Persistent organic pollutants and risk of cutaneous malignant melanoma among women

Abstract Background Despite the increasing trend of cutaneous malignant melanoma (CMM) incidence in Canada, especially among females, few risk factors other than ultraviolet radiation exposure, have been identified. Aim We conducted a case–control study of 406 CMM cases and 181 controls to evaluate the potential impact of body burdens of various persistent organic pollutants on CMM risk. Methods Detailed data on potential confounding factors, including lifetime repeated sun exposure and skin reaction to repeated sun exposure, were collected. Gas chromatography tandem mass spectrometry was used to assay plasma levels of 14 polychlorinated biphenyl (PCB) congeners and 11 organochlorine (OC) pesticides among cases and controls. Results Statistically significant trends of increased CMM risk were observed with increasing plasma concentrations of multiple PCB congeners, including PCBs 138, 153, 170, 180, 183 and 187. For example, compared to lowest plasma concentration quartile of PCB‐138, the second, third and fourth quartiles were associated with 1.7 (95% CI: 0.9–2.9), 2.3 (95% CI: 1.3–4.1) and 2.4 (95% CI: 1.3–4.5) ‐fold increased risks of CMM, respectively. Similarly, increasing plasma concentrations of several OC pesticides (i.e., β‐HCH, HCB, Mirex, oxychlordane and trans‐Nonachlor) showed statistically significant trends with increased CMM risk. For example, compared to lowest plasma concentration quartile of β‐HCH, the second, third and fourth quartiles were associated with 1.3 (95% CI: 0.7–2.3), 2.1 (95% CI: 1.2–3.7) and 2.3 (95% CI: 1.2–4.4) ‐fold increased risks of CMM, respectively. Conclusion Plasma levels of several persistent organic pollutants were highly correlated, suggesting that observed associations were not necessarily independent of each other. Given the highly correlated nature of exposure to PCB and OC analytes, sophisticated analyses that consider complex mixtures should be considered in future studies.


| INTRODUCTION
As in other parts of the world, there has been an increasing trend of cutaneous malignant melanoma (CMM) incidence in Canada. Since 1994, Canadian females have the second highest annual percent change in age-standardize incidence rate of melanoma over other cancers. 1,2 Despite the increasing incidence, few risk factors, other than ultraviolet radiation exposure, have been identified. 3 Polychlorinated biphenyls (PCBs) are a class of persistent organic pollutant (POP; i.e., compounds that bio-accumulate in the environment, animals and humans and can still be detected in the general population despite their use being banned decades ago) and have been classified as known human carcinogens by the International Agency for Research on Cancer (IARC). [4][5][6] This was primarily based on evidence of increased melanoma risk among occupationally exposed individuals; however, more recently conducted meta-analyses, that primarily include highly exposed occupational groups, do not confirm IARC's classification. 7,8 Few studies of organochlorine (OC) pesticides, another class of POP, and CMM risk have also been conducted and, as with PCBs, the primary focus has been on occupational exposures.
For example, in the Agricultural Health Study (AHS), toxaphene exposure, as assessed via questionnaire, was found to be associated with increased CMM risk, but this was not replicated in the AHS after additional years of follow-up. 9,10 To assess the potential associations of PCB and OC pesticide exposures with risk of CMM in the general population, we previously conducted a preliminary population-based case-control study in BC. 4 Comparing 80 CMM cases to 310 controls, we observed strong associations between plasma levels of various PCB congeners and OC pesticides and risk of CMM. 4 We have now followed-up these findings with another study that includes a much larger number of cases. In addition, the study focuses on women given the steeper increases in CMM incidence observed in this group.

| Study population
The protocol for this investigation was approved by the Research Ethics Board of the University of British Columbia and the BC Cancer Agency.
Between July 2011 and January 2013, females aged 20-79 with CMM, were identified through the population-based BC Cancer Registry. A total of 703 cases were successfully contacted by telephone (up to three contact attempts were made), with 241 (34%) refusing to participate, leaving 462 (66%) that consented to participate in the study ( Figure 1).
From December 2012 to May 2014, 2078 cancer-free controls, frequency matched to cases on age and residential district, were randomly selected from the consolidation file of the British Columbia Ministry of Health, which contains identifying information on all participants in the population-based health insurance plan. A total of 778 were successfully contacted by telephone, with 526 (67%) refusing to participate, leaving 252 (32%) that consented to participate in the study (Figure 1).

