Conserved gene expression patterns in the placenta underlie pregnancy complications associated with hypoxia in mice and humans

Abstract

During pregnancy hypoxia leads to complications like fetal growth restriction (FGR). Maternal inhalation hypoxia during late murine pregnancy impairs placental phenotype and fetal development in a severity-dependent manner. To identify the molecular mechanisms and sex-specificity of these effects, placentas from pregnant mice exposed to moderate (13% O₂) or severe (10% O₂) hypoxia were analysed using RNA-sequencing, qPCR, western blotting, histology, and nutrient transport assays. Transcriptomic profiling of male 13% O₂ placentas revealed differential gene expression regulating calcium binding, lipid metabolism, and peroxisome proliferator-activated receptor (PPAR) signalling. In both sexes, hypoxia reduced fetal weight and altered placental nutrient transport in a severity-dependent manner. Abundance of PPARα, PPARγ, and associated targets varied with sex and hypoxia severity. Placental calcium deposition was significantly increased by hypoxia irrespective of severity. Human placental datasets revealed that orthologues of key hypoxia-responsive genes in the mouse placenta are also associated with pregnancy outcomes in human pregnancy. These findings implicate placental PPAR signalling and calcium dysregulation as potential mediators of FGR in compromised pregnancies.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-026-06110-7.

Introduction

Hypoxia during pregnancy is a major contributor to complications such as pre-eclampsia and fetal growth restriction (FGR), both of which are associated with increased maternal, fetal and neonatal morbidity and mortality [1, 2]. These conditions are prevalent in pregnancies at high altitude [3, 4], but can also occur at sea level due to smoking, anaemia, and respiratory disease [5–8]. Abnormal placental development, such as insufficient decidual invasion or spiral artery remodelling can also result in hypoxia of the placenta and is associated with adverse pregnancy outcomes [9–12]. In addition, a growing body of evidence has linked placental hypoxia to adverse offspring health outcomes in adulthood [13–16].

The placenta is the critical interface between the mother and fetus, responsible for nutrient and gaseous exchange, endocrine signalling and immunological regulation for successful pregnancy [17]. It acts as a nutrient sensor and responds to environmental cues, including stress and oxygen availability, through dynamic changes in structure and function mediated by intracellular signalling pathways [15, 18–20]. In early gestation, placental development occurs under a low oxygen environment [21, 22]. However, as fetal growth accelerates in later gestation, oxygen demand increases, and placental function becomes more dependent on adequate oxygen delivery. Under adverse conditions, like hypoxia in late gestation [23], the placenta can adapt to optimise fetal growth, though the mechanisms governing these adaptations remain incompletely defined.

In the mouse, the labyrinthine zone, responsible for nutrient and gaseous exchange, undergoes structural and functional remodelling in response to maternal inhalation hypoxia in a severity-dependent manner [16]. Moderate hypoxia (normobaric 13% O₂; equivalent to oxygen levels at ~ 3,700 m altitude) enhances fetal capillary expansion, reduces trophoblast barrier thickness, and increases glucose transport to the fetus with only a modest (~ 5%) reduction in fetal growth [24]. In contrast, severe hypoxia (normobaric 10% O₂; equivalent to oxygen levels at ~ 5,800 m altitude), thickens the trophoblast barrier, reduces amino acid transport, suppresses mitochondrial oxidative metabolism [23] and results in a 20% reduction in fetal growth (4-times that seen in moderate hypoxia). Despite these clear phenotypic changes, the molecular signalling pathways driving these severity-dependent adaptations, and their implications for human pregnancy, remain poorly understood.

Differences in gene expression have been observed between placentas of male and female fetuses [25, 26], and may reflect different growth strategies between fetal sexes [27–30]. Emerging evidence also indicates that fetal sex influences the susceptibility to gestational insults and pregnancy outcomes [31]. For instance, female fetuses have been linked to higher rates of preeclampsia and preterm birth [27], whereas males may be more vulnerable to environmental challenges due to prioritisation of growth over placental reserves [27–30]. Sex-specific responses have been documented in both human and animal models in response to developmental cues, as well as maternal asthma [32], diet [33], glucocorticoid exposure [34], aging [35], and hypoxia [36–39]. Yet, the extent to which fetal sex modulates hypoxia-induced placental signalling and its consequences for fetal development remains insufficiently analysed [16, 40].

Here, we aimed to identify placental signalling pathways that are regulated by moderate gestational hypoxia (13% O₂), and tested the hypothesis that fetal sex and severity of hypoxia can impact these pathways with functional consequences for fetal growth. More specifically, in an initial discovery phase, we performed transcriptomic analysis (RNA-sequencing) on the placental labyrinthine zone (which is specialised in substrate transport) solely from male fetuses exposed to moderate hypoxia (13% O₂) or normoxia. We then validated changes in candidate genes and signalling pathways using qPCR, western blotting, and histological analysis and related these changes to in vivo measures of placental nutrient transport and fetal growth with respect to both sex and hypoxia severity (13% and 10% O₂). To assess translational relevance, we interrogated human placental datasets to determine whether the key hypoxia-responsive genes identified in mouse placenta were associated with fetal growth outcomes and preeclampsia in human pregnancy.

Materials and methods

Animals

This study utilised fetal and placental weight measurements, in vivo placental transport data, and placental samples obtained from a previous published study in which mice were exposed to hypoxia during the final third of gestation [24, 41]. All animal procedures were conducted in accordance with the UK Animal (Scientific Procedures) Act 1986. In brief, age-matched, time-mated pregnant C57BL6/J female mice with ad libitum access to food and water were exposed to either normoxia (21% O₂; N), moderate hypoxia (13% O₂; 13% hypoxia) or severe hypoxia (10% O₂; 10% hypoxia) from gestational day (D) 14 to 19 (term=D20). On D19, mice underwent a placental transport assay (described below), schedule-1 killed by cervical dislocation, and fetal and placental tissues were collected and weighed. Fetal sex was retrospectively determined from placental samples by PCR amplification of the sex-determining region Y (Sry; Fwd 5’-GTGGGTTCCTGTCCCACTGC-3’, Rev 5’-GGCCATGTCAAGCGCCCCAT-3’and an autosomal gene (Csa; Fwd 5’-TGGTTGGCATTTTATCCCTAGAAC-3’, Rev 5’- GCAACATGGCAACTGGAAACA-3’) as control, using RedTaq PCR reaction mix (Sigma, USA). Previously published fetal and placental weight and transport data [24] were re-analysed to include fetal sex as a biological variable after the current genetic determination of sex.

