REVIEWS Contributions of the microbial hydrogen economy to colonic homeostasis Franck Carbonero, Ann C. Benefiel and H. Rex Gaskins Abstract | Colonic gases are among the most tangible features of digestion, yet physicians are typically unable to offer long-term relief from clinical complaints of excessive gas. Studies characterizing colonic gases have linked changes in volume or composition with bowel disorders and shown hydrogen gas (H 2), methane, hydrogen sulphide, and carbon dioxide to be by-products of the interplay between H 2-producing fermentative bacteria and H2 consumers (reductive acetogens, methanogenic archaea and sulphate- reducing bacteria [SRB]). Clinically, H2 and methane measured in breath can indicate lactose and glucose intolerance, small intestinal bacterial overgrowth and IBS. Methane levels are increased in patients with constipation or IBS. Hydrogen sulphide is a by-product of H2 metabolism by SRB, which are ubiquitous in the colonic mucosa. Although higher hydrogen sulphide and SRB levels have been detected in patients with IBD, and to a lesser extent in colorectal cancer, this colonic gas might have beneficial effects. Moreover, H 2 has been shown to have antioxidant properties and, in the healthy colon, physiological H2 concentrations might protect the mucosa from oxidative insults, whereas an impaired H2 economy might facilitate inflammation or carcinogenesis. Therefore, standardized breath gas measurements combined with ever-improving molecular methodologies could provide novel strategies to prevent, diagnose or manage numerous colonic disorders. Carbonero, F. et al. Nat. Rev. Gastroenterol. Hepatol. advance online publication 15 May 2012; doi:10.1038/nrgastro.2012.85 Introduction Despite a long history of dealing with bowel gas and a remaining being metabolized (H 2 only) or passed basic accounting of its components and their origins, as flatus. 4 H 2 metabolism—reflecting the balance our understanding of the mechanisms by which these between H 2-producing (hydro­g enogenic) bacteria gases are produced and metabolized remains rudimen- and H2-utilizing (hydrogenotrophic) microbes—has a tary. Human colonic gases are comprised of hydrogen primary influence on the final composition of colonic (H2), carbon dioxide (CO2), methane (CH4), nitrogen gases. Here, we review the microbiology of H2 produc- and oxygen as well as several odiferous trace gases. tion and utilization, and mechanisms by which these Nitrogen and oxygen are exclusively derived from swal- pathways influence colonic health and disease. lowed air; on the other hand—making up ~74% of the flatus—H2, CO2 and CH4 are produced solely by colonic Colonic gases microbes, which ferment dietary components that escape Early measurements digestion by host enzymes and endogenous substrates Early attempts to measure excretion of colonic gases derived from the colonic mucosa.1 Other microbial gases focused on total volume and composition of flatus are present in flatus in trace concentrations; for example, gases under baseline and varying dietary conditions hydrogen sulphide (H2S) at 1.06 μmol/l, methanethiol at (Figure 1).5–9 Ruge is credited with the first attempts to 0.21 μmol/l and dimethyl sulphide at 0.08 μmol/l in one collect colonic gases from humans in the 1800s, using a study.1 Intracolonic concentrations of these trace gases glass tube from the anus to a water displacement system might be substantially higher than detected in flatus, beneath a specialized chair.10,11 In 1942, Beazell and Ivey 8 as H2S and methanethiol rapidly permeate the colonic measured flatus gases from healthy individuals over a mucosa and are detoxified.2 Marked individual differ- 24 h period and determined that the daily excretion rate ences also exist in the proportional composition of major was in the range of 380–655 ml per day. Kirk demon- colonic gases.1,3 strated that total flatus production increased with dietary The accumulation of gas in the colonic lumen is fibre consumption.5 In later studies, Steggerda6 collected University of Illinois at Urbana-Champaign, dependent on the interplay between various micro- an average of 360 ml flatus per day, of which 7.4% was 1207 W. Gregory Drive, bial metabolic pathways and host physiology. Most CH4 and 19.8% H2, from individuals on a baseline diet Urbana, IL 61801, USA CO 2, and 30–40% of H 2 and CH 4, are absorbed by of foods typically considered non-flatus-­producing (F. Carbonero, A. C. Benefiel, the colonic mucosa and recirculated in the blood, the (such as boiled eggs, lean beef and apple sauce), demon­ H. R. Gaskins). strating the microbial origin of these two gases in the Correspondence to: Competing interests colon. 6,12 By adding varying types and amounts of H. R. Gaskins The authors declare no competing interests. beans to the diet over subsequent 7‑day trials—whilst hgaskins@illinois.edu NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY ADVANCE ONLINE PUBLICATION  |  1 © 2012 Macmillan Publishers Limited. All rights reserved REVIEWS Key points methane excretion also varies markedly demonstrates ■■ The colonic gases hydrogen (H2), carbon dioxide and methane (CH4) are the key role of H2 microbial consumption in determining end products of microbial fermentation; their concentrations depend on the intraluminal concentrations of this gas. interplay between host physiology and H2-producing (hydrogenogenic) and H2-using (hydrogenotrophic) microbes Current understanding ■■ Colonic H2 production is most readily measured via excretion in breath; Whole-body calorimetry demonstrated the dynamic clinically, breath H2 and CH4 are commonly measured to assess lactose and relationship between flatus and breath gas excretion of glucose intolerance and small intestinal bacterial overgrowth, and increasingly both H2 and CH4.4 Improvements in means of meas- IBS uring H2 from exhaled breath have enabled reliable, ■■ Improved understanding of microbial H2 metabolism and its relation to expired gas concentrations will reinforce the breath gas test as a widely applicable, non­invasive estimates of colonic H 2 levels in a clini- easy and cost-effective diagnostic or prognostic tool cal setting. 19 Gas chromatography systems have now ■■ Use of breath gas tests in diagnosis could enable novel therapeutic or become common for detecting the somewhat low con- preventative measures for a wide array of colonic diseases centrations of H2 (1–200 parts per million) in breath.11 ■■ Although emphasis has been given to the potential inflammatory or Calloway 20 demonstrated changes in respiratory H2 carcinogenic properties of colonic gases, emerging evidence suggests these and CH4 levels with consumption of gas-forming food gases might have a beneficial effect in colonic health such as beans, linking early direct measurements from the lumen or in flatus with levels detectable in breath. maintaining the baseline level of fat, protein, carbo- Calloway and Murphy 21 replicated the findings of Levitt hydrates and calories—Steggerda also examined the and Ingelfinger—that a subgroup of individuals seem to effect of altering the type, but not overall amount, of produce little to no colonic CH4—using breath testing. dietary carbohydrate on production of flatus. 