Oral vitamin B₁₂ supplement is delivered to the distal gut, altering the corrinoid profile and selectively depleting Bacteroides in C57BL/6 mice

ABSTRACT

Vitamin B₁₂ is a critical nutrient for humans as well as microbes. Due to saturable uptake, high dose oral B₁₂ supplements are largely unabsorbed and reach the distal gut where they are available to interact with the microbiota. The aim of this study was to determine if oral B₁₂ supplementation in mice alters 1) the concentration of B₁₂ and related corrinoids in the distal gut, 2) the fecal microbiome, 3) short chain fatty acids (SCFA), and 4) susceptibility to experimental colitis. C57BL/6 mice (up to 24 animals/group) were supplemented with oral 3.94 µg/ml cyanocobalamin (B₁₂), a dose selected to approximate a single 5 mg supplement for a human. Active vitamin B₁₂ (cobalamin), and four B₁₂-analogues ([ADE]CN-Cba, [2Me-ADE]CN-Cba, [2MeS-ADE]CN-Cba, CN-Cbi) were analyzed in cecal and fecal contents using liquid chromatography/mass spectrometry (LC/MS), in parallel with evaluation of fecal microbiota, cecal SCFA, and susceptibility to dextran sodium sulfate (DSS) colitis. At baseline, active B₁₂ was a minor constituent of overall cecal (0.86%) and fecal (0.44%) corrinoid. Oral B₁₂ supplementation increased active B₁₂ at distal sites by >130-fold (cecal B₁₂ increased from 0.08 to 10.60 ng/mg, fecal B₁₂ increased from 0.06 to 7.81 ng/ml) and reduced microbe-derived fecal corrinoid analogues ([ADE]CN-Cba, [2Me-ADE]CN-Cba, [2MeS-ADE]CN-Cba). Oral B₁₂ had no effect on cecal SCFA. Microbial diversity was unaffected by this intervention, however a selective decrease in Bacteroides was observed with B₁₂ treatment. Lastly, no difference in markers of DSS-induced colitis were detected with B₁₂ treatment.

Introduction

Vitamin B₁₂ is synthesized exclusively by microbes.¹ In the human gut, such synthesis occurs distal to the site of absorption, therefore this nutrient must be obtained from exogenous sources. The quantity of vitamin B₁₂ required by adults, 2 µg/day, is lower than any other micronutrient.² Commercially available supplements contain up to 5 mg/tablet, greatly exceeding the daily requirement. Unlike other water-soluble vitamins, which are largely absorbed and enter the circulation, receptor-mediated B₁₂ absorption in the ileum becomes saturated around 2 µg/meal¹ and only a small fraction of high dose oral supplements are absorbed by non-specific pathways. Following supplement ingestion, non-absorbed vitamin B₁₂ reaches the large intestine where it is available to interact with the colonic microbiota. Microbes require vitamin B₁₂ or related corrinoids for their own functions. It is estimated that 20% of gut bacteria are able to synthesize vitamin B12, yet >80% of gut bacteria require this nutrient for their own metabolic purposes.^3,4 Further, microbes produce various B₁₂ analogues by altering the lower ligand base (Figure 1(a)), and only a small fraction of the corrinoid in human feces is active B₁₂.⁵ Auxotrophic bacteria rely on B₁₂ transporters for acquisition, even encoding multiple transporters with varied preference for different B₁₂ analogues to gain competitive advantage.³ B₁₂ destined for utilization by the host is chaperoned throughout the digestive tract and then during circulation by glycoproteins (haptocorrin, intrinsic factor, transcobalamin II), possibly as a means of preserving them from microbial acquisition.⁶ The importance of B₁₂ for microbial physiology is illustrated by early B₁₂ assays which derive B₁₂ concentration from the rate of bacterial growth.⁷ Collectively, these observations led us to hypothesize that high dose oral B₁₂ could fundamentally alter the corrinoid economy and microbial composition in the distal gut.

Figure 1.

B₁₂ supplement alters the corrinoid profile in the cecum and feces. (a) Base moieties of the most abundant corrinoids. Cobinamide lacks the base as well as ribose and phosphate moieties present in other corrinoids. (b) Cecal corrinoid profile overall and (c) by individual corrinoid in control (n = 5) and B₁₂ treated (n = 4) mice by unpaired two-tailed Student’s t-test. (d) Fecal corrinoid profile overall and (e) by individual corrinoid at baseline and after 16-day B₁₂ supplementation (ADE n = 16/group, others n = 22/group, paired two-tailed t-test).

