Abstract
In mammals, sex differences in the adult brain are established very early in development, when the brain is still very immature. In the case of having inherited the SRY gene, during embryogenesis, testosterone secreted by the testes enters the brain and is converted to estradiol by the aromatase. Then the estradiol acts by binding to intracellular estrogen receptors (ERs) located predominantly in neurons, masculinizing specific brain regions. But ERs are also transcription factors that, when they are exposed to their ligand, dimerize and form complexes with coactivator proteins and corepressors, modifying the transcription of multiple target genes in a cascade effect and ultimately neuronal function. Given the intimate relationship between steroids and brain dimorphism, and steroid coactivators and gene transcription, in the present work, we further explore the implication of ERs α and β, and steroid coactivators NCoA-1, NCoA-2, NCoA-3, NCoA-4, NCoA-5 and p300-CREBBP, in the genesis of brain dimorphism. Based on our data, we believe that the coactivators NCOA-1, NCOA-2 and p300-CREBBP could be considered as candidate genes for GI.
Keywords
- estrogens
- gender incongruence
- steroid coactivators
1. Gender incongruence
The term
Most people present an alignment of gender identity with natal sex (cisgender individuals), but in some cases (transgender individuals) gender identity differs (in varying degrees) from the sex assigned at birth. Thus, Gender Dysphoria (GD) in the Diagnostic and Statistical Manual of Mental Disorders DSM-5 [3] or Gender Incongruence (GI) in the International Classification of Diseases ICD-11 [4] are characterized by a marked incongruence between one’s experienced gender and the sex assigned at birth.
2. The genetic and epigenetic basis of the gender incongruence
The origin of GI is complex and appears to be multifactorial. Current hypotheses point out that GI could be associated with a characteristic neurodevelopmental processes of the brain [5, 6], not concordant with gender, due to the influence of testosterone, converted into estradiol in the brain.
Traditionally, this process of brain masculinization
2.1 The genetic component
A genetic component should also be taken into account since different genes start to express before the formation of the testes [7, 8]. In fact, in mammals, sexual differentiation begins at the time of fertilization, through a different chromosomal complement in males and females, and will be driven by the
Most studies about the genetic basis of GI analyzed the implication of some DNA polymorphisms related to ERs, α and β, the AR, the aromatase CYP19A1 or the CYP17A1 [11, 12, 13, 14, 15, 16, 17, 18, 19, 20] as well as the interaction effects (epistasis) among them [15, 21]. A summary of the principal studies about the genetic component of GI is shown in Table 1. This gene selection is based on the hypothesis that a small variation in the DNA sequence of these genes would imply a high variability in the sensitivity of the hormonal receptors to their ligands.
Investigations | Genes | Populations |
---|---|---|
Henningsson et al. [11] | ERβ, AR, CYP19A1 | 29 transwomen |
Hare et al. [12] | ERβ, AR, CYP19A1 | 112 transwomen |
Ujikce et al. [13] | ERβ, AR, CYP19A1 | 168 transmen 74 transwomen |
Fernández et al. [14, 16] | ERβ, AR, CYP19A1 | 273 transmen 442 transwomen |
Bentz et al. [19] | CYP17A1 | 49 transmen 104 transwomen |
Fernández et al. [20] (2016) | CYP17A1 | 223 transmen 317 transwomen |
Cortés Cortés et al. [17] | ERα: (TA)n-rs3138774, PvuII-rs2234693, XbaI-rs9340799 | 183 transmen 184 transwomen |
Fernández et al. [21] | ERα, ERβ, AR, CYP19A | 425 transmen 549 transwomen |
Foreman et al. [15] | AR, ERα, ERβ, SRD5A2, STS, SULT2A1, PGR, COMT, CYP17, SRD5A2 | 380 transwomen |
Fernández et al. [18] | ERα: rs9478245, rs3138774 rs2234693, rs9340799 | 226 transmen 273 transwomen |
Aranda et al. [24] | Epigenetics: ERα, ERβ, AR | 12 transmen 6 transwomen |
Fernández et al. [23] | Epigenetics: ERα promoter | 10 transmen 10 transwomen |
Table 1.
Mean investigations about the genetic basis of gender incongruence.
Henningsson et al. [11] were the first group to analyze three repeat polymorphisms, located in the estrogen receptor beta, the androgen receptor, and the aromatase genes in a trans female population. They found a relationship between the number of repetitions and gender incongruence. They found longer estrogen receptor and androgen receptor polymorphisms in the trans female population. Later, Hare et al. [12] replicated Henningsson’s study in a bigger population, finding longer androgen receptor polymorphisms. However, when Ujike et al., [13] analyzed the same polymorphisms in a Japanese population, they did not find any statistical difference. These and others polymorphisms were analyzed in a Spanish population by our group. Our results confirmed the involvement of both estrogen receptors (alpha and beta) in gender incongruence. Part of this data was confirmed by Foreman et al. [15].
