Oxytocin pathway gene networks in the human brain

Abstract

Oxytocin is a neuropeptide involved in animal and human reproductive and social behavior. Three oxytocin signaling genes have been frequently implicated in human social behavior: OXT (structural gene for oxytocin), OXTR (oxytocin receptor), and CD38 (oxytocin secretion). Here, we characterized the distribution of OXT, OXTR, and CD38 mRNA across the human brain by creating voxel-by-voxel volumetric expression maps, and identified putative gene pathway interactions by comparing gene expression patterns across 20,737 genes. Expression of the three selected oxytocin pathway genes was enriched in subcortical and olfactory regions and there was high co-expression with several dopaminergic and muscarinic acetylcholine genes, reflecting an anatomical basis for critical gene pathway interactions. fMRI meta-analysis revealed that the oxytocin pathway gene maps correspond with the processing of anticipatory, appetitive, and aversive cognitive states. The oxytocin signaling system may interact with dopaminergic and muscarinic acetylcholine signaling to modulate cognitive state processes involved in complex human behaviors.

Fig. 1

Oxytocin pathway gene expression in the human brain. Each point represents mean expression from six donors with standard errors for a given brain region for a OXTR and b CD38. The bolded dashed lines represent mean expression across the all regions with 1 standard deviation (+/−) also shown. Compared to average brain expression, there is increased expression of OXTR and CD38 in central and temporal brain structures, along with the olfactory region. Lower than average expression is observed in the cerebellum. *p < 0.05 (FDR corrected for 54 tests)

Fig. 2

Expression of OXTR, CD38, and OXT in hypothalamic and thalamic substructures. Each point represents mean expression of a OXTR, b CD38, and c OXT with standard errors for a given brain region. The bolded dashed lines represent mean expression across the all regions with 1 standard deviation (+/−) also shown. Z-scores for d OXTR, e CD38, and f OXT expression are also presented in anatomic maps for hypothalamic and thalamic substructures (see Supplementary Fig. 7 for map legend). Substructures with expression more than 1 standard deviation greater than the mean are labeled. DMH: dorsomedial hypothalamic nucleus, LMH: lateral hypothalamic area (mammillary region), LT: lateral tuberal nucleus, Pa: paraventricular nucleus of the hypothalamus, PR: Preoptic region, SON: supraoptic nucleus, TM: tuberomammillary nucleus, VMH: ventromedial hypothalamic nucleus, AN: anterior group of nuclei, DLG: dorsal lateral geniculate nucleus, DL: lateral group of nuclei (dorsal division), VL: lateral group of nuclei (ventral division), MG: medial geniculate complex, MN: medial group of nuclei, PG: posterior group of nuclei, RT: reticular nucleus of thalamus. * p < 0.05 (FDR corrected for 16 tests)

Fig. 3

Co-expression of selected oxytocin, dopaminergic, muscarinic acetylcholine, vasopressin, and opioid gene sets in the brain. a Co-expression patterning for the expression of selected oxytocin, dopaminergic, muscarinic acetylcholine, vasopressin, and opioid gene sets. The complete linkage method was used to identify 3 clustering groups (black squares). OXTR and CD38 are clustered together, along with DRD2, DRD5, COMT, CHRM4, and CHRM5. b Whole-brain co-expression of OXTR, CD38, and OXT with selected vasopressin, muscarinic acetylcholine, and dopaminergic gene sets. Spearman’s correlations are visualized on a density distribution demonstrating all co-expression correlations between the of OXTR, CD38, and OXT and all genes in the complete oxytocinergic (n = 94), cholinergic (n = 79), and dopaminergic (n = 63) gene sets (see Supplementary Data 1 for full lists). c Cortex-only co-expression of OXTR, CD38, and OXT with selected vasopressin, muscarinic acetylcholine, and dopaminergic gene sets. d Differential expression of the OXTR/CD38 gene sets in 30 tissue types from the GTEx dataset, with -log 10 p-values representing the probability of the hypergeometric test. ***p < 0.001, **p < 0.01, *p < 0.05

Fig. 4

Co-expression patterns of the full oxytocin signaling pathway. a Weighted gene co-expression network analysis was used to construct a hierarchical clustering tree (also see Supplementary Fig. 9) and network heatmap plot of gene–gene connectivity of full oxytocin signaling pathway in the brain. Darker red colors represent stronger topological overlap. This analysis identified a module (dark blue) containing both OXTR and CD38 (see Supplementary Fig. 9 and Supplementary Data 1 for a full list of genes and modules). b Differential expression of the OXTR/CD38 module in 30 tissue types from the GTEx dataset, with -log 10 p-values representing the probability of the hypergeometric test. *p < 0.05

Fig. 5

Differential stability of protein-coding genes. a Differential stability for all protein coding genes (n = 20,737) was calculated to assess the similarity of gene expression patterns from donor-to-donor. OXTR and CD38 were in the top decile of all genes (ranking in gray boxes), suggesting high reproducibility. Donor-to-donor differential stability was also assessed for b OXTR, c CD38, and d OXT. Each individual association for all three genes were statistically significant, suggesting that gene expression patterns were stable between each donor (ethnicity and sex are presented for each donor). *p < 0.001

Fig. 6

Voxel-by-voxel brain gene expression maps for OXTR, CD38, and OXT. mRNA expression maps were created using left hemisphere data from the Allen Human Brain Atlas dataset. These maps were submitted to the NeuroSynth framework to assess associations with cognitive state activation maps. SD: standard deviation

Fig. 7

Cognitive state correlates of central oxytocin pathway gene expression patterns. a Cognitive states were meta-analytically decoded from central oxytocin pathway mRNA maps (Fig. 6) using the NeuroSynth framework. The top five strongest relationships for OXTR, CD38, and OXT (Spearman’s r) are shown, with duplicates removed. b The distribution of Spearman’s correlations between each protein coding gene map (n = 20,737) and cognitive state maps. The absolute ranking for each oxytocin pathway gene out of 20,737 correlations are shown (also see Supplementary Table 2). mRNA expression of c OXTR, d CD38, and e OXT in brain regions associated with social and non-social mental states (see Supplementary Data 1 for lists of social and non-social terms)