Germline Variants in the Immune Response-Related Genes: Possible Modifying Effect on Age-Dependent BRCA1 Penetrance in Breast Cancer Patient

Simple Summary

BRCA1/2 mutations are associated with highly elevated but still not fatal risk of breast and ovarian cancer. The identification of modifiers BRCA1/2 penetrance is essential for the personalization of medical management of carriers of BRCA1/2 pathogenic alleles. There are two novelties related to this investigation. Firstly, all relevant studies in this area have a shortage of non-affected BRCA1/2 mutation carriers. We have proposed an alternative approach, which is based on the comparison of early-onset versus late-onset cancer patients, thus assuming that the age at diagnosis may serve as a surrogate of BRCA1/2 penetrance. Secondly, we focused not on common polymorphisms but on rare pathogenic variants in immune response-related genes, which cannot be analyzed upon conventional genome-wide association studies. This effort led to the demonstration of the BRCA1 penetrance-modifying role for the PRF1 p.Ala91Val variant. PRF1 p.Ala91Val homozygotes are known to be affected by familial hemophagocytic lymphohistiocytosis, while heterozygous carriers of this allele may have a subclinical immune deficiency.

Abstract

Background: BRCA1/2 mutations are the most recognized causes of hereditary breast cancer (BC), but their penetrance is incomplete. BRCA1-driven tumors are often chromosomally unstable and exhibit increased antigenicity. We hypothesized that inherited variations in immune-related pathways may influence BRCA1 penetrance. Methods: Case–control comparison of BC-affected versus non-affected BRCA1-mutated women is generally complicated because the latter groups are often low in numbers, represented by younger subjects and may contain relatives of the analyzed patients. We utilized a novel approach, i.e., we compared young-onset and late-onset BRCA1-associated BC cases, assuming that the early age at disease manifestation may be an indicator of increased BRCA1 penetrance. Results: NGS for 353 genes implicated in inborn errors of immunity was performed on 42 young-onset (<39 y.o.) and 35 late-onset (>57 y.o.) BC patients carrying BRCA1 pathogenic variants. This effort identified 22 potentially relevant variants, which were further analyzed in an extended cohort (up to 90 patients per group). The PRF1 p.Ala91Val variant, associated with familial hemophagocytic lymphohistiocytosis, was found in 9.6% of young-onset patients and none of the late-onset group (7/73 vs. 0/78, p = 0.005). The significance of this allele was further validated in an additional group of Russian patients (14/164 (8.5%) vs. 8/236 (3.4%), p = 0.042). This trend also retained upon the pooled analysis of Russian and Polish subjects (24/278 (8.6%) vs. 15/337 (4.4%), p = 0.045). Conclusions: Rare variants in immune-related genes, such as PRF1 p.Ala91Val, may influence BRCA1 penetrance. Broader exome-wide analyses comparing affected vs. unaffected BRCA1/2 mutation carriers, or women stratified by age at cancer onset, could help identify additional genetic modifiers of cancer risk.

1. Introduction

BRCA1/2 pathogenic variants are the most common genetic cause of hereditary breast cancer (BC) syndrome, being responsible for up to 5–8% of the total BC morbidity [1,2,3]. The penetrance of pathogenic BRCA1/2 variants is usually within the range of 60–90%. Furthermore, BRCA1/2 mutation carriers are characterized by huge variations with regard to the age of disease onset [4,5,6,7]. The individual lifetime risk of BRCA1-associated BC depends on the interplay between lifestyle and genetic factors, many of which are still unknown. There are several dozen single-nucleotide polymorphisms (SNPs), which have been identified mainly by the efforts of the Consortium of Investigators of Modifiers of BRCA1 and BRCA2 (CIMBA) as potential contributors to BC risk in BRCA1/2 mutation carriers [8,9,10,11]. Many of them have been recommended for inclusion in “polygenic risk scores” to assist in the estimation of individual cancer susceptibility for BRCA1/2-carriers [12,13,14]. However, these SNP studies focused mainly on relatively common variations, e.g., SNPs with a minor allele frequency (MAF) exceeding 5%. Meanwhile, each individual genome contains a huge number of rare variants, many of which significantly affect the function of involved genes. These rare alleles cannot be efficiently analyzed by conventional genome-wide association studies (GWAS) and therefore require specifically designed investigations.

