Search

Background: Evidence suggests that consumption of dark green leafy vegetables may influence the decrease in the risk of cutaneous squamous cell carcinoma (SCC). Dark green leafy vegetables contain folate as a main component among other nutrients; thus, we hypothesised that their possible observed protective effect on SCC, observed in previous studies, would be more evident in persons with specific genotypes related to folate metabolism. Methods: Genotyping of methylenetetrahydrofolate reductase (MTHFR) gene variants rs1801133 (C677T) and rs1801131 (A1298C) was carried out for 1,128 participants in an Australian community-based longitudinal study of skin cancer. Dietary intakes were assessed through repeated Food Frequency Questionnaires (1992–1996), and all incident skin cancers were recorded in 1992–2007 and histologically confirmed. We assessed associations between intake of dark green leafy vegetables and SCC development in strata defined by genotype, by calculating relative risks (RRs) with 95% confidence intervals (CIs) using generalised linear models with negative binomial distribution and person-years of follow-up as offset. Results: High versus low intake of dark green leafy vegetables was associated with a lower risk of SCC tumours in carriers of the C677T variant allele (RR = 0.42, 95% CI = 0.23–0.75), and within wild-type A1298C homozygotes (RR = 0.43, 95% CI = 0.22–0.85). Conclusion: The protective effect of dark green leafy vegetables on cutaneous SCC may be genotype-dependent. Folate metabolism-related gene polymorphisms should be considered when assessing the relation of green leafy vegetables to cancer risk.

Gain in weight is common after heart transplantation but the magnitude of usual weight gain and whether this varies by country is unknown. We systematically reviewed all relevant studies to quantify weight change among heart transplant recipients (HTRs) in the years after transplantation and assess variation with geographic location. We searched PubMed, Cumulative Index to Nursing and Allied Health Literature, and Excerpta Medica Database databases to September 2020. Eligible studies reported adult HTRs' mean/median weight and/or body mass index (BMI) up to time of transplantation (baseline) and posttransplantation in any language. Weighted mean differences (WMDs) (95% confidence intervals [CIs]) of weight/BMI from baseline to posttransplantation were estimated using a random-effects model. Ten studies met the inclusion criteria. Pooled analysis showed weight gain of 7.1 kg (95% CI, 4.4-9.8 kg) in HTRs 12 months posttransplant, with corresponding BMI increase of 1.69 kg/m2 (95% CI, 0.83-2.55 kg/m2). Greatest weight gain at 12 months posttransplant occurred in US HTRs (WMD weight 10.42 kg, BMI 3.25 kg/m2) and least, in European HTRs (WMD weight 3.10 kg, BMI 0.78 kg/m2). In conclusion, HTRs gain substantial weight in the years after transplantation, but varying widely by geographic location.

Human papillomaviruses (HPVs) cause superficial epidermal infections and are only cleared if they trigger an immunological response. We analysed SNPs that had previously been investigated for association with HPV infection to determine whether they play a role in the serological response to cutaneous beta-HPVs in an Australian population. Serum samples from 1,142 participants were analysed for seropositivity against the L1 protein of 21 beta-HPV types. Associations between seropositivity to beta-HPV types and the SNPs rs9264942 (HLA-C; HPV-9, p = 0.022, HPV-15, p = 0.043 and HPV-17, p = 0.004), rs12449858 (EVER1; HPV-23, p = 0.029), and rs2981451 (FGFR2; HPV-22, p = 0.049) were identified. We found that certain SNPs could be involved in the serological response to beta-HPVs.

