eTable 1. VDR Polymorphism Primer Details
eTable 2. Genotype and Allele Frequencies of the VDR Polymorphisms in Men
eTable 3. Genotype and Allele Frequencies of the VDR Polymorphisms in Women
eTable 4. Genotype Frequencies by Non-Melanoma Skin Cancer Type
Customize your JAMA Network experience by selecting one or more topics from the list below.
Burns EM, Guroji P, Ahmad I, et al. Association of Vitamin D Receptor Polymorphisms With the Risk of Nonmelanoma Skin Cancer in Adults. JAMA Dermatol. 2017;153(10):983–989. doi:10.1001/jamadermatol.2017.1976
Is there an association between vitamin D receptor single-nucleotide polymorphisms and the risk of nonmelanoma skin cancer?
This case-control study involving 100 case patients and 100 control patients proposed a model for predicting the incidence of skin cancer and found that individuals with the BsmI single-nucleotide polymorphism were twice as likely to develop nonmelanoma skin cancer than those with no mutation.
A screening for the BsmI single-nucleotide polymorphism may emphasize the need for skin cancer prevention for individuals with this polymorphism.
Protective effects of UV-B radiation against nonmelanoma skin cancer (NMSC) are exerted via signaling mechanisms involving the vitamin D receptor (VDR). Recent studies have examined single-nucleotide polymorphisms (SNPs) in the VDR, resulting in contradictory findings as to whether these polymorphisms increase a person’s risk for NMSC.
To examine whether the polymorphisms in the VDR gene are associated with the development of NMSC and the demographic characteristics of the participants.
Design, Setting, and Participants
This case-control study recruited 100 individuals who received a diagnosis of and were being treated for basal cell carcinoma or squamous cell carcinoma (cases) and 100 individuals who were receiving treatment of a condition other than skin cancer (controls) at the dermatology clinics at the Kirklin Clinic of the University of Alabama at Birmingham Hospital between January 1, 2012, and December 31, 2014. All participants completed a questionnaire that solicited information on skin, hair, and eye color; skin cancer family history; and sun exposure history, such as tanning ability and number of severe sunburns experienced throughout life. Blood samples for DNA genotyping were collected from all participants.
Main Outcomes and Measures
Polymorphisms in the VDR gene (ApaI, BsmI, and TaqI) were assessed to determine the association of polymorphisms with NMSC development and demographic characteristics. χ2 Analysis was used to determine whether genotype frequencies deviated significantly from Hardy-Weinberg equilibrium. Logistic regression was used to calculate odds ratios (ORs) and associated 95% CIs for the identification of factors associated with NMSC diagnosis. A model was created to predict NMSC diagnoses using known risk factors and, potentially, VDR SNPs.
A total of 97 cases and 100 controls were included. Of the 97 cases, 68 (70%) were men and 29 (30%) were women, with a mean (SD) age of 70 (11) years. Of the 100 controls, 46 (46%) were men and 54 (54%) were women, with a mean (SD) age of 63 (9) years. All participants self-identified as non-Hispanic white. A model including age, sex, and skin color was created to most effectively predict the incidence of skin cancer. Risk factors that significantly increased the odds of an NMSC diagnosis were light skin color (OR, 5.79 [95% CI, 2.79-11.99]), greater number of severe sunburns (OR, 2.59 [95% CI, 1.31-5.10]), light eye color (OR, 2.47 [95% CI, 1.30-4.67]), and less of an ability to tan (OR, 2.35 [95% CI, 1.23-4.48]). The risk factors of family history of NMSC (OR, 1.66 [95% CI, 0.90-3.07]) and light hair color (OR, 1.17 [95% CI, 0.51-2.71]) did not reach statistical significance. Participants with the BsmI SNP were twice as likely to develop NMSC than participants with no mutation (OR, 2.04 [95% CI, 1.02-4.08]; P = .045).