2.2 The epigenetic component
An epigenetic component may also be involved since there is evidence that some environmental factors play a role in the sexual differentiation of the brain. For example, in mice, the sex difference in maternal anogenital licking of male compared with female pups produces a different methylation of the estrogen receptor α promoter in the preoptic area [22]. And in humans, certain environmental factors, such as a short crossover hormonal treatment (only 6 months), can modify the methylation profile of the estrogen receptor α promoter in a trans population [23, 24].
3. Estrogens, androgens and their receptors and coactivators
3.1 Estrogens
Estradiol (E2) exerts a wide variety of effects on growth, development, the function of reproductive systems and regulation in the central nervous system [25, 26]. The mechanism of action of the two ER isoforms α and β consists of binding with the E2 ligand to obtain the receptor’s dimerization (αα, αβ or ββ), originating the necessary conformational changes in the ligand binding domain (LBD) [27] and binding with high affinity to specific DNA sequences called estrogen response elements (EREs) [27] in the genes that are regulated by E2 (Figure 1). This conformational change in the LBD allows coactivators and other co-regulating proteins to be recruited. We must point out that this step is critical for the transcriptional regulation of genes induced by E2 [28].

Figure 1.
Molecular mechanisms of action of ERs α and β. hormone 17β-estradiol (E2) binds to the nuclear receptor (ERα or ERβ), and after dimerization and translocation to the nucleus, the nuclear receptor complex binds to a specific sequence of DNA known as an estrogen response element (ERE). The nuclear receptor complex in turn recruits the coactivators NCOAs, p300 and the CREBBP that activate the transcription of target genes.
Furthermore, estrogens are produced in many regions of the brain including the cortex, the hippocampus, the cerebellum, the hypothalamus and the amygdala, among others [29]. The actions of estrogens in the developing brain are generally permanent and range from the establishment of sexual differences to generalized trophic and neuroprotective effects [30]. In addition, estrogens are an important regulator of brain growth and differentiation, and ERs α and β are found in both the developing [31] and the adult human brain [32].
3.2 Estrogen and androgen receptors are transcription factors
We must point out that the packaging of DNA into chromatin causes a general repression of gene activity, and transcription factors function to relieve this chromatin-mediated repression [33].Thereby, once attached to their ligands, the receptors dimerize, enter the nucleus, and interact with the promoter regions of the target genes, modulating the expression of multiple genes in collaboration with some steroid coactivators (Figure 1). In the case of the AR, its ligand is androgen [34] while for the ERs, it is estrogen, 17β-Estradiol (E2) in particular [26].
3.3 Steroid receptor coactivators
Proteins called SRCs (Steroid Receptor Coactivators) serve as primary coactivators that interact with the complex formed by E2 and the hormonal receptor. Additionally, SRCs recruit multiple secondary coactivators such as p300 and the CREB-binding protein (also known as CREBBP) [35]. Both SRCs and p300 are the first coactivators that are coupled to the E2-ER complex [36] to activate the transcriptional process of the genes that are E2 targets (Figure 1).
Coactivators are proteins that influence the ability of the transcription factors to activate or inhibit expression of multiple genes in a cascade mode [37]. Given the intimate relationship between steroids and brain dimorphism, and coactivators and gene transcription, and since ERs α-β and the AR are hormonal receptors that act as transcription factors, it was clear that we should hypothesize the implication of DNA coactivators in the process of GI.
4. The role of E2-coactivators in the genetic basis of gender incongruence
To our knowledge, no studies have been published about the role of steroid receptor coactivators in the genetic basis of GI. Nevertheless, given the importance of estrogens in GI, and the critical role of coactivators in the transcriptional gene regulation induced by E2, our team deemed it interesting to analyze 247 single nucleotide polymorphisms (SNPs) located at the coactivators NCoA-1(or SRC-1), NCoA-2 (or SRC-2), NCoA-3 (or SRC-3), NCoA-4, NCoA-5 and p300-CREBBP, in a transgender
4.1 The characteristics of the study
Genomic DNA was extracted from 94 Spanish transgender individuals (47 transmen and 47 transwomen)
We analyzed 247 polymorphisms distributed in the coactivators NCOA-1 (63 SNPs), NCOA-2 (64 SNPs), NCOA-3 (30 SNPs), NCOA-4 (4 SNPs), NCOA-5 (8 SNPs), p300 (9 SNPs) and CREBBP (69 SNPs) (Table 2), in a population of 94 Spanish transgender individuals
Gene | Chromosome | Function | Analyzed SNPs | SNPs with significant differences |
---|---|---|---|---|
NCOA-1 | 2 | The protein encoded by this gene acts as a transcriptional coactivator for steroid and nuclear hormone receptors. | 63 | 3 |
NCOA-2 | 8 | The encoded protein acts as an intermediary factor for the ligand-dependent activity of nuclear receptors, which regulate their target genes upon binding of cognate response elements. | 64 | 5 |
NCOA-3 | 20 | The protein encoded by this gene is a nuclear receptor coactivator that interacts with nuclear hormone receptors to enhance their transcriptional activator functions. | 30 | 0 |
NCOA-4 | 10 | This gene encodes an androgen receptor coactivator. | 4 | 0 |
NCOA-5 | 20 | This gene encodes a coregulator for the α and β estrogen receptors. | 8 | 0 |
p300 | 22 | This gene encodes a transcriptional coactivator protein. | 9 | 2 |
CREBBP | 16 | This gene is ubiquitously expressed and is involved in the transcriptional coactivation factors. | 69 | 1 |
Table 2.
