Sex differences emerge early in human development and are detectable from the embryonic stage to parturition [1, 2]. While sex differences in pregnancy outcomes such as fetal birth weight and infant mortality have been recognized for centuries [3], the biological mechanisms through which fetal sex shapes prenatal development remain to be determined. As the interface with maternal circulation and the central coordinator of fetal growth, the placenta is likely to play a starring role in the production of sex-linked prenatal phenomena.

The framing of sex differences in prenatal development generally centers on male vulnerability. Indeed, male fetuses have been reported to be at elevated risk for early preterm birth, term preeclampsia (PE), placental inflammation, premature rupture of membranes (PPROM) and a variety of other gestational complications [4,5,6,7,8,9]. However, female fetuses show a higher incidence of preterm PE, intrauterine growth restriction (IUGR) and being small for gestational age (SGA) across multiple populations [10,11,12,13,14], suggesting that sex-biased prenatal vulnerability can manifest in unique and context-specific ways [15,16,17,18,19,20].

How does sex manifest in the human placenta?

The placenta is a critical determinant of both fetal and maternal health throughout gestation. In addition to providing the interface for the exchange of nutrients and waste, the placenta is also a source of hormones and immune factors that facilitate pregnancy maintenance and fetal growth [21]. During the process of human placentation, trophoblast cells from the outer trophectoderm layer of the blastocyst invade maternal decidua to form the placenta and chorionic membranes. The resulting extraembryonic compartment shares the biological sex of the developing embryo. Because fetal cells can express paternal antigens that are immunologically distinct from maternal cells, the successful establishment of maternal immune tolerance to the fetal “graft” is an essential requirement for successful placentation and pregnancy. If these finely tuned interactions become dysregulated, placental dysfunction can result, leading to complications such as spontaneous abortion, preterm birth, preeclampsia and intrauterine growth restriction [22]. Given the sex differences observed in common gestational pathologies, the sex of the trophoblast and the other cells that compose the placenta is likely to influence the interactions between fetal and maternal cells.

The primary known factor giving rise to sex differences in early embryogenesis is sex chromosome-linked gene expression (23). In addition to differential expression of X and Y transcripts themselves, differences in autosomal gene expression in early embryonic and extraembryonic tissues have been observed and are likely to play a role in sex-biased fetal outcomes. Towards the end of the first trimester, differential expression between the sexes is likely to reflect the interaction of cell-intrinsic chromosome complement with extrinsic endocrine signals from the fetal compartment that accompany gonadal differentiation. Both potential mechanisms will be expanded upon later in this review.

The dynamic placental transcriptome

Placentas from term pregnancies have been frequently examined for transcriptomic differences based on fetal sex over the past decade, and these studies have revealed widespread differences in hormone signaling, immune signaling, and metabolic functions [24,25,26,27,28]. Sood et al. [24] first observed sex differences in both sex chromosome and autosomal gene expression in term placentas by microarray, identifying JAK-STAT-related immune regulation as a central signaling hub. Osei-Kumah et al. [25] also highlighted sex differences in cytokine signaling in placentas from pregnancies complicated by asthma, as well as glucocorticoid hormone signaling. Cvitic et al. [26] performed the first cell-type specific analysis, isolating and culturing different trophoblast and endothelial cell types from male and female placentas and subjecting them to microarray analysis. TNFα and NFкB signaling pathways emerged as a major node of sexually dimorphic gene expression patterns, showing elevation in male placental endothelium. Sex-linked alterations in pro-inflammatory signaling at the mRNA and protein levels are a theme across multiple studies in both healthy and inflamed placenta [24, 25, 29]. In line with this, cultured male trophoblasts from healthy term placentas produce more TNFα and less IL-10 than female trophoblasts in response to lipopolysaccharide [30].

Beyond differences in inflammatory signaling, a meta-analysis of transcriptome data from term placentas by Buckberry et al. [27] observed 142 genes differentially expressed (DE) between male and female placentas, with > 60% being autosomal, including genes related to gene transcription, cell growth, proliferation, and hormone signaling. Higher female expression from the LHB-CGB cluster was detected, which includes genes involved in placental development, maintenance of pregnancy and maternal immune tolerance of the conceptus. Osei-Kumah et al. [25] and Sedlmeier et al. [28] reported that female placentas at term are more responsive to changes to both maternal inflammation and diet, with male placental gene expression appearing less sensitive to environmental perturbations. Given the central roles of hormonal and immune regulation in the production of pregnancy pathologies like preterm birth, these placental differences likely play a major precipitating role in sex-biased pregnancy complications and fetal outcomes if they are also present in the placenta earlier in gestation.

