Gut Colonization with Methanogenic Archaea Lowers Plasma Trimethylamine N-oxide Concentrations in Apolipoprotein e−/− Mice

A mechanistic link between trimethylamine N-oxide (TMAO) and atherogenesis has been reported. TMAO is generated enzymatically in the liver by the oxidation of trimethylamine (TMA), which is produced from dietary choline, carnitine and betaine by gut bacteria. It is known that certain members of methanogenic archaea (MA) could use methylated amines such as trimethylamine as growth substrates in culture. Therefore, we investigated the efficacy of gut colonization with MA on lowering plasma TMAO concentrations. Initially, we screened for the colonization potential and TMAO lowering efficacy of five MA species in C57BL/6 mice fed with high choline/TMA supplemented diet, and found out that all five species could colonize and lover plasma TMAO levels, although with different efficacies. The top performing MA, Methanobrevibacter smithii, Methanosarcina mazei, and Methanomicrococcus blatticola, were transplanted into Apoe−/− mice fed with high choline/TMA supplemented diet. Similar to C57BL/6 mice, following initial provision of the MA, there was progressive attrition of MA within fecal microbial communities post-transplantation during the initial 3 weeks of the study. In general, plasma TMAO concentrations decreased significantly in proportion to the level of MA colonization. In a subsequent experiment, use of antibiotics and repeated transplantation of Apoe−/− mice with M. smithii, led to high engraftment levels during the 9 weeks of the study, resulting in a sustained and significantly lower average plasma TMAO concentrations (18.2 ± 19.6 μM) compared to that in mock-transplanted control mice (120.8 ± 13.0 μM, p < 0.001). Compared to control Apoe−/− mice, M. smithii-colonized mice also had a 44% decrease in aortic plaque area (8,570 μm [95% CI 19587–151821] vs. 15,369 μm [95% CI [70058–237321], p = 0.34), and 52% reduction in the fat content in the atherosclerotic plaques (14,283 μm [95% CI 4,957–23,608] vs. 29,870 μm [95% CI 18,074–41,666], p = 0.10), although these differences did not reach significance. Gut colonization with M. smithii leads to a significant reduction in plasma TMAO levels, with a tendency for attenuation of atherosclerosis burden in Apoe−/− mice. The anti-atherogenic potential of MA should be further tested in adequately powered experiments.

Atherosclerotic vascular disease is the leading cause of death in the US 1 . The gut microbiome is now recognized as a mediator of numerous host physiological processes. Changes in the gut microbiome have been causally linked to several metabolic, inflammatory, and cardiovascular diseases, including atherosclerosis 2,3 . Systemic concentrations of the gut microbe-derived metabolite, trimethylamine-N-oxide (TMAO), is associated with atherosclerosis and major adverse cardiovascular events [4][5][6] . Choline diet dependent enhancement in atherosclerosis could be prevented by either gut microbiota suppression with broad spectrum antibiotics 6 , or in more recent studies, administration of an inhibitor of microbial choline trimethylamine (TMA) lyase activity and therefore TMAO generation (4,4-dimethyl-1-butanol) 7 , indicating a causal link between TMAO and atherosclerosis. In this study, we tested a novel microbiota-based approach to reduce systemic exposure of TMAO.
Gut microbes generate TMA from ingested precursors, including choline 6 , carnitine 4 , phosphatidylcholine 5 , betaine 8 , and trimethyllysine 9 . TMA is converted to TMAO by the host hepatic flavin monooxygenase 3 (FMO3) 10 . Multiple phylogenetically distinct bacteria are involved in TMA production 11 . There are many microbial families, genus, and species which possess microbial enzymes that can make TMA. There have been three microbial enzyme genes/sources of TMA thus far identified; the cutC/D 11 , the cntA/B gene 12 , and the YeaW/Z gene 13 .
In humans, FMO3 gene mutations cause the inherited disorder primary trimethylaminuria (TMAU) 14 , also known as the fish odor syndrome, which severely reduces the ability to convert TMA to TMAO. Consequently, the affected individuals excrete large amounts of odorous TMA in their urine, sweat, and breath 15 . Although the disorder is not known to affect patient health, it can have profound social and psychological consequences. Development of a probiotic that can catalytically consume TMA could also serve as a therapeutic approach for treating this genetic disorder.