Immune-mediated mechanisms of tumor development may be particularly relevant to BRCA1-driven neoplasms, as BRCA1 inactivation is associated with a deficiency in homologous DNA repair, chromosomal instability, and, consequently, increased tumor antigenicity [15,16,17]. There are several hundred genes whose homozygous inactivation is associated with primary immune deficiencies (PID), and which may, in theory, act as modifiers of BRCA1/2 penetrance. We hypothesized that congenital defects in the functioning of some immune response genes may manifest as subclinical variations in immunity (including antitumor immunity) and influence the disease risk in BRCA1/2 mutation carriers.

While designing the study, we aimed to overcome several limitations. First of all, the distribution of BRCA1/2 pathogenic alleles is a subject of interethnic variations [18]. We acknowledged the low number of BRCA2 mutation carriers in our collection as well as the differences between BRCA1- and BRCA2-driven cancers [19], and intentionally limited the study to the BRCA1 gene. Secondly, while BRCA1 mutations predispose both to breast and ovarian carcinomas, BCs are significantly more frequent [6,20]; therefore, we focused only on this category of cancer patients. Thirdly, the correct estimation of BRCA1 penetrance requires a comparison of affected and non-affected subjects. The collection of a critical number of healthy persons with BRCA1 mutations is complicated, especially given that “true” cancer-free status can be assigned only after achieving a certain age threshold. Furthermore, “control” BRCA1 mutation carriers are usually obtained via the analysis of family members of patients with BRCA1-driven cancer, which creates a bias. In our previous study, we have implemented an alternative approach for the analysis of BRCA1 penetrance [21]. We assumed that BRCA1 carriers with unfavorable genomic context are likely to develop cancer disease at an earlier age when compared to women with disease-protective allele combinations. Hence, the analysis of genotypes of young-onset vs. late-onset patients with breast cancer has a potential for identification of BRCA1 penetrance modifiers. Here, we present the results of the study on pathogenic or likely pathogenic mutations in immune response genes in these patients’ groups.

2. Materials and Methods

We initially considered all patients with BRCA1-driven BC treated in the N.N. Petrov Institute of Oncology (St. Petersburg, Russia). The study was approved by the Local Ethics Committee (approval date: 22 October 2021; reference No. 36/302). To define the borderline values for the “young-onset” versus “late-onset” groups, we evaluated the age distribution of cancer manifestation in the hospital database, including the clinicopathological data on 1200 consecutive BRCA1 mutation carriers who were forwarded to the local diagnostic facility between years 2008 and 2022 for genetic testing. The age cutoffs between the 1st and 4th quartiles were taken as thresholds, being <39 years for young onset and >57 years for late-onset. NGS for immune response genes was initially performed for 42 young-onset versus 35 late-onset BRCA1-mutated BC patients (the “discovery” cohort,

Table 1. Characteristics of patients with BRCA1-driven breast cancer (BC) subjected to targeted NGS of immunodeficiency genes (the “discovery” cohort).

