Associations between complex OHC mixtures and thyroid and cortisol hormone levels in East Greenland polar bears

Abstract

The multivariate relationship between hair cortisol, whole blood thyroid hormones, and the complex mixtures of organohalogen contaminant (OHC) levels measured in subcutaneous adipose of 23 East Greenland polar bears (eight males and 15 females, all sampled between the years 1999 and 2001) was analyzed using projection to latent structure (PLS) regression modeling. In the resulting PLS model, most important variables with a negative influence on cortisol levels were particularly BDE-99, but also CB-180, -201, BDE-153, and CB-170/190. The most important variables with a positive influence on cortisol were CB-66/95, α-HCH, TT3, as well as heptachlor epoxide, dieldrin, BDE-47, p,p'-DDD. Although statistical modeling does not necessarily fully explain biological cause-effect relationships, relationships indicate that (1) the hypothalamic-pituitary-adrenal (HPA) axis in East Greenland polar bears is likely to be affected by OHC-contaminants and (2) the association between OHCs and cortisol may be linked with the hypothalamus-pituitary-thyroid (HPT) axis.

Figure 1

a) Loading score plot showing loadings of the significant PLS model between Y=cortisol and 58 X-variables (49 individual organohalogens in adipose tissue [ng/g lipid weight], lipid content, girth, length, estimated BM, age, sex, capture day, and circulating TT3 and TT4 levels) of 23 polar bears (Ursus maritimus) including subadult and adult males and females. The PLS model had significant components: R²X=0.2, R²Y=0.5, Q²=0.3, and CV-ANOVA: p=0.03. The loading plot displays the correlation structure of the variables and shows how the X-variables relates to Y. Herein, variable loading of the single PLS component is plotted against their listed number in the data matrix. Thus, this loading plot must only be interpreted vertically. The variables with the highest loading (w*c) above zero have the highest positive influence on Y. The variables with the lowest loading below zero have the highest negative influence on Y. Loadings with value close to zero are unimportant. b) All bars show variable importance for the projection (VIP) values that summarizes the importance of the variables in explaining the X-matrix and to correlate with Y=cortisol. The error bars represent 95% confidence intervals. VIP ≥ 1 indicates important X-variables (dark grey bars) and the open bars indicate X-variables with 1 > VIP > 0.5. The X-variables that obtained VIP ≤ 0.5 are considered unimportant and not included in the figure. c) The regression coefficient plot of the PLS model. All bars show regression coefficient (CoeffCS) values of each variable indicating the direction and strength of the relationships between individual X-variables and Y=cortisol. The error bars represent the 95 % confidence intervals. The full bars present CoeffCS values of variables with VIP values ≥1 from Fig. 1b, which indicate high importance for the model. Please see text for further details.

Figure 1

a) Loading score plot showing loadings of the significant PLS model between Y=cortisol and 58 X-variables (49 individual organohalogens in adipose tissue [ng/g lipid weight], lipid content, girth, length, estimated BM, age, sex, capture day, and circulating TT3 and TT4 levels) of 23 polar bears (Ursus maritimus) including subadult and adult males and females. The PLS model had significant components: R²X=0.2, R²Y=0.5, Q²=0.3, and CV-ANOVA: p=0.03. The loading plot displays the correlation structure of the variables and shows how the X-variables relates to Y. Herein, variable loading of the single PLS component is plotted against their listed number in the data matrix. Thus, this loading plot must only be interpreted vertically. The variables with the highest loading (w*c) above zero have the highest positive influence on Y. The variables with the lowest loading below zero have the highest negative influence on Y. Loadings with value close to zero are unimportant. b) All bars show variable importance for the projection (VIP) values that summarizes the importance of the variables in explaining the X-matrix and to correlate with Y=cortisol. The error bars represent 95% confidence intervals. VIP ≥ 1 indicates important X-variables (dark grey bars) and the open bars indicate X-variables with 1 > VIP > 0.5. The X-variables that obtained VIP ≤ 0.5 are considered unimportant and not included in the figure. c) The regression coefficient plot of the PLS model. All bars show regression coefficient (CoeffCS) values of each variable indicating the direction and strength of the relationships between individual X-variables and Y=cortisol. The error bars represent the 95 % confidence intervals. The full bars present CoeffCS values of variables with VIP values ≥1 from Fig. 1b, which indicate high importance for the model. Please see text for further details.

Figure 1

a) Loading score plot showing loadings of the significant PLS model between Y=cortisol and 58 X-variables (49 individual organohalogens in adipose tissue [ng/g lipid weight], lipid content, girth, length, estimated BM, age, sex, capture day, and circulating TT3 and TT4 levels) of 23 polar bears (Ursus maritimus) including subadult and adult males and females. The PLS model had significant components: R²X=0.2, R²Y=0.5, Q²=0.3, and CV-ANOVA: p=0.03. The loading plot displays the correlation structure of the variables and shows how the X-variables relates to Y. Herein, variable loading of the single PLS component is plotted against their listed number in the data matrix. Thus, this loading plot must only be interpreted vertically. The variables with the highest loading (w*c) above zero have the highest positive influence on Y. The variables with the lowest loading below zero have the highest negative influence on Y. Loadings with value close to zero are unimportant. b) All bars show variable importance for the projection (VIP) values that summarizes the importance of the variables in explaining the X-matrix and to correlate with Y=cortisol. The error bars represent 95% confidence intervals. VIP ≥ 1 indicates important X-variables (dark grey bars) and the open bars indicate X-variables with 1 > VIP > 0.5. The X-variables that obtained VIP ≤ 0.5 are considered unimportant and not included in the figure. c) The regression coefficient plot of the PLS model. All bars show regression coefficient (CoeffCS) values of each variable indicating the direction and strength of the relationships between individual X-variables and Y=cortisol. The error bars represent the 95 % confidence intervals. The full bars present CoeffCS values of variables with VIP values ≥1 from Fig. 1b, which indicate high importance for the model. Please see text for further details.

Figure 2

Scatter plots of the Spearman’s rank order correlation tests conducted on some of the parameters deemed most important according to the PLS model; a) TT3, b) α-HCH, c) CB-180. Please see text for further details.