Deviations in the Gut Microbiota of Neonates Affected by Maternal Group B Streptococcus Infection

Background: Group B Streptococcus (GBS) infection is the leading cause of septicemia, meningitis, and pneumonia in neonates. Aberrant gut colonization in early life may predispose children to various diseases in adulthood. However, the associations between gut microbial changes and GBS infection is still unclear. Methods: We adopted a new microarray-based technique on neonatal meconium samples collected from 13 participants with mother’s GBS infection/colonization in vaginas and 73 uninfected controls. Results: The composition and diversity of meconium microbiota in GBS group were similar to that of healthy controls. However, we identied several specic taxa that were differentially abundant between the two groups (linear discriminant analysis (LDA) effect size (LEfSe): p<0.05, LDA>2.0). Particularly, the relative abundance of Lactobacillus paracasei was signicantly reduced, indicating a role in GBS infection. Conclusions: Our study presented a series of bacterial species affected by GBS, thus providing novel evidence in support of initial intestinal microbiota dysbiosis in the neonates with mother’s GBS infection/colonization.


Background
Group B Streptococcus (GBS) are β-hemolytic and Gram-positive bacteria, which are recognized as a leading cause of neonatal early-onset sepsis (EOS), meningitis, and pneumonia [1,2]. The mother-to-child vertical transmission is the major GBS infection route in neonatal periods. Previous studies have shown that the prevalence of GBS colonization in vagina during pregnancy is approximately 10-30% [3,4]. And the neonatal morbidity rate for acquiring GBS through birth canal is 60% [5]. The implementation of intrapartum antibiotic prophylaxis (IAP) in pregnant women with GBS colonization is an preventive treatment for reducing the risk of GBS-induced neonatal EOS [6]. However, IAP might also disrupt the balance between microbial members of the gut microbiota [7][8][9].
Recent evidences indicated that the disturbance of gut microbiome has been involved in potential prenatal and early life of infant [10,11]. For example, Cassidy-Bushrow et al. observed that Clostridiaceae, Ruminococcoceae, and Enterococcaceae were signi cantly enriched in infants of GBS positive (GBS+) mothers compared to infants of GBS negative (GBS-) ones [12]. Rosen et al. reported 18 taxa that were found to be signi cantly associated with GBS carriage [13]. Although previous studies provided clues about how GBS alters the vaginal microbiome of pregnant women, the relationship between the gut microbiota of infants and maternal GBS infection remains largely undetermined.
In this study, we adopted a new microarray-based technique [14] to characterize the fecal samples of the neonates in GBS + group as compared with control group. The study aimed to investigate the in uence of maternal GBS infection on the gut microbiome of newborns, with the intention of improving perinatal infant care.

Results
The clinical information in the study A total of 104 neonatal fecal specimens were collected during the study period. Of these, 12 fecal samples were undetectable due to inadequate amount of total DNA after extraction and 6 fecal samples were further excluded due to low content of 16S rDNA after ampli cation. Finally, 86 fecal samples from neonates were analyzed. The owchart of the study was shown in Figure 1.
The clinical characteristics of the 86 neonates were shown in Table 1. There were no differenference between two groups except in the term of gestational age and antibiotics exposure after birth. Comparison of α-and β-diversity between two groups To evaluate the differences in composition of gut microbiota, we performed α-and β-diversity analyses.
Several α-diversity indexes including Chao, Ace, Shannon and Simpson (Fig. 2) indicated no signi cant difference in species richness and diversity between GBS + and control groups. As to β-diversity, as βdiversity indicators were applied to estimate the dissimilarity between samples. PCoA plots based on weighted Unifrac distance and Bray Curtis distance showed that the controls clustered more tightly than the infants in GBS + group (Fig. 3), indicating similar bacterial compositions in the controls .

Alteration of taxa in the GBS + and control groups
To identify the speci c taxa associated with GBS infection, a comparison of the microbiota between the infants in GBS + and control groups was conducted by the Linear discriminant (LDA) and effect size (LEfSe) approach. A cladogram represented the signi cant structure of the gut microbiota from phylum level to species level (Fig. 4A), which listed a collection of the differential abundant bacteria between two groups. Particularly, the abundance of Staphylococcus lugdunensis, Lactobacillus helveticus, Lactobacillus mudanjiangensis, Staphylococcus paracasel in the infants with GBS + group were reduced as compared to the controls.

