Stool donor screening for FMT
Online pre-screening survey
Interested donors were recruited via local and internet media (public web platform, paper, local news) in Xiamen and Guangzhou, China. Donor candidates filled out an online pre-screening form to document the general health information, such as health status, occupation, disease history, medication history, family history, and risk of infectious disease. Empirically, it is suggested that potential FMT donors should be ≥18 and ≤40 years old and have a body mass index (BMI) between >18.5 and <28 kg/m2. Individuals who were active smokers and drinkers were excluded because these factors perturb the gut microbiome[36]. As it is important to limit the time of collection and preparation of faeces to preserve as many anaerobes as possible, donors were requested to live in the local region and defaecate in the appointed clean rooms. We included an additional lifestyle questionnaire survey, which has not been used in other reported methods. Using the data from this survey, we evaluated the temperament, physical activity, and dietary habits of donor candidates, preferring those candidates with the best health status and an enthusiastic and self-disciplined personality.
In-person evaluation and clinical assessment
This evaluation was performed by a trained physician and nurse and supervised by a senior internal medicine specialist. The exclusion criteria in this assessment included atopic, allergic, gastrointestinal, autoimmune, metabolic, neurologic, and psychiatric conditions, as these are known to be associated with an abnormal intestinal microbiome profile. A positive response to a history of chronic pain syndromes, malignancy, receiving growth hormone or receiving an experimental medicine resulted in exclusion from further consideration as a donor. Our screening was inclusive of oral screening as an increasing number of diseases are associated with oral microbiomes[37]. Caries, periodontal diseases, mucosal diseases, and oral cancer were rejected. These assessments were conducted to exclude risk factors for potential microbiome-perturbated conditions and transmissible diseases. Additionally, all donor candidates were also assessed based on the Hamilton Anxiety Rating Scale (HAMA) and Hamilton Depression Rating Scale (HAMD). The inclusion and exclusion criteria are listed in table 1. With the help of these tests, we sought to improve our screening program to incorporate subjective and scaled assessments consistently.
Laboratory screening
Mandatory stool and blood screening of suitable donors was performed four weeks before donation. To ensure the suitability for inclusion as a donor during the test, laboratory screening was repeated regularly. Periodic testing typically required the donor stool and blood to be tested for all potential pathogens of concern at least every eight weeks (Table 1). A new donor sample of faeces was frozen and quarantined until all tests were confirmed to be negative. HIV testing was performed two weeks after the last donation was received. Stool samples of donors were only cleared for use if the test was negative.
Ethical principles and requirements in donor selection and screening
We documented all the necessary health records including general information, online survey, clinical evaluation report, stool and blood test reports, and informed consent and financial incentives records, which conformed to the ethical principles. The issue of informed consent for FMT has become particularly prominent because there is limited information on the potential side effects of FMT, especially related to mental health-associated microbiota, including depression, anxiety, and mood. Studies on the informed consent of patients who accept FMT treatment or subjects participating in human microbiome research highlight the difficulties in identifying and explaining the potential risks and benefits of FMT[38], as well as the vulnerability of recipients[39]. We offered 200 RMB to stool donors for each donation.
Collection and preparation of faeces
On the day of donation, donors filled out a questionnaire that assessed their general health, new gastrointestinal symptoms, stool pattern/frequency, use of antibiotics, travel history, and sexual behaviour. We recommend that a health-related questionnaire be provided at the time of each stool donation. After passing the questionnaire, we gave donors a sterile plastic box that could be opened over a toilet to collect the stool. Cooler boxes and bags could be used so that the samples could be delivered to the specified site within 1 hour after defecation. Stool samples were stored for up to 4 h at 4°C in an air-tight box.
A minimum amount of 100g of fresh stool sample was required for each donation. Fresh stool (25%) should be blended with normal saline (60%) and pharmaceutical grade glycerol (15%), and placed in an automatic stirring and separation machine (Treatgut Biotechnology Co., Ltd.). The process of blending was performed in an anaerobic chamber to prevent obligate anaerobic bacteria from exposure to oxygen. After blending, the stool mixture was aliquoted into individual sterile cryotolerant pots in a biological safety cabinet. The specifications, appearance, quantity, and weights of all the products were checked thoroughly. All samples were immediately frozen at -80°C.