6,12 He Although all intestinal CH4 derives from microbial concluded that the low-molecular-weight fraction of methanogenesis, estimates of colonic production based carbohydrates (mono­s accharide, disaccharide and on breath measurements should be interpreted with oligosaccharides) were responsible for an increase in caution, as proportions of both H2 and CH4 excreted overall gas excretion, up to an average of 4,224 ml per in breath are influenced by the production rates in the day when a commercial pork and bean diet comprised colon; 65% of the gas is excreted in breath at low pro- 57% of the diet.6 Using a constant infusion technique, duction rates (<200 ml per day) and 25% at high rates Levitt and Ingelfinger 13 measured the rate of H2 and (>500 ml per day).4 Several studies show that breath H2 CH4 production in healthy individuals directly from concentrations are lower in CH 4-excretors than non- the intestinal lumen, demon­strating that H2 was pro- excretors.4,22–24 The clinical value of breath testing for duced primarily in the large intestine of all individuals, colonic gases was further demonstrated by Calloway correlated with breath excretion levels, and was almost et al. 25 and Levitt and Donaldson 26 who linked an completely dependent upon fermentation of dietary increase in breath H2 concentration with fermentation substances. Of the nine study participants, four were in the colon of malabsorbed lactose, suggesting breath CH4 producers (producing colonic CH 4 after lactose H 2 levels as a measure of lactose intolerance. Metz infusion at a rate of 0.5–0.6 ml/min); the remaining et al.27 compared intestinal lactase activity with symp- participants produced no detectable CH4.13 toms such as abdominal pain and bloating, increases in Less than 20% of carbohydrate remains unabsorbed in blood glucose concentration, and breath H2 production people consuming a typical Western diet.14 Theoretically, in a group of patients with diarrhoea and found end- with this amount of substrate available for bacterial expiratory breath H2 to be as reliable as blood glucose fermentation and H2 production occurring at a rate of concentration, and better than symptoms, in the 340 ml/g of glucose,15 the potential exists for >13 l of H2 diagnosis of lactase deficiency. to be generated daily. However, Strocchi and Levitt 16 Easy to administer, H2 breath tests (typically follow- measured a mean absolute H2 production rate during ing a period of fasting and ingestion of lactulose) have glucose fermentation of 80 ml/g, suggesting that fermen- become one of the primary means of detecting small tation by stool bacteria might involve metabolic path- intestinal bacterial overgrowth (SIBO).28 Theoretically, ways that do not liberate H2. Hammer,17 likewise, found an early peak in breath H2 indicates fermentation of lact- H2 production in flatus after fasting to be very low and ulose in the small intestine;11 however, conditions such in the range of 50–200 ml per 6 h period after ingestion as rapid intestinal transit, in which lactulose would reach of 12.5 g lactulose. Indeed, H2 excretion varies markedly the colon much earlier than is typical, can confound with different substrates, throughout the day and among results. In addition, breath H2 peaks do not always cor- individuals. 6,7,17,18 Although yet to be quantitatively relate with results of gold-standard tests for SIBO (such determined, such interindividual differences probably as cultures from jejunal aspirate), and high H2 concen- reflect complex interactions between host physiologi- trations can be found in apparently healthy individu- cal and microbial factors, including: quantity and type als with no other indication of SIBO.29,30 Although the of substrate; the ability of colonic microbes to ferment value of breath testing in detecting SIBO—and even the carbohydrates; the abundance and location of different importance of SIBO in disorders such as IBS, immuno- types of H2-producing and H2-consuming microbes; the deficiency syndromes and motility disorders—is contro- efficiency of stirring of gut contents; and environmen- versial, breath testing to detect H2 and CH4 production tal factors such as pH and sulphate availability. 11 That in the colon remains a valuable indicator of microbial 2  |  ADVANCE ONLINE PUBLICATION www.nature.com/nrgastro © 2012 Macmillan Publishers Limited. All rights reserved  REVIEWS 1992 Christl 1968 Levitt & Ingelfinger 1975 and 1976 Metz Whole-body H2 production H2 in end-expiratory calorimetry used primarily colonic; rates air linked to to compare rate of measured and correlated lactase deficiency H2 and CH4 excretion with breath excretion and SIBO in flatus and breath 2002 Olson & Maier 1868 Ruge 1966 Calloway 1970 Levitt & Bond 1984 Pique Gastric mucosa contains Contribution to Respiratory H2 and Composition of Increased breath H2 and is used by 2005 Dear et al. the knowledge of CH4 affected by intestinal gas CH4 level linked to Helicobacter pylori Reducing fermentation, intestinal gases gas-forming foods 0.06–47% H2 colorectal cancer as energy source improves IBS symptoms 1870 1940 1950 1960 1970 1980 1990 2000 2010 1941 Beazell & Ivey 1968 Steggerda 1979 Tedesse et al. 1987 Bujanover et al. 1998 King et al. Colonic gas excretion Composition of flatus Breath H2 test used as Increased breath CH4 Increased volume and 380–655 ml per 24 h determined under measure of transit time found in cystic fibrosis rate of H2 excretion varying diets and lactose intolerance in patients with IBS 1949 Kirk 1977 Tadesse & Eastwood 1983 Solomons et al. 1993 Hammer 2003 Pimentel et al. CH4 gas measured in Smoking causes acute Described portable Colonic H2 absorption CH4 linked to human flatus; total increase in breath breath gas analyzer affected by accumulation constipation- volume increased with H2 and CH4 levels with pump to rate; varies based on predominant IBS increased fibre eliminate need for individual differences (600 g Brussels sprouts) large gas tank in bacterial metabolism Figure 1 | Milestones in the measurement of colonic gases in breath and flatus. Colonic gases in breath and flatus have been investigated since at least the late 1800s. The clinical utility of breath testing, however, emerged much later in the mid-to-late 1970s when it was demonstrated to be an accurate measurement of disaccharide malabsorption. Since then, the sensitivity and practicality of instrumentation has improved considerably, and the measurement of H2 and CH4 in breath is now generally accepted by most gastroenterologists as a reliable indicator of lactose intolerance and SIBO. Its value in evaluating IBS has been controversial, but might be increased by standardization of substrate, sampling frequency and duration and routinely including CH4 as well as H2 testing. This noninvasive tool is fairly inexpensive, easy to use and is one of few methods that provides real-time assessment of fermentation and colonic gas physiology. The utility of breath gas analysis could be expanded substantially with greater knowledge of the microbiology of gas production relative to diet or disease states. Abbreviations: CH 4, methane; H2, hydrogen; SIBO, small intestinal bacterial overgrowth. activity. Early studies by Christl confirmed the predict- Microbial guilds in the hydrogen economy able relationship between H2 and CH4 production and Host colonic cells derive energy from aerobic respiration, breath excretion of these two colonic gases.4 Likewise, in which nutrients are fully oxidized in mitochondria, breath testing remains a validated means of determin- with oxygen serving as the terminal electron acceptor. ing intestinal transit time 31 and distinguishing two Similar to fermentation, the metabolic reactions involved fairly stable phenotypes, excretors versus non­excretors in respiration are catabolic and based on redox reactions of CH4.32 (that is, oxidation of one molecule coupled to the reduc- A particularly provocative finding regarding the dis- tion of another). In fermentation—the anaerobic process tinction between CH4-excretors and non-CH4-excretors by which most colonic microbes gain energy—nutrient is the variation in percentages of CH4-excretors among substrates are incompletely oxidized and the reduced fer- different ethnic and racial populations (range 34–87%; mentation products serve as terminal electron acceptors. summarized in Levitt et al.33). Black Africans tend to be In such cases, the amount of energy (ATP) that can be highly methanogenic compared with North American produced depends on the difference in redox potential white or Asian individuals (including those of Indian between the substrate and the reduced end products. origin).34–36 Moreover, the percentage of a given popula- Another distinction is that in fermentation the reduced tion that is methanogenic seems to remain stable over pyridine (NADH) and flavin (FADH) nucleotides must time. 32,33 However, distinguishing the multifactorial be reoxidized