Differences in transcript and protein abundance in term placentas are informative but given the rapid and dynamic development of the human placenta [31], sex differences that exist at term are unlikely to align precisely with those present at earlier stages when the fetal–maternal interface is being formed and the foundations of a healthy or diseased pregnancy are established. Several groups have performed transcriptomic profiling of the first-trimester human placenta (reviewed by Yong and Chan [32], 2020: Table 1), however fetal sex as a variable in these datasets is often left unreported or, if reported, is not directly analyzed. In the recent years, four studies have examined fetal sex in early human placental development (Table 1) [33,34,35,36]. In the following sections, we review these findings and consider the evidence for the genetic, hormonal and immune mechanisms that are theorized to account for sex differences in early human placenta and highlight the cellular and molecular processes that are most likely to be impacted by fetal sex.

Table 1 Summary of early human placenta transcriptome studiesThe sex chromosomes

Sex chromosomes account for the earliest, most pronounced, and most reproducible sex differences in gene expression (23). Potential sources of variation related to sex chromosome complement include expression from the Y chromosome in XY males and selective expression from the second X chromosome in XX females.

An early model of sex chromosome gene dosage compensation held that male cells contain one X chromosome and one Y chromosome, while female cells contain one active and one compacted and inactive X known as a Barr body [37, 38]. This inactivation of one X theoretically ensures that XX transcription matches the dosage in XY males, leading to the model: 1 active X → female phenotype, 1 active X + Y → male phenotype. This simple model is complicated by the existence of the two human pseudoautosomal regions (PARs 1&2), which are not inactivated in XX cells [39, 40], and later by the discovery of other X transcripts located outside of the X chromosome PARs that escape inactivation in a variable and cell-specific manner [41]. Extraembryonic tissues share the sex karyotype of the fetus, and genes in all these categories are likely to play some role in producing sex differences in human placenta.

Pseudoautosomal genes in placenta

Pseudoautosomal regions (PARs) are short nucleotide sequences on the ends of sex chromosomes exhibiting homology between the X and Y. They are not inactivated in XX cells and exhibit X and Y variants that can be distinguished via PCR and high-throughput sequencing. PAR1, the larger and better characterized PAR, is located at the ends of Xp/Yp and contains at least 24 genes encoding proteins involved in functions including transcriptional regulation, RNA splicing, signal transduction, and cell adhesion [42]. PAR2, located at the ends of Xq/Yq, is evolutionarily recent and unique to humans, making humans the only species known to have 2 distinct PARs [43]. Interestingly, both PARs are enriched for genes that underlie immune signaling, including IL3RA, IL9R, CSF2RA, and CD99 (Table 2). PAR1 gene ASMTL, PAR1 pseudogene CD99P1, and PAR2 gene VAMP7 were upregulated in the male placenta at mid-gestation [35], likely attributable to increased expression from the Yq allele. In a sex-based reanalysis of Soncin et al. (21, GEO accession number: GSE107824) both CD99 and VAMP7 trended towards upregulation in cytotrophoblasts from across gestation in male fetuses compared to females. In a meta-analysis of sex differences in human term placenta, PAR1 genes were shown to be elevated in males [27], suggesting that this may be a persistent bias throughout development. PAR gene dosage is altered in sex chromosome aneuploidies such as XXY (Klinefelter syndrome) and monosomy X (Turner syndrome), and abnormal gene dosage of PAR genes is thought to contribute to the elevated risk of pregnancy complication and spontaneous abortion in aneuploid sex chromosome karyotype pregnancies [44,45,46].

Table 2 Gene symbols and full gene names of all transcripts mentionedMechanisms of XX protection in placenta

In regions of the X chromosome outside of the PARs, X-chromosome inactivation (XCI) normalizes gene dosage between XY males and XX females. Prior to XCI, which occurs between implantation and tissue differentiation, X-linked genes in XX cells are expressed from both the paternal and maternal alleles. XCI is essential for the development of XX conceptuses past the early embryonic stage [47] and occurs via epigenetic compaction of either the paternal or maternal X chromosome into inactive heterochromatin. In monotremes and marsupials, X inactivation is imprinted, with the paternally derived X inactivated in every cell. In Eutherians, the process of X inactivation has become more nuanced, sometimes parentally imprinted and other times randomized via stochastic expression of the X-linked non-coding RNA XIST. Murine extraembryonic tissues retain the selective paternal X inactivation seen in non-Eutherian mammals, however this imprint is lost and reset in the cells of the inner cell mass, where a new round of random inactivation allows for mosaic paternal and maternal X expression (48,49,50).