Description of the analyzed polymorphisms.
4.2 The findings
4.2.1 Similar prevalence rates for the analyzed polymorphisms and comparison with the global and European 1000 genomes
As expected, the prevalence rates for all analyzed polymorphisms in our population were similar to those found in the Global 1000 genomes and the European 1000 genomes http://www.1000genomes.org (Table 3).
Gene | Polymorphism | Alias | DNA variation | Our study frequency | Global 1000 genomes frequency | European 1000 genomes frequency |
---|---|---|---|---|---|---|
NCOA-1 | rs10495747 | P1 | T/C | C = 0.11 | C = 0.1330 | C = 0.1153 |
rs2584940 | P2 | T/G | G = 0.38 | G = 0.4605 | G = 0.4125 | |
rs6756785 | P3 | A/G | G = 0.32 | G = 0.2115 | G = 0.2883 | |
NCOA-2 | rs76968380 | P4 | G/A | A = 0.06 | A = 0.1138 | A = 0.0646 |
rs34406737 | P5 | G/A | A = 0.15 | A = 0.1300 | A = 0.1262 | |
rs1963250 | P6 | G/T | T = 0.57 | T = 0.5691 | T = 0.5368 | |
rs10755950 | P7 | G/A | A = 0.42 | A = 0.5655 | A = 0.4483 | |
rs56055423 | P8 | A/G | G = 0.05 | G = 0.0132 | G = 0.0457 | |
p300 | rs133084 | P9 | T/C | C = 0.5956 | C = 0.5447 | C = 0.59 |
rs11806 | P10 | A/G | G = 0.3894 | G = 0.3499 | G = 0.42 | |
CREBBP | rs2191416 | P11 | G/A | A = 0.2660 | A = 0.2555 | A = 0.26 |
Table 3.
Description of the polymorphisms with significant differences.
4.2.2 Eleven polymorphism showed differences in the distribution of the allele and genotype frequencies
When we compared the distribution of the allele and genotype frequencies, we found significant differences in 11 polymorphisms, that correspond to 4.45% of the total analyzed: three polymorphisms located in NCOA-1, five in NCOA-2, two in p300 and one in CREBBP (Tables 2 and 3). The description of the significant association analyses with GI in different models of inheritance is in Table 4.
Model | Genotype | Cis (%) | Trans (%) | OR | |
---|---|---|---|---|---|
P1 polymorphism (rs10495747) | |||||
Codominant | T/T | 79 (84%) | 67 (71.3%) | 1.00 | 0.068 |
T/C | 15 (16%) | 26 (27.7%) | 2.04 (1.00–4.17) | ||
C/C | 0 (0%) | 1 (1.1%) | NA (0.00-NA) | — | |
Dominant | T/T | 79 (84%) | 67 (71.3%) | 1.00 | |
T/C-C/C | 15 (16%) | 27 (28.7%) | 2.12 (1.04–4.32) | ||
Recessive | T/T–T/C | 94 (100%) | 93 (98.9%) | 1.00 | 0.24 |
C/C | 0 (0%) | 1 (1.1%) | NA (0.00-NA) | — | |
Overdominant | T/T-C/C | 79 (84%) | 68 (72.3%) | 1.00 | 0.051 |
T/C | 15 (16%) | 26 (27.7%) | 2.01 (0.99–4.11) | 0,054 | |
Log-additive | — | — | — | 2.15 (1.07–4.29) | |
P2 polymorphism (rs2584940) | |||||
Codominant | T/T | 43 (45.7%) | 27 (28.7%) | 1.00 | |
T/G | 42 (44.7%) | 52 (55.3%) | 1.97 (1.05–3.70) | ||
G/G | 9 (9.6%) | 15 (16%) | 2.65 (1.02–6.91) | ||
Dominant | T/T | 43 (45.7%) | 27 (28.7%) | 1.00 | |
T/G-G/G | 51 (54.3%) | 67 (71.3%) | 2.09 (1.14–3.83) | ||
Recessive | T/T–T/G | 85 (90.4%) | 79 (84%) | 1.00 | 0.19 |
G/G | 9 (9.6%) | 15 (16%) | 1.79 (0.74–4.33) | 0.198 | |
