PCSK9 c.1420G>A variant compared with the combined AG/GG genotypes was 1.17 for all cases versus controls (p = 0.05).

Table 1.

Comparison of the PCSK9 c.1420G>A genotype distributions among the breast cancer cases and healthy controls

Genotype	BRCA1-positive patients	Familial breast cancer patients	Controls	Univariable OR (95% CI, p-value) BRCA1 patients vs controls	Univariable OR (95% CI, p-value) familial cases vs controls	Univariable OR (95% CI, p-value) all cases vs controls
AA	759 (72.2%)	346 (74.1%)	976 (69.5%)	1.14 (0.95–1.35, 0.16)	1.25 (0.99–1.59, 0.06)	1.17 (1.0–1.37, 0.05)
AG	269 (25.6%)	103 (22.1%)	396 (28.2%)	NA	NA	NA
GG	24 (2.3%)	18 (3.9%)	32 (2.3%)	NA	NA	NA
AG/GG	293 (27.9%)	121 (25.9%)	428 (30.5%)	1.0 (Ref)	1.0 (Ref)	1.0 (Ref)
Total	1052	467	1404

The characteristics of patients with BRCA1 mutations are summarized in Table 2. Among the 1052 BRCA1-positive breast cancer patients, 404 had received neoadjuvant chemotherapy, and response data were available for 384 of these patients. In total, 173 patients (45.1%) achieved a pCR. The pCR rates were 45.5% for patients with the AA genotype, 44.0% for those with the AG genotype, and 42.9% for those with the GG genotype (Table 3).

Table 2.

Characteristics of the BRCA1-positive breast cancer patients

Characteristic	N = 1052
Age at diagnosis	44.7 (23.1–69.3)
Year of diagnosis	2010.6(1996.0–2022.8)
Follow-up duration	11.0 yrs (0.1–26.9)
Receptor status
ER-positive	298 (28.7%)
ER-negative	740 (71.3%)
Missing	14
Tumor size
0–2 cm	501 (50.9%)
2.1–5 cm	413 (41.9%)
>5 cm	71 (7.2%)
Mean size	2.7 cm (0.05–40)
Missing	67
Nodal status
Negative	714 (69.9%)
Positive	307 (30.1%)
Missing	31
HER2 status
Negative	786 (82.6%)
Positive	166 (17.4%)
Missing	100
Chemotherapy
Neoadjuvant	404 (38.4%)
Adjuvant	569 (54.1%)
None	77 (7.5%)
Missing	2
Oophorectomy
No	461 (44.8%)
Yes	568 (55.2%)
Missing	23
Age at oophorectomy	46.7 yrs (31.1–71.1)
Tamoxifen	283 (28.2%)
Aromatase inhibitor (AI)	70 (7.0%)
None	740 (70.3%)
PCSK9 genotype
AA	759 (72.2%)
AG	269 (25.6%)
GG	24 (2.3%)
Vital status
Alive	852 (81.0%)
Dead	200 (19.0%)
Cause of death
Due to breast cancer	140 (70.0%)
Due to other cause	60 (30.0%)

Table 3.

pCR rates according to PCSK9 c.1420G>A genotype among 404 BRCA1-positive patients who underwent neoadjuvant chemotherapy

Genotype (all patients)a	Patients with neoadjuvant therapya	pCR number	%pCR
AA (N = 757)	290 (38.1%)	126	45.5%
AG (N = 268)	106 (39.6%)	44	44.0%
GG (N = 24)	8 (33.3%)	3	42.9%
All (N = 1049)	404 (38.5%)	173	45.1%

^a20 patients had missing data on response to chemotherapy and were excluded from the table.

