Genetic Markers of Tumor Multiplicity in Non-melanoma Skin Cancer: Associations of 19 SNPs in an Italian Cohort

Abstract

Introduction

Ultraviolet exposure is the main environmental risk factor for non-melanoma skin cancers (NMSCs), but genetic susceptibility also contributes to variability in disease burden among individuals.

Methods

This study analyzed selected single-nucleotide polymorphisms (SNPs) in genes related to vitamin D and nicotinamide adenine dinucleotide metabolism, DNA repair, inflammation, and pigmentation as potential biomarkers for NMSC risk in an Italian cohort. Participants were stratified into low- and high-risk groups on the basis of tumor burden.

Results

No significant differences were observed in demographic or phenotypic factors between the groups; however, chronic sun exposure was associated with an increased risk. A total of 19 SNPs showed significant associations with NMSC multiplicity. Most of these SNPs were noncoding variants that likely influence gene expression or transcript stability. Specific variants in the NNMT, NFKBIA, ERCC6, XPA, LIG1, LIG3, and ZNF365 genes were more prevalent in individuals at high risk. Literature-based functional data indicate that the identified NFKBIA and ERCC6 SNPs are particularly relevant to NMSC risk, as NFKBIA variants may promote inflammation, and the ERCC6 variant can impair DNA repair.

Conclusions

These results underscore the significance of regulatory genetic variation in NMSC susceptibility. The identified SNPs could represent useful biomarkers for genetic risk stratification and support the development of personalized prevention strategies based on genetic profiles.

Trial Registration

Protocol no.: 518-2/19-12-2019.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13555-026-01740-y.

Plain Language Summary

Non-melanoma skin cancers are very common tumors. Although sun exposure is the main cause, not everyone who receives similar amounts of sunlight develops the same number of tumors. This suggests that individual genetic differences may also influence tumor risk. This study examined whether small differences in DNA can explain why some people develop multiple non-melanoma skin cancers. These differences, known as “genetic variants,” are a normal part of human biology and can influence how the body responds to sun damage. We examined genetic variants in some genes responsible for metabolism, inflammation, DNA repair, and skin pigmentation in a group of Italian patients. We compared individuals with a low number of skin tumors with those with many tumors. The two groups were similar in terms of age, sex, and skin type. However, those with long-term sun exposure had more tumors. We identified 19 genetic variants that were more prevalent among individuals with multiple skin cancers. Most of these variants were found in regions of DNA that control how genes are activated, rather than in regions that alter the structure of proteins. Some of these genetic variants were associated with increased inflammation or reduced ability to repair DNA damage in response to sunlight. Both processes can increase the risk of developing skin cancer. Overall, these results suggest that genetic differences, together with sun exposure, help to explain why some individuals are more prone to developing multiple non-melanoma skin cancers. This information may support more personalized prevention strategies in the future.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13555-026-01740-y.

Key Summary Points

Why carry out this study?

Non-melanoma skin cancers (NMSCs) are common, and individual susceptibility varies owing to both sun exposure and genetic factors.

Understanding genetic contributors to multiple NMSC development could improve risk stratification and prevention strategies.

The study investigated whether selected genetic variants in genes involved in DNA repair, inflammation, metabolism, and pigmentation are associated with the risk of developing multiple NMSCs.

What was learned from the study?

19 genetic variants were significantly associated with NMSC multiplicity, most located in regulatory regions that are likely to influence gene expression rather than protein structure.

Variants in the NFKBIA and ERCC6 genes may increase tumor susceptibility through inflammation and impaired DNA repair.

Integrating genetic and environmental factors may support personalized prevention strategies for individuals at higher risk of developing multiple NMSCs.

Introduction

Non-melanoma skin cancers (NMSCs), including basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and actinic keratosis (AK), are the most common malignancies in fair-skinned populations [1, 2]. BCC accounts for ~80% of cases and is locally invasive and leads to disfigurement, particularly on the face and other cosmetically sensitive areas [3]. SCC represents ~20% of cases and displays a higher risk of metastasis [4].

Ultraviolet (UV) radiation, especially UVB, is the main environmental factor that causes DNA damage and mutations, increasing skin cancer risk. Adequate topical and systemic photoprotection remains the most effective tool in preventing NMSC onset [5, 6]. Additional risk factors include fair skin, older age, male sex, immunosuppression, and cutaneous beta-human papillomavirus infection [1]. However, not all individuals with similar environmental exposures develop NMSC, suggesting a significant role for genetic susceptibility. Single-nucleotide polymorphisms (SNPs), common heritable DNA sequence variants, are increasingly recognized as important modulators of cancer risk [7]. SNPs may influence individual responses to environmental stress, immune regulation, and genomic stability [8, 9]. Genome-wide association studies identified multiple SNPs associated with NMSC risk, particularly in genes related to pigmentation (e.g., MC1R, ASIP, SLC45A2), immune regulation (e.g., IFIH1, CCR5), cell cycle control (e.g., CDKL1), and tumor suppression (e.g., TP53, CDKN2A) [10–12]. DNA repair pathways safeguard genomic stability, and polymorphisms in repair genes such as XPC, ERCC2, and FANCA have been linked to reduced repair efficiency and increased skin cancer risk, particularly in interaction with UV exposure [13–15]. Chronic activation of the Nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) pathway impairs DNA repair and promotes inflammation. Variants in NFKB1, NFKBIA, and TNFAIP3 are linked to other cancers [16–18], but remain poorly studied in NMSC. SNPs in metabolic regulators, including vitamin D receptor- (VDR) [19, 20] and nicotinamide adenine dinucleotide (NAD)-associated genes (e.g., NAMPT, PARP1, SIRT1) [21, 22], may also modulate DNA repair and cancer susceptibility, but their role in NMSC is largely unexplored.

