Mice and vitamin B12 supplementation
Animal experiments were conducted in accordance with guidelines set by the Canadian Council on Animal Care and approved by the Animal Care and Use Committee at the University of Alberta (Edmonton, AB, Canada). All mice were raised and maintained under specific pathogen-free (SPF) or germ-free (GF) conditions. Six to seven-week-old female C3H/HeOuJ mice (Jackson Laboratories, Maine, USA) were randomly housed four or five per cage. Eight-week old female C57BL/6J mice (University of Alberta, AB, Canada) were housed three per cage using the Tecniplast Isocage-P bioexclusion system (Buguggiate, VA, Italy). Eight-week old germ-free female C57BL/6J mice were housed in an isolator (CEP Standard Safety, McHenry, Illinois, USA) with open cages in the University of Alberta Axenic Mouse Research Unit. All mice were allowed to acclimatize for one week with ad libitum access to water and standard chow containing approximately 0.08 mg of cyanocobalamin per kilogram of diet post-autoclaving (2020SX; Envigo-teklad, Indiana, USA). Calculated based on the average consumption of 5 g of diet per mouse per day, the standard chow diet alone contributed approximately 0.4 mg of cyanocobalamin. Mice received filter-sterilized drinking water supplemented with or without B12 in the form of cyanocobalamin (V2876, Sigma-Aldrich, St. Louis, MO, USA) at 40 mg/ml, approximately 100 times the amount in the 2020SX diet. After two weeks of water treatment, survival (SURV) and early-stage pathogen colonization (EPC) experiments were performed in C3H/HeOuJ mice using C. rodentium. A 20% loss of initial body weight was selected as a humane endpoint for mice in the SURV experiment as previously described [16]. For the EPC experiment, two mice from each cage were euthanized (naïve_CON & naïve_CNCbl40) and the remaining mice were challenged with C. rodentium while continuing water treatment (inf_CON & inf_CNCbl40). In addition, different forms of B12 were investigated in conventionalized C57BL/6J mice by supplementing cyanocobalamin and methylcobalamin (Thermo Fisher Scientific, Massachusetts, USA) at 10 mg/ml and 40 mg/ml in drinking water. Germ-free C57BL/6J mice were used to test the direct role B12 at the 40 mg/ml dose to corroborate the change in colonic inflammation markers found in the conventional mice. In all experiments, cages were randomly assigned to treatment groups: control (CON) and B12 supplemented (cyanocobalamin at 10 mg/ml, CNCbl10 and 40 mg/ml, CNCbl40; methylcobalamin at 10 mg/ml, MeCbl10, and 40 mg/ml, MeCbl40) groups accordingly.
Water intake and B12 dose estimation
A pilot study (data not shown) was conducted to determine daily water consumption of mice supplemented with B12 (cyanocobalamin) in drinking water. Cyanocobalamin at 40 mg/ml or control drinking water was provided to mice (n = 15; five mice per cage) ad libitum and water consumption was monitored for a week. Drinking water was changed every two days and water consumption was measured. The water consumed per cage at each timepoint was considered a replicate and the average was used to compare water intake between treatments. B12 supplemented in drinking water did not impact water consumption. Daily water intake was approximately 3 ml per mouse per day.
According to water intake, B12 supplemented in drinking water at 10 mg/ml and 40 mg/ml were estimated to reach a total dose of 30 mg and 120 mg per mouse a day, respectively. The human equivalent dose for the cyano- and methyl- cobalamin drinking water used in these studies are ~5,000 mg/day (CNCbl10 & MeCbl10) and ~25,000 mg/day (CNCbl40 & MeCbl40), which was calculated using the standard human equivalent dose parameters [17]. A mega-dose was used in the present study to push the phenotype and ensure high B12 levels were maintained in the gut to disrupt competition, cross-feeding and microbe-host interactions related to B12 utilization by the gut microbiota. These levels were considered feasible within the human population, where 10,000 mg B12 capsules are available and would be taken as a single bolus. Comparing B12 supplementation between humans and mice using a multiplicity factor of their respective recommended dietary allowance (RDA) shows that a dose of 5,000 - 10,000 tablet per day in an adult human is equal to 2,083 - 4,167X their RDA of 2.4 mg/day, whereas mice receiving a 120 mg/day in the present study is equal to 2,400X their RDA. The RDA of B12 for mice is approximately 0.05 mg/day and is based on a diet containing 10 mg of B12 per kilogram that was deemed adequate for mice [18].
