2.1 Subjects
All subjects in this study were required to sign an informed consent form. According to the Helsinki Declaration, the sample collection and data analysis were performed with the approval of the Ethics Committee of the Demobilized Soldiers Kangning Hospital in Liaoning Province.
Referring to previous research standards, all subjects were from Han residents in Huludao, Liaoning Province, China. They had no special religious beliefs or eating habits. 97 schizophrenia patients and 69 matched healthy controls were recruited. All subjects were aged between 18 and 65 years; BMI was between 18 and 35 kg / m2, and their weight was stable, no significant change in the past three months. The members of patient group were diagnosed as schizophrenia according to the SCID- IV-TR diagnostic manual, and were not accompanied by other types of mental disorders, personality disorders, and mental retardation; the control group was free of any mental disorders, personality disorders, or mental retardation. According to WHO standards,[18] 18.5 ≤ BMI <25 is considered as normal; BMI ≥ 25 is considered as overweight or obesity.
Based on our previous research,[19] patients with the following conditions were excluded: 1) In addition to obesity, there are physical diseases reported in the literature that can affect the gut microbiota, such as hypertension, diabetes, and digestive diseases; 2) in the last 6 months taking drugs that might affect the gut microbiota, such as antibiotics, glucocorticoids and high-dose probiotics; 3) Medical examinations of the digestive tract, such as gastrointestinal, barium meal, etc. in the last 6 months 4) surgery on the digestive tract and biliary tract in the last 5 years; 5) there are obvious changes in eating habits in the past 3 months; 6) there are obvious restrictions on movement due to physical diseases, such as bedridden.
2.2 Clinical evaluation
For patients who met the inclusion and exclusion criteria, a questionnaire survey was conducted on all subjects using a Case Report Form, including: age, gender, ethnicity, occupation, height, weight, previous medical history, medication history, history of surgery, consumptive history of tobacco and alcohol.
2.3 Collection of stool samples
Patients were firstly instructed to urinate before defecation to prevent urine from diluting or contaminating feces. The feces were drained into a clean container or on a clean urine pad. After defecation, the staff opened the sterile sampling bottle, peeled off the feces with a small spoon which on the inner cover of the sampling bottle, and dug the middle part of the feces. Repeatedly dug feces into the sampling bottle until about 2g samples were collected. The cap of the sampling bottle was then screwed tightly and quickly placed in a container containing liquid nitrogen for transportation. And make sure that it was transferred to -80℃ refrigerator and stored frozen within 1 hour. The samples must be placed in a container filled with liquid nitrogen and attended by a special person no matter a short or long-distance transportation.
2.4 16S rRNA amplification of V3-V4 region and Illumina sequencing
Fresh fecal samples were taken from 166 subjects, and all samples were stored in a -80℃ refrigerator until DNA extraction. According to the manufacturer’s instructions, PowerSoil DNA kit (MoBio, USA) was used to extract 200 mg feces per sample for DNA extraction. The 16S rRNA (V3-V4) gene marker was amplified using KAPA HiFi HotStart ReadyMix (KAPA, USA). Each DNA sample of the bacterial 16S rRNA gene was amplified with primers 341F (GGACTACHVGGGTWTCTAAT) and 805R (ACTCCTACGGGAGGCAGCAG). Amplification was performed in triplicate by PCR. The amplicons were analyzed on a 1.5% agarose gel electrophoresis, and a band of a desired size was purified using a QIAquick gel extraction kit (QIAGEN, Germany). The products were sequenced on the Illumina HiSeq 2500 platform and submitted to the second-generation sequencing laboratory of the Beijing Institute of Bioinformatics.
2.5 Statistical analysis and bioinformatics analysis
2.5.1 Processing of sequencing data
Raw sequence data was processed and analyzed with QIIME software (Quantitative Analysis of Microbial Ecology, Version 1.9.1) [20]. Separation, ligation, and quality filtering of the forward and reverse sequencing fragments of each sample. Fragments that contain ambiguous characters in the sequence or that contain more than two nucleotide mismatched primers need to be removed. The "open reference" QIIME protocol and other default parameters of the UCLUST method were used to select operational taxonomic units (OTUs). When performing statistical analysis of biological information, to understand the number of bacteria and genus in a sample sequencing result, it was necessary to perform the classification operation and OTU division on all sequences according to the specified similarity (95%, 97%, or 98%, etc.). This study brought together sequences with at least 97% similarity, and used representative sequences from each cluster to identify bacterial taxa from the Greengenes database that was launched on August 13, 2013.[21] OTUs containing less than 2 sequences or an overall relative abundance of <0.00005 were deleted and no further analysis was performed. Because the sequence number of each sample obtained from the sequencing results was variable, the sequence data of each sample was refined into 10,000 sequences to consider the variation in sequencing depth.
2.5.2 Statistical analysis of clinical data
Statistical analysis was performed using SPSS19.0 software. Gender, tobacco and alcohol consumption of all participants were expressed as a proportion or percentage, and the chi-square test was used for the count data. Measurement data, such as age and body mass index (BMI), conformed to the normal distribution for independent sample t test. The study took P <0.05 as statistically significant.
2.5.3 Statistical Analysis of Sequencing Data
Statistical analysis was performed using R-3.3.1 and metagenomic data statistical analysis software.[22] Participants whose gut microbiota mainly predominantly Prevotella was marked as enterotype P, and whose gut microbiota mainly predominantly Bacteroides was marked as enterotype B. Based on this, the subjects could be divided into four groups: patients with enterotype P group (SCH-P), patients with enterotype B group (SCH-B), controls with enterotype P group (HC-P), and controls with enterotype B Group (HC-B). Independent t-test, Welch t-test, and White nonparametric t-test were used in continuous variables. For categorical variables between groups, Pearson's chi-square test or Fisher's exact test were used based on the validity of the hypothesis.
Visualization of the relationship between samples was performed with a principal coordinate analysis (PCoA) based on an unweighted UniFrac distance matrix, and significant differences in the composition of the microbial community were tested with ANOSIM.[23] All significance tests were two-sided tests, and p <0.05 or adjusted p <0.05 was considered statistically significant.
2.5.4 Gut microbiota abundance
Linear discriminant analysis (LDA) effect size (LEfSe, v1.0) was used to analyze the significant differences in relative abundance of gut microbiota categories related to patients with enterotype P group and controls with enterotype P group.[24] Moreover, the work further compared the differences in metabolic pathways of patients with enterotype P with the body mass index of more than 25kg/m2 (BMI-A group) and less than 25kg/m2 (BMI-N group) . Wilcoxon rank sum test for α = 0.05, and the log value of LDA analysis was set to <2.0.[25]