Susceptibility to Cutaneous Squamous Cell Carcinoma in Renal Transplant Recipients Associates with Genes Regulating Melanogenesis Independent of their Role in Pigmentation.

The highly polymorphic melanocortin 1 receptor (MC1R) gene plays a crucial role in pigmentation. Variants of the gene have been implicated in risk of cutaneous squamous cell carcinoma (SCC) in the general population. In renal transplant (RT) recipients these cancers are more aggressive and very common. To evaluate the risk of SCC relative to MC1R and the pigmentation-associated genes ASIP, TYR, and TYRP1, a group of 217 RT recipients with and without SCC was genotyped. Associations with SCC risk were indicated in carriers of the red hair color associated MC1R variant p.Arg151Cys (OR = 1.99; 1.05-3.75), and in carriers of two of any of the MC1R variants disclosed (OR = 2.36; 1.08-5.15). These associations appeared independent of traditionally protective phenotypes, also supported by the stratifications from skin phototype and hair color. A tendency towards an increased SCC risk was observed for a specific ASIP haplotype (OR = 1.87; 0.91-3.83), while no such associations were observed for the TYR and TYRP1 variants. Thus, the risk of developing SCC in RT patients is modulated by MC1R variation irrespective of phenotypes considered to be protective. Heterozygous combinations of MC1R variants appear to be more relevant in assessing SCC risk than the effects of variants individually.

Introduction

Improved long term graft survival and better quality of life in organ transplant recipients are explained by increasingly better graft versus host compatibility and harmonization between allograft functionality and immunosuppression.1 However, this progression comes with an overall increased risk of developing cancer, in particular non-melanoma skin cancer.2,3 Squamous cell carcinoma (SCC) represents the major challenge conferring both increased morbidity and mortality.4 Standardized incidence ratios for developing SCC in graft recipients are 65–250 times higher than in the general population.2,5 The main reason is immunosuppressive therapy causing reduced host immunosurveillance and direct oncogenic effects.6

The melanocortin 1 receptor (MC1R) gene encodes a seven pass transmembrane G protein-coupled receptor in melanocytes with a key role in regulation of melanogenesis.7,8 The gene is highly polymorphic in populations with lighter skin and some variants have been associated with both melanoma and non-melanoma skin cancer.9–11 The ability of MC1R to bind α-melanocortin stimulating hormone (αMSH) and activate adenylate cyclase to catalyze the production cAMP is crucial to pigmentation. UV induced αMSH stimulation regulates the ratio between the black-brown eumelanin and the yellow-red pheomelanin and increases melanocyte proliferation, dendricity, and melanosome transfer to keratinocytes.12 Other cytoprotective roles of MC1R involve antioxidant defense, DNA repair, and regulation of inflammatory responses through NF-κB signaling pathways.7,13,14 More than 100 MC1R variants resulting in alternative amino acids at specific codons, normally termed as non-synonymous variants, have been described with highly variable frequencies.15 Some are referred to as the RHC (red hair color) variants (p.Asp84Glu, p.Arg151Cys, p.Arg160Trp, and p.Asp294His) because of their associations with red hair, fair skin, and poor tanning. In contrast, those less associated with these phenotypic traits (p.Val60Leu, p.Val92Met, and p.Arg163Gln) are categorized as non-RHC (NRHC) variants.8,16 The only study committed to the impact of MC1R variation on risk of SCC in RT patients is a thesis stating that an increased SCC risk was observed in carriers of the RHC associated variants p.Asp84Glu and p.Arg151Cys, and that the increased risk was independent of pigmentation.17

The agouti signaling protein encoded by the ASIP gene antagonizes ligand-induced basal cAMP activity resulting in default pheomelanin pigmentation.18 A specific ASIP haplotype (AH) based on two single nucleotide polymorphisms has been associated with red and blond hair, tanning sensitivity, and increased risk of skin cancer.19,20 Downstream of MC1R, a tyrosinase (TYR) is involved in both eumelanin and pheomelanin synthesis, while a tyrosinase related protein (TYRP1) is involved in eumelanin synthesis exclusively.21 Polymorphic variants of TYR and TYRP1 associate with blue eye color, increased skin sensitivity to sun, and melanoma risk.20