Incorporating genetic information, including functionally relevant SNPs, may improve risk prediction beyond traditional clinical factors, especially in high-risk populations. In patients with NMSC, tumor multiplicity reflects both cumulative UV exposure and individual genetic susceptibility that affects DNA repair and the immune response. Therefore, the number of tumors can serve as a proxy phenotype to explore genetic determinants of cutaneous carcinogenesis beyond environmental exposure. Identifying SNPs could contribute to improved risk stratification and potentially support more personalized strategies for prevention and early detection.

This study examined SNPs in genes related to vitamin D and NAD⁺ metabolism, DNA repair, pigmentation, and inflammation to assess their association with multiple NMSC development in an Italian cohort and their potential contribution to genetic risk stratification.

Methods

Study Subjects

A total of 219 patients with NMSC were recruited from the Istituto Dermopatico Dell’Immacolata–Istituto di Ricovero e Cura a Carattere Scientifico (IDI–IRCCS) (n = 158) and the Policlinico Tor Vergata (PTV), University Hospital (n = 61) between February 2020 and June 2021. All patients were white, over 18 years of age, and had at least two NMSCs in their clinical history. The clinical study protocol was approved by the local ethics committee at IDI–IRCCS, Rome, Italy, and the study (protocol no. 518-2/19-12-2019) was executed in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from each participant before inclusion in the study. Patient data were anonymized before analysis. All participants completed a questionnaire covering demographics, age, gender, habits (e.g., sun exposure, smoking), and clinical history. A 5-mL blood sample was collected from patients. The study design is schematized in Fig. 1.

Fig. 1.

Study design. The study followed an observational, cross-sectional design. All analyses were based on patients with at least two histologically confirmed tumors. To assess associations with tumor burden, a two-step stratification approach was employed: (1) patients were categorized by total tumor count (two versus > 2); (2) patients were stratified, on the basis of the 25th–75th percentiles of the tumor count distribution, into low risk (n = 2 tumors) and high risk (n ≥ 4 tumors). This classification was used for both the full cohort (n = 219) and the subgroup with available genotyping data (n = 190). Patients with percentile-based tumor counts (n = 157) were considered for demographic/clinical association analyses. Only patients with genotyping data and tumor counts within the defined percentile thresholds (n = 139) were included for genotype–risk association analyses. SNPs Single-nucleotide polymorphisms

SNP Analysis

Of the 219 patients enrolled in the study, genomic DNA suitable for SNP genotyping was only available for 190 (IDI, n = 140; PTV, n = 50) owing to insufficient starting DNA or library issues. Genomic DNA for the genotyping of SNPs was obtained from peripheral blood (EDTA-treated) using the Maxwell^® CSC Blood DNA Kit (Promega). The DNA obtained was quantified on a Qubit Fluorometer instrument, and 12 ng was used for sequencing by next-generation sequencing (NGS) technology.

The custom panel was designed after an extensive review of the literature on PubMed until January 2021, using the following search terms: “NAD Enzyme,” “VitD metabolism,” “DNA repair enzyme,” “NFKB pathway,” “Skin SCC biomarker” (comprehensive reference list provided in Supplementary Table S1).

The selected SNPs were validated using the Ion AmpliSeq Designer software (Thermo Fisher Scientific) with the hg19 genome assembly as a reference. The software generated a custom panel consisting of 161 hotspots for a total of 158 amplicons, suitable for genomic analysis using NGS techniques. Sequencing of amplicons permitted the identification of additional 244 SNPs located near the primary SNPs. Thus, the SNP array analysis identified a total of 405 genetic variants.

NGS analysis was performed on genomic DNA using the Ion GeneStudio™ S5 Plus platform (Thermo Fisher Scientific, Waltham, MA, USA). Libraries were prepared with the Ion AmpliSeq™ Library Kit Plus and quantified with a Qubit™ 4 Fluorometer (dsDNA HS assay) and an Agilent 2100 Bioanalyzer (High Sensitivity DNA kit). Sequencing was executed in four independent runs encompassing 48 multiplexed libraries. Primary processing, alignment to the human reference genome (GRCh37/hg19), variant calling, and annotation were performed in Ion Reporter™ (latest available release at the time of analysis). Variants were retained if supported by read depth > 30× and variant allele frequency (VAF) ≥ 0.30, and annotations were independently verified with ANNOVAR on the hg19 assembly (UCSC gene models).

Statistical Analysis

Descriptive statistics were used to summarize the study population. Continuous variables were described as means ± standard deviations (SD) or median with interquartile range (IQR) according to data distribution, while categorical variables were reported as absolute and relative frequencies. Normality of continuous variables was assessed by the Shapiro–Wilk test. Differences between groups were evaluated using the Mann–Whitney U test for continuous variables and Fisher’s exact test for categorical variables. A p-value ≤ 0.05 was considered statistically significant.