Overall, 140 of the 1052 BRCA1 mutation carriers died from breast cancer (Table 2). Among these patients, the mean time from diagnosis to death was 5.6 years (range: 0.2–21.5 years). We conducted a survival analysis of BRCA1-postive patients from the date of diagnosis until the date of death from breast cancer, death from another cause, or last follow-up. A Kaplan–Meier curve was constructed to compare breast cancer survival between patients with the AA genotype and those with the AG/GG genotypes (Fig. 1). The 15-year breast cancer survival was 84.5% in the AA genotype group compared to 79.6% in the AG/GG genotype group (p = 0.14; Fig. 1). The multivariable HR was adjusted for age at diagnosis, tumor size, lymph node status, ER status, chemotherapy, endocrine therapy and oophorectomy. Oophorectomy was modeled as a time-dependent covariate. Using the AG/GG group as the refence, the unadjusted HR for breast cancer–specific death in the AA group was 0.81 (95% CI: 0.57–1.16, p = 0.25). After adjustment, the HR for breast cancer–specific death was 0.94 (95% CI: 0.65–1.36, p = 0.75; Table 4).

Fig. 1. Breast cancer survival according to PCSK9 c.1420G>A genotype among BRCA1-positive patients.

Kaplan–Meier analysis of 1052 women with stage I–III breast cancer carrying a pathogenic BRCA1 mutation. Patients were stratified under a recessive model comparing AA homozygotes with AG/GG genotype carriers. Follow-up was from diagnosis to breast cancer–specific death, death from other causes, or last follow-up. The blue solid line represents the AA genotype, and the red dashed line represents AG/GG genotypes. The 15-year breast cancer–specific survival was 84.5% in the AA group and 79.6% in the AG/GG group. Survival curves were compared using the log-rank test. Multivariable Cox proportional hazards models were adjusted for age at diagnosis, tumor size, nodal status, estrogen receptor status, chemotherapy, endocrine therapy, and oophorectomy.

Table 4.

Hazard ratios for breast cancer–specific mortality in BRCA1-positive patients according to clinical characteristics and the PCSK9 c.1420G>A genotype

Characteristic	Deaths/Total	Univariable HR (95% CI, p-value)	Multivariablea HR (95% CI, p-value)
Age at diagnosis
<40	36/349	1.0	1.0
41–50	57/403	1.30 (0.86–1.98, 0.22)	1.47 (0.96–2.26, 0.08)
51–60	35/235	1.36 (0.85–2.16, 0.20)	1.92 (1.18–3.13, 0.01)
>61	12/65	1.95 (1.01–3.75, 0.05)	2.42 (1.21–4.83, 0.01)
ER status
ER- negative	94/740	1.0 (Ref)	1.0 (Ref)
ER positive	45/298	1.28 (0.89–1.82, 0.18)	0.90 (0.50–1.61, 0.72)
Missing	1/14
Size (cm)
0–2	44/501	1.0 (Ref)	1.0 (Ref)
2.1–5.0	63/413	1.76 (1.20–2.59, 0.004)	1.62 (1.08–2.41, 0.02)
>5	20/71	3.88 (2.28–6.58, <0.0001)	2.89 (1.64–5.10, 0.0002)
Missing	13/67
Nodal status
Negative	64/714	1.0 (Ref)	1.0 (Ref)
Positive	67/307	2.57 (1.83–3.62, <0.0001)	2.45 (1.70–3.52, <0.0001)
Missing	9/31
Oophorectomyb
No	90/461	1.0 (Ref)	1.0 (Ref)
Yes	46/568	0.48 (0.32–0.68, <0.0001)	0.48 (0.33–0.70, 0.0002)
Treatmentc
No tamoxifen or AI	6/43	1,0 (Ref)	1.0 (Ref)
Tamoxifen only	27/193	0.74 (0.31–1.81, 0.51)	0.75 (0.28–1.99, 0.56)
AI only	3/33	0.47 (0.12–1.88, 0.29)	0.44 (0.10–1.94, 0.28)
Both tamoxifen and AI	9/29	1.51 (0.53–4.27, 0.44)	1.17 (0.37–3.68, 0.79)
PCSK9 genotype
AA	96/759	0.81 (0.57–1.16, 0.25)	0.94 (0.65–1.36, 0.75)
AG/GG	44/293	1.0 (Ref)	1.0 (Ref)

^aAdjusted for age at diagnosis, tumor size, lymph node status, ER status, chemotherapy, endocrine therapy and oophorectomy.

^bOophorectomy was treated as a time-dependent variable.

