In the current study, we combined data from three case-control studies and one family study conducted in Italy, and examined the associations of 39 genes related to the biology of telomeres, and 475 SNPs within these genes, with the risk of melanoma, presence of dysplastic nevi, number of nevi and telomere length.
The gene regions around RECQL4, RTEL1 and TERF2 were associated with melanoma, the presence of dysplastic nevi and number of nevi, respectively, in the Italian combined study. However, we did not find the same associations in two American studies with different study designs.
Previously, among the telomere-related genes, two SNPs in TERT, rs2853676 and rs2242652, were reported to have a significant association with melanoma in a Caucasian population 
. However, these SNPs were not associated with melanoma in our population (OR
1.17; 95% CI: 0.84, 1.63 for rs2853676; OR
0.96; 95% CI: 0.66, 1.39 for rs2242652). Furthermore, the same study also reported two SNPs in TERF2 (rs153045 and rs251796) that were significantly correlated with number of nevi, but they did not report the size of the association after adjustment for confounding variables. We found that these two SNPs were significantly associated with higher number of nevi in our study. In the gene-based analysis, the TERF2 region had the strongest association with nevus count, suggesting that further research into this gene region and its role in nevi development may prove informative.
Shorter telomere length measured in prediagnostic blood samples was previously found to be correlated with number of nevi (P
0.002) and associated with a reduced risk of melanoma, although not significantly (P-trend
. In a recent combined analysis of 3 prospective studies, shorter telomere length was associated with a reduced risk of melanoma (P-trend
. In our study, shorter telomere length was not associated with number of nevi, but similar to the previous two analyses, telomere length was associated with a decreased risk of melanoma, although it did not reach statistical significance. Measurements of telomere length in our study were performed in samples collected retrospectively. However, this is unlikely to play an important role in our findings since one would have expected to observe a positive association with the risk of melanoma and we did not observe such an association.
This is the most comprehensive study of telomere-related genes and melanoma risk to date. In addition, we explored the associations between telomere length and nevi (dysplastic nevi and number of nevi). Further, we used gene-set and pathway approaches that allowed us to maximize our ability to detect effects that were only significant at the aggregate level, and not at the SNP level. We combined both case-control and family studies and assessed multiple variants in detail. However, the assessment of different characteristics, such as the number of nevi, could vary across studies and could have affected our measurements. For this reason, we performed analyses regarding number of nevi in the two studies (CCS1 and FS) in which skin examinations were performed by the same doctor. Moreover, we only considered the variable of presence/absence of dysplastic nevi instead of their number. Furthermore, we accounted for the potential heterogeneity in the studies by computing meta-analytic estimates, instead of pooling the data together. It is known that the number of nevi changes through the lifetime of an individual, with nevi count reaching its maximum around 35–40 years of age 
. This could result in misclassification, leading to attenuated results. However, we included an interaction term between the number of nevi and age in our models to account for this effect modification by age.
Our findings may suggest a role for telomere-related genes in melanoma risk as observed for other cancers 
and add insights into the complex biology of telomeres in melanoma development. However, despite the strengths of our study, our results were not confirmed in the two US populations, including a melanoma family study and a hospital-based case-control study. Differences in the ascertainment of nevi in the replication studies and related analyses could partially explain these results. Mediterranean populations generally differ from American and Australian populations, which experience a higher incidence of melanoma, in terms of pigmentation characteristics, intensity of sun exposure and genetic susceptibility 
. These factors may also explain some discrepancies, but as in any epidemiological investigation, chance findings cannot be completely ruled out. Larger studies are needed to explore the role of these factors in relation to telomere genes and melanoma risk.
In summary, our results suggest that telomere-related genes might be related to the susceptibility of melanoma, dysplastic nevi and nevus count. Additional studies across different populations are warranted.