A growing number of studies show systematically poorer social, economic and health outcomes in gender minorities, in addition to many other difficulties, compared to cisgender individuals.5 This constitutes a risk factor for mental health disorders and other unfavorable experiences and outcomes, such as stigmatization, bullying, self-harm, suicidal ideation and suicide attempts, etc.6,7

People’s beliefs regarding the cause of certain conditions have a marked impact on how they perceive other individuals, so that the more “controllable” a trait or condition is perceived to be, the more that individuals are construed to be responsible for it.8,9 In contrast, the current literature suggests that biomedical research and educational approaches that integrate knowledge on biological contributions to complex conditions or traits (eg, anorexia nervosa or obesity) can help reduce negative stereotypes and stigma and increase solidarity toward these groups.5,9–11

In recent decades, our understanding of the sexual differentiation of the mammalian brain has changed radically. The simple model according to which testosterone masculinizes the brain of males away from a default female form has been replaced by a more complex scenario, according to which genetic, epigenetic, hormonal and environmental factors intervene in the sexual differentiation of the brain, acting through multiple, partly independent mechanisms.12

The broad spectrum of gender identity is not linked to variations in a single gene, but constitutes a multifactorial, complex or polygenic trait. In other words, multiple genetic variants, each with a small individual effect, contribute additively to the variance of complex traits (but do not determine them) along with other nongenetic factors.5,13 In this model, the contributing factors follow a continuous distribution in the population, so that while two individuals may have very different phenotypes (eg, gender identities), the entire population exists along a single spectrum without clear divisions (that is, there is no clear boundary between cisgender and transgender identities).5 This model contradicts the notion that gene variants could be used to identify or predict the gender identity of an individual.5,13

At the neuroanatomical level, the multifactorial model translates to the substantial variation observed in the degree of masculinization/feminization of different features in a single brain, with considerable overlap between different gender identities at the individual level, which has led to the development of the “mosaic” brain hypothesis.12,14–16

Both the multifactorial model and the mosaic brain hypothesis have the potential to reduce the stigma endured by gender minorities by highlighting the continuous, as opposed to dichotomous, nature of gender identity.5 They also entail that the search for significant associations between underlying biological causes and the resulting phenotype is complicated, further hindered by the fact that many relevant gender identity variables cannot be modelled in animals, as they can only be recordered in humans. In the investigation of multifactorial traits, the extremes of the distribution are usually studied first, as this makes differences between groups more evident. In the study of gender identity, this involves the comparison of cisgender individuals with binary transgender individuals with clinically significant distress, even with a heterosexual orientation.17 In addition to more exploratory and validation studies in this subset of transgender individuals, research needs to be expanded to study other combinations of gender identity, expression and sexual orientation and other minorities, in addition to the small percentage of transgender individuals that transition back to a cisgender identity or to a nonbinary identity.18

Although the interrelated nature of all the elements at play may make the division into sections seem artificial, for the sake of clarity we are going to structure the discussion of the biological basis of the variation of gender identity into three broad sections: hormonal component of sexual differentiation, neuroanatomic and neurodevelopmental component and genetic and epigenetic components.

Many important psychological traits exhibit sex-based differences and are influenced by sex hormones at different stages of development. The differentiation of genitalia occurs much earlier in human development (within two months of conception) than the sexual differentiation of the brain (second half of gestation). Thus, these two processes may be affected independently by different interactions between genes, sex hormones and the developing brain cells.19,20

Unfortunately, studying how hormones influence gender development is challenging, as hormones cannot be manipulated experimentally in human beings, and human behavior is strongly influenced by the social context.21 As a result, this field advances much lower than others and has to borrow, to a great extent, from the work conducted in other species, in which there has been significant progress in the past 20 years.

Theisen et al.13 carried out a review on the subject that we will summarize here. What has been observed in rodents is that activation of estrogen receptors (ERs) during neurodevelopment gives rise to sex differences in the brain that are specific to each region.22 Contrary to what may be expected, this activation of ERs takes place in male, not female, individuals.22 The reason is that, during the perinatal period, there is a rapid and transient surge in the testicular secretion of testosterone, which readily converts to estradiol, thus stimulating the ER-activated neurodevelopmental pathways that give rise to a male-sex pattern of behavior.22,23 On the other hand, during the same period, ovaries are inactive in females, and the resulting lack of stimulation of ERs steers sex-specific neurodevelopment toward feminization.22 In rodents and at least nine other mammal species (including nonhuman primates), the effects of this process on sex-specific behavior only emerge with the reactivation of the hypothalamic-pituitary-gonadal axis, that is, in puberty.13,22,24,25

Although we have not been able to study the sexual development of the human brain as thoroughly, the pattern described above is consistent with the average timeline of transgender identity development, in which there is frequently significant worsening of dysphoria during puberty, and ample evidence suggests that intrauterine exposure to sex hormones affects sex-specific behavior in humans.5,25,26

Several studies based on magnetic resonance imaging (MRI) data have found that most sex/gender differences are small at the group level, with the distribution of these features exhibiting considerable overlap with the broad spectrum that characterizes gender identity.12 Brains in which the volume of all regions is consistently in one end of the maleness-femaleness continuum are rare.14,27,28 “Mosaic” brains are much more frequent, both in cisgender and transgender individuals as well as heterosexual and homosexual individuals, who exhibit a mix of these regions.14,27–29 This is consistent with the conclusion derived from animal models that, to a large extent, the sexual differentiation of different brain tissues progresses independently (see the previous section on the hormonal component of sexual differentiation). Therefore, although it is possible to distinguish typical male and female neuroimaging patterns at the group level, studies to date have failed to identify specific brain features that differ consistently between cisgender and transgender individuals (reviewed in Frigerio et al.30).