With reference to skin type, hair and eye color, the aim of this study was to assess the impact of MC1R variation on SCC risk in RT patients. Also the relationship between the naturally occurring variants of ASIP, TYR, and TYP1 and susceptibility to SCC was explored. This is the first comprehensive report dedicated to the correlation between genetic variation associated with phenotypic traits and susceptibility to SCC risk in RT patients.

Materials & Methods

Study subjects

All participants invited (n = 555) were recruited through the Norwegian Renal Registry. They were all above the age of 18 with functional renal grafts at time of invitation. At least one invasive SCC was diagnosed in 185 (33.3%) of the invitees. Two SCC negative controls were matched by gender, year of birth (+/− 3 years), and duration of grafts (+/− 3 years) for each case according to the incidence density sampling method.22 Among those who eventually volunteered to participate (n = 217; 39.1%), SCC was diagnosed in 80 patients (36.9%). All study participants provided informed consent, delivered EDTA-blood for DNA analyses, and responded to a questionnaire reporting skin phototype, hair and eye color, and the presence of skin lesions considered as warts or “warty-like” lesions (Table 1). In short, the typing of skin refers to skin phototype 1 (SPT1) as white, always burns and never tans; SPT2 as having some tanning response, otherwise as SPT1; SPT3 have white skin with a gradual and moderate tanning potential with minimal burns; whereas SPT4 has light brown skin with good tanning response and minimal burns.23 The representation of transplant recipients with and without SCC were evenly distributed throughout the different therapeutic eras administrating azathiopurine (Aza), prednisone (Pred), cyclosporine A (CsA), mycophenolate mofetil, and tacrolimus (1968–1983; Aza+Pred; 1983–1985: CsA+Pred; 1985–1987: CsA+Pred, or Csa+Aza+Pred, randomized; 1987–2000: Csa+Aza+Pred; 2000-onwards: gradually conversion from Aza to mycophenolat mofetil, and CsA to tacrolimus). Initial immunosuppression is normally used during the grafts entire life. The study was approved by the Regional Ethical Committee and performed according to the Helsinki declaration.

Table 1

Clinical variables and phenotypic characteristics of renal transplant patients with (SCC positive) and without (SCC negative) squamous cell carcinoma of the skin.

CLINICAL VARIABLES	SCC POSITIVE (n = 87) % (n)		SCC NEGATIVE (n = 130) % (n)
Males (n)	63.2 (55)		62.3 (81)
Females (n)	36.8 (32)		37.7 (49)
Mean age at transplantation	Years	Range/St.dev.	Years	Range/St.dev.	p-value
All	45.0	13–74/16.2	45.1	6–80/14.9	0.94
Males	47.3	19–74/15.9	46.9	6–72/13.7	0.89
Females	41.1	13–70/16.3	42.3	14–80/16.4	0.75
Mean age at inclusion	Years	Range/St.dev.	Years	Range/St.dev.	p-value
All	56.4	26–75/11.4	56.8	32–89/10.7	0.77
Males	58.0	27–75/10.7	58.2	37–87/8.7	0.89
Females	53.6	26–74/12.3	54.6	32–89/13.1	0.75
Mean duration of graft	Years	Range/St.dev.	Years	Range/St.dev.	p-value
All	11.7	1–32/7.7	11.2	0–35/7.6	0.65
Males	11.1	1–30/7.6	10.6	1–35/7.3	0.70
Females	12.8	1–32/7.9	12.3	0–30/8.0	0.79
Phenotypes	n	%	n	%	p-value
Skin phototypes
1	6	7.0	13	10.1
2	18	20.9	17	13.2
3	51	59.3	70	54.3	0.15
4	11	12.8	29	22.5
Hair color
Dark Blond	40	46.0	66	51.2
Blond	30	34.5	43	33.3	0.56
Light blond	8	9.2	13	10.1
Red	9	10.3	7	5.4
Eye color
Blue	68	78.2	102	80.3	0.70
Non blue	19	21.8	25	19.7
Warts
Negative	12	14.8	47	38.2	0.00
Positive	69	85.2	76	61.8