Methanogenic archaea (MA) represents a distinct group of anaerobic archaea that produce methane as the end-product of their anaerobic respiration. The abundance and diversity of gut MA in humans is highly variable and limited to a few species [16][17][18] . Methanobrevibacter smithii is the dominant methanogen in the human gut, detected in 95.7% of individuals, whereas Methaomassiliicoccus luminyensis is detected only in 4% of individuals 19 . Certain MA such as Methanosarcina species are known to use methylated amines as growth substrates 20,21 . A catabolic microbe that literally consumes TMA preventing/intercepting TMA prior to when it can be absorbed and converted into TMAO, would serve as a therapeutic intervention. However, up until now this potential has largely remained a theoretical concept. However, up until now this potential has largely remained a theoretical concept. In this study, we tested the effect of MA colonization on blood TMAO level and atherosclerosis burden in the atherosclerosis prone Apoe −/− mouse model.

Materials and Methods
All methods were performed in accordance to guidelines and regulations, and under approval of the Institutional Biosafety Committee of The George Washington University.  22 . For comparison, we selected M. blatticola (isolated from the hindgut of cockroach Periplaneta americana) 23 , M. mazei (isolated from a laboratory digester) 24 , and M. portucalensis (isolated from sediments at a saline lake) 25 . Animal Studies. All animal studies were performed under approval of the Animal Research Committee of The George Washington University. To test the efficacy of MA to colonize gut and metabolize TMA, 4-week-old female C57BL6J (Jackson Laboratory, Bar Harbor, Maine) were randomly selected for transplantation with one of the MA or sham, and placed on high TMA/choline diet. Because of the small sample size (n = 3/group), considering that the male mice have lower plasma TMAO levels, we used only female mice in this pilot study 6,10 . Prior to transplantations, endogenous gut microbiota were suppressed using an oral cocktail of poorly absorbed antibiotics previously reported to suppress TMA and TMAO levels 6 (0.5 g/L vancomycin, 1 g/L neomycin sulfate, 1 g/L metronidazole, 1 g/L ampicillin) administered in drinking water ad lib for three weeks, refreshed 3 times per week. Each mouse was inoculated with a single gavage of 10 8 MA 24 hours after the antibiotic treatment was completed. The test groups (n = 3/group) received a 0.1 ml gavage of corresponding MA in liquid growth media. The control groups (n = 3 each) received a 0.1 ml gavage of liquid growth media. Following transplantations, mice were supplemented with choline (1.0%, Sigma Aldrich) and TMA (0.12%, Sigma Aldrich) in drinking water ad lib, refreshed 3 times per week. The negative control group (NC) received untreated water. Both choline and TMA were given to animals since antibiotic treatment depletes most of the TMA-generating bacteria. All animals received standard chow diet for the entire study. Blood plasma and fresh stool samples were collected from the mice on days 2, 10, and 30 post-transplantation and stored at −80 °C, until processed. In a second study, 4-week-old female C57BL6J mice with Apoe −/− background (n = 5/group) (Jackson Laboratory) were treated with antibiotics, transplanted with M. smithii, M. mazei, and M. blatticola and fed with high TMA/choline diet as described above to determine the impact of MA on TAMO concentration. In a third study, we examined the effect of sustained high colonization with M. smithii on TMAO in Apoe −/− mice (n = 5/group). The experiment was performed as described above, except that M. smithii transplantation was repeated every three weeks, until the end of the study. One group of mice (n = 5) was maintained throughout the study on an antibiotic regimen (0.5 g/L vancomycin and 1 g/L ampicillin) to which MA are resistant, in order to chronically suppress the endogenous gut microbiota and enhance MA colonization ( Quantification of MA. Gut colonization by the MA was determined using quantitative real-time PCR (q-PCR). Briefly, microbial DNA was extracted from stool samples using the Qiagen DNeasy PowerSoil kit. The stool specimens were tested for the levels of engraftment by q-PCR as previously described 26 . Arch344f (ACGGGGYGCAGCAGGCGCGA) 26 and Arch806r (GGACTACCCGGGTATCTAAT) 27 primers were used to target the archaea 16 S rRNA gene present in MA. Q-PCR was performed for detection and quantification of microbes using SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA) in an Applied Biosystems 7900HT Sequence Detection system (Applied Biosystems) 28 . PCR products amplified from the bacterial and archaeal 16 s rRNA genes were cloned and grown in transformed E. coli to establish a standard curve, as described previously 29 . The accuracy of the q-PCR assay was confirmed through melting curve analysis and agarose gel electrophoresis. Calibration curves were obtained using 10-fold serial dilutions of known concentrations of cloned DNA samples that were prepared in the laboratory.