| Data and blood sample collection
All study participants were asked to complete a computer-assisted telephone interview (CATI) to provide information on sun exposure, medical history, lifestyle factors, family history of melanoma, occupational history, and residential history. Participants were also asked to provide a whole blood sample, which was collected in EDTA tubes at a community laboratory and shipped, on ice, to BC Cancer within 24 h of collection for processing. A volume of 2 ml of plasma was separated from whole blood by centrifugation, transferred to vials with a Teflon stopper and frozen at À80 C. β-hexachlorocyclohexane (β-HCH), α-chlordane, γ-chlordane, cis-Nonachlor, trans-Nonachlor, p,p'-DDT, p,p'-DDE, hexachlorobenzene (HCB), mirex, and oxychlordane), which were selected based on observations in the preliminary study, 4,11 were measured using the gas chromatography-mass spectrometry (GC-MS)

| PCB and OC pesticide assays
where the analytical method was described in Fisher et al. 12 The limit of detection (LOD) varied from 0.01 to 0.3 μg/L for PCBs and varied from 0.005 to 0.09 μg/L for OCs (Table 1).Concentrations below the LOD were assigned a value of the LOD divided by √2. 4,13 Using enzymatic methods, free cholesterol (FC), total cholesterol (TC), triglycerides (TG), and phospholipids (PL) were measured in each plasma sample. 4 Total lipid concentration was calculated using the T A B L E 1 Number of samples measured above limit of detection Excluded from further analysis due to ≤20% values above the detection limit. b Median lipid-adjusted concentration among controls of analytes with >80% of measurements above the limit of detection.
Akins summation formula. 4,12,14 Lipid-adjusted concentrations (μg/kg of lipids) were calculated by dividing the whole-weight measurements of each analyte by the total lipid concentration. 4,14 Blood sample collection and processing used the same methods for both melanoma cases and NHL controls as previously described.
All organochlorine assays were performed at the Centre de Toxicologie in Quebec, Canada; assays for the con-trols between 2002 and 2005, and for CMM cases in 2008.

| Statistical analysis
For those analytes with greater than 80% of samples above the LOD, Spearman rank correlations between the various analytes were examined.

| RESULTS
As shown in Figure 1, 39 cases and 43 controls were excluded due to missing samples and incomplete CATI. After further exclusion of participants with missing assay results (1 case and 1 control), non-Caucasian participants (7 cases and 27 controls) and those residing outside the Greater Vancouver Regional District and Capital Regional District (9 cases), 406 cases and 181 controls were included in the analyses (as shown in Tables 1 and 3). For multivariate analyses (Tables 4 and 5), only 375 cases and 171 controls were used due to missing covariate data.
Information on the number of samples measured above the LOD for each analyte is provided in Table 1. Statistically significant correlations between plasma levels of various PCB and pesticide analytes were observed (p < .001). The strongest correlations were observed between PCB congeners 170, 180, and 187 (r > 0.90) and between oxychlordane and trans-Nonachlor (r = 0.939) ( Table 2).
In general, cases were slightly younger than controls (mean age: 55 vs. 60) and were less educated (bachelor's degree or higher: cases 34.0 vs. 40.9%; Table 3). On average, cases reported a 1.1 pound increase in weight in the time between study entry and 2 years before study entry (i.e., before diagnosis and treatment), while controls reported, on average, a 0.1 pound increase in weight during this time (Table 3). Additionally, compared to controls, a higher proportion of cases had light hair color, fair skin color, and reported many moles.
These differences reflect well known differences seen in virtually all studies of melanoma and UV exposure.