Two known B₁₂-dependent enzymes have been described in humans, but at least 15 B₁₂-dependent enzymes exist among gut microbes.⁸ B₁₂ has additional roles in microbial gene regulation, functioning in B₁₂-dependent riboswitches (regulatory RNA elements),⁹ and as a cofactor for gene regulatory proteins.¹⁰ L-Methylmalonyl-CoA mutase is a B₁₂-dependent enzyme shared by humans and microbes (though the direction of product-substrate is reversed). Anaerobic microbes use this enzyme to generate CO₂, a critical electron acceptor in the anaerobic lumen.¹¹ Propionate, a short-chain fatty acid (SCFA) generated as a byproduct of this reaction, is important in host physiology and immunity.^4,12 In one study, the addition of B₁₂ to bacterial growth media increased propionate production by three bacterial species.¹³ This knowledge prompted the hypothesis that oral B₁₂ might alter microbial SCFA production and thereby, resistance to colitis in the distal gut.

Here we report results from murine experiments that evaluated the effect of oral B₁₂ supplementation on luminal B₁₂ and related corrinoids in the distal gut. In addition, we assessed the effect of B₁₂ supplement on the fecal microbiome, cecal SCFA, and resistance to experimental colitis.

Results

As a starting point to profile B₁₂ metabolites in mice, we measured the influence of oral B₁₂ supplementation (3.94 µg/ml) on corrinoids in the distal gut. Cobalamin, and four other corrinoid analogues were selected for analysis based on their abundance in human feces⁵ and the availability of standards for LC/MS. These included cobalamin, [ADE]CN-Cba, [2Me-ADE]CN-Cba, [2MeS-ADE]CN-Cba, and CN-Cbi, which differ from cobalamin in the lower ligand base moiety (Figure 1(a)). Cobalamin was found to be a minor constituent representing just 0.86% of cecal and 0.44% of fecal corrinoid at baseline (Figure 1(b–d)). Supplemented animals had 132-fold higher cobalamin concentration in cecal contents (Figure 1(c). 0.08 vs 10.60 ng/mg, p = 0.0005) and 136-fold higher concentration in fecal contents (Figure 1(e). 0.06 vs 7.81 ng/mg, p = 0.0001) with Bonferroni-corrected α of 0.0125 and 0.01 respectively. Likewise, MeS-Ade was 5.6-fold higher in cecal contents of supplemented animals (Figure 1(b,c). 0.26 vs 1.45 ng/mg) but this was non-significant after Bonferroni correction. In contrast, there was a robust decrease in cobamide analogues (Figure 1(d,e)) in feces of supplemented mice except for cobinamide which was increased. To control for change over time we compared corrinoids in non-supplemented control animals at days 1 vs 16, and found a marginal decrease in fecal cobalamin, which was non-significant after Bonferroni correction (α = 0.01; Fig S1).

Given the widespread microbial requirement for exogenous corrinoid, we hypothesized that B₁₂ supplementation would select for microbes whose growth was otherwise limited by this nutrient. Stool samples were collected for 16S rRNA gene sequencing at baseline (T1) and after 16 days (T2) of treatment. Including samples from DSS colitis experiments, described below, 54 stool and cecal samples were collected for microbiome analysis, of which 52 were successfully profiled (two animals given DSS had inadequate cecal contents to generate adequate 16S rRNA gene amplicons). In total 3,242,801 sequences were obtained (average 60,051 per sample). Goods coverage exceeded 99.8% for all samples analyzed demonstrating adequate sequencing. Beta diversity analysis at phylum level (Figure 2(a)) showed similar overall composition before and after B₁₂ supplementation, while a significant difference was seen at the genus level (Figure 2(b), p = 0.048). Additionally, this analysis yielded one conspicuous finding: the closely related genera Bacteroides (2.50% OTU baseline vs 0.54% OTU post, p = 0.027) was significantly reduced after B₁₂ supplementation (Figure 2(c)). Furthermore, while alpha diversity in these samples demonstrated no difference by treatment, principal coordinates analysis (PCoA) revealed that the control vs. B₁₂ groups post-supplementation differed significantly (Fig S2).

Figure 2.