Overdominant | T/T-G/G | 52 (55.3%) | 42 (44.7%) | 1.00 | 0.14 |
T/G | 42 (44.7%) | 52 (55.3%) | 1.53 (0.86–2.72) | 0.148 | |
Log-additive | — | — | — | 1.72 (1.10–2.69) | |
P3 polymorphism (rs6756785) | |||||
Codominant | A/A | 50 (53.2%) | 32 (34%) | 1.00 | |
A/G | 38 (40.4%) | 54 (57.5%) | 2.22 (1.21–4.08) | ||
G/G | 6 (6.4%) | 8 (8.5%) | 2.08 (0.66–6.56) | 0.213 | |
Dominant | A/A | 50 (53.2%) | 32 (34%) | 1.00 | |
A/G-G/G | 44 (46.8%) | 62 (66%) | 2.20 (1.22–3.97) | ||
Recessive | A/A-A/G | 88 (93.6%) | 86 (91.5%) | 1.00 | 0.58 |
G/G | 6 (6.4%) | 8 (8.5%) | 1.36 (0.45–4.10) | 0.598 | |
Overdominant | A/A-G/G | 56 (59.6%) | 40 (42.5%) | 1.00 | |
A/G | 38 (40.4%) | 54 (57.5%) | 1.99 (1.11–3.55) | ||
Log-additive | — | — | — | 1.77 (1.10–2.87) | |
P4 polymorphism (rs76968380) | |||||
Codominant | G/G | 88 (93.6%) | 80 (85.1%) | 1.00 | 0.075 |
G/A | 6 (6.4%) | 12 (12.8%) | 2.20 (0.79–6.14) | 0.132 | |
A/A | 0 (0%) | 2 (2.1%) | NA (0.00-NA) | — | |
Dominant | G/G | 88 (93.6%) | 80 (85.1%) | 1.00 | 0.055 |
G/A-A/A | 6 (6.4%) | 14 (14.9%) | 2.57 (0.94–7.00) | 0.065 | |
Recessive | G/G-G/A | 94 (100%) | 92 (97.9%) | 1.00 | 0.095 |
A/A | 0 (0%) | 2 (2.1%) | NA (0.00-NA) | — | |
Overdominant | G/G-A/A | 88 (93.6%) | 82 (87.2%) | 1.00 | 0.13 |
G/A | 6 (6.4%) | 12 (12.8%) | 2.15 (0.77–5.98) | 0.144 | |
Log-additive | — | — | — | 2.57 (1.01–6.55) | |
P5 polymorphism (rs34406737) | |||||
Codominant | G/G | 63 (67%) | 72 (76.6%) | 1.00 | |
G/A | 31 (33%) | 17 (18.1%) | 0.48 (0.24–0.95) | ||
A/A | 0 (0%) | 5 (5.3%) | NA (0.00-NA) | — | |
Dominant | G/G | 63 (67%) | 72 (76.6%) | 1.00 | 0.14 |
G/A-A/A | 31 (33%) | 22 (23.4%) | 0.62 (0.33–1.18) | 0.142 | |
Recessive | G/G-G/A | 94 (100%) | 89 (94.7%) | 1.00 | |
A/A | 0 (0%) | 5 (5.3%) | NA (0.00-NA) | — | |
Overdominant | G/G-A/A | 63 (67%) | 77 (81.9%) | 1.00 | |
G/A | 31 (33%) | 17 (18.1%) | 0.45 (0.23–0.88) | ||
Log-additive | — | — | — | 0.85 (0.49–1.49) | 0.57 |
P6 polymorphism (rs1963250) | |||||
Codominant | T/T | 20 (21.3%) | 37 (39.4%) | 1.00 | |
T/G | 54 (57.5%) | 46 (48.9%) | 0.46 (0.24–0.90) | ||
G/G | 20 (21.3%) | 11 (11.7%) | 0.30 (0.12–0.74) | ||
Dominant | T/T | 20 (21.3%) | 37 (39.4%) | 1.00 | |
T/G-G/G | 74 (78.7%) | 57 (60.6%) | 0.42 (0.22–0.79) | ||
Recessive | T/T–T/G | 74 (78.7%) | 83 (88.3%) | 1.00 | 0.075 |
G/G | 20 (21.3%) | 11 (11.7%) | 0.49 (0.22–1.09) | 0.080 | |
Overdominant | T/T-G/G | 40 (42.5%) | 48 (51.1%) | 1.00 | 0.24 |
T/G | 54 (57.5%) | 46 (48.9%) | 0.71 (0.40–1.26) | 0.245 | |
Log-additive | — | — | — | 0.53 (0.34–0.83) | |
P7 polymorphism (rs10755950) | |||||
Codominant | G/G | 33 (35.1%) | 31 (33%) | 1.00 | 0.064 |
A/G | 50 (53.2%) | 40 (42.5%) | 0.85 (0.45–1.62) | 0.632 | |
A/A | 11 (11.7%) | 23 (24.5%) | 2.23 (0.93–5.31) | 0.071 | |
Dominant | G/G | 33 (35.1%) | 31 (33%) | 1.00 | 0.76 |
A/G-A/A | 61 (64.9%) | 63 (67%) | 1.10 (0.60–2.01) | 0.770 | |