to maintain redox balance, a reaction that interactions between host ethnicity, environment and is the primary source of H2 in the colon (Figure 2). diet is difficult. Segal et al.35 reported that the percentage Accordingly, the production of H 2 by hydrogeno- of CH4-producers was lower in urban (72%) than rural genic microbes is crucial to the efficiency of fermen- black Africans (84%). Moreover, O’Keefe et al.37 found tation. However, H2 accumulation would rapidly lead that the H2 and CH4 breath emission patterns of African to a H2 partial pressure that would thermodynamically Americans and white Americans—both consuming a restrict further fermentation. Such an outcome typi- typical Western diet—were more similar to each other cally is prevented by the simultaneous oxidation of H2 than the excretion patterns of native Africans consuming by three groups of hydrogenotrophic (H 2-utilizing) a maize-based diet low in animal-based protein and high microbes that conserve energy through anaerobic respi- in resistant starch. ration: reductive acetogens, methanogenic archaea and NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY ADVANCE ONLINE PUBLICATION  |  3 © 2012 Macmillan Publishers Limited. All rights reserved REVIEWS Glucose and other monosaccharides Ferredoxin-H2 2H+ Polysaccharides NAD+ NADH: ferredoxin Hydrogenase NADH+ oxidoreductase Dietary and host-derived H2 carbohydrates Ferredoxin 1 CO2 Pyruvate Succinate Propionate Pyruvate:formate lyase Formate Formate hydrogen lyase complex: formate Ferredoxin dehydrogenase and 2H+ hydrogenase 3 Pyruvate:ferredoxin Hydrogenase 3 oxidoreductase CO2, H2 2 H2 Ferredoxin-H2 Acetyl-CoA Acetyl-P Acetoacetyl-CoA Acetate Butyryl-CoA Butyrate Figure 2 | Biochemical pathways of H2 production from bacterial fermentation. The primary source of H2 derives from (1) the reoxidation of reduced pyridine and flavin nucleotides. This pathway is inhibited by a high partial pressure of H 2, whereas the other two pathways of H2 production are not. These two pathways are (2) the cleavage of pyruvate to formate and subsequent metabolism by formate hydrogen lyase (primarily Clostridia), and (3) the generation from pyruvate through the activity of pyruvate: ferredoxin oxidoreductase and hydrogenase (primarily Enterobacteria). These biochemical pathways were defined mainly from in vitro cultivation of fermentative bacteria. The availability of genomics and molecular-based methods provides opportunities for further metabolic characterization of colonic H2 production in situ. Abbreviations: CO2, carbon dioxide; H+, hydrogen ion; H2, hydrogen. sulphate-reducing bacteria (SRB; Figure 3). Microbial Of note, the colon is composed of different sites with hydrogenotrophy—together with excretion in flatus various physiological and chemical characteristics, and breath (15–20% for each route)—results in effi- which have a role in shaping microbial communities.50 cient removal of H2, and shifts fermentation to more Specifically, it has been suggested that the more acidic oxidized end products, hence, increasing the energy right colon is mainly colonized by reductive acetogens, yield of fermentative microbes.4 Although the existence whereas methanogenic archaea would thrive better in the of H2 disposal mechanisms is crucial to colonic fermen- more neutral pH of the distal colon.51–54 Indeed, coherent tation, the extent to which the three metabolic guilds gradients of microbial abundance were demonstrated by (reductive acetogens, methanogenic archaea, SRB) co- a 2011 molecular survey,39 but all microbial communi- occur in the healthy human colon has been the subject ties were ubiquitous throughout the colon. However, of debate. Reductive acetogenesis has been presented this observation does not preclude regional differentia- as a facultative H2 disposal pathway owing to sulphate tion, which could also occur on a microscale, enabling reduction and methanogenesis being thermodynami- acid-intolerant methanogens to grow in microniches cally more favourable reactions for this requirement exhibiting neutral pH, in the right colon for example. of fermentation.38 In addition, models of competitive exclusion between methanogenic archaea and SRB in Hydrogen gas producers the human colon have been suggested to explain the In addition to the reoxidation of reduced pyridine and observed disparities in detectable breath CH 4 excre- flavin nucleotides, H2 can be produced by cleavage of tion and in differential outcomes of culture-­dependent pyruvate to formate and subsequent metab­olism by approaches.13,39–48 Culture-based and molecular-based formate hydrogen lyase, or by generation from pyruvate studies have demonstrated that SRB persistently colonize through the activity of pyruvate:ferredoxin oxidoreduc- the healthy human colon.39–41,49 tase and hydrogenase (Figure 2).55 With H2 production 4  |  ADVANCE ONLINE PUBLICATION www.nature.com/nrgastro © 2012 Macmillan Publishers Limited. All rights reserved  REVIEWS Methanogens pH2 H2 CH4 Methanosphaerae Methanobrevibacter Ruminococcus spp. SCFA stadtmanae smithii Colonocytes, pH2 blood, Reductive acetogens breath, H2 Acetate flatus Acetate Roseburia spp. Ruminococcus spp. Clostridium spp. H2S H2 Proinflammatory Genotoxic pH2 Blautia hydrogenotrophica Clostridium spp. SRB Bacteroides spp. Desulfovibrio spp. Desulfobacter spp. Desulfobulbus spp. Desulfotomaculum spp. Hydrogenogens Hydrogenotrophs Figure 3 | H2 gas is inherently produced during microbial fermentation in the human colon. The accumulation of H 2 leads rapidly to a pH2 that thermodynamically restricts further fermentation. Accordingly, three groups of hydrogenotrophic (H 2- utilizing) microbes possess the crucial ability to simultaneously oxidize H 2, which lowers pH2 enabling fermentation to proceed. Given the centrality of H2 production in fermentative pathways, this metabolic feature is almost certainly widespread among colonic bacteria. The genera depicted are among those that have been characterized in vitro for H2 production. Hydrogenotrophic microbes include methanogenic archaea, reductive acetogens and SRB. Abbreviations: CH 4, methane; H2, hydrogen; pH2, partial pressure of hydrogen; SCFA, short-chain fatty acids; SRB, sulphate-reducing bacteria. All images shown are microscopy images with the exception of a schematic representation of Desulfobacter spp. Ruminococcus spp. image courtesy of Harry Flint and Sylvia Duncan. Desulfovibrio spp. image courtesy of Grahame Bradley. Clostridium spp. and Bacteroides spp. images both courtesy of CDC. Desulfobulbus spp. image taken from Pagani, I. et al. Stand. Genomic Sci. 4, 100–110 (2011), which is published under an open-access license by the Genomic Standards Consortium. Desulfotomaculum spp. and Roseburia image used with permission from the Society for General Microbiology © Fardeau et al. Int. J. System. Bacteriol. 45, 218–221 (1995) and Duncan, S. H. et al. Int. J. System. Evol. Microbiol. 56, 2437–2441 (2006), respectively. M. stadtmaniae and B. hydrogenotrophica image used with permission from Springer © Miller, T. L. & Wolin, M. J. Arch. Microbiol. 141, 116–122 (1985) and Bernalier et al. Arch. Microbiol. 