Influence of oral B₁₂ supplement on the fecal microbiome. Beta diversity analysis of (a) phylum and (b) genus level (n = 12 mice/time point) at baseline and following 16-day B₁₂ supplementation. *p < 0.05. (c) Significant reduction in the relative abundance of Bacteroides following B₁₂ supplementation (paired two-tailed t-test). (d) Cecal short-chain fatty acids in animals administered H₂O (Control) and experimental colitis (DSS) with and without B₁₂. (Control n = 15, B₁₂ n = 15, DSS n = 20, DSS/B₁₂ n = 18; paired two-tailed t-test (c) and ANOVA with Tukey’s post-test (d).

Culture experiments have shown increased propionate synthesis by some bacteria with the addition of B₁₂ to culture media.¹³ Therefore, we tested the hypothesis that B₁₂ would alter cecal SCFA concentrations. Cecal contents rather than stool was used for SCFA analysis because a large portion of SCFA produced in the gut is absorbed in the colon before luminal contents are expelled as fecal pellets. This analysis did not identify any effect of B₁₂ on cecal acetate, propionate, or butyrate levels (Figure 2(d)).

Published work implicates commensal Bacteroides species in the pathogenesis of murine colitis,^14,15 in contrast, lower levels of Bacteroides are associated with human inflammatory bowel disease (IBD).^16–18 Given our finding that oral B₁₂ supplementation decreased the proportion of Bacteroides in feces, we sought to determine the effect of oral B₁₂ in murine colitis. The DSS model of experimental colitis was chosen because prior studies using this model demonstrated a role for SCFA-mediated signaling.¹⁹ As expected, induction of experimental colitis with the addition of 2.25% DSS in drinking water resulted in lower weight, shorter colon length, and increased gut permeability as reflected by the appearance of serum fluorescence following gavage of FITC-dextran (Figure 3(a–c)). However, B₁₂ supplementation, which was sustained during DSS administration, did not significantly influence these endpoints (Figure 3(a–c)). Similarly, there was no significant difference in IL-1β, TNF-α, IL-6, IL-10, IL-12p70, IFN-γ, or murine KC in colonic mucosal scrapings with oral B₁₂ supplementation in DSS colitis (Figure 3(d)). Given the potential for host B₁₂ status to influence response to colitis,²⁰ we included a control group that received intraperitoneal B₁₂ injection (parenteral B₁₂) – an intervention that did not alter fecal corrinoids (Fig S3).

Figure 3.

B₁₂ supplement in DSS colitis. (a) Vitamin B₁₂ had no effect on weight loss in DSS colitis (sum of three replicate experiments, Control (n = 15), B₁₂ (n = 15), DSS (n = 23), DSS/B₁₂ (n = 24) by ANOVA and Dunnett’s multiple comparisons (Control vs DSS or DSS/B₁₂: p < 0.0001, but DSS vs DSS/B₁₂: p = ns) or measures of disease including (b) colon length, Control (n = 15), B₁₂ (n = 15), DSS (n = 20), DSS/B₁₂ (n = 19) by ANOVA and Tukey’s multiple comparisons, or (c) enteral administered FITC-dextran detected in circulation of Control (n = 14), B₁₂ (n = 15), DSS (n = 20), DSS/B₁₂ (n = 19) by ANOVA and Tukey’s multiple comparisons. (d) Colon tissue cytokines were not significantly different comparing DSS vs DSS/B₁₂ by ANOVA and Tukey’s multiple comparisons. Unique letter represents p < 0.05.

Discussion

It is predicted that 80% of gut microbes require B₁₂ or similar corrinoids, though a minority harbor the B₁₂ biosynthetic pathway.⁸ Therefore, most gut microbes rely on acquisition of exogenous corrinoid or precursor molecules from their environment and successful competition confers a survival advantage.³ Therefore, we hypothesized that oral supplementation would deliver excess B₁₂ to the distal gut, disrupting the corrinoid economy and thereby alter the microbiome, SCFA, and resistance to colitis.

To our knowledge, this is the first study to profile corrinoids in the distal gut of mice since the LC/MS method was originally described by Allan and Stabler.⁵ Importantly, it demonstrates that active B₁₂ is a minor constituent of the fecal corrinoid pool in mice (0.44%, 59 ng/g stool) which is similar to published results from human feces (1.4%, 19 ng/g stool).⁵ We show that oral B₁₂ supplementation in mice, at a level corresponding to human consumption of a 5 mg daily supplement, increased B₁₂ in the distal gut >130 fold. B₁₂ supplementation decreased the absolute concentration of several non-B₁₂ corrinoids. Cobinamide was increased in fecal contents which indicates microbial modification of cobalamin salvaging the corrin ring. Given the ability of microbes to sense and acquire B₁₂,^3,9 this likely reflects negative regulation of microbial corrinoid synthesis in the setting of environmental excess.