Placental transport assay

As described previously [24, 42], dams were anaesthetised via intraperitoneal injection of fentanyl-fluanisone and midazolam in sterile water (1:1:2, 10 µg/ml; Janssen Animal Health, High Wycombe, UK). Non-metabolisable radiolabelled analogues of glucose (methyl-d-glucose, MeGlu, Perkin Elmer, Waltham, MA, USA) and amino acid (methyl-amino isobutyric acid, MeAIB, Perkin Elmer, Waltham, MA, USA) were administered via the jugular vein. Radioactivity was quantified in maternal plasma, fetuses, and placentas to calculate placental clearance (µl/min/g of placenta) or tracer accumulation (µl/g fetus or placenta) for MeGlu and MeAIB.

RNAseq

Total RNA was extracted from dissected placental labyrinthine zones, from four male placentas per group (13% hypoxia and normoxia), each from different litters using RNeasy Plus Mini Kit (Qiagen), following the manufacturer’s instructions. RNA-sequencing (RNAseq) was performed on samples from male fetuses (discovery phase), with analysis by quantitative real time PCR (qPCR) in both sexes and hypoxia severities. RNAseq libraries were prepared and sequenced using 50 bp single-end reads on a HiSeq4000 platform. Sequencing yielded over 400 million reads in total and passed all quality control (QC) metrics. After QC and trimming, reads were aligned to the GRCm38 reference genome (Ensembl). Gene quantification and differential expression analyses were performed with featureCounts and DESeq2. Genes with an adjusted p-value (padj) < 0.05 were considered differentially expressed genes (DEGs). Pathway analyses were performed using MsigDB [43], incorporating gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) terms. The complete RNAseq dataset (12,782 genes), ranked by fold change and p-value, was further analysed by gene set enrichment analysis (GSEA) using the Fgsea R package.

To explore regulatory mechanisms, promoter regions of the DEGs were analysed for PPAR, EPAS1, HIF1α and ARNT binding motifs using the MoLoTool [44]. This was supported by a database of experimentally validated and predicted PPAR binding sites (Fang et al., 2016). Additionally, a curated list of extracellular matrix (ECM) components (Naba et al., 2017) and UniProt annotations were used to identify genes associated with ECM and calcium-related functions, respectively. Human orthologues of DEGs were identified using Ensembl and queried in the Pregnancy Outcome Prediction Study (POPs) ‘Placentome’ database [45], to assess dysregulation in placentas from pregnancies complicated by preeclampsia (PE) or small for gestational age (SGA) births. Correlation between the expression of each orthologue and birth weight was assessed [45] using linear modelling in R.

qPCR and analysis

Total RNA was extracted from frozen placental labyrinthine zones of male and female fetuses across all experimental groups and reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Thermofisher, UK). Real time quantitative PCR (qPCR) was performed using gene-specific forward and reverse primers (Table 1), PowerUp SYBR Green Master Mix (Applied Biosystems), and a 500 Fast Real-Time PCR System (Applied Biosystems, Foster City, US). The cycling protocol consisted of an initial denaturation step at 95 °C for 3 min, followed by 40 cycles of 95 °C for 1 min, 95 °C for 15 s and 60 °C for 1 min. Expression levels of target genes were normalised to the geometric mean expression of Ywhaz, Gapdh and Ubc, which showed stable expression across experimental groups. Relative expression was calculated using the ΔΔCt method, with values expressed relative to the normoxic female controls.

Table 1.

Primer sequences used for qPCR. Primers designed using Primer-BLAST

Target genes	Forward primer 5’-3’	Reverse primer 5’-3’
Apoc1	CTGAGAGATCCTTAGATCCAGGGTG	CGGGCCTTGTCTTCCAAAGT
Bcat2	CTATGGACCCACTGTGGCTGT	CCAGCTCCAGTACTCCGTCT
Bmp2	CTTCCATCACGAAGAAGCCGT	TCTGTTCCCGGAAGATCTGGA
Cxcl12	CCTTCAGATTGTTGCACGGC	CTTGCATCTCCCACGGATGT
Gkn1	CGAGGAGATTCCAGGACCAAAC	ATTGCGATGGTCAGAGCCC
Gkn2	CGAAGGTGGACTGGTTCCTG	TCCAGTGCTCACCTCTTTTGG
Hmgcs2	CAGCTACTGGGATGGTCGC	CCATGTGAGTTCCCCTCAGC
Lpl	CAGATGCCCTACAAAGTGTTCCA	ATTTGTGGAAACCTCGGGCA
Ndrg1	ATGGTAGAGGGTCTCGTGCT	TGTGTATCTCCTCCTTGCCGA
Ntrk2	GTTGCTCCCCAGCCCTGA	GACCAAATGGTTCACAGTGGC
Pdgfb	GGCGCGCTCCGTCTAC	AATGGGATCCCCCTCGGC
Plcd3	CTGTGCGGTGGCTGGAAAA	ATCCTCCGTCAGGCCCATC
Pparα	TGCAGCCTCAGCCAAGTTGAA	TTCCCGAACTTGACCAGCCA
Pparγ	CAGGTCAGAGTCGCCCCG	GCGACCCAATACCCCAGC
Serpine1	GGCACAACCCGACAGAGAC	TCGTCCCAAATGAAGGCGTC
Slc20a2	GAGGGAACGAGAAGCCAGAAAG	TCAGTGGTACACCGCATTCC
Slc25a23	CGGAACCAAGATGGTCACATAGA	CATCGCGGTCCATGCTGT
Slc39a2	TGCTGCTTCTAGTTACGTGGAGT	GCCAGCGACTCCAAAAGGAA
Tgfa	CTGTGTGCTGATCCACTGCT	GGCCAAATTCCTCCTCTGGG
Housekeeper genes:
Ywhaz	AAACAGCTTTCGATGAAGCCA	CATCTCCTTGGGTATCCGATGT
Ubc	GGAGTCGCCCGAGGTCA	AAAGATCTGCATCGTCTCTCTCAC
Gapdh	GGGAAATGAGAGAGGCCCAG	GAACAGGGAGGAGCAGAGAG

