Effect of intestinal ora on the formation of endometritis of sow

Aberration of birth canal microbiota is one of the most important factors in the etiology of sow endomentritis. Nevertheless, reports about the structure and composition of birth canal microbiota in endomentritis sow and their relationship with intestinal microbes is limited. Therefore, understanding the relationship between birth canal microbiota and intestinal microbiota of the host has become exceedingly crucial. In this study, 4 healthy and 4 endomentritis sows were selected basing on whether the sow had endometritis or not in a farm of China. The microora of their birth canal secretions and fresh feces were analyzed via sequencing the V3 + V4 region of bacterial 16S rDNA gene. The results showed that the signicant difference between endometritis and healthy sows birth canal ora in the composition and abundance. Sow endomentritis was associated with increasing in the relative abundance of Porphyromonas, Clostridium_sensu_stricto_1, Streptococcus, Ezakiella, Fusobacterium, Actinobacillus, Bacteroides, and Prevotella as well as Anaerococcus. On the contrary, the majority of benecial bacteria that belonging to Firmicutes phylum (e.g., Lactobacillus and Enterococcus ) declined in endomentritis sow. The increased relative abundance of Porphyromonas in the vaginal secretions might correlate with the decrease of Lactobacillusin the feces of endometritis sows. Moreover, the experimental result also found that some intestinal bacteria (such as Escherichia-Shigella and Bacteroides) may be bound up with the onset of sow endometritis.


Abstract Background
Aberration of birth canal microbiota is one of the most important factors in the etiology of sow endomentritis. Nevertheless, reports about the structure and composition of birth canal microbiota in endomentritis sow and their relationship with intestinal microbes is limited. Therefore, understanding the relationship between birth canal microbiota and intestinal microbiota of the host has become exceedingly crucial.

Results
In this study, 4 healthy and 4 endomentritis sows were selected basing on whether the sow had endometritis or not in a farm of China. The micro ora of their birth canal secretions and fresh feces were analyzed via sequencing the V3 + V4 region of bacterial 16S rDNA gene. The results showed that the signi cant difference between endometritis and healthy sows birth canal ora in the composition and abundance. Sow endomentritis was associated with increasing in the relative abundance of Porphyromonas, Clostridium_sensu_stricto_1, Streptococcus, Ezakiella, Fusobacterium, Actinobacillus, Bacteroides, and Prevotella as well as Anaerococcus. On the contrary, the majority of bene cial bacteria that belonging to Firmicutes phylum (e.g., Lactobacillus and Enterococcus ) declined in endomentritis sow. The increased relative abundance of Porphyromonas in the vaginal secretions might correlate with the decrease of Lactobacillusin the feces of endometritis sows. Moreover, the experimental result also found that some intestinal bacteria (such as Escherichia-Shigella and Bacteroides) may be bound up with the onset of sow endometritis.

Conclusion
Sow endometritis is closely related to the microbiota of birth canal, and that some intestinal bacteria may promote the onset of endometritis. The above results can supply a theoretical basis to research the pathogenesis of endometritis and the microbiota of sow's birth canal and gut.

