Alteration of the fecal microbiota in Chinese patients with Schistosoma japonicum infection

Schistosoma japonicum infection causes pathological injury to the host. Multiple studies have shown that intestinal helminth infection causes dysbiosis for the gut microbial community and impacts host immunology. However, the effect of acute S. japonicum infection on the gut microbiome structure (abundance and diversity) is still unclear. We collected fecal samples from healthy and infected patients from a single hospital in Hunan Province, China. The bacterial community was analyzed using 16S ribosomal RNA gene sequencing of the V4 hypervariable region using the HiSeq platform. Compared with healthy subjects, infected patients exhibited an increase in relative abundance of the TM7 phylum. At the genus level, there were seven differentially abundant genera between groups. The most significant finding was a Bacteroides enterotype in patients with acute schistosomiasis. These results suggest that S. japonicum infection has a significant effect on microbiome composition characterized by a higher abundance of the TM7 phylum and development of a Bacteroides enterotype.


Introduction
Schistosomiasis remains a devastating and highly prevalent neglected tropical disease that is endemic mainly in poor and undeveloped regions [50]. Of the three major pathogenic species causing schistosomiasis, Schistosoma japonicum is responsible for human and animal infections in parts of East and Southeast Asia, primarily China, the Philippines, and Indonesia [58]. Adult S. japonicum worms inhabit the mesenteric veins of the gut and trigger a cellular immune response in the host when eggs are released, which leads to a wide range of clinical manifestations including gut inflammation [25]. This calso affects the gut microbiota [28].
The human gut is colonized by an enormous community of microbes, termed the microbiota [8], which impacts the host immune system. The gut microbiota interacts with interleukin-17 to induce T helper cell differentiation in the lamina propria of the small intestine to drive autoimmune disease [53]; the microbiota also provides resistance to colonization by enteric pathogens [4]. Alterations in the human microbiome have been associated with a range of conditions in the developed world, including diabetes [55], non-alcoholic fatty liver disease (NAFLD) [45], inflammatory bowel disease (IBD), cardiovascular disease, cancer [3], and refractory Clostridium difficile infection [38]. Microbiome changes have been observed in people with parasitic nematode infections, including higher relative abundance of Paraprevotellaceae in patients infected with Trichuris trichiura [34] and an increased abundance of Sphingobacteria, Deltaproteobacteria, and Erysipelotrichia after deworming [44]. Moreover, these microbiome changes have been observed in children with parasitic protozoa infections [23,40,52]. Children [32] and adults [1] infected with S. haematobium had altered gut microbiota, urogenital schistosomiasis, and altered bladder pathologies. In addition, a mouse model of S. mansoni infection showed that depletion of gut bacteria resulted in a reduction in schistosome egg excretion, an alteration in the specific immune response [28], and an alteration in inflammatory response and gut pathology caused by specific bacteria [30].
Studies analyzing the effect of S. japonicum infection on the gut microbiota are lacking. We directly addressed this issue by examining the intestinal microbial community during acute S. japonicum infection, and identifying associations between bacteria and enterotypes with acute schistosomiasis in an epidemic area of China.

Ethics statement
All experiments were approved by the Ethics Committee of the National Institute of Parasitic Diseases, Chinese Center for Disease Control and Prevention (No. 2015-011). All participants were informed of the objectives, procedures and potential risks of the study. Written informed consent forms were personally signed by all adult subjects. The personal information of all the participants has been kept confidential.