Western blotting

Frozen placental labyrinthine zones were homogenised using RIPA buffer with protease inhibitor and protein concentrations quantified by BCA assay. Samples were diluted to 2.5 µg/µl in SDS loading gel. Proteins were separated on hand-cast polyacrylamide gels (10% resolving gel with 4% stacking gel) by SDS-PAGE and transferred onto nitrocellulose membranes. Total protein loading was confirmed by Ponceau S staining. Membranes were incubated overnight with primary antibodies against PPARα (ab8934, Abcam) or PPARγ (sc-7273, Santa Cruz), each diluted 1:1,000 in tris buffered saline containing tween-20 (TBST). This was followed by incubation with HRP-conjugated secondary antibodies (NA934 & NA931, Amersham ECL, Cyvita, UK) diluted 1:10,000. Protein bands were visualised using enhanced chemiluminescence (ECL) and imaged using on an iBright Imaging System (Thermo Fisher). Band intensities were quantified using FIJI (ImageJ), normalised to Ponceau staining, and expressed relative to the normoxic control group mean in females.

Histology

Paraffin-embedded placentas, fixed in 4% paraformaldehyde, were sectioned at 7 μm for histological analysis. Mid-line sections of four samples per group, each from different litters, two male and two female, were stained with 1% Alizarin Red (Sigma Aldrich) to detect calcium deposition [46] or with Masson’s Trichrome (Chromaview™ Thermo Scientific) to assess fibrosis, following the manufacturer’s protocol. Sections were scanned using a Hamamatsu Nanozoomer at 40x magnification. Sections were visually ranked from highest to lowest for the extent of calcium deposition, and the mean rank across the sections for each placenta was calculated.

Immunohistochemistry staining was performed on sections of four placentas per group from individual litters to detect lipid droplets via PLIN2/Perilipin-2 [47]. Antigen retrieval was carried out using citrate buffer, followed by blocking with BSA and goat serum. Sections were incubated with anti-PLIN2 primary antibody (Abcam; ab108323; 1:1,000 in PBST), followed by goat anti-rabbit biotin-conjugated secondary antibody (Abcam; ab6720; 1:10,000 in PBST), and then streptavidin-HRP (Rockland, 1:250 dilution in PBST). Staining was developed with 3,3′-diaminobenzidine (DAB, Abcam) and counterstained with nuclear fast red (Sigma Aldrich). Slides were imaged at 40x using the Nanozoomer.

Statistical analysis

Statistical analyses were conducted using GraphPad Prism v9.0 and RStudio V1.4.1717 (R version 4.1.0). Data were stratified by fetal sex where possible. For fetal and placental weights, and placental transport assay, estimated marginal means were calculated from a linear mixed-effects model with litter as a random effect, followed by two-way ANOVA for treatment and sex (using the lmer and emmeans packages in R). For qPCR and western blot analyses, group comparisons among normoxia (N), moderate hypoxia (13% hypoxia) and severe hypoxia (10% hypoxia) groups were performed using a two-way ANOVA with Tukey’s post hoc test (factors: treatment and sex). Gene expression levels from qPCR were correlated with fetal weight within each treatment group (not stratified by sex) using linear modelling in R. Alizarin Red staining quantification (not split by sex due to sample size limitations) were analysed using Student’s t-test. A p-value < 0.05 was considered statistically significant.

Results

Maternal inhalation hypoxia reduces fetal weight similarly in both sexes

Maternal inhalation hypoxia (days 14–19; term day 20; Fig. 1A) reduced fetal weight in a dose-dependent manner, with similar effects in both sexes. Compared to normoxia, fetal weights were decreased by 5% in females and 7% in males under 13% hypoxia, and by 21% in both sexes under 10% hypoxia (Fig. 1B). Across conditions, male fetuses were consistently heavier than females, reaching statistical significance under normoxic and 10% hypoxic conditions (Fig. 1B). Placental weight was unaffected by hypoxia in either sex but was significantly higher in males across all groups (Fig. 1C). Placental efficiency, defined as the fetal-to-placental weight ratio, was reduced under both hypoxic conditions in both sexes (13% hypoxia: 7.5% in females, 13% in males; 10% hypoxia: 25.5% in females, 25.5% in males; Fig. 1D). Across all treatments, males exhibited lower placental efficiency than females, with a significant sex difference observed only under 13% hypoxia. Collectively, these findings indicate that while fetal sex does not alter the overall impact of hypoxia on fetal weight, placental efficiency is modulated by oxygen availability in a sex-specific manner.

Fig. 1.

