- Email: [email protected]
Dietary polychlorinated biphenyls, long-chain n-3 polyunsaturated fatty acids and incidence of malignant melanoma
European Journal of Cancer 72 (2017) 137e143
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Original Research
Dietary polychlorinated biphenyls, long-chain n-3 polyunsaturated fatty acids and incidence of malignant melanoma Carolina Donat-Vargas a, Marika Berglund a, Anders Glynn b, ˚ kesson a,* Alicja Wolk a, Agneta A a b
Institute of Environmental Medicine, Karolinska Institutet, Box 210, SE-171 77, Stockholm, Sweden Department of Risk and Benefit Assessment, National Food Agency, Box 622, SE-751 26, Uppsala, Sweden
Received 28 September 2016; received in revised form 25 October 2016; accepted 21 November 2016
KEYWORDS Polychlorinated biphenyls (PCBs); Long-chain n-3 polyunsaturated fatty acids (LC n-3 PUFAs); Cutaneous malignant melanoma; Nutritional epidemiology; Cancer prevention
Abstract Background: For malignant melanoma, other risk factors aside from sun exposure have been hardly explored. Polychlorinated biphenyls (PCBs)dmainly from fatty fishd may affect melanogenesis and promote melanoma progression, while long-chain n-3 polyunsaturated fatty acids seem to exert antineoplastic actions in melanoma cells. Objectives: We aimed to assess the association of validated estimates of dietary PCB exposure as well as the intake of eicosapentaenoic acid and docosahexaenoic acid (EPA-DHA), accounting for sun habits and skin type, with the risk of malignant melanoma in middle-aged and elderly women. Methods: We included 20,785 women at baseline in 2009 from the prospective populationbased Swedish Mammography Cohort. Validated estimates of dietary PCB exposure and EPA-DHA intake were obtained via a food frequency questionnaire. Incident melanoma cases were ascertained through register-linkage. Results: During 4.5 years of follow-up, we ascertained 67 incident cases of melanoma. After multivariable adjustments, exposure to dietary PCBs was associated with four-fold increased risk of malignant melanoma (hazard ratio [HR], 4.0 [95% confidence interval {CI}, 1.2e13; P for trend Z 0.02]), while EPA-DHA intake was associated with 80% lower risk (HR, 0.2 [95% CI, 0.1e0.8; P for trend Z 0.03]), comparing the highest exposure tertiles with the lowest.
* Corresponding author: Karolinska Institutet, Institute of Environmental Medicine, Nutritional Epidemiology, Box 210, 171 77, Stockholm, Sweden. ˚ kesson). E-mail address: [email protected] (A. A http://dx.doi.org/10.1016/j.ejca.2016.11.016 0959-8049/ª 2016 Elsevier Ltd. All rights reserved.
138
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
Conclusion: While we found a direct association between dietary PCB exposure and risk of melanoma, EPA-DHA intake showed to have a substantial protective association. Question of benefits and risk from fish consumption is very relevant and further prospective studies in the general population verifying these findings are warranted. ª 2016 Elsevier Ltd. All rights reserved.
1. Introduction The incidence of malignant melanomadthe most lethal skin cancer due its high potential for metastasisdhas increased abruptly over the past 50 years [1]. In Europe, incidence rates are particularly high in the Nordic countries and generally higher among women than among men [1], although with a female survival advantage [2]. Despite ultraviolet (UV) radiation is the major risk factor [3], other environmental factors such as chemical exposures may play a part in the aetiology of melanoma [4]. Polychlorinated biphenyls (PCBs) are extremely persistent synthetic organochlorine chemicals, with a well-known potential toxicity, that have been widely dispersed into the environment for decades, bioaccumulating and magnifying in the food chain. The general population is exposed to PCBs primarily via food, where fatty fish is a major contributor to dietary PCB exposure in populations with relatively high fish consumption [5,6]. PCBs are readily absorbed, distributed in the body and accumulated in adipose tissue, with half-life for the most persistent congeners ranging from a couple of years to decades [7]. The International Agency for Research on Cancer has recently upgraded PCBs to group 1, i.e. carcinogenic to humans [8]. The strongest evidence on PCB carcinogenicity in humans comes from epidemiological data on risk of malignant melanoma. Virtually all studies on occupational and accidental PCB exposure report excess risk [8]. To date, the only study conducted in a general population is a caseecontrol study with 80 malignant melanoma cases and 310 controls [9]. A clear statistically significant higher odds was observed for the highest compared with the lowest quartile of total plasma PCB concentrations (odds ratio, 6.0; 95% confidence intervals [CI], 2.0e18.2), after adjusting for phenotypic factors, sun sensitivity and sun exposure. Apart from being a major source of PCBs, fatty fish is also the main source of long-chain n-3 polyunsaturated fatty acids (LC n-3 PUFAs), proposed to protect against several types of cancer [10] including melanoma [11]. Although in vitro and animal studies have provided convincing evidence to support these favourable effects of LC n-3 PUFAs (mainly eicosapentaenoic acid [EPA; 20:5 n-3] and docosahexaenoic acid [DHA, 22:6 n-
3]) against melanoma [12e15], the existing evidence from epidemiological studies exploring this relationship is very limited [11,16]. The aim of the present study was to assess the association between validated estimates of dietary PCB exposure and LC n-3 PUFA intake with the risk of malignant melanoma in a population-based prospective cohort study of middle-aged and elderly Swedish women. 2. Methods 2.1. Study population The Swedish Mammography Cohort is a large population-based longitudinal cohort established in 1987e1990. All women born between 1914 and 1948, resident in two counties in Central Sweden (n Z 90,303), were invited to complete a self-administrated questionnaire concerning diet (response rate 74%). In 1997, a more detailed questionnaire was sent to all participants still alive and living in the study area (response rate 70%). In 2009, an additional questionnaire that sought information on sun habits and skin type (information not available until then) was distributed to all remaining cohort members (response rate 84%). More details on the study design have been published elsewhere [17]. In the present study, the 1997 questionnaire was used for the exposure assessment, as there was sufficient documentation of the PCB content in different foods at that time [6]. In order to be able to control for sun habits and skin type (the major known risk factors for melanoma), we used the 2009 questionnaire as baseline in the present study (n Z 21,818). For the final analyses, we excluded those women with implausible total energy intake (3 standard deviation from the log-transformed mean), those who had a diagnosis of cancer before 1997 and those with prevalent malignant melanoma at baseline (1 May 2009), resulting in 20,785 women available at the start of follow-up. 2.2. Dietary PCB exposure and other covariates The 1997 questionnaire included a 96-item food frequency questionnaire (FFQ), constructed to reflect the women’s average consumption of different foods and beverages during the last year. The FFQ has been
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
validated, obtaining Pearson correlation coefficients (r) between the average of four 1-week dietary records and the FFQ of 0.6 for fatty fish and 0.5e0.7 for dairy products (Wolk A, unpublished results). The dietary exposure to PCBs was estimated through an extensive recipe-based database created for the 1997 FFQ, described in detail elsewhere [5]. This database was based on concentrations of the PCB-153 congener, which is the most abundant congener in food and therefore a very good indicator of both total PCBs and dioxin-like PCBs in food as well as in human serum [6,8,18]. Daily dietary exposure to PCBs (ng day1) and to EPA-DHA (g day1) was estimated by multiplying the average concentration in various foods (obtained from the Food Agency’s food control and monitoring programmes and the Swedish food composition database) with the respective consumption frequency and portion size. In order to insure that the association of PCB and EPA-DHA intakes with the outcome is independent of the total energy intake, estimated PCBs were adjusted for total daily energy intake (mean 1700 kcal) using the residual-regression method [19]. The FFQ-based dietary estimate of PCB exposure has been extensively validated against six serum PCB congeners in a representative subsample of the cohort (Spearman correlation coefficients [r] ranged from 0.30 to 0.58) [6]. The FFQ-based dietary intake of EPADHA has also been validated against adipose tissue concentrations of EPA and DHA (r Z 0.32 and 0.48, respectively, for concurrent exposure, and r Z 0.21 and 0.33, respectively, for past exposure [6 years prior to the adipose tissue sampling]) [21]. From the 2009 questionnaire we obtained information on skin type and sun habits, comprising use of sunblock, travel to sunny resorts during winter time and sun behaviour. The other variables were obtained from the 1997 questionnaire. 2.3. Case ascertainment Incident cases of malignant melanoma (code C43 in International Classification of Diseases-10) were ascertained from the Swedish Cancer Register, which is close to 100% completion [20]. 2.4. Statistical analyses Person-time of follow-up was calculated for each participant from 1 May 2009 until the date of melanoma diagnosis, death (through register-linkage) or end of follow-up (December 31, 2013), whichever occurred first. To assess the relationship between tertiles of dietary PCB and EPA-DHA intakes and the subsequent risk of developing malignant melanoma, hazard ratios and 95% CI were estimated using Cox
139
proportional hazard regression models, with attained age as the underlying timescale. No evidence of departure from the proportional hazards assumption was detected. To test for linear trends across increasing categories of dietary PCB exposure and EPA-DHA intake we assigned the median exposure within each category and included it as a continuous variable. The models were adjusted for the following potential confounders: level of education, factors related to the retention of PCB in the body (i.e. parity and body mass index), skin type (based on the skin reaction to the sun), sun habits (i.e. use of sunblock, travel to sunny resorts during winter time and sun behaviour), smoking habits, alcohol consumption and total energy intake. In a separate model mutual adjustments for dietary PCB exposure and EPA-DHA intake were performed. We fitted the multivariable-adjusted models including the missing values (below 3% for all covariates with the exception of skin type, 6%) in a separate category. Analyses were run in Stata version 13.0 (StataCorp LP), with statistical significance set at the two-sided 0.05 level. 3. Results During 4.5 years of follow-up (94,200 person-years), 67 cases of incident malignant melanoma were ascertained. The average age at diagnosis of melanoma was 73 (8) years. The median energy-adjusted dietary PCB exposure was 162 ng day1 (5the95th percentile 74e346 ng day1) and the median energy-adjusted dietary EPA-DHA intake was 0.3 g day1 (5the95th percentile 0.1e0.7 g day1). No major differences in age-standardised characteristics were observed across tertiles of dietary PCB exposure and EPA-DHA intake (Table 1). We observed no association between dietary PCB exposure and risk of melanoma before adjusting for LC n-3 PUFAs. However, in the complete multivariableadjusted model additionally adjusted for EPA-DHA intake, dietary PCB exposure was associated with a statistically significant four-fold higher risk of melanoma comparing the highest PCB tertile with the lowest tertile (P for trend Z 0.02). Likewise, the inverse relationship between dietary EPA-DHA intake and melanoma development only reached statistical significance after adjustment for dietary PCB exposure, resulting in an 80% lower risk of melanoma when comparing women in the highest and lowest tertiles of EPA-DHA intake (P for trend Z 0.03) (Table 2; Fig. 1). 4. Discussion Melanoma is a potentially fatal malignancy that continues to increase in incidence despite public education measures to limit sun exposure. Conceivable risk
140
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
Table 1 Age-standardised baseline characteristics of 20,785 women from the Swedish Mammography Cohort by tertiles of energy-adjusted dietary PCB and EPA-DHA exposures. Characteristicsa
Tertiles of dietary PCB exposure (ng day1)
Tertiles of EPA-DHA intake (g day1)
Range (median) n Age, yearsb Postsecondary education, % Nulliparous, % Body mass index, kg/m2 Skin reaction to the sun, %c Always red/never tanned Always tanned/never red Use of sunblock always or most of the time, %c Travelling yearly to sunny resorts during winter time, %c Preference on a sunny day, %c Sun all the time Shade all the time Current smoker, % Alcohol consumption, % No use >15 g/day Total energy intake, kcal day1 Dietary PCB exposure, ng day1 EPA-DHA intake, g day1
<141 (113) 6929 70 8 31 9 24.9 0.05
141e188 (162) 6928 70 7 35 8 24.8 0.04
>188 (246) 6928 72 8 32 9 25.2 0.05
<0.24 (0.18) 6929 71 8 31 9 24.8 0.05
0.24e0.35 (0.29) 6928 70 7 34 8 24.9 0.04
>0.35 (0.46) 6928 72 8 33 9 25.3 0.05
5 17 35
4 16 38
5 19 38
5 17 34
4 15 38
4 19 39
7
7
8
6
7
8
4 15 23
3 13 21
3 15 22
4 15 23
3 13 21
3 14 22
15 9 1820 6
9 10 1735 5
10 10 1710 6
0.17 <0.01
0.29 <0.01
0.53 <0.01
16 8 1840 6 110 0.4
9 10 1750 5 163 0.3
10 11 1675 6 276 1.6
All variables are expressed as mean standard deviation or percentage (%); EPA Z eicosapentaenoic acid; DHA Z docosahexaenoic acid; PCB Z Polychlorinated biphenyl. a Unless otherwise indicated, acquired from the 1997 questionnaire. b At the beginning of follow-up (1 May 2009). c Acquired from the 2009 questionnaire.
factors aside from sun exposure such as diet and dietary contaminants are scarcely explored. Within this large population-based prospective cohort, while exposure to dietary PCBs was associated with a statistically
significant four-fold increased risk of malignant melanoma, LC n-3 PUFAs were associated with 80% lower risk in mutually-adjusted models, comparing the highest exposure tertiles with the lowest.
Table 2 Hazard Ratio (95% confidence intervals) of malignant melanoma by tertiles of energy-adjusted dietary PCB and EPA-DHA exposures in 20,785 women from the Swedish Mammography Cohort. Tertiles of dietary PCB exposure 1
Range (median), ng day Cases Person-years Age-adjusted HR Multivariable-adjusted HRa Multivariable-adjusted HRb
<141 (113) 22 31,403 1 (Ref.) 1 (Ref.) 1 (Ref.)
141e188 (162) 19 31,561 0.9 (0.5e1.6) 0.8 (0.5e1.6) 1.5 (0.7e3.5)
P trend >188 (246) 26 31,236 1.2 (0.7e2.1) 1.2 (0.7e2.1) 4.0 (1.2e13)
Tertiles of EPA-DHA intake 1
Range (median), g day Cases Person-years Age-adjusted HR Multivariable-adjusted HRa Multivariable-adjusted HRc
<0.24 (0.18) 26 31,379 1 (Ref.) 1 (Ref.) 1 (Ref.)
0.47 0.44 0.02 P trend
0.24e0.35 (0.29) 19 31,525 0.7 (0.4e1.3) 0.7 (0.4e1.3) 0.4 (0.9e1.0)
>0.35 (0.46) 22 31,296 0.9 (0.5e1.5) 0.9 (0.5e1.5) 0.2 (0.1e0.8)
0.63 0.67 0.03
EPA Z eicosapentaenoic acid; DHA Z docosahexaenoic acid; PCB Z Polychlorinated biphenyl. a Adjusted for attained age (years), postsecondary education (yes/no), parity (0, 1e2, 3 child), body mass index (<18.5, 18.5e25, 25e30, >30 kg/m2), skin reaction to the sun (always red/never tanned, always red/sometimes tanned, sometimes red/always tanned, never red/always tanned), use of sunblock (always, most of the time, sometimes, never), travel to sunny resorts during winter time (no, sometimes, every year), preference on a sunny day (sun all the time, both sun and shade, shade all the time), smoking habits (current, former, never), alcohol consumption (no use, 0.1e5, 5.1e15, >15 g/day) and total energy intake (continuous, kcal/day). b Additionally adjusted for dietary EPA-DHA intake (tertiles). c Additionally adjusted for dietary PCB exposure (tertiles).
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
141
Fig. 1. Hazard ratios (95% confidence intervals) of malignant melanoma by tertiles of energy-adjusted dietary PCB (ng day1) and EPADHA (g day1) exposures in 20,785 women from the Swedish Mammography Cohort. EPA Z eicosapentaenoic acid; DHA Z docosahexaenoic acid; PCB Z Polychlorinated biphenyl. Adjusted for attained age (years), postsecondary education (yes/no), parity (0, 1e2, 3 child), body mass index (<18.5, 18.5e25, 25e30, >30 kg/m2), skin reaction to the sun (always red/never tanned, always red/sometimes tanned, sometimes red/always tanned, never red/always tanned), use of sunblock (always, most of the time, sometimes, never), travel to sunny resorts during winter time (no, sometimes, every year), preference on a sunny day (sun all the time, both sun and shade, shade all the time), smoking habits (current, former, never), alcohol consumption (no use, 0.1e5, 5.1e15, >15 g/day), total energy intake (continuous, kcal/day) and mutually adjusted for dietary EPA-DHA intake (tertiles) and dietary PCB exposure (tertiles).
This is the first prospective study to evaluate the association between validated estimates of dietary PCB exposure and the incidence of melanoma. The results are in line with the only previous study on melanoma and PCB exposure conducted in the general population [9], which found an almost five-fold higher odds (95% CI, 1.7e14.1) of melanoma for those in the highest quartile of the plasma PCB-153 congener compared to those in the lowest. Animal studies have shown evidence of PCB carcinogenicity in different tissues and several common PCB carcinogenic mechanisms have been established, including the aryl hydrocarbon receptor (AhR)emediated effects, oxidative stress, chronic inflammation and immunosuppression [8]. It has been observed that the AhR receptordwhose activation is one of the key events linked to carcinogenesis mediated by dioxin-like PCBs [22]dnot only mediates xenobiotic effects on melanogenesis [23], but is also involved in melanoma progression [24]. PCBs also modulate cellular signalling pathways leading to increased formation of reactive oxygen species and subsequent increase in oxidative stress [25,26] as well as the induction of chronic inflammation [27,28]. Whereas that oxidative stress appears to be crucial in the pathogenesis of melanocyte transformation and melanoma progression [2,29], the
potential role of the inflammatory microenvironment in the pathogenesis and progression has also been highlighted [30e32]. Furthermore, PCBs have been shown to have a suppressor effect on the human immune system [33] which is of particular relevance as immunosuppressed individuals are more prone to develop melanoma than the general population [34]. Interestingly, increased sensitivity to develop chloracne and nonHodgkin lymphomaeassociated chromosomal translocation was indicated after accidental dioxin-like compound exposure in Seveso, in light hair/eye coloured individuals as compared to those with darker pigmentation [35,36]. Whether this indicates a higher sensitivity to PCB-induced melanoma in the Nordic countries can only be speculated. The limited number of cases did, however, not allow for any meaningful stratification based on skin type in the present study. In contrast to PCBs, EPA and DHA have shown to exert different antineoplastic actions in melanoma cells [12,13,15,37] and the biological rationale for their protective role in the genesis of cancer [10], and specifically in melanoma [11] has been well established. The antiinflammatory activity of EPA and DHA involves inhibition of the cyclooxygenase-2 [10], whose overexpression has been observed in malignant melanoma [30]. Moreover, EPA and DHA have also been
142
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
associated with decreased markers of UV damage [38,39] and have shown to reduce UV-induced cutaneous immune suppression in a randomised clinical trial [40]. Consistent with this evidence, we observed a statistically significantly 80% lower risk of melanoma in women in the highest tertile of EPA-DHA intake compared with those in the lowest. This inverse association was only apparent after adjustment for dietary PCB exposure, suggesting that the protection exerted by EPA and DHA against melanoma development might be masked by the higher melanoma risk derived from PCB exposure. This fact could explain the conflicting results reported from observational studies on fish consumption and melanoma risk [41]. Unfortunately, because of the limited number of cases, we were not able to assess the effects of dietary PCB and EPA-DHA exposures separately in stratified analysis, and consequently, we have to tread carefully the interpretation of our results due to the potential collinearity derived from the high correlation between dietary PCBs and EPA-DHA intake (r Z 0.95). Strengths of this study include its prospective population-based design, the linkage to the Swedish Cancer Registry which allows us to perform virtually complete follow-up of melanoma cases, the validated data on dietary PCB and EPA-DHA exposures, as well as the availability of detailed information on wellknown factors related to melanoma risk such as sun habits and skin type. Furthermore, in view of the long latency of melanoma, as much as 10 years [42], to conduct the exposure assessment several years (in 1997) before starting the follow-up (in 2009) is a positive factor as we can be fairly certain that exposure precedes sufficiently the development of melanoma. Although we have to assume that dietary habits remained the same over the entire period, it has been recognised that eating patterns evolve over periods of years and that diets of individuals tend to be reasonably correlated year to year [43]. Despite acknowledged limitations such as the inevitable measurement errors derived from self-reported dietary and sun habits, which can lead to exposure misclassification, the results from this study may suggest that dietary PCB exposure could be positively associated with an increased risk of melanoma at the levels to which the general population is exposed. At the same time, EPA and DHA, obtained mainly through the intake of fatty fish, showed to have a substantial protective association against melanoma development. To adequately balance the risk and benefits of fish consumption is a key public health issue, and thus, further prospective studies in the general population corroborating these findings are warranted. Moreover, exploring any genetic predisposition for PCB-induced melanoma may provide important insights.
Funding The Swedish Cancer Society (120840) and the Swedish Research Council/Research Infrastructures (8252008-5997) supported the study. Conflict of interest statement None declared.
References [1] Greinert R, de Vries E, Erdmann F, Espina C, Auvinen A, Kesminiene A, et al. European code against cancer 4th edition: ultraviolet radiation and cancer. Cancer Epidemiol 2015; 39(Suppl. 1):S75e83. [2] Joosse A, de Vries E, van Eijck CH, Eggermont AM, Nijsten T, Coebergh JW. Reactive oxygen species and melanoma: an explanation for gender differences in survival? Pigment Cell Melanoma Res 2015;23(3):352e64. [3] IARC. IARC monographs on the evaluation of carcinogenic risks to humans. 100: a review of human carcinogens. Part D: radiation. Lyon: IARC; 2012. [4] Berwick M, Buller DB, Cust A, Gallagher R, Lee TK, Meyskens F, et al. Melanoma epidemiology and prevention. Cancer Treat Res 2016;167:17e49. [5] EFSA. Update of the monitoring of dioxins and PCBs levels in food and feed. EFSA J 2012;10(7):1e82. [6] Bergkvist C, Akesson A, Glynn A, Michae¨lsson K, Rantakokko P, Kiviranta H, et al. Validation of questionnairebased long-term dietary exposure to polychlorinated biphenyls using biomarkers. Mol Nutr Food Res 2012;56(11):1748e54. [7] Milbrath MO, Wenger Y, Chang C-W, Emond C, Garabrant D, Gillespie BW, et al. Apparent half-lives of dioxins, furans, and polychlorinated biphenyls as a function of age, body fat, smoking status, and breast-feeding. Environ Health Perspect 2009;117(3):417e25. [8] IARC. Polychlorinated biphenyls and polybrominated biphenyls. IARC Monogr Eval Carcinog Risk Chem Hum 2015;107. [9] Gallagher RP, Macarthur AC, Lee TK, Weber J-P, Leblanc A, Mark Elwood J, et al. Plasma levels of polychlorinated biphenyls and risk of cutaneous malignant melanoma: a preliminary study. Int J Cancer 2011;128(8):1872e80. [10] Larsson SC, Kumlin M, Ingelman-Sundberg M, Wolk A. Dietary long-chain n-3 fatty acids for the prevention of cancer: a review of potential mechanisms. Am J Clin Nutr 2004;79(6):935e45. [11] Serini S, Fasano E, Celleno L, Cittadini A, Calviello G. Potential of long-chain n-3 polyunsaturated fatty acids in melanoma prevention. Nutr Rev 2014;72(4):255e66. [12] Reich R, Royce L, Martin GR. Eicosapentaenoic acid reduces the invasive and metastatic activities of malignant tumor cells. Biochem Biophys Res Commun 1989;160(2):559e64. [13] Albino AP, Juan G, Traganos F, Reinhart L, Connolly J, Rose DP, et al. Cell cycle arrest and apoptosis of melanoma cells by docosahexaenoic acid: association with decreased pRb phosphorylation. Cancer Res 2000;60(15):4139e45. [14] Serini S, Fasano E, Piccioni E, Monego G, Cittadini ARM, Celleno L, et al. DHA induces apoptosis and differentiation in human melanoma cells in vitro: involvement of HuR-mediated COX-2 mRNA stabilization and b-catenin nuclear translocation. Carcinogenesis 2012;33(1):164e73. [15] Nehra D, Pan AH, Le HD, Fallon EM, Carlson SJ, Kalish BT, et al. Docosahexaenoic acid, G protein-coupled receptors, and melanoma: is G protein-coupled receptor 40 a potential therapeutic target? J Surg Res 2014;188(2):451e8.
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143 [16] Noel SE, Stoneham ACS, Olsen CM, Rhodes LE, Green AC. Consumption of omega-3 fatty acids and the risk of skin cancers: a systematic review and meta-analysis. Int J Cancer 2014;135(1): 149e56. ˚ kesson A, [17] Harris H, Ha˚kansson N, Olofsson C, Julin B, A Wolk A. The Swedish Mammography Cohort and the Cohort of Swedish Men: study design and characteristics of two populationbased longitudinal Cohorts. OA Epidemiol 2013;1(2):16. [18] Covaci A, Koppen G, Van Cleuvenbergen R, Schepens P, Winneke G, van Larebeke N, et al. Persistent organochlorine pollutants in human serum of 50e65 years old women in the Flanders Environmental and Health Study (FLEHS). Part 2: correlations among PCBs, PCDD/PCDFs and the use of predictive markers. Chemosphere 2002;48(8):827e32. [19] Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 1986;124(1):17e27. [20] Wallin A, Di Giuseppe D, Burgaz A, Ha˚kansson N, Cederholm T, Michae¨lsson K, et al. Validity of food frequency questionnaire-based estimates of long-term long-chain n-3 polyunsaturated fatty acid intake. Eur J Nutr 2014;53(2):549e55. [21] Mattsson B, Wallgren A. Completeness of the Swedish Cancer Register. Non-notified cancer cases recorded on death certificates in 1978. Acta Radiol Oncol 1984;23(5):305e13. [22] Beischlag TV, Luis Morales J, Hollingshead BD, Perdew GH. The aryl hydrocarbon receptor complex and the control of gene expression. Crit Rev Eukaryot Gene Expr 2008;18(3):207e50. [23] Luecke S, Backlund M, Jux B, Esser C, Krutmann J, Rannug A. The aryl hydrocarbon receptor (AHR), a novel regulator of human melanogenesis. Pigment Cell Melanoma Res 2010;23(6): 828e33. [24] Villano CM, Murphy KA, Akintobi A, White LA. 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) induces matrix metalloproteinase (MMP) expression and invasion in A2058 melanoma cells. Toxicol Appl Pharmacol 2006;210(3):212e24. [25] Green RM, Hodges NJ, Chipman JK, O’Donovan MR, Graham M. Reactive oxygen species from the uncoupling of human cytochrome P450 1B1 may contribute to the carcinogenicity of dioxin-like polychlorinated biphenyls. Mutagenesis 2008; 23(6):457e63. [26] Schlezinger JJ, Struntz WDJ, Goldstone JV, Stegeman JJ. Uncoupling of cytochrome P450 1A and stimulation of reactive oxygen species production by co-planar polychlorinated biphenyl congeners. Aquat Toxicol Amst Neth 2006;77(4):422e32. [27] Hennig B, Meerarani P, Slim R, Toborek M, Daugherty A, Silverstone AE, et al. Proinflammatory properties of coplanar PCBs: in vitro and in vivo evidence. Toxicol Appl Pharmacol 2002;181(3):174e83. [28] Kwon O, Lee E, Moon TC, Jung H, Lin CX, Nam K-S, et al. Expression of cyclooxygenase-2 and pro-inflammatory cytokines induced by 2,20 ,4,40 ,5,50 -hexachlorobiphenyl (PCB 153) in human mast cells requires NF-kappa B activation. Biol Pharm Bull 2002; 25(9):1165e8.
143
[29] Meyskens FL, Farmer PJ, Anton-Culver H. Etiologic pathogenesis of melanoma: a unifying hypothesis for the missing attributable risk. Clin Cancer Res 2004;10(8):2581e3. [30] Denkert C, Ko¨bel M, Berger S, Siegert A, Leclere A, Trefzer U, et al. Expression of cyclooxygenase 2 in human malignant melanoma. Cancer Res 2001;61(1):303e8. [31] Melnikova VO, Bar-Eli M. Inflammation and melanoma metastasis. Pigment Cell Melanoma Res 2009;22(3):257e67. [32] Sanz-Motilva V, Martorell-Calatayud A, Nagore E. Non-steroidal anti-inflammatory drugs and melanoma. Curr Pharm Des 2012;18(26):3966e78. [33] Haase H, Fahlenkamp A, Schettgen T, Esser A, Gube M, Ziegler P, et al. Immunotoxicity monitoring in a population exposed to polychlorinated biphenyls. Int J Environ Res Public Health 2016;13(3):295. [34] Hollenbeak CS, Todd MM, Billingsley EM, Harper G, Dyer AM, Lengerich EJ. Increased incidence of melanoma in renal transplantation recipients. Cancer 2005;104(9):1962e7. [35] Baccarelli A, Pesatori AC, Consonni D, Mocarelli P, Patterson DG, Caporaso NE, et al. Health status and plasma dioxin levels in chloracne cases 20 years after the Seveso, Italy accident. Br J Dermatol 2005;152(3):459e65. [36] Baccarelli A, Hirt C, Pesatori AC, Consonni D, Patterson DG, Bertazzi PA, et al. t(14;18) translocations in lymphocytes of healthy dioxin-exposed individuals from Seveso, Italy. Carcinogenesis 2006;27(10):2001e7. [37] Serini S, Fasano E, Piccioni E, Cittadini ARM, Calviello G. Differential anti-cancer effects of purified EPA and DHA and possible mechanisms involved. Curr Med Chem 2011;18(26): 4065e75. [38] Rhodes LE, Shahbakhti H, Azurdia RM, Moison RMW, Steenwinkel M-JST, Homburg MI, et al. Effect of eicosapentaenoic acid, an omega-3 polyunsaturated fatty acid, on UVRrelated cancer risk in humans. An assessment of early genotoxic markers. Carcinogenesis 2003;24(5):919e25. [39] van der Pols JC, Xu C, Boyle GM, Hughes MC, Carr SJ, Parsons PG, et al. Serum omega-3 and omega-6 fatty acids and cutaneous p53 expression in an Australian population. Cancer Epidemiol Biomark Prev 2011;20(3):530e6. [40] Pilkington SM, Massey KA, Bennett SP, Al-Aasswad NM, Roshdy K, Gibbs NK, et al. Randomized controlled trial of oral omega-3 PUFA in solar-simulated radiation-induced suppression of human cutaneous immune responses. Am J Clin Nutr 2013; 97(3):646e52. [41] de Waure C, Quaranta G, Gualano MR, Cadeddu C, JovicVranes A, Djikanovic B, et al. Systematic review of studies investigating the association between dietary habits and cutaneous malignant melanoma. Public Health 2015;129(8):1099e113. [42] Liu T, Soong SJ. Epidemiology of malignant melanoma. Surg Clin North Am 1996;76(6):1205e22. [43] Rothman KJ, Greenland S, Lash TL. Modern epidemiology. 3rd ed. Philadelphia: Lippincott, Williams & Wilkins; 2008.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Original Research
Dietary polychlorinated biphenyls, long-chain n-3 polyunsaturated fatty acids and incidence of malignant melanoma Carolina Donat-Vargas a, Marika Berglund a, Anders Glynn b, ˚ kesson a,* Alicja Wolk a, Agneta A a b
Institute of Environmental Medicine, Karolinska Institutet, Box 210, SE-171 77, Stockholm, Sweden Department of Risk and Benefit Assessment, National Food Agency, Box 622, SE-751 26, Uppsala, Sweden
Received 28 September 2016; received in revised form 25 October 2016; accepted 21 November 2016
KEYWORDS Polychlorinated biphenyls (PCBs); Long-chain n-3 polyunsaturated fatty acids (LC n-3 PUFAs); Cutaneous malignant melanoma; Nutritional epidemiology; Cancer prevention
Abstract Background: For malignant melanoma, other risk factors aside from sun exposure have been hardly explored. Polychlorinated biphenyls (PCBs)dmainly from fatty fishd may affect melanogenesis and promote melanoma progression, while long-chain n-3 polyunsaturated fatty acids seem to exert antineoplastic actions in melanoma cells. Objectives: We aimed to assess the association of validated estimates of dietary PCB exposure as well as the intake of eicosapentaenoic acid and docosahexaenoic acid (EPA-DHA), accounting for sun habits and skin type, with the risk of malignant melanoma in middle-aged and elderly women. Methods: We included 20,785 women at baseline in 2009 from the prospective populationbased Swedish Mammography Cohort. Validated estimates of dietary PCB exposure and EPA-DHA intake were obtained via a food frequency questionnaire. Incident melanoma cases were ascertained through register-linkage. Results: During 4.5 years of follow-up, we ascertained 67 incident cases of melanoma. After multivariable adjustments, exposure to dietary PCBs was associated with four-fold increased risk of malignant melanoma (hazard ratio [HR], 4.0 [95% confidence interval {CI}, 1.2e13; P for trend Z 0.02]), while EPA-DHA intake was associated with 80% lower risk (HR, 0.2 [95% CI, 0.1e0.8; P for trend Z 0.03]), comparing the highest exposure tertiles with the lowest.
* Corresponding author: Karolinska Institutet, Institute of Environmental Medicine, Nutritional Epidemiology, Box 210, 171 77, Stockholm, Sweden. ˚ kesson). E-mail address: [email protected] (A. A http://dx.doi.org/10.1016/j.ejca.2016.11.016 0959-8049/ª 2016 Elsevier Ltd. All rights reserved.
138
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
Conclusion: While we found a direct association between dietary PCB exposure and risk of melanoma, EPA-DHA intake showed to have a substantial protective association. Question of benefits and risk from fish consumption is very relevant and further prospective studies in the general population verifying these findings are warranted. ª 2016 Elsevier Ltd. All rights reserved.
1. Introduction The incidence of malignant melanomadthe most lethal skin cancer due its high potential for metastasisdhas increased abruptly over the past 50 years [1]. In Europe, incidence rates are particularly high in the Nordic countries and generally higher among women than among men [1], although with a female survival advantage [2]. Despite ultraviolet (UV) radiation is the major risk factor [3], other environmental factors such as chemical exposures may play a part in the aetiology of melanoma [4]. Polychlorinated biphenyls (PCBs) are extremely persistent synthetic organochlorine chemicals, with a well-known potential toxicity, that have been widely dispersed into the environment for decades, bioaccumulating and magnifying in the food chain. The general population is exposed to PCBs primarily via food, where fatty fish is a major contributor to dietary PCB exposure in populations with relatively high fish consumption [5,6]. PCBs are readily absorbed, distributed in the body and accumulated in adipose tissue, with half-life for the most persistent congeners ranging from a couple of years to decades [7]. The International Agency for Research on Cancer has recently upgraded PCBs to group 1, i.e. carcinogenic to humans [8]. The strongest evidence on PCB carcinogenicity in humans comes from epidemiological data on risk of malignant melanoma. Virtually all studies on occupational and accidental PCB exposure report excess risk [8]. To date, the only study conducted in a general population is a caseecontrol study with 80 malignant melanoma cases and 310 controls [9]. A clear statistically significant higher odds was observed for the highest compared with the lowest quartile of total plasma PCB concentrations (odds ratio, 6.0; 95% confidence intervals [CI], 2.0e18.2), after adjusting for phenotypic factors, sun sensitivity and sun exposure. Apart from being a major source of PCBs, fatty fish is also the main source of long-chain n-3 polyunsaturated fatty acids (LC n-3 PUFAs), proposed to protect against several types of cancer [10] including melanoma [11]. Although in vitro and animal studies have provided convincing evidence to support these favourable effects of LC n-3 PUFAs (mainly eicosapentaenoic acid [EPA; 20:5 n-3] and docosahexaenoic acid [DHA, 22:6 n-
3]) against melanoma [12e15], the existing evidence from epidemiological studies exploring this relationship is very limited [11,16]. The aim of the present study was to assess the association between validated estimates of dietary PCB exposure and LC n-3 PUFA intake with the risk of malignant melanoma in a population-based prospective cohort study of middle-aged and elderly Swedish women. 2. Methods 2.1. Study population The Swedish Mammography Cohort is a large population-based longitudinal cohort established in 1987e1990. All women born between 1914 and 1948, resident in two counties in Central Sweden (n Z 90,303), were invited to complete a self-administrated questionnaire concerning diet (response rate 74%). In 1997, a more detailed questionnaire was sent to all participants still alive and living in the study area (response rate 70%). In 2009, an additional questionnaire that sought information on sun habits and skin type (information not available until then) was distributed to all remaining cohort members (response rate 84%). More details on the study design have been published elsewhere [17]. In the present study, the 1997 questionnaire was used for the exposure assessment, as there was sufficient documentation of the PCB content in different foods at that time [6]. In order to be able to control for sun habits and skin type (the major known risk factors for melanoma), we used the 2009 questionnaire as baseline in the present study (n Z 21,818). For the final analyses, we excluded those women with implausible total energy intake (3 standard deviation from the log-transformed mean), those who had a diagnosis of cancer before 1997 and those with prevalent malignant melanoma at baseline (1 May 2009), resulting in 20,785 women available at the start of follow-up. 2.2. Dietary PCB exposure and other covariates The 1997 questionnaire included a 96-item food frequency questionnaire (FFQ), constructed to reflect the women’s average consumption of different foods and beverages during the last year. The FFQ has been
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
validated, obtaining Pearson correlation coefficients (r) between the average of four 1-week dietary records and the FFQ of 0.6 for fatty fish and 0.5e0.7 for dairy products (Wolk A, unpublished results). The dietary exposure to PCBs was estimated through an extensive recipe-based database created for the 1997 FFQ, described in detail elsewhere [5]. This database was based on concentrations of the PCB-153 congener, which is the most abundant congener in food and therefore a very good indicator of both total PCBs and dioxin-like PCBs in food as well as in human serum [6,8,18]. Daily dietary exposure to PCBs (ng day1) and to EPA-DHA (g day1) was estimated by multiplying the average concentration in various foods (obtained from the Food Agency’s food control and monitoring programmes and the Swedish food composition database) with the respective consumption frequency and portion size. In order to insure that the association of PCB and EPA-DHA intakes with the outcome is independent of the total energy intake, estimated PCBs were adjusted for total daily energy intake (mean 1700 kcal) using the residual-regression method [19]. The FFQ-based dietary estimate of PCB exposure has been extensively validated against six serum PCB congeners in a representative subsample of the cohort (Spearman correlation coefficients [r] ranged from 0.30 to 0.58) [6]. The FFQ-based dietary intake of EPADHA has also been validated against adipose tissue concentrations of EPA and DHA (r Z 0.32 and 0.48, respectively, for concurrent exposure, and r Z 0.21 and 0.33, respectively, for past exposure [6 years prior to the adipose tissue sampling]) [21]. From the 2009 questionnaire we obtained information on skin type and sun habits, comprising use of sunblock, travel to sunny resorts during winter time and sun behaviour. The other variables were obtained from the 1997 questionnaire. 2.3. Case ascertainment Incident cases of malignant melanoma (code C43 in International Classification of Diseases-10) were ascertained from the Swedish Cancer Register, which is close to 100% completion [20]. 2.4. Statistical analyses Person-time of follow-up was calculated for each participant from 1 May 2009 until the date of melanoma diagnosis, death (through register-linkage) or end of follow-up (December 31, 2013), whichever occurred first. To assess the relationship between tertiles of dietary PCB and EPA-DHA intakes and the subsequent risk of developing malignant melanoma, hazard ratios and 95% CI were estimated using Cox
139
proportional hazard regression models, with attained age as the underlying timescale. No evidence of departure from the proportional hazards assumption was detected. To test for linear trends across increasing categories of dietary PCB exposure and EPA-DHA intake we assigned the median exposure within each category and included it as a continuous variable. The models were adjusted for the following potential confounders: level of education, factors related to the retention of PCB in the body (i.e. parity and body mass index), skin type (based on the skin reaction to the sun), sun habits (i.e. use of sunblock, travel to sunny resorts during winter time and sun behaviour), smoking habits, alcohol consumption and total energy intake. In a separate model mutual adjustments for dietary PCB exposure and EPA-DHA intake were performed. We fitted the multivariable-adjusted models including the missing values (below 3% for all covariates with the exception of skin type, 6%) in a separate category. Analyses were run in Stata version 13.0 (StataCorp LP), with statistical significance set at the two-sided 0.05 level. 3. Results During 4.5 years of follow-up (94,200 person-years), 67 cases of incident malignant melanoma were ascertained. The average age at diagnosis of melanoma was 73 (8) years. The median energy-adjusted dietary PCB exposure was 162 ng day1 (5the95th percentile 74e346 ng day1) and the median energy-adjusted dietary EPA-DHA intake was 0.3 g day1 (5the95th percentile 0.1e0.7 g day1). No major differences in age-standardised characteristics were observed across tertiles of dietary PCB exposure and EPA-DHA intake (Table 1). We observed no association between dietary PCB exposure and risk of melanoma before adjusting for LC n-3 PUFAs. However, in the complete multivariableadjusted model additionally adjusted for EPA-DHA intake, dietary PCB exposure was associated with a statistically significant four-fold higher risk of melanoma comparing the highest PCB tertile with the lowest tertile (P for trend Z 0.02). Likewise, the inverse relationship between dietary EPA-DHA intake and melanoma development only reached statistical significance after adjustment for dietary PCB exposure, resulting in an 80% lower risk of melanoma when comparing women in the highest and lowest tertiles of EPA-DHA intake (P for trend Z 0.03) (Table 2; Fig. 1). 4. Discussion Melanoma is a potentially fatal malignancy that continues to increase in incidence despite public education measures to limit sun exposure. Conceivable risk
140
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
Table 1 Age-standardised baseline characteristics of 20,785 women from the Swedish Mammography Cohort by tertiles of energy-adjusted dietary PCB and EPA-DHA exposures. Characteristicsa
Tertiles of dietary PCB exposure (ng day1)
Tertiles of EPA-DHA intake (g day1)
Range (median) n Age, yearsb Postsecondary education, % Nulliparous, % Body mass index, kg/m2 Skin reaction to the sun, %c Always red/never tanned Always tanned/never red Use of sunblock always or most of the time, %c Travelling yearly to sunny resorts during winter time, %c Preference on a sunny day, %c Sun all the time Shade all the time Current smoker, % Alcohol consumption, % No use >15 g/day Total energy intake, kcal day1 Dietary PCB exposure, ng day1 EPA-DHA intake, g day1
<141 (113) 6929 70 8 31 9 24.9 0.05
141e188 (162) 6928 70 7 35 8 24.8 0.04
>188 (246) 6928 72 8 32 9 25.2 0.05
<0.24 (0.18) 6929 71 8 31 9 24.8 0.05
0.24e0.35 (0.29) 6928 70 7 34 8 24.9 0.04
>0.35 (0.46) 6928 72 8 33 9 25.3 0.05
5 17 35
4 16 38
5 19 38
5 17 34
4 15 38
4 19 39
7
7
8
6
7
8
4 15 23
3 13 21
3 15 22
4 15 23
3 13 21
3 14 22
15 9 1820 6
9 10 1735 5
10 10 1710 6
0.17 <0.01
0.29 <0.01
0.53 <0.01
16 8 1840 6 110 0.4
9 10 1750 5 163 0.3
10 11 1675 6 276 1.6
All variables are expressed as mean standard deviation or percentage (%); EPA Z eicosapentaenoic acid; DHA Z docosahexaenoic acid; PCB Z Polychlorinated biphenyl. a Unless otherwise indicated, acquired from the 1997 questionnaire. b At the beginning of follow-up (1 May 2009). c Acquired from the 2009 questionnaire.
factors aside from sun exposure such as diet and dietary contaminants are scarcely explored. Within this large population-based prospective cohort, while exposure to dietary PCBs was associated with a statistically
significant four-fold increased risk of malignant melanoma, LC n-3 PUFAs were associated with 80% lower risk in mutually-adjusted models, comparing the highest exposure tertiles with the lowest.
Table 2 Hazard Ratio (95% confidence intervals) of malignant melanoma by tertiles of energy-adjusted dietary PCB and EPA-DHA exposures in 20,785 women from the Swedish Mammography Cohort. Tertiles of dietary PCB exposure 1
Range (median), ng day Cases Person-years Age-adjusted HR Multivariable-adjusted HRa Multivariable-adjusted HRb
<141 (113) 22 31,403 1 (Ref.) 1 (Ref.) 1 (Ref.)
141e188 (162) 19 31,561 0.9 (0.5e1.6) 0.8 (0.5e1.6) 1.5 (0.7e3.5)
P trend >188 (246) 26 31,236 1.2 (0.7e2.1) 1.2 (0.7e2.1) 4.0 (1.2e13)
Tertiles of EPA-DHA intake 1
Range (median), g day Cases Person-years Age-adjusted HR Multivariable-adjusted HRa Multivariable-adjusted HRc
<0.24 (0.18) 26 31,379 1 (Ref.) 1 (Ref.) 1 (Ref.)
0.47 0.44 0.02 P trend
0.24e0.35 (0.29) 19 31,525 0.7 (0.4e1.3) 0.7 (0.4e1.3) 0.4 (0.9e1.0)
>0.35 (0.46) 22 31,296 0.9 (0.5e1.5) 0.9 (0.5e1.5) 0.2 (0.1e0.8)
0.63 0.67 0.03
EPA Z eicosapentaenoic acid; DHA Z docosahexaenoic acid; PCB Z Polychlorinated biphenyl. a Adjusted for attained age (years), postsecondary education (yes/no), parity (0, 1e2, 3 child), body mass index (<18.5, 18.5e25, 25e30, >30 kg/m2), skin reaction to the sun (always red/never tanned, always red/sometimes tanned, sometimes red/always tanned, never red/always tanned), use of sunblock (always, most of the time, sometimes, never), travel to sunny resorts during winter time (no, sometimes, every year), preference on a sunny day (sun all the time, both sun and shade, shade all the time), smoking habits (current, former, never), alcohol consumption (no use, 0.1e5, 5.1e15, >15 g/day) and total energy intake (continuous, kcal/day). b Additionally adjusted for dietary EPA-DHA intake (tertiles). c Additionally adjusted for dietary PCB exposure (tertiles).
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
141
Fig. 1. Hazard ratios (95% confidence intervals) of malignant melanoma by tertiles of energy-adjusted dietary PCB (ng day1) and EPADHA (g day1) exposures in 20,785 women from the Swedish Mammography Cohort. EPA Z eicosapentaenoic acid; DHA Z docosahexaenoic acid; PCB Z Polychlorinated biphenyl. Adjusted for attained age (years), postsecondary education (yes/no), parity (0, 1e2, 3 child), body mass index (<18.5, 18.5e25, 25e30, >30 kg/m2), skin reaction to the sun (always red/never tanned, always red/sometimes tanned, sometimes red/always tanned, never red/always tanned), use of sunblock (always, most of the time, sometimes, never), travel to sunny resorts during winter time (no, sometimes, every year), preference on a sunny day (sun all the time, both sun and shade, shade all the time), smoking habits (current, former, never), alcohol consumption (no use, 0.1e5, 5.1e15, >15 g/day), total energy intake (continuous, kcal/day) and mutually adjusted for dietary EPA-DHA intake (tertiles) and dietary PCB exposure (tertiles).
This is the first prospective study to evaluate the association between validated estimates of dietary PCB exposure and the incidence of melanoma. The results are in line with the only previous study on melanoma and PCB exposure conducted in the general population [9], which found an almost five-fold higher odds (95% CI, 1.7e14.1) of melanoma for those in the highest quartile of the plasma PCB-153 congener compared to those in the lowest. Animal studies have shown evidence of PCB carcinogenicity in different tissues and several common PCB carcinogenic mechanisms have been established, including the aryl hydrocarbon receptor (AhR)emediated effects, oxidative stress, chronic inflammation and immunosuppression [8]. It has been observed that the AhR receptordwhose activation is one of the key events linked to carcinogenesis mediated by dioxin-like PCBs [22]dnot only mediates xenobiotic effects on melanogenesis [23], but is also involved in melanoma progression [24]. PCBs also modulate cellular signalling pathways leading to increased formation of reactive oxygen species and subsequent increase in oxidative stress [25,26] as well as the induction of chronic inflammation [27,28]. Whereas that oxidative stress appears to be crucial in the pathogenesis of melanocyte transformation and melanoma progression [2,29], the
potential role of the inflammatory microenvironment in the pathogenesis and progression has also been highlighted [30e32]. Furthermore, PCBs have been shown to have a suppressor effect on the human immune system [33] which is of particular relevance as immunosuppressed individuals are more prone to develop melanoma than the general population [34]. Interestingly, increased sensitivity to develop chloracne and nonHodgkin lymphomaeassociated chromosomal translocation was indicated after accidental dioxin-like compound exposure in Seveso, in light hair/eye coloured individuals as compared to those with darker pigmentation [35,36]. Whether this indicates a higher sensitivity to PCB-induced melanoma in the Nordic countries can only be speculated. The limited number of cases did, however, not allow for any meaningful stratification based on skin type in the present study. In contrast to PCBs, EPA and DHA have shown to exert different antineoplastic actions in melanoma cells [12,13,15,37] and the biological rationale for their protective role in the genesis of cancer [10], and specifically in melanoma [11] has been well established. The antiinflammatory activity of EPA and DHA involves inhibition of the cyclooxygenase-2 [10], whose overexpression has been observed in malignant melanoma [30]. Moreover, EPA and DHA have also been
142
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143
associated with decreased markers of UV damage [38,39] and have shown to reduce UV-induced cutaneous immune suppression in a randomised clinical trial [40]. Consistent with this evidence, we observed a statistically significantly 80% lower risk of melanoma in women in the highest tertile of EPA-DHA intake compared with those in the lowest. This inverse association was only apparent after adjustment for dietary PCB exposure, suggesting that the protection exerted by EPA and DHA against melanoma development might be masked by the higher melanoma risk derived from PCB exposure. This fact could explain the conflicting results reported from observational studies on fish consumption and melanoma risk [41]. Unfortunately, because of the limited number of cases, we were not able to assess the effects of dietary PCB and EPA-DHA exposures separately in stratified analysis, and consequently, we have to tread carefully the interpretation of our results due to the potential collinearity derived from the high correlation between dietary PCBs and EPA-DHA intake (r Z 0.95). Strengths of this study include its prospective population-based design, the linkage to the Swedish Cancer Registry which allows us to perform virtually complete follow-up of melanoma cases, the validated data on dietary PCB and EPA-DHA exposures, as well as the availability of detailed information on wellknown factors related to melanoma risk such as sun habits and skin type. Furthermore, in view of the long latency of melanoma, as much as 10 years [42], to conduct the exposure assessment several years (in 1997) before starting the follow-up (in 2009) is a positive factor as we can be fairly certain that exposure precedes sufficiently the development of melanoma. Although we have to assume that dietary habits remained the same over the entire period, it has been recognised that eating patterns evolve over periods of years and that diets of individuals tend to be reasonably correlated year to year [43]. Despite acknowledged limitations such as the inevitable measurement errors derived from self-reported dietary and sun habits, which can lead to exposure misclassification, the results from this study may suggest that dietary PCB exposure could be positively associated with an increased risk of melanoma at the levels to which the general population is exposed. At the same time, EPA and DHA, obtained mainly through the intake of fatty fish, showed to have a substantial protective association against melanoma development. To adequately balance the risk and benefits of fish consumption is a key public health issue, and thus, further prospective studies in the general population corroborating these findings are warranted. Moreover, exploring any genetic predisposition for PCB-induced melanoma may provide important insights.
Funding The Swedish Cancer Society (120840) and the Swedish Research Council/Research Infrastructures (8252008-5997) supported the study. Conflict of interest statement None declared.
References [1] Greinert R, de Vries E, Erdmann F, Espina C, Auvinen A, Kesminiene A, et al. European code against cancer 4th edition: ultraviolet radiation and cancer. Cancer Epidemiol 2015; 39(Suppl. 1):S75e83. [2] Joosse A, de Vries E, van Eijck CH, Eggermont AM, Nijsten T, Coebergh JW. Reactive oxygen species and melanoma: an explanation for gender differences in survival? Pigment Cell Melanoma Res 2015;23(3):352e64. [3] IARC. IARC monographs on the evaluation of carcinogenic risks to humans. 100: a review of human carcinogens. Part D: radiation. Lyon: IARC; 2012. [4] Berwick M, Buller DB, Cust A, Gallagher R, Lee TK, Meyskens F, et al. Melanoma epidemiology and prevention. Cancer Treat Res 2016;167:17e49. [5] EFSA. Update of the monitoring of dioxins and PCBs levels in food and feed. EFSA J 2012;10(7):1e82. [6] Bergkvist C, Akesson A, Glynn A, Michae¨lsson K, Rantakokko P, Kiviranta H, et al. Validation of questionnairebased long-term dietary exposure to polychlorinated biphenyls using biomarkers. Mol Nutr Food Res 2012;56(11):1748e54. [7] Milbrath MO, Wenger Y, Chang C-W, Emond C, Garabrant D, Gillespie BW, et al. Apparent half-lives of dioxins, furans, and polychlorinated biphenyls as a function of age, body fat, smoking status, and breast-feeding. Environ Health Perspect 2009;117(3):417e25. [8] IARC. Polychlorinated biphenyls and polybrominated biphenyls. IARC Monogr Eval Carcinog Risk Chem Hum 2015;107. [9] Gallagher RP, Macarthur AC, Lee TK, Weber J-P, Leblanc A, Mark Elwood J, et al. Plasma levels of polychlorinated biphenyls and risk of cutaneous malignant melanoma: a preliminary study. Int J Cancer 2011;128(8):1872e80. [10] Larsson SC, Kumlin M, Ingelman-Sundberg M, Wolk A. Dietary long-chain n-3 fatty acids for the prevention of cancer: a review of potential mechanisms. Am J Clin Nutr 2004;79(6):935e45. [11] Serini S, Fasano E, Celleno L, Cittadini A, Calviello G. Potential of long-chain n-3 polyunsaturated fatty acids in melanoma prevention. Nutr Rev 2014;72(4):255e66. [12] Reich R, Royce L, Martin GR. Eicosapentaenoic acid reduces the invasive and metastatic activities of malignant tumor cells. Biochem Biophys Res Commun 1989;160(2):559e64. [13] Albino AP, Juan G, Traganos F, Reinhart L, Connolly J, Rose DP, et al. Cell cycle arrest and apoptosis of melanoma cells by docosahexaenoic acid: association with decreased pRb phosphorylation. Cancer Res 2000;60(15):4139e45. [14] Serini S, Fasano E, Piccioni E, Monego G, Cittadini ARM, Celleno L, et al. DHA induces apoptosis and differentiation in human melanoma cells in vitro: involvement of HuR-mediated COX-2 mRNA stabilization and b-catenin nuclear translocation. Carcinogenesis 2012;33(1):164e73. [15] Nehra D, Pan AH, Le HD, Fallon EM, Carlson SJ, Kalish BT, et al. Docosahexaenoic acid, G protein-coupled receptors, and melanoma: is G protein-coupled receptor 40 a potential therapeutic target? J Surg Res 2014;188(2):451e8.
C. Donat-Vargas et al. / European Journal of Cancer 72 (2017) 137e143 [16] Noel SE, Stoneham ACS, Olsen CM, Rhodes LE, Green AC. Consumption of omega-3 fatty acids and the risk of skin cancers: a systematic review and meta-analysis. Int J Cancer 2014;135(1): 149e56. ˚ kesson A, [17] Harris H, Ha˚kansson N, Olofsson C, Julin B, A Wolk A. The Swedish Mammography Cohort and the Cohort of Swedish Men: study design and characteristics of two populationbased longitudinal Cohorts. OA Epidemiol 2013;1(2):16. [18] Covaci A, Koppen G, Van Cleuvenbergen R, Schepens P, Winneke G, van Larebeke N, et al. Persistent organochlorine pollutants in human serum of 50e65 years old women in the Flanders Environmental and Health Study (FLEHS). Part 2: correlations among PCBs, PCDD/PCDFs and the use of predictive markers. Chemosphere 2002;48(8):827e32. [19] Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses. Am J Epidemiol 1986;124(1):17e27. [20] Wallin A, Di Giuseppe D, Burgaz A, Ha˚kansson N, Cederholm T, Michae¨lsson K, et al. Validity of food frequency questionnaire-based estimates of long-term long-chain n-3 polyunsaturated fatty acid intake. Eur J Nutr 2014;53(2):549e55. [21] Mattsson B, Wallgren A. Completeness of the Swedish Cancer Register. Non-notified cancer cases recorded on death certificates in 1978. Acta Radiol Oncol 1984;23(5):305e13. [22] Beischlag TV, Luis Morales J, Hollingshead BD, Perdew GH. The aryl hydrocarbon receptor complex and the control of gene expression. Crit Rev Eukaryot Gene Expr 2008;18(3):207e50. [23] Luecke S, Backlund M, Jux B, Esser C, Krutmann J, Rannug A. The aryl hydrocarbon receptor (AHR), a novel regulator of human melanogenesis. Pigment Cell Melanoma Res 2010;23(6): 828e33. [24] Villano CM, Murphy KA, Akintobi A, White LA. 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) induces matrix metalloproteinase (MMP) expression and invasion in A2058 melanoma cells. Toxicol Appl Pharmacol 2006;210(3):212e24. [25] Green RM, Hodges NJ, Chipman JK, O’Donovan MR, Graham M. Reactive oxygen species from the uncoupling of human cytochrome P450 1B1 may contribute to the carcinogenicity of dioxin-like polychlorinated biphenyls. Mutagenesis 2008; 23(6):457e63. [26] Schlezinger JJ, Struntz WDJ, Goldstone JV, Stegeman JJ. Uncoupling of cytochrome P450 1A and stimulation of reactive oxygen species production by co-planar polychlorinated biphenyl congeners. Aquat Toxicol Amst Neth 2006;77(4):422e32. [27] Hennig B, Meerarani P, Slim R, Toborek M, Daugherty A, Silverstone AE, et al. Proinflammatory properties of coplanar PCBs: in vitro and in vivo evidence. Toxicol Appl Pharmacol 2002;181(3):174e83. [28] Kwon O, Lee E, Moon TC, Jung H, Lin CX, Nam K-S, et al. Expression of cyclooxygenase-2 and pro-inflammatory cytokines induced by 2,20 ,4,40 ,5,50 -hexachlorobiphenyl (PCB 153) in human mast cells requires NF-kappa B activation. Biol Pharm Bull 2002; 25(9):1165e8.
143
[29] Meyskens FL, Farmer PJ, Anton-Culver H. Etiologic pathogenesis of melanoma: a unifying hypothesis for the missing attributable risk. Clin Cancer Res 2004;10(8):2581e3. [30] Denkert C, Ko¨bel M, Berger S, Siegert A, Leclere A, Trefzer U, et al. Expression of cyclooxygenase 2 in human malignant melanoma. Cancer Res 2001;61(1):303e8. [31] Melnikova VO, Bar-Eli M. Inflammation and melanoma metastasis. Pigment Cell Melanoma Res 2009;22(3):257e67. [32] Sanz-Motilva V, Martorell-Calatayud A, Nagore E. Non-steroidal anti-inflammatory drugs and melanoma. Curr Pharm Des 2012;18(26):3966e78. [33] Haase H, Fahlenkamp A, Schettgen T, Esser A, Gube M, Ziegler P, et al. Immunotoxicity monitoring in a population exposed to polychlorinated biphenyls. Int J Environ Res Public Health 2016;13(3):295. [34] Hollenbeak CS, Todd MM, Billingsley EM, Harper G, Dyer AM, Lengerich EJ. Increased incidence of melanoma in renal transplantation recipients. Cancer 2005;104(9):1962e7. [35] Baccarelli A, Pesatori AC, Consonni D, Mocarelli P, Patterson DG, Caporaso NE, et al. Health status and plasma dioxin levels in chloracne cases 20 years after the Seveso, Italy accident. Br J Dermatol 2005;152(3):459e65. [36] Baccarelli A, Hirt C, Pesatori AC, Consonni D, Patterson DG, Bertazzi PA, et al. t(14;18) translocations in lymphocytes of healthy dioxin-exposed individuals from Seveso, Italy. Carcinogenesis 2006;27(10):2001e7. [37] Serini S, Fasano E, Piccioni E, Cittadini ARM, Calviello G. Differential anti-cancer effects of purified EPA and DHA and possible mechanisms involved. Curr Med Chem 2011;18(26): 4065e75. [38] Rhodes LE, Shahbakhti H, Azurdia RM, Moison RMW, Steenwinkel M-JST, Homburg MI, et al. Effect of eicosapentaenoic acid, an omega-3 polyunsaturated fatty acid, on UVRrelated cancer risk in humans. An assessment of early genotoxic markers. Carcinogenesis 2003;24(5):919e25. [39] van der Pols JC, Xu C, Boyle GM, Hughes MC, Carr SJ, Parsons PG, et al. Serum omega-3 and omega-6 fatty acids and cutaneous p53 expression in an Australian population. Cancer Epidemiol Biomark Prev 2011;20(3):530e6. [40] Pilkington SM, Massey KA, Bennett SP, Al-Aasswad NM, Roshdy K, Gibbs NK, et al. Randomized controlled trial of oral omega-3 PUFA in solar-simulated radiation-induced suppression of human cutaneous immune responses. Am J Clin Nutr 2013; 97(3):646e52. [41] de Waure C, Quaranta G, Gualano MR, Cadeddu C, JovicVranes A, Djikanovic B, et al. Systematic review of studies investigating the association between dietary habits and cutaneous malignant melanoma. Public Health 2015;129(8):1099e113. [42] Liu T, Soong SJ. Epidemiology of malignant melanoma. Surg Clin North Am 1996;76(6):1205e22. [43] Rothman KJ, Greenland S, Lash TL. Modern epidemiology. 3rd ed. Philadelphia: Lippincott, Williams & Wilkins; 2008.