#	Age Group	Age y.o.	BRCA1 Mutation	Family History	BC Sub Type	Candidate Variant from the “Discovery” Study
1	late-onset	58	BRCA1 p.Ala457fs	s-BC 49, m-ThC 49	TNBC
2	late-onset	66	BRCA1 5382insC	m-OC 50	nd
3	late-onset	76	BRCA1 5382insC	d-BC 49	LumA
4	late-onset	61	BRCA1 5382insC	m-BC, d-BC	nd	DOCK8 p.Tyr1340Cys
5	late-onset	58	BRCA1 5382insC	m-BC, aunt(f)	TNBC
6	late-onset	83	BRCA1 5382insC	d-BC(BRCA), s-GaCa	LumA
7	late-onset	64	BRCA1 5382insC	No	nd	DDX58 rs61752945; DDX41 p.Val408Asp
8	late-onset	62	BRCA1 5382insC	No	nd
9	late-onset	70	BRCA1 5382insC	f-CRC, m-BC, aunt(f)-BC	LumA	TPP1 p.Arg208Ter
10	late-onset	60	BRCA1 5382insC	No	TNBC
11	late-onset	60	BRCA1 5382insC	m-OC, s-OC	TNBC
12	late-onset	68	BRCA1 5382insC	No	LumA	AK2 rs138577419; DNAJC21 p.Arg539Gln
13	late-onset	61	BRCA1 5382insC	aunt(f)-BC	nd	SP110 p.Gly483Arg
14	late-onset	64	BRCA1 5382insC	s-BC 50	TNBC
15	late-onset	65	BRCA1 5382insC	f-LC	nd	JAGN1 p.Met1?
16	late-onset	61	BRCA1 5382insC	m-BC, aunt(m)-UtCa	TNBC
17	late-onset	62	BRCA1 5382insC	No	TNBC
18	late-onset	61	BRCA1 5382insC	f-EsophCa, aunt (m)-BC 56	LumA
19	late-onset	61	BRCA1 4153delA	f-HNSSC	nd
20	late-onset	68	BRCA1 5382insC	m-GaCa, aunt(f), uncle(m)-GaCa	TNBC	DDX41 p.Val408Asp
21	late-onset	62	BRCA1 5382insC	m-BC, f-CRC, gm(m)-OC	TNBC
22	late-onset	58	BRCA1 5382insC	s-BC	LumB	DDX41 p.Val408Asp
23	late-onset	64	BRCA1 4153delA	gm(m)-GaCa 60, uncle(m)-GaCa 60, aunt(f)-UtCa 60	TNBC	TMC6 p.Pro502Leu
24	late-onset	63	BRCA1 p.L1205fs	m-BC 38	nd
25	late-onset	61	BRCA1 4153delA	m-UtCa	LumA	IL12B rs3213119
26	late-onset	59	BRCA1 5382insC	no	LumA
27	late-onset	61	BRCA1 p.D435fs	m-BC 53, b-BraCa 29	LumA
28	late-onset	61	BRCA1 5382insC	no	nd	DDX58 rs61752945
29	late-onset	59	BRCA1 4153delA	no	nd
30	late-onset	60	BRCA1 5382insC	aunt(m)-GaCa	nd
31	late-onset	61	BRCA1 5382insC	nd	nd	IL12B rs3213119; TMC6 p.Pro502Leu
32	late-onset	61	BRCA1 5382insC	nd	nd
33	late-onset	59	BRCA1 5382insC	no	nd
34	late-onset	63	BRCA1 5382insC	no	nd
35	late-onset	62	BRCA1 5382insC	m-BC	TNBC	ATP6AP1 p.Arg15Ter
36	young-onset	27	BRCA1 5382insC	gm(m)-LC 45	LumA	NOP10 p.Asp12His; NLRP1 p.Phe629Leu
37	young-onset	36	BRCA1 2080delA	m-BC 62, gm(m)-BC 70	TNBC
38	young-onset	37	BRCA1 185delAG	m-BC, s-BC	nd	PEPD p.Arg237Cys
39	young-onset	35	BRCA1 c.3629_3630delAG	nd	TNBC
40	young-onset	38	BRCA1 5382insC	m-BC 55, gm(m)-OC 48, aunt(f)-OC 50, gf(f)-HCC 80, s-ThC 32	LumA
41	young-onset	38	BRCA1 c.5215+1G>T	s-BC 40	TNBC
42	young-onset	33	BRCA1 c.3304_3307delAATT	m-BC	TNBC	NOP10 p.Asp12His
43	young-onset	36	BRCA1 4153delA	m-BiBC, OC 30, aunt(m)-CRC, gf(m)-LC 70	nd
44	young-onset	27	BRCA1 p.S281fs	m-OC 36, gm(f)-RenC	TNBC	STAT4 p.Thr446Ile
45	young-onset	30	BRCA1 5382insC	no	LumB	NOP10 p.Asp12His
46	young-onset	27	BRCA1 5382insC	m-BC 46, gm(m)-CRC 55, gm(f)-HCC 60	HER2+++
47	young-onset	29	BRCA1 p.R1726fs	no	TNBC
48	young-onset	29	BRCA1 C61G	m-BC	TNBC
49	young-onset	30	BRCA1 5382insC	m-BC, gm(m)-BC	TNBC
50	young-onset	30	BRCA1 5382insC	aunt(m)-CaUt	nd	RORC p.Arg10Ter; NLRC4 p.Arg310Ter
51	young-onset	26	BRCA1 5382insC	aunt(f)-BC, gf(m)-LC	TNBC
52	young-onset	27	BRCA1 5382insC	gm(f)-small intestine Ca	TNBC	PRF1 p.Ala91Val
53	young-onset	34	BRCA1 5382insC	gm-BC	nd
54	young-onset	28	BRCA1 5382insC	gm(f)-BC, aunts (m,f)-BC	TNBC
55	young-onset	33	BRCA1 5382insC	no	LumB	NOD2 p.Gly908Arg
56	young-onset	31	BRCA1 5382insC	m-BC 27, gm(m)-BC	TNBC	NOD2 p.Gly908Arg
57	young-onset	37	BRCA1 5382insC	no	LumA
58	young-onset	35	BRCA1 5382insC	gm(m)-melanoma 60, gm(f)-BC 50	TNBC
59	young-onset	36	BRCA1 4153delA	aunt(f)-OC 45, gm(f)-CRC 68, gf(f)-CaLarynx 62, aunt(cousin,f)-OC 55	LumA	PRF1 p.Ala91Val (homo)
60	young-onset	37	BRCA1 5382insC	aunt(f)-GaCa 39	LumA
61	young-onset	36	BRCA1 5382insC	m-BC 38, aunt(m)-BraCa 42	nd
62	young-onset	27	BRCA1 5382insC	no	TNBC
63	young-onset	34	BRCA1 c.5075-1G>C	m-LC 54, aunt(m)-BC 60	LumA
64	young-onset	31	BRCA1 5382insC	gf(m)-LC	TNBC
65	young-onset	37	BRCA1 5382insC	gm(m)-BC 43	nd	PNP rs104894453
66	young-onset	31	BRCA1 4153delA	m-BC, s-BC, gm(m)-OC	nd	IL17RC rs148575246
67	young-onset	32	BRCA1 5382insC	m-OC	TNBC
68	young-onset	34	BRCA1 5382insC	f-PrC	nd
69	young-onset	34	BRCA1 5656del20	no	TNBC
70	young-onset	33	BRCA1 5382insC	gf(f)-GaCa	LumA
71	young-onset	38	BRCA1 5382insC	m-BC, s-BC, aunt-BC	TNBC
72	young-onset	32	BRCA1 5382insC	f-CaLarynx, ggm(m)-BraCa	TNBC	RORC p.Arg10Ter
73	young-onset	35	BRCA1 5382insC	gm(f)-BC	TNBC
74	young-onset	33	BRCA1 5382insC	gm-CRC, gf(m)-LC	LumA	PRF1 p.Ala91Val
75	young-onset	37	BRCA1 5382insC	gm(f)-PanCa	TNBC
76	young-onset	32	BRCA1 5382insC	nd	TNBC
77	young-onset	38	BRCA1 5382insC	m-BC	TNBC