^cER-positive cancers only.

We performed an exploratory analysis of survival following distant recurrence among BRCA1-positive patients (Fig. 2). However, this endpoint reflects prognosis after metastasis has occurred and is different from distant metastasis–free survival, which was investigated by Mei et al. Among BRCA1-positive patients who experienced distant recurrence, the probability of survival was lower in the AG/GG group (22.2%) compared to the AA group (34.9%), but this difference was not statistically significant (p = 0.17; Fig. 2).

Fig. 2. Time to death following distant recurrence according to PCSK9 c.1420G>A genotype among BRCA1-positive patients.

Kaplan–Meier analysis of survival following distant recurrence among BRCA1 mutation carriers. Time was measured from recurrence to breast cancer–specific death or last follow-up. Patients were stratified as AA versus AG/GG genotypes. The blue solid line represents the AA genotype, and the red dashed line represents AG/GG genotypes. Post-recurrence survival was 34.9% in the AA group and 22.2% in the AG/GG group. Survival curves were compared using the log-rank test.

We also followed 467 breast cancer patients with a family history of the disease who tested negative for mutations in known cancer susceptibility genes. Vital status was obtained from the Polish vital statistics registry, and during follow-up, 108 patients had died (Table 5). Using the AG/GG group as the reference, the unadjusted HR for all-cause mortality in the AA group was 1.27 (95% CI: 0.80–2.02; p = 0.31). After adjustment for age at diagnosis, the HR was 1.29 (95% CI: 0.81–2.05; p = 0.28) for the AA genotype compared with the AG/GG group (Table 5). The 15-year all-cause survival was 75.7% for carriers of the PCSK9 c.1420G>A AA genotype compared with 80.8% for carriers of the AG/GG genotypes among familial breast cancer patients without mutations in known susceptibility genes. However, this difference was not statistically significant (p = 0.43; Fig. 3).

Table 5.

Hazard ratios for all-cause mortality among familial breast cancer patients without mutations in known cancer susceptibility genes according to the PCSK9 c.1420G>A genotype

Characteristic	Deaths/Total	Univariable HR (95% CI, p-value)	Multivariablea HR (95% CI, p-value)
Age at diagnosis
<40	8/60	1.0	1.0
41–50	35/168	1.45 (0.67–3.13, 0.34)	1.43 (0.66–3.09, 0.36)
51–60	35/162	1.69 (0.78–3.64, 0.18)	1.69 (0.78–3.65, 0.18)
>60	30/77	2.80 (1.28–6.11, 0.01)	2.80 (1.28–6.11, 0.01)
PCSK9 genotype
AA	84/346	1.27 (0.80–2.02, 0.31)	1.29 (0.81–2.05, 0.28)
AG/GG	24/121	1.0 (Ref)	1.0 (Ref)

^aAdjusted for age at diagnosis.

Fig. 3. All-cause survival according to PCSK9 c.1420G>A genotype in mutation-negative familial breast cancer patients.

Kaplan–Meier analysis of 467 women with familial breast cancer without mutations in known susceptibility genes. Follow-up was from diagnosis to death from any cause or last follow-up. Patients were stratified as AA versus AG/GG genotypes. The blue solid line represents the AA genotype, and the red dashed line represents AG/GG genotypes. The 15-year all-cause survival was 75.7% for AA and 80.8% for AG/GG. Cox models were adjusted for age at diagnosis.

Discussion

The study by Mei et al. explored the role of the PCSK9 rs562556 variant (NM_174936.4:c.1420G>A, p.Val474Ile) in breast cancer metastasis and reported that the homozygous AA genotype was associated with reduced survival across multiple cohorts⁷. To replicate these findings, we analyzed two independent cohorts: familial breast cancer patients who tested negative for mutations in known cancer susceptibility genes and BRCA1-positive breast cancer patients. In neither cohort did we find evidence that AA genotype of the PCSK9 c.1420G>A variant was associated with breast cancer–specific mortality, all-cause mortality, or time to death following distant recurrence.