Recessive | G/G-A/G | 83 (88.3%) | 71 (75.5%) | 1.00 | |
A/A | 11 (11.7%) | 23 (24.5%) | 2.44 (1.11–5.36) | ||
Overdominant | G/G-A/A | 44 (46.8%) | 54 (57.5%) | 1.00 | 0.14 |
A/G | 50 (53.2%) | 40 (42.5%) | 0.65 (0.37–1.16) | 0.14 | |
Log-additive | — | — | — | 1.35 (0.90–2.04) | 0.15 |
P8 polymorphism (rs56055423) | |||||
Codominant | A/A | 91 (96.8%) | 81 (86.2%) | 1.00 | |
A/G | 3 (3.2%) | 12 (12.8%) | 4.49 (1.22–16.49) | ||
G/G | 0 (0%) | 1 (1.1%) | NA (0.00-NA) | — | |
Dominant | A/A | 91 (96.8%) | 81 (86.2%) | 1.00 | |
A/G-G/G | 3 (3.2%) | 13 (13.8%) | 4.87 (1.34–17.69) | ||
Recessive | A/A-A/G | 94 (100%) | 93 (98.9%) | 1.00 | 0.24 |
G/G | 0 (0%) | 1 (1.1%) | NA (0.00-NA) | — | |
Overdominant | A/A-G/G | 91 (96.8%) | 82 (87.2%) | 1.00 | |
A/G | 3 (3.2%) | 12 (12.8%) | 4.44 (1.21–16.29) | ||
Log-additive | — | — | — | 4.69 (1.32–16.63) | |
P9 polymorphism (rs133084) | |||||
Codominant | C/C | 27 (28.7%) | 42 (44.7%) | 1.00 | 0.056 |
T/C | 46 (48.9%) | 39 (41.5%) | 0.55 (0.29–1.04) | 0.066 | |
T/T | 21 (22.3%) | 13 (13.8%) | 0.40 (0.17–0.93) | ||
Dominant | C/C | 27 (28.7%) | 42 (44.7%) | 1.00 | |
T/C-T/T | 67 (71.3%) | 52 (55.3%) | 0.50 (0.27–0.91) | ||
Recessive | C/C-T/C | 73 (77.7%) | 81 (86.2%) | 1.00 | 0.13 |
T/T | 21 (22.3%) | 13 (13.8%) | 0.56 (0.26–1.19) | 0.135 | |
Overdominant | C/C-T/T | 48 (51.1%) | 55 (58.5%) | 1.00 | 0.3 |
T/C | 46 (48.9%) | 39 (41.5%) | 0.74 (0.42–1.32) | 0.307 | |
Log-additive | — | — | — | 0.61 (0.41–0.93) | |
P10 polymorphism (rs11806) | |||||
Codominant | A/A | 37 (39.8%) | 25 (26.6%) | 1.00 | 0.076 |
A/G | 45 (48.4%) | 49 (52.1%) | 1.61 (0.84–3.08) | 0.151 | |
G/G | 11 (11.8%) | 20 (21.3%) | 2.69 (1.10–6.58) | ||
Dominant | A/A | 37 (39.8%) | 25 (26.6%) | 1.00 | 0.055 |
A/G-G/G | 56 (60.2%) | 69 (73.4%) | 1.82 (0.98–3.38) | 0.058 | |
Recessive | A/A-A/G | 82 (88.2%) | 74 (78.7%) | 1.00 | 0.08 |
G/G | 11 (11.8%) | 20 (21.3%) | 2.01 (0.91–4.48) | 0.086 | |
Overdominant | A/A-G/G | 48 (51.6%) | 45 (47.9%) | 1.00 | 0.61 |
A/G | 45 (48.4%) | 49 (52.1%) | 1.16 (0.65–2.06) | 0.627 | |
Log-additive | — | — | — | 1.63 (1.06–2.52) | |
P11 polymorphism (rs2191416) | |||||
Codominant | G/G | 53 (57%) | 47 (50%) | 1.00 | 0.075 |
A/G | 38 (40.9%) | 38 (40.4%) | 1.13 (0.62–2.05) | 0.702 | |
A/A | 2 (2.1%) | 9 (9.6%) | 5.07 (1.04–24.67) | ||
Dominant | G/G | 53 (57%) | 47 (50%) | 1.00 | 0.34 |
A/G-A/A | 40 (43%) | 47 (50%) | 1.32 (0.74–2.36) | 0.354 | |
Recessive | G/G-A/G | 91 (97.8%) | 85 (90.4%) | 1.00 | |
A/A | 2 (2.1%) | 9 (9.6%) | 4.82 (1.01–22.93) | ||
Overdominant | G/G-A/A | 55 (59.1%) | 56 (59.6%) | 1.00 | 0.95 |
A/G | 38 (40.9%) | 38 (40.4%) | 0.98 (0.55–1.76) | 0.951 | |
Log-additive | — | — | — | 1.49 (0.92–2.41) | 0.1 |
Table 4.
Polymorphism association analysis with gender incongruence, in different models of inheritance (Codominant, Dominant, Recessive, Overdominant and Log-additive) (n = 188, crude analysis).