In the human placenta, the existence of parentally imprinted X inactivation has been a topic of controversy. Some reports indicate skewed X inactivation [51,52,53], and others report unbiased expression that suggests a more random process matching that of the inner cell mass [54, 55]. An archived report awaiting peer review by Phung et al. [56] provides evidence for a clonal model of X inactivation in placenta, suggesting that there are regions of paternal X inactivation and maternal X inactivation with an overall skew towards paternal inactivation. Sampling bias including the timing, the exact tissue compartment and other confounding factors may help to resolve the puzzling and contradictory findings in these studies. Two regions of opposite X inactivation pooled together may appear randomly biallelic, and sampling from only one region may lead to inaccurate reports of single-parent X inactivation. In the future, a single-nucleus approach will be essential to fully understanding XCI in the placenta especially given the multinucleated nature of the syncytium [57, 58].

The evolutionary movement away from strict paternal X inactivation in Eutherians highlights the dynamic balance of selection pressures on placental X expression. While inactivation of the paternal X through strict imprinting may minimize contact between surveilling maternal immune cells and possible foreign paternal antigens, the genetic robustness conferred by mosaicism may provide a survival benefit. Skewed but non-imprinted X inactivation in placenta may reflect a process of internal selection, particularly in tissues where development involves cellular competition for growth factors resulting in differential cell survival [41]. If one parent’s copy of the X proves more advantageous to the survival of a given cell type, those clones will prevail in that tissue and X inactivation will appear skewed toward one parent without the necessity of a priori genomic imprinting [59, 60]. This may help to explain observations of skewed, clonal X inactivation in tissues like the placenta. The wide heterogeneity within and between individual placentas is also suggestive of an ongoing evolutionary process, where different strategies are viable in different contexts [56].

In addition to the potential mutation-masking effect of X mosaicism, XX karyotype comes with another advantage. While the original model of X inactivation held that the inactivated X is functionally silenced outside of the PARs, it is now well established that select transcripts can be expressed from the inactive X, showing biallelic expression in a chromosome dosage-dependent manner [41]. These XCI “escapees” account for a large portion of the genes upregulated in the early XX placenta compared to the XY placenta. In 2018, Gonzalez et al. showed that out of 58 differentially expressed genes in the late first-trimester placenta, over a third were X-linked genes upregulated in female samples, and half of those genes were known to escape X chromosome inactivation [33, 61]. Among this group, DDX3X, EIF1AX, KDM5C, KDM6A, OFD1, RPS4X, SMC1A, and ZFX were confirmed as upregulated in XX females (Fig. 1) a subsequent analysis suggested that these genes escape X inactivation in chorionic villus (CV) in a robust manner [35]. X-linked genes BRCC3, CHM, EGFL6, EIF2S3, HDAC8, MXRA5, NUDT10, PUDP, RBM41, SMARCA1, STS, THOC2, TRAPPC2, YIPF6, ZMAT1, and ZRSR2 were also found to be upregulated in females in one of the two datasets, indicating that these genes may escape X inactivation in a less robust fashion, potentially varying by cell type. Indeed in a single cell transcriptomic analysis of the maternal–fetal interface, trophoblast cells appeared to have unique X chromosome genes upregulated in females compared to males, which included MAGEA4 (melanoma associated antigen 4) and TMSB4X (thymosin beta 4) [34]. Phung et al. have suggested that while a small group of X genes reproducibly escape inactivation across individuals and tissue regions (PLCXD1, GTPBP6, PUDP, CSF2RA, SLC25A6, ASMTL, AKAP17A, DHRSX, STS, EIF2S3, ZFX, DDX3X, KDM6A, DIPK2B, UBA1, SMC1A, RENBP, FLN4), others exhibit variable and heterogeneous escape that varies between individuals and between tissue regions in the same individual (CD99, EGFL6, RPS6KA3, MBTPS2, SEPT6, CYBB, MED14, USP9X, CDK16, TIMP1, WDR13, MAGED2, OPHN1, EFNB1M, PIN4, RPS4X, ATRX, TSPAN6, ACSL4, PLS3, DOCK11, IL13AR1, Cxorf56, GPC4, HTATSF1, GABRE, BGN, AVPR2, ARHGAP4, HCFC1, IRAK1, MECP2) [56]. This finding may help explain variability in reports of the specific transcripts that escape X inactivation in the placenta and suggests that escape from X inactivation could act as a tunable protective mechanism, providing unique benefits in different cell types.

Fig. 1

Ideogram visualization of significant (p ≤ 0.05) sex-biased gene expression on sex chromosomes comparing two datasets (G: Gonzalez et al., B: Braun et al.). Includes protein-coding RNA only. M: male (green triangles), F: female (orange circles)

It is currently unknown whether XCI escapees play a functional role in the early stages of human placentation, however X inactivation escapees have been shown to contribute to female protection against mitochondrial stressors in human third trimester placenta. Gong et al. [58] demonstrated that the propylamine transferring enzyme spermidine synthase (SMS) shows X inactivation escape in term placentas, and that its relative insufficiency in male placenta is associated with vulnerability to mitochondrial stressors. It was observed that polyamine metabolite diacetylspermine is higher in the female placenta and in the serum of women pregnant with a female fetus and correlated both with an increased risk of preeclampsia and a decreased risk of fetal growth restriction (FGR). To our knowledge, this study provides the first direct connection between sex differences in placental gene expression, changes in metabolism, and pregnancy outcome. In addition, Howerton et al. [62] demonstrated that escapee gene O-GlcNAc transferase (OGT) mediates female resilience to the effects of maternal psychosocial stress in mice and found its abundance to be higher in human female term placenta as well. OGT and SMS were not found to be significantly differentially expressed in 11- to 16-week CV, however expression patterns indicate a trend toward higher expression in females which is likely to increase over time [35], and may be more pronounced when looking at individual cell types in the CV. Functional studies of early human placentation to identify genes that escape XCI are ethically and technically challenging but will become more feasible given recent advances in 2D trophoblast stem cell models and 3D trophoblast organoid models [63,64,65].

The Y chromosome

The Y chromosome is much smaller and gene-sparse compared to the X, with most of its contents primarily involved in spermatogenesis and male fertility [66]. Chief among these is sex determining region Y (SRY), which triggers a developmental cascade that converts the nondifferentiated fetal gonad to testis, leading to the production of androgens such as testosterone from Leydig cells, which then trigger the canonical hormonal masculinization of the male fetus. A subgroup of Y chromosome genes with X homologs outside the pseudoautosomal regions have been shown to be essential for embryonic development, primarily through chromatin modification and RNA splicing [66].

Gene expression from the Y chromosome is detected in early CV samples, with DDX3Y, EIF1AY, KDM5D, PCDH11Y, RPS4Y1, USP9Y, UTY, and ZFY expressed consistently in males across datasets (Fig. 1). Additionally, transcripts NLGN4Y, TBL1Y, and TMSB4Y were detected in a similar analysis which sequenced a broader variety of RNA types [33]. Single cell sequencing of villus tissues collected from late first-trimester healthy pregnancies showed that Y transcripts such as DDX3Y, EIF1AY, RPS4Y1 were specifically upregulated in male placental cell types including trophoblasts, stromal cells and Hofbauer cells [34].

Importantly, male chorionic villus expresses several Y chromosome transcripts that correspond to peptides that compose the human H-Y antigen, which is detectable in syncytiotrophoblast debris [67]. Several of these Y transcripts correspond to peptides that are presented by the major histocompatibility complex (MHC), including KDM5D, DDX3Y, ZFY, and UTY, which were expressed in early CV in two separate analyses (Fig. 1). In single cell analysis, DDX3Y was expressed consistently across different cell types such as trophoblasts, stromal cells and Hofbauer cells in early pregnancy with a male fetus [33,34,35].

X chromosome expression in the male placenta

While the majority of upregulated genes in early male placenta were Y-linked, three X-linked genes (ARHGEF9, ARMCX3, and ARMCX6) were also identified as upregulated by Gonzalez et al. [33], perhaps due to upstream Y-linked genes or downregulation of the second X chromosome in females. Interestingly, the protein encoded by ARMCX3 regulates migration and invasion in tumor cells, functions which are also relevant to placentation. ARMCX6 was also found to be upregulated in male term placenta [26], suggesting that this trend may persist throughout placental development. Additionally, Braun et al. [35] found the X chromosome genes CAPN6, GLA, IQSEC2, MAOA, PORCN, and TMEM164 to be upregulated in males relative to females. These comparisons suggest that complex sex-linked regulation of gene expression beyond the effect of X dosage compensation likely occurs.

Autosomal sex differences in early placenta

As expected, the most pronounced differences in gene expression between male and female placentas have consistently been localized to the sex chromosomes, however widespread differences in autosomal gene expression have been detected as well, and account for some of the most dynamic developmental sex differences in gene expression (33, 35) found that 31% of differentially expressed sex chromosome genes detected in the late first trimester were also DE in term placenta tissue. However, there was no overlap in the sex-based DE autosomal genes between early and term placenta, suggesting that sex-selective transcriptional programs correspond to specific developmental stages. This observation highlights the need to examine placental gene expression and function across gestation.

Chromosomal location of DE genes

Braun et al. [35] found that the group of autosomal genes upregulated in females was distributed widely throughout the genome, while the autosomal genes upregulated in males were clustered in particular loci including Chr11q13.1 (CYB561A3, FAU, GPR137, SCYL1, TMEM258, VPS51