In addition, it is important to avoid confusing a difference and its source when discussing brain differences in men and women, as the brain is a plastic organ that continues to change throughout life.12 Thus, there is no way to determine whether observed differences between groups reflect differences in the life experiences of individuals with different identities or predated those experiences.31,32 It is also impossible to determine whether differences in specific brain structures are responsible for different identities or to establish a linear correlation between anatomy and behavior.27,33 These cause-effect relationships are further complicated by the evidence that brain functions are generally not localized to a particular cerebral structure, but distributed over neural circuits that involve the interaction of a large number of regions.12

In fact, the analysis of the structure of the entire brain carried out by Joel et al.27 revealed that a male and a female were nearly as likely to have the same pattern of brain architecture as two females or two males. This is true even if it is possible to predict whether a person is male or female based on their brain architecture27,34 and despite the fact that various studies using structural and functional MRI have described different cerebral phenotypes in transgender and cisgender women and men (eg,28,34,35).

The fact that gender identity is a multifactorial trait poses a challenge for genetic research, as, rather than approaching investigation as the search for a single so-called “transgender gene”, it has to focus on the polygenic milieu that contributes to its development, which may differ completely from one individual to the next.13

As is the case of other complex traits, twin studies, which are among the most powerful designs available to estimate genetic and environmental effects on trait variation, provided the earliest evidence that genetic factors contribute to the development of gender identity. A review analysis of these studies yielded estimates of the heritability of transgender identity ranging from 11% to 47% in natal femaled and 25%–47% in natal male individuals.5 This was consistent with the values obtained for other behavioral and personality traits, with heritability estimates that generally range between 30% and 60%.36

Given the contribution of hereditary factors to variation in gender identity, several studies have focused on individual candidate genes in recent decades.5 However, the small effect of each of the multiple genes that contribute to a complex character complicates the identification of conclusive associations. Approaches based on whole-genome data, more suitable for teasing out the genomic and epigenomic components of complex traits, are starting to be applied in gender identity studies, contributing insights into the complexities of its development, but keeping in mind that each associated gene has to be viewed as one among the multiple potential genetic factors that contribute to the development of the final phenotype, as opposed to an important causal genetic mechanism. We also must not forget that any findings need to be replicated in larger, independent cohorts. These limitations are evident in the lack of overlap between some of the results of the studies published to date, which are detailed in Table 1.

In the context of the hormonal component of sexual differentiation described in the first section, it has been hypothesized that functional genetic variants may affect sex hormone signaling, leading to variation in the sexual differentiation of developing brains. Concordantly, different studies have found associations between transgender populations and the DNA sequence of the genes encoding androgen receptors (AR), estrogen receptors (ERα and ERβ) and their ligands and coactivators (Table 1). There is also evidence of associations between the transgender identity and variations in other genes that may be associated with the sexual differentiation process at both the hormonal and the neural level (Table 1).

On the other hand, the investigation of epigenetics complements genetic studies by reflecting the interaction between genes and the environment, and there is now evidence that epigenetic regulation plays a key role in sexual differentiation processes in the brain.12,15,37 Therefore, and given the sensitivity of gender identity development to environmental stimuli, epigenetic regulation is one of the many factors at play in this process.15,16,24

Studies on the epigenetic regulation of gender identity have focused on DNA methylation, an epigenetic mark that has a context-dependent influence on gene expression. DNA methylation is very dynamic during preimplantation embryo development and in cellular differentiation,38 but it remains susceptible to environmental factors throughout the lifespan.39 Sex-specific methylation patterns have been found in various types of cells, some potentially attributable to changes in hormones. However, it is difficult to differentiate hormonal versus genetic influences in methylation, as hormones are themselves regulated by the chromosomal sex.

Two studies of DNA methylation carried out in samples of hormonal treatment-naïve transgender individuals found differences in methylation patterns compared to cisgender individuals, overall and in genes involved in brain development (Table 1). Fernández et al.17 also found differences in methylation of two genes, CBLL1 and DLG1, correlating to differences in cerebral cortical thickness in trans men.

Last of all, research has also been conducted on the impact of gender-affirming hormone therapy (GAHT) on epigenetic marks. Studies have found changes in methylation within six months of treatment initiation in both the ERα, ERβ and AR genes and at the whole-genome level, with sex- and age-specific differentially methylated regions exhibiting a shift towards the methylation signature of the GAHT-naïve opposite sex (Table 1).

As a final reflection, we ought to take into account that the evolutionary advantage of sexual reproduction stems from the enormous increase in inter-individual variation that it produces, resulting from the combination of maternal and paternal DNA and subsequent developmental processes.12 The continuous distribution of gender identity in the population, in opposition to the dichotomous differentiation of the reproductive system into male and female phenotypes, appears to be just another aspect contributing to the role of sexual reproduction as a source of variation.40