Maternal hypoxia reduces fetal and placental weight and placental nutrient transport in a severity- and sex-dependent manner. A Schematic of experimental design; dams were exposed to normoxia, moderate 13% hypoxia, or severe 10% hypoxia from gestational day 14 to 19. Fetal weight (B) placental weight (C), fetal-to-placental weight ratio (D), and placental transport of radiolabelled analogues of glucose (methyl-d-glucose, MeGlu; E-F) and amino acid (methyl-amino isobutyric acid, MeAIB; G-H) were assessed at day 19 and analyzed by two-way ANOVA, with litter as a random effect variable. Data are presented as estimated marginal means ± SEM. Asterisks indicate differences between hypoxia groups within each sex; hash symbols indicate differences between sexes within each treatment group (P < 0.05, P < 0.01, P < 0.001, P < 0.0001; same thresholds apply for #, ##, ###, ####). N = 31-48 fetuses per sex per group from 13-16 litters

Maternal inhalation hypoxia alters placental glucose and amino acid transport in both sexes

Maternal hypoxia influenced placental transport and fetal accumulation of non-metabolisable glucose (methyl-d-glucose; MeGlu) and amino acid (methyl-amino isobutyric acid; MeAIB) analogues in vivo, with broadly similar effects across sexes. Placental MeGlu clearance was significantly higher under 13% hypoxia compared to 10% hypoxia in both sexes (Fig. 1E). Fetal MeGlu accumulation increased with 13% hypoxia relative to normoxia in females (+ 60%) and was elevated in males compared to both normoxia and 10% hypoxia (+ 83%; Fig. 1F). Across all treatment groups, MeGlu clearance was approximately 11% lower in males than females, although fetal MeGlu accumulation was not significantly affected by sex (Fig. 1E & F).

Placental clearance of MeAIB was significantly reduced by 10% hypoxia in both sexes (females − 48%, males − 38%) and was lower compared to 13% hypoxia (Fig. 1G). However, these reductions did not translate to significant changes in MeAIB accumulation per gram of fetus with either hypoxic severity or sex (Fig. 1G). Notably, across all treatment, fetal MeAIB accumulation was higher in males than females. These findings together indicate that severity dependent effects of maternal inhalation hypoxia are largely consistent between male and female fetuses, with both sexes displaying reduced amino acid transport with severe maternal hypoxia and improved glucose delivery to the fetus in response to moderate maternal hypoxia.

RNAseq analysis identifies hypoxia-sensitive placental genes involved in development, metabolism and PPAR signalling

To investigate gene expression changes underlying placental adaptations to moderate hypoxia, whole-transcriptome RNA sequencing (RNAseq) was performed on micro-dissected labyrinthine zones (specialised in substrate exchange) of placentas from male fetuses exposed to maternal 13% hypoxia. RNAseq identified 21 differentially expressed genes (DEGs; adjusted p < 0.05), including 7 upregulated and 14 downregulated transcripts (Fig. 2A–B). Of these, 20 were protein-coding genes; the remaining gene, 2010204K13Rik, is a long non-coding RNA significantly reduced by 13% hypoxia.

Fig. 2.

Maternal hypoxia alters placental gene expression linked to developmental and metabolic pathways. A Volcano plot showing differential gene expression in placentas from dams exposed to 13% oxygen from D14 to 19 compared to normoxia controls. Log₂(fold change) is plotted against -log₁₀ (P value); significantly differentially expressed genes (DEGs; adjusted P < 0.05, DESeq2) are shown in red. Additional genes selected for qPCR analysis are shown in pink; non-significant genes are in grey. B Heatmap of normalized expression values (log₁₀ counts) for 21 DEGs (padj < 0.05), calculated using the apeglm shrinkage estimator within DESeq2. C Knockout mouse phenotypes [48] for 17 of the 21 DEGs are associated with developmental (e.g., preweaning lethality, abnormal organ development), body composition (e.g., altered weight, lean mass), or metabolic function (e.g., insulin or lipid metabolism) traits based on MGI annotations. D Enrichment analysis of Gene Ontology (GO: Biological Process) and KEGG pathways for the 21 DEGs. GO enrichment was performed using MSigDB [49]; KEGG pathway [50] enrichment was conducted using the ClusterProfiler [51]. Colours denote direction of regulation in response to hypoxia (blue: up-regulated; red: down-regulated). E Gene set enrichment analysis (GSEA) across the full transcriptome (n = 12,782 genes) using GO biological processes. Genes were ranked by adjusted P value and effect size. GSEA was performed using the fgsea R package; enriched pathways were collapsed to representative terms and the top 20 positive/negative enriched terms are shown. Selected terms of interest in bold. F Hypoxia-inducible factor (HIF) binding site analysis. Venn diagram of predicted binding sites for HIF1A, HIF2A (EPAS1), and ARNT, based on in silico motif analysis. Colours denote direction of regulation in response to hypoxia (blue: up-regulated; red: down-regulated). Data are from placentas of male fetuses (n = 4/group from independent litters)

To assess the functional relevance of these DEGs, we cross-referenced them with the International Mouse Phenotyping Consortium (IMPC) database [48]. Six DEGs (Mmp12, Slc20a2, Ndrg1, Ntrk2, Lpl, and Fat4) were associated with perinatal lethality, abnormal birth phenotypes, or reduced neonatal survival in knockout mouse models (Fig. 2C, and Supplementary Table 1, which contains further details of the genes). Additional DEGs also linked to developmental phenotypes, including Gkn1, Mal, Tspan8, Tgfa, and Slc39a2, were associated with impaired organ development and morphological abnormalities. Common phenotypes among adult knockout mice included altered body composition, such as reduced body weight (Fat4, Ntrk2, Tgfa), increased lean mass (Slc20a2, Gkn1, Gkn2, Mmp12, Ndrg1), and reduced obesity susceptibility (Bcat2), as well as disruptions in lipid metabolism (Apoc1, Bcat2, Hmgcs2, Lpl) and insulin handling (SerpinB1a, Gkn1, Bcat2, Slc20a2, Ndrg1, Ntrk2).

Gene ontology (GO) term enrichment (MsigDB) revealed significant enrichment for processes related to cell population proliferation, regulation of lipase activity, and small molecule biosynthesis (Fig. 2D). KEGG pathway analysis identified PPAR signalling, cholesterol metabolism, and branched chain amino acid (valine, leucine, and isoleucine) degradation pathways (Fig. 2D). Together, these findings demonstrate that moderate hypoxia alters the expression of placental genes involved in growth, metabolic regulation, and PPAR signalling; pathways with established roles in fetal development and viability.