Our analysis did not reveal major alterations of the fecal microbiome with B₁₂ supplementation. Use of samples from replicate experiments reduced the chance of type I error in this analysis, the study likely was limited in its power to detect changes in the microbiome. The primary finding was the selective decrease in Bacteroides, a highly prevalent constituent of the mammalian gut microbiome.²¹ Bacteroides require exogenous B₁₂,^13,22 so the depletion of this genus with supplementation is paradoxical. Bacteroides derive competitive advantage from redundant B₁₂ transporters that enable them to acquire environmental corrinoid. This leads us to speculate that Bacteroides may lose this advantage in an environment replete with B₁₂. Conflicting data exist regarding the role of Bacteroides in colitis. Bacteroides in general^16,23 and enterotoxigenic B. fragilis specifically²⁴ are associated with IBD in humans and contribute to pathogenesis of colitis in animal models.^15,25,26 To the contrary, monocolonization of germ free mice with B. fragilis is protective in DSS colitis,²⁷ likely through immunomodulatory influences of its capsular polysaccharide-A²⁸ and modulating secreted outer membrane vesicles.²⁹ Given this selective decrease of Bacteroides, we subjected B₁₂ supplemented mice to DSS colitis to evaluate their susceptibility to disease. Despite some trends toward B₁₂-mediated protection of barrier function, no significant influence of B₁₂ supplementation was discernable during DSS administration (Figure 3(a–d)-D). These findings agree with recent data regarding the role of cyanocobalamin in murine colitis. Zhu et al. show that cyanocobalamin supplementation during DSS administration had a negligible effect on colitis outcomes, including weight loss and colon length reduction.³⁰ Furthermore, cyanocobalamin supplementation during DSS did not have a significant effect on microbial alpha diversity,³⁰ similar to our findings. This is in contrast with work in other organ systems, which has shown that cyanocobalamin exhibits anti-inflammatory effects again acute and chronic inflammation in mice.³¹

The finding that cecal SCFA concentrations are unaffected by B₁₂ supplementation (Figure 2(d)) contrasts with results of bacterial culture experiments in which addition of B₁₂ increased propionate production.¹³ This may reflect resilience of the metabolic network in the distal gut lumen.¹¹ Alternately, an influence of B₁₂ on cecal SCFA may have been obfuscated in the background of B₁₂ provided by the standard diet that all animals received. The decision to use standard B₁₂-containing chow and the selection of the B₁₂ dose for these studies was intended to reflect typical human exposures. One strength of this work is the number of mice included and replication of experiments. For example, we report that DSS reduces butyrate in cecal contents (Figure 2(d), n = 20/group). This finding corroborates results from other studies that employed just five animals/group.^32,33

In summary, we report that oral B₁₂ supplement, as cyanocobalamin, was effective in delivering B₁₂ to the distal gut of C57BL/6 mice as well as modulating B₁₂ analogues and selectively depleting Bacteroides, but did not appreciably influence cecal SCFA levels or protection in DSS colitis.

Methods

Mice and B₁₂ treatment

All experimental protocols were approved by the University of Colorado Institutional Animal Care and Use Committee. Male and female C57BL/6 mice were bred and maintained in a specific-pathogen-free facility and used at age 8–16 weeks. Animals were evenly distributed amongst treatment groups by age and sex. Mice were fed a standard diet (Teklad Global Soy Protein-Free Extruded Rodent Diet) containing .08 mg/kg vitamin B₁₂ as cyanocobalamin. Mice were supplemented with B₁₂ in the form of cyanocobalamin (Sigma-Aldrich, St. Louis, MO, USA) in drinking water at a concentration of 3.94 µg/mL, which provided 1.025 µg B₁₂/g, or on average, 30.46 µg B₁₂ per mouse, based on water disappearance. This is the human equivalent of a single 5,000 µg oral supplement based on published data on C57BL/6 water consumption using standard body surface area based conversion between human and mouse.^34,35 Mice were supplemented with B₁₂ throughout the duration of the experiment with baseline and post supplement samples collected at day 1 and day 16. Water bottles containing B₁₂ were replaced every other day. As a control in colitis experiments, parenteral B₁₂ (via intraperitoneal injection) at a dose of 2.65 µg/mouse, every 3 days, was provided to the control group to account for potential systemic influences of B₁₂ on colitis.²⁰ This dose was selected to provide an optimistic 2.9% absorption/day.³⁶

