iPSC-derived neurons were generated from men with different sexual orientation
Unlike previous studies, we did not study how genotype is associated with sexual orientation but investigated actual neuronal gene expression phenotype related to male homosexuality, by using induced pluripotent stem cell (iPSC)-derived cortical neurons corresponding to maturity at late first/early second trimester of pregnancy. We generated and characterized iPSC-derived cortical neurons from six Finnish homosexual men, six heterosexual men, and five heterosexual women (Fig. 1a and Supplementary Table 1). We used dual SMAD inhibition to induce neuronal differentiation and after 2–3 weeks in sphere cultures, we dissociated neural progenitors and matured them for one week before harvesting. All the lines differentiated to neurons, which were on average 72.1% (SD 12.1%) positive for pan-neuronal marker Tubulin Beta 3 (TUBB3) and 75.5% (SD 14.3%) positive for Microtubule-associated protein 2 (MAP2) (Fig. 1b-c, Supplementary Fig. 5).
Homosexual men displayed downregulation of histone genes and upregulation of an X chromosome-residing gene FIRRE
We performed bulk RNA sequencing of iPSC-derived neurons and compared the transcriptomic profiles between the sexes and sexual orientation groups. As expected, the largest differences between men (both homosexual and heterosexual) and women were observed in the expression of sex chromosome genes (DDX3Y, ZFY, KDM5D, TXLNGY, UTY, XIST, and NLGN4Y) (based on cut-off Benjamini-Hochberg adjusted p-value p < 0.05 and absolute log2 fold change > 1) (Fig. 1d-f, Supplementary Tables 4–6). The most significant differentially expressed gene (DEG) found between men and women was DEAD-Box Helicase 3 Y-Linked (DDX3Y), which has been assumed to modulate neuronal differentiation (31). Overall, 33 DEGs overlapped between hetero- or homosexual men and heterosexual women (Fig. 1g). The sex chromosomal genes were dominating both comparisons. On the other hand, the comparison between homosexual and heterosexual men resulted 19 DEGs, of which 17 DEGs did not overlap with the men-women comparisons. Compared to the heterosexual men, the homosexual men displayed downregulation of ten histone genes (the most significant H4C13, H2AC17, H3C12, H2BC11, and H2AC4). Notably, we found that all the differentially expressed histone genes were located in the same locus 6p22.1-2. Moreover, most of these genes resided in the same cluster of 13 consecutive histone genes. Only three genes (FIRRE, ZP3, and PRSS33) were upregulated in homosexual men compared to heterosexual men. One of these genes: Firre Intergenic Repeating RNA Element (FIRRE), is an X chromosome-residing gene that was also significantly upregulated in heterosexual women compared to heterosexual men (adj. p-value 0.018). Given the significance of these findings, we discovered that even the expression of one of the histone genes (H4C13) and FIRRE was enough to separate homosexual and heterosexual men (Fig. 1h).
miR-199a-5p, upregulated in homosexual men, interacts with H3C11 and FBXO17
As sex-differences in microRNA (miRNA) expression were previously noted(32), we also performed miRNA sequencing from the same neuronal cultures. The differentially expressed miRNA for each comparison are presented in Fig. 2a-c and in Supplementary Tables 7–9. In the comparison between homosexual and heterosexual men, the most profound differences were up-regulation of miR-214-3p, miR-412-5p, miR-10a-5p, miR-199a-5p, miR-206, and miR-199a-3p (Fig. 2c). miR-199a-5p was statistically significantly down-regulated also in the comparison between heterosexual men versus heterosexual women (Fig. 2a; FDR 0.012), but not in the comparison between homosexual men versus heterosexual women (Fig. 2b), i.e., both of these latter groups differed significantly from heterosexual men. Given that miRNAs are known to target mRNAs and subsequently downregulate their expression, we compared the expression patterns of the differentially expressed miRNAs and mRNAs in our datasets for negative correlation using miRNA target prediction analysis from Ingenuity Pathway Analysis (IPA). Predicted interactions were found between three up-regulated miRNAs (miR-206, miR-199a-5p and miR214-3p), and down-regulated histone genes (H3C11 and H2BC17) and F-Box Protein 17 (FBXO17) between homosexual and heterosexual men (Fig. 2d, Supplementary Table 10).
Differentially expressed genes in homosexual men enrich in pathways related to DNA methylation and tRNA expression regulation
To further analyze differentially regulated signaling pathways between the groups, we performed Gene Ontology (GO) enrichment analysis and IPA canonical pathway analysis for differentially expressed genes (Fig. 2e and Supplementary Tables 11–16). In case of both GO and IPA analyses, the top differentially regulated pathways between male and female groups contained mostly sex chromosome-residing genes (KDM5D, UTY, EIF1AY, RPS4Y1, and TBL1Y). Interestingly, the IPA analysis between homosexual and heterosexual men revealed broad differences in pathways related to DNA methylation, transcriptional repression signaling, sonic hedgehog signaling, and IL-33 signaling pathways (Fig. 2e and Supplementary Table 13).
Proteomics results revealed downregulation of collagen proteins in homosexual men
Next, we sought to understand if the transcriptomic changes found in the neurons from homosexual men were translated to protein level. We performed global proteomics for untargeted detection of differentially expressed proteins (DEPs) between our study groups. As a result, we identified several hundreds of DEPs. Totally 53.8% of DEPs found between homosexual and heterosexual men overlapped with the DEPs found between homosexual men and heterosexual women(Fig. 3a, Supplementary Tables 17–19). Despite a large number of detected proteins, the DEGs and DEPs overlapped only by two genes/proteins (Fig. 3b). Similar to the transcriptomics results, the proteomics comparisons between either hetero- or homosexual men and heterosexual women revealed DEPs located in the Y-chromosome (Fig. 3c-d; cut-off Benjamini-Hochberg adjusted p-value p < 0.05 and absolute log2 fold change > 1). The most significant of these DEPs were DDX3Y and RPS4Y1, which were also the only DEPs that were differentially expressed in both omic data sets (Fig. 3b). Among the most significant DEPs between homosexual and heterosexual men were downregulated collagen genes (COL8A1, COL6A1, and COL18A1), which were also enriched in the top GO and IPA pathways (Fig. 3f, Supplementary Tables 20–25). Upregulated proteins found in homosexual men were related to neuronal firing (GNAO1), plasticity (CAMK2A), synapse modulation (SULT4A1), and dendrite development (ACTL6B), which could indicate improved neuronal maturation. The most significant DEP was NEDD4 Like E3 Ubiquitin Protein Ligase (NEDD4L) belonging to the ubiquitination machinery, which was upregulated in homosexual men compared to heterosexual men (Fig. 3e). Together with the most down-regulated protein (COL8A1), we were able to separate men based on their sexual orientation (Fig. 3g).