P1 polymorphism: The genotype T/T was overrepresented in the cis population (
P2 polymorphism: The genotype T/T was overrepresented in the cis population (
P3 polymorphism: The genotype A/A was overrepresented in the cis population (
P4 polymorphism: The P4 polymorphism was only significant for the log-additive model. The genotype G/G was overrepresented in the cis population, while the genotypes G/A and A/A were overrepresented in the trans population (OR = 2.57;
P5 polymorphism: This polymorphism was significant for the codominant, the recessive, and the overdominant models. The G/G and the A/A genotypes were overrepresented in the trans population (
P6 polymorphism: The T/T genotype was overrepresented in the trans population while the T/G and G/G were overrepresented in the cis population (OR = 0.42;
P7 polymorphism: This polymorphism was only significant for the recessive model. The A/A genotype was overrepresented in the trans population (OR = 2.44;
P8 polymorphism: The genotype A/A was overrepresented in the cis population (
P9 polymorphism: The C/C genotype was overrepresented in the trans population, while the T/C and T/T were overrepresented in the cis population (OR = 0.50;
P10 polymorphism: This polymorphism was only significant for the log-additive model. The G/G genotype was overrepresented in the trans population (OR = 2.69;
P11 polymorphism: The A/A polymorphism was overrepresented in the trans population (OR = 4.82;
4.2.3 The three polymorphisms (P2, P9 and P10) showed significant differences in the interaction analysis with the covariate “sex”
Furthermore, polymorphisms P2, P9 and P10 showed significant differences in the interaction analysis with covariate “sex”. For the P2 polymorphism, the genotype T/G was more frequent in the trans population assigned as females at birth than in the cis female population (OR = 2.76;
While for the P9 polymorphism, the genotype T/T was more frequent in the cis female population than in the trans population assigned as females at birth (OR = 0.34;
4.2.4 The haplotype analysis of the coactivators NCOA-1, NCOA-2 and p300, and comparison between cis and trans population
The simultaneous analysis of multiple loci (haplotypes) was carried out in those coactivators with two or more polymorphisms with statistical significance (NCOA-1, NCOA-2 and p300) using logistic regression models.
4.2.4.1 Polymorphisms in NCOA-1
For the three polymorphisms located in NCOA-1 (Table 5), the T allele for P1 was linked to the T allele for P2, and to the A allele for P3 (haplotype 1: T–T-A) with a total frequency of 0.45. This haplotype was more frequent in the cis than in the trans population. The haplotype 5: C-G-A was overrepresented in the trans population and showed statistical significance (OR = 2.62;
Haplotype frequencies estimation and haplotype association with GI (n = 188, adjusted by sex) | |||||||||
---|---|---|---|---|---|---|---|---|---|
Haplotypes | P1 | P2 | P3 | Total | Cis population | Trans population | Cumulative frequency | OR (95% CI) | |
1 | T | T | A | 0.4501 | 0.5201 | 0.377 | 0.4501 | 1.00 | — |
2 | T | T | G | 0.1495 | 0.134 | 0.1699 | 0.5996 | 2.25 (0.99–5.13) | 0.054 |
3 | T | G | G | 0.147 | 0.1319 | 0.1601 | 0.7466 | 1.73 (0.81–3.71) | 0.16 |
4 | T | G | A | 0.139 | 0.1341 | 0.144 | 0.8856 | 1.80 (0.81–3.97) | 0.15 |
5 | C | G | A | 0.069 | 0.0531 | 0.0897 | 0.9546 | 2.62 (1.00–6.83) | |
6 | C | T | A | 0.0228 | 0.0267 | 0.0169 | 0.9774 | 1.39 (0.23–8.34) | 0.72 |
7 | C | G | G | 0.0226 | 0 | 0.0423 | 1 | 379142884.10 (379142883.16–379142885.04) |
Table 5.
Haplotype analysis for polymorphisms located in NCOA-1 (P1, P2 and P3 polymorphisms).
Global haplotype association
4.2.4.2 Polymorphisms in NCOA-2
For the five polymorphisms located in NCOA-2 (Table 6), the significant haplotypes were the haplotype 2: (G-G-T-A-A) (OR = 2.49;
Haplotype frequencies estimation and haplotype association with GI (n = 188, adjusted by sex) | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Haplotypes | P4 | P5 | P6 | P7 | P8 | Total | Cis population | Trans population | Cumulative frequency | OR (95% CI) | P-value |
1 | G | G | G | G | A | 0.2546 | 0.3239 | 0.1963 | 0.2546 | 1.00 | — |
2 | G | G | T | A | A | 0.2206 | 0.2142 | 0.236 | 0.4752 | 2.49 (1.16–5.34) | |
3 | G | G | T | G | A | 0.2022 | 0.1721 | 0.2303 | 0.6773 | 2.00 (0.93–4.31) | 0.079 |
4 | G | G | G | A | A | 0.0891 | 0.0861 | 0.0793 | 0.7664 | 1.05 (0.38–2.88) | 0.92 |
5 | G | A | G | G | A | 0.0474 | 0.0656 | 0.0235 | 0.8138 | 0.55 (0.11–2.89) | 0.48 |
6 | G | A | T | A | A | 0.0419 | 0.0474 | 0.0423 | 0.8557 | 1.11 (0.24–5.24) | 0.89 |
7 | G | A | T | G | A | 0.041 | 0.0393 | 0.0426 | 0.8967 | 3.00 (0.10–86.09) | 0.52 |
8 | G | G | T | A | G | 0.0217 | 0 | 0.036 | 0.9184 | 12.86 (1.34–123.38) | |
9 | A | G | T | G | A | 0.0173 | 0.0085 | 0.0228 | 0.9357 | 5.62 (0.56–56.71) | 0.15 |
10 | A | G | G | G | A | 0.0165 | 0.0077 | 0.0205 | 0.9522 | 3.53 (0.40–31.50) | 0.26 |
11 | G | A | T | A | G | 0.0138 | 0.0028 | 0.0189 | 0.9659 | 2.68 (0.27–26.20) | 0.4 |
12 | A | G | T | A | A | 0.0108 | 0.0158 | 0.0093 | 0.9767 | 0.00 (-Inf - Inf) | 1 |
rare | * | * | * | * | * | 1 | 26887674246665641394185292538194458249584765033713636843978752.00 (26887674246622846799367197037120806603076791371544336569204736.00–26887674246708435989003388039268109896092738695882937118752768.00) |
Table 6.