C. rodentium-challenge model
From a glycerol stock, C. rodentium (DBS100) was plated on MacConkey agar (BD Difco, NJ, USA) and a single colony was picked and incubated overnight at 37°C in Luria-Bertani broth (Sigma-Aldrich) with shaking at 200 rpm. Mice were infected with 100 µl of the overnight culture (1 x 109 CFU/ml) by oral gavage. All mice were confirmed to be free of coliforms by plating a fecal sample on MacConkey agar prior to infection. Pathogen load was determined daily in fecal samples and in the GI content at day 5 post-infection by plating serial dilutions of sample homogenates in 1 x PBS on MacConkey agar. Plates were incubated at 37°C overnight and colonies were counted and normalized to sample weight.
Sample collection
Fresh fecal samples were collected daily or every second day post-infection directly in 1 ml of sterile 1 x PBS for plating. All mice were euthanized using carbon dioxide and sampling was done aseptically. Prior to infection, fecal samples were collected from mice for baseline microbiome analysis. Mouse tissues and intestinal content (ileum, cecum and colon) were snap-frozen in liquid nitrogen and stored at -80°C until use.
Intestinal vitamin B12 level measurement
Snap-frozen cecal and colonic digesta samples were weighed and homogenized with two rounds of beating (30 s at 4 m/s with a cooling step on ice) in a proprietary buffer provided by Calgary Laboratory Services: Diagnostic and Scientific Research Centre (Calgary, AB, Canada). Samples were subsequently centrifuged at 10,000 rpm and the supernatant was collected and stored at -20 °C. Vitamin B12 was quantified via electrochemiluminescence using the Roche Diagnostics Vitamin B12 II assay performed on the Roche Diagnostics e602 (Calgary Laboratory Services). This technique measures total vitamin B12 using a ruthenium-label recombinant porcine intrinsic factor as the reporter probe.
Short-chain fatty acids (SCFA) analysis
Snap-frozen cecal content was thawed on ice, weighed (30 mg/sample) and homogenized in 600 ml of 25% phosphoric acid. Samples were centrifuged at 15,000 rpm at 4°C for 10 min and the supernatant was passed through a 0.45 mm syringe filter (Fisher). A 200 ml aliquot of filtered sample was combined with 50 ml of internal standard (23 mmol/ml, isocaproic acid) and analyzed on a Scion 456-GC instrument.
Cecal microbial metatranscriptome analysis
Total RNA was extracted from frozen cecal samples as previously described [19]. Approximately 50 mg of frozen cecal content was added to 0.1 mm glass bead-containing tubes (PowerBead Tubes, Qiagen) prefilled with 300 ml RLT buffer (RNeasy mini kit, Qiagen) supplemented with β-mercaptoethanol (10 ml/ml, Sigma-Aldrich) and 1 ml Trizol (Invitrogen). Cell disruption was accomplished using a FastPrepâ-24 bead-beating machine (MP biomedicals) with two rounds of beating (30 s at 6.5 m/s). After incubating for 5 min at room temperature, samples were centrifuged (1 min, 12,000 x g, 4°C) and supernatants were transferred into tubes containing 300 ml of chloroform, vortexed and incubated for 3 min. After centrifugation (15 min, 12,000 x g, 4°C), the upper aqueous phase was carefully collected and transferred into a new tube containing 1 ml of freshly prepared 70% ethanol solution, mixed by pipetting, and loaded onto a RNeasy spin column (RNeasy mini kit, Qiagen). RNA extraction and on-column DNA digestion (Qiagen) were completed as described by the manufacturer’s protocol. The quality and quantity of RNA were measured using an Agilent Bioanalyzer. Samples with an RNA integrity number (RIN) ³ 7.0 were used to generate metatranscriptome libraries at Génome Québec Innovation Centre (Montréal, QC). Samples were diluted to 100 ng/ml and host rRNA-depletion (NEBNextÒ Human/Mouse/Rat) was conducted. The libraries were sequenced as 100 bp paired-end reads on a NovaSeq 6000 system (Illumina).