Importantly, the absence of a significant association in our study is unlikely to be explained by a lack of sufficient statistical power. Mei et al. reported large effect sizes for the PCSK9 rs562556 AA genotype, with HRs of ~2.0–2.5 for distant metastasis and survival outcomes under a recessive genetic model. Given the size of our BRCA1-positive cohort (n = 1052) and the observed number of events, our study had greater than 90% statistical power to detect a hazard ratio of 2.0 or greater at a two-sided significance level of 0.05. Therefore, if an effect of the magnitude previously reported were present in this population, it would likely have been observed.

The PCSK9 gene encodes proprotein convertase subtilisin/kexin type 9 (PCSK9), an enzyme that regulates cholesterol metabolism by promoting the degradation of low-density lipoprotein receptors (LDLR) on hepatocytes. Variants in PCSK9 can alter LDL cholesterol (LDL-C) levels and thereby modify the risk of cardiovascular disease (CVD)^8,9.

Beyond its canonical role in lipid metabolism, PCSK9 has recently been implicated as a potential oncogenic factor, with missense mutations identified that promote either gain-of-function activity or exert dominant-negative effects^10–13. PCSK9 overexpression has been shown to promote colon cancer progression by driving EMT and PI3K/AKT signaling, and to support hepatocellular carcinoma growth by inhibiting apoptosis. In head and neck squamous cell carcinoma, PCSK9 enhances cancer cell stemness and immune evasion, while its inhibition boosts antitumor immunity and responsiveness to PD-1 blockade^10–12. Collectively, these findings support the hypothesis that PCSK9 may influence cancer biology through cholesterol-independent mechanisms, including immune modulation and signaling cross-talk. While PCSK9 inhibitors are primarily recognized as cholesterol-lowering agents, emerging evidence suggests that they may also bolster cancer immunotherapy by enhancing the antitumor effects of PD-1/PD-L1 blockade. Mechanistically, PCSK9 suppression preserves MHC-I recycling on tumor cells, promotes LDL receptor–mediated T-cell receptor function, and reshapes the tumor microenvironment, leading to increased cytotoxic CD8⁺ T-cell infiltration and decreased regulatory T-cells. Preclinical studies and early clinical trials combining PCSK9 inhibitors with checkpoint inhibitors have demonstrated promising synergistic activity in colorectal, lung, breast, and head and neck cancers, although clinical validation remains limited and under active investigation^14,15.

The PCSK9 rs562556 variant (p.Val474Ile, NM_174936.4:c.1240G>A; allele frequency = 0.83) affects exon 9 of PCSK9, causing a substitution of valine (Val) to isoleucine (Ile) at position 474 of the PCSK9 protein¹⁶. This variant has been studied for its potential impact on lipid levels and cardiovascular health, with findings suggesting it may be associated with modest reductions in LDL-C levels¹⁷. Functional studies have explored the molecular impact of this variant, showing reduced binding affinity to LDLR and a decrease in PCSK9-mediated LDLR degradation⁹. Genome-wide association studies (GWAS) have linked p.Val474Ile to lipid profiles in multiple populations, with some evidence suggesting a protective effect against hypercholesterolemia¹⁸. Furthermore, a study in a European cohort found that individuals carrying this variant exhibited significantly lower LDL-C levels and a decreased risk of atherosclerotic cardiovascular disease, supporting the hypothesis that p.Val474Ile may have a functional role in lipid metabolism¹⁷.

The study by Mei et al. reported a strong association between the PCSK9 rs562556 variant and the risk of metastasis and death in multiple breast cancer patient cohorts, including a large cohort of early-stage breast cancer patients from Sweden, highlighting the potential of PCSK9 inhibition in metastasis prevention. Using functional assays and mechanistic studies in cell lines and mouse models, the researchers demonstrated that antibody-based inhibition of PCSK9 effectively suppressed metastasis across multiple breast cancer models. Their findings suggest that PCSK9 inhibitors, which are already FDA-approved for treating hypercholesterolemia, could be repurposed for preventing and treating breast cancer metastasis. The study also proposed that the PCSK9 rs562556 variant may serve as a prognostic biomarker to identify patients at higher risk of metastatic relapse, enabling personalized therapeutic strategies⁷.