16S rRNA-based taxonomic survey. The V4 region of the 16 S rRNA gene was PCR amplified using barcoded primers 30 essentially as previously described 31 . Briefly, for each reaction, 2 μL of extracted DNA was combined with 12.5 μL 2X Phusion Hot Start II High-Fidelity PCR Master Mix (Thermo Fisher Scientific), and 100 nM each of forward and reverse primers. The primers were designed containing the Illumina linker adaptor, a unique index sequence, followed immediately by a variable sequence spacer (0-6 bases) and 16 S rRNA gene primers ( Table 2). The PCR was carried out in a 25 µl reaction in a Thermal Cycler (Applied BioSystem GeneAmp PCR system 9700) with the following parameters: initial denaturation at 94 °C for 3 min., followed by 35 cycles of 94 °C for 45 s, 50 °C for 60 s, and 72 °C for 90 s with a final extension at 72 °C for 10 min. The PCR was performed in triplicates with two PCR controls, a negative water control and a positive Mock Community control (M.G. Serrano, unpublished). All amplicons were quantified using Picogreen (Invitrogen/Thermo Scientific) and a spectrofluorimeter (Biotek). Amplicons were combined in equal volumes, followed by removal of unincorporated primers, salts and enzymes using Agencourt AMpure XP beads as described by the manufacturer. The DNA concentration of this concentrated pool was confirmed by qPCR using the KAPA Library Quantification Kit for Illumina platform. The library pool was diluted and denatured as described in the Illumina MiSeq library preparation guide. The sequencing run was conducted on the Illumina MiSeq using 600 cycles reagent kit (version 3) and 2 × 300 b paired end sequencing. Demultiplexing of sequence reads was performed using an in-house Python script. Raw paired-end sequence data was merged and quality-filtering using the MeFiT software 32 which invokes CASPER 33 for merging paired-end sequences and quality filters them using a meep-score (maximum expected error rate) cutoff of 1.0. Non-overlapping high quality paired end reads were retained for downstream analysis by linking them artificially with 15 Ns. High-quality sequences were taxonomically assigned using the Ribosomal Database Project (RDP) Naïve Bayesian Classifier (version 2.9) 34 with a bootstrap cutoff of 80%. PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) 35 was used to infer the abundance of metabolic pathways from 16 S rRNA gene sequencing data. The predicted functions (KOs) were collapsed into hierarchical KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways using the catego-rize_by_function step provided in the PICRUSt pipeline.

Assessment of atherosclerosis.
At the end of each study, mice were euthanized, and the heart was removed just proximal to the aortic arch. The heart was cut transversely at the level of the atria and placed ventricle down into a tissue mold in optimal cutting temperature compound (OCT compound) and stored at −20 °C until sectioning. Multiple cryosections at 10 μ thickness were taken of the aortic sinus and aortic root and stained for fat content with Oil-Red-O (ORO). Six sections from each heart were selected for image analysis, including the first section that contained the three leaflets of the aortic valve and the next 5 sections at 40 μ intervals along the aortic root distal to the valve over a total distance of 200 μ. The area of the plaque and the content of ORO was quantified by image analysis (ImageJ software) and calculated as an average of the area of plaque found in each of

MA can colonize normal C57BL/6 mice guts and lower their plasma TMAO concentrations.