BC—breast cancer, ThC—thyroid cancer, OC—ovarian cancer, GaCa—gastric cancer, CRC—colorectal cancer, UtCa—uterine cancer; EsophCa—esophageal cancer, HNSSC—head and heck squamous cell carcinoma, BraCa—brain cancer, HCC—hepatocellular carcinoma, LC—lung cancer, RenC—renal cancer, PrC—prostate cancer, PanCa—pancreatic cancer, TNBC—triple-negative BC, LumA—luminal A BC, LumB—luminal B BC, m—mother, gm—grandmother, f—father, gf—grandfather, s—sister, b—brother, d—daughter.

; clinicopathological characteristics and the spectrum of BRCA1 pathogenic variants for these patients are presented in Supplementary Table S1). The study workflow is shown in Figure 1.

High-molecular-weight blood-derived DNA served as a source for germline mutation analysis. The custom NGS panel included 353 genes associated with 354 inborn errors of immunity (IEIs), listed in the 2019 version of the IUIS classification [22]. The panel was designed via the NimbleDesign tool and ordered from Roche [23]. A list of genes and corresponding diseases is presented in Supplementary Table S2a,b. DNA libraries were prepared using the Kapa HyperPlus Kit (Roche, Mannheim, Germany). Target enrichment was performed with the SeqCapEZ System (Roche, Mannheim, Germany). NGS analysis was carried out using the Illumina MiSeq platform with 70–90× coverage. The sequences were aligned to the GRCh37 (hg19) reference genome via the BWA 0.7.15 tool. Variant calling was performed using the GATK 3.6 instrument. Quality filtering was carried out with BCFtools 1.2 software. The multisample file was annotated using the snpEff v.4.3t tool [doi: 10.4161/fly.19695], and variants with predicted high or moderate impact were selected for further consideration. Clinical interpretation of the detected variants was performed under the ACMG/AMP 2022 guidelines. High priority was assigned to candidate mutations with the following pathogenicity attributes: (i) status “pathogenic/likely pathogenic” by the ClinVar database [24,25]; (ii) status “deleterious” by in silico tools CADD v1.7 (universal predictor; https://cadd.gs.washington.edu/, accessed on 18 November 2025) and fitCons V1.01 (the tool which integrates functional assays with selective pressure, http://compgen.cshl.edu/fitCons/, accessed on 18 November 2025); (iii) statistically increased mutation prevalence in the cancer cohort against the cancer-free population (MAF data presented in gnomAD, version 2); alleles producing OR per allele > 2 at p < 0.05 [26,27]; (iv) recurrent variants, which occurred two or more times in our collection, being exceptionally rare in the general population; and (v) potentially relevant functions of the candidate genes (i.e., involvement in the DNA damage response, proliferation, apoptosis, cell mobility, stress response, etc.).