An important distinction between our study and that of Mei et al. lies in cohort composition and underlying genetic context. A key limitation of our study is the lack of systematically collected data on time to distant metastasis, which precluded analysis of distant metastasis–free survival. This represents an important difference from the study by Mei et al., which directly evaluated this endpoint. While Mei et al. primarily analyzed largely unselected early-stage breast cancer cohorts drawn from multiple populations, our study focused on BRCA1 mutation carriers and a highly selected group of familial breast cancer patients without mutations in known susceptibility genes. These cohorts represent biologically and clinically distinct populations, characterized by differing germline genetic backgrounds, tumor biology, and treatment trajectories. It is therefore plausible that the prognostic impact of common modifiers such as the PCSK9 rs562556 variant may differ between unselected breast cancer populations and genetically defined high-risk groups. This represents a key methodological difference, as Mei et al. assessed distant metastasis–free survival in largely unselected cohorts, whereas our study was limited to survival-based endpoints in genetically defined high-risk populations.

In conclusion, while our study did not find evidence to support p.Val474Ile (PCSK9 rs562556) as a breast cancer prognostic biomarker in two high-risk Polish cohorts, it contributes important data to the growing body of literature on PCSK9 in cancer biology. The lack of replication in our cohorts does not negate prior findings; rather, it highlights the need for larger and more diverse studies that integrate epidemiological and functional data to better clarify the genetic and mechanistic role of PCSK9 in breast cancer progression and metastasis prevention. Collectively, these findings highlight the complexity of PCSK9 biology and support continued investigation into its potential as a multifaceted therapeutic target in breast cancer.

Methods

BRCA1-positive breast cancer cohort

This cohort comprised of women with breast cancer who carried a pathogenic BRCA1 mutation. Participants were eligible if they were diagnosed with stage I to stage III breast cancer at age 70 or younger, between 1995 and 2023. Patients were excluded if they had a prior diagnosis of cancer other than breast cancer, if genetic testing was conducted five years or more after the date of diagnosis, or if they lacked key clinical or follow-up information. A total of 1,052 patients were recruited, all of whom had previously tested positive for one of the six most common Polish BRCA1 founder mutations (c.5266dupC, c.181T>G, c.4035delA, c.68_69delAG, c.3700_3704delGTAAA, c.5251C>T) at the Pomeranian Medical University using previously described methods¹⁹. All patients were diagnosed and treated at the Hospital of the Pomeranian Medical University or an affiliated center in the Szczecin region of northern Poland. Clinical and treatment information was gathered through medical chart review and consultation with referring physicians. Data collected included tumor characteristics (size, lymph nodes involvement, ER-status, PR status, HER2 status and tumor grade). Details regarding the type and timing of chemotherapy (neoadjuvant and adjuvant) were recorded, as well as whether a pathological complete response (pCR) was achieved. For those who had underwent an oophorectomy, the date of surgery was recorded. Date of death was documented through linkage with the Polish Vital Statistics Registry, and cause of death was determined by the review of medical record or by interview with the treating physician or next of kin.

Familial breast cancer cohort

The second cohort included 467 women with highly familial breast cancer who tested negative for mutations in BRCA1, ATM, BRCA2, BARD1, CHEK2, TP53, PALB2, PTEN, NBN, RECQL, RAD50, and CDH1 through exome sequencing. These patients were selected from a registry of 3,519 familial breast cancer cases housed at the Hereditary Cancer Center in Szczecin, based on the number of first- or second-degree affected relatives and age at diagnosis.

Among the 467 familial breast cancer patients, 411 women were from families with at least three affected relatives, and 56 women came from families with two affected relatives, including at least one diagnosed with bilateral breast cancer or with breast cancer before age 50. The mean number of breast cancer cases per family was 3.4. The mean age at the time of breast cancer diagnosis among the 467 probands was 46 years (range: 28–76 years). Detailed information regarding cancer stage and treatment was not available for these patients. Cause of death was also unavailable; therefore, follow-up was confined to all-cause mortality, assessed using the Kaplan–Meier method.