MC1R DNA sequencing

The complete coding region of MC1R (NM_002386.3) was amplified and sequenced to anticipate MC1R variation in the Norwegian population (details available upon request to communicating author). Seven non-synonymous variants were identified, all previously recognized as common single nucleotide polymorphisms (SNPs) (Supplemental data; Table S1). These were the red hair (RHC) associated variants p.Asp84Glu (rs1805006), p.Arg151Cys (rs1805007), p.Arg160Trp (rs1805008), and p.Asp294His (rs1805009), and the non-red hair color (NRHC) variants p.Val60Leu (rs1805005), p.Val92Met (rs2228479), and p.Arg163Gln (rs885479).

The common SNPs of ASIP, TYR, and TYRP1 were analyzed on a MALDI-TOF mass spectrometry platform (Sequenom, San Diego, CA) platform, at Centre of Integrative Genetics (CIGENE) at the Norwegian University of Life Sciences, Ås, Norway (www.cigene.no). To assess the reproducibility of MALDI-TOF, all the initially sequenced RT patient samples were included and retested. Two intergenic SNPs near the ASIP locus, rs1015362 (G>A) and rs4911414 (G>T), define a specific ASIP haplotype (AH; alleles G and T, respectively). The TYR SNP results in an alternative amino acid at codon 402 (p.Arg402Gln; rs1126809), and the TYRP1 variant is an inter-gene nucleotide transition with unknown consequence (rs1408799; C>T). PCR and extension primers were designed using the Sequenom Spectro DESIGNER software (version 3.0).

Statistics

The statistical models were adjusted for gender, age at transplantation, and age at inclusion. Only the results from the unconditional analyses are presented as they resulted in essentially the same risk estimates as the conditional analyses. All analyses were conducted using Stata statistical software, release 11 (Stata Corporation; 2009. College Station, TX). Cross-tabulations were used to assess clinical (SCC versus non-SCC) relative to phenotypic categorization and carrier status. Two-sided P values less than 0.05 indicated statistical significance. Hardy Weinberg equilibrium (HWE) was assessed for MC1R genotype distribution among the participants (Supplemental data, Table S1). MC1R genotype categorization; Wild type: absence of any of the identified variants; Any variant: combinations of any or number of variants; 1 NRHC/RHC: NRHC/WT or RHC/WT; 2 NRHC/RHC: NRHC/NRHC, NRHC/RHC, or RHC/RHC; 1–2 NRHC: NRHC/WT or NRHC/NRHC; 1–2 RHC: RHC/WT or RHC/RHC. Odds ratio (OR) with a 95 % confidence interval (CI) according to multivariable unconditional logistic regression models were used to assess interactions between genotypes, phenotypic traits, and relative risk of SCC. Estimation of ASIP haplotype distribution (based on genotyping the rs1015362 and rs4911414 SNPs) was done by an expectation-maximization algorithm using the Haploview program.24

Results

Based on Hardy Weinberg equilibrium the entire group of RT patients appeared representative of the general (Norwegian) population (Supplemental data; Table S1). As presented in Table 1 the matching between RT cases with and without SCC appeared unaffected by the overall low response rate (39.1%). The representation of males dominated equally (~63%) in both groups. Mean ages at first transplantation (~45 years) and age of allograft (11–12 years) did not differ significantly between the groups. No significant associations between skin phototypes, hair and eye color, on one side, and SCC on the other, were observed.