Analysis of unfiltered raw data (~35M read average per sample) was completed using the Simple Annotation of Metatranscriptomes by Sequence Analysis 2.0 (SAMSA2) pipeline [20] as follows: PEAR (version 0.9.10) to merge reads, Trimmomatic (version 0.36) to trim low quality reads, ShortMeRNA (version 2.1) to remove rRNA, and DIAMOND (version 2.0.2) to annotate mRNA data to the RefSeq database [21]. The merging step resulted in ~25M merged reads per sample. Bacterial rRNA made up ~9M reads per sample, and ribodepleted reads (mRNA transcripts) led to ~2M annotated reads with ~13M unknown reads per sample. In addition, the SEED Subsystems hierarchical database [22] was used to categorize and compare functional activities of the microbiota. Analysis of annotated reads was completed using the DESeq2 package and visualized in R with the ggplot2 package.
Cytokine and chemokine assays
Protein was extracted using a 2-cm piece of distal colon and homogenized in 300 μl of Meso Scale Discovery lysis buffer with protease and phosphatase inhibitors as described in the assay protocol. The homogenates were centrifuged at 15,000 rpm for 10 min, and the protein concentration in the supernatant was determined using the Pierce Bicinchoninic Acid assay Kit (Thermo Scientific). Sample homogenates from naïve mice and infected mice were loaded into wells at 150 µg and 100 µg of total protein, respectively. The U-Plex Biomarker Group 1 (mouse) assay platform (Meso Scale Discovery, Gaithersburg, MD, USA) was used to measure interferon gamma (INFγ), interleukins (IL1β, IL4, IL6, IL10, IL12/23p40, IL17A, and IL22), keratinocytes-derived chemokine (KC), tumour necrosis factor alpha (TNFα), granulocyte-macrophage colony-stimulating factor (GM-CSF), matrix metalloproteinase-9 (MMP9), chemokine protein known as regulated on activation normal T cell expressed and secreted (RANTES), interferon gamma-induced protein-10 (IP10), monocyte chemoattractant protein-1 (MCP1), and macrophage inflammatory proteins (MIP1α, MIP2, and MIP3α). Final concentrations were presented as pg/ml in 100 μg of total colon protein.
Microbial community analyses
Total DNA was extracted from ileum, cecum, and colon contents using the QIamp Fast DNA Stool Mini Kit (Qiagen, Valencia, CA) with an additional bead-beating step using ~200 mg of garnet rock at 6.0 m/s for 60 s on a FastPrep-24 5G instrument (MP Biomedicals). Amplicon libraries were constructed according to the protocol from Illumina (16S Metagenomic Sequencing Library Preparation) that amplified the V3-V4 region of the bacterial 16S rRNA gene: 341F (5’- TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG - 3’) and 805R (5’-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC-3’). Paired-end sequencing was accomplished using an Illumina MiSeq Platform (2 x 300 cycles; Illumina Inc., San Diego, CA). Raw sequences were processed with Quantitative Insight into Microbial Ecology 2 (QIIME) [23] pipeline using the divisive amplicon denoising algorithm 2 (DADA2) to filter, trim and merge paired-end reads into amplicon sequence variants (ASVs). Ribosomal RNA data from cecal metatranscriptomic sequencing was processed as pre-merged single-end reads in QIIME2 using deblur denoise-16S function and trimmed at 160 bp. Phylogenetic trees were constructed using the qiime alignment (mafft; mask) and qiime phylogeny (fasttree; midpoint-root) function. Taxonomy was assigned using the qiime feature-classifier classify-sklearn function using the SILVA v138 database trained for the specific amplicon region [24]. QIIME2 files (.qza) were imported into R using qiime2R (version 0.99.4) package and analyzed with phyloseq (version 1.34.0) package [25].