Potentially relevant mutations were subjected to manual inspection using the Integrative Genomics Viewer (IGV) browser [https://bioviz.org/ (accessed on 15 June 2024)]. Wherever appropriate, variants were confirmed by Sanger sequencing.

Newly identified candidate variants were genotyped in the two-step validation study, which included 368 young-onset (<39 y.o.) and 427 late-onset (>57 y.o.) patients affected by BRCA1-driven BC (Figure 2). The group of “young” BC patients (median age: 33 years; range: 25–38 years) included 254 Russian women treated at the N.N. Petrov Institute of Oncology (SPb, St. Petersburg, Russia) and 114 Polish patients from Pomeranian Medical University (PUM, Szczecin, Poland). The “late-onset” group (median age: 61 years; range: 58–80 years) was composed of 326 Russian patients (“SPb”) and 101 Polish patients (“PUM”).

In the first step, the most promising candidate genetic variants selected upon NGS analysis were genotyped in 90 young-onset vs. 90 late-onset Russian BC patients using high-resolution melting (HRM) analysis coupled with Sanger sequencing of abnormally melted DNA fragments (“pilot” study). The oligonucleotides, which were utilized for this effort, are listed in Supplementary Table S3. PRF1 p.Ala91Val genotype frequencies were subjected to second round of validation, which involved 278 young-onset and 337 late-onset Russian and Polish BRCA1-associated BC cases (the “enlarged” study).

The prevalence of candidate at-risk alleles and genotypes in the studied groups was statistically analyzed by the SPSS software (version 22) via two-sided Fisher’s exact, chi-square, or Mantel-Haenszel tests. The Bonferroni–Holm method was applied for multiple-test correction.

3. Results

NGS genotyping of 353 immune response genes was performed for 42 young-onset and 35 late-onset Slavic BC patients carrying BRCA1 pathogenic alleles. A total of 2054 non-synonymous coding gene variants were detected (Supplementary Table S4a). The process of variant filtering is described in Figure 3. We considered as candidates only rare (gnomAD MAF < 5%) protein-truncating variants (n = 80) and missense variants with in silico pathogenicity CADD scores >/= 25 (n = 105) (Supplementary Table S4b, sheet 1). Further, we selected those mutations that were found exclusively in either young- or late-onset patients (Supplementary Table S4b, sheets 2 and 3). According to our hypothesis, the “young-onset” associated alleles increase the penetrance of BRCA1, whereas the “late-onset” variants are likely to have a protective effect, delaying the age of BC manifestation.

Based on the predefined criteria, 29 variants were classified as likely “protective” alleles and 42 as potential modifiers increasing BRCA1 penetrance (Supplementary Table S5). The prevalence of 22 top candidate variants was assessed in a “pilot” molecular epidemiological case–control study comprising 90 young-onset and 90 late-onset Russian BRCA1-driven BC cases (

Table 2. The frequency of 22 top candidate allelic variants discovered by NGS in 90 young-onset vs. 90 late-onset BRCA1-driven BC patients (the “pilot” study).