To evaluate associations with tumor burden, a two-step stratification approach was applied (Fig. 1). First, patients were grouped according to the total number of tumors (two versus > 2). As no significant differences emerged, a second stratification based on the 25th and 75th percentiles of the tumor count distribution was adopted: Patients with two tumors were defined as low risk, and those with ≥ 4 tumors as high risk. This stratification was applied to the full cohort (n = 219) and to the subset with available genotyping data (n = 190). The subset of patients with both genotyping data and tumor counts in the 25th–75th percentiles (n = 139) was used for genotype–risk association analyses.

Associations between clinical or genetic variables and risk categories were tested using Fisher’s exact test. Odds ratios (OR) and 95% confidence intervals (CI) were calculated where appropriate. Only variants passing predefined quality filters were included in the analysis. Statistical analyses were performed using R (version 4.3.2) and SPSS (version 28.0).

Results

Characteristics of the Study Population

A total of 219 patients were included in the study. The demographic, anthropometric, phenotypic, and phototype characteristics of the study population are summarized in Tables 1 and 2. The majority of patients were male (143, 65.3%), while 76 individuals (34.7%) were female. The average age was 69.4 years, with the 65–79-year age group being the most represented. The average body mass index (BMI) was 26.72 kg/m², and most patients had a BMI ≤ 24.9 kg/m². Most patients had phototypes I–III and thus a higher risk of sunburn. Table 2 summarizes the smoking habits, alcohol consumption, sun exposure, and personal history of skin cancer of the recruited subjects. Most patients were nonsmokers (45.2%) or former smokers (37.9%). More than half of the population (54.8%) drank alcohol. Most patients (93.6%) were frequently exposed to the sun, and this exposure was mainly acute (87.7%). Many patients used sunscreen regularly (66.2%) and rarely used sunbeds (5.5%). Previously developed skin cancer lesions were primarily BCC (89.5%), followed by SCC (37%), AK (21%), and melanoma (1.8%).

Table 1.

Characteristics of the population enrolled in the study: demographic, anthropometric, phenotypic, and phototype characteristics

Characteristics	N = 219
	Na/mean	%/(SD)
Center
IDI-IRCCS	158	72.1%
PTV	61	27.9%
Sex
Male	143	65.3%
Female	76	34.7%
Age, years
Mean (SD)	69.40	(11.77)
Median (IQR)	71.10	(60–78.7)
< 65	76	34.7%
65–79	98	44.7%
≥ 80	45	20.5%
Body mass index (kg/m2)
Mean (SD)	26.72	(22.94)
Median (IQR)	25.20	(22.7–27.5)
< 24.9	106	48.4%
25.0–29.9	96	43.8%
≥ 30.0	17	7.8%
Hair color
Black/dark brown	86	39.3%
Light brown	85	38.8%
Blond/red	41	18.7%
Eye color
Black/dark brown	52	23.7%
Light brown	64	29.2%
Green/blue	93	42.5%
Phototypeb
I–II	102	46.5%
III	103	47%
≥ IV	11	5.1%

^aTotals may vary owing to missing data

^bI—ivory, always burns, never tans; II—white, burns easily, tans with difficulty; III—cream, burns little, tans gradually and moderately; IV—beige/olive, rarely burns, tans easily; V—light brown, burns very rarely, tans heavily easily; VI—dark brown or black, never burns, tans heavily easily

SD standard deviation, IQR interquartile range

Table 2.

Characteristics of the population enrolled in the study: smoking habit, sun exposure, and personal history of skin cancer

Characteristics	N = 219
	Na/mean	%/(SD)
Smoking habit
Nonsmokers	99	45.2%
Smokers	34	15.5%
Ex-smokers	83	37.9%
Alcohol
No	97	44.3%
Yes	120	54.8%
Sun exposure
No	8	3.7%
Yes	205	93.6%
If yes,
Acute	192	87.7%
Chronic	106	48.4%
Use of sunbeds
No	198	90.4%
Yes	12	5.5%
Use of sunscreens
No	58	26.5%
Yes	145	66.2%
Tumor history
Squamous cell carcinoma	81	37%
Median (IQR)	0	(0–1)
Basal cell carcinoma	196	89.5%
Median (IQR)	2	(2–3)
Actinic keratosis	46	21%
Median (IQR)	1	(0–2)
Melanoma	4	1.8%
Median (IQR)	1	(1–1)

^aTotals may vary owing to missing data

IQR interquartile range

Association Analysis between Clinical Variables and NMSC Risk

Patients were stratified by tumor burden using a two-step approach to evaluate associations between clinical variables and NMSC risk (Fig. 1).

First, groups were defined by the total number of tumors (two versus > 2). As no significant differences emerged, a percentile-based stratification was adopted: low-risk (two tumors; n = 76) and high-risk (≥ 4 tumors; n = 81) (Table 3). Standard demographic and phenotypic variables, such as age, gender, phototype, smoking status, and alcohol consumption, did not differ between groups, indicating homogeneous study cohorts (Table 3). Chronic sun exposure was significantly associated with higher tumor burden (p = 0.02) and, specifically, with an increased number of BCCs (p = 0.03; Supplementary Table S2). No associations were observed with occupational or recreational exposures (Table 4; Supplementary Table S3).

Table 3.

