The human intestine harbors a complex population of microorganisms that co-evolve with the host by maintaining a symbiotic relationship (Cheng et al., 2019). These microorganisms, including bacteria, fungi, and viruses, are present in numbers far exceeding those of host cells and are collectively referred to as the microbiota. The gut, vaginal, and oral microbiota undergo changes during pregnancy (Koren et al., 2024). The notion that the placenta harbors its own microbiome was first proposed by Aagaard et al. in 2014. The researchers, who studied placental samples after excision of maternal decidua and fetal chorion-amnion to avoid potential contamination, identified multiple phyla within the placental microbiome, namely Firmicutes, Tenericutes, Proteobacteria, Bacteroidetes, and Fusobacteria (Aagaard et al., 2014). Since then, numerous studies have either described the presence of a placental microbiome or argued against its existence (de Goffau et al., 2019; Theis et al., 2019; Williams et al., 2022; Stupak et al., 2023). Summarizing, Zakis et al. conducted a systematic review of studies investigating the existence and composition of the placental microbiome in healthy pregnancies (Zakis et al., 2022). They concluded that the placenta might be characterized by low microbial biomass, and that some of the microbial taxa identified in placental tissue may originate from the maternal oral cavity. However, they also noted that many of the microorganisms reported in placental samples were also detected in negative controls, suggesting that their presence may be attributable to contamination (Zakis et al., 2022). The authors of this review carefully assessed the risk of bias and selected few high-quality studies (Zakis et al., 2022). Among these, for instance, Lager et al. found no evidence of eukaryotic organisms in the placentas of either patients with adverse pregnancy outcomes or healthy controls (Lager et al., 2018). In contrast, Seferovic et al., based on their own research, concluded that the placenta is characterized by low abundance, low biomass, and sparse microbial populations and that its taxonomic composition is distinct from that observed in contamination controls (Seferovic et al., 2019). On the other hand, De Goffau et al. demonstrated that most bacterial deoxyribonucleic acid (DNA) detected in placental samples likely originates from contaminants, a conclusion also supported by Theis et al (de Goffau et al., 2018; Theis et al., 2019). It is important to highlight that studies on the placental microbiome often differ in their sampling techniques, which are critical for accurate interpretation of results. While some researchers, such as Seferovic et al., provide a detailed description of their methodology—including the removal of membranes from the fetal side —many studies lack such methodological transparency (Seferovic et al., 2019; Zakis et al., 2022). Thus, the question of whether the placenta possesses its own microbiome remains unresolved. For many decades, however, the uterus was also believed to be a sterile environment. Several studies have investigated the putative endometrial microbiota using 16S rRNA sequencing, each documenting the presence of a uterine microbiome (Franasiak et al., 2016; Moreno et al., 2016; Winters et al., 2019). Therefore, it is plausible that the placenta also harbors a unique microbiome on its maternal surface, as it is co-formed by the decidua.

A fetus classified as small-for-gestational-age (SGA) is typically defined as having an estimated weight below the 10^th centile for gestational age, while fetal growth restriction (FGR) refers to a condition in which the fetus is unable to reach its genetic growth potential (Lees et al., 2020). In pregnancies complicated by FGR, increased beta diversity, with a particular abundance of bacteria from the genera Bacteroides, Faecalibacterium, and Lachnospira, has been previously described (Tu et al., 2022). However, the maternal gut microbiota profile in pregnancies affected by FGR has rarely been reported. Abnormalities in the maternal cytokine profile have also been observed in women with FGR. Raghupathy et al. found elevated levels of interleukin 8 (IL-8) and reduced levels of IL-13 in the blood of women with FGR. Additionally, levels of IL-8, interferon γ (IFNγ), and tumor necrosis factor α (TNFα) were increased, while IL-10 levels were lower in women with FGR and placental insufficiency (Raghupathy et al., 2012). A dominance of pro-inflammatory cytokines has also been observed in other studies (Bartha et al., 2003; Al-Azemi et al., 2017).

The aim of the study was to analyze the relationship between the gut and placental microbiome and the cytokine profile of pregnant women with appropriate-for-gestational-age (AGA) and SGA fetuses.

This pilot study investigated the maternal microbiome and its relationship with the cytokine profile in pregnancies complicated by SGA. The study was funded by the National Science Centre in Poland (grant no. 2022/06/X/NZ5/01127).