Discussion
The imbalance of bacterial communities in infants has a profound impact on host's health, but there is insu cient evidence to suggest the associations between dysbiosis in meconium and bacterial infections such as GBS infection. In this context, identi cation of the factors affecting the morbidity for gestational GBS-related infection is an important issue that needs to be addressed.
In this study, bacterial composition and diversity showed no signi cant differences between infants in GBS + and control groups. However, we found a lower abundance of Staphylococcus and Lactobacillus in the infants with GBS + mothers, which was in line with previous studies. Staphylococcus lugdunensis is a coagulase-negative Staphylococcus [17], which has been implicated as the main pathogen in various infections, including central nervous system infections, urinary tract infections, and systemic infections [17][18][19][20]. A study from Japan reported that the GBS detection was correlated with signi cantly lower probability of coagulase-negative Staphylococcus [21]. Furthermore, Altoparlak et al. reported that the decreased level of Lactobacillus species was associated with detection of GBS infection [15]. And Kubota et al. demonstrated that GBS positive women had lower percentages of Lactobacillus than GBS negative women [21]. It should be noted that certain Lactobacillus such as Lactobacillus paracasei had the capabilities to prevent GBS adherence to vaginal epithelial cells [22], and antimicrobial activity of Lactobacillus against GBS had been documented in vitro [23]. Moreover, this lower Lactobacillus species had been detected in the neonatal EOS recently [24]. Thus, the reduced abundance of above genera might limit the protective role of microbiome so as to increase susceptibility to infection.
Currently, intrapartum antibiotics prophylaxis (IAP) implementation is considered as the most effective measurement for reducing the risk of GBS-induced neonatal EOS [6]. However, previous studies indicated that IAP could impact not only infant meconium microbiota at the time of birth [25] but also intestinal microbiota during the rst 3 months of life, thus increasing the prevalence of antibiotic resisitance genes [26]. In the present study, although there were a higher proportion of IAP in GBS + group than control group (84.6% vs 60.3%), there were no difference in the aspect of IAP implementation between two groups (Table 1). It was further indicated that this imbalance of bacterial communities within 24 h of life in the neonate in our study might be closely correlated with mother's GBS colonization in vaginal tract, which was in line with previous studies [12].
There were also certain limitations of our results. First, this study was conducted in a single center with a relatively small sample set. Second, potential in uence of nutrition intakes during pregnancy was not taken into consideration. Furthermore, we do not have access to the matched maternal microbiome samples, thus evidence tracing the origin of meconium microbiome is required in further study.

Conclusion
In summary, our ndings add to a growing body of knowledge about the association between GBS infection and neonatal meconium.Our results demonstrated the potential features of microbiota in GBSaffected infants, which may lead to new biomarkers and innovative therapeutic approaches for perinatal infant care.

Study Design and Sample Collection
The Ethics Committee of Shenzhen Luohu Maternity and Child Health Hospital has approved all of the research procedures. Under the procedure approved by the Institutional Review Board (registry number: LL2018006), informed consent was given by the parents of the newborns.
The high-risk neonates admitted to the department of neonatology from May, 2018 to Jul, 2019 were enrolled after receiving informed consent from the parents. Preterm infants with extreme asphyxia (stage III), fetal chromosomal abnormalities, cyanotic congenital cardiac failure, congenital intestinal atresia, gastroschisis, omphalocele, excessive upper gastric intestinal bleeding, or parental permission de ciency/refusal were excluded from the study era.
The fecal samples were collected in 30 50 g within 24 hours after birth by senior nurse and transported immediately to lab on ice and stored at − 80℃ for further studies. The GBS culture from mother's vaginal swab were conduected at 36 week gestational age for term labor or prior to delivery for premature labor.
The clinical information, treatment and lab data of mothers and neonates were extracted from medical records.

Dna Extraction And Labeling
Bacterial DNA from the stool samples was processed following a previously published protocol [14]. In brief, DNA was collected using the Stool DNA Extraction Kit (Halgen, Ltd., Zhongshan, China) and ampli ed in a PCR reaction with standardized primers that covered the 16S rRNA gene V1-V9 regions. The PCR products were explicitly labeled without puri cation for array hybridization.

Microarray Hybridization
Again, previous protocols were followed to perform microarray hybridization [14]. In general, Cy5-and Cy3-labeled sample DNA were combined and loaded into a hybridization tank. After 3.5 h incubation, the slides were manually washed and automatically screened using a dual-channel (Genepix 4000B) scanner to calculation the mean signal strength of Cy5/Cy3 ratio, by which the relative abundance of each bacterial species is given.

Data analysis
Alpha-diversity was measured using default parameters and QIIME tools [15]. Wilcoxon rank-sum test was used to measure the disparities in alpha-diversities between classes. Analyzes of PCoA and NMDS were performed using QIIME modules and visualized with R packages (version 3.5.2). Linear Discriminant Analysis (LDA) Impact Size (LEfSe) tool [16] was adopted to analyze the disparity between classes of bacterial organisms.