166, 176–183 (1996), respectively. M. smithii image used with permission from National Academy of Sciences © Samuel, B. S. et al. Proc. Natl Acad. Sci. USA 104, 10643–10648 (2007). being integral to microbial fermentation, a broad levels of H2. Thus, it seems that colonic H2 is produced assemblage of hydrogenogens must exist in the human mainly by members of the Firmicutes and much less by colon—probably most abundant in the right colon, the members of the Bacteroidetes. colonic region with the greatest extent of microbial fer- Commonly, H2 production has been characterized mentation.50 Few studies have focused on the phylo- by breath testing; however, data indicating the extent genetic diversity of these microbes and the metabolic to which such tests accurately reflect the abundance or niches they occupy. Among abundant bacterial genera activity of hydrogenogens in luminal or mucosal micro- detected in the colon, several strains of Roseburia biota are scarce. A proof-of-principle description of a spp.56,57 and Ruminococcus spp.44,45 produce substantial protocol for selective enrichment of hydrogenogens concentrations of H2 in vitro. Other prominent colonic offers promise of successful culture-based approaches for bacteria known to produce H 2 include Anaerostipes application to the colonic environment.63 A more readily caccae, 46 Clostridium spp. 47,48 Eubacterium rectale, 58 available approach would be the direct characterization Bacteroides spp.43,59 and Victivallis vadensis.60 Conversely, and quantification of microbial H2 production by target- Akkermansia muciniphila61 and Faecalibacterium praus- ing the hydrogenase enzymes involved in H2 production nitzii 62 were clearly shown to produce no detectable (Figure 2). Functional gene approaches are particularly NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY ADVANCE ONLINE PUBLICATION  |  5 © 2012 Macmillan Publishers Limited. All rights reserved REVIEWS useful for measuring the DNA-based abundance or the Methanogenic archaea RNA-based activity of phylogenetically diverse metabolic Colonic methanogenic archaea derive all (or most) of groups such as fermentative hydrogenogens. Of the two their metabolic energy from methanogenesis by reducing main types of hydrogenases, [Fe-Fe]-hydrogenases are CO2 or methanol to CH4 using H2 or formate as electron involved primarily in H2 production and can be used donors.79 Substantial interindividual differences exist as molecular targets.64 The utility of this approach was in colonic methanogenesis. A threshold value of 1 × 108 demon­strated for the characterization of H2 producers methanogens per g of stool for CH4 to be detected in in an acidic fen65 and earthworm gut,66 and thus offers breath was defined by Miller and Wolin80 and confirmed promise for unravelling greater detail on the extent to in a subsequent study in which breath CH4 excretors har- which the abundance and activities of hydrogenogens boured an average of 1 × 109 CFU per g of methanogens vary among individuals and in response to diet or disease. in stool, whilst nonexcretors harboured ~1 × 104 CFU per g.49 Chassard et al.81 demonstrated that the struc- Reductive acetogens ture of the cellulose-degrading bacterial community in The reductive acetogens are a group of obligately individuals differs according to their CH 4 status. The anaerobic bacteria that utilize the acetyl-CoA (Wood– dominant cellulose degraders isolated from non-CH4- Ljungdhal) pathway to synthesize acetate from CO2 and excreting individuals belonged mainly to Bacteroidetes; H2.67 Colonic acetogenesis was first demonstrated using CH4-excreting individuals harboured predominantly slurries from human stool,68 and later cultivation-based Firmicutes. Methanogens were also observed to co- studies estimated that the number of acetogens ranged occur predominantly with members of the Clostridiales from 1 × 10 2–1 × 10 8 colony-forming units (CFUs) order (Firmicutes) in individuals harbouring abundant per g wet human stool. 69,70 Acetogens isolated from methano­gens.82 This same study confirmed that host human stool were related to the genera Ruminococcus, genotype influences methanogen carriage, as the con- Clostridium or Streptococcus.70–72 The gene sequences of cordance rate for carriage of methanogens was markedly formyl tetra­hydrofolate synthetase (fhs) and acetyl-CoA greater in monozygotic adult twin pairs than in di­zygotic synthase (acs) are highly conserved among acetogens twins. However, carriage of methanogens between and thus serve as useful functional gene targets. 48,73,74 mother and daughter was discordant. 82 Thus, host Analysis of fhs sequences amplified from human stool genotype and various environmental factors are among identified Blautia producta (formerly Ruminococcus the potential determinants of persistent colonization productus) as the predominant acetogen, and detected by methanogens. several fhs sequences that had not been identified previ- The two methanogenic species isolated from ously.75 Both fhs and acs were persistently detected in the human colon, Methanobrevibacter smithii and colonic biopsies from 25 healthy individuals, with values Methanosphaera stadtmanae, have different biochemi- ranging from 1.8 × 103 to 8.8 × 106 and from 9.8 × 103 to cal characteristics. M. smithii converts CO2 and H2 to 3.8 × 107 gene copy numbers per g of tissue, respectively.39 CH4, but M. stadtmanae uses H2 to reduce methanol to It should be noted, however, that the presence of these CH4.83,84 To date, studies (using both culture-based and functional genes does not necessarily correlate with molecular-based approaches) indicate that M. smithii is reductive acetogenesis. the predominant methanogen in the human colon;85–90 Acetogenesis has been postulated to be a less important M. stadtmanae has been isolated from the human intes- hydrogenotrophic pathway in the colon than methano­ tinal tract at a lower abundance. 83 Several different genesis and sulphate reduction4 because both the con- phylo­types closely related to M. smithii, M. stadtmanae, version of H2 and CO2 into CH4 and sulphate reduction M. oralis or Methanosarcinales have been identified using are thermodynamically more favourable than reductive molecular fingerprinting studies targeting 16S ribo­ acetogenesis.38,76 Nevertheless, a metagenomic study of somal DNA and functional gene coenzyme M reductase the in vivo metabolic potential of human gut acetogens (mcrA).89–91 The mcrA gene was persistently detected concluded that acetogenesis was the most prevalent with values ranging from 3.0 × 102 to 4.5 × 109 in colonic pathway of H2 disposal in the human colon.77 This model biopsy samples from 25 healthy individuals.39 These data is supported by a previous study that used radioisotope further confirm initial observations that breath CH4 con- analysis to quantify the fraction of acetate derived from centration reflects the relative abundance or activity of the reduction of CO2 by H2 in stool suspensions from two colonic methanogens, and not the presence or absence individuals.78 However, the model conflicts with the the- of this hydrogenotrophic group. oretical consideration of the thermo­dynamics of H2 uti- lization among hydrogenotrophs as well as experimental Sulphate-reducing bacteria evidence of the relative prevalence of methanogenic SRB are a diverse group sharing the ability to use sul- archaea and SRB in the human colon. Without question, phate as a terminal electron acceptor for respiration, with polyphyletic reductive acetogens are more metabolically the concomitant production of H2S. Colonic SRB gener- versatile than methanogenic archaea or SRB, and defin- ally use H2 as their electron donor, but electrons can also ing the nature of their interactions with these two groups be provided from the oxidation of organic compounds, of hydrogenotrophic microbes and the extent to which such as lactate.92 SRB are ubiquitously present in the they contribute to H2 disposal in CH4-excretors versus human intestinal mucosa40,41,93 and have been enumer- nonexcretors needs to be resolved. ated from human stool within the range 1 × 103–1 × 1011 6  |  ADVANCE ONLINE PUBLICATION www.nature.com/nrgastro © 2012 Macmillan Publishers Limited. All rights reserved  REVIEWS Table 1 | Studies on the link between IBS and colonic H2 Study No. of participants Major findings Levitt & Donaldson (1970) 26 55 Abnormal breath H2 and CH4 in carbohydrate malabsorption Metz et al. (1975)194 24 IBS linked with lactose intolerance Metz et al. (1976)195 17 IBS symptoms linked to SIBO Rhodes et al. (1979) 28 64 Lactulose breath testing for SIBO El Oufir et al. (1996)196 8 Altering intestinal transit times leads to a change in bacterial abundance and activity King et al. (1998)113 12 Increased volume and rate of H2 excretion in patients with IBS Pimentel et al. (2000)116 202 Reduction of SIBO with antibiotics reduces IBS symptoms Pimentel et al. (2003) 153 551 CH4 linked to constipation-predominant IBS Pimentel et al. (2004)197 18 Patients with