Conclusions and Relevance
The results of this study are especially useful in the early treatment and prevention of NMSC with chemopreventive agents (for those with the BsmI SNP). A screening for the BsmI SNP may emphasize the importance of sun protection and facilitate skin cancer prevention and, therefore, decrease the skin cancer burden.
More than 3.5 million new diagnoses of nonmelanoma skin cancer (NMSC) are reported by 2 million Americans each year.1 While the prevalence of cutaneous squamous cell carcinoma (SCC) is only about 16% of all skin cancers, SCC can metastasize, killing 4000 to 8000 people each year.2 Basal cell carcinoma (BCC) is the most common skin cancer and can be highly disfiguring if left untreated. Furthermore, advanced BCC contributes to 3000 deaths annually in the United States. Recent studies have reported that NMSC is the fifth most costly cancer to treat in the Medicare population.3
Risk factors for NMSC include white race, male sex, and sun exposure.4 Epidemiological studies have shown that men have twice the risk for BCC and 3 times the risk for SCC than women do.5 UV-B radiation (290-320 nm) is the major etiologic agent for NMSC.6 On the other hand, UV-B induces cutaneous vitamin D production. Vitamin D has been shown to exert protective effects in several cancer types,7-9 including NMSC,10 via an unknown signaling mechanism involving the vitamin D receptor (VDR).
Polymorphisms (DNA sequence variations) occur frequently in the population and may explain individual variation in disease risk. More than 60 single-nucleotide polymorphisms (SNPs) in the VDR have been detected in or near exons 2 to 9, as well as in the 3′ UTR (untranslated) region of the promoter. Owing to the large size of this gene and the difficulty in analyzing it, the most extensively studied VDR SNPs in the context of skin cancers are TaqI, BsmI, FokI, and ApaI.11TaqI (rs731236) occurs at codon 352 in exon 9 of the VDR gene (NM_000376) and is a restriction fragment length polymorphism. The products are digested into fragments on the basis of the presence or absence of a TaqI restriction site in each allele. In the absence of a restriction site, there will be 2 fragments (T allele); in the presence of a restriction site, there will be 3 fragments (t allele). The TaqI polymorphism results in a silent codon change from ATT to ATC, which both result in an isoleucine.12 The BsmI (rs1544410) and ApaI (rs7975232)13 polymorphisms are located at the 3′ end of the gene in intron 8. BsmI, like TaqI, is considered to be silent because it does not change the amino acid sequence.14 However, the BsmI SNP is able to affect messenger RNA (mRNA) stability and gene expression.15
Previous studies have examined SNPs (ApaI, BsmI, and TaqI) in the VDR,11,16,17 resulting in contradictory findings as to whether these polymorphisms increase risk for NMSC. Previous studies indicated that the heterozygous genotype for the TaqI polymorphism may be associated with an increased risk for the development of BCC.17,18 Few studies have examined the relationship between SNPs and known skin cancer risk factors. One study found no correlation between SNPs and the known NMSC risk factors of light skin, hair, or eye color,17 whereas another study did not examine these variables.18 Another study found a significant association between the wild type, homozygous genotype for the BsmI VDR polymorphism and SCC risk.16
Because most of these studies were carried out in Europe, where sun exposure is moderate, we set out to study a US population with a high mean UV exposure index19 and, therefore, a high risk for the development of NMSC. Homburg, Germany, and Lodz, Poland, have a yearly mean solar-radiation exposure of 778 J/cm2, whereas Birmingham, Alabama (in the United States) has a yearly mean sun radiation exposure of 1711 J/cm2. We examined 3 polymorphisms in the VDR (ApaI, BsmI, and TaqI) to determine the association of polymorphisms with NMSC development and demographic characteristics.