Pathway analysis of additional hypoxia-sensitive placental genes implicates calcium signalling, cell adhesion and HIF regulation

To gain further information on the impacts of moderate hypoxia on genes and pathways in the placenta, we conducted gene set enrichment analysis (GSEA) across the entire transcriptome (12,782 genes), ranked by fold change and statistical significance (Fig. 2E). GSEA identified coordinated upregulation of immune-related pathways, calcium signalling, and cell adhesion, while revealing significant downregulation of pathways involved in carbohydrate, lipid (fatty acid and steroid), and vitamin metabolism (Fig. 2E).

Given the central role of hypoxia-inducible factors (HIFs) in oxygen sensing, we investigated whether differentially expressed genes harboured hypoxia response elements (HREs) in their promoter regions. Promoter motif analysis (MoLoTool) revealed that 16 out of 21 DEGs contained HREs specific for either HIF1α or HIF2α/EPAS and HIF1β/ARNT (Fig. 2F).

Taken together, these results reveal that maternal 13% hypoxia induces a coordinated placental transcriptional response involving calcium signalling, cell adhesion, immune activation, and metabolic reprogramming. The enrichment of HIF binding motifs among these genes supports a central role for HIF signalling in mediating these hypoxia-induced adaptations.

Fetal sex partly determines placental gene expression responses to maternal inhalation hypoxia

To assess if the transcriptomic signature identified in 13% hypoxia exposed male placentas is also present in placentas of female fetuses, and to identify if DEGs may be affected with hypoxic severity, genes were selected for assessment by real-time quantitative PCR (qPCR) in both sexes and severities.

Specifically, gene expression of 8 DEGs (Apoc1, Bcat2, Hmgcs2, Lpl, Ndrg1, Ntrk2, Slc20a2, Slc39a2; Fig. 3A), and 6 genes which were close to the adjusted pvalue cut off by RNAseq (Bmp2, Cxcl12, Pdgfb, Plcd3, Serpine1 and Slc25a23; Fig. 3B) was measured in labyrinthine zones from male and female placentas exposed to either 13% or 10% maternal hypoxia or normoxia (see Supplementary Table 2 for further details of these genes).

Fig. 3.

Impact of maternal hypoxia and fetal sex on placental gene expression and association with fetal growth. Expression of 14 genes; 8 DEGs A and 6 genes close to the adjusted pvalue cut off by RNAseq B measured by qPCR in the placental labyrinth zone from normoxic, 13% hypoxia, and 10% hypoxia pregnancies. Expression is shown relative to the normoxic mean (females). Each point represents one placenta per sex each from different litters (n = 4–7/sex/group); bars indicate mean ± SEM. Two-way ANOVA was used to assess effects of hypoxia, sex, and their interaction; overall P values are shown when < 0.05. * indicates significant differences between hypoxia groups within a sex; # indicates significant differences between sexes within a group (*, # = p < 0.05; **, ## = p < 0.01; ***, ### = p < 0.001). C Correlation between placental Ntrk2 expression and fetal weight. Pearson correlation coefficient (R²) is shown; bold font and asterisk indicate statistically significant correlation (p < 0.05)

Several genes displayed similar directional changes in expression at both levels of hypoxia (Fig. 3A-B, p.hyp), including Bcat2, Hmgcs2, Slc20a2, Slc39a2, Ntrk2, Bmp2, Cxcl12, and Serpine1. Notably, placental Ntrk2 was upregulated by hypoxia in a severity dependent manner, consistent with a graded transcriptional response to hypoxia severity. However, both severity and sex-specific effects were identified for some genes. For example, Ndrg1 was significantly upregulated by 13% hypoxia in male placentas only, while Pdgfb expression was reduced in males, but not females, under 10% hypoxia. Additionally, baseline sex differences in gene expression were observed, with Slc25a23 lower in males compared to female placentas, irrespective of treatment. These findings indicate that fetal sex modulates both baseline (under normoxic conditions) and hypoxia-induced expression of specific placental genes. Taken together, these data demonstrate that fetal sex contributes to the heterogeneity of select placental transcriptional responses to maternal hypoxia.

To assess potential functional links to fetal growth, the expression of the 14 validated genes was correlated with fetal weight under normoxia, 13% hypoxia, and 10% hypoxia conditions. A significant inverse correlation was identified between Ntrk2 expression and fetal weight in the 10% hypoxia group, with the highest placental expression observed in the lightest fetuses regardless of sex (Fig. 3C). No significant correlations were detected for other genes or under normoxic or 13% hypoxic conditions. Hence, Ntrk2 may play a mechanistic role in the fetal growth restriction observed under severe (10%) hypoxic conditions.

Maternal inhalation hypoxia alters PPAR signalling and lipid handling

Peroxisome proliferator-activated receptor (PPAR) signalling emerged as one of the top GO terms enriched in placentas exposed to 13% hypoxia (Fig. 2D). PPARs are ligand-activated transcription factors that regulate lipid metabolism and energy homeostasis. Analysis using the PPAR gene database [52] identified that 10 out of 21 DEGs in response to 13% hypoxia contained known (e.g., Hmgcs2, Lpl, Mmp12) or predicted (e.g., Ntrk2, Serpinb1a, Ndrg1) PPAR response elements in their promoter regions (Fig. 4A). qPCR analysis revealed no significant differences in Pparα or Pparγ mRNA expression in the placental labyrinthine zone with either 13% or 10% hypoxia compared to normoxia (Fig. 4B). However, Western blot analysis demonstrated changes in PPAR protein abundance. Specifically, PPARα protein levels were significantly decreased and PPARγ levels increased in placentas from 13% hypoxic pregnancies (Fig. 4C). These effects were greater in male placentas, where PPARα was reduced by 26% and PPARγ increased by 60%, while in females the changes were smaller (PPARα − 9%; PPARγ + 11%) and not statistically significant (Fig. 4C). In 10% hypoxia, PPARα abundance was also reduced, but this effect was only observed in female placentas (− 29%). PPARγ protein levels remained unchanged under 10% hypoxia, regardless of fetal sex.