Background
Endometritis is a common and frequent reproductive system disease in female domestic animal. It can lead to abnormal estrus, repeated infertility or miscarriage in female animals, which may bring enormous economic losses [1]. Although antibiotics have a certain therapeutic effect on livestock endometritis, that the long-term use of antibiotics has made problems of veterinary drug residues in animal products become increasingly prominent. Therefore, it is imperative to exploit a new therapy of curing endometritis. Many prevenient researches have shown that birth canals ora plays a vital role on the formation of biologic barriers that protect the health of the reproductive tract by the production of substances such as lactic acid, hydrogen peroxide and bacteriocin [2]. Recent researches have shown that the intestinal microbiota is strongly associated with diseases such as metabolic syndrome, in ammatory bowel diseases (IBD) and colorectal cancer [3]. Hence, understanding the gut and the birth canal ora is signi cant for the prevention and treatment of livestock diseases.
In recent years, with the development of high-throughput sequencing technology of 16S rDNA gene, which is possible to analyze the diversity of bacterial community in different ecological niches. Growing evidence suggests that the intestinal micro ora affects many important physiological functions of the host, such as immune system activation, metabolism, epithelial cell proliferation and anti-infection [4].
We found that the risk of endometritis was increased in the sows with constipation in clinical practice. Nevertheless, there are few reports on the association between birth canal secretions and intestinal micro ora of sows. In this research, high-throughput sequencing of 16S rDNA gene was used to analyze the composition and differences in the vaginal secretions and intestinal bacterial communities of the sows in the health and endometritis. It is a new attempt to unveil the effect of intestinal ora on the pathogenesis of endometritis. Samples total DNA were extracted and sequenced on Ion S5 XL platform. After cutting off the barcodes, primers and ltering low-quality reads, chimeras, a total of 1212768 high-quality sequences were acquired from all samples. These high-quality sequences were clustered into 7392 OTUs on the basis of 97% similarity. Each sample contained 75798 reads and 462 OTUs on average (see Additional le 1). In this study, six alpha diversity measures were calculated including observied-species (observied OTUs), Shannon, Simpson, Chao1, ACE and PD-whole-tree (see Additional le 2, Additional le 3,Additional le 4 and Additional le 5).

Analysis of the birth canal microbial community of endometritis sows and healthy sows
Under the condition that the similarity was 97%, a total of 1102 OTUs were observed in samples from HV and EV (Fig. 1A). The microbiota of HV and EV samples shared 219 OTUs, with 874 and 9 OTUs uniquely identi ed from EV and HV samples, respectively (Fig. 1A). Of which, among the 874 OTUs unique to the EV group include 301 bacterial genera, the 9 OTUs unique to the HV group contain 3 bacterial genera, the 219 OTUs shared by EV and HV contain 98 bacteria genera. Different from the analysis of other micro oras, the diversity of HV micro oras was signi cantly lower than that of EV. In the PCoA gure, both the unweighted UniFrac distance ( Fig. 2A) and the weighted UniFrac distance (Fig. 2B) could distinguish the signi cantly difference in microbiota communities between EV and HV samples (Fig. 2). Figure 3 showed the average relative abundance of the top 20 phylum and top 30 genus bacteria in the EV and HV samples. At the level of phylum, ve predominant phyla were identi ed in the bacterial communities of EV and HV samples. On average, the relative abundance of these bacteria is over 1%. Firmicutes (41.26%) was the most predominant phyla in EV samples, followed by Proteobacteria (30.47%), Bacteroidetes (17.78%) and Actinobacteria (5.48%) (Fig. 3A). The most dominant bacteria in HV samples were Firmicutes (74.36%) and Proteobacteria (24.68%) (Fig. 3A). The relative abundance of Firmicutes in HV was signi cantly higher than that of in EV (P < 0.05).Compared with HV samples,the relative abundance of Bacteroidetes, Actinobacteria and Fusobacteria in EV samples were signi cantly increased (P < 0.05) (see Additional le 6).
The relationship between the birth canal and intestinal ora of endometritis sows A total of 1115 OTUs were observed in samples from EV and EF (Fig. 1B). The microbiota of EV and EF samples shared 413 OTUs, with 680 and 22 OTUs uniquely identi ed from EV and EF samples, respectively (Fig. 1B). On the PCoA plot, the unweighted UniFrac distance ( Fig. 2A) could completely separate the EV samples from EF samples. However, the weighted UniFrac distance does not completely separate the EV sample from the EF sample (Fig. 2B). This indicates that, to some extent, there is a certain similarity between the birth canal and intestinal ora of endometritis sows. At the levels of phylum, both Proteobacteria and Firmicutes have high relative abundance in EV and EF samples (see Additional le 7).