In an initial experiment, we screened five MA for colonization efficacy and TMAO lowering potential using mice were similar to that obtained in C57BL/6 mice, ranging from 0.6 0 × 10 7 to 3.0 × 10 7 copies/mg stool on day 2 post-transplant ( Fig. 2A). TMAO concentrations in mice colonized with M. smithii (9.5 ± 5.9 μM), M. mazei (7.4 ± 6.1 μM) and M. blatticola (5.8 ± 0.6 μM) were significantly lower on day 2 compared to positive control mice (98.5 ± 26.5 μM) (Fig. 2B). The plasma TMAO concentrations remained low up to 10 days post-transplantation. The group x time interaction was significant (p = 0.004), indicating that the pattern of TMAO across time differed by species.
Repeated transplantation of Apoe −/− mice with M. smithii in presence of antibiotics leads to stable and high level gut colonization and diminishes the plasma TMAO concentrations. In order to improve gut colonization efficacies, we tested whether repeated transplantation with M. smithii leads to sustained high colonization level and persistent reduction in TMA concentration in Apoe −/− mice. In one group of repeatedly transplanted mice (n = 5/group) we also examined whether chronic suppressive antibiotic therapy further enhances colonization with M. smithii. As shown in Fig. 3A, repeated transplantation of M. smithii led to high engraftment levels throughout the study period. High and stable gut colonization with M. smithii led to lower average plasma TMAO concentrations (18.2 ± 19.6 μM) throughout the study period compared to that in positive control mice (120.8 ± 13.0 μM, p < 0.001) (Fig. 3B). The group of mice which were maintained on antibiotics, in addition to repeated transplantations, had a higher MA engraftment level compared to the mice not receiving antibiotics (1.3 ± 7.7 × 10 7 vs. 4.5 ± 6.4 × 10 6 average copies/mg stool) and lower average TMA concentrations (7.41 ± 9.38 vs. 29.1 ± 31.2 μM, p < 0.001).
Antibiotic depletion of Firmicutes enhances gut colonization by M. smithii. 16  pre-antibiotic microbiome profile of the mice consist of 53% Bacterioidetes, 43% Firmicutes, and 4.9% Verrucomicrobia (Fig. 4A). However, following the 3-week antibiotic treatment, the Bacterioidetes, Firmicutes and Verrucomicrobia levels decreased to 2.4%, 0.2%, and 0.1%, respectively, and Proteobacteria constituted the majority of the remaining bacteria that colonized in the gut (97.1%). In the negative control group, after stopping antibiotics, within 4 weeks, the Firmicutes returned and constituted the main surviving phylum (100%). Mice that were fed with Choline/TMA also showed 100% colonization by Firmicutes on week-4, but during the next 4 weeks, the level decreased to 32.8% with the return of Verrucomicorbia, which constituted 67.2% of the gut bacteria. Gut colonization levels with M. smithii were 2.4% and 2.2% in weeks 4 and 8, respectively in mice that were not on antibiotics, and 64% and 84% in weeks 4 and 8, respectively, in mice that received antibiotics. At the genus level, the untreated mouse microbiome was dominated by Bacteroides at the baseline, while Clostridium and Turicibacter trended higher among the remaining bacteria in antibiotic-treated mice (Fig. 4B). Choline/TMA supplemented mouse microbiome was dominated by Akkermansia, which was replaced by Clostridium and Turicibacter in M. smithii-colonized mice. There was a significant difference in the colonization levels of M. smithii between the antibiotic treated and untreated mice, and the proportional distribution of classes of bacteria varied significantly. Functional analysis using PICRUSt 35 showed that purine and pyrimidine metabolism, methane metabolism, and "phenylalanine, tyrosine and tryptophan biosynthesis pathways were active. (Fig. 4C).

Discussion
In this study, we investigated the in vivo TMA-metabolizing efficacy of several gut and not-gut associated methyl-trophic methanoarchaea of the genera Methanomicrococcus (Methanomicrococcus blatticola); Methanosarcina (Methanosarcina mazei), Methanohalophilus (Methanohalophilus portucalensis), and Methanomassiliicoccus (Methanomassiliicoccus luminyensis). The ability of all of these species to grow on TMA in culture has been previously demonstrated 23,25,38,39 . Also included in this study, was M. smithii, a member of Methanobrevibacter genus which is the most predominant methanoarchaea in the human gut. Unlike the other methanoarchaea used in this study, M. smithii is not known to grow on TMA in culture. However, a recent study demonstrated that many essential genes involved in methanogenesis, including methyltransferases are present and significantly up-regulated in vivo when M. smithii was co-colonized with a prominent human gut symbiont 40 .