Fig. 4.

Effect of maternal hypoxia on PPAR signalling and lipid droplet accumulation in the placenta. A Genes with predicted PPAR binding sites [52] among the differentially expressed genes. Each node on the outer circle represents a gene; fill colour indicates direction of change compared to normoxia (blue, up-regulated; red, down-regulated). Genes with experimentally confirmed binding by PPARα and/or PPARγ are connected to purple (PPARα) and/or green (PPARγ) nodes. B mRNA expression of Pparα and Pparγ in the placental labyrinth zone, measured by qPCR. C Protein expression of PPARα and PPARγ in the placental labyrinth zone, shown relative to the normoxic mean (females), with representative Western blots and Ponceau S staining as loading control. (B, C) Each data point represents one placenta per sex per litter (n = 6–7/sex/group); bars represent mean ± SEM. Two-way ANOVA was used to test for effects of hypoxia, sex, and their interaction; overall P values are shown when < 0.05. * indicates significant differences between hypoxia groups within a sex; # indicates differences between sexes within a group (*, # = p < 0.05; **, ## = p < 0.01; ***, ### = p < 0.001). D Immunohistochemistry for the lipid droplet marker PLIN2 (brown) in placental sections, counterstained with haematoxylin (blue). Representative images of four replicates shown for each group; scale bars, 50 μm. Inset in each image is at higher magnification with arrows to indicate example lipid droplets

Given the central role of PPARs in lipid metabolism and storage [47, 53, 54], immunostaining for adipophilin (PLIN2), a lipid droplet-associated protein [47] was used to assess lipid deposition. Lipid droplets were detected in the placental labyrinthine zone from normoxic and 13% hypoxic pregnancies, but appeared reduced under 10% hypoxia (Fig. 4D). Together, these findings suggest that maternal hypoxia alters placental PPAR signalling and lipid storage in a severity dependent manner.

Maternal inhalation hypoxia disrupts calcium handling and extracellular matrix (ECM) remodelling in the placenta

Gene set enrichment analyses identified ‘calcium-mediated signalling using intracellular calcium’ as an enriched pathway in placental labyrinthine zone in response to maternal 13% hypoxia (Fig. 2E), with upregulated gene Ntrk2 contributing to this enrichment. Closer examination also revealed that three DEGs; Lpl, Mmp12 and Fat4, encode proteins with calcium binding domains based on UniProt database annotations [55].

Consistent with these transcriptional changes, histological analysis using Alizarin Red staining demonstrated significantly increased calcium deposition in the placental labyrinthine zone in both 13% and 10% hypoxia-exposed pregnancies. This effect was observed in placentas of both male and female fetuses (Fig. 5A-B).

Fig. 5.

Calcium signalling, deposition, and extracellular matrix (ECM) remodelling in the hypoxic placental labyrinth zone. A Histological staining for calcium and ECM content. Representative images of four placentas per group. Top panels: Alizarin red staining for calcium deposition (red), with lower magnification (left) and magnified labyrinth zone (LZ; right). Bottom panels: Masson’s trichrome staining, showing collagen (blue) and cytoplasm (red). Scale bars: left panels = 500 μm; middle panels = 250 μm; right panels 50 μm. B Semi-quantitative analysis of Alizarin red staining. Each placenta (n = 4/group) was scored by averaging ranks from three blinded slides, with higher rank indicating greater calcium deposition. Bars represent mean ± SEM. * indicates significant differences between groups by one-way ANOVA (data not stratified by sex due to low sample size). C Heatmap of genes encoding ECM-related proteins, including core matrisome and matrisome-associated genes [56]. The top 20 ECM-related proteins by adjusted pvalue are shown. DEGs (genes with padj < 0.05) are underlined. Colour bar to the left of the heatmap denotes functional ECM category for each gene

In parallel, several pathways related to ECM regulation were also enriched under hypoxic conditions. These included the biological process ‘negative regulation of cell adhesion’ (Fig. 2E). Cross-referencing curated ECM gene lists, including both core matrisome and matrisome-associated components [56], with our dataset identified 20 relevant genes, including 4 DEGs; Coch, Mmp12, Serpinb1a and Tgfa (Fig. 5C). Masson’s Trichrome staining revealed increased collagen deposition (blue staining) in the placental labyrinthine zone of pregnancies exposed to both 13% and 10% hypoxia (Fig. 5A; representative images of male placentas). Notably, regions with enhanced collagen deposition frequently overlapped with areas of calcium accumulation, suggesting spatial coupling of these pathological processes.

Together, these findings indicate that maternal inhalation hypoxia induces ECM remodelling and abnormal calcium handling in the placenta, irrespective of fetal sex. These changes, which include increased collagen and calcium deposition in the labyrinthine zone, point to a hypoxia-driven fibrotic response that may impair placental function and fetal development.

Expression of human orthologues of hypoxia-responsive genes in mouse placenta reveals relevance to fetal growth and pregnancy complications

To investigate the potential translational significance of hypoxia-responsive placental genes identified in mice, we examined the expression of human orthologues of the 21 differentially expressed genes from mouse placentas exposed to 13% maternal inhalation hypoxia. Mouse genes were mapped to their human counterparts using Ensembl (Supplementary Table 3), and expression profiles were interrogated in term placentas from 298 human pregnancies using the Placentome RNAseq dataset [45].