Seven non-synonymous variants of MC1R were characterized by minor allele frequencies ranging from 1.6% (p.Asp84Glu) to 12.2% (p.Arg151Cys). The only variant apparently associated with risk of SCC specifically was the RHC allele p.Arg151Cys (OR = 1.99; CI: 1.05–3.75) (Supplemental data; Table S1). When adjusted for the concurrent presence of other MC1R variants or red hair, the basis for estimating the significance of p.Arg151Cys on an individual basis, diminished (not shown). Carriers of any variant or combination of variants implied a non-significant risk (OR = 1.85; CI: 0.92–3.69), while carriers of two variants reached a significant association (OR = 2.36; CI: 1.08–5.15) (Table 2). These estimates were unaffected when adjusted for eye, skin, or hair phenotypes (Supplemental data; Table S2). However, when stratified by phenotypic traits (Table 3), a significant elevation in SCC risk was observed in carriers of any MC1R variant combination with the darker skin phototype (SPT3) (OR = 3.94; CI: 1.37–11.30). Against this background, it appeared sufficient to carry one of any variant combined with the wild type allele (OR = 3.48; CI: 1.14–10.60); not differing significantly from those being carriers of two MC1R variants of any type (OR = 4.62; CI: 1.47–14.60) (Table 3 and Supplemental data; Table S3). When stratified by hair color, two of any of the MC1R variants indicated a higher risk of SCC in blond haired individuals (OR = 10.50; CI: 1.86–59.27). Assessing the NRHC and RHC genotype groups individually revealed that only carriers of NRHC alleles reached significance (OR = 7.29; CI: 1.39–38.20) (Table 3). All red-haired individuals (n = 16) were consistently positive for at least one RHC variant and negative for any NRHC variant; a representation observed evenly distributed between those with and without SCC (Supplemental data; Table S4). When stratified by eye color, carriers of two of any MC1R variant and carriers of 1–2 NRHC variants indicated an increased risk within the blue-eyed group of patients (OR = 2.80; CI: 1.15–6.83, and OR = 2.50; CI: 1.02–6.16, respectively) (Table 3 and Supplemental data; Table S5). The presence of self-reported warts correlated with a relatively high SCC risk apparently independent of MC1R (Tables 1 and 3, and Supplemental data; Table S6). This was consistent with the observed independence between MC1R and warts (Supplemental data; Table S7).

Table 2

Odds ratio (OR) with 95% confidence intervals (CI) for risk of SCC in renal transplant patients related to MC1R variation (RHC variants: p.Asp84Glu, p.Arg151Cys, p.Arg160Trp, and p.Asp294His. NRHC-variants: p.Val60Leu, p.Val92 Met, and p.Arg163Gln). Adjustments for phenotypic traits are provided in supplemental data (Table S2).

VARIABLE	SCC POSITIVE n (%)	SCC NEGATIVE n (%)	ORa	95% CI
Wild type	14 (16.1)	34 (26.2)	1.00
Any variant	73 (83.9)	96 (73.8)	1.85	0.92–3.69
Number of variants
1 NRHC/RHC	39 (44.8)	61 (46.9)	1.55	0.74–3.26
2 NRHC/RHC	34 (39.1)	35 (26.9)	2.36	1.08–5.15
Type of variant
1–2 NRHC	31 (36.6)	39 (30.0)	1.93	0.88–4.21
1–2 RHC	42 (48.3)	57 (43.8)	1.79	0.85–3.75

Significant odds ratios are indicated by bold numbers.

Table 3

Associations between MC1R genotypes and SCC risk reflected by odds ratios after stratifications by phenotypic traits.