This is the first study to investigate the maternal gut and placental microbiota and its relationship with the cytokine profile in women with SGA and AGA newborns. We observed a lower Chao index in placental samples collected from the maternal side, while a higher Chao index was observed in placental samples collected from the fetal side in SGA pregnancies. Taxonomic analysis identified four genera that were significantly less abundant in placental samples from the maternal side in the SGA group, whereas no taxa remained significant after correction on the fetal side. However, several taxa showed trends towards differing abundance. Stool sample analysis revealed no significant differences in α-diversity between the SGA and AGA groups based on the Shannon and Chao indices. However, β-diversity analysis indicated a significant difference in the first principal coordinate. Additionally, the F/B ratio was significantly lower in the SGA group compared to the AGA group. Veillonella showed a trend towards higher abundance in SGA stool samples, while other taxa were significant only at a lower threshold. Metabolite analysis identified hexanoic acid as the only significantly elevated compound in the stool of women from the SGA group. Proteobacteria (unclassified) and Halomonadaceae correlated with stool metabolites, while IL-6 and TNF-α correlated with specific bacterial groups.

The maternal gut microbiota plays a crucial role in metabolism, immunity, and nutrient absorption. Several studies have identified links between gut microbiota and fetal growth. Published data indicate differences in the gut microbiome between pregnant women with SGA and AGA fetal growth; however, findings are sometimes inconclusive. Tang et al. observed an increase in the microbial population, microbial richness, and community richness, alongside a decrease in microbial diversity in the maternal gut microbiome of women carrying growth-restricted fetuses. PCoA analysis showed both overlap and separation between the AGA and FGR groups. The researchers found that Bacteroides, Akkermansia, Eubacterium coprostanoligenes group, Phascolarctobacterium, Parasutterella, Odoribacter, Lachnospiraceae UCG_010, and Dielma were significantly more abundant in the FGR group, whereas Dialister, Tyzzerella, Collinsella, Roseburia, Intestinibacter, Monoglobus, Clostridium sensu stricto 1, Veillonella, Corynebacterium, Anaerococcus, Staphylococcus, Eubacterium, DTU089, and Eubacterium brachy group were significantly less abundant in the FGR group (Tang et al., 2024). He et al. identified significant differences in the abundance of 20 gut microbial taxa in the gut microbiome of mothers with FGR and AGA newborns. Their findings indicated a positive correlation between the genus Roseomonas and unclassified Propionibacteriaceae, while a negative correlation was observed between the genus Marinisporobacter and Sphingomonas and neonatal birthweight percentile (He et al., 2023). Tu et al. compared the gut microbiota of 14 pregnant women with SGA and 18 with AGA fetal growth using 16S rDNA amplicon sequencing. The authors identified significant differences in β-diversity between the groups. At the genus level, Bacteroides, Faecalibacterium, and Lachnospira were highly abundant in women from the SGA group (Tu et al., 2022). Similarly, according to Xiao et al., enriched bacterial operational taxonomic units (OTUs) of the genus Bacteroides were observed in the maternal gut microbiome of the FGR group (Xiao et al., 2024). A high abundance of Bacteroides from the phylum Bacteroidota was also noted by Tu et al. in the maternal gut microbiome of women with SGA fetal growth (Tu et al., 2022). Bacteroides in the intestines is associated with lipid metabolism and has been linked to maternal dyslipidaemia during pregnancy (Yang et al., 2023). Previous research has suggested that maternal dyslipidaemia may be associated with accelerated placental epigenetic ageing, which in turn can lead to placental insufficiency and pregnancy complications such as preeclampsia, preterm delivery, or SGA infants (Shrestha et al., 2019). Therefore, it is possible that an increased abundance of Bacteroides in the maternal gut contributes to the development and progression of FGR through lipid metabolic pathways.

Additionally, Bacteroides, Faecalibacterium, and Lachnospira are significantly enriched in Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways related to glycometabolism (Tu et al., 2022), suggesting that this specific shift in the gut microbiome may also influence glucose metabolism. Bacteroidetes is a major contributor to lipopolysaccharide (LPS) biosynthesis in the intestines (Tu et al., 2022), and its high abundance may induce inflammation during pregnancy (Wang et al., 2020). Animal studies have shown that maternal LPS exposure in late gestation results in intrauterine FGR in mice (Xu et al., 2005; Zhao et al., 2013). The data on maternal gut microbiota in FGR pathology is, however, varied and sometimes contradictory. Tao et al. investigated a cohort of 35 women with FGR and 35 with AGA pregnancies. At the species level, the abundances of Lactobacillus and Catenibacterium were elevated in the SGA group, whereas the abundances of Ruminococcaceae, Bacteroides uniformis, Mollicutes RF39, and Alistipes onderdonkii were reduced in the gut microbiota of pregnant women carrying FGR fetuses. The researchers also observed that the abundance of Catenibacterium was negatively correlated, while Lachnospiraceae was positively correlated, with neonatal birth weight (Tao et al., 2023).