Faecal DNA extraction, library generation, and sequencing
Faecal samples from each donor were collected to obtain the gut microbiota information. Total DNA was extracted from the samples weighing 0.25 g, using the QIAamp Fast DNA Stool Mini Kit (Qiagen, CA, USA). The resulting DNA yield and quality were checked with Qubit® dsDNA HS Assay kit (Thermo Fisher Scientific, MA, USA). For shotgun metagenomics, DNA was fragmented to an average insert size of 350 bp using a Bioruptor NGS sonicator (Diagenode, BE) and further selected using VAHTSTM DNA Clean Beads (Vazyme, NJ, China). Metagenomic libraries were generated using the NEBNext® Ultra™ II DNA Library Prep Kit for Illumina. Library size and quality were assessed using an HS-DNA chip on an Agilent Bioanalyzer 2100 (Agilent Technologies, Palo Alto, CA). Library concentrations were measured by quantitative PCR with KAPA Illumina Library Quantification Kits (KAPA Biosystems, MA, USA) on an ABI 7300 Plus machine (Thermo Fisher Scientific, MA, USA). Each library was sequenced with 150 bp paired-end reagents on Illumina platforms (Illumina, Inc., San Diego, CA). Amplicon metagenomics was adopted to explore the gut bacterial stability of the frequently donating donors over time. Samples were consecutively (with an interval of one week) collected from 16 donors from March 2018 to August 2019. They were then amplified using the 16S V4 primer set: 515F (5’-GTGCCAGCMGCCGCGGTAA-3’) and 806R (5’-GGACTACNVGGGTWTCTAAT-3’), following the methods reported[40], and sequenced on the Illumina MiniSeq with 150 bp paired-end reagents. All the protocols were followed according to the respective manufacturer’s instructions.
Bioinformatic analyses, microbiota evaluation, and stability
For shotgun metagenomics, raw sequencing data and human genome reads were trimmed using KneadData tool (https://huttenhower.sph.harvard.edu/kneaddata), set at default parameters. Taxonomic profiles were generated by MetaPhlan2[41] with default parameters, except for “stat_q: 0.0” to obtain taxa as many as possible. Additional metagenomic data of 78 donors and healthy subjects from four published studies[42-45], were used as references for downstream analyses. A total of 13 harmful bacterial taxa (Campylobacter, Haemophilus, Veillonella, Enterobacter, Listeria, Plesiomonas, Pseudomonas aeruginosa, Klebsiella, Desulfovibrio, Methanobrevibacter mithii, Fusobacterium nucleatum, Mycobacterium tuberculosis, Escherichia coli), 16 beneficial taxa (Bifidobacterium, Lactobacillus acidophilus, Clostridium butyricum, Akkermansia muciniphila, Propionibacterium freudenreichii, Lactobacillus helveticus, Faecalibacterium prausnitzii, Lactobacillus reuteri, Pediococcus acidilactici, Enterococcus hirae, Megamonas, Odoribacter, Citrobacter, Butyricimonas, Alistipes, Bacteroides thetaiotaomicron), and microbial richness were taken into account to build a donor microbial evaluation index (DoMEI). Each beneficial taxon scored one point if its abundance was greater than the median of the abundance of the 78 donors; otherwise, it received a negative point. Similarly, the harmful taxon earned one point when its abundance was less than the median; otherwise, it received a negative point. Likewise, microbial richness scored five points if it was greater than the median; otherwise, it scored zero points. The sum of each feature was tabulated as the DoMEI score of the donor. Second, donors were further assessed in terms of thirteen clinal pathogens: Helicobacter pylori, Clostridium difficile, Vibrio cholerae, Listeria monocytogenes, Campylobacter coli, Staphylococcus aureus, Rotavirus, Norovirus, Mastadenovirus, Atadenovirus, Shigella, Salmonella, and Yersinia). The relative abundance of each pathogen was checked whether more than 0.01%.
For 16S amplicon metagenomics, raw paired-end reads were assembled using FLASH[46]. Operational taxonomic units (OTU) clustering at 97% similarity was performed using USEARCH[47]. Representative sequences were annotated against the Silva database[48] for taxonomic classification by the Ribosomal database project (RDP) Classifier[49]. Each sample was rarefied to 24598 reads. Ecological diversity estimates and distance were calculated using the vegan package[50] and visualised using the ggplot2 package[51] in R 3.5.3[52]. Significance was tested using the Kruskal Wallis test or PERMANOVA with 999 permutations using vegan package[50].