Association of characteristics of the population with non-melanoma skin cancer risk

Characteristics	N = 157
	Low risk (n = 76) N (%)	High risk (n = 81) N (%)
Center			p = 0.67
Istituto Dermopatico Dell’Immacolata–Istituto di Ricovero e Cura a Carattere Scientifico	54 (71.1%)	55 (67.9%)
Policlinico Tor Vergata	22 (28.9%)	26 (32.1%)
Sex			p = 0.66
Male	50 (65.8%)	56 (69.1%)
Female	26 (34.2%)	25 (30.9%)
Age, years			p = 0.31
< 65	30 (39.5%)	23 (28.4%)
65–79	32 (42.1%)	38 (46.9%)
≥ 80	14 (18.4%)	20 (24.7%)
Body mass index (kg/m2)			p = 0.57
< 24.9	33 (43.4%)	38 (46.9%)
25.0–29.9	39 (51.3%)	36 (44.4%)
≥ 30.0	28 (38.4%)	33 (42.3%)
Hair color			p = 0.25
Black/dark brown	18 (24.7%)	11 (14.1%)
Light brown	27 (37.0%)	34 (43.6%)
Blond/red	4 (5.3%)	7 (8.6%)
Eye color			p = 0.16
Black/dark brown	38 (52.8%)	28 (37.3%)
Light brown	19 (26.4%)	28 (37.3%)
Green/blue	15 (20.8%)	19 (25.3%)
Phototypea			p = 0.32
I–II	32 (42.1%)	36 (45.6%)
III	38 (50.0%)	41 (51.9%)
≥ IV	6 (7.9%)	2 (2.5%)
Smoking habit			p = 0.60
Nonsmokers	42 (55.3%)	38 (48.1%)
Smokers	9 (11.8%)	13 (16.5%)
Ex-smokers	25 (32.9%)	28 (35.4%)
Alcohol			p = 0.82
No	36 (47.4%)	36 (45.6%)
Yes	40 (52.6%)	43 (54.4%)
Sun exposure			p = 0.46
No	2 (2.7%)	4 (5.1%)
Yes	71 (97.3%)	75 (94.9%)
Acute sun exposure			p = 0.23
No	5 (6.8%)	10 (12.7%)
Yes	68 (93.2%)	69 (87.3%)
Chronic sun exposure			p = 0.02
No	44 (60.3%)	32 (40.5%)
Yes	29 (39.7%)	47 (59.5%)
Use of sunbeds			p = 0.79
No	69 (94.5%)	72 (93.5%)
Yes	4 (5.5%)	5 (6.5%)
Use of sunscreens			p = 0.49
No	24 (35.3%)	23 (29.9%)
Yes	44 (64.7%)	54 (70.1%)

Statistically significant results are displayed in bold

^aI—ivory, always burns, never tans; II—white, burns easily, tans with difficulty; III—cream, burns little, tans gradually and moderately; IV—beige/olive, rarely burns, tans easily; V—light brown, burns very rarely, tans heavily easily; VI—dark brown or black, never burns, tans heavily easily

Table 4.

High-risk and low-risk patients stratified by chronic sun exposure

Characteristics	N = 157		Significance
	Low-risk patients N (%)	High-risk patients N (%)	Chronic sun exposure (no)	Chronic sun exposure (yes)
Job			p = 0.85	p = 0.37
Employed	22 (31.4%)	21 (27.6%)
Unemployed	1 (1.4%)	0
Retired	42 (60.0%)	50 (65.8%)
Homemaker	5 (7.1%)	5 (6.6%)
Type of work			p = 0.91	p = 0.24
Indoor	21 (84.0%)	24 (92.3%)
Outdoor	2 (8.0%)	2 (7.7%)
Both	2 (8.0%)	0
Gardening			p = 0.77	p = 0.77
Yes	25 (59.5%)	29 (51.8%)
No	17 (40.5%)	27 (48.2%)
Agriculture			p = 0.30	p = 0.43
Yes	6 (14.3%)	9 (15.8%)
No	36 (85.7%)	48 (84.2%)
Sport			p = 0.56	p = 0.24
Yes	36 (49.3%)	37 (46.8%)
No	37 (50.7%)	42 (53.2%)
Running			p = 0.31	p = 0.82
Yes	15 (45.5%)	20 (57.1%)
No	18 (54.5%)	15 (42.9%)

Genomic Profiles

Targeted next-generation sequencing (NGS) was performed across four sequencing runs (48 multiplexed libraries) and yielded high-quality data, with a mean ± SD of 140,000 ± 15,700 mapped reads per sample, on-target rate 98.0% ± 4.0%, mean depth 846× ± 100×, and uniformity 96.3% ± 0.4%. After applying the prespecified filters (read depth ≥ 30× and VAF ≥ 0.30), zygosity was assigned at each locus (heterozygous/homozygous versus wild type). Summary outcomes for predefined hotspot (HS) variants are reported in Supplementary Table S4. Beyond HS loci (n = 161), sequencing of the targeted regions identified 244 additional non-HS variants.

Association Analysis of Genomic Profiling with NMSC Risk

We assessed the clinical relevance of identified SNPs by examining their association with the risk of multiple NMSCs in 139 patients (Fig. 1) stratified into low- (n = 70) and high-risk (n = 69) groups. Overall, 15 SNPs in known hotspot genes (NNMT, VDR, NFKBIA, NFKB2, ERCC3, ERCC6, LIG1, LIG3, POLB, XPA, HERC2, and MC1R) showed significant associations with NMSC (Table 5).