Subjects and sample collection
Twenty-six participants living in the same geographical area were included in this cross-sectional study (Table 1). They were local male fishermen without hepatitis B or C infections, who had not received any medical therapy within the previous three months. All of the study participants were >30 years old, and had no other parasitic infection. The study was conducted between September and December 2016.
Eleven patients were initially screened for Schistosoma spp. eggs in feces using the Kato-Katz method [31] and by quantifying levels of indirect hemagglutination antibody (IHA) in the serum. They also had a fever or fatigue accompanied by tenderness in the liver region, and had been in contact with cercariae in water in the previous three months. They had no other helminth eggs (e.g., hookworm, roundworm, whipworm, pinworm, or Taenia) according to microscopic examination, but fecal samples were found to contain Schistosoma spp. eggs. Moreover, molecular examinations were not indicative of any emerging and important protozoa (such as Blastocystis spp., Giardia intestinalis, Entamoeba spp., Cryptosporidium spp., Enterocytozoon bieneusi, and Cyclospora cayetanensis). After sample collection, all infected patients were treated with praziquantel.
The 15 healthy subjects (control) did not display clinical symptoms of schistosome infection and yielded negative laboratory results according to the Kato-Katz method and IHA test, and they had no history of schistosomiasis.
A total of 26 fresh fecal samples (15 controls, 11 patients) were stored in 2.0-mL Eppendorf tubes and frozen at À80°C until DNA extraction.

DNA extraction
DNA was extracted from a frozen aliquot (200 mg) of each fecal sample using a QIAamp DNA Mini Stool Kit (Qiagen, Valencia, CA, USA). DNA concentration was measured by a Qubit 2 (Invitrogen) and its molecular size was estimated by agarose gel electrophoresis. DNA libraries were constructed according to the manufacturer's instructions (Illumina). The V4 region of the 16S rDNA gene was amplified using the 515F forward (5 0 -GTGCCAGCMGCCGCGGTAA -3 0 ) and 806R reverse primers (5 0 -GGACTACHVGGGTWTC-TAAT -3 0 ) (BGI company, China). Sequencing was performed using Illumina HiSeq (Illumina, paired end, 250 bp reads).
Analytical processing and annotation of the 16S rRNA sequences Sequence reads were joined using fastq-join and quality filtered (phred score Q ! 25). Sequences were assembled using the FLASh assembler [36]. Assembled reads were filtered to remove sequences shorter than 200 bp and any sequences containing fewer than 42 N-free 8-mers. Chimera sequences were identified and removed using ChimeraSlayer, with default parameters [16]. As a result, the total dataset comprised approximately 0.9 million reads, and the number of reads of individual samples ranged from 30,387 to 40,304. Taxonomic calls were assigned using the classify.seqs program with the Mothur 16S rRNA gene data processing pipeline [46]. The Ribosomal Database Project Naïve Bayesian Classifier (version 2.5 with training set 9 [9]) (Release9 201203 [10]) was used with a 0.5 confidence level. Reads with <0.5 confidence of classification was considered to be "unclassified" at a given taxonomical level.

Statistical analysis
Microbiota richness was determined using the Chao1 index and Shannon's diversity index. These indices were calculated with QIIME and displayed with R software (version 2.15.3), and the Kruskal-Wallis test was used for comparison between groups. Beta diversity analysis (i.e., differences in samples in terms of bacterial community composition) was visualized using Principal Coordinates Analysis (PCoA) with the "cmdscale" function in R. Samples were clustered according to relative taxa abundance values across all taxa (i.e., read counts normalized by total reads mapped per sample) with the "hclust" function in R to interpret the distance matrix using complete linkage. DESeq2 was used to identify specific taxa that are significantly different between patients and healthy subjects [35] (unpaired differential analysis) in R for abundance testing of differential taxa. DESeq2 was the preferred approach based on its negative binomial statistical design and high performance over a full range of effect sizes, replicate numbers, and library sizes [39]. All DESeq2 input data (read counts) and output results are available in Supplemental Table 1. The phylum level comparison between the healthy participants and the patients was done with the Wilcox rank sum test. Enterotype assignment, encoded as "ET_B" (Bacteroides enriched), "ET_P" (Prevotella enriched), and "ET_F" (Firmicutes enriched), were generated using the classification tool (http://enterotypes.org/) and are independent of de novo clustering. Differences were considered significant if the p-value was <0.05 using the chi-square test. Figures were prepared using GraphPad Prism software, version 6.

Characteristics of the study population
Eleven subjects were diagnosed with acute schistosomiasis based on positive Schistosoma egg and IHA tests; 15 participants were healthy subjects from the same area (controls). All participants were between 39 and 68 years old, and they were divided according to age into the !50-year-old group and the <50-year-old group. They could be divided into three groups based on the serum IHA test results (Table 1).