Difference of intestinal community of healthy sows and endomentritis sows
In the study, a total of 1069 OTUs were observed in samples from HF and EF (Fig. 1C). The microbiota of HF and EF samples shared 251 OTUs, with 634 and 184 OTUs uniquely identi ed from HF and EF samples, respectively. The community richness index (Chao1 and ACE) and community diversity index (Shannon) were extremely signi cant higher in HF samples than those of EF samples (P < 0.01) (see Additional le 4), only the Simpson Index difference is not signi cant (P < 0.05), indicating that both community richness and community diversity were dramatically higher in HF samples than in EF samples. Next, the unweighted UniFrac distance showed remarkable segregations of microbiota between HF samples and EF samples ( Fig. 2A). Similar discrimination was also observed, via weighted UniFrac distances, in the PCoA (Fig. 2B), suggesting that the Beta diversity of HF samples were also obvious higher than EF samples.
A total of 165 genera were identi ed in the intestines bacterial communities of endometritis sows compared to 211 genera in healthy sows. To identify the signi cant differences in gut bacteria between HF and EF samples, we compared the relative abundance of gut bacteria in HF and EF samples. The result indicated that the relative abundance of six genera (Lactobacillus, Pseudomonas, Psychrobacter, Escherichia-Shigella, Brochothrix and Bacteroides) exhibited obvious increased in the EF samples, although the difference is not signi cant(see Additional le 8). LEfSe analysis also revealed that Brochothrix_thermosphacta, Lactococcus_ra nolactis, Psychrobacter_faecalis, Pseudomonas_fragi, and Psychrobacter_maritimus were the signi cant higher in EF samples compared with HF samples (Fig. 4B ).In brief, compared with HV samples,there were 874(238 + 76 + 394 + 166) unique OTUs in EV samples, of which 242(76 + 166)OTU were shared with EF; while among these 242 OTUs, 166 OTUs were shared with HF, and only 76 OTUs were shared with EF. 76 OTUs were the key to understanding the relationship between the occurrence of sow endometritis and its intestinal ora (Fig. 1D).It was found that there were 5 bacterial phylum, mainly including Firmicutes, Bacteroidetes, Actinobacteria, and 46 bacterial genera in 76 OUTs. These bacterial genera including Bacteroides, Peptoniphilus, Flavobacterium, Anaerococcus, etc, among which Bacteroides was the most abundant genus

Discussion
Differences between the birth canal microbiota of endometritis sows and healthy sows Despite documented evidence indicating that birth canal ora has a key function in the etiology of endometritis, understanding for the structure and composition of the birth canal microbiota in endometritis sows was still limited. The dramatically difference in the relative abundance of microbiota communities in endometritis and healthy sows' birth canal secretion supports the standpoint that endometritis of sows are associated with alterations in the interactions among microorganisms of the birth canal.
At the phylum level, the relative abundance of Firmicutes was the highest in the EV samples. Bacteroidetes, Actinobacteria, and Fusobacteria were only present in the EV samples, when compared with HV samples. At the geneus level, the relative abundance of Clostridium_sensu_stricto_1, Streptococcus, Fusobacterium, Escherichia-Shigella, Actinobacillus and Bacteroides were remarkablely higher in EV samples than those of HV samples (P < 0.05). Growing evidence suggests that Clostridium_sensu_stricto_1, Streptococcus, Fusobacterium, Actinobacillus and Bacteroides are closely correlated with diseases in animal. For instance, Correlation studies showed that the mRNA expression of IL-1β and TNF-α were positively correlated with the enrichments in Clostridium_sensu_stricto_1 in the colon mucosal of sheep, this enrichment eventually leads to in ammation of the colonic epithelium in sheep [5]. A study by Xiaojing Xia et al. reported that the pathogenic protein secreted by Streptococcus can escape host phagocytosis and complement-mediated immune destruction leading to the onset of the body [6]. Moreover, Wang et al. reported that Clostridium_sensu_stricto_1, Fusobacterium and Bacteroides are more abundant in the vagina of endometritis sows compared to healthy sows [7]. Similar results have been obtained in studies of human bacterial vaginosis, compared with healthy women, Bacteroidetes, Actinobacteria and Fusobacteria were more abundant in bacterial vaginosis women [8]. These investigations suggested that the onset of sow endometritis is inextricably linked to the increase relative abundance of these bacteria in the birth canal.
Another conspicuous difference in EV samples was the decreased relative abundance of Firmicutes members, including Lactobacillus and Enterococcus. While, we observed obvious statistically predominance of Lactobacillus in 9 OTUs unique to HV samples. Firmicutes was reported to have been the highest abundant phenotype in birth canals samples of the healthy sows [7], and a few members of this phylum are considered to adjust systemic immune responses [9]. Therefore, these bene cial bacteria may involve in regulating bacteria balance, inhibiting conditional pathogens, and preventing colonization of pathogenic microorganisms. For instance, Lactobacillus are generally studied as probiotic agents, which affect pathogenicity of opportunistic pathogens and host immune regulation [10]. L. sakei releases spherical membrane vesicles (MVs) through its cell wall components by activating host TLR2 signals, thereby enhancing the production of IgA, and then preventing the incursion of pathogenic microorganisms and regulating the composition of intestines microbiota [11]. In this study, Lactobacillus and Enterococcus were found at lower levels in EV samples, which are consistent with the previous studies. These data indicate that under the condition of endometritis, the environment of the birth canal maintains the competition and abundance of different bacteria, and the continuous reduction of these groups may affect the stability of bacterial balance and the immune regulation of the host.