Gene	Description	Protein/ rs-id dbSNP	Effect	CADD *	fitCons **	MAF, %	Discovery Study N Carriers		Validation “Pilot” Study [wt/wt-mut/wt-mut/mut]
							Young- Onset (<39 y.o.)	Late- Onset (>57 y.o.)	Young- Onset (<39 y.o.)	Late- Onset (>57 y.o.)
AK2	Adenylate kinase 2	-/ rs138577419	Structural interaction	25.8	0.73 (del)	0.221	0	1	90-0-0	90-0-0
SP110	SP110 nuclear body protein	p.Gly483Arg/rs149485401	Missense	28.9	0.72 (del)	0.949	0	1	90-0-0	90-0-0
JAGN1	Jagunal homolog 1	p.Met1?/rs143438463	Start lost	25.9	0.44 (be)	0.132	0	1	84-0-0	79-0-0
IL12B	Interleukin 12B	-/ rs3213119	Structural interaction	25	0.53 (del)	3.002	0	2	88-2-0	98-0-0
DOCK8	Dedicator of cytokinesis 8	p.Tyr1340Cy/rs116920018	Missense	32	0.71 (del)	0.327	0	1	89-1-0	98-0-0
DDX58	DExD/H-box helicase 58	-/ rs61752945	Structural interaction	27.7	0.71 (del)	1.927	0	2	83-1-0	86-2-0
TPP1	Tripeptidyl peptidase 1	p.Arg208Ter/rs119455955	Stop gained	36	0.72 (del)	0.04	0	1	88-2-0	90-0-0
TMC6	Transmembrane channel like 6	p.Pro502Leu/rs75400929	Missense	25.8	0.71 (del)	0.976	0	2	89-0-0	90-0-0
ATP6AP1	ATPase H+ transporting accessory protein 1	p.Arg15Ter/rs201620814	Stop gained	28.8	n/a	0.342	0	1	90-0-0	90-0-0
DDX41	DEAD-box helicase 41	p.Val408Asp/no ID	Missense	32	0.71 (del)	.	0	3	87-3-0	87-3-0
DNAJC21	DnaJ heat shock protein family (Hsp40) member C21	p.Arg539Gln/rs146933471	Missense	33	0.74 (del)	0.053	0	1	90-0-0	90-0-0
RORC	RAR related orphan receptor C	p.Arg10Ter/rs17582155	Stop gained	36	0.5 (be)	0.368	2	0	90-0-0	90-0-0
NLRC4	NLR family CARD domain containing 4	p.Arg310Ter/rs199475953	Stop gained	35	0.55 (del)	0.031	1	0	90-0-0	90-0-0
STAT4	Signal transducer and activator of transcription 4	p.Thr446Ile/rs141331848	Missense	34	0.62 (del)	0.11	1	0	86-0-0	88-0-0
IL17RC	Interleukin 17 receptor C	-/ rs148575246	Splice donor	29.7	0.11 (be)	1.02	1	0	89-0-0	89-1-0
PNP	Purine nucleoside phosphorylase	-/ rs104894453	Structural interaction	26.5	0.67 (del)	0.004	1	0	89-0-0	86-0-0
NOP10	NOP10 ribonucleoprotein	p.Asp12His/rs146261631	Missense	28	0.44 (be)	1.218	3	0	87-3-0	87-3-0
NOD2	Nucleotide binding oligomerization domain containing 2	p.Gly908Arg/rs2066845	Missense	29.8	0.56 (del)	1.427	2	0	89-1-0	82-5-0
NLRP1	NLR family pyrin domain containing 1	p.Phe629Leu/rs149035689	Missense	25.9	0.71 (del)	1.225	1	0	90-0-0	90-0-0
PEPD	Peptidase D	p.Arg237Cys/rs766107449	Missense	33	0.71 (del)	0.002	1	0	89-0-0	90-0-0
PRF1	Perforin 1	p.Ala91Val/rs35947132	Missense	25	0.55 (del)	4.662	3	0	67-7-0 ***	78-0-0
AIRE	Autoimmune regulator	p.Arg257Ter/rs121434254	Stop gained	39	n/a	0.061	1	0	88-2-0	89-1-0

* CADD—universal tool for scoring the deleteriousness of single nucleotide variants, multi-nucleotide substitutions, and insertion/deletions variants in the human genome [https://cadd.gs.washington.edu/ (accessed on 10 May 2025)]; ** fitCons—in silico pathogenicity predictor, which integrates the results of functional assays (such as ChIP-Seq) with the data on selective pressure [http://compgen.cshl.edu/fitCons/ (accessed on 23 October 2025).], del—deleterious (score > 0.50), be—benign (score ≤ 0.50); *** Statistically significant differences between age groups have been detected: 7/73 (9.6%) vs. 0/78 (0%), Fisher exact test p = 0.005.

). Twelve of these variants were absent in both cohorts; apparently, their impact on BRCA1-driven BC risk could not be evaluated within a reasonably powered case-control study. Ten variants were observed at low frequencies (1–5%) and showed comparable distributions between the two groups, suggesting no apparent effect on the age of BC onset. Notably, only one variant, PRF1 p.Ala91Val (rs35947132), was detected exclusively in the young-onset group, with a borderline statistically significant enrichment [7/73 (9.6%) vs. 0/78 (0%), Fisher’s exact test p = 0.005; p after adjustment for multiple comparisons = 0.055]. This mutation was selected for extended case-control analysis in independent BRCA1-mutated BC cases, which included 278 young-onset and 337 late-onset patients from Poland (“PUM” cohort) and Russia (“SPb” cohort) (

Table 3. The frequency of PRF1 p.Ala91Val mutations in BRCA1-associated BC patients with early- vs. late disease onset (the “enlarged” study).