Table 5.

Association of single-nucleotide polymorphism genotypes with non-melanoma skin cancer risk

Genotype	Low risk (n = 70)	High risk (n = 69)	p‑Value
rs12285641 (NNMT)
CC	53 (75.7%)	38 (55.1%)	0.038
CT	15 (21.4%)	27 (39.1%)
TT	2 (2.9%)	4 (5.8%)
rs4760648 (VDR)
CC	21 (30.0%)	31 (44.9%)	0.039
CT	33 (47.1%)	32 (46.4%)
TT	16 (22.9%)	6 (8.7%)
rs1421509936 (VDR)
AA	32 (45.7%)	45 (65.2%)	0.02
AAG	31 (44.3%)	23 (33.3%)
AGAG	7 (10.0%)	1 (1.4%)
rs2233406 (NFKBIA)
GG	40 (57.1%)	25 (36.2%)	0.035
GA	26 (37.1%)	35 (50.7%)
AA	4 (5.7%)	9 (13.0%)
rs3138053 (NFKBIA)
TT	39 (55.7%)	25 (36.2%)	0.05
TC	27 (38.6%)	35 (50.7%)
CC	4 (5.7%)	9 (13.0%)
rs4919632 (NFKB2)
CT	3 (4.3%)	10 (14.5%)	0.039
TT	67 (95.7%)	59 (85.5%)
rs4150403 (ERCC3)
CC	56 (80.0%)	65 (94.2%)	0.04
CT	13 (18.6%)	4 (5.8%)
TT	1 (1.4%)	0 (0.0%)
rs4253160 (ERCC6)
TT	33 (47.1%)	16 (23.2%)	0.012
TA	22 (31.4%)	33 (47.8%)
AA	15 (21.4%)	20 (29.0%)
rs3793784 (ERCC6)
GG	34 (48.6%)	17 (24.6%)	0.008
GC	21 (30.0%)	36 (52.2%)
CC	15 (21.4%)	16 (23.2%)
rs156641 (LIG1)
CC	36 (51.4%)	25 (36.2%)	0.015
CT	30 (42.9%)	29 (42.0%)
TT	4 (5.7%)	15 (21.7%)
rs3136027 (LIG3)
AA	68 (97.1%)	61 (88.4%)	0.046
AT	2 (2.9%)	8 (11.6%)
rs3136797 (POLB)
CC	65 (92.9%)	69 (100.0%)	0.024
CG	5 (7.1%)	0 (0.0%)
rs1800975 (XPA)
TT	12 (17.1%)	3 (4.3%)	0.05
TC	29 (41.4%)	33 (47.8%)
CC	29 (41.4%)	33 (47.8%)
rs12916300 (HERC2)
CC	8 (11.4%)	20 (29.0%)	0.028
CT	42 (60.0%)	30 (43.5%)
TT	20 (28.6%)	19 (27.5%)
rs885479 (MC1R)
GG	65 (92.9%)	69 (100.0%)	0.024
GA	5 (7.1%)	0 (0.0%)
rs10761652 (chr10:64397656)
GG	11 (15.7%)	2 (2.9%)	0.034
GA	20 (28.6%)	22 (31.9%)
AA	39 (55.7%)	45 (65.2%)
rs11568820 (chr12:48302545)
CC	35 (50.0%)	44 (63.8%)	0.039
CT	30 (42.9%)	25 (36.2%)
TT	5 (7.1%)	0 (0.0%)
rs1946710317 (chr12:48304552)
CC	35 (50.0%)	44 (63.8%)	0.039
CT	30 (42.9%)	25 (36.2%)
TT	5 (7.1%)	0 (0.0%)
rs16846876 (chr4:72592491)
AA	25 (35.7%)	35 (50.7%)	0.038
AT	42 (60.0%)	27 (39.1%)
TT	3 (4.3%)	7 (10.1%)

Genotype abbreviations indicate homozygous major allele, heterozygous, and homozygous minor allele genotypes

Analyses were extended to all SNPs not included in the HS target list. These “additional” SNPs located near the primary SNPs may have functional or clinical significance. Sequencing may also reveal previously unreported variants. Overall, 4 of the 244 “additional” SNPs were associated with NMSC risk (Table 5): rs10761652 (located in a regulatory enhancer near ZNF365); rs11568820 and rs1946710317 (located in VDR regulatory regions); and rs16846876 (located in chromosome 4, with no clear gene assignment in available public datasets).

Overall, 19 SNPs across genes involved in metabolism, inflammation, DNA repair, UV susceptibility, and pigmentation were associated with NMSC risk (Table 5; Supplementary Table S5).

Discussion

This study examined the contribution of selected SNPs in biologically relevant pathways to interindividual variability in NMSC burden. The findings indicate that several genetic variants are associated with the risk of developing multiple NMSCs, supporting the study hypothesis. Demographic and phenotypic characteristics were comparable between risk groups, suggesting a homogeneous cohort, while cumulative sun exposure was significantly associated with tumor burden, particularly in relation to BCCs.