The patient group consisted of 100 participants with a BCC or SCC diagnosis who were visiting the dermatology clinics at the Kirklin Clinic of the University of Alabama at Birmingham Hospital between January 1, 2012, and December 31, 2014 (Table 1). This group comprised non-Hispanic white men (68 [70%]) and women (29 [30%]) with a mean (SD) age of 70 (11) years; 3 patients were excluded. One hundred non-Hispanic white men (46 [46%]) and women (54 [54%]), with a mean (SD) age of 63 (9) years, who visited these clinics for purposes other than skin cancer were recruited as the control group (Table 1). All participants completed a questionnaire that solicited information on skin, hair, and eye color; skin cancer family history; and sun exposure history, such as tanning ability and number of severe sunburns experienced throughout life. Blood samples for DNA genotyping were collected from all participants. This study was approved by the institutional review board of the University of Alabama at Birmingham Hospital. All participants provided written informed consent prior to entering the study.
DNA was extracted from blood samples using commercially available blood kits (QIAamp DNA Blood Mini Kit; Qiagen) and following the manufacturer instructions. The presence of the 3 VDR polymorphisms was assessed using polymerase chain reaction (eTable 1 in the Supplement) and Sanger sequencing (samples processed were based on the protocol of BigDye Terminator v3.1 Ready Reaction kit [Thermo Fisher Scientific]). Sequences for polymorphisms were examined using 4peaks software, version 1.7 (Nucleobytes). Approximately 15% of the samples were sequenced a second time, with the initial and repeated results in total agreement, demonstrating the reliability of our sequencing method.
For purposes of analysis, self-reported skin color was divided into 2 groups, with ivory and pale skin defined as the “light” group and beige, olive, and dark brown skin as the “dark” group. Age was categorized into quartiles (50-58, 59-65, 66-73, and 74-91 years). New SNP variables were created for Apa, Bsm, and Taq, with the genotype of homozygous/no SNP in one group and the genotypes of heterozygous/SNP and homozygous/SNP in the other group.
Univariate analysis was conducted on all variables of interest to assess for normality. To statistically compare the association between the factors of interest and NMSC diagnosis, the unpaired, 2-tailed t test or analysis of variance was used for continuous variables and χ2 analysis was used for categorical variables—specifically, to determine whether genotype frequencies deviated significantly from Hardy-Weinberg equilibrium. Logistic regression was used to calculate odds ratios (ORs) and associated 95% CIs for identifying factors associated with NMSC diagnosis. Following the example of Han et al,16 we set out to create a model for predicting NMSC diagnoses using known risk factors and, potentially, VDR SNP information.16 The logistic model was constructed using the categorical variables already mentioned and a backward selection process with a significance level retained at 0.1. For analyses, 2-sided P < .05 was considered statistically significant, and SAS software, version 9.3 (SAS Institute Inc), was used.
Initial analyses examining demographic factors included all cases (n = 97) and controls (n = 100). Three case-eligible participants were excluded owing to the poor quality of their samples, which did not provide valid sequencing data. Two of the known risk factors for NMSC, older age (70 years for cases vs 63 years for controls; P < .001) and male sex (70% men vs 46% women; P = .001), were significantly different between cases and controls. Because some participants did not complete the questionnaires, the number of participants for some comparisons was less than the total number of participants in the study. After adjusting for age and sex, we found that the following risk factors significantly increased the odds of an NMSC diagnosis: light skin color (OR, 5.79 [95% CI, 2.79-11.99), greater number of severe sunburns (OR, 2.59 [95% CI, 1.31-5.10]), light eye color (OR, 2.47 [95% CI, 1.30-4.67]), and lesser ability to tan (OR, 2.35 [95% CI, 1.23-4.48]) (Table 2). The risk factors of family history of NMSC (OR, 1.66 [95% CI, 0.90-3.07]) and light hair color (OR, 1.17 [95% CI, 0.51-2.71]) did not reach statistical significance.