Haplotype analysis for polymorphisms located in NCOA-2 (P4, P5, P6, P7 and P8 polymorphisms).
Global haplotype association
4.2.4.3 Polymorphisms in p300
For the two polymorphisms located in p300 (Table 7), the significant haplotype was haplotype 2 (T-A) (OR = 0.57;
Haplotype frequencies estimation and Haplotype association with GI (n = 188, adjusted by sex) | ||||||||
---|---|---|---|---|---|---|---|---|
Haplotypes | P9 | P10 | Total | Cis population | Trans population | Cumulative frequency | OR (95% CI) | |
1 | C | G | 0.4103 | 0.3468 | 0.4734 | 0.4103 | 1.00 | — |
2 | T | A | 0.4004 | 0.4549 | 0.3457 | 0.8107 | 0.57 (0.36–0.90) | |
3 | C | A | 0.1828 | 0.1851 | 0.1809 | 0.9935 | 0.69 (0.38–1.27) | 0.24 |
4 | T | G | 0.0065 | 0.0132 | 0 | 1 | 0.00 (-Inf - Inf) | 1 |
Table 7.
Haplotype analysis for polymorphisms located in p300 (P9 and P10 polymorphisms).
Global haplotype association
Haplotype and sex cross-classification interaction table (n = 188, crude analysis) | |||||
---|---|---|---|---|---|
Females | Males | ||||
0.4102 | 1.00 | — | 0.63 (0.19–2.07) | 0.457 | |
0.4003 | 0.43 (0.22–0.83) | 0.49 (0.19–1.27) | 0.141 | ||
0.1829 | 0.60 (0.26–1.40) | 0.237 | 0.52 (0.18–1.52) | 0.232 | |
0.0066 | 0.00 (0.00 - Inf) | — | 0.00 (-Inf - Inf) | — | |
0.4102 | 1.00 | — | 1.00 | — | |
0.4003 | 0.43 (0.22–0.83) | 0.77 (0.40–1.49) | 0.444 | ||
0.1829 | 0.60 (0.26–1.40) | 0.237 | 0.82 (0.35–1.96) | 0.665 | |
0.0066 | 0.00 (0.00 - Inf) | — | 0.00 (0.00 - Inf) | — | |
0.4102 | 1.00 | 0.63 (0.19–2.07) | 0.457 | ||
0.4003 | 1.00 | 1.12 (0.55–2.30) | 0.769 | ||
0.1829 | 1.00 | 0.87 (0.32–2.34) | 0.796 | ||
0.0066 | 1.00 | Inf |
Table 8.
Haplotype interaction analysis with covariate sex for polymorphisms located in p300 (P9 and P10 polymorphisms). Haplotype frequency, Odds ratio (OR) and P-value in female and male populations.
4.2.4.4 Summary of findings
In summary, when we analyzed the allele and genotype frequencies, we found significant differences in 11 polymorphisms located in NCOA-1, NCOA-2, p300 and CREBBP. Being the NCOA-2 and p300 the coactivators with the highest percentages of polymorphisms with significant differences (5/64 and 2/9 respectively). Furthermore, only P2 (located at NCOA-1), P9 (located at p300) and P10 (located at p300) showed a different distribution of the genotypes in males and females, that is, they showed significant differences in the interaction analysis with covariate “sex”.
Regarding the haplotype analysis, there were four polymorphisms with significant differences: the haplotype 5 (C-G-A) in NCOA-1, the haplotype 2 (G-G-T-A-A) in NCOA-2, the haplotype 8 (G-G-T-A-G) also in NCOA-2, and the haplotype 2 (T-A) in p300. These NCOA-1 and NCOA-2 significant haplotypes were more frequent in the trans population (OR = 2.62, OR = 2.49 and OR = 12.86, respectively) while the haplotype 2 (T-A) in p300 was more frequent in the cis population (OR = 0.57). The NCOA-2 haplotype 8 (OR = 12.86;
4.3 Concordance of our findings with the literature about receptor coactivators
To our knowledge, no studies have been published about the role of steroid receptor coactivators in the genetic basis of GI. Our data are in concordance with a recent work that showed that the nuclear receptor coactivators, NCOA-1, NCOA-2 and p300, are essential for efficient ER transcriptional activity in the brain [33, 38]. Furthermore, NCOA-1 and NCOA-2 are distributed in several specific areas of the brain in different proportions, such as the hypothalamus and the hippocampus, showing at the same time, difference in the coupling with the ERs [38, 39]. These differential interactions between NCOA-1 and NCOA-2 with the ER subtypes α and β suggest that these brain regions have distinct expression pattern of co-regulators, and understanding how nuclear receptor coactivators function with various steroid receptors is critical to understanding how hormones act in different brain regions.