methanogenic IBS have reduced postprandial serotonin levels relative to patients with hydrogenogenic IBS Dear et al. (2005)198 12 Reduction of fermentation with metronidazole or exclusion diet improves IBS symptoms Abbreviations: CH4, methane; H2, hydrogen; SIBO, small intestinal bacterial overgrowth. bacteria per g.34,49,94 In vitro, colonic SRB are metaboli- in particular (Table 1).109,110 Malabsorption of carbo­ cally flexible and can oxidize a variety of short-chain hydrates from the intestine and SIBO are commonly, but fatty acids.94,95 Detection of SRB in the stool of infants inconsistently, associated with IBS, and their relevance <1 month indicates that these bacteria are facultative H2 to the cause of the disorder has been controversial.111 utilizers.40,96 A range of nutritionally and physiologically In particular, CH4 has been linked to decreased colonic distinct SRB has been detected in human stool.34,94,97,98 transit time in patients with IBS‑C; meanwhile H2 accu- The genes encoding adenosine 5'-phosphosulfate reduc- mulation, possibly due to a failure of colonic microbes tase (apsr1) and dissimilatory sulphite reductase (dsrAB), to dispose of H2 produced via fermentation, has been which are enzymes in the sulphate reduction pathway, postulated to account for the bloating and pain that are also useful molecular markers owing to their highly often distinguish IBS‑C from chronic constipation.110,112 conserved nature and congruence with the evolutionary Using whole-body calorimetry, King et al.113 found that history of SRB.99–102 However, few studies have been pub- patients with IBS excreted substantially more H2 overall lished that have examined the diversity or abundance of than healthy individuals. Although H2 plus CH4 excre- human colonic SRB using molecular-based techniques. tion was slightly higher in those with IBS than healthy The functional gene dsrAB was persistently detected with controls, the difference in volume was not statistically values ranging from 1.8 × 102–1.4 × 109 copy numbers in significant; however, the rate of excretion of both gases colonic biopsy samples from 25 healthy individuals.39 was 400% higher in patients with IBS than controls. Furthermore, in the same set of biopsy samples, four Dietary changes that improved IBS symptoms decreased different SRB genera identified previously by cultur- gas excretion levels, particularly H2.113 ing 34 were consistently detected with relevant 16S rRNA Because bloating and abdominal pain are suggested gene probes.39 to result from carbohydrate malabsorption and subse- quent bacterial fermentation in the colon, clinicians and Colonic gases and human disease scientists have sought to correlate elevated breath levels Links with IBS of H2 produced by colonic bacteria during fermentation IBS is categorized as a functional intestinal disorder and with IBS.114,115 Several studies have measured increases in afflicts ~14% of the US population;103 it is the first-listed breath H2 concentration in patients with IBS,116 as well diagnosis in >1.6 million individuals according to the as premature peaks in breath H2, implicating SIBO as a National Hospital Ambulatory Medical Care Survey cause of abdominal pain and bloating. However, evidence for 2004–2005.104 However, based on findings from a exists that early H2 peaks associated with IBS reflect dif- 2005 survey of 5,009 individuals, Hungin et al.103 esti- ferences in small-bowel transit time rather than SIBO.117 mated that >76% of IBS sufferers may go undiagnosed. Quigley 118 suggests that IBS owing to SIBO might present IBS subtypes include constipation-predominant IBS a distinct category of this disorder. Nonetheless, reliable (IBS‑C), diarrhoea-predominant, and mixed or unspeci- data verifying a link between colonic microbiota, H 2 fied.105,106 Primary symptoms—in addition to diarrhoea metabolism and IBS are sparse.119 or constipation in 30–40% of patients—include pain, Finally, a correlation was observed between high bloating and/or abdominal distension, flatulence and breath CH4 levels and the occurrence of motility dis­ belching.107,108 Despite its prevalence and severity, no orders.120,121 As demonstrated in human and mammalian confirmatory diagnostic test exists for IBS. model systems, high levels of CH 4 are correlated with The role of gas byproducts of microbial fermenta- decreased intestinal motility;121 however, it has not been tion has been implicated in IBS in general, and IBS‑C confirmed that this increase in breath gases is associated NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY ADVANCE ONLINE PUBLICATION  |  7 © 2012 Macmillan Publishers Limited. All rights reserved REVIEWS with increased abundance of colonic methanogens. sulphide concentrations was observed between patients For example, increased CH4 production could result with ulcerative colitis who were treated with 5‑ASA and from increased bacterial H2 production. Global and noncolitic individuals,141 whereas stool sulphide concen- deep microbial analysis of stool samples from patients trations were markedly higher, by comparison, in patients with IBS demonstrated the presence of methanogenic with ulcerative colitis who were not administered archaea in a higher percentage of those with IBS‑C 5‑ASA.139 Aminoglycoside antibiotics also inhibit SRB than in healthy controls or patients with diarrhoea- growth and are of therapeutic benefit in active ulcerative predominant or alternating IBS. However, among those colitis.136 Further supporting a role for H2S in ulcerative patients harbouring methanogens, their abundance was colitis is the observation that SRB were found in surgi- fourfold greater in healthy individuals than in either the cally constructed ileoanal pouches of patients with ulcer- IBS group as a whole or those with IBS‑C.122 In addi- ative colitis but not in pouches of patients with familial tion, patients with IBS who excreted CH 4 produced adenomatous polyposis.141 Moreover, H2S production in lower levels of serotonin in response to glucose than ulcerative colitis pouches was 10 times greater than that those excreting primarily H2.123 Serotonin in IBS has in familial adenomatous polyposis pouches.142 In addi- been linked to abnormal gut motility as well as visceral tion, the severity of pouchitis is positively correlated hypersensitivity; however, the potential success of treat- with stool concentrations of H2S,143 possibly reflecting ments capitalizing on this relationship has been hindered a pathogenic role for this gas. by serious adverse effects.124–127 This lack of knowledge Early studies used culture-based approaches to on the relationship between measures of microbial abun- examine colonic SRB in patients with IBD (Table 2). dance and activities relative to breath gas concentrations The numbers of SRB and rate of sulphidogenesis were highlights the need for more systematic investigation of greater in those with ulcerative colitis than in healthy host–microbe aspects of H2 metabolism. controls.144,145 In another in vitro study, production of H2S has been shown to modulate peripheral noci­ H2S from stool of patients with ulcerative colitis was ceptive (pain-related) signals.128 As abdominal pain is a 3–4 times greater than that from controls.146 Molecular- primary symptom of IBS, these data implicate a poten- based techniques have been used to evaluate the tial, direct role for colonic H2S. Indeed, a pro-nociceptive prevalence and diversity of SRB species (Table 2). For effect has been suggested from mouse model studies that example, Desulfovibrio piger was more abundant in could support this hypothesis.129,130 Nevertheless, anti­ stool of patients with IBD than in healthy individuals nociceptive effects have also been reported,131 and, thus, or in patients with other gastrointestinal symptoms.147 the importance of the role of H2S in abdominal pain in However, Fite and coworkers40 reported that patients IBS remains controversial. with active ulcerative colitis did not harbour more Desulfovibrio spp. than healthy controls in either stool Links with IBD or rectal mucosal samples (as measured