Differences between the intestines microbiota of endometritis and healthy sows
The imbalances of micro ora and abnormal immune responses to intestines bacteria can destroy intestinal homeostasis and host homeostasis [12]. Many research results indicated that the higher the diversity of intestinal ora, the stronger its ability to maintain the balance of intestinal ora [13].In this study, we observed that the diversity of intestinal ora of sows with endometritis were signi cantly reduced. It may be related to the This may be related to the occurrence of endometritis.
In order to better understand the relationship between them, we conducted a comparative analysis of EF and HF samples. The results showed that: Lactobacillus, Psychrobacter, Pseudomonas and Escherichia-Shigella were the most dominated genera in EF samples, compared with HF samples. It is worth noting that there were 184 unique OTUs in EF samples.It is worth noting that there were 184 unique OTUs in EF samples, which were composed of 101 bacterial genera. Bacteroides, Prophyromonas, Pseudomonas, Streptococcus, etc. were the main bacterial genera. Many studies have shown that Pseudomonas, and Psychrobacter are associated with some diseases of the animal. For instance, some scholars have reported that Pseudomonas can be as an oral and tracheal pathogens in premature infants [14]. The members of genus Psychrobacter are considered to be opportunistic pathogens, as they are occasionally isolated from infected animals, as well as from human patients [15]. Speci cally, the pathogenic mechanism of these bacteria needs further study.

Effect of sow birth canals and gastrointestinal ora on endometritis
The birth canal and gastrointestinal tract of mammals are extremely complex ecosystems that play an important role in animal health and disease. In this study, we found that Firmicutes, Proteobacteria and Bacteroidetes are the main bacterial phylum in the birth canals and gastrointestinal ora of endometritis sows, which is similar to the study by Koh [16]. At the genus level, Escherichia-Shigella and Bacteroides were bacterial genus shared in EF and EV samples. Both of their abundance were 0.02% in HV samples, and the abundance in HF samples is 0.50% and 0.63%, respectively. Neis et al. found that members of Escherichia-Shigella play a vital role in amino acid utilization in animals and prefer to live in the weakly alkaline environment [17]. In our study, the relative abundance of Lactobacillus was low in EV samples, creating a weakly alkaline environment, which may provide a favorable condition for the growth and reproduction of Escherichia-Shigella members. Remarkably, the relative abundance of Bacteroides was also high in EV samples. A study by Wang et al. reported that Bacteroides can lead to an endogenous infection when the immune system or intestinal microbiota is dysfunctional [18]. In addition, the high abundance of Bacteroides was found in cows with endometritis also indicated that Bacteroides were highly associated with uterine disease [19]. We speculate about that the intestinal microbiota(Escherichia-Shigella and Bacteroides) of the sow may be affected the balance of the ora of the birth canal, promote the growth and reproduction of opportunistic pathogens, which leads to endometritis. Speci cally, how it affects requires further research to clarify.
In addition, some scholars have found that the rst colonization ora of humans originates from maternal microorganisms [20]. In our previous study, the bacteria that E. coli, Shigella and Clostridium existed in endometritis sows were also the main dominant bacteria in the intestines of a group of diarrhea piglets [21]. More interestingly, the lactobacillus, which is more abundant in the birth canal of healthy sows, has been reported to have the effect of alleviating diarrhea in piglets [22]. The bacteriocin secreted by Lactobacillus can promote the absorption of intestinal uid and reduce the secretion of intestinal uid by activating phosphodiesterase activity and reducing cAMP and cGMP levels [23]. These indicate that the sow's birth canal microbiota may be related to the health of the piglets.
All in all, the imbalance of sow intestinal ora may affect the balance of the birth canal microbiota and lead to endometritis, while the vertical transmission of birth canal microbes will affect the health of piglets. Speci cally, the way of the sow gut microbiota affects its birth canal and piglet microbiota (or the regulatory mechanism of the impact) needs further study to clarify.