PHENOTYPIC TRAITS	SKIN PHOTOTYPE				HAIR COLOR				EYE COLOR		WARTS (UNSPECIFIED)
	1	2	3	4	DARK BLOND	BLOND	LIGHT BLOND	RED	BLUE	NON-BLUE	NEGATIVE	POSITIVE
SCC/non-SCCa	6/13	18/17	51/70	11/29	40/46	30/43	8/13	9/7	68/102	19/25	12/47	69/76
OR 95% CIb
Wild type	1.00	1.00	1.00	1.00	1.00	1.00	1.00	–	1.00	1.00	1.00	1.00
Any variant	0.360.04–3.52	1.080.22–5.22	3.941.37–11.30	0.850.18–4.10	1.500.60–3.72	5.420.11–26.40	0.300.04–2.42	–	2.190.99–4.88	0.940.21–4.10	1.350.25–7.19	3.081.31–7.21
1 NRHC/RHC	0.430.04–4.64	0.670.12–3.75	3.481.14–10.60	0.930.18–4.86	1.410.54–3.69	3.650.71–18.79	0.400.04–3.96	–	1.830.78–4.30	0.800.17–3.82	0.800.13–5.08	2.841.14–7.05
2 NRHC/RHC	0.250.01–4.73	2.000.32–12.10	4.621.47–14.60	0.670.08–5.30	1.710.55–5.34	10.501.86–59.27	0.220.02–2.45	–	2.801.15–6.83	1.250.22–7.08	2.500.41–15.23	3.431.32–8.91
1–2 NRHC	0.250.01–4.73	2.000.22–17.90	3.831.23–12.00	1.040.17–6.23	1.440.50–4.14	7.291.39–38.20	0.170.01–2.82	–	2.501.02–6.16	0.830.16–4.44	0.710.09–5.96	3.471.32–9.08
1–2 RHC	0.430.04–4.64	0.910.18–4.64	4.041.32–12.40	0.720.12–4.16	1.540.57–4.14	3.880.73–20.80	0.380.04–3.34	–	2.010.86–4.70	1.020.21–4.98	0.220.04–1.27	3.651.39–9.61

SCC/non-SCC: Number of patients within the SCC positive (SCC) and negative (non-SCC) groups with informative genotypes.

Significant odds ratio estimates indicated in bold numbers.

In the SCC-positive and -negative groups, 80 to 86, and 118 to 128, respectively, were informative for the assessment of ASIP, TYR, and TYRP1 variants relative to SCC (Supplemental data; Table S8). None of the individual ASIP polymorphisms were significantly associated with SCC, although an overrepresentation was observed for the ASIP G>T transversion (rs4911414) in the SCC group. The non-synonymous TYR (p.Arg402Gln) and TYRP1 variants appeared insignificant in imposing risk (Supplemental data; Table 8). Seven of nine possible ASIP genotypes were observed of which the estimated ASIP haplotype (AH) representation tended towards increased risk of SCC (OR = 1.87; CI: 0.91–3.83) (Table 4 and Supplemental data; Table S8).

Table 4

The significance of the ASIP haplotype (AH) relative to SCC risk as indicated by odds ratio (OR) with 95% confidence intervals in RT patients with (SCC) and without (Non-SCC) squamous cell carcinoma.

ASIP HAPLOTYPE (AH)a	SCC-POSITIVE n = 160	%	SCC-NEGATIVE n = 236	%	OR	95% CI
Non-AH	142	88.8	221	93.6	1.0
AH	18	11.2	15	6.4	1.87	0.91–3.83

Background data are given under Supplemental data (Table S8).

Discussion

Among the more than 100 genes implicated in pigmentation, the key signaling regulator MC1R, its antagonist ASIP, and the downstream melanization regulatory genes TYR and TYRP1, remain the most extensively studied in relation to skin cancer.25–27 Here we demonstrate that MC1R variation has a significant impact on risk of developing SCC in RT patients independent of conventional risk phenotypes. Carriers of the RHC variant p.Arg151Cys, or carriers of two MC1R variants independent of being represented by NRHC or RHC alleles, indicated a significant risk of SCC. After stratifications by phenotypic traits the significance of MC1R was supported even further as traits traditionally associated with skin cancer protection were shown inferior to the impact of MC1R. None of the other pigmentation-associated genes were found to have a significant influence on SCC risk in RT patients.

