Subject clinical features description
A total of 100 women were recruited, which included 21 healthy women (NW group), 28 healthy pregnant women (NP group), 25 pregnant women with abnormal placental growth (APG group, decreased PIGF MoMs) and 26 pregnant women with pre-eclampsia (PE group). There were no significant differences on age and body mass index (BMI) between the four groups. The gestational age of APG is significantly lower than that of NP and PE group. Setting of APG group is more conducive for us to estimate the gut microbiota changes at earlier gestational weeks and, the diastolic blood pressure (DBP), systolic blood pressure (SBP) and urinary protein concentration (UP) of PE group were significantly higher than those of other three groups (Table 1).
Table 1
༎Summary of subject characteristics.
Groups | NW | NP | APG | PE |
Subjects, n | 21 | 28 | 25 | 26 |
Age range, years (SD) | 31.04 ± 3.73 | 29.03 ± 3.80 | 27.50 ± 0.57 | 29.23 ± 4.85 |
gestational week (SD) | - | 36.50 ± 7.43 | 23.71 ± 5.80* | 36.73 ± 3.44 |
BMI (SD) | 20.49 ± 1.30 | 21.87 ± 2.62 | 25.01 ± 3.55 | 24.27 ± 4.34 |
SBP, mmHg (SD) | 110.95 ± 9.38 | 113.46 ± 8.79 | 115.00 ± 9.37 | 145.61 ± 10.44** |
DBP, mmHg (SD) | 71.38 ± 5.62 | 73.21 ± 7.28 | 70.27 ± 6.99 | 95 ± 7.14** |
Urine protein, scores (SD) | 0 | 0 | 0 | 1.57 ± 0.70** |
#All values are mean ± standard deviation. An asterisk indicated a significant difference (* at p < .05, ** p < .01) between the NP and the labelled group. |
Effects of PE on the Overall Structure of Gut Microbiota.
In order to better analyze the reads range from 74949 to 201741 (about 124761 clean reads per sample) obtained by fecal sample sequencing, ≥ 97% similarity of sequences were clustered into 1 operational taxonomic units (OTUs). This was an efficient and widely used strategy for the investigation of the fecal microbiota.
Principal component analysis (PCoA) based on the unweighted UniFrac index showed no significant difference in bacterial composition between NW and all pregnancies (Fig. 1A, B). As in the three pregnant woman groups, the gut microbiota composition of APG group and PE group and all the abnormal pregnancies (APG group plus PE group) were significantly shifted compared with the NP group (Adonis, P = 0.002, P = 0.015, P = 0.001, respectly) (Fig. 1B). There was no significant difference in the composition of gut microbiota between APG and PE groups even a notable gestational age gap existed (Adonis, P = 0.463). Although the pathophysiological state of pregnant women affected their gut microbiota, the alpha diversity, which included shannon and chao1 index, exhibited only marginally difference (Fig. 1C, D). The total numbers of OTUs (detected in more than 2 samples) in each group did not changed (Fig. 1E). A total of 2332 OTUs were observed as the core shared feature in all the four groups (Fig. 1F), a higher number of detectable unique OTUs were observed in the NW and APG groups (228 unique OTUs in the NW group and 495 in the APG group).
Gut microbiota compositional shifts in APG and PE at phylum level.