SNPs in Metabolic Pathway Genes and NMSC Risk

Elevated nicotinamide N-methyltransferase (NNMT) expression was observed in several cancers, including SCC, and is associated with enhanced migration, proliferation, and resistance to therapy. NNMT can promote tumorigenesis by influencing cell metabolism and epigenetic regulation [23, 24]. The NNMT gene carries numerous somatic mutations in tumors, though no germline SNPs have been tied to skin cancer susceptibility [23]. In our cohort, the wild-type genotype at rs12285641 in the NNMT gene was more frequently observed in individuals with lower NMSC counts, while the presence of variant genotypes CT and TT was enriched among those with a higher tumor burden. Thus, our findings indicate a potential role for rs12285641 in NMSC susceptibility.

The VDR gene plays a protective role against NMSCs, as demonstrated by the high incidence of skin tumors in VDR knockout mice [25]. Few VDR polymorphisms have been associated with NMSC susceptibility [19, 20]. Here, two VDR SNPs, rs4760648 and rs1421509936, displayed increased frequency of wild-type genotypes linked to higher NMSC burden. At rs4760648, the homozygous genotype was more frequent in individuals with a low NMSC burden. No significant differences were found between the two risk groups for the heterozygous genotype. At rs1421509936, subjects with a low NMSC burden displayed increased frequency of heterozygous or homozygous genotypes.

Two “additional” SNPs (rs11568820 and rs1946710317) near the VDR promoter had similar frequencies. Subjects with more tumors predominantly carried the CC genotype for both. rs11568820 resides in a promoter regulatory region, affecting CDX-2 binding and modulating VDR expression [26]. CDX-2 is a transcription factor that is typically expressed in the intestinal epithelium. In intestinal Caco2 cells, the rs11568820 variant allele increases VDR promoter activity by ~15% [27]. However, in monocytes, where CDX2 is absent, the variant allele correlates with increased promoter methylation and reduced VDR expression [28]. Although normal skin does not express CDX2, some primary skin carcinomas, particularly those with basaloid morphology, brisk mitotic activity, atypia, or ambiguous histology, show nuclear CDX2 positivity [29]. The CC genotype might increase the risk of tumor development by reducing VDR expression and impairing downstream protective mechanisms, including DNA repair, apoptosis, and immune surveillance.

SNPs in NF-κB Pathway Genes and NMSC Risk

NFKBIA encodes IκBα, an inhibitor of NF-κB, a transcription factor involved in immune responses, inflammation, cell survival, and tumor suppression [30]. In our study, heterozygous and homozygous genotypes at rs2233406 and rs3138053 of the NFKBIA gene were over-represented among high-burden patients. These SNPs are functionally relevant as they lie in the NFKBIA promoter region and are known to reduce its transcriptional activity. In reporter assays, promoters harboring the minor variants show approximately 50% lower transcriptional activity than common variant constructs in CHO-K1 cells. Peripheral blood mononuclear cells (PBMCs) from individuals with heterozygous genotypes exhibited lower IκBα transcript and protein expression as well as altered NF-κB-mediated inflammatory responses and innate immunity, including increased TNF-α secretion, compared with wild-type individuals [31, 32]. Notably, the rs2233406 heterozygous genotype is also associated with increased IL-1β production following immune stimulation, and reduced responsiveness to radioactive iodine therapy [16, 18]. As dysregulated NF-κB signaling may foster a proinflammatory, tumor-promoting microenvironment [33], our findings support a functional link between elevated NF-κB activity and increased NMSC burden. Previous studies have associated rs2233406 with an increased risk of cancer, particularly hepatocellular and gastrointestinal tumors. However, meta-analyses have reported inconclusive evidence regarding overall tumor susceptibility, suggesting context-dependent effects. Conversely, meta-analyses indicated that the rs3138053 is a good candidate for susceptibility to overall cancers, especially in hepatocellular cancer [17, 34, 35].

The NFKB2 gene encodes the p100/p52 protein, which plays a role in the noncanonical NF-κB signaling pathway and is important for processes such as antigen presentation, cytokine production, and tumor immune surveillance. Downregulation of NFKB2 has been reported in various tumors, including melanoma, and contributes to immune escape and tumor progression [30]. No enrolled subjects had the wild-type genotype at rs4919632 of the NFKB2 gene. High-risk individuals displayed increased CT genotype, while low-risk subjects more often had the TT genotype. The lack of functional data limits the ability to draw mechanistic conclusions.

SNPs in DNA Repair Genes and NMSC Risk

The present study identified several SNPs in DNA repair genes significantly associated with NMSC risk. Polymorphic variations may compromise genomic stability under chronic UV exposure, leading to higher tumor susceptibility [36].

Mutations in the ERCC3 and XPA genes cause xeroderma pigmentosum, an autosomal recessive disorder characterized by extreme UV sensitivity and early onset skin cancers [36].

The ERCC3 gene, also known as XPB, encodes a subunit of transcription factor IIH (TFIIH), which unwinds the double helix surrounding a DNA adduct [36], and the variant rs4150403 has been associated with HNSCC risk [37]. In our study, high-risk individuals more frequently carried the wild-type genotype at rs4150403. The TT and CT genotypes were more prevalent among low-risk individuals. rs4150403 is an intronic SNP that could influence splicing efficiency, messenger RNA (mRNA) stability, or enhancer/silencer interactions.