by quantitative A role for bacterial-generated H 2 S in IBD aetio­ PCR). Thus, either increased sulphidogenic activities or pathogenesis has support from both clinical and reduced sulphide detoxification in the colonic epithelium experimental studies (Table 2). H 2S is highly toxic to might explain the increased H2S concentrations in these colonocytes and impairs their metabolic function, espe- patients and their potential inflammatory impact. cially butyrate oxidation.132,133 In aqueous solutions, H2S Indirect support for a role of H 2S and SRB in the dissociates into hydrosulphide anion (pKa 7.04) and sul- aetiology of ulcerative colitis (but not Crohn’s disease) phide ion (pKa 11.96).134 Generally, sulphide exists in the is the observed increased activity of mucin sulphatase, human colon in the volatile, highly toxic undissociated an enzyme that cleaves sulphate groups from mucosal form (H2S), which is quickly absorbed by the mucosa or sulphomucins.148 Additionally, in most patients, fluc- passed as flatus.2 The majority of sulphate (>90%) dis- tuations in stool sulphatase activities correspond to appears during passage through the colon of individu- clinical disease activity.148 Therefore, a model in which als lacking SRB; thus, a number of colonic processes, in increased sulphatase activity results in increased avail- addition to sulphate reduction by resident bacteria, must ability of endogenous sulphate for H2S production by compete for sulphate.135 SRB might contribute to perpetuation of the disease. Anionic sulphide concentrations were elevated in This model is supported by the finding of an associa- the colon of patients with ulcerative colitis. 136 These tion between the abundance of sulphomucins and the patients also ingested more protein, and thereby more quantity of several SRB genera in the colonic mucosa of sulphur amino acids, than healthy control individuals.137 healthy individuals.149 Removing foods rich in sulphur amino acids (milk, eggs A few, but consistent, reports indicate that the preva- and cheese) has proven therapeutic benefits in those with lence of the methanogenic phenotype is markedly lower ulcerative colitis.138 However, other studies examining in patients with Crohn’s disease or ulcerative colitis than SRB-related variables have not confirmed a possible link in healthy individuals. However, these intriguing find- with IBD. These conflicting observations might reflect ings have received limited attention, and it is not known the common use of 5‑aminosalicylic acid (5-ASA), whether the potentially reduced prevalence of CH4 excre- which has a proven therapeutic value for the treatment tion in IBD is a cause or consequence of, for example, of ulcerative colitis136 but also diminishes H2S production reduced transit time or pH. Breath CH4 was detected in by colonic bacteria.139,140 Indeed, no difference in stool 44% of healthy white individuals but absent in patients 8  |  ADVANCE ONLINE PUBLICATION www.nature.com/nrgastro © 2012 Macmillan Publishers Limited. All rights reserved  REVIEWS Table 2 | Studies on the link between IBD, SRB and H2S Study No. of Major findings participants Gibson et al. (1991)97 30 Faecal slurries from patients with ulcerative colitis produce markedly higher rates of H2S than healthy controls Roediger et al. (1993)132 31 H2S impairs butyrate acquisition by colonocytes Christl et al. (1996) 167 10 NaHS induces mucosal hyperproliferation Levine et al. (1996)199 100 No difference in SRB stool carriage between populations at high risk of IBD (Ashkenazi Jews) and control lower-risk populations Roediger et al. (1997)136 NA* Therapeutic value of 5-ASA for ulcerative colitis might result from SRB inhibition Pitcher et al. (2000)139 49 5-ASA inhibits H2S production from SRB and might confound correlation between ulcerative colitis and SRB Zinkevich et al. (2000)41 74 SRB cultivated successfully from 92% of patients with ulcerative colitis and only 52% of noncolitic ones. SRB-specific PCR successful from all biopsy samples, indicating differences in abundance Duffy et al. (2002)200 25 SRB colonize pouches formed for ulcerative colitis, but not for familial adenomatous polyposis Kleessen et al. (2002)93 38 Biopsy samples from patients with ulcerative colitis always positive for SRB detection by FISH; only some patients with Crohn’s disease had positive samples Loubinoux et al. 151 Higher prevalence of cultivable SRB (68%) and Desulfovibrio piger DNA (55%) in stool from patients with IBD (2002)147 compared with healthy individuals or patients with other digestive diseases (24–37% for SRB and 12–25% for D. piger, respectively) Bullock et al. (2004)201 12 No difference in SRB abundance detected between patients with active ulcerative colitis and those in remission Ohge et al. (2005)143 50 Stool H2S markedly higher in patients with ulcerative colitis who have a recent history of pouchitis than patients with ulcerative colitis who have an older or no history of pouchitis Smith et al. (2005)202 14 SRB only colonizes ulcerative colitis pouches, but not adenomatous polyp pouches Bambury et al. (2008)203 21 Ulcerative colitis pouches characterized by high sulphomucin expression that correlates with high SRB colonization Coffey et al. (2009)204 NA* Model for ulcerative colitis pouchitis suggested in which an increase in sulphomucin production enables SRB development and H2S-induced inflammation Lim et al. (2009)205 20 Two Desulfosporosinus sequences and other uncultivated Proteobacteria (potential SRB) detected only in inflamed colonic pouches Rowan et al. (2010)206 39 Desulfovibrio absolute and relative abundance increased in patients with acute and chronic ulcerative colitis compared with healthy controls Verma et al. (2010)207 149 SRB (and Methanobrevibacter smithii) genes more abundant in patients with ulcerative colitis and those with Crohn’s disease compared with healthy controls Strauss et al. (2011)208 56 H2S-producer Fusobacterium nucleatum isolated more often from IBD biopsy samples than from healthy controls. F. nucleatum strains originating from inflamed biopsy tissue were more invasive in a Caco-2 invasion assay than strains isolated from healthy tissue *Review article. Abbreviations: 5‑ASA, 5‑aminosalicylic acid; FISH, fluorescence in situ hybridization; H2S, hydrogen sulphide; NA, not applicable; NaHS, sodium hydrogen sulphide; SRB, sulphate-reducing bacteria. with Crohn’s ileitis.22,150 McKay et al.151 reported a 13% to healthy controls, statistical significance was observed prevalence of CH4 excretion in patients with Crohn’s only for those with ulcerative colitis. disease and 15% in those with ulcerative colitis compared with 54% in healthy controls. Peled and coworkers152 Links with colorectal cancer found that among healthy individuals, 50% produced Multiple lines of evidence for a possible association of CH4, whereas breath CH4 was detected in only 6.1% of both methanogens and SRB with sporadic colorectal patients with Crohn’s disease and 31.4% of patients with cancer (CRC) have been reported (Table 3). In the 1970s ulcerative colitis. A 2003 study 153 compared the excre- and 1980s, numerous studies reported a higher preva- tion of either H2 or CH4 alone to combined excretion lence of methane CH4 excretion among patients with of these two gases following a lactulose breath test. The CRC compared with healthy individuals and, in some predominant gas excreted by patients with IBD was cases, patients with other gastrointestinal disease.154–159 H2 alone (76 of 78 individuals with Crohn’s disease or However, subsequent studies did not find major dif- ulcerative colitis). By contrast, breath CH4 was detected ferences in CH4 status between patients with CRC and as the predominant gas in only two of 78 individuals with healthy individuals,160,161 and the use of the breath test IBD in this study.153 To date, only a single report, using a was apparently abandoned as a possible CRC diagnostic molecular-based approach, has compared the incidence tool. It was suggested that observations of higher breath and density of colonic methanogens in healthy