Groups	Age at Onset	PRF1 p.Ala91Val Carriers (%)	Total (Patients)	p- Value *	OR (95% CI) per Carrier	PRF1 p.Ala91Val mut/mut (%)	Total (Patients)	p-Value *	OR (95% CI) per Homozygote	PRF1 p.Ala91Val alleles (%)	Total (Alleles)	p- Value *	OR (95% CI) per Allele
BC SPb	<39 y.o. young- onset	14 (8.5%)	164	0.042	2.7 [1.08–6.51] p = 0.032	2 *** (1.2%)	164	0.047	7.1 [0.35–152.58] p = 0.201	16 (4.9%)	328	0.01	3.0 [1.25–7.03] p = 0.013
	>57 y.o. late- onset	8 (3.4%)	236			0 (0%)	236			8 (1.7%)	427
BC PUM	<39 y.o. young- onset	10 (8.8%)	114	0.800	1.3 [0.47–3.52] p = 0.618	1 *** (0.9%)	114	1.000	2.7 [0.11–66.60] p = 0.540	11 (4.8%)	228	0.630	1.4 [0.54–3.71] p = 0.484
	>57 y.o. late- onset	7 (7.6%)	101			0 (0%)	101			7 (3.5%)	202
BC PUM + SPb	<39 y.o. young- onset	24 (8.6%)	278	0.045	1.9 ** [1.00–3.76] p = 0.068	3 *** (1.1%)	278	0.273	8.6 [0.44–146.55] p = 0.156	27 (4.9%)	556	0.017	2.14 ** [1.13–4.07] p = 0.024
	>57 y.o. late- onset	15 (4.4%)	337			0 (0%)	337			15 (2.2%)	674

* Fisher exact test; ** OR adjusted, Mantel Haenszel chi-square test; *** The carriers of homozygous PRF1 p.Ala91Val genotype manifested with BC at age 27, 34, and 36, respectively.

).

In the “SPb” cohort, PRF1 p.Ala91Val carriers occurred significantly more frequently among early-onset patients than among late-onset patients [14/164 (8.5%) vs. 8/236 (3.4%), p = 0.042, Fisher’s exact test]. The increase in the prevalence of the PRF1 p.Ala91Val allele in the early-onset group was also statistically significant [16/328 (4.9%) vs. 8/472 (1.7%), p = 0.01, Fisher’s exact test]. This trend was not replicated in the “PUM” cohort, although a borderline numerical increase in the frequency of the PRF1 p.Ala91Val allele and corresponding genotypes was also observed. The significance of the association between the presence of the PRF1 p.Ala91Val allele and earlier age of BC manifestation [24/278 (8.6%) vs. 15/337 (4.4%), p = 0.045] was retained upon the pooled analysis of both groups (Table 3). The adjusted OR per carrier was 1.9 [1.00–3.76], and the OR per allele was 2.14 [1.13–4.07] (p = 0.068 and p = 0.024, Mantel Haenszel chi-square test).

For comparison, we analyzed a cohort of BRCA1-positive breast cancer patients with ages at onset ranging from 39 to 57 years (n = 84). In this intermediate group, the prevalence of heterozygous PRF1 p.Ala91Val genotypes was 5/84 (6.0%), with no homozygotes detected. As expected, this frequency falls between those observed for the young-onset (8.6%) and late-onset (4.4%) patient groups.

4. Discussion

PRF1 gene encodes perforin, which is a toxin responsible for the lysis of infected or neoplastic cells. Perforin induces the formation of pores in the attacked cell and acts in combination with protease granzymes. These proteins are stored in specialized secretory lysosomes, which are characteristic of cytotoxic T lymphocytes and “natural killers” (NKs) [28]. When these lytic granules are released, targeted cells undergo apoptosis [29,30]. Biallelic germline inactivation of the PRF1 gene (10q22.1) is associated with familial hemophagocytic lymphohistiocytosis (FHL), a severe hereditary syndrome characterized by excessive inflammation; this condition is caused by the inability of NK and CD8+ T cells to eliminate target cells through perforin-dependent and granule-mediated cytotoxicity [31,32]. Partial perforin deficiency, which is likely to be observed in heterozygous carriers of PRF1 alleles with impaired function, may be the cause of delayed FHL or other inflammatory or neoplastic disorders [33].

