This was a case-control study conducted at Zhongda Hospital Affiliated to Southeast University. All biological samples were collected in 2024. A total of 26 participants were enrolled in this study, including 10 women diagnosed with preeclampsia and 16 participants in the control group, whose ages and gestational weeks at the time of biological sample collection were matched with those of the preeclampsia patients. In addition, we collected the demographic information of all participants from the electronic medical record system.

The inclusion criteria for preeclampsia were as follows: (1) Pregnant women with singleton pregnancy diagnosed with preeclampsia according to the American College of Obstetricians and Gynecologists (ACOG) guidelines [1]. The criteria for exclusion included the following conditions: (2) Use of preeclampsia-related therapeutic drugs before the collection of biological samples; (3) Presence of digestive system diseases or abnormalities, such as diarrhea; (4) Using drugs that may affect the gut microbiota, such as antibiotics, in the past month; (5) Pathogen infection at any site; (6) Complications with other pregnancy-related diseases, such as gestational diabetes mellitus or intrahepatic cholestasis of pregnancy; (7) Unable to cooperate with the collection of biological samples for any reason; (8) Any fetal abnormalities unrelated to preeclampsia; (9) Any history of serious diseases in pregnant women.

To be included in the healthy control group, participants had to meet the criteria of having a single pregnancy and not having preeclampsia. Participants were excluded if they met any of the following criteria: (1) Presence of digestive system diseases or abnormalities, such as diarrhea; (2) Using drugs that may affect the gut microbiota, such as antibiotics, in the past month; (3) Pathogen infection at any site; (4) Complications with other pregnancy-related diseases, such as gestational diabetes mellitus, intrahepatic cholestasis of pregnancy, etc.; (5) Unable to cooperate with the collection of biological samples for any reason; (6) Any fetal abnormalities unrelated to preeclampsia; (7) Any history of serious diseases in pregnant women.

The flow chart (Fig. 1) shows the research design.

Fig. 1.

Flow chart of the research design. HPLC-MS, High Performance Liquid Chromatography-Tandem Mass Spectrometry.

For the detection of red blood cell folate, the samples were pretreated by incubation with mouse serum and solid-phase extraction method. The analytes were detected using the Multiple Reaction Monitoring (MRM) mode of High Performance Liquid Chromatography-Tandem Mass Spectrometry (HPLC-MS/MS) (Waters Masslynx 4.1, Waters Corporation, Milford, MA, USA). The items to be detected include red blood cell folate, 5-methyltetrahydrofolate, 5,10-methylenetetrahydrofolate, 5-formyltetrahydrofolate, and unmetabolized red blood cell folate.

The total DNA of the samples was extracted using a DNA extraction kit (DP302, Tiangen Biotech, Beijing, China), and specific primers with a Barcode label were synthesized according to the full-length primer sequence for PCR amplification. The PCR products were then purified, quantified, and combined to construct the sequencing library (SMRT Bell). Ultimately, high-quality libraries were sequenced on the online platform BMKCloud (www.biocloud.net).

Statistical analyses were conducted with SPSS software version 22.0 (International Business Machines Corporation, Armonk, NY, USA). The normality of continuous variables was assessed using the Shapiro-Wilk test. Normally distributed count data were summarized as the mean $\pm$ standard deviation (SD) and analyzed using Student’s t-test. As for data that did not follow a normal distribution, results were reported as the median with the interquartile range (IQR), and the Mann-Whitney U test was used for analysis. Categorical variables are expressed as frequency (n) and percentage (%). The chi-square test or Fisher’s exact probability test was used for comparisons between groups.

Based on the analysis results of features in 16S rRNA sequencing, taxonomic classification was performed across different levels, generating community structure charts, species clustering heatmaps, phylogenetic trees, and taxonomic dendrograms for each sample at the phylum, class, order, family, genus, and species levels. Alpha diversity analysis was carried out to study the species diversity within a single sample. After counting the abundance-based coverage estimator (ACE), Chao1, Shannon and Simpson indexes of each sample, the researcher drew sample dilution curves and grade abundance curves accordingly and then compared the differences in species diversity among different samples by using Beta diversity analysis. According to the distance matrix, the sample hierarchical clustering (UPGMA) tree, non-metric multidimensional scaling (NMDS) analysis, sample clustering heatmap, and sample principal component analysis (PCA) and principal coordinates analysis (PCoA) plots at the corresponding distances were obtained. At the species taxonomy level, intergroup differences in species abundance were further assessed through significance testing, and biomarkers showing statistically significant variations between groups were identified. Based on 16S rRNA or internal transcribed spacer (ITS) gene sequencing data, the researchers performed functional predictive analyses to infer the gene function or phenotype of the sample and calculated the abundance of the predicted functional gene or phenotype.

Finally, Pearson’s correlation analysis was conducted to assess the association between folate levels and gut microbiota alpha diversity. Statistical significance was defined as a p-value of 0.05 or less.

There were no significant differences in age (p = 0.554), pre-pregnancy body mass index (BMI, p = 0.134), and gestational weeks at biological sample collection (p = 0.776) between the 2 groups of subjects (Table 1), indicating that the 2 groups of subjects were comparable. The gestational weeks at delivery in the preeclampsia group were shorter than those in the control group (p 0.001). There were no significant differences in the proportions of primiparas, spontaneous pregnancies, smoking, and alcohol consumption between the 2 groups (all p $>$ 0.05).

We detected and analyzed the red blood cell folate levels in 10 patients with preeclampsia and 16 healthy pregnant women. The results revealed that the levels of total erythrocyte folate (p 0.001), 5-methyltetrahydrofolate (p = 0.001), and erythrocyte 5,10-methylenetetrahydrofolate (p = 0.002) in red blood cells of patients with preeclampsia were significantly lower than those in the control group (Table 2). There were no significant differences in 5-acyltetrahydrofolate (p = 0.816) and unmetabolized folate (p = 0.241) between the 2 groups.