The XPA gene encodes a zinc-finger protein that plays a central scaffolding role in repair [36]. In our cohort, the TC and CC genotypes at rs1800975 displayed increased frequency in high-risk individuals, whereas the wild-type genotype was more prevalent among those with fewer tumors. rs1800975 has been evaluated in lung cancer, though findings are not entirely consistent across all studies. Some meta-analyses have found that the C allele is associated with an increased risk of cancer, particularly in smoking-related cases. However, others did not find significant associations, likely due to population differences or interaction with environmental factors [38–40]. rs1800975 lies in the 5′ untranslated region (UTR), which can affect mRNA stability, transcription efficiency, and translation initiation. Thus, the TC and CC genotypes could lead to reduced XPA expression, impaired repair efficiency, and increased UV susceptibility and NMSC development.

The ERCC6 gene, also known as CSB, is a DNA repair protein involved in DNA repair and chromatin remodeling in both nuclear and mitochondrial DNA [36]. Although mutations in the ERCC6 gene cause Cockayne syndrome, which is not typically associated with cancer, ERCC6-deficient mice are more susceptible to skin cancer. This finding suggests that ERCC6 plays a complex role in tumor prevention [36]. In our study, wild-type genotypes at rs4253160 and rs3793784 were more prevalent in the low-risk group. Conversely, the TA and AA genotypes at rs4253160, as well as the GC genotype at rs3793784, were more common in the high-risk group. Functional data have demonstrated that the allele C at rs3793784 reduces transcription in vitro and in vivo [41, 42] and influences susceptibility to lung cancer [42]. Thus, our data corroborate these functional findings, indicating that individuals with a greater tumor burden are more likely to carry genotypes associated with diminished DNA repair capacity.

The LIG1 and LIG3 genes encode DNA ligases involved in replication and high-fidelity DNA repair [36]. The LIG1 rs156641 variant was more prevalent in the high-burden group, whereas the wild-type genotype was associated with lower tumor burden. Moreover, low-risk individuals had a higher prevalence of the wild-type genotype at LIG3 rs3136027. The AT genotype was more frequent among high-risk individuals. No subjects carried the TT genotype. Since LIG3 is involved in both nuclear and mitochondrial DNA repair, the variant could contribute to genomic instability.

The POLB gene encodes the DNA polymerase β, a critical DNA repair enzyme that repairs single-nucleotide damage and maintains mitochondrial DNA integrity. Mutations or overexpression of POLB are common in cancers and can affect prognosis, therapy response, and tumor immune environment [36]. Here, only wild-type genotypes were detected in high-risk individuals at rs3136797, while the CG genotype was more frequent in the low-risk group.

Among “additional” SNPs, the rs10761652 is located within an enhancer near the ZNF365 gene, encoding the zinc-finger protein 365 (ZFP365) involved in DNA binding, transcription regulation, and genome stability. The ZNF365 gene has been associated with breast cancer and inflammatory disorders [43]. Here, the homozygous genotype at rs10761652 was more frequent among high-risk individuals, suggesting a possible role in genome stability.

SNPs in Pigmentation and UV-Susceptibility Genes and NMSC Risk

The HERC2 protein and the melanocortin 1 receptor (MC1R) regulate skin pigmentation and indirectly affect susceptibility to UV damage [11, 12].

The HERC2 gene encodes the ubiquitin-protein ligase HERC2. This multifunctional protein is involved in DNA damage response, cell cycle progression, and protein degradation via the ubiquitin–proteasome system. It also regulates pigmentation. Indeed, the HERC2 gene contains a regulatory region that modulates the expression of the OCA2 gene, which is located nearby on chromosome 15 and affects melanin synthesis [44]. The rs12916300 SNP resides in a regulatory region of the HERC2 gene, affecting OCA2 expression [45]. Previous studies have reported that rs12916300 may increase the risk of cutaneous SCC through mechanisms beyond pigmentation [45]. The rs12916300 site is located just two base pairs from the FOXI1 transcription factor binding motif; however, there is no clear evidence of a direct disruptive effect on binding. Thus, its functional impact remains uncertain. Owing to the multifunctional role of the HERC2 protein, this variant could influence susceptibility by altering the regulation of the DNA damage response or modulating the p53 pathway [45]. In our cohort, the wild-type genotype displayed increased frequency in individuals with multiple NMSCs, whereas the CT genotype was more prevalent in the low-risk group.

The MC1R gene encodes a highly polymorphic G protein-coupled receptor that modulates pigmentation and inflammation [46]. The rs885479 SNP, located in the MC1R gene, is nonsynonymous, resulting in a change in amino acids (R163Q). In biochemical assays, the R163Q missense variant exhibits reduced basal signaling capacity. However, it maintains surface receptor expression and responds normally to synthetic agonists, even when α-MSH-mediated signaling is impaired. The MC1R variant results in attenuated receptor signaling and reduced cAMP/MAPK activation in melanocytes, which may reduce UV sensitivity [46, 47]. Computational tools also support a functional impact, confirming possible deleterious effects [48]. However, the clinical or phenotypic consequences of R163Q appear to be modest. There are only weak associations with pigmentation traits in Europeans and ambiguous effects in diverse populations [49, 50]. No enrolled patients had the AA genotype. High-risk individuals exclusively presented the wild-type allele, whereas the heterozygous genotype was only found in low-risk subjects.