Moreover, our results are also in concordance with the study of the functional significance of the nuclear receptor coactivator NCOA-1 in the developing brain [40]. The authors, Auger et al., investigated the consequence of reducing NCOA-1 protein during sexual differentiation of the brain, and reported that reducing this protein interferes with the defeminizing actions of estrogen in neonatal rat brains. Their data indicated that NCOA-1 expression is critically involved in the hormone-dependent development of normal male reproductive behavior and brain morphology. Thus, our data are in agreement with the results of Auger et al., [40] since the polymorphic analysis of this coactivator showed significant differences when allelic and genotypic frequencies and haplotypes analyses were carried out.
Our data are also in concordance with other studies about the critical role of p300 and CREBBP in ERα transcription. p300 and CREBBP are two of multiple secondary coactivators recruited by NCOA1, NCOA2 or NCOA3 to form a receptor-coactivator complex that can promote chromatin remodeling and facilitate transcriptional activation [35]. In our work, we found statistical significances in p300, CREBBP, NCOA1 and NCOA2, but not in NCOA3.
Transcription by RNA polymerase II requires the coordinated action of multiple factors such as DNA-binding factors, coactivators, chromatin remodeling, with the basal transcriptional machinery. Futhermore, p300 and CREBBP, do not bind DNA on their own, but they play an essential role in the transcription process mediated by E2 [33, 41]. Thus, ERα functions cooperatively with p300 and CREBBP to increase transcription [42]. Yi et al. [35] demonstrated the quaternary structure of an active complex of DNA-bound ERα, steroid receptor coactivator, and p300 as secondary coactivator. The structural model suggests that the ER binds the ERE-DNA as a dimer and then recruits two NCOAs; these two coactivators, in turn, secure one molecule of p300 to the complex through multiple contacts.
It is very important to maintain the nucleotide sequence of the genes encoding the coactivators in order to maintain the interactions of the ER-E2 -NCOA -p300-CREBBP complex and thus perform the genomic function of estrogens. Therefore, our data are in concordance because finding significant polymorphisms in the sample analyzed may result in ineffective or low effective interactions affecting the E2 target genes involved in brain dimorphism.
In our work, we found 2/9 polymorphisms with statistically significant differences (P9 and P10) in p300 in the interaction analysis with covariate “sex”. This implied differences in haplotype distribution according to sex, and thus, the haplotype 2 (T-A) (Table 8) only showed significant differences in the population assigned as females at birth. The other haplotypes did not show differences in the distribution between cis and trans population, nor in males or females.
Based on experiments in rodents, it is believed that male sexual differentiation of the brain is caused by androgens, after conversion to estrogens by the aromatase. Moreover, observations in human subjects show that the direct effects of testosterone on the developing fetal brain and also during puberty, are of great importance for the development of male gender identity [43]. However, the analysis of the androgen coactivator NCOA-4 did not show any significant data.
Currently, it is still very difficult to interconnect molecular, brain, and behavioral findings [44] due to the complex interactions among behavior, genes, hormones, receptors and enzymes. But we must point out that MRI studies in people with GI, show characteristic brain profiles [5]. Both trans populations (females and males) share some common features: firstly, the involvement of the two ERs in neurobiological origin [21] and, secondly, their cortex, in some regions, is thicker than in cismen [5]. These observations support the hypothesis that transmen and transwomen undergo an atypical developmental process with respect to the sexual differentiation of their cortex [5], hypothetically, under the influence of brain estrogens, androgens, their receptors and some of their coactivators.
4.4 Consistency of the findings with the current hypothesis of a multiplicity of mechanisms involved in the complex “mosaic” model of the mammalian brain
Finally, our data are also consistent with the current hypothesis about the existence of a complex “mosaic” model of the mammalian brain [45], with a multiplicity of mechanisms involved, allowing a variable degree of masculinization/feminization within the brain. The simple model according to which testosterone masculinizes the brain of men away from a predetermined female profile, has been replaced by a complex model, according to which sexual effects on the brains of women and men are exerted by a complex combination of behavior, genetic, epigenetic and hormonal factors [45].
5. Conclusions
Based on the data presented here, we believe that it can be stated that there is a genetic basis for GI. Thus, the coactivators, NCOA-1, NCOA-2 and p300-CREBBP could be considered as candidates for increasing the list of potential “susceptibility” genes for GI. Furthermore, our data continue to support the hypothesis that GI is a multifactorial complex trait, involving intricate interactions among genes, steroids, steroids receptors and coactivators.
Acknowledgments
This work was supported by grants: ED431B 019/02 (EP), PGC2018-094919-B-C21 (AG), PGC2018-094919-B-C22 (RF, EP). We are grateful to everyone who contributed to the study, and to the trans and cis individuals who participated in particular.
Conflict of interest
The authors declare no conflict of interest.