First, we used multiple indexes to evaluate the alpha diversity of the gut microbiota in preeclampsia patients and the control group, including ACE index (Fig. 2A), Chao1 index (Fig. 2B), Simpson index (Fig. 2C), Shannon index (Fig. 2D), and PD_whole_tree index (Fig. 2E). The results revealed significant differences between the 2 groups in ACE (p = 0.011), Chao1 (p = 0.010), Simpson (p = 0.043), Shannon (p = 0.015), and PD_whole_tree (p = 0.046) indexes, suggesting notable variations in gut microbiota abundance between the preeclampsia and control groups. The rarefaction curves of the 2 groups also showed that as the amount of sequencing data increased, a large number of microorganisms were discovered in the community, meaning that the sequencing volume of each sample was sufficient (Fig. 2F). The rarefaction curve of the diversity index plotted based on the Shannon index also indicated that the data volume of the sequencing results of both groups was large enough, and the number of feature species would not increase with the increase in sequencing volume (Fig. 2G). The rank-abundance curve also demonstrated that the species composition of the sequenced samples was rich and the evenness of the species composition was relatively high (Fig. 2H).

Fig. 2.

Alpha diversity of fecal gut microbiota in preeclampsia vs. control groups. (A) Abundance-based Coverage Estimator (ACE) index. (B) Chao1 index. (C) Simpson index. (D) Shannon index. (E) PD_whole_tree index. (F) Rarefaction curve. (G) Shannon index curve. (H) Rank abundance curve. (I) Species accumulation curve graph (Phylum). (J) Species accumulation curve graph (Genus). PE, preeclampsia; Ctrl, control. Note: *p 0.05.

Finally, we utilized cumulative relative abundance curves to illustrate the relationship between the number of samples and the number of annotated microbial taxa. The red boxes formed the cumulative curve, which reflected the rate of new species appearance under continuous sampling. Within a certain range, as the sample size increased, a sharp upward curve indicated that a large number of new species were discovered in the community. When the curve tended to be stable, the number of species in the environment no longer increased significantly with the increase in samples. A single green box represented the common species in the sampled data set. Together, these green boxes formed the familiar species curve, reflecting the rate of common species occurrence during continuous sampling. The curve showed a downward trend within a specific range as the sample size increased, indicating that the number of newly detected common species gradually decreased. When the curve flattened out, it meant that the common species in this environment were approaching saturation. The results indicated that the sample sizes were adequate across 6 taxonomic levels: phylum (Fig. 2I), class (Supplementary Fig. 1), order (Supplementary Fig. 2), family (Supplementary Fig. 3), genus (Fig. 2J), and species (Supplementary Fig. 4).

PCA, PCoA, and NMDS were applied to examine variations among sample groups. The first principal component accounted for 29.10% of the variation in the PCA results, while the second principal component explained 12.75%. The characteristic differences among the control group were relatively small, while those among the preeclampsia sample group were relatively large (Fig. 3A). In the results of PCoA, the first principal component contributed 9.90%, and the second accounted for 5.98%. Notable differences in species diversity were observed between the 2 groups from Fig. 3B. The NMDS analysis revealed a stress value of 0.1518 for the 2 groups, indicating that the results of the NMDS analysis had a certain degree of reliability (Fig. 3C). Based on the binary (Fig. 3D) and bray (Fig. 3E) distance algorithms respectively, we plotted the sample clustering heatmaps to intuitively reflect the differences between every 2 samples.

Fig. 3.

Beta diversity of fecal gut microbiota in preeclampsia vs. control groups. (A) Principal component analysis (PCA). (B) Principal coordinates analysis (PCoA). (C) Non-metric multi-dimensional scaling (NMDS). (D) Sample clustering heatmaps based on the binary distance algorithms. (E) Sample clustering heatmaps based on the bray distance algorithms. (F) Permutational multivariate analysis of variance (PERMANOVA/Adonis). (G) Analysis of similarities (Anosim). PC1, principal component 1; PC2, principal component 2.

Permutational Multivariate Analysis of Variance (PERMANOVA, Adonis) and Analysis of Similarities (Anosim) were used to evaluate the reliability of the beta diversity between the 2 groups of samples. The PERMANOVA results yielded an R² of 0.068 with a p-value of 0.001 (Fig. 3F), and the Anosim results indicated an R-value of 0.302 and a p-value of 0.002 (Fig. 3G). The results of the beta diversity analysis of the 2 groups of samples have a high degree of reliability.

We analyzed the differential microbiota of the two groups of gut microbiota at the phylum (Fig. 4A), class (Supplementary Fig. 5), order (Supplementary Fig. 6), family (Supplementary Fig. 7), genus (Fig. 4B), and species levels (Supplementary Fig. 8). Supplementary Figs. 9–14 display the specific differentially abundant microbiota at the 6 levels. The analysis revealed significant differences in the composition of the gut microbiota across all taxonomic levels.

Fig. 4.

Heatmap of differentially abundant microbiota. (A) Phylum. (B) Genus.

We used Pearson’s correlation analysis to evaluate the relationship between 5 alpha diversity indices and 5 forms of folate. The results revealed that total erythrocyte folate and erythrocyte 5-methyltetrahydrofolate were significantly correlated with the 5 alpha diversity indices. Erythrocyte 5,10-methylenetetrahydrofolate was not correlated with the Simpson index. However, Erythrocyte 5-acyltetrahydrofolate and Erythrocyte unmetabolized folate were not correlated with any of the 5 alpha diversity indices (Table 3).