Conclusion
In conclusion, our study unveiled differences in birth canals microbiota between endometritis and healthy sows and described the correlation between the birth canal and the gut microbiota of the endometritis sow. The results showed that Clostridium_sensu_stricto_1, Streptococcus, Fusobacterium, Escherichia-Shigella, Actinobacillus and Bacteroides may be related to the occurrence of sow endometritis. Of which, Escherichia-Shigella and Bacteroides may be from the intestinal tract of endometritis sows. Simultaneously, we have also found that a decrease in the abundance of Lactobacillus could lead to a diversity increase of the ora of the birth canal, which the latter has the risk of causing endometritis. These ndings can provide a theoretical basis to research endometritis and the sow's birth canal and gut microbiota, and will be helpful to establish an effective strategy to reduce postpartum disease generating in sows.

Experimental design and sample collection
This research was approved via the Institutional Animal care and use committee of Jiangxi Agricultural University and performed according to its guidelines. According to the health status of the sows (Whether have endometritis), 8 fecal samples (from 4 healthy sows and 4 endometritis sows) and 8 vaginal secretions samples (from 4 healthy sows and 4 endometritis sows) were randomly collected from a farms in Jingdezhen, Jiangxi, China. During collecting samples, these animals were not disturbed. And there were no clinical abnormalities such as constipation and diarrhea in the intestines of sows with endometritis in this study. Fecal samples and vaginal secretions samples were kept at 4℃ and transported to the laboratory, and then stored at -80℃ until DNA extraction were performed.

DNA extraction and sequencing
Total bacterial genomic DNA from samples was extracted using CTAB/SDS method, and stored at -80 ℃ until further analysis. Sequencing was performed at the Novogen Bioinformatics Technology Co., Ltd, Beijing, China. Brie y, the V3 + V4 region of the bacterial 16S rDNA gene was ampli ed from the total extracted DNA via using the 314F/806R primer set. All PCR reactions were carried out with Phusion® High Fidelity PCR Master Mix (New England Biolabs, USA) with the following programs: initial predenaturation 1 cycle of 94℃ for 3 min, followed by 38 cycles of 94℃ 45 s, 55℃ 60 s, and 72℃ 90 s, and a nal extension step 1 cycle of 72℃ 10 min. Using electrophoresis on 2% agarose gel and GeneJETTM Gel Extraction Kit (Thermo Scienti c) to separate and purify the PCR products. Sequencing libraries were generated using Ion Plus Fragment Library Kit 48 rxns (Thermo sher) following manufacturer's recommendations and were assessed on Qubit@2.0 Fluorometer (Thermo sher), and then sequenced on an Ion S5TM XL platform. 16S rDNA sequencing data was saved in the European Nucleotide Archive (ENA) under the Accession Number ERS3526284-ERS3526299.

Statistical analysis of microbial community
Low-quality partial of the reads were sheared by using Cutadapt ( V1.9.1, http://cutadapt.readthedocs.io/en/stable/ ), then splited the sample reads from the obtained reads according to barcode, and the original reads was obtained by cutting off the initial quality control of barcode and primer sequences. Quality ltering on the original reads were performed under speci c ltering conditions to obtain the high-quality clean reads on the basis of the Cutadapt (V1.9.1) quality controlled process.