Altered gut microbiota may activate inflammatory pathways, leading to dysbiosis and increased intestinal permeability. Both bacteria and their metabolites can pass into the maternal circulation and either migrate to the placenta or influence its function (La et al., 2022; Lopez-Tello et al., 2022). Abnormal fetal growth is often associated with placental insufficiency and preeclampsia. Chen et al. conducted a case-control study comparing the gut microbiome of normotensive and preeclamptic women. The authors found reduced bacterial diversity in preeclamptic women, with an enrichment of opportunistic pathogens, particularly Fusobacterium and Veillonella, and a marked depletion of beneficial bacteria, including Faecalibacterium and Akkermansia. To investigate the causative relationship between gut dysbiosis and the development of preeclampsia, the researchers performed fecal microbiota transplantation in an antibiotic-treated mouse model. Animals transplanted with microbiota from preeclamptic donors exhibited lower fetal and placental weights (Chen et al., 2020).

During our study, we observed a positive correlation between plasma IL-6 and TNF-α and Xanthomonadales and Firmicutes. Similar associations have been reported in the literature. Orbe-Orihuela et al., in a study involving children with obesity, found a high relative abundance of Firmicutes correlated with increased levels of TNF-α (Orbe-Orihuela et al., 2018). Bahar-Tokman et al., investigating a population of patients with type 2 diabetes, reported an elevated Firmicutes-to-Bacteroidetes ratio, with Firmicutes positively correlating with the expression of pro-inflammatory cytokine genes (Bahar-Tokman et al., 2022). Although data on the Xanthomonadales are limited, previous study have demonstrated an association between the presence of Xanthomonadales in the vaginal microbiome and preterm prelabor rupture of membranes (Mu et al., 2023). Furthermore, the abundance of Veillonella, which we identified in the gut microbiota of pregnant women from SGA group, has previously been associated with pro-inflammatory dietary patterns, gastric cancer, and autoimmune hepatitis (Wei et al., 2020; Rocha et al., 2023; Nath and Natarajan, 2024).

The metabolomic composition in SGA pregnancies differs from that in AGA pregnancies. To date, metabolomic profiles of FGR/SGA in maternal feces have been examined in only a few studies. Tao et al. investigated the fecal metabolome in women with SGA and AGA fetal growth and identified 23 differential metabolites, including 16 downregulated and 7 upregulated metabolites. Physagulin E, ginkgolide C, and pyrraline were found to be associated with neonatal birth weight. The researchers performed a pathway analysis and found that lipid, amino acid, sphingolipid, fatty acid, and steroid hormone metabolism pathways were enriched in the FGR group (Tao et al., 2023).

Our study is the second to analyze the maternal stool metabolome in SGA fetal growth pregnancies. We compared seven SCFAs and eight AAs in maternal stool samples and found that only one SCFA, hexanoic acid, had a significantly higher concentration in stool samples from women in the SGA group.

SCFAs are produced through the fermentation and decomposition of dietary fiber by anaerobic gut microbiota (Rios-Covian et al., 2020). They play important roles in maintaining gut barrier and modulation of the immune system, among other, by stimulating the secretion of antimicrobial factors and reducing the production of reactive oxygen species and proinflammatory cytokines (Mirmonsef et al., 2012; Bilotta and Cong, 2019; Kim, 2023). They support the development of regulatory T cells (Tregs), which are essential for maintaining immune tolerance. SCFAs can also activate PPAR-γ, a transcription factor involved in suppressing inflammation (Liberato et al., 2012). Additionally, SCFAs bind to free fatty acid receptor 2 (FFAR2) (GPR43) thus regulating the function of the colonic Treg pool (Smith et al., 2013). Notably, SCFAs have been shown to exert strong positive effects on B cells, promoting their activation and differentiation into plasma cells. This function, while beneficial in many contexts, was postulated to contribute in some context to the pathogenesis of autoimmune diseases such as lupus (Kim, 2023).