individu- CH4 levels in patients with CRC might have resulted als versus patients with IBD. Targeting the mcrA gene, from reduced transit time owing to at least partial Scanlan et al.90 reported that although the abundance of obstruction by tumour tissue.154 Karlin et al.155 found methanogens was reduced in both IBD groups relative that, among a group of 55 patients with unresected CRC NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY ADVANCE ONLINE PUBLICATION  |  9 © 2012 Macmillan Publishers Limited. All rights reserved REVIEWS Table 3 | Studies on the link between CRC, SRB and H2S Study No. of study Major findings participants/ patient samples Kanazawa et al. (1996)171 27 Higher H2S concentration in high-risk patients compared with controls Deplancke et al. (2003)165 5 H2S induces cell-cycle entry in rat intestinal epithelial cells Attene-Ramos et al. (2006) 168 NA* Physiological H2S concentrations genotoxic to mammalian cells Ramasamy et al. (2006)172 NA Levels of sulphide-detoxifying enzymes in the human colon decreased in patients with cancer Attene-Ramos et al. (2007)169 NA* H2S induces direct radical-associated DNA damage Balamurugan et al. (2008)173 46 No major differences in abundance of Desulfovibrio between stool from patients with CRC and healthy individuals Scanlan et al. (2009)209 90 No major differences in abundance of Desulfovibrio between stool from patients with CRC and healthy individuals Attene-Ramos et al. (2010)170 NA* Physiological H2S concentrations genotoxic to colonic epithelial cells Cai et al. (2010)164 NA* H2S induces human colon cancer cell proliferation Castellarin et al. (2011)176 99 H2S-producer Fusobacterium nucleatum DNA sequences over-represented in CRC tumour tissues Kostic et al. (2011)175 95 H2S-producer Fusobacterium nucleatum sequences over-represented in microbial metagenomes from CRC tumour tissues Marchesi et al. (2011)174 6 H2S-producer Fusobacterium nucleatum sequences more abundant in 16S pyrosequencing from CRC tumour tissues than in healthy tissue *In vitro studies. See references for details. Abbreviations: CRC, colorectal cancer; H 2S, hydrogen sulphide; NA, not applicable; SRB, sulphate-reducing bacteria. in varying regions of the colon and rectum, no difference from patients with CRC (and ulcerative colitis). In a was observed in the frequency or amount of CH4 excre- limited study targeting only a few microbial taxa in stool tion for the groups; however, of those with more distal from patients with CRC and healthy individuals, major cancer (descending or sigmoid), the likelihood of excret- differences were not observed for Desulfovibrio abun- ing breath CH4 was twice as high as for patients with dance.173 In 2011, Marchesi et al.174 described a potential proximal colon or rectal cancer. Holma et al.162 reported CRC tumour-associated microbiome, based on high- an extensive assessment of methanogenic status in CRC. throughput sequencing. Intriguingly, the highest H 2 Again, breath and stool CH4 levels were similar in healthy producers belong to genera over-represented in colonic individuals and those with CRC. However, patients with tumour tissue (Eubacterium and Roseburia). This study, right-sided CRC exhibited lower methanogenesis, lower as well as two subsequent reports, also consistently stool pH and increased abdominal discomfort compared detected an enrichment of Fusobacterium spp. in colonic with patients with left-sided CRC.162 tumour tissue, a genus known to produce H2S via deg- H 2S can damage the intestinal epithelium leading radation of sulphur-containing amino acids.174–176 Thus, to chronic inflammation,132,136,163 as well as perturbing it is possible that the observed association of increased the balance between cellular proliferation and apop- H2S with CRC derives from cysteine fermentation rather tosis.164–167 At concentrations similar to those found in than sulphate respiration. the human and mouse intestine, the H2S donor sodium Endogenously produced H2S has been demonstrated hydrosulphide produced genomic DNA damage in from a wide range of tissues, including the gastro­ Chinese hamster ovary and human HT29-Cl.16E intestinal tract,177 and this molecule is receiving a lot of colonic epithelial cells when DNA repair was inhibited.168 attention as a possible intracellular ‘gaseous trans­mitter’.178 Sodium hydrosulphide also induced DNA damage in the Numerous examples exist whereby the exo­genous admin- absence of cellular metabolism, and this damage was at istration of H2S donors (such as sodium hydrosulphide) least in part produced by free radicals.169 A subsequent exerts anti-inflammatory effects in a wide range of in vitro study confirmed the genotoxic properties of sodium and in vivo settings.179 However, at present, a large dis- hydrosulphide in nontransformed human intestinal crepancy remains between the concentrations of H2S in epithelial cells with intact DNA repair pathways, and tissue versus those needed for alteration of tissue function demonstrated that H 2S modulates the expression of in vitro.180 One hypothesis that has not been tested is a genes involved in cell-cycle progression and triggers both direct modulation of the gut microbiota by H2S rather inflammatory and DNA repair responses.170 than physiological effects. H2S is known to be toxic to Kanazawa171 measured higher stool H2S levels in indi- microbes in general and also to help maintain anaerobic viduals with high risk of CRC than in healthy controls. status.181,182 Thus, it is possible that the suggested benefi- Ramasamy 172 found that thiosulphate sulphotransferase, cial effects of H2S might actually reflect their shaping of an enzyme purported to be involved in H2S detoxifica- a more beneficial microbiome, possibly by inhibiting the tion, was present in lower abundance in biopsy tissue growth of pathogenic facultative anaerobes. 10  |  ADVANCE ONLINE PUBLICATION www.nature.com/nrgastro © 2012 Macmillan Publishers Limited. All rights reserved  REVIEWS What is clear is that SRB are tightly associated with the reduced colitis (induced by dextran sodium sulphate) in colonic mucosa and presumably their presence provides a rat model.192 Evidence that H2-infused water and H2 gas a source of H2S that directly influences the colonic epi­ exerted beneficial effects in animal models of, or patients thelium. Hence, numerous host and microbial compo- suffering from, obesity, diabetes mellitus and metabolic nents might fit multifactorial models that could explain syndrome expands the potential clinical relevance of the gene–environment interactions that predispose to H2 economy.193 Thus, it would seem that the H2 endo­ sporadic CRC. genously produced by the resident microbiota might exert a similar protective antioxidant role in the healthy Links with obesity colon, and that decreases in the net production of H2 Obesity has been hypothesized to correlate with elevated might increase the risk of inflammatory or metabolic dis- levels of colonic CH4 and H2.86 This hypothesis is based eases. It is intriguing to consider whether manipulation on the assumption that increased methanogenesis would of microbial H2 metabolism, either through enhanced improve fermentation efficiency, resulting in increased production or diminished utilization, might provide production of short-chain fatty acids, which potentially a novel means of regulating colonic homeostasis. At promotes adipogenesis by the host. The idea was first present, technological limitations preclude interroga- suggested by observations that methanogens were more tion of microbial H2 metabolism at temporal and spatial abundant in homozygous ob/ob mice than in their non- scales relevant to the colonic mucosal ecosystem. obese heterozygotic littermates and wild-type controls.183 Subsequently, an intriguing study detected markedly Conclusions higher numbers of methanogenic archaea in obese indi- Despite long-standing evidence demonstrating the impor- viduals than in normal-weight individuals or patients after tance of microbial H2 metabolism, minimal efforts have gastric bypass.86 Alternatively, four reports