We show that gut colonization with MA, consistently and significantly reduces plasma TMAO concentrations in two different mouse models that were maintained on a high choline and TMA diet. The reduction in plasma TMAO concentrations, in general paralleled the abundance of MA in the gut. We initially screened five species of MA isolated from gut and non-gut environments and noted that the colonization efficacy and the TMAO lowering efficacy is highest with M. smithii, a normal inhabitant of the human gut. Unlike M. smithii, the other indigenous human gut methanoarchaea, M. luminyensis, was very poor at gut colonization in mice and lowering blood TMAO levels. This observation is however, in accordance with the reports on low prevalence of M. luminyensis which is detected only in 4% of individuals, compared to that of M. smithii which has a high prevalence of nearly 96% 19 .  content in the atherosclerotic plaques, but these differences did not reach statistical significance possibly due to the small sample size.
Increased systemic exposure of TMAO is associated with atherosclerosis and major adverse cardiovascular events 4-6 . TMAO originates from the microbiota-dependent breakdown of dietary phosphatidylcholine to TMA. Functional studies have shown that TMAO decreases expression of bile acid transporters in the liver and reduces synthesis of bile acids from cholesterol 4 . TMAO also inhibits reverse cholesterol transport and promotes accumulation of cholesterol in macrophages through increasing cell surface expression of proatherogenic scavenger receptors CD36 and Scavenger Receptor A 6 , thus, creating foam cells that subsequently accumulate in the endothelial wall causing inflammation and plaque formation. Another mechanism related to atherosclerosis is the increase in thrombosis mediated by TMAO 41 .
It is therefore, not surprising that there is considerable interest in treatments designed to lower TMAO concentrations 7 . Targeting gut microbial production of TMA using a structural analog of choline, 3,3-dimethyl-1-butanol, inhibits TMA formation and lowers TMAO concentrations in mice fed a high-choline or L-carnitine diet 7 . 3,3-dimethyl-1-butanol inhibited choline diet-enhanced endogenous macrophage foam cell formation and atherosclerotic lesion development in Apoe −/− mice 7 . In this study, we took a different approach and aimed at depleting TMA as it is being formed by MA. In vitro studies have shown that certain MA can utilize TMA as a substrate for growth 38 . However, this is the first in vivo study that examines the utility of MA in decreasing plasma TMAO concentrations. We used a logical step-wise approach including screening, discovery and validation. For our initial screening, we chose normal C57/BL6 mice to evaluate the gut colonization capability and TMAO lowering efficacy of five selected species of MA. Subsequently, we tested the TMAO lowering efficacy of the three top performing MA in Apoe −/− mice. In the final experiment, we examined the effect of repeated transplantation with M. Smithii, and chronic suppression of endogenous bacteria by antibiotics, on colonization efficacy, TMAO concentrations and atherosclerosis level in Apoe −/− mice.
Our initial study in C57BL/6 mice showed that all MA tested were able to colonized the gut and reduce TMAO concentrations irrespective of their original habitat, but with large inter-individual variations. Efficient gut colonization by M. smithii was evident, but the colonization level of the other human gut MA, M. luminyensis, was poor. In the second experiment, a significant reduction in plasma TMAO concentrations were also observed in Apoe −/− mice transplanted with M. smithii (90.4%), M. mazei (92.5%), and M. blatticola (94%) on the 2 nd day, compared to non-transplanted control animals fed with high choline/TMA diet. Our initial experiments in both C57BL/6 and Apoe −/− mice revealed that, even after a 3-week-long antibiotic depletion of the gut bacteria, colonization was only transient with a single dose of MA. In a 3 rd experiment, we show that repeated transplantations significantly improved the gut colonization levels, which was further improved by antibiotic suppression of endogenous gut bacteria. There was a corresponding sustained reduction in TMAO concentrations. 16 S sequencing analysis confirmed that maintaining the transplanted mice on antibiotics greatly enhances M. smithii engraftment, but depleted Firmicutes and enriched Proteobacteria.