Microbial communities by host infection status
A total of 939,453 quality-filtered reads were obtained for an average of 34,795 reads per sample. Reads were clustered into 419 unique genera and assigned to 27 bacterial phyla. The Chao1 indexwhich is used to estimate the total number of observed species in each communityindicated that there was no significant difference between the healthy controls and infected subjects (Fig. 1A). The Shannon diversity indexwhich indicates the diversity of species in every samplewas also not significantly different between the assessed groups (Fig. 1B). Principal coordinate analysis (PCoA) of the fecal microbial communities showed strong clustering of the samples by individual rather than infection status, age group, or IHA test result (Figs. 1C-1E). This suggests that the microbial community composition in each subject remained relatively stable regardless of S. japonicum infection status.

Differentially abundant taxa
Proteobacteria and Firmicutes were the dominant phyla in the fecal microbial communities of both groups ( Fig. 2A). The mean relative abundance of the top five phyla in infected subjects and healthy controls based on the Wilcoxon rank   [11,42] and is referred to as TM7 in this study for comparison with the reference) (1.0% vs. 5.9% in infected subjects and healthy controls, respectively, p = 0.003, p-value < 0.01) (Fig. 2B). At the genus level, there were seven differentially abundant genera. Five genera (Comamonas, Psychrobacter, Clostridium, Veillonella, and Butyricimonas) had significantly lower relative abundance in infected patients than in the healthy controls (adjusted p-value < 0.05); meanwhile, two genera (Methylophilus and Turicibacter) had significantly higher relative abundance in infected patients than in the healthy controls (adjusted p-value < 0.05) (Fig. 3A and Table 2).

Detection of the gut microbial enterotypes in healthy subjects and patients
Some reports have indicated that enterotypes provide a new perspective for microbial markers related to certain diseases or specific host traits [54]. Based on online enterotype classification and compared with large-scale projects such as MetaHIT and HMP [2,33], enterotypes were significantly differentially distributed between healthy subjects and acute schistosomiasis patients (chi-square test for independence: n = 26, v 2 = 6.61, p = 0.037), with the patients harboring the unique Bacteroides enterotype. A comparison between healthy subjects and acute schistosomiasis patients was made in Figure 3B.