No more than two of any of the MC1R variants were carried simultaneously among the participants in this study. This is in line with previous population- and haplotyping-based studies indicating lack of linkage disequilibrium between Caucasian-prevalent MC1R variants.26,28 In accordance with an anticipated south to north gradient in Western Europe, Norway appears to have the highest frequencies of common European MC1R alleles.28,29 Because of the high MC1R variability and the low allele frequencies for some of the variants, the categorization into genotypes related to strong or weak associations with risk phenotypes (RHC and NRHC, respectively), seemed appropriate. The risk of developing SCC in the general population has been evaluated based on similar genotype categorizations.26,30

The impact on SSC risk appears to depend on combinations of alleles rather than on the individual allele or whether the allelic variants are associated with a particular phenotype. This may explain the discordance between SCC risk and traditional risk phenotypes. When stratified by skin phenotypes, patients with the darker skin phototype SPT3 had significant and similar odds ratios for all allelic combinations when compared to wild type MC1R. Being a carrier of any of these MC1R genotypes appeared to negate the protection normally afforded by darker skin. In the general population, individuals with darker or olive colored skin being carriers of loss-of-function MC1R have been shown to be at an increased risk of developing skin cancer.10,11,31 The influence of NRHC variants was prominent when stratified by hair color. Initially, when assessing the distribution of the individual NRHC variants independent of phenotypes, only p.Arg163Gln was overrepresented in the SCC group, but without indicating significant risk. The other two (p.Val60Leu and p.Val92Met) were more or less equally distributed between those with and without SCC. This suggests that the effects on risk associate with variable dominant negative action where the impact of the individual NRHC alleles are masked and affected by the combination of alleles.

Unless stratified by phenotypic traits only the strongly red hair associated p.Arg151Cys variant and a dual representation of any MC1R variants were found associated with SCC. This is consistent with the only previous report that to our knowledge addresses the correlation between MC1R and SCC in RT patients.17 However, the significance of p.Arg151Cys is obscured when adjusting for the concurrent presence of other MC1R variants or red hair; again pointing towards the necessity of assessing the overall MC1R genotype. In the general population a three-fold increase in risk has been indicated in carriers of NRHC/RHC compound heterozygotes followed by a two-fold increase in heterozygous WT/NRHC and WT/RHC carriers, and with the least impact on risk associated with homozygous representation of variants, all independent of traditional risk phenotypes.26 Signaling conveyed by dimeric and oligomeric proteins like MC1R is vulnerable to structural changes caused by non-synonymous polymorphisms in the parental alleles. Such polymorphisms apparently have an effect on signaling because of variation in cell surface expression and density, organelle retention affecting trafficking, and G-protein coupling.12,16,32–34 The considerable variation in residual signaling generated from the numerous allelic combinations complicates the establishment of meaningful correlations between MC1R variation and phenotypes.15 Similar challenges in assessing risk of SCC are conceivable. Also because RT recipients are treated with combinations of immunosuppressive drugs at different strengths and that these regimens have changed over time, the impact of changes in therapeutic conditions on cancer risk is difficult to assess.

The ASIP haplotype (AH) based on two specific polymorphisms tended towards an increased SCC risk. However, it appeared that only one of the polymorphisms (ASIP G>T; rs4911414) contributed to this tendency. An overrepresentation was observed for this variant among the ASIP genotypes and estimated haplotypes in the SSC group of patients (Supplemental data; Table S8). This particular polymorphism has been associated with increased SCC risk in the general population.30 A strong association between AH, red hair, and reduced tanning ability supports gain-of-function for this haplotype with the potential of inhibiting MC1R signaling.20,35 A broader systematic study is necessary to conclude on the significance of this haplotype relative to SCC risk in RT patients. The Gln-allele of the TYR p.Arg402Gln variant was overrepresented in the SCC positive patients but without reaching a significant association. A haplotype based on p.Arg402Gln and another TYR variant (p.Ser192Tyr) has been associated with SCC risk in the general population and should be assessed together in future studies.30 The lack of a correlation between SCC and TYRP1 was consistent with observations in the general population.20,21