Our study demonstrated that the prevalence of the p.Ala91Val missense variant in the PRF1 gene was nearly twofold greater among young-onset BRCA1-driven BC patients than among late-onset patients (4.9% vs. 2.2%). The hypomorphic p.Ala91Val allele (rs35947132) is the most common variant observed in the Caucasian population, with a minor allele frequency (MAF) of 7–9%. The pathogenic potential of this mutation is well documented. The monoallelic p.Ala91Val substitution results in a 10-fold reduction in the lytic activity of the enzyme due to protein misfolding, leading to a notable decline in NK-cell cytotoxicity in healthy carriers of this variant [34,35]. Individuals who are homozygous for p.Ala91Val develop familial hemophagocytic lymphohistiocytosis type 2 (FHL2), although the penetrance of this genotype is not complete [36]. Some data indicate that PRF1 p.Ala91Val heterozygosity is associated with subclinical immunodeficiency symptoms [37]. Notably, PRF1 p.Ala91Val has also been linked to an increased risk of various hematological cancers, including B and T-cell lymphoma and acute lymphoblastic leukemia [38,39,40]. While the presence of a single copy of the p.Ala91Val allele is unlikely to significantly impact cancer risk, it may compromise the antitumor immune response, potentially contributing to the earlier development of certain solid tumors that are known to be controlled by the immune system, such as BRCA1-driven breast carcinomas. One may hypothesize that in young females, where hormonal stimulation drives rapid mammary epithelial proliferation, the combined effects of defective DNA repair resulting from BRCA1 dysfunction and compromised immune surveillance due to PRF1 deficiency may synergistically accelerate oncogenic transformation. Furthermore, subclinical impairment of natural killer cell function and other cytotoxic effector mechanisms associated with PRF1 p.Ala91Val heterozygosity may promote the release of pro-inflammatory cytokines, including TNF, IL-6, and IFN-γ [41,42], fostering a microenvironment conducive to tumor initiation and progression. The elevated incidence of hematologic malignancies among PRF1 p.Ala91Val carriers further supports the broader oncogenic potential of partial perforin deficiency [38,39,40].

In the discovery cohort, three carriers of the PRF1 p.Ala91Val variant were identified within the young-onset group; two of these cases (67%) exhibited the Luminal A subtype, compared with 14 of 51 (27%) among PRF1 wild-type cases (Table 1). Among the five additional PRF1 carriers with available receptor status information, four presented with Luminal A tumors. These trends are interesting, given that usually no more than quarter of BRCA1-associated tumors belong to the Luminal A subtype [43,44,45].

This study has several limitations. It is noteworthy that although the potential significance of the PRF1 p.Ala91Val allele has been proven by statistical analysis, its role was highly evident only in Russian patients but not in Polish patients. These discrepancies may result from random variation or reflect the influence of local environmental factors and underlying ethnic heterogeneity. The Russian cohort is likely to involve individuals with diverse ancestral backgrounds, including populations from the Caucasus, Siberia, and the Far East, whereas the Polish cohort is apparently more genetically homogeneous, consisting predominantly of women of Western Slavic descent [46]. The role of geographic, ethnic or other variations in determining BRCA1/2 penetrance has not yet been properly addressed in available studies [47,48,49,50]. Furthermore, although the distribution of the PRF1 p.Ala91Val alleles appears to differ between early-onset and late-onset BRCA1-associated BC patients, these data do not strictly support the penetrance-modifying role of this genetic variation. Finally, as this study focused exclusively on genes associated with inborn errors of immunity, it cannot exclude the possibility that additional generic modifiers may influence the age-related BC risk in BRCA1 mutation carriers [51]. Properly designed case-control multicenter investigations are needed for the validation of the role of variations in immune-related genes in determining BRCA1 penetrance.

5. Conclusions

Our findings suggest that rare genetic variations may play a role in modulating cancer risk among BRCA1/2 mutation carriers, underscoring the potential value of including these variants in future analyses. Previous investigations revealed that pathogenic variants in low-penetrance cancer-predisposing genes contribute to breast cancer development in individuals with BRCA1/2 mutations [52,53,54,55,56]. In this study, we obtained suggestive evidence for the penetrance-modifying role of genes associated with primary immune deficiencies. Further systematic exome-wide comparisons between affected and unaffected BRCA1/2 heterozygotes, or between young-onset and late-onset BRCA1/2-driven cancer cases, are required to reveal novel genetic determinants which influence cancer risk BRCA1/2 mutation carriers.