Discussion
This study is the first description of gut microbiota changes in S. japonicum-infected patients in an epidemic area in China by high-throughput 16S rDNA gene sequencing. Comparing the alpha diversity index using Kruskal-Wallis tests, the Chao1 index and Shannon diversity index, no significant differences were observed between patients and healthy subjects. Although these data were the same as for subjects infected with S. haematobium and controls [1], the relatively small sample size was a limitation of our study, and may have affected the statistical power to identify minor changes in the fecal microbiota following S. japonicum infection. However, our most intriguing finding was that there was a higher proportion of TM7 and a Bacteroides-rich enterotype in patients with acute S. japonicum infection than in healthy controls.
As a unique phylum, TM7, which exhibited an increase in relative abundance in acute S. japonicum infected patients, has not previously been linked to schistosome infection [47]. TM7 is globally distributed and is often associated with human inflammatory mucosal diseases. In particular, TM7 is a kind of oral resident bacteria [14] that is dominant in cases of intra-oral halitosis [48]. Cultivation of a human-associated TM7 phylotype revealed a complete lack of capacity for amino acid biosynthesis; this suggests a potential for immune suppression through the repression of TNF-alpha production in macrophages [26]. Other studies have shown that periodontal disease is a risk factor for human colorectal cancer [41], and have identified orally associated bacteria as biomarkers for cancer. Therefore, compared with our data, we suggest that the increase in the relative abundance of TM7 may be a novel biomarker associated with S. japonicum infection.
In our study, Proteobacteria and Firmicutes were the most altered phyla in response to S. japonicum infection; this is consistent with previous studies [37]. An increasing amount of data identifies Proteobacteria as a possible microbial signature of disease [43]. Proteobacteria are present in various human body sites, including the skin, oral cavity, tongue, vaginal tract, and gut [12], and were found to be correlated with the genesis of endotoxemia and in the development of metabolic disorders [49]. It is possible that alterations in pH, bile flow, ratio of obligate anaerobes to facultative anaerobes, and intestinal hormone levels influence the abundance of Proteobacteria in the feces [27]. Specifically, some researchers consider Methylophilus to be an aerobic, methanol-utilizing bacteria [19] that utilizes a number of organic carbon compounds for growth; additionally, it has traditionally been regarded as one of the most important saccharolytic species and expresses multiple anti-oxidative enzymes [18]. In addition, it has also been linked to the stimulation of TNF production in vitro from peripheral blood mononuclear cells of healthy patients, as well as a higher circulating IgG response in patients affected by  inflammatory bowel disease [5,18]. This might indicate that the abundance of Methylophilus in infected patients in our study is the result of an inflammatory response to acute schistosomiasis. We also found that the relative abundance of Comamonas and Psychrobacter were significantly decreased in infected patients.
Comamonas is an obligate aerobe [51]. Psychrobacter is a kind of cold adapted bacteria that can help to moderate temperature for other bacteria and enzymes [15,17] to maintain a suitable environment for infection. The alterations in the relative abundance of these three genera of Proteobacteria suggest that S. japonicum infection induces a systemic inflammation response, possibly by potentially highly efficient xenobiotic metabolizing species. The relative abundance of Turicibacter was increased in infected patients, which suggests it may be involved in the development of systemic inflammation through alterations in immune cell activation [24]. However, the relative abundance of Clostridium, Butyricimonas and Veillonella were decreased in infected patients. Some studies have reported that decreased levels of Clostridium are correlated with total cholesterol levels [22], which suggests that these taxa may play a role in the infection through effects on lipid metabolism. In addition, the relative abundance of Butyricimonas and Veillonella have been reported to be negatively correlated with the severity of a number of diseases [20]. Thus, the alterations to the relative abundance of these taxa show that the microbiota is correlated with changes in metabolism and immune response during S. japonicum infection.
The identification of gut microbial clusters in the two groups was confirmed in our study. There are at least three gut microbial enterotypes, dominated by Bacteroides, Faecalibacterium (classified as Ruminococcus at the time), or Prevotella [13]. Similarly, we found one cluster of fecal communities that was distinguished by Bacteroides levels in acute schistosomiasis patients. This enterotype has been linked to alterations in nutrient processing, and also corresponds with the presence of genes that code for enzymes such as proteases, hexoaminidases, and galactosidases [29]. Growing evidence supports the notion that Bacteroides activate an infectious response in the intestine [6] and they were significantly more abundant in a mouse model of S. mansoni infection [30]. Based on the altered immune response in acute S. japonicum infection [56] and the data from mouse gut microbial modulation after infection with S. japonicum cercaria [57], future studies should examine whether the shift in enterotypes is associated with the Bacteroides enterotype as a possible biomarker for acute schistosomiasis [21].

Conclusions
In this study, we identified differentially abundant taxa in the gut microbiota of infected subjects and healthy controls, which may be associated with inflammation in patients with acute schistosomiasis. Further studies are required to unveil the functional roles conserved in these significantly altered taxa, to determine the interspecies interactions with the gut microbiota of patients with acute schistosomiasis, to determine the interactions between gut microbiota members and S. japonicum, as well as with host, and to determine the link between systemic inflammation, alterations in fecal microbiota structure, and disease activity.
Acknowledgements. We would like to thank Professor Yongkang He for help with collecting samples at the Hunan Institute of Schistosomiasis Control, China.

Data availability statement
The raw sequencing data from 16S rRNA gene sequencing used in this study are available in the short reads archive (SRA) database (accession number PRJNA625383).

Funding
This study was funded by the National Natural Science Foundation of China (Nos. 81772225 and 81971969 to JC),

Supplementary material
Supplementary material is available at https://www.parasitejournal.org/10.1051/parasite/2020074/olm Table S1. Bacterial taxa associated with healthy subjects and patients at the phylum and genus levels.