Warts reported by the patients themselves appeared significantly associated with SCC. No dosage effect was observed when estimated in concert with MC1R variants indicating that these lesions and MC1R are mutually independent. This was also supported by the observation that MC1R apparently does not modulate risk of contracting warts (Supplemental data; Table S7). As self-reported, no reference to histological classification exists and the correlations should therefore be evaluated accordingly. A recent study of organ transplant patients demonstrated a significant association between rare subtypes of warts like verrucokeratotic lesions and verrucous papilloma and SCC, but not with the more common verrucae vulgares and flat warts.36 Solar keratosis is an indicator of excessive sun exposure and SCC risk in the immunocompetent population where the majority of keratotic lesions regress spontaneously.37 Whether the potential of regression in immunocompromised patients is affected by marginal or substantial differences in keratinocyte differentiation imposed by MC1R remains to be seen. To address this issue further a histological differentiation between SCC and other possible non-pigmented epithelial tumors of the skin is necessary.38

The increasing incidence of skin cancers and the general lack of public awareness encourage the institution of guidance protocols for supervising patients and health personnel about SCC risk factors and precautionary measures. The criteria for stratifying risk in relation to cumulative sun-exposure, pigmentation, individual and family history of solar-associated skin cancer, are uncontroversial.39 Before implementing genetic testing of candidate genes like MC1R as an integral part of risk assessment, broader and more systematic studies searching meaningful correlations between MC1R variation, phenotypes, and risk of histological verified SCC, are essential. At this point a sober message to be carried forward is that traditionally low and high risk phenotypes do not necessarily imply a reduction or an increase in SCC risk, respectively. The modest statistical power in this study as indicated by wide confidence intervals warrants cautious consideration of the presented data. However, we conclude that a correlation exists between MC1R genotype variation and the risk of developing SCC in RT patients independent of traditional risk phenotypes. The current data do not allow an explicit differentiation between the SCC risk in immunocompromised patients and the risk in the general population.

Supplement table 1. The distribution and frequencies of MC1R alleles and genotypes in all renal transplant (RT) patients (All), with (SCC positive), and without (SCC negative) squamous cell carcinoma, and estimated odds ratios (OR) with 95% confidence intervals (95% CI) as indicators of SCC risk.

Supplement table 2. Adjustments for phenotypic traits and odds ratio (OR) with 95% confidence intervals (CI) indicative of SCC in RT patients related to MC1R.

Supplement table 3. Stratifications by skin phototypes (SPT); Associations between MC1R genotypes and SCC risk indicated by odds ratio (OR) with 95% confidence intervals (95% CI).

Supplement table 4. Stratifications by hair color; Associations between MC1R genotypes and SCC risk indicated by odds ratio (OR) with 95% confidence intervals (95% CI).

Supplement table 5. Stratifications by eye color; Associations between MC1R genotypes and SCC risk indicated by odds ratio (OR) with 95% confidence intervals (95% CI).

Supplement table 6. Relation between the presence of wart-like lesions (WLL) and risk of SCC as reflected by odds ratios (OR) and 95% confidence intervals (95% CI) independent (A) and dependent on MC1R genotypes (B).

Supplement table 7.MC1R genotypes related to the incidence of self-reported warty-like lesions (WLL).

Supplement table 8. Odds ratios with confidence intervals after distribution of specific ASIP, TYR, and TYRP1 alleles (A), observed ASIP genotypes (B), and estimated ASIP haplotypes (C) in patents informative for assessment (All), SCC positive (SCC), and SCC negative (Non-SCC) patients.