We performed separate analyses for men and women on the basis of the difference in the numbers of cases. Adjusting for age and other factors in the model, we found that the following risk factors significantly increased the odds of an NMSC diagnosis in men (Table 3): light eye color (OR, 3.26 [95% CI, 1.24-8.56]), light skin color (OR, 4.10 [95% CI, 1.46-11.47]), and lesser ability to tan (OR, 2.31 [95% CI, 0.87-6.12]). Adjusting for age and other factors in the model, we found the following risk factors significantly increased the odds of an NMSC diagnosis in women (Table 4): light skin color (OR, 8.43 [95% CI, 1.70-41.86]) and greater number of severe sunburns (OR, 8.00 [95% CI, 2.01-31.84]).
The frequencies for the selected VDR SNPs’ alleles and genotypes are shown in Table 5. Genotype distributions for all polymorphisms were found to be in Hardy-Weinberg equilibrium (P = .16 for ApaI; P = .26 for BsmI; and P = .35 for TaqI). Based on the genotype distributions, only the ApaI SNP differed significantly between the cases and controls (Table 3). The ApaI A allele was significantly more frequent among controls than cases (0.500 vs 0.405; P = .01), whereas the BsmI b allele was significantly less frequent among controls than cases (0.360 vs 0.464; P = .04). The TaqI t allele was also less frequent among controls, although not significantly so (0.345 vs 0.382; P = .09).
Individuals with the A genotypes were only about half as likely to develop NMSC (OR, 0.45 [95% CI, 0.25-0.84]). However, after adjusting for age and sex in the model, we found that individuals with the A genotypes were 0.57 times as likely to develop NMSC (OR, 0.57 [95% CI, 0.29-1.10]), which was not statistically significant. In contrast, the b (Bb + bb; P = .04) and t (Tt + tt; P = .09) genotypes were less frequent among controls than cases, demonstrating that individuals with the b or t genotypes, respectively, were 1.86 or 1.66 times as likely to develop NMSC as were individuals with B (OR, 1.86 [95% CI, 1.03-3.36]) or T (OR, 1.66 [95% CI, 0.93-2.96]) genotypes. After adjusting for age and sex in the model, we found that neither B genotypes (OR, 1.65 [95% CI, 0.87-3.11]) nor T genotypes (OR, 1.50 [95% CI, 0.81-2.81]) reached statistical significance. We performed separate analyses for men (eTable 2 in the Supplement) and women (eTable 3 in the Supplement) based on the difference in the numbers of cases, finding no significant differences in allele or genotype frequencies in the sexes separately.
The frequencies for the selected VDR polymorphism genotypes by NMSC type are shown in eTable 4 in the Supplement. We found no significant differences in genotype distribution within each NMSC type (BCC, SCC, BCC and SCC), between BCC and SCC, or among all 3 types.
Our final model included the following parameters (C = 0.799): age (OR, 1.18 [95% CI, 0.49-2.85] for 59-65 years vs 50-58 years; OR, 1.49 [95% CI, 0.58-3.80] for 66-73 years vs 50-58 years; OR, 6.20 [95% CI, 2.20-17.48] for 74-91 years vs 50-58 years; P = .003); sex (OR, 2.89 [95% CI, 1.45-5.61] for male vs female; P = .002); skin color (OR, 5.79 [95% CI, 2.79-11.99] for light skin vs darker skin; P < .001); and BsmI SNP status (OR, 2.04 [95% CI, 1.02-4.08] for SNP vs no SNP; P = .045).
The final model can be described by the following equation: ln [odds (NMSC)] = 0.16 (59-65 years of age) + 0.40 (66-73 years of age) + 1.82 (≥74 years of age) + 1.16 (male) + 1.79 (light skin color) + 0.71 (BsmI SNP) − 2.9135.