| DISCUSSION
In this population-based case-control study, we observed statistically significant increased odds of CMM in association with several individual PCB congeners and OC pesticides after adjustment for known risk factors including constitutional factors and sunlight exposure. As shown in Table 2, plasma levels of several of these analytes were highly correlated, suggesting that observed associations were not necessarily independent of each other.
T A B L E 2 Spearman's rank correlation between individual organochlorine pesticide analytes and PCB congeners   Our findings are consistent with data reported in our preliminary case-control study. 4 Median concentrations of most analytes were slightly lower than in the previous study. This is consistent with previously published observations of declining POP concentrations over time. 19 Also, in the current study, the population was restricted to females. In general, only a few studies previously assessed sex-stratified associations of PCBs and melanoma, and no significant sex differences were observed. 7,20,21 No studies evaluating sex stratified associations between OC pesticides and CMM were identified.
In a recently conducted prospective study among Swedish women, exposure to dietary PCBs (based on food frequency questionnaire) was associated with a four-fold increased risk of malignant melanoma. 22 In contrast, recently conducted meta-analyses do not  7,8,23 While the studies included in the meta-analyses were mostly from occupational settings with presumably much higher exposure levels, few of the studies involved direct assessment of POP body burdens. Variability in mixtures of PCBs and OCs to which populations are exposed, 6 as well as heterogeneous study designs (case-control vs. cohort) as well as variability in adjustment for potential confounders (e.g., UV/sun exposure) could potentially explain the mixed associations. 6,7,22 In addition to chronic inflammation and immunosuppressive effects, it has been shown that PCBs can induce carcinogenesis through prolonged impact on cell receptors (e.g., aryl hydrocarbon receptor [AhR]) leading to cell proliferation, and deregulation of the endocrine system. 5 While oxidative stress and chronic inflammation have been speculated as mechanisms by which pesticides may increase cancer risk, an actual mechanism has not been identified. 23,24 PCBs and OC pesticides may also induce carcinogenesis through epigenetic mechanisms. Multiple studies have reported significant associations between DNA methylation and exposure to PCBs and OC pesticides. [25][26][27][28] Though focused on early life exposures, a recent study found PCBs and OC pesticides to drive sex-specific changes in DNA methylation. 29 The large case group and highly detailed covariate data, particularly related to sun susceptibility and sun exposure, are major strengths of this study. Limitations include the use of post-diagnostic blood samples, in which OC levels may have been impacted by the occurrence of cancer or its treatment. For example, it is known that cancer development and/or its treatment weight loss may lead to increased blood levels of OCs, leading to potential reverse causation. 4,30 However, as noted earlier, very little weight change between study entry and 2 years before study participation was reported among cases. Furthermore, since among the cases with a reported Breslow thickness value, over two-thirds (241 out of 359) were diagnosed with thin CMM (Breslow thickness < = 1 mm) (data not provided), it is likely that the majority of cases only received surgical treatment which would have minimal impact on OC levels. Additional limitations include the fact that controls were recruited 1-2 years after cases and the small number of controls available for the analysis.
Participation rates in epidemiologic studies, particularly among controls in case-control studies have been declining since 1990. 31 This has been attributed to an increase in unlisted phone numbers, cell phone usage and screening of calls due to widespread availability of caller identification. 32,33 While an inadequately representative control population may bias risk estimates, 34 our findings were consistent with the preliminary study which included a larger number of controls (n = 309) 3 and, in exploratory analyses, associations with known melanoma risk factors were consistent with those previously reported (results not shown). For example, as compared to having a dark skin color, an OR of 2.0 (95% CI 1.3-3.0) for CMM among those with fair skin was observed in our study. This is comparable to the association reported for fair compared to dark skin in a previously conducted meta-analysis (OR = 1.89; 95% CI, 1.49-2.39). 35 These findings help reduce concerns about any biases resulting from our control group.
In conclusion, in this study we demonstrated significant associations between various PCB and OC pesticides with CMM risk. Given the highly correlated nature of exposure to PCB and OC analytes, sophisticated analyses that take into account complex mixtures should be considered in future studies.

ACKNOWLEDGMENT
The study was supported by research grant from the Canadian Institutes of Health Research (CIHR) Operating grant MOP-102613.

CONFLICT OF INTEREST
The authors declare there is no conflict of interest.

ETHICAL STATEMENT
Ethics was approved by UBC Clinical REB H10-02669. All participants signed an informed consent.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.