Clinical and Biological Implications

The SNPs identified in this study span key biological pathways, including metabolism, inflammation, DNA repair, and pigmentation, and represent candidate genetic markers that may help stratify patients at risk of multiple NMSCs. Their cumulative effect likely contributes to interindividual differences in susceptibility. Most variants reside in noncoding or regulatory regions, suggesting an impact on gene expression or transcript stability rather than protein structure, consistent with large-scale genomic studies highlighting the predominance of regulatory variants in cancer risk [10–12].

Several SNPs in our cohort have known functional roles. For example, NFKBIA variants may promote a proinflammatory, tumor-supportive microenvironment, while ERCC6 variants may increase tumor susceptibility through impaired DNA repair. Conversely, a subset of SNPs appeared more frequent in low-risk individuals, suggesting a limited impact on susceptibility or potential protective effects.

Overall, these findings support a polygenic contribution to NMSC susceptibility and highlight specific genetic variants with potential clinical relevance.

Study Limitations

This study has some limitations. The moderate sample size may limit the detection of rare variants and the execution of subtype-specific analyses. In addition, the analysis was performed in a relatively homogeneous Italian cohort, which may restrict the generalizability of the findings to other ethnic groups. Furthermore, several of the identified variants still require functional validation to clarify their mechanistic roles in NMSC susceptibility.

Despite these limitations, our results support a role for regulatory genetic variation in modulating NMSC burden and highlight variants that warrant future functional investigation and validation in larger and more ethnically diverse cohorts.

Conclusions

This study assessed the association between selected SNPs and the risk of developing multiple NMSCs in a clinically homogeneous cohort, in which cumulative sun exposure, particularly for BCCs, emerged as the main nongenetic factor associated with tumor burden.

Our findings support a polygenic contribution to NMSC susceptibility and identify biologically relevant genetic variants associated with higher tumor burden. The two-step stratification approach and the use of exact tests strengthen the robustness of the observed associations despite the moderate sample size.

Overall, these findings highlight the potential value of integrating genetic information with clinical and environmental factors to improve risk stratification and support the development of precision prevention approaches. If validated in larger and multiethnic cohorts, these variants may contribute to the development of integrated genetic risk prediction models and to the advancement of precision medicine in NMSC management.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Conceptualization: Elena Dellambra; data curation: Carola Valente, Nidia Margot Salcedo, and Simona Mastroeni; formal analysis: Giovanni Luca Scaglione, Simona Mastroeni, and Damiano Abeni; funding acquisition: Luca Fania and Elena Dellambra; investigation: Irene Campana, Ylenia Aura Minafò, Giulia Giovanardi, Terenzio Cosio, Caterina Lanna, Fabio Artosi, Sara Lambiase, Martina Morelli, Marilena Minieri, Alfredo Giovannelli, and Valeria Bartolocci; project administration: Elena Dellambra; resources: Enzo Palese, Cinzia Mazzanti, Luca Fania, Cristina Albanesi, and Daniele Avitabile; supervision: Elena Dellambra, Luca Fania, and Elena Campione; validation: Elena Dellambra, Luca Fania, Giovanni Luca Scaglione, and Damiano Abeni; visualization: Elena Dellambra and Valeria Bartolocci; writing—original draft preparation: Elena Dellambra; writing—review and editing: Elena Dellambra, Luca Fania, Giovanni Luca Scaglione, and Elena Campione. All authors read and approved the final manuscript. We are grateful to Drs. Denise Campagna, Alessia Barcherini, and Laura Bilotti for help in biological sample processing, and to Dr. Giulia Raimondi for statistical analysis. The authors thank all study participants for their valuable contribution to this research.

Funding

This research was supported by the Italian Ministry of Health under grant “Ricerca Corrente RC2020-23” and grant “RF2016-02362541” to E.D., by European Union Next Generation EU–PNRR M6C2–Investimento 2.1 Valorizzazione e potenziamento della ricerca biomedica del SSN, under grant “PNRR-MCNT2-2023-12378474” and by IDI Farmaceutici s.r.l. to L.F. and E.D. The journal’s Rapid Service Fee was funded by European Union Next Generation EU–PNRR M6C2–Investimento 2.1 Valorizzazione e potenziamento della ricerca biomedica del SSN, under grant “PNRR-MCNT2-2023-12378474.”

Data Availability

NGS data supporting the findings of this study can be found in the Supplementary Materials or obtained on reasonable request from the corresponding author.

Declarations

Conflict of Interests

Irene Campana, Ylenia Aura Minafò, Giovanni Luca Scaglione, Giulia Giovanardi, Terenzio Cosio, Caterina Lanna, Fabio Artosi, Sara Lambiase, Carola Valente, Valeria Bartolocci, Martina Morelli, Cristina Albanesi, Enzo Palese, Marilena Minieri, Alfredo Giovannelli, Daniele Avitabile, Nidia Margot Salcedo, Simona Mastroeni, Cinzia Mazzanti, Damiano Abeni, Elena Campione, Luca Fania, and Elena Dellambra have nothing to disclose.

Ethical Approval

The clinical study protocol was approved by the local ethics committee at IDI–IRCCS, Rome, Italy, and the study (protocol no. 518-2/19-12-2019) was executed in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from each participant before inclusion in the study.