Corrinoid analysis

Cobalamin and B₁₂ analogues were measured in cecal contents and stool by stable isotope dilution liquid chromatography/mass spectrometry (LC/MS) as previously described by Allen and Stabler.⁵ Briefly, frozen samples were thawed and 1 mL 0.1M KPO₄ buffer at pH 8.0 was added with stable isotope labeled internal standards. Samples were heated to 90°C for 45 min, then cooled on ice and centrifuged and eluted fraction added to prepared C18 reverse phase resin columns. The eluant was dried overnight before rehydration in 1 mL H₂O and then passed through R-protein resin which was prewashed with 1 mL phenol, 3 mL H₂O and 1 mL 0.5M glycine. Sample was eluted with 100 µl phenol, and the aqueous layer was extracted and analyzed by LC/MS as detailed previously.⁵

Microbiome and SCFA analysis

Bacterial profiles were determined by broad-range amplification and sequence analysis of the bacterial 16S rRNA gene V3V4 variable region following our previously described methods. Fecal pellets and cecal contents were collected directly into sterile Eppendorf tubes and flash frozen at −80°C until samples were processed for DNA isolation and sequencing. Bacterial DNA from cecal and fecal samples were isolated using the Power Fecal DNA Isolation Kit (MO BIO Laboratories, Carlsbad, CA). Paired-end sequencing was performed on the Illumina MiSeq platform using the 600-cycle v3 kit. Sequences were merged and quality filtered as previously described.^37,38 Merged sequences were aligned and classified with SINA (1.3.0-r23838).^39,40 Operational taxonomic units (OTUs) were produced by clustering sequences with identical taxonomic assignments. Explicet software (v2.10.5) was used for microbiome analyses, including calculation of alpha and beta diversity measures such as Goods Coverage index.⁴¹ Short chain fatty acids were analyzed by stable isotope GC/MS using an adaption of a published method.⁴² Briefly, cecal samples were collected directly into pre-weighed, sterile Eppendorf tubes and flash frozen at −80°C until processing. Samples were then subject to an alkylation procedure in which sample and alkylating reagent were added, vortexed for 1 min, and incubated at 60°C for 15 min. Following cooling and addition of n-hexane to allow for separation, 200 µL of the organic phase was transferred to glass inserted and analyzed by GC/MS. Results were quantified by reference to a standard curve and normalized to sample weight.

Experimental colitis

Colitis was induced using 2.25% dextran sodium sulfate (DSS), (MW 36,000–50,000, MP Biochemicals, Santa Ana, CA, USA) added to drinking water for 9 days, as previously reported.⁴³ DSS was made fresh daily. Animals were weighed daily and on the day of euthanasia were gavaged with 3-5KD FITC dextran (Sigma Aldrich, St. Louis, MO, USA). Serum was collected at euthanasia, four hours after gavage, and fluorescence quantified. Mucosal scrapings from the colon were placed in phosphate buffered solution containing Halt Protease Inhibitor Cocktail (Fisher Scientific) and stored at −80°C. Protein concentration was determined using the Pierce BCA Protein Assay Kit (Thermo Scientific, Waltham, MA, USA). Tissue cytokines were quantified as previously reported⁴⁴ using the Mouse Proinflammatory 7-Plex Tissue Culture Kit and according to manufacturer instructions (Meso Scale Discovery, Rockville, MD, USA).

Statistical Analyses were performed using GraphPad Prism 7 (GraphPad Software, La Jolla, CA, USA). Tests include paired and unpaired two-tailed Student’s t-test and ANOVA with Tukey’s and Dunnett’s multiple comparisons as indicated in figure legends. Statistical differences reported as significant when p < 0.05 with Bonferroni correction applied when multiple comparisons were made.

Funding Statement

This work was supported by the National Institutes of Health [DK1047893, DK50189, DK095491, DK103712] and by the Veterans Administration [Merit Award BX002182]. 

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website