Healthy controls

The control group included 1404 cancer-free women residing in the Szczecin region. The mean age at the time of blood draw was 48.5 years. Control participants were not tested for mutations in the breast cancer susceptibility genes listed above. We compared the frequency distribution of the PCSK9 rs562556 genotypes (AA, AG, GG) between breast cancer cases and controls to assess whether the PCSK9 genotype was associated with the risk of developing breast cancer among the Polish population. All case and control subjects were of Polish ethnicity.

PCSK9 c.1420G>A Genotyping

In this study, we investigated the PCSK9 c.1420G>A (rs562556) missense variant across the three cohorts: BRCA1-positive breast cancer patients, familial breast cancer patients without mutations in known susceptibility genes, and cancer-free controls. For the BRCA1-positive cohort and control subjects, genotyping of the PCSK9 c.1420G>A variant was performed using TaqMan Genotyping Assays (Applied Biosystems, Thermo Fisher Scientific). Germline DNA was extracted from peripheral blood leukocytes, and genotyping was carried out according to the manufacturer’s protocol. Rigorous quality control measures were applied to ensure high specificity and accuracy in variant detection. For the familial breast cancer cohort, PCSK9 genotypes were derived from existing whole-exome sequencing (WES) data²⁰. Briefly, Germline DNA was previously isolated from peripheral blood leukocytes and sequenced using the Agilent SureSelect Human All Exon V6 kit on the Illumina NextSeq 500 platform, generating 150 bp paired-end reads. Raw reads were aligned to the human reference genome (GRCh37/hg19) using BWA-MEM, and variants were called with Sentieon Haplotyper (Sentieon Inc, San Jose, CA, USA). Variant annotation was performed with SVS (Golden Helix Inc, Bozeman, MT, USA), and the PCSK9 c.1420G>A genotype was extracted directly from the annotated variant data for each participant.

Statistical analysis

For the BRCA1-positive cohort, patients were followed for survival from the date of breast cancer diagnosis until the date of death due to breast cancer, death from another cause, or the date of last follow-up. The crude 15-year all-cause mortality and breast cancer–specific mortality were calculated using the Kaplan–Meier method. For the crude analysis, patient cohorts were stratified into two groups based on their PCSK9 genotype: AA homozygotes versus AG/GG (heterozygotes and GG homozygotes combined).

In the Cox proportional hazards model, we compared mortality rates among individuals with the AA genotype of the PCSK9 rs562556 variant with those carrying the AG or GG genotypes. These categories were based on the results reported by Mei et al.⁷. Additional analyses using dominant and additive genetic models were performed and did not reveal any significant associations with survival outcomes (data not shown). In the multivariable analysis, the hazard ratio (HR) was adjusted for established prognostic factors, including age at diagnosis, ER status, tumor size, nodal status, oophorectomy, and chemotherapy, all of which are known as predictors of survival²¹. We also performed a survival analysis to evaluate the time from distant recurrence to breast cancer–specific death according to the genotype in the BRCA1-positive cohort.

For the familial breast cancer cohort, patients were followed from the date of diagnosis until death or last follow-up. The primary endpoint was all-cause mortality. The HR for death according to the PCSK9 rs562556 genotype (AA versus AG/GG) was adjusted for age at diagnosis. The mean age at diagnosis was 51.2 years (range: 25–83 years), and the mean follow-up duration was 15 years (range: 7–31 years).

The study was conducted in accordance with the Declaration of Helsinki and was approved by the ethics committees of Pomeranian Medical University in Szczecin, Poland (BN-001/174/05), and Women’s College Hospital, University of Toronto, Canada (2017-0098-E). Informed consent was obtained from all individual participants included in the study.

Author contributions

M.R.A., S.N., C.C., J.G., and J.L. conceived the study and contributed to the methodological design. M.R.A., W.K., D.W., K.S., T.H., N.Z., M.F., and K.G. curated the data. P.S. performed the formal analysis. N.S., M.R.A., and N.Z. drafted the original manuscript. All authors contributed to validation, investigation, visualization, and critically reviewed and approved the final manuscript.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to consent restrictions from the study participants but are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.