For comparison purposes, individuals older than 73 years of age were about 6 times as likely to develop NMSC as were individuals 58 years of age or younger (OR, 6.20 [95% CI, 2.20-17.48]; P = .003). Men were about 3 times as likely to develop NMSC as were women (OR, 3.17 [95% CI, 1.60-6.29]; P < .001). Individuals with light skin color were about 6 times as likely to develop NMSC as were individuals with dark skin (OR, 5.97 [95% CI, 2.84-12.53]; P < .001). Individuals with a BsmI SNP were twice as likely to develop NMSC as were individuals with no mutation (OR, 2.04 [95% CI, 1.02-4.08]; P = .045).
In a population of patients with a high mean UV exposure index,19 this study examined 3 known VDR polymorphisms—ApaI, BsmI, and TaqI. We found that the known NMSC risk factors of older age, male sex, light eye color, light skin color, lesser ability to tan, and greater number of lifetime severe sunburns increased the odds of developing NMSC by 2 to nearly 6 times. Individuals with ApaI A genotypes were 0.45 times as likely to develop NMSC, whereas individuals with BsmI b or TaqI t genotypes were 1.86 and 1.66 times as likely to develop NMSC. This finding suggests that the alleles A, B, and T may contribute to decreased carrier risk of developing NMSC, whereas a, b, and t may contribute to increased risk. We found no significant differences in the VDR SNP genotype distribution among NMSC types. Our final model for predicting NMSC diagnoses included the known NMSC risk factors of older age, male sex, and light skin color; including BsmI SNP status contributed significantly to the model, with individuals with the SNP being twice as likely to develop NMSC, whereas including ApaI and TaqI SNP status did not.
Although the known NMSC risk factors of older age, male sex, and light skin color were included in the final model, other risk factors were not significant in the study. Blonde and red hair colors are usually included in NMSC risk factors, but, interestingly, the distribution of hair color between cases and controls was not significantly different in this study. This finding could be due to errors in the questionnaire answers or could highlight an area for future investigation into the combined effects of hair and skin colors. A family history of skin cancer may increase individual risk for a skin cancer diagnosis. In this study, 58 case participants (29%) were unsure whether a family member had a skin cancer diagnosis, leading to inconclusive results.
Previous studies examining the association of VDR polymorphisms with skin cancer risk have reached various conclusions. These differences can be attributed to variations in populations, types of skin cancer, and types of data collected. Consistent with our results suggesting that the BsmI B allele may be protective against the development of NMSC, previous studies have found that this allele is protective against melanoma20 and reduces cancer risk at any site, with great reductions in risk observed in the skin.21 In contrast, the BsmI BB genotype was found to be significantly associated with increased SCC risk, with an interaction observed between this genotype and high vitamin D intake.16
Several studies have investigated the potential interactions between VDR polymorphisms and vitamin D levels. Zeljik et al22 found that the TaqI genotypes Tt and tt were significantly associated with increased melanoma risk, but no association with clinical or pathological characteristics was observed. Furthermore, no associations were found with vitamin D serum levels.
Zhao et al23 demonstrated that TaqI and ApaI polymorphisms significantly increased the risk of skin cancer, but our study showed that when adjusted for age and sex, the TaqI and ApaI polymorphisms did not achieve statistical significance. The only SNP that significantly contributed to the final model was BsmI. This SNP may be contributing to skin cancer risk in that it is able to affect mRNA stability and gene expression,15 thus promoting the development of skin cancer.
The present study did not find any significant differences in any of the SNP genotypes between NMSC types. Köstner et al18 found an association of genotype combination AaTtBb with BCC risk, and the genotype combination aaTTbb was associated with both BCC and SCC. The authors concluded that ApaI and TaqI may be important for BCC risk but not SCC risk. UV-B exposure is a risk factor for both BCC and SCC, but basal cell nevus syndrome or Gorlin syndrome (a rare congenital condition) significantly increases the risk of the development of BCC, which often begins in childhood or adolescence.