Gain in weight is common after heart transplantation but the magnitude of usual weight gain and whether this varies by country is unknown. We systematically reviewed all relevant studies to quantify weight change among heart transplant recipients (HTRs) in the years after transplantation and assess variation with geographic location. We searched PubMed, Cumulative Index to Nursing and Allied Health Literature, and Excerpta Medica Database databases to September 2020. Eligible studies reported adult HTRs' mean/median weight and/or body mass index (BMI) up to time of transplantation (baseline) and posttransplantation in any language. Weighted mean differences (WMDs) (95% confidence intervals [CIs]) of weight/BMI from baseline to posttransplantation were estimated using a random-effects model. Ten studies met the inclusion criteria. Pooled analysis showed weight gain of 7.1 kg (95% CI, 4.4-9.8 kg) in HTRs 12 months posttransplant, with corresponding BMI increase of 1.69 kg/m 2 (95% CI, 0.83-2.55 kg/m 2 ). Greatest weight gain at 12 months posttransplant occurred in US HTRs (WMD weight 10.42 kg, BMI 3.25 kg/m 2 ) and least, in European HTRs (WMD weight 3.10 kg, BMI 0.78 kg/m 2 ). In conclusion, HTRs gain substantial weight in the years after transplantation, but varying widely by geographic location.

Background Sonic Hedgehog (SHH) pathway dysregulation is implicated in basal cell carcinoma (BCC) development. To evaluate the possible wider role of SHH gene variants in skin carcinogenesis, we assessed associations of genes in the SHH pathway with lifetime development of any keratinocyte cancer (KC), and with developing either BCCs or squamous cell carcinomas (SCCs) exclusively, in a 25-year prospective, population-based study of 1,621 Australians. Methods We genotyped 795 unrelated adults with available blood samples: 311 cases with any KC (186 developing BCCs-only, 55 SCCs-only, 70 BCCs and SCCs) and 484 controls. We compared allele frequencies of 158 independent SNPs across 43 SHH genes between cases and controls, and performed a gene-based analysis. Results We found associations between SNP rs4848627 (GLI2) (related to DNA synthesis in keratinocytes) and development of any KC (OR = 1.53; 95% CI = 1.06–2.13, P P CCND2 was associated with exclusive BCC development (OR = 1.43, CI = 1.12–1.82, P PRKACG (protein kinase cAMP-activated catalytic subunit gamma) with any KC (P = 0.013). Conclusion We conclude that variants located in genes in the SHH pathway may are involved in SCC as well as BCC development.

Background A personal history of keratinocyte carcinoma (KC) has been reported as a risk factor for developing subsequent primary cutaneous and non-cutaneous malignancies. However most evidence to date stems from observational studies which are prone to bias, confounding and reverse causation. Our aim was to examine this association using different Mendelian randomization (MR) approaches. Methods We performed a one-sample MR analysis using individual-level data from the UK Biobank (n = 394,306). This analysis was then validated in an independent dataset in the QSkin cohort (n = 16,896). Using 64 independent genetic variants known to be associated with KC, we generated a polygenic risk score (PRS) for each participant in the UK Biobank and the QSkin cohort. We then performed two-sample MR analyses using genome-wide association study (GWAS) summary statistics. We tested the association between genetically predicted KC and risk of subsequent cancer using logistic regression. Results Results from one-sample MR analyses in the UK Biobank indicated that a personal history of KC was significantly associated with cancer overall (excluding melanoma) (OR: 1.15, 95% CI: 1.10-1.20, per doubling the prevalence of KC). The results from the two-sample MR corroborate the findings from the one-sample MR, although the risk estimate was lower (OR: 1.05, 95% CI: 1.03-1.07). Conclusions Our MR analyses suggest that genetically predicted KC is a risk factor for developing subsequent primary malignancies. Key messages A personal history of KC may serve as a proxy marker of inherited cancer risk.