We chose to test M. smithii because it is known to be the most predominant MA in the human gut. M. smithii thrives in the distal intestine through its versatility in consuming the fermentation products made by saccharolytic bacteria, and by its ability to produce surface glycans and adhesion-like proteins 40 . M. luminyensis also colonizes human gut, although at much lower levels compared to M. smithii 19 . In human gut, TMA-metabolizing MA depend on the gut microbiota for converting choline to TMA which serves as a carbon-source for MA. In return, MA facilitates the continuation of bacterial fermentation process in the gut by metabolizing and removing H 2 .
On a chow diet, Apoe −/− mice demonstrate a total cholesterol level >500 mg/dL, with fatty streaks first observed in the proximal aorta at 12 weeks of age, and fibrous plaques appearing at 20 weeks of age 42 . A Western diet 43 , or a choline-rich diet 6 , induce a marked increase in plaque size and aggressive plaque morphology. We found that gut colonization with M. smithii resulted in a tendency for reduced atherosclerosis burden in Apoe −/− mice fed with high choline/TMA diet. The results were not statistically significant, possibly due to small number of animals studied. In addition, we cannot exclude the possibility that factors other than TMAO contribute to the pathogenesis of atherosclerosis, which were unmodified by this intervention 44 .
Since our data on the lowering of plaque characteristics were not statistically significant, either this study needs to be repeated in a much larger cohort of animals, or it could be concluded that lowering TMAO (as significant as our study demonstrates), does not lower TMAO-induced atherosclerosis in Apoe −/− mice.
We acknowledge that this is a pilot study in small number of animals and that the study could have benefited from measurement of TMA and choline concentrations; however, the study has a number of strengths. Strengths include: (a) first study to explore the utility of gut colonization by MA to lower plasma TMAO concentrations; (b) the evaluation of several candidate gut and non-gut MA in two mouse models; (c) the examination of single vs. repeated transplantation effect of MA; and, (d) the determination of MA engraftment efficiency after suppression of the endogenous gut microbiome.
While we do not know and did not investigate the factors that support increased engraftment level of MA in mice maintained on antibiotics, we speculate that there may be other bacteria and/or bacteria-derived factors that would either compete with or inhibit MA gut colonization. Clearly, the use of antibiotic in this study was to enhance gut colonization in our mouse model and facilitate the study of MA impact on blood TMAO levels, and not a recommended method to enhance the MA gut colonization in human. Future studies are needed to identify the exact mechanism and factors that support or compete with MA engraftment in human gut, and the consequence of increased MA abundance on the human host physiology and health.
To date, the impact of methanogens on human physiology and health is poorly understood. Association between MA and inflammatory bowel disease, colorectal cancers and obesity has been reported, however, there is no convincing evidence supporting the pathogenic properties of the MA, beyond occasional reports of constipation 45,46 . M. smithii has been shown to enhance energy retrieval by another sacharolytic bacterial species when co-transplanted into germ-free mice 47 . A study of fecal DNA extracts from patients with colorectal cancer, polypectomised, irritable bowel syndrome and control subjects found no significant association between these diseases and the presence of MA 48 . There is also no conclusive evidence that MA is associated with colon cancer 49,50 .
To summarize, our study shows that gut colonization with certain MA reduces plasma concentrations of TMAO. Among the MA tested, M. smithii, an endogenous human gut MA, was most effective in lowering plasma TMAO levels in Apoe −/− mice fed with high choline/TMA, and showed a tendency to attenuate atherosclerosis. The anti-atherogenic potential of the MA should be confirmed in studies using larger number of animals. The field of microbiome research is still nascent, but is evolving rapidly. Therapeutic use of specific toxin-degrading gut commensal microbes as probiotics to lower the levels of uremic toxins such as TMAO is a novel concept, but if proven effective, it will have a significant impact on cardiovascular disease and progression of chronic kidney disease. Safety of MA should be tested in pilot studies prior to launching efficacy studies in human subjects.