Alpha diversity (Observed, Shannon, Phylogenetic diversity (PD)) and beta diversity (weighted and unweighted UniFrac) indices were analyzed with rarefied samples (ileum at 21,497, cecum at 34,012, and colon at 7,435 reads) in C3H/HeOuJ mice. In addition, we analyzed the cecal microbial community by analyzing the rRNA (rarefied at 1,247,061 reads) from the metatranscriptome sequencing data. Statistical significance for alpha diversity indices was determine with ANOVA and Tukey correction. Principal coordinate analyses (PCoA) was plotted using the phyloseq package and clustering significance was determined using the ‘betadisper’ function [26] for dispersion and ‘pairwiseAdonis.dm’ function [27] for orientation. Differential abundance analysis was done with DESeq2 using non-rarefied reads and ‘tree_glom’ (or tax_glom for cecum rRNA) function. The ‘log2foldchange’ of only the ASVs with a P value less than 0.05 were plotted with bolded ASVs signifying the significant adjusted P value < 0.10, < 0.05 (*), < 0.01 (**) and < 0.001 (***). Plotted ASVs were assigned according to their lowest classifiable taxonomic rank and are distinguishable by their corresponding ASV number from most to least abundant. Spearman’s correlation of the colonic microbiota (DESeq2 normalized ASV counts) and immune profiles in colon tissues from naïve (SURV and EPC) and infected (EPC) mice experiments were analyzed and visualized using the ‘psych’ and ‘pheatmap’ packages in R. Significant correlations were plotted on the heatmap as “*” for P value < 0.05 and “#” to denote a trend (P value < 0.10).
In vitro culture experiments
All in vitro culture experiments were done in an anaerobic chamber (5% CO2, 5% H2, and 90% N2) and cultures were incubated at 37 °C without shaking. B. thetaiotaomicron was isolated from C3H/HeOuJ mice fecal samples by serially diluting in 1 x PBS with 0.1% L-cysteine and plating on pre-reduced brain heart infusion (BHI; Difco) agar plus 10% calf blood (Cedarlane, ON, Canada) supplemented with 200 µg/ml of gentamicin [28]. Isolates were identified by amplifying and Sanger sequencing the 16S rRNA gene, and sequences were matched using BLAST web-based tool [29]. C. rodentium (10 ml of overnight culture) was inoculated alone or in competition with B. thetaiotaomicron (100 ml overnight culture) in 10 ml of pre-reduced low-glucose Dulbecco’s modified Eagle’s medium (Gibco life technologies, Grand Island, NY, USA) supplemented with cyanocobalamin at 0 ppm, 0.01 ppm and 15 ppm, which was subsequently incubated for 6 hours. Overnight cultures grown from a single colony of C. rodentium grown in Luria-Bertani broth at 1.4x107 CFU/ml and B. thetaiotaomicron grown in BHI broth (Difco) at 5.5x107 CFU/ml were used as inoculums. Counts were determined by plating on MacConkey agar for C. rodentium and the BHI with calf blood agar for B. thetaiotaomicron. Total RNA was immediately extracted from 1 ml of pelleted cells with 1 ml of Trizol reagent and purified using the spin columns as described above.
Reverse-transcription quantitative PCR
Colon tissues were homogenized in 600 ml of lysis buffer via bead beating and RNA was extracted using the GeneJET RNA Purification Kit (Thermo Scientific). Sample were treated with DNase as manufacturer’s protocols. RNA samples extracted from both colon tissue and in vitro culture experiment were reverse transcribed using the qFlex cDNA Synthesis Kit (Quanta Bioscience). Primers used for quantitative PCR (Table S1) were previously validated [28, 30, 31]. The qPCR was performed using PercfeCTa SYBR Green Super-mix (Quantabio) conducted on an ABI StepOne real-time System following the cycles: 95°C for 3 min and 40 cycles of 95°C for 10 s, 60°C for 30 s. Gene expression was calculated using the delta-delta Ct (2-ΔΔCt) method that showed the fold change relative to a housekeeping gene.
Statistical analysis
Significance testing was conducted using GraphPad Prism 6 (Graphpad Software, La Jolla, CA, USA). Student’s t-test or ANOVA was used for parametric and Kruskal-Wallis test was used for nonparametric data. Data were presented as mean ± standard deviation. Survival curve analysis was done using Mantel-Cox test with data up to day 10 post-infection. Differences between multiple treatments were corrected by conducting either the Bonferroni’s, Tukey’s, or Dunn’s post-hoc comparison test.