The limitations of the present study include participant recall bias and lack of correlative vitamin D intake or circulating plasma levels. Because the cases and controls provided information regarding skin cancer diagnosis, sunburns, and tanning ability, recall bias could exist. We were not able to examine vitamin D intake or circulating plasma vitamin D levels. Therefore, no correlation with VDR polymorphisms could be investigated. It has been suggested that it is difficult to reach conclusions regarding VDR SNPs and skin cancer risk because of potential interactions with calcium, 25-hydroxyvitamin D plasma levels, vitamin D intake, and UV exposure.24 We did not determine the human immunodeficiency virus status or immunosuppression status for organ transplant for these participants, which can be a major confounder for skin cancer. We also did not assess the use of systemic medication, which can affect the results of this study. Further studies with larger sample sizes, verified NMSC diagnoses, immune status, medication use, and correlative vitamin D measures are necessary to more fully investigate the association of vitamin D polymorphisms with NMSC risk.
This study, which involved a population with a high mean UV exposure index, found that known NMSC risk factors, such as older age, male sex, light eye color, light skin color, lesser ability to tan, and greater number of lifetime severe sunburns, increased the odds of an NMSC diagnosis by 2 to nearly 6 times. Individuals with ApaI A genotypes were only half as likely to develop NMSC, whereas individuals with BsmI b or TaqI t genotypes were 1.86 or 1.66 times as likely to develop NMSC, respectively. This finding suggests that alleles A, B, and T may contribute to decreased carrier risk of developing NMSC and that alleles a, b, and t may contribute to increased risk. We found no significant differences in VDR SNP genotype distribution when skin cancers were subcategorized as BCC and SCC. The final model that we developed for predicting NMSC diagnoses included age, sex, and skin color, and individuals with the BsmI SNP were twice as likely to develop NMSC. This information can be especially useful in the early treatment and prevention of NMSC given that individuals presenting with actinic keratosis may need a more aggressive form of treatment with chemopreventive agents if they have the BsmI SNP. A family history of skin cancer was not found to be significant in our model, but if an individual has known skin cancer risk factors in combination with family members with skin cancer diagnoses, a screening for the BsmI SNP may emphasize the importance of sun protection and facilitate skin cancer prevention in that individual, which will therefore decrease the skin cancer burden.
Corresponding Author: Nabiha Yusuf, MS, PhD, Department of Dermatology, University of Alabama at Birmingham, 1670 University Blvd, VH 566A, PO Box 202, Birmingham, AL 35294-0019 (email@example.com).
Accepted for Publication: April 29, 2017.
Published Online: August 23, 2017. doi:10.1001/jamadermatol.2017.1976
Author Contributions: Drs Burns and Griffin had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Burns, Bush, Hurst, Griffin, Elmets, Yusuf.
Acquisition, analysis, or interpretation of data: Burns, Guroji, Ahmad, Nasr, Wang, Tamimi, Stiefel, Abdelgawwad, Shaheen, Muzaffar, Bush, Griffin, Elmets, Yusuf.
Drafting of the manuscript: Burns, Nasr, Wang, Tamimi, Stiefel, Shaheen, Muzaffar, Hurst, Griffin, Elmets, Yusuf.
Critical revision of the manuscript for important intellectual content: Burns, Guroji, Ahmad, Abdelgawwad, Bush, Griffin, Yusuf.
Statistical analysis: Burns, Tamimi, Griffin.
Obtained funding: Elmets.
Administrative, technical, or material support: Guroji, Ahmad, Nasr, Wang, Tamimi, Stiefel, Abdelgawwad, Shaheen, Bush, Yusuf.
Study supervision: Tamimi, Hurst, Yusuf.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported in part by the University of Alabama at Birmingham (UAB) Department of Dermatology, by grant P30 CA13148 from the UAB Comprehensive Cancer Center, and by training grant R25 CA47888 from the National Institutes of Health Division of Cancer Prevention and Control (Dr Burns).
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: J. W. Waterbor, MD, DrPH, University of Alabama at Birmingham, provided critical review of the manuscript. Dr Waterbo was not compensated for his contribution.
Create a personal account or sign in to: