Associations between testosterone, estradiol, and androgen receptor genotype with amygdala subregions in adolescents

Abstract

Much is known about the development of the whole amygdala, but less is known about its structurally and functionally diverse subregions. One notable distinguishing feature is their wide range of androgen and estrogen receptor densities. Given the rise in pubertal hormones during adolescence, sex steroid levels as well as receptor sensitivity could influence age-related subregion volumes. Therefore, our goal was to evaluate the associations between the total amygdala and its subregion volumes in relation to sex hormones - estradiol and free testosterone (FT) - as a function of age and genetic differences in androgen receptor (AR) sensitivity in a sample of 297 adolescents (46% female). In males, we found small effects of FT-by-age interactions in the total amygdala, portions of the basolateral complex, and the cortical and medial nuclei (CMN), with the CMN effects being moderated by AR sensitivity. For females, small effects were seen with increased genetic AR sensitivity relating to smaller basolateral complexes. However, none of these small effects passed multiple comparisons. Future larger studies are necessary to replicate these small, yet possibly meaningful effects of FT-by-age associations and modulation by AR sensitivity on amygdala development to ultimately determine if they contribute to known sex differences in emotional neurodevelopment.

Keywords: Adolescence; Amygdala; Androgen receptor; Estradiol; Lateral amygdala; Testosterone.

Figure A.1:

3D and graphical representation of results in females. 1) Coronal 3D representation of the androgen receptor sensitivity results. 2) Graphical representation of subregions with small AR effects from panel (1); for visualization purposes, each individual is plotted (black dots) and lines represent predicted values of volume in relationship to AR repeat length (inversely related to sensitivity), with mean values of each covariate of non-interest (i.e., SES and ICV) held constant; dashed lines represent the 95% Confidence Intervals of the model prediction line; ω² represents the effect size of the highest order interaction term. Abbreviations: AR, androgen receptor; LA, lateral nucleus; BLDI, basolateral dorsal and intermediate subdivision; BLVPL, basolateral ventral and paralaminar subdivision; BM, basomedial nucleus. * FDR p-value ≤ 0.05 (significant)

Figure A.2:

3D and graphical representation of results in males. 1) Coronal 3D representation of androgen receptor sensitivity and free testosterone interaction results. 2) Graphical representation of small and medium effects from panel plotted in the low AR sensitivity group and the high (1); FT concentrations plotted were based on FT intra-quartile range values (Q1 = 0.12; Q3 = 0.35); for visualization purposes, each individual is plotted (black dots) and lines represent predicted values of volume at a given FT level, with mean values of each covariate of non-interest (i.e., SES and ICV) held constant; dashed lines represent the 95% Confidence Intervals of the model prediction line; ω² represents the effect size of the highest order interaction term. Abbreviations: FT, free testosterone; AR, androgen receptor; BLVPL, basolateral ventral and paralaminar subdivision; CMN, cortical and medial nuclei. * FDR p-value ≤ 0.05 (significant)