Microbial metabolic alterations, including changes in hexanoate levels, have also been associated with various pathological conditions. For example, Zhang et al. reported a positive correlation between fecal hexanoate levels and the severity of symptoms in patients with irritable bowel syndrome (IBS), as measured by the IBS Symptom Severity Score (Zhang et al., 2019). De Preter et al. found reduced concentrations of hexanoic acid in patients with Crohn’s disease and ulcerative colitis, and levels of this metabolite negatively correlated with Crohn’s disease activity (De Preter et al., 2015). C5–C8 fatty acids (e.g., pentanoate, hexanoate, heptanoate, and octanoate) were identified as key discriminatory metabolites between healthy individuals and those with inflammatory bowel disease (De Preter et al., 2015). In another study, lower levels of caproic acid were observed in individuals with colorectal polyps compared to controls (Ruiz-Saavedra et al., 2023). Similarly, in people living with HIV, significantly reduced fecal levels of caproic acid were found in those with subclinical atherosclerosis, linked to chronic inflammation, compared to those without (El-Far et al., 2021). In light of these findings, the results of our study appear unexpected, especially considering the commonly reported pro-inflammatory cytokine profile in FGR. However, these observations might contribute to a better understanding of SCFAs’ roles in immune regulation in the future.

Tang et al. investigated gut microbiota metabolism in AGA and FGR pregnancies. The authors identified significant functional changes in metabolites, including methionine, alanine, L-tryptophan, 3-methyl-2-oxovalerate, and ketoleucine. Methionine and alanine were found to be upregulated in FGR cases and were associated with alterations in circulating mRNA expression and microbial abundance (Tang et al., 2024).

The existence of a placental microbiome remains a subject of debate in the literature. As mentioned before, in 2021, Zakis et al. conducted a systematic review of the available published data on the placental microbial composition in healthy pregnancies. The authors identified 24 studies with a low (N=12) to moderate (N=12) risk of bias, which were included in the analysis. A total of 22 studies reported the presence of microorganisms in placental tissues. The most frequently identified genera were Lactobacillus (11 studies), Ureaplasma (7), Fusobacterium (7), Staphylococcus (7), Prevotella (6), and Streptococcus (6) (Zakis et al., 2022). In our study, we investigated the placental microbiome in SGA and AGA groups. We found significantly lower species richness in placental samples collected from the maternal side, whereas species richness was higher in placental samples collected from the fetal side in the SGA group. LINDA-based analyses demonstrated that, at the genus level, the abundances of Ruminiclostridium 6, Lachnospiraceae A2, Corynebacteriaceae, and Faecalibaculum were significantly lower in placental samples from the SGA group.

Zheng et al. investigated placentas collected from low birth weight and normal birth weight full-term neonates born consecutively at Peking Union Medical College Hospital. They utilized 16S ribosomal DNA amplicon high-throughput sequencing to identify bacteria within placental tissues. The researchers observed significantly lower bacterial richness and evenness in the placentas of low-birth-weight newborns compared to those of normal-weight newborns. Lactobacillus was positively associated with birth weight (Zheng et al., 2019). Stupak et al. examined the placental microbiome in pregnancies affected by late FGR. The microbiome was analyzed using LC-ESI-MS/MS mass spectrometry, and bacterial identification was performed through the analysis of bacterial protein sets. Microbiological screening revealed significantly higher relative abundances of pathogenic bacteria (e.g., E. coli, Listeria costaricensis, and Clostridiales bacterium) in placental samples from the FGR group compared to the controls (Stupak et al., 2023). Hu et al. applied 16S sequencing to assess α- and β-diversity and to identify differential taxa features associated with fetal growth. They found a diverse range of flora, predominantly comprising Proteobacteria, Fusobacteria, Firmicutes, and Bacteroidetes. Neither α- nor β-diversity showed significant differences based on fetal growth status. However, at the taxa level, bacteria associated with a hypoxic environment were observed in women with FGR. A significantly higher prevalence of Neisseriaceae, known for its ability to uptake iron-bound host proteins such as hemoglobin, along with an increase in anaerobic bacteria including Desulfovibrio, and hydrogen peroxide (H₂O₂)-producing Bifidobacterium and Lactobacillus, reflected the hypoxic status of the placenta in FGR (Hu et al., 2021).

Our study identified a wide range of bacterial taxa in placental samples from both the maternal and fetal sides, with distinct differences observed between AGA and SGA pregnancies. Although our primary objective was to compare microbial composition, it is important to consider the potential functional relevance of these identified bacterial taxa in the context of fetal development and placental function. Emerging evidence suggests that some of the genera identified in our study, such as Faecalibaculum and Lachnospiraceae, may play roles in regulating immune responses and metabolic processes within the placenta. For example, members of the Lachnospiraceae family have been associated with the production of short-chain fatty acids (SCFAs), which can modulate inflammatory pathways and impact immune tolerance (Bui et al., 2025). Similarly, Faecalibaculum has been linked to metabolic functions that could potentially influence nutrient transfer between the mother and fetus (Pessa-Morikawa et al., 2022).