demonstrate been exerted to manipulate relevant host–microbe interac- a reduced number of CH4 excretors among obese indi- tions to affect colonic health. The few studies that demon- viduals compared with lean individuals,184 a lower level strate the ease with which colonic gases can be regulated of M. smithii in obese individuals,185,186 and greater abun- by dietary substrate offer promise for relatively innocu- dance of methanogens in those with anorexia compared ous strategies to alter the balance between hydrogenogenic with obese and lean individuals.187 Clearly, much addi- and hydrogenotrophic microbes, and thereby broaden tional work is needed to determine the extent to which preventative and therapeutic options for managing mul- colonic H2 metabolism might influence the develop­ment tiple colonic and possibly metabolic dis­orders. However, of obesity. Determining the relationship between colonic for this promise to be fulfilled, additional research will microbial populations and breath gas measurements in be required to better understand the microbial and mol­ obese individuals, populations at risk of obesity, and in ecular bases of colonic H2 production and utilization as response to diet would help to clarify this matter. well as the effectiveness of the H2 breath test as a reliable readout of microbial metabolic activities. Such findings Hydrogen as a therapeutic gas are expected to reveal explanations for the marked varia- Perhaps the most exciting link between H2 and human tion among individuals in both host and microbial aspects disease is emerging evidence that this microbial- of H2 metabolism, thus enabling these parameters to be derived gas has potent antioxidative, antiapoptotic and utilized for novel biomarkers of health and disease. anti-inflammatory activities in a wide range of disease Review criteria models.188 The breakthrough in this area of research was the report that molecular H2 selectively reduced the levels This Review is based upon data from systematic of hydroxyl radicals in vitro and that H2 molecules also reviews, review papers and individual studies known to exerted therapeutic antioxidant activities in a rat model of the authors. Other relevant studies were identified by MEDLINE and ISI Web of Knowledge searches of English- middle cerebral artery occlusion.189 H2 selectively reduces language papers published up to November 2011 using hydroxyl radicals and peroxynitrite, which are strong the search terms: “acetogen”, “breath test”, “colorectal oxidants that react indiscriminately with nucleic acids, cancer”, “Crohn’s disease”, “hydrogen”, “hydrogen lipids and proteins resulting in DNA fragmentation, sulfide”, “hydrogenotroph”, “inflammatory bowel lipid peroxidation and protein inactivation. disorder”, “irritable bowel syndrome”, “methanogen”, Accumulating evidence suggests that H2 can protect “methane”, “obesity”, “sulfate-reducing bacteria”, various cells, tissues and organs against oxidative “ulcerative colitis”, and any relevant combination of terms. When appropriate, the reference lists of key papers were injury.190,191 Of specific interest to the field of gastro­ checked to identify additional articles of interest. enterology is the demonstration that H2-enriched water 1. Suarez, F., Furne, J., Springfield, J. & Levitt, M. 3. Levitt, M. D. & Bond, J. H. Jr. Volume, 5. Kirk, E. The quantity and composition of Insights into human colonic physiology composition, and source of intestinal gas. human colonic flatus. Gastroenterology 12, obtained from the study of flatus composition. Gastroenterology 59, 921–929 (1970). 782–794 (1949). Am. J. Physiol. 272, G1028–G1033 (1997). 4. Christl, S. U., Murgatroyd, P. R., Gibson, G. R. 6. Steggerda, F. R. Gastrointestinal gas following 2. Suarez, F., Furne, J., Springfield, J. & Levitt, M. & Cummings, J. H. Production, metabolism, food consumption. Ann. NY Acad. Sci. 150, Production and elimination of sulfur-containing and excretion of hydrogen in the large 57–66 (1968). gases in the rat colon. Am. J. Physiol. 274, intestine. Gastroenterology 102, 1269–1277 7. Tomlin, J., Lowis, C. & Read, N. W. G727–G733 (1998). (1992). Investigation of normal flatus production in NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY ADVANCE ONLINE PUBLICATION  |  11 © 2012 Macmillan Publishers Limited. All rights reserved REVIEWS healthy volunteers. Gut 32, 665–669 27. Metz, G., Jenkins, D. J., Peters, T. J., 45. Miller, T. L. & Wolin, M. J. Formation of hydrogen (1991). Newman, A. & Blendis, L. M. Breath hydrogen and formate by Ruminococcus albus. J. Bacteriol. 8. Beazell, J. M. & Ivey, A. C. The quantity of as a diagnostic method for hypolactasia. 116, 836–846 (1973). colonic flatus excreted by the “normal” Lancet 1, 1155–1157 (1975). 46. Schwiertz, A. et al. Anaerostipes caccae gen. individual. Am. J. Dig. Dis. 8, 128–129 (1941). 28. Rhodes, J. M., Middleton, P. & Jewell, D. P. The nov., sp. nov., a new saccharolytic, acetate- 9. Askevold, F. Investigations on the influence of lactulose hydrogen breath test as a diagnostic utilising, butyrate-producing bacterium from diet on the quantity and composition of test for small-bowel bacterial overgrowth. human faeces. Syst. Appl. Microbiol. 25, 46–51 intestinal gas in humans. Scand. J. Clin. Lab. Scand. J. Gastroenterol. 14, 333–336 (1979). (2002). Invest. 8, 87–94 (1956). 29. Khoshini, R., Dai, S. C., Lezcano, S. & 47. Steer, T., Collins, M. D., Gibson, G. R., Hippe, H. 10. Ruge, E. Beitrag sur kennuness der darmgase Pimentel, M. A systematic review of diagnostic & Lawson, P. A. Clostridium hathewayi sp. nov., [German]. Sitsber. Kaiserlicken Akad. 44, 739 tests for small intestinal bacterial overgrowth. from human faeces. Syst. Appl. Microbiol. 24, (1861). Dig. Dis. Sci. 53, 1443–1454 (2008). 353–357 (2001). 11. Levitt, M. D., Gibson, G. R. & Christl, S. U. in 30. Bures, J. et al. Small intestinal bacterial 48. Kamlage, B., Gruhl, B. & Blaut, M. Isolation and Human Colonic Bacteria: Role in Nutrition, overgrowth syndrome. World J. Gastroenterol. characterization of two new homoacetogenic Physiology, and Pathology (eds Gibson, G. R. & 16, 2978–2990 (2010). hydrogen-utilizing bacteria from the human Macfarlane, G. T.) 131–154 (CRC Press, Boca 31. Caride, V. J. et al. Scintigraphic determination intestinal tract that are closely related to Raton, 1975). of small intestinal transit time: comparison Clostridium coccoides. Appl. Environ. Microbiol. 12. Steggerda, F. R. & Dimmick, J. F. Effects of with the hydrogen breath technique. 63, 1732–1738 (1997). bean diets on concentration of carbon dioxide Gastroenterology 86, 714–720 (1984). 49. Pochart, P., Dore, J., Lemann, F., Goderel, I. & in flatus. Am. J. Clin. Nutr. 19, 120–124 (1966). 32. Bond, J. H., Engel, R. R. & Levitt, M. D. Factors Rambaud, J. C. Interrelations between 13. Levitt, M. D. & Ingelfinger, F. J. Hydrogen and influencing pulmonary methane excretion in populations of methanogenic archaea and methane production in man. Ann. NY Acad. Sci. man—indirect method of studying in situ sulfate-reducing bacteria in the human colon. 150, 75–81 (1968). metabolism of methane-producing colonic FEMS Microbiol. Lett. 77, 225–228 (1992). 14. Levitt, M. D., Hirsh, P., Fetzer, C. A., bacteria. J. Exp. Med. 133, 572–588 (1971). 50. Macfarlane, G. T., Gibson, G. R. & Sheahan, M. & Levine, A. S. H2 excretion after 33. Levitt, M. D., Furne, J. K., Kuskowski, M. & Cummings, J. H. Comparison of fermentation ingestion of complex carbohydrates. Ruddy, J. Stability of human methanogenic reactions in different regions of the human Gastroenterology 92, 383–389 (1987). flora over 35 years and a review of insights colon. J. Appl. Bacteriol. 72, 57–64 (1992). 15. Wolin, M. J. & Miller, T. L. in Human Intestinal obtained from breath methane measurements. 51. Pochart, P. et al. Pyxigraphic sampling to Microflora in Health and Disease (ed. Clin. Gastroenterol. Hepatol. 4, 123–129 enumerate methanogens and anaerobes in the Hentges, D. J.) 147–165 (Academic Press, New (2006). right colon of healthy humans. Gastroenterology York, 1983). 34. Gibson, G. R., Macfