Xenohormone transactivities are inversely associated to serum POPs in Inuit

Background The persistent organic pollutants (POPs) are highly lipophilic and resistant to biodegradation and found in e.g. seafood and marine mammals. Greenlandic Inuit have high intake of marine food and thus high POP burden that varies according to local conditions and dietary preference. We do for the very first time report the serum POP related non-steroidal xenohormone activity of Inuit across Greenland. The aims were 1) to determine the integrated xenohormone bioactivities as an exposure biomarker of the actual lipophilic serum POP mixture measuring the effect on estrogen (ER) and androgen receptor (AR) transactivity in citizens from different Greenlandic districts and 2) to evaluate associations to serum POP markers (14 PCBs and 10 pesticides) and lifestyle characteristics. Methods Serum samples from 121 men and 119 women from Nuuk, Sisimiut and Qaanaaq were extracted using SPE-HPLC fractionation to obtain the serum POP fraction free of endogenous hormones. The serum POP fraction was used for determination of xenohormone transactivity using ER and AR reporter gene assays. Results In overall, the xenohormone transactivities differed between districts as well as between the genders. Associations between the transactivities and age, n-3/n-6 and smoker years were observed. The xenoestrogenic and xenoandrogenic transactivities correlated negatively to the POPs for the combined female and male data, respectively. Conclusion The non-steroidal xenohormone transactivities can be used as an integrated biomarker of POP exposure and lifestyle characteristics. The actual serum POP mixtures antagonized the age adjusted sex hormone receptor functions. Comparison of different study populations requires in addition to age inclusion of diet and lifestyle factors.


Introduction
The increasing load of the environment by man-made pollutants is of concern for the human health. Due to long-range transport by atmospheric and oceanic currents [1,2] the human exposure is not limited to individuals living close to the sources of the contaminants. The lipophilic persistent organic pollutants (POPs) includes polychlorinated dibenzo-p-dioxins/furans (PCDDs/ PCDFs), polychlorinated biphenyls (PCBs) and certain organochlor pesticide residues e.g. 2,2-bis(p-chlorophenyl)-1,1,1-trichloroethane (DDT), its metabolite 1,1dichloro-2,2-bis (p-chlorophenyl)-ethylene (p,p'-DDE), toxaphenes and chlordanes. Due to their high lipophilicity and resistance to biodegradation POPs are biomagnified through the food chain and found in fatty tissues at high concentrations in predator fish and birds, seals, whales and polar bears [3]. Greenlandic Inuit display high body burden of POPs [4], that significantly correlate with age, smoking and the level of n-3 polyunsaturated fatty acids in plasma, being strong indicators of the main source of POP contaminants in their traditional marine food [5][6][7][8][9].
In some Greenlandic districts the PCB body burden levels exceed Health Canada guidelines of concern (5 -100 μl/ L plasma) [10]. Also within Greenland strong regional differences are observed with the highest contaminant levels found in Inuit at the East coast [10], where also higher levels of PCBs and DDT are found in marine species and birds compared to the West coast [11,12]. Some districts and settlements such as Qaanaaq (North West) still primarily rely on traditional foods, whereas the diet in the cities Nuuk and Sisimiut (South West) are more westernized [10,13].
During the last two decades, a number of POPs have been identified and characterized as being estrogenic, antiestrogenic and/or antiandrogenic [14][15][16][17][18], and to have a multitude of potential health effects on wildlife and humans, including immunotoxicity, carcinogenicity, and adverse effects on reproductive, neurobehavioral, and endocrine functions [19][20][21][22][23][24]. Recently also endocrine-related human health effects of POPs on child development were reported at the individual or population level [25][26][27]. However, the toxicological assessment of POPs on humans is complicated. Not only is the analytical chemical approach for the detection of all xenobiotics practically impossible especially on large-scale surveys but additive enhancement of hormone actions has been reported in vitro for xenoestrogen mixtures [28,29] and recently in vivo for antiandrogens [30]. Therefore, the assessment of the integrated biological effect of the actual chemical mixture in human blood is important and ex vivo cell systems have recently been introduced to enable the assessment of the integrated level of xenobiotic transactiv-ity in human adipose tissue [20,31,32] or in human serum [33][34][35][36][37]. We have previously validated the SPE-HPLC serum POP extraction method and it was demonstrated that the ER-and AR-CALUX assays respond to a wide range of different compounds including PCBs and pesticides [14,15] as well as to PCB-spiked serum extracts [35].
The aims were 1) to determine the integrated xenohormone bioactivities as an exposure marker of the actual lipophilic serum POP mixture measuring the effect on estrogen (ER) and androgen receptor (AR) transactivity in cerum of citizens from different Greenlandic districts and 2) to evaluate associations of the xenohormone transactivities to serum POP markers (14 PCBs and 10 pesticides) and lifestyle characteristics.

SPE-HPLC fractionation of the serum samples
To obtain the serum fraction containing the actual mixture of bio-accumulated lipophilic POPs a solid phase extraction (SPE) and high performance liquid chromatography (HPLC) fractionation was performed on 3.6 ml serum [35]. The first fraction (F1: 0.00 -5.30 min.) was defined to include most POPs and free of endogenous hormones [35]. This SPE-HPLC F1 extract was evaporated and stored at -80°C for later ER or AR mediated CALUX assays.
A set of serum control samples were prepared by combining batches of male and female serum (blood bank of Aarhus Sygehus, DK), respectively, and distributed into 3.6 ml portions and stored at -80°C. One serum sample from each sex was, on a weekly basis, processed by the SPE-HPLC method in parallel with the study samples serving as serum control for the cleanup procedure.
On the day of analysis the SPE-HPLC F1 extracts (samples and controls) were thawed and processed as previously described [33,36]. The samples were analyzed randomly attempting to analyze samples from different districts in each independent assay.

Receptor chemically activated luciferase gene expression (CALUX) assays
SPE-HPLC serum extracts were analyzed in order to assess the ability of the lipophilic POPs to affect the ER or AR transactivity (referred to as XER and XAR, respectively). Moreover, we determined the effect of the serum extracts in the presence of the respective high affinity ligand of the receptors (the ER agonist 17β-estradiol (E2); the synthetic AR agonist methyltrienolone (R1881)) to mimic the effect in vivo assessing the ability of the lipophilic POPs to compete with the endogenous hormones/ligands (called XERcomp and XARcomp).  [36]. The determined luciferase activity per well was corrected for cell density using protein measurements and expressed in relative luciferase units per microgram protein (RLU/μg protein) equal to a response of 0.32 ml serum. The controls and the SPE-HPLC F1 serum extracts were determined at least in triplicate. If one of the triplicate values deviated more than 30% from the other two values, the mean was calculated from two wells only. The mean of the samples were related to the respective solvent controls.
In each independent assay a concentration-response control for the receptor ligand (E2 or R1881) was analyzed in parallel. The solvent controls ± E2 or R1881 were handled and analyzed in the same way as the SPE-HPLC F1 extracts [33,36]. The average ER-CALUX intra-coefficient of variation (CV) of the serum extracts and inter-CV of solvent controls were below 5%, and for AR-CALUX the intra-CV of the serum extracts and inter-CV of solvent controls were 11% and 14%, respectively. No cell toxicity was measured upon exposure of cells to SPE-HPLC extracts determined by CellTiter 96 Cell Proliferation assay from Promega (Madison WI, US) [33,36].

Statistical analysis
In each independent CALUX-assay transactivity differences between the triple serum extract determinations and their respective solvent controls (% agonistic, % decreased, % further increased and % antagonistic) were tested by the Student's T-test in Microsoft Excel (p ≤ 0.05).
To improve normal distribution the serum ER and AR transactivity and the POP data were natural logarithmic transformed for the statistical analysis. The analyses were performed on continuous data.
The comparisons of means between men and women were performed by Student's t-test. One-way ANOVA was used to compare levels of xenohormone transactivities and POPs among the districts. If differences were observed, multiple comparison ad hoc test were performed using the least significant difference (LSD) pair wise multiple comparison test for the variables with equal variance (p > 0.05) and Dunnett's T3 test for the variables with an unequal variance (p ≤ 0.05). The homogeneity of variance was tested by Levene's test.
Pearson's correlation analyses (two-tailed) were used to evaluate 1) the interrelationship between the 14 PCB congeners and the 10 pesticides and 2) associations between xenohormone transactivity and lifestyle characteristics.
Homogeneity or heterogeneity between the POP markers and xenohormone transactivity for each gender across the study groups were evaluated using multiple linear regressions analysis. Both the combined male or female data across the districts were found to be homogenous except for the XER activity of the combined men (see Additional file 1). Therefore, to obtain a better statistical power and the overall trend of the studied groups, the data based on the combined genders across the districts were evaluated assuming a common intercept.
To our knowledge no experience about which dietary and/ or other life-style characteristics that might influence the xenohormone transactivity is reported. We hypothesized that a potential predictor of POP bioaccumulation might also be a potential predictor for serum xenohormone transactivity. As known from literature age and seafood intake affect the serum POP levels [40]. Also body mass index (BMI), smoking years and bird intake might influence the serum POP levels [7]. Using the multivariate linear regression model, assessing the relation between the xenohormone transactivities and the POPs the impact of the lifestyle characteristics were evaluated by entering variables together with the POPs. Due to the small sample sizes and to obtain a higher power this was done for the combined genders across the districts.
The statistical analysis was performed in SPSS 13.0 (SPSS Inc, Chicago, IL). The term statistically significant level is used to denote a p-value ≤ 0.05.

Lifestyle characteristics of the study groups
Lifestyle characteristics that potentially influence the POP related ER and AR mediated transactivities of the volunteers are shown in Table 1. The age range of the volunteers was 18-72 years. The Nuuk men were older than the rest of the study groups, having similar age range. Furthermore, Nuuk men had the highest smoking year value, the highest n-3/n-6 fatty acid ratio and the highest level of seabird intake. BMI did not differ between districts or gender. Generally male participants had lower dairy food intake than women. Age and n-3/n-6 ratio was significantly correlated (p < 0.05) except for Nuuk women.

POP levels and intercorrelations
For PCB28, PCB52 and PCB128 more than 68% of the samples were at the detection limit (0.01 μg), whereas 37% and 26% of the samples were at the detection limit for PCB101 and PCB105, respectively. 6% of the samples were at the detection limit for PCB156 and PCB183, while less than 1% of the remaining PCBs were at the detection limit.
High intercorrelations between the PCBs (r > 0.66) were observed except for PCBs of lower concentration: PCB28 (0.24 < r < 0.64), PCB52 (0.24 < r < 0.64) and PCB128 (0.54 < r < 0.74). The pesticides were highly intercorrelated (r > 0.75). However, different POP intercorrelations were observed for the study groups. Generally in Nuuk and Qaanaaq the intercorrelations among PCBs and pes-ticides were higher than in Sisimiut, and higher intercorrelations were observed for men compared to women.
Male participants had significantly higher POP levels than women for all districts ( Table 2). For the men, similar median levels of ΣPCB, Σpesticide and ΣPOP were observed in Nuuk and Qaanaaq, being significantly higher than in Sisimiut. Qaanaaq women had significantly higher POP levels than that of Nuuk and Sisimiut women as also reported by Deutch et al [7].
Because of the differences between the genders and to improve the statistical power we mainly focus on the combined male and combined female data across districts.

Xenoestrogenic transactivities
In overall the serum POP extracts of both sexes predominantly decreased/antagonized the ER (XER) and E2-ER induced transactivities (XERcomp), respectively, and for all districts except Nuuk women (having equal % of further increased and antagonized samples) the median XERcomp value was below the E2 reference control level (Table 3). No differences in the xenoestrogenic activities were observed between men and women except for the XERcomp activity of Sisimiut men being significantly lower than Sisimiut women ( Table 3). The ANOVA test showed in overall significant differences between the districts for both sexes (Table 3).
Scatter plots of the xenoestrogenic transactivities versus ΣPOP for the single districts are shown in figure 1.

Correlation between xenoestrogenic transactivities and lifestyle characteristics
For the combined men positive correlations were observed for XER to n-3/n-6 fatty acid ratio and smoker years and for XERcomp to age and smoker years (Table 4).
For the combined women negative correlations were observed for XERcomp to n-3/n-6 fatty acid ratio (Table  4).

Association between xenoestrogenic transactivities and POPs
Few and scattered significant correlations were observed between the xenoestrogenic transactivities and the single PCBs and pesticides (data not shown).
For the XERcomp data of the combined men significant positive correlations were observed to non-adjusted POPs but the correlations disappeared upon adjustment for age and/or n-3/n-6 fatty acid ratio (Table 5). For the XERcomp data of the combined women negative correlations were observed to non-adjusted as well as age adjusted POPs (Table 5). No XERcomp-POP correlations were observed in the single districts. Table 3 shows that in overall the serum POP extracts agonized the XAR transactivities while antagonized XARcomp transactivities were observed except for Nuuk and Sisimiut men. The XAR transactivities for Sisimiut men were significantly higher than for Sisimiut women and the XARcomp transactivities for Nuuk and Sisimiut men were significantly higher than for Nuuk and Sisimiut women, respectively. Furthermore, the ANOVA analysis showed significant differences between districts for the XARcomp transactivities of both genders (Table 3).

Xenoandrogenic transactivities
Scatter plots of the xenoandrogenic transactivities and ΣPOP for the single districts are shown in Figure 1.

Correlation between xenoandrogenic transactivities and lifestyle characteristics
The XAR transactivities did not correlated to any lifestyle characteristics for the combined men or the combined women, whereas the XARcomp transactivities correlated positively to age for the combined men. Few and scattered correlations were found for the single districts (Table 4).

Association between xenoandrogenic transactivities and POPs
Few and scattered significant correlations were observed between the xenoandrogenic transactivities and the single PCBs and pesticides (data not shown). For the XAR transactivity data no correlations were observed to the sums of POPs neither for the combined men, the combined women nor in the single districts (data not shown).
The combined male data across the districts showed negative correlations between XARcomp and the POPs upon adjustment for age and age plus n-3/n-6 ratio. Excluding the Nuuk men, being atypically with respect to age and diet (see section 3.1) compared to Sisimiut and Qaanaaq men, a significant negative correlation between XARcomp and the POPs were observed both before and after adjustment for age. No XARcomp-POP correlations were observed for the combined women, neither in the single districts (Table 5).

The ratio of xenoandrogenic to xenoestrogenic serum transactivity
The XAR/XER ratio suggests a net xenoandrogenicity (ratio > 1) for both genders and also the XARcomp/XERcomp ratio for men. However, for the women the XARcomp/ XERcomp ratio being < 1 suggests a net xenoestrogenic activity (Table 3).

Discussion
In the present study we determined the actual xenohormone transactivity as a biomarker of serum POP exposure and evaluated whether the transactivity was associated to serum POP concentrations and the lifestyle characteristics across the Greenlandic districts Nuuk, Sisimiut and Qaanaaq. For the age and/or n-3/n-6 adjusted data we found an inverse correlation between xenoestrogenic and xenoandrogenic transactivities for combined female and combined male, respectively. XER and XAR: ER and AR transactivity of serum extract alone, respectively; XERcomp and XARcomp: competitive receptor transactivity upon co-treatment with EC 50 of E2 or R1881, respectively. % agonistic and % decreased indicates a significant increase or decrease, respectively, for the XER and XAR activity compared to the activity of the solvent control which was set to 1. % further increased and % antagonist indicates a further significant increase or decrease, respectively, for the XERcomp and XARcomp transactivity compared to the transactivity of the E2-EC 50 or R1881-EC 50 solvent control, which was set to 1. RLU/μg protein: The determined luciferase activity was corrected for cell density using protein measurements and expressed in relative luciferase units per microgram protein (RLU/μg protein). *: p value for men compared to women within the same district (e.g. Nuuk men and Nuuk women) using Student's t-test. **: p value for comparison between the districts for the separate genders (e.g. Nuuk vs Sisimiut vs Qaanaaq and vice versa) using ANOVA, result in the brackets (post hoc LSD/Dunnett T3 test).
Xenohormone serum transactivities in the Greenlandic districts   In general, serum POP extracts, free of endogenous hormones, showed a trend to decrease the ER transactivity (XER) and a low trend to agonize the AR transactivity (XAR) for both genders. In order to assure that the effect of the actual serum POP mixture on the receptor activation does reflect the biological activity in serum and the in vivo interaction between endogenous hormones and the POPs, we did evaluate the effect of the serum POP extracts in the presence of relevant concentrations of high affinity ligands (E2 and R1881) for the receptors to mimic the in vivo potential of the serum samples (XERcomp and XARcomp).
In the single districts XERcomp and XARcomp transactivities were predominantly antagonized except for serum extracts of the older Nuuk men where 50% elicited further increased XARcomp transactivities. No significant differences in the XER and the XAR were observed between men and women except the XAR activity of Sisimiut men being higher than Sisimiut women supporting that the endogenous hormones were effectively separated from the serum POP fractions.
Because of differences between the genders in lifestyle characteristics and POP levels and to improve the statistical power, we evaluated the combined data of each gender across the districts.
In the present study the serum samples of men from Nuuk contained a higher level of POPs due to their higher average age and higher intake of seafood and seabird [7]. However, in accordance with previous studies, Qaanaaq inhabitants, having a high intake of marine and traditional diet, had in general significantly higher levels of POPs compared to Nuuk women and Sisimiut inhabitants, eating a more westernized diet. For all three districts the serum POP levels of men were significantly higher than for women, reflecting the higher intake of traditional food by men [7]. Moreover, despite a high intercorrela-tion between POPs, the profile differed between districts and genders reflecting different exposure pattern.
In Arctic populations higher serum POP concentrations is determined by age and intake of marine food as reflected by an increased n-3/n-6 fatty acid ratio [8,42], and a higher level of POPs was reported for smokers [7,9]. We observed positive associations between the xenohormone transactivities and age, n-3/n-6 and smoker years for the combined male data whereas the xenoestrogenic activities were negatively associated to n-3/n-6 for the combined women. Thus, the determinants of the serum POP concentration seem also to be predictors of the xenohormone transactivities supporting the use of xenohormone transactivities as an integrated biomarker of POP exposure and lifestyle characteristics.
For combined men the positive correlation between the XERcomp transactivities and the POPs disappeared upon adjustment for age. POP concentrations increase by age, and age was found to be a determinant of the XERcomp transactivities for men. Thus, the observed association between XERcomp and the unadjusted POP data might be a reflection of the broad range of age for the included men and the correlation between XERcomp and age.
The XERcomp transactivities for the combined women correlated inversely to both non-adjusted and ageadjusted POPs. We found in a previous study [33], evaluating the xenoestrogenic serum transactivity of male study groups from Europe and Greenland, a negative correlation between the ER transactivity and the two POP proxy markers PCB153 and p,p'-DDE for Inuit.
In the same study population as used in this study, the serum dioxin-like transactivity mediated via the aryl hydrocarbon receptor (AhR) was determined [45]. Interestingly, the xenoestrogenic transactivities correlated neg-   Data on the xenoestrogenic activity in human samples are still scarce. Few studies using a different setup of measurement of xenoestrogenic transactivities in adipose tissue from Spanish women and serum samples from pregnant Danish and Faroese women have been reported, although, in contrast to our studies no correlation between POPs and MCF-7 proliferation were found [32,20,31,34]. However, recent data showed, for nonpregnant female greenhouse workers in Denmark that the E2-induced MCF-7 proliferation response was reduced, indicating an antiestrogenic effect of the serum extract containing major xenoestrogens but without pharmaceutical and endogenously produced estrogens [46]. We believe that the high impact of POP burden in the Inuit is responsible for the antagonized ER function in line with earlier in vitro studies [15,34].
In the present study only a small fraction of the serum POP extracts agonized the XAR transactivities. Numerous POPs have been shown to be antiandrogenic in vivo and in vitro [17,47,14,18,48,15,[49][50][51]. However, also androgenic effects have been found in the environment, e.g. craft mill effluent from a river in Florida contained a chemical mixture that induced AR-dependent gene expression and masculinized female fish in vitro [52]. . Thus, the serum extracts of Nuuk men seems to include some POPs further increasing the XARcomp transactivity but inhibiting the AhRcomp transactivity. Whether these POPs are related to the higher age of Nuuk men or their higher intake of seafood and seabird can at this point not be determined.
Taking the broad age range of the men across the districts into account the combined male XARcomp data correlated negatively to the POPs upon adjustment for age. Because of the higher age of Nuuk men we also analyzed the combined data of Sisimiut and Qaanaaq men separately, and significant negative correlations to the POPs were observed both before and after adjustment for age. In support to the present results, we previously observed negative correlations between XARcomp and the POP proxy marker p,p'-DDE for combined European study groups [36].
The data showed correlations between POPs and female xenoestrogenic and male xenoandrogenic transactivities. We cannot explain this gender pattern but have the following suggestions: The serum POP concentration was approximately 3 times higher in male serum compared to female serum mainly because of a different diet pattern, men eating more traditional and women more westernized [7], resulting in a different compositions of the serum POP mixture between the two sexes. +age  ---------99  -.07  .50  +n-3/n-6  ---------97  -.04  .75  +age+n-3/n-6  --------- Adjustment for BMI, smoker years and bird intake did not further change the data. The overall association between the POP markers and xenohormone transactivity across the study groups were evaluated using multiple linear regressions analysis.
Homogeneity across the districts was observed (see Additional file 1) and to improve the statistical power and the overall trend of the study groups the combined male and female data across the districts was evaluated. *: The combined data of Sisimiut and Qaanaaq men was significantly negatively correlated to non-adjusted as well as age adjusted POPs.
-: The data of the single districts were not adjusted for lifestyle characteristics because of the few cases in each district (n = 31 -40). We believe that the presented data reflects the effect of POPs on the sex hormone receptors as supported by the following: the observed correlations between xenohormone transactivities and the POPs were all observed in the in vivo mimic by the presence of a high affinity receptor ligand (E2 for ER and R1881 for AR). No correlations were observed for the serum POP extracts alone and furthermore, the observed correlations were negative. Thus, it does not seem plausible that the correlations between XERcomp/XARcomp and the age-adjusted POPs result from remnant endogenous hormones in the serum POP fraction. However, we cannot assess whether the correlations in the present study are only by chance findings. 25 out of 276 correlations (<10%) showed statistically significance, whereas the expected number of significant findings under the null hypothesis of no difference would be 14 out of 276 correlations (5%). We did observe fewer significant findings for the xenoandrogenic transactivities. This could be a result of the similar coefficients of variation within and between days for the AR-CALUX assay making the assay less robust than the ER assay.
Interestingly, we found higher xenoandrogenic/xenoestrogenic ratio for the male Inuit compared to the daily serum controls of Danish men (see section 2.3) as well as compared to European male in another study [36]. At this point it is not possible to assess whether this change in xenoandrogenicity can have any impact in vivo on the hormone homeostasis. It was proposed, that exposure to a mixture of chemicals with antiandrogenic or estrogenic properties may affect the androgen-estrogen ratio and thereby influence the risk of cryptorchidism in Danish and Finnish boys [56]. Whether a lower rate of cryptorchidism is found in Inuit awaits further research. We did previously observe negative correlations for xenoestrogenic and dioxin-like transactivity to sperm chromatin integrity and DNA damage for Inuit and positive correlations between xenoandrogenic and dioxin-like transactivity and sperm chromatin integrity and DNA damage for Europeans [57,58]. Further studies are needed to elucidate the possible association between POP related xenohormone transactivity and human health risk. Furthermore, other types of lipophilic endocrine disrupters in serum like the brominated flame retardants (BFRs) and perfluorooctanesulfonated (PFOS) affect the receptor function and should in future studies be taken into consideration.
Whether our ex vivo data mimic the cellular and molecular effects in vivo awaits additional studies. The presented data gives the final integrated net effect on the given receptor alone and other factors such as receptor and/or co-factor interactions might precede the final data. Our biomarker approach gives the possibility to estimate the combined impact of complex mixtures on health risks mediated via sex hormone receptors.

Conclusion and Perspectives
Determinants of serum POP concentration are also predictors of the xenohormone transactivities and can be used as an integrated biomarker of POP exposure and lifestyle characteristics. The actual serum POP mixture antagonize age adjusted sex hormone actions as supported by the inverse correlations to the xenoestrogenic and the xenoandrogenic transactivities for the combined female and the combined male Inuit, respectively.
The data of the presented manuscript is the first of its kind ever published for Inuit having very high levels of POPs. The emerging changes in diet in Greenland and thus the observed district differences in POP concentrations as well as mixture profile will be of great importance for the future elucidation of how mixtures of different POP levels and mixtures among different geographical population might affect the risk of health.