Our findings contribute to the ongoing debate regarding the presence and nature of a placental microbiome. Several previous studies have reported the detection of microbial DNA in placental tissues, suggesting a low-biomass but distinct microbial community (Aagaard et al., 2014; Parnell et al., 2017). However, others have challenged these results, attributing microbial signals to contamination or technical artifacts (Leiby et al., 2018; de Goffau et al., 2019). In our dataset, we detected a diverse range of bacterial taxa in placental samples, with differences observed between the maternal and fetal sides, and between SGA and AGA groups. Importantly, our analysis included stringent controls and a comparative assessment with negative controls, allowing us to better distinguish true biological signals from potential contaminants. Several taxa commonly associated with maternal mucosal and cutaneous sites were detected in the placental samples, including Lactobacillus, Staphylococcus, Escherichia/Shigella, Ureaplasma, and Gardnerella. The presence of Lactobacillus aligns with findings from Aagard et al., who reported a non-pathogenic, low-biomass microbiota in term placentas dominated by taxa resembling the oral and vaginal microbiomes (Aagaard et al., 2014). Similarly, Gardnerella and Ureaplasma—both associated with bacterial vaginosis and intrauterine infections—have been implicated in preterm birth and adverse pregnancy outcomes (Kenfack-Zanguim et al., 2023; Oh et al., 2024). Their presence, particularly in SGA placentas, may suggest a role in altered intrauterine environments contributing to impaired fetal growth. However, the detection of genera such as Staphylococcus and Escherichia/Shigella—which are frequent contaminants in low-biomass microbiome studies—necessitates cautious interpretation. de Goffau et al. emphasized that many such taxa identified in previous placental studies may arise from reagent or environmental contamination (de Goffau et al., 2019). Our inclusion of AGA controls and relative abundance thresholds helps to mitigate, though not eliminate, this concern. Taken together, these observations support the need for further studies employing metagenomic or culture-based confirmation and rigorous contamination control. The distinct taxonomic signatures observed between SGA and AGA samples, including in both maternal and fetal sides of the placenta, underscore the potential biological relevance of these findings and warrant deeper investigation. Although our results do not resolve the controversy, they support the idea that microbial DNA is detectable in the placenta under certain conditions and may differ by fetal growth status. Further studies with rigorous contamination control and standardized pipelines will be essential to clarify whether these signals represent a resident microbiota or transient microbial DNA exposure.

The strength of our study lies in the homogeneous study group of SGA pregnancies, excluding cases with genetic or major anatomical abnormalities in the fetus or known intrauterine infections causing fetal growth restriction. All potential factors that could influence the maternal microbiome were carefully excluded to minimize possible bias. Only SGA newborns assessed according to the same growth charts were included in the study. Placental samples were collected under sterile conditions and immediately frozen. Stool samples were self-collected by participants following detailed instructions on the collection technique, ensuring no contact with the toilet and using sterile containers. Stool and blood samples were collected within seven days before delivery, allowing for all collections to take place within a short and consistent timeframe. Furthermore, all women delivered between 35 and 41 weeks of gestation, minimizing the impact of potential microbiome changes occurring throughout pregnancy. However, there are several limitations to this study. As a pilot study, the sample size was relatively small. Additionally, we did not collect data on maternal diet and lifestyle. All participants were recruited from the same perinatal center, meaning that potential regional differences in the maternal gut microbiota cannot be entirely ruled out. In our study, we included women with hypertension and pre-eclampsia, conditions that are known to potentially influence the maternal microbiome. However, the prevalence of these conditions in our cohort was low and did not differ significantly between the SGA and AGA groups. Given the small number of affected cases and the lack of significant differences between groups, we did not conduct separate subgroup analyses for these conditions. Nevertheless, we recognize that the presence of hypertension or pre-eclampsia may act as potential confounders in microbiome studies. Despite the limited occurrence of these conditions in our study, it is important to acknowledge their potential impact on the microbiota composition. Future studies with larger sample sizes and more uniform distribution of these conditions would allow for a more detailed assessment of their role in shaping the maternal and placental microbiome in pregnancies complicated by fetal growth restriction.

Our findings reveal significant differences in microbiota composition and immune interactions between AGA and SGA pregnancies. The altered microbial diversity in placental samples, alongside elevated hexanoic acid levels in maternal stool in the SGA group, suggests a potential link between microbiota and fetal growth. Correlations between specific bacterial taxa and cytokines further support the role of microbiome-immune crosstalk in pregnancy outcomes. Despite being a pilot study, these findings provide valuable insights into the microbiota-immune interplay in pregnancy and lay the groundwork for future research on microbiota-based strategies to support fetal health in at-risk pregnancies.