As many other studies indicated, the mainland of human gut microbiota were the microbial communities mainly belonged to Firmicutes, Bacteroidetes, Proteobacteria, Actinobacteria and Verrucomicrobia (26). The same results were found in our data (Fig. 2A). The ratio of the phyla Firmicutes and Bacteroidetes (F/B) have been associated with obesity, type 2 diabetes and systemic inflammation, we compared the F/B ratios between the four groups. Although with no significance, an obvious increasing trend of F/B ratios among NP, APG and PE groups were observed (Fig. 2B). Two phyla, TM7 and Proteobacteria, were disturbed. The percentage of TM7 was significantly increased in APG group (Fig. 2C), and the rates of patients with higher content of Proteobacteria were observed in APG and, especially, PE groups (Fig. 2D). Because the APG and NW groups carried more unique OTUs, Venn diagram combined with phylogenetic relationships of OTUs at the phylum level was exhibited. Bacteria belonged to Firmicutes dominanted the core shared microbial community (Fig. 1F and 2E). In the NP group, bacterium belonged to Firmicutes were the main driver which constituted the unique microbial community (Fig. 2E). But in the APG group, bacterium belonged to Bacteroidetes and Actinobacteria were prevailed (Fig. 2E). Just as the APG group, the PE group showed same unique OTUs composition (Fig. 2E).
Disturbance of gut microbiota was associated with the clinical characteristics of PE.
The composition of the gut microbiome at the genus level revealed by OTUs showed some more detail changes on microbes. The composition of the most abundant 10 genera in all subjects was shown in Fig. 3A. Apparently, g_Prevotella was the most common and prevailed genus in the NP group when compared with NW or any patient groups (Fig. 3A). For further identify the relationship between the gut microbiota changes and the development of PE, linear discriminant analysis (LDA) was used to catch the core bacterial differences and identified a total of ten genera with significant differences.
Detailly, the relative abundance of g_Prevotella, g_WAL_1855D, g_1_68, g_Porphyromonas, g_Varibaculum and g_Lactobacillus between NP and PE groups were significant decreased (Fig. 3B). Between the APG and the NP groups, g_Prevotella, g_1_68, g_Porphyromonas,g_Lactobacillus༌g_Mobiluncus༌ g_Campylobacter and g_Peptostreptococcus were decreased significantly (Fig. 3B). When consider the abnormal pregnant women as a whole group, 6 genera with significant difference were detected whereby the significant decreased abundance of g_Prevotella, g_1_68, g_Porphyromonas, g_Lactobacillus, g_Varibaculum and the significant increased abundance of g_Lactococcus (Fig. 3B).
Obviously, g_Lactobacillus showed unanimous correlations with both blood pressure and UP (Fig. 3C). g_1_68, g_Porphyromonas, g_Mobiluncus, and g_Lactococcus were significant correlated with SBP and DBP. Although with no significant LDA difference in PE group compared with NP group (Fig. 3B), the relative abundance of g_Staphylococus, a potential pathogen, was inversely correlated with BMI index in the NP and the PE groups (Fig. 3C).
The loss of g_Lactobacilli is only related to the abnormal clinical indicators of Pre-eclampsia patients.
Over the past decades, Lactobacillus has been regarded as the most common probiotics in human intestine and been world-widely used in food processing, drug development and clinical treatment (27). In our cohort, g_Lactobacillus was the most detectable genus in the four groups with detection rates range from 68–100% (Fig. 4A). The detection rate and the relative abundance of g_Lactobacillus were significantly higher in NP than any other groups (Fig. 4A) implied that the g_Lactobacillus played an important role in maintaining the health of pregnant women. Moreover, we analyzed the correlations between the relative abundance of g_Lactobacillus and the blood pressure and UP in different grouping modes. The abundance of g_Lactobacillus was significant inversely correlated with SBP, DBP and UP in NP and PE groups (Fig. 4b). But in the NW and NP, NP and APG or all samples groups, no statistically significant correlations were found (Fig. 4B). We further found 20 OTUs (species) belonged to g_Lactobacillus and identified 2 OTUs, otu255 and otu784, showing significant differences (both of them with negative LDA score) (Fig. 4C). The relative abundance of otu255 was significant decreased in PE group when compared with NP group and, the otu784 was decreased with significance only between the APG and NP groups (Fig. 4C). Only otu255 was negatively related to DBP in the PE and NP groups with statistical significance (Fig. 4D). These results reveal profound changes in the intestinal microbiome structure of the APG and PE groups, indicating the importance of gut microbiota changes in the development of PE.