Study design
Enrolled patients included those aged > 18 years undergoing maintenance dialysis for more than 3 months at Korea University Anam Hospital, a tertiary hospital in Seoul, South Korea, from Nov to Dec 2018. Exclusion criteria were patients who: (a) were on HD for acute kidney injury; (b) were on HD less than twice per week; (c) had uncontrolled diarrhea or gastrointestinal infection; (d) were treated with oral or intravenous antibiotics within 4 weeks of enrollment; and (e) were actively being treated for cancer, or with immunosuppressive drugs. Among the 23 patients providing informed consent, 1 patient withdrew from the study and 22 patients completed the trial. These patients were treated with sachets [2 g mixtures of probiotics containing 7.0×109 colony forming units (CFU)/g of Bifidobacterium bifidum BGN4 and 2.0×109 CFU/g of Bifidobacterium longum BORI] twice per day for 3 months. B. bifidum BGN4 and B. longum BORI were isolated from the feces of healthy breast-fed infants, and have been used as probiotics since 2000 [12–16]. Its complete genomic sequence was reported to GenBank [17]. B. longum BORI has been shown to shorten the duration of diarrhea in a clinical study of infants infected with rotavirus [12]. A safety assessment of B. bifidum BGN4 and B. longum BORI concerning ammonia production, hemolysis of blood cells, biogenic amine production, antimicrobial susceptibility patterns, antibiotic resistance gene transferability, polymerase chain reaction (PCR) data on antibiotic resistance genes, mucin degradation, genome stability, and the presence of virulence factors has been recently reported [18] Both strains have been considered Generally Recognized As Safe by the United States Food and Drug Administration (US FDA) (GRN813 for B. longum BORI and GRN814 for B. bifidum BGN4; https://www.accessdata.fda.gov/scripts/fdcc/?set=GRASNotices). Blood and fecal samples were obtained at baseline and at 3 and 6 months. The study protocol was approved by the Korea University Medical Center Institutional Review Board (IRB No. 2018AN0346) Written informed consent was provided by all participants.
Laboratory measurements
Routine laboratory measurements including complete blood counts with white blood cell differentials, CRP, albumin, blood urea nitrogen, creatinine, and electrolytes were obtained before the dialysis session. Patients’ demographic factors and current medications were also recorded.
Blood and fecal samples were centrifuged at 2,500×g for 15 min and stored at -80 OC for calprotectin measurements. Samples were analyzed using an enzyme-linked immunosorbent assay (ELISA - LSBio, LS-F9275, Seattle, WA, USA) according to the manufacturer’s instructions.
Flow cytometric detection of Tregs and pro-inflammatory monocyte subsets
Blood samples were collected before the dialysis session. Whole blood (2.5 ml) was collected in a heparinized tube and 200 µl aliquots of heparinized blood were stained for 15 min at room temperature with either anti-human CD4 conjugated with allophycocyanin (CD4-APC), anti-human CD25 conjugated with phycoerythrin (CD25-PE) antibodies for detection of Tregs, anti-human CD14 conjugated with fluorescein isothiocyanate (CD14-FITC), or anti-human CD16 conjugated with allophycocyanin (CD16-APC) antibodies for the detection of monocytes (BD Biosciences, San Jose, CA, USA). After red blood cell lysis and washing, flow cytometric detection of Tregs and the CD14+ CD16+ pro-inflammatory monocyte subset from among 106 cells were performed (FACSCalibur™; BD Biosciences, San Jose, CA, USA) and analyzed by FlowJo v8.5.2. The percentage of cells or the actual number of cells at baseline were compared to the values 3 months after initiation of probiotic supplementation as well as 4 months after the discontinuation of probiotic treatment.
Quantification of Cytokines before and after a Lipopolysaccharide (LPS) challenge
Fold changes of cytokine production upon an LPS challenge were compared before and after 3 months of probiotic supplementation. Quantification of plasma cytokines were performed using human inflammation cytometric bead array kits (BD Biosciences, San Jose, CA, USA) and cytometric bead arrays (Human inflammation kit, BD Biosciences, San Jose, CA, USA) according to the manufacturer's instructions to simultaneously detect levels of interleukin (IL)-1β, IL-6, IL-8, IL-10, IL-12p70, and tumor necrosis factor (TNF)-α.
Measurement of fecal SCFAs
The concentrations of SCFAs in fecal samples were analyzed using high-performance liquid chromatography in the National Instrumentation Center for Environmental Management at Seoul National University (Seoul, Korea). Feces were prepared in normal saline (300 µl per 1 g of feces). The fecal slurry supernatants were obtained through centrifugation (2,500 rpm for 10 min). SCFAs, including acetic, butyric, propionic, isovaleric, and valeric acids were measured before and 3 months after probiotic supplementation.
Microbiome analysis
Sample DNA extraction and next generation sequencing (NGS)
Bacterial genomic DNA was extracted from stool samples using a QIAamp Fast DNA Stool Mini Kit (Qiagen, Hilden, Germany). DNA extraction was performed after homogenization at 30 s for 1 minute using a TissueLyser system (Qiagen) and quantified using a QUBIT 3.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). For NGS, 16S rRNA gene amplifications and index PCRs were performed following the Illumina 16S metagenomic Sequencing Library preparation guide (Illumina, San Diego, CA, USA). The V3 and V4 regions of 16S rRNA were amplified using the following primer pair (forward 5’–TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG–3’, reverse 5’–TCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC–3’). Nextera XT index kits (Illumina, San Diego, California, USA), using 8 cycles, were then used to fragment DNA and add adapter sequences onto the DNA template. Each PCR product was purified using AMPure XP beads (Beckman Coulter, Pasadena, CA, USA). The amplicon library was sequenced by the 2×300 bp paired-end method on a Miseq instrument (Illumina) according to the Illumina protocol.
Bioinformatics Analysis
Raw sequencing data were analyzed via QIIME2 (https://docs.qiime2.org/2019.7/) [19]. The fastq files were imported to QIIME2 using the ‘Casava 1.8 paired-end demultiplexed method' and merged by DADA2 [20]. Filter parameters for trimming and truncating using the DADA2 plugin were 0 and 140 to remove low quality regions of sequences. Feature tables and data generation (‘qiime feature-table summarized’, and ‘qiime feature-table tabulate-seqs’), and phylogenetic tree construction (‘qiime phylogeny-align-to-tree-mafft-fasttree’) were then performed. QIIME 2’s statistical analyses were also performed using the diversity plugin (“core-metrics-phylogenetic”, “alpha-group-significance”, and “beta-group-significance”). To perform taxonomic classification, 'Greengenes 13_8 99% OTUs full-length sequences' were used as 16S rRNA gene databases [21, 22].
Statistical analysis
All the analyses were performed using SPSS software, version 25.0 (IBM Corporation, USA). Data are expressed as medians [interquartile ranges] according to the distribution. Continuous variables (baseline, after 3 months) were compared using the Wilcoxon signed-rank test. A p-value < 0.05 was considered significant.