First, the Placentome dataset was interrogated to determine if these human orthologues were dysregulated in placentas from pregnancies with preeclampsia (PE) and from small for gestational age (SGA) births. Four human orthologues were significantly dysregulated in placentas from complicated pregnancies (Fig. 6A). Specifically, NTRK2, COCH, and TSPAN8 were differentially expressed in placentas from pregnancies with preeclampsia (PE), while GKN1 was significantly altered in placentas from small for gestational age (SGA) births. Notably, NTRK2 had the highest fold change across the whole placentome dataset.

Fig. 6.

Expression of hypoxia-responsive mouse placental genes in human placenta and their association with birthweight and pregnancy complications. A Volcano plot of human orthologs of the 21 mouse DEGs in placentas from pregnancies complicated by preeclampsia (PE) and/or small for gestational age (SGA). Data from the POPs study placentome dataset [45]. Highest and lowest Log2FC of all genes in the placentome dataset are shown. B Correlation between human orthologs of the 21 mouse hypoxia-responsive differentially expressed genes (DEGs) and birthweight in human placental samples [45]. Only genes with significant correlations (p < 0.05) are shown

To determine if our hypoxia-responsive genes are relevant to human growth outcomes, we performed correlation analysis of placental expression of human orthologues with birth weight. This analysis revealed that five human orthologues were significantly associated with birth weight (Fig. 6B). Of these, SERPINB1, NTRK2, and GKN1 were negatively correlated, suggesting that higher expression is associated with reduced fetal growth, while ACOT2 and PDE9A showed positive correlations with birth weight. These findings support the hypothesis that gene expression changes induced by hypoxia may contribute to growth restriction phenotypes observed in both mice and humans.

Together, these data indicate that genes altered by maternal hypoxia in the mouse placenta are also dysregulated in human pregnancies with adverse outcomes. This cross-species overlap highlights a subset of candidate genes, such as NTRK2 and GKN1 that may play conserved roles in placental adaptation to stress and in regulating fetal growth, offering potential targets for future diagnostic or therapeutic interventions.

Discussion

This study demonstrates that maternal inhalation hypoxia induces severity-dependent alterations in fetal and placental growth, as well as placental glucose and amino acid transport with broadly similar physiological outcomes across fetal sexes. However, transcriptomic and molecular analyses revealed important sex- and severity-dependent differences in placental gene expression and protein abundance, which may underlie distinct adaptive strategies to adverse conditions.

Placental gene expression in response to 13% hypoxia in male fetuses showed enrichment of pathways related to calcium handling, lipid metabolism, PPAR signalling, and ECM remodelling, each of which is tightly regulated during healthy placental development. Despite in vivo up-regulation of glucose transport, glucose metabolism pathways were not significantly enriched among the differentially expressed genes based on adjusted p-value thresholds. However, gene set enrichment analysis revealed negative enrichment for carbohydrate and monosaccharide metabolic processes, suggesting a possible shift toward favouring glucose transport over local metabolism. Several differentially expressed genes (e.g., Ndrg1, Bcat2) have previously been implicated in glucose handling [57], further supporting this adaptive hypothesis. Notably, Ndrg1 was upregulated only under 13% hypoxia, which may support enhanced glucose transport to the fetus. This aligns with prior findings that fetuses lacking Ndrg1 exhibit increased vulnerability to hypoxic stress [58].

Beyond glucose, placental amino acid handling was also impacted. Transcriptomic enrichment of branched-chain amino acid (BCAA) catabolism pathways in male placentas exposed to 13% hypoxia suggests altered nutrient processing. Bcat2, encoding mitochondrial BCAT2, was downregulated. This enzyme generates glutamate from BCAAs [59], a process critical for fetal growth [60]. Decreased BCAA levels and reduced placental uptake have been associated with fetal growth restriction in human and animal models [61–64], implying that hypoxia-induced Bcat2 suppression may impair amino acid supply to the fetus beyond the effects on system A transport (as indicated by placental MeAIB clearance in vivo) alone.

Altered placental lipid metabolism emerged as another central feature of hypoxia adaptation. Transcriptomic analyses showed enrichment of cholesterol metabolism and lipase activity pathways, with upregulation of Lpl under both 13% and 10% hypoxia, suggesting a compensatory mechanism to support fetal fatty acid demand [65]. Hmgcs2, a key regulator of ketogenesis allowing utilisation of lipids as an energy source for the conceptus [66], was down-regulated in a severity-dependent fashion. In line with previous findings describing altered placental lipid metabolism in response to hypoxia [67], our data also suggest that hypoxia disrupts placental capacity to utilize lipids for energy. Analysis of lipid droplet accumulation, by staining for PLIN2, suggests reduced lipid droplet accumulation under 10% hypoxia, which may reflect metabolic impairment. This highlights the need for further study into the lipid composition of the hypoxic placenta.

PPAR signalling, a central regulator of placental lipid handling [47, 68–70] and pregnancy outcome [71–73], was strongly implicated in the response to hypoxia. Promoter analyses identified PPAR binding sites in DEGs including Lpl and Hmgcs2, with corresponding changes in PPARα and PPARγ protein abundance. Hypoxia reduced placental PPARα and increased PPARγ levels in a severity- and sex-specific manner, without affecting transcript levels, which suggests post-transcriptional regulation. These protein changes align with previous reports linking decreased PPARα with developmental problems [53, 74, 75], and are consistent with compromised placental fatty-acid supported respiration seen in hypoxia-exposed pregnancies [23]. The protein changes also align with work linking PPARγ activity to favourable placental adaptations, including increased nutrient transport and vascular development that were seen with 13% hypoxia [24, 54]. Moreover, PPARγ agonists have been shown to partially rescue fetal growth under hypoxic conditions [73], and improve placental outcomes in a rat model of preeclampsia [76] supporting the hypothesis that PPAR modulation may be therapeutically exploitable.