Background: Observational studies evaluating the link between polyunsaturated fatty acids (PUFA) and cancers have yielded mixed findings. We used Mendelian randomization (MR) to evaluate whether genetic evidence supports a causal role for PUFAs on overall cancer outcomes. Methods: We identified genetic instruments for six PUFAs from previous literature and evaluated their association with overall cancer risk (46,155 cases, 270,342 controls) and cancer mortality (6,998 deaths, 270,342 controls) among the UK Biobank cohort. We used the inverse variance weighted model to combine SNP estimates, and derived log (OR) estimates per SD change in each PUFA. Results: None of the six PUFAs showed association with overall cancer risk or mortality, with narrow confidence interval (CI) ruling out all but very small effects, for example, arachidonic acid (AA) overall cancer risk (OR, 1.02; 95% CI, 1.00–1.03). Sex-specific analysis revealed no associations except α-linolenic acid for potentially reducing cancer risk in men (OR, 0.92; 95% CI, 0.86–0.98; P = 0.02); however, this was nonsignificant after multiple testing correction. From individual cancers, only colorectal cancer showed evidence for a causal association for higher AA levels (OR, 1.05; 95% CI, 1.03–1.07), with similar results for the other correlated PUFAs. Conclusions: Our study provides no support for the hypothesis that PUFAs reduce overall cancer risk or mortality. Higher AA levels increased the risk for colorectal cancer. Impact: Our well-powered MR study provides robust causal inferences for the PUFAs on overall cancer risk and mortality. Future larger studies are warranted to replicate the individual cancer findings.

The keratinocyte cancers (KC), basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) are the most common cancers in fair-skinned people. KC treatment represents the second highest cancer healthcare expenditure in Australia. Increasing our understanding of the genetic architecture of KC may provide new avenues for prevention and treatment. We first conducted a series of genome-wide association studies (GWAS) of KC across three European ancestry datasets from Australia, Europe and USA, and used linkage disequilibrium (LD) Score regression (LDSC) to estimate their pairwise genetic correlations. We employed a multiple-trait approach to map genes across the combined set of KC GWAS (total N = 47 742 cases, 634 413 controls). We also performed meta-analyses of BCC and SCC separately to identify trait specific loci. We found substantial genetic correlations (generally 0.5–1) between BCC and SCC suggesting overlapping genetic risk variants. The multiple trait combined KC GWAS identified 63 independent genome-wide significant loci, 29 of which were novel. Individual separate meta-analyses of BCC and SCC identified an additional 13 novel loci not found in the combined KC analysis. Three new loci were implicated using gene-based tests. New loci included common variants in BRCA2 (distinct to known rare high penetrance cancer risk variants), and in CTLA4, a target of immunotherapy in melanoma. We found shared and trait specific genetic contributions to BCC and SCC. Considering both, we identified a total of 79 independent risk loci, 45 of which are novel.

Genome-wide association studies (GWAS) have identified a number of risk loci for cutaneous melanoma. Cutaneous melanoma shares overlapping genetic risk (genetic correlation) with a number of other traits, including its risk factors such as sunburn propensity. This genetic correlation can be exploited to identify additional cutaneous melanoma risk loci by multitrait analysis of GWAS (MTAG). We used bivariate linkage disequilibrium-score regression score regression to identify traits that are genetically correlated with clinically confirmed cutaneous melanoma and then used publicly available GWAS for these traits in a multitrait analysis of GWAS. Multitrait analysis of GWAS allows GWAS to be combined while accounting for sample overlap and incomplete genetic correlation. We identified a total of 74 genome-wide independent loci, 19 of them were not previously reported in the input cutaneous melanoma GWAS meta-analysis. Of these loci, 55 were replicated (P0.05/74, Bonferroni-corrected P-value in two independent cutaneous melanoma replication cohorts from Melanoma Institute Australia and 23andMe, Inc. Among the, to our knowledge, previously unreported cutaneous melanoma loci are ones that have also been associated with autoimmune traits including rs715199 near LPP and rs10858023 near AP4B1. Our analysis indicates genetic correlation between traits can be leveraged to identify new risk genes for cutaneous melanoma.

Background: Epidemiological data suggest that cancer patients have a reduced risk of subsequent Parkinson’s disease (PD) development, but the prevalence of PD in melanoma patients is often reported to be increased. Causal relationships between cancers and PD have not been fully explored. Objective: To study causal relationship between different cancers and PD. Methods: We used GWAS summary statistics of 15 different types of cancers and two-sample Mendelian randomization to study the causal relationship with PD. Results: There was no evidence to support a causal relationship between the studied cancers and PD. We also performed reverse analyses between PD and cancers with available full summary statistics (melanoma, breast, prostate, endometrial and keratinocyte cancers) and did not find evidence of causal relationship. Conclusion: We found no evidence to support a causal relationship between cancers and PD and the previously reported associations could be a result of genetic pleiotropy, shared biology or biases.

Human papillomaviruses (HPVs) cause superficial epidermal infections and are only cleared if they trigger an immunological response. We analysed SNPs that had previously been investigated for association with HPV infection to determine whether they play a role in the serological response to cutaneous beta-HPVs in an Australian population. Serum samples from 1,142 participants were analysed for seropositivity against the L1 protein of 21 beta-HPV types. Associations between seropositivity to beta-HPV types and the SNPs rs9264942 (HLA-C; HPV-9, p = 0.022, HPV-15, p = 0.043 and HPV-17, p = 0.004), rs12449858 (EVER1; HPV-23, p = 0.029), and rs2981451 (FGFR2; HPV-22, p = 0.049) were identified. We found that certain SNPs could be involved in the serological response to beta-HPVs.

Background Epidemiological studies have consistently documented an increased risk of developing primary non-cutaneous malignancies among people with a history of keratinocyte carcinoma (KC). However, the mechanisms underlying this association remain unclear. We conducted two separate analyses to test whether genetically predicted KC is related to the risk of developing cancers at other sites. Methods In the first approach (one-sample), we calculated the polygenic risk scores (PRS) for KC using individual-level data in the UK Biobank (n = 394 306) and QSkin cohort (n = 16 896). The association between the KC PRS and each cancer site was assessed using logistic regression. In the secondary (two-sample) approach, we used genome-wide association study (GWAS) summary statistics identified from the most recent GWAS meta-analysis of KC and obtained GWAS data for each cancer site from the UK-Biobank participants only. We used inverse-variance-weighted methods to estimate risks across all genetic variants. Results Using the one-sample approach, we found that the risks of cancer at other sites increased monotonically with KC PRS quartiles, with an odds ratio (OR) of 1.16, 95% confidence interval (CI): 1.13–1.19 for those in KC PRS quartile 4 compared with those in quartile 1. In the two-sample approach, the pooled risk of developing other cancers was statistically significantly elevated, with an OR of 1.05, 95% CI: 1.03–1.07 per doubling in the odds of KC. We observed similar trends of increasing cancer risk with increasing KC PRS in the QSkin cohort. Conclusion Two different genetic approaches provide compelling evidence that an instrumental variable for KC constructed from genetic variants predicts the risk of cancers at other sites.

Background and objectivesMost cancers appear with reduced frequency in Parkinson’s disease (PD), but the prevalence of melanoma and brain cancers are often reported to be increased. Shared genetic architecture and causal relationships to explain these associations have not been fully explored.MethodsLinkage disequilibrium score regression (LDSC) was applied for five cancer studies with available genome-wide association studies (GWAS) summary statistics to examine genetic correlations with PD. Additionally, we used GWAS summary statistics of 15 different types of cancers as exposures and two-sample Mendelian randomization to study the causal relationship with PD (outcome).ResultsLDSC analysis revealed a potential genetic correlation between PD and melanoma, breast cancer and prostate cancer. There was no evidence to support a causal relationship between the studied cancers and PD.ConclusionsOur results suggest shared genetic architecture between PD and melanoma, breast, and prostate cancers, but no obvious causal relationship between cancers and PD.

Three major types of skin cancers (melanoma, basal cell carcinoma (BCC) and squamous cellcarcinoma (SCC)) are a significant health burden mainly for the fair-skinned population.Melanoma is the most lethal form of skin cancer while BCC is the commonest form of skincancer. Cutaneous squamous cell carcinoma is more frequent than melanoma and moreaggressive than BCC. Ultraviolet radiation, pigmentation phenotypes and genetic factorsplay a role in skin cancer etiology. The objective of this work is to expand the currentunderstanding of genetic architecture of skin cancer. Furthermore, finding the causalassociation of various risk factors and skin cancer may strengthen preventive efforts. Thus, Ihave used genetic data to help evaluate the causal association of polyunsaturated fattyacids (PUFAs), vitamin D and melanoma and overall-cancer risk and mortality.In Chapter 1, I have included an introduction to skin cancer and its various risk factors aswell as a general overview of the statistical and computational methods used in the studiesincluded in the thesis.In Chapter 2, I performed Mendelian randomisation (MR) of melanoma on PUFAs. The studysample comprised 12,874 melanoma cases and 23,203 controls. PUFA Instrumentalvariables were derived from publicly available CHARGE consortium data. By using theinverse variance weighted method, I identified that even with a large change of fatty acidlevels (Docosapentaenoic acid (DPA- moving a person from the 17 th percentile to themedian)) there is no causal association between PUFAs and risk of melanoma odds ratio[OR] = 1.03 (95% confidence interval [CI] = 0.96-1.10). Other PUFAs yielded similar results.In Chapter 3, I used the same instruments of PUFAs to determine the causal association ofoverall cancer risk and mortality (46,155 cases for risk, 6,998 cases for morality and 270,342controls from UK Biobank). I further performed individual cancer analysis for 6 cancer types.I identified that there is no causal association between PUFAs and overall cancer risk andmortality except arachidonic acid (AA) overall cancer risk (OR = 1.02, 95% CI = 1.00 – 1.03).Colorectal cancer risk was increased with higher AA levels (OR = 1.05, 95% CI =1.03 – 1.07).In Chapter 4, I present MR analysis of vitamin D on melanoma risk. Observational studieshave shown many health benefits of vitamin D. To determine the causality of melanoma andvitamin D, I used melanoma meta-analysis consortium data with vitamin D instruments fromprevious publications. I found that changing vitamin D levels by a large amount (20nmol/Ldecrease), did not change the melanoma risk, OR =1.06 (95% CI = 0.95 – 1.19).In Chapter 5, I present the multi-trait analysis of GWAS (MTAG) of keratinocyte cancer (KC)which revealed 29 novel loci for KC risk. From individual cancer meta-analysis of either BCCor SCC, I further identified 12 loci and 1 locus, respectively. I used data from UKBB, QSkin,23andMe and eMERGE (47,742 cases, 634,413 controls) for KC MTAG analysis. I identifiedseveral drug targets; these included the important immune regulation locus CTLA4, which isa target of immunotherapy in melanoma treatment.In Chapter 6, I present the MTAG analysis of melanoma and correlated traits (biochemicaltraits, autoimmune traits) which revealed 74 loci. 18 of them are novel for melanoma risk.New loci included NDUFC2 is a drug target for diabetes mellitus, CTLA4 which is already adrug target for melanoma immunotherapy. My studies may provide future avenues formelanoma and KC treatments. Finally, in Chapter 7 I provide a general discussion withfuture directions for genetic epidemiology, particularly about GWAS, MR and MTAG and theusage of polygenic risk scores.In conclusion, I identified new genetic variants associated with KC and melanoma risk whichwill be potentially useful for future MR and polygenic risk score construction. My causalinference on overall cancer risk and melanoma may inform public health policies. Thesefindings may provide insight for future personalised treatments for skin cancer.

Previous Mendelian randomization (MR) studies on 25-hydroxyvitamin D (25(OH)D) and cancer have typically adopted a handful of variants and found no relationship between 25(OH)D and cancer; however, issues of horizontal pleiotropy cannot be reliably addressed. Using a larger set of variants associated with 25(OH)D (74 SNPs, up from 6 previously), we perform a unified MR analysis to re-evaluate the relationship between 25(OH)D and ten cancers. Our findings are broadly consistent with previous MR studies indicating no relationship, apart from ovarian cancers (OR 0.89; 95% C.I: 0.82 to 0.96 per 1 SD change in 25(OH)D concentration) and basal cell carcinoma (OR 1.16; 95% C.I.: 1.04 to 1.28). However, after adjustment for pigmentation related variables in a multivariable MR framework, the BCC findings were attenuated. Here we report that lower 25(OH)D is unlikely to be a causal risk factor for most cancers, with our study providing more precise confidence intervals than previously possible., Studies of the genetic association between vitamin D and cancer risk have typically been underpowered. Here the authors analyse this using Mendelian Randomisation with more than 70 vitamin D variants obtained from the UK Biobank and large-scale data from various consortia, confirming null associations between vitamin D and most cancers.

Background: Basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) (collectively referred to as keratinocytic skin cancers (KC)) are the most common forms of cancer in fair-skinned people. KCs account for the highest expenditure of any cancer type in the Australian healthcare system. Knowledge about the genetic architecture of skin cancer may help elucidate the biology of the disease. Methods: We conducted a genome-wide meta-analysis of KC, including 28248 cases and 353855 controls from United Kingdom, United States and Australian populations. We performed LD score regression to determine the genetic correlation between KC, BCC, SCC, and melanoma. Results: We identified 41 independent genome-wide significant SNPs associated with KC risk, 16 of which have not been reported previously. A further 10 loci were implicated using gene-based tests. New loci included common variants in BRCA2 (rare variants increase risk of various cancers) and CTLA4 (immune response; antibodies targeting CTLA-4 are used to treat melanoma). We found a strong genetic correlation between BCC and SCC, 0.93 (95% confidence interval [CI] 0.71 - 1.16). The genetic correlation between melanoma and BCC was 0.66 (95% CI: 0.41 - 0.91). Comparatively lower genetic correlations were identified between melanoma and SCC 0.43 (95% CI: 0.12 - 0.74), and melanoma and KC 0.52 (95% CI: 0.27 - 0.76). Conclusions: We have identified novel genetic loci associated with KC risk which will lead to a greater understanding of the biology of KC. Future work will leverage the genetic correlations between BCC, SCC and melanoma to further advance our understanding of these cancers. Citation Format: Upekha E. Liyanage, Matthew H. Law, Xikun Han, Jiyuan An, Jue-Sheng Ong, Puya Gharahkhani, Mark M. Iles, Rachel E. Neale, Catherine Olsen, Stuart Macgregor, David C. Whiteman. Genome-wide meta-analysis of keratinocytic cancers identifies 26 novel risk loci [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1592.

Several preclinical studies have identified the antiproliferative effects of 25-hydroxyvitamin D [25(OH)D; vitamin D]. Ultraviolet radiation (UVR) is essential for vitamin D synthesis yet increases the risk of melanoma. Observational studies on the association of vitamin D levels with melanoma risk have reported inconclusive results, and are difficult to interpret owing to the potential confounding from the dual role of UVR.To determine whether there is a causal association between genetically predicted 25(OH)D concentrations and melanoma using a Mendelian randomization (MR) approach.We performed MR using summary data from a large genome-wide association study (GWAS) meta-analysis of melanoma risk, consisting of 12 874 cases and 23 203 controls. Five single nucleotide polymorphisms associated with 25(OH)D concentration - rs12785878, rs10741657, rs2282679, rs6013897 and rs116970203 - were selected as instrumental variables. An inverse variance weighted method was used to access the evidence for causality. MR results from the melanoma meta-analysis were combined with results from an MR study based on a melanoma risk GWAS using UK Biobank data.A 20 nmol LThe results suggest that vitamin D levels may not be causally associated with the risk of melanoma. What's already known about this topic? Antitumour activity of vitamin D has been identified in preclinical studies. Observational studies link vitamin D deficiency with an increased risk of a range of cancers. There is a growing public interest for vitamin D supplementation. Observational studies of melanoma are fraught with difficulties because while higher ultraviolet radiation levels increase vitamin D levels, such exposure is also associated with increased melanoma risk. Results from observational studies are inconclusive regarding the effect of vitamin D on melanoma risk. What does this study add? Using Mendelian randomization, an approach to causal inference, which is analogous to a natural randomized controlled trial, we found no causal association between vitamin D levels and melanoma.