This study also implicates ECM remodelling in the placental response to hypoxia. MMP12, upregulated by 13% hypoxia, plays a role in elastin degradation [77] and uterine vascular remodelling [78]. Elevated Mmp12 expression has also been reported in hypoxic human trophoblasts and rodent models [79], and its deletion impairs fetal viability under hypoxia [78], underscoring its importance in placental adaptation. Additional protease-related changes included down-regulation of Serpinb1a and upregulation of Serpine1, the latter encoding PAI-1, a fibrinolysis inhibitor [80, 81]. These changes were associated with increased fibrosis in the labyrinthine zone under both hypoxic conditions, likely compromising placental function. Notably, Serpine1 was highly induced by 10% hypoxia, and in humans SERPINE1 has been shown to be upregulated in human placenta complicated by FGR and pre-term birth [82]. In addition, SERPINB1 expression correlated negatively with birth weight in human placentas, implicating these Serpins as markers or mediators of placental insufficiency. Fibrosis coincided with calcium deposition in the labyrinthine zone, and proteins with calcium binding domains were identified in differentially expressed genes, including Lpl, Mmp12, and Fat4, raising the possibility that placental calcium mis-handling may act both upstream and downstream of transcriptional changes in hypoxia.

One gene of particular interest is Ntrk2, which was robustly induced (> 10-fold) by 10% hypoxia and negatively correlated with fetal weight. Its human orthologue, NTRK2, is also upregulated in placentas from pregnancies complicated by intrauterine growth restriction and preeclampsia [83, 84], and negatively correlates with birth weight. Ntrk2 encodes the TrkB receptor, a modulator of calcium signalling via phospholipase-C, and mitochondrial function [85, 86]. Notably, prior studies show that TrkB inhibition impairs placental growth and structure [87], while TrkB agonists improve mitochondrial respiration [88]. The functional relevance of TrkB induction in hypoxic mouse placenta, particularly in relation to calcium handling and mitochondrial adaptations, warrants further investigation.

However, a limitation of maternal inhalation hypoxia is the impact of systemic hypoxia on maternal physiology. Although this approach allows modelling of pregnancy at high altitude, most human populations experiencing high altitude pregnancy have at altitude prior to conception and often for many generations; consequently some degree of physiological adaptation is observed [89, 90]. Whilst maternal physiological adaptations that help maintain internal oxygen supply can also occur acutely in response to hypoxia, fetal growth restriction (FGR) remains a common consequence of both acute and chronic hypoxic exposure. In rodent models, numerous studies using maternal inhalation hypoxia (see recent comprehensive review [16] and [91]) have demonstrated reduced fetal weight, and multiple measurements have confirmed that maternal hypoxia induces local hypoxia within the placenta. Thus, maternal inhalation hypoxia represents a useful model of FGR associated with placental hypoxia, and more broadly, a valuable tool for studying the role of oxygen in placental function.

Indeed, previous studies have reported similar transcriptional responses in the human placenta in high-altitude pregnancies, pre-eclampsia (PE), and in vitro trophoblast cultures exposed to hypoxia [12], exemplifying the link between placental hypoxia and pregnancy complications. To further investigate the complex placental response to reduced oxygen availability and its relationship to pregnancy outcomes, gene expression changes in placentas from mice exposed to 13% oxygen were compared with a large dataset of human placental RNAseq samples [45]. This enabled comparison with placentas from human PE and FGR pregnancies, as well as direct correlation with fetal weight. These analyses revealed that similar biological pathways may be affected in hypoxic mouse placentas and in placentas from human PE/FGR pregnancies. However, it is important to note that although whole-animal hypoxic exposure is expected to induce transcriptomic changes in the placenta, this may not fully replicate the effects of impaired uteroplacental blood flow and resulting localized placental hypoxia characteristic of PE and FGR. Nonetheless, the present study highlights the need for further investigation into hypoxia-related placental gene expression changes and their relationship to fetal growth outcomes.

As our analyses identified broadly similar physiological responses to maternal inhalation hypoxia across both fetal sexes, RNAseq analysis was performed in only one sex. Additionally, RNAseq was conducted exclusively on placentas from mice exposed to 13% oxygen. While targeted molecular and histological follow-up analyses were performed across both sexes and hypoxia severities, comprehensive RNAseq analysis across all groups would have been required to fully define sex- and severity-dependent effects. Such an approach would be valuable for elucidating the contribution of sex-specific placental plasticity in response to hypoxia [16, 40], as well as for identifying pathways underlying the biphasic response to hypoxia, for example, adaptive responses observed at 13% oxygen that may fail or be overwhelmed at more severe hypoxia (10% oxygen). This would also clarify whether differences exist between male and female placentas or fetuses in their capacity to adapt to hypoxic stress relative to their growth/survival strategies.

In summary, this study identifies severity- and sex-specific placental responses to maternal inhalation hypoxia that may underpin observed changes in fetal growth and nutrient transport. Key dysregulated pathways include PPAR signalling, lipid and amino acid metabolism, calcium handling, and ECM remodelling. Specific genes, such as Ntrk2/NTRK2 and Serpinb1a/SERPINB1, were altered in both mouse and human placentas and correlated with birth weight, highlighting their potential relevance for translational applications. Although overall growth and transport phenotypes were similar between sexes, distinct molecular signatures support the concept of sex-specific placental plasticity. Continued examination of these mechanisms may yield diagnostic biomarkers and therapeutic targets for hypoxia-related pregnancy complications.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: ANSP, EJS, ALF.

Investigation: EJS, JSH, ANSP, ALF, ORV.

Visualization: EJS.

Supervision: ANSP.

Writing—original draft: EJS, ANSP.

Writing—review & editing: ANSP, ALF, EJS, ORV, JSH.

Funding

BBSRC studentship and in vivo skills (ALF, ASP, JSH).

Loke Centre for Trophoblast Research Next Generation Fellowship (ANSP).

Loke Centre for Trophoblast Research Studentship (EJS, ORV).

Data Availability

All data are available in the main text or the supplementary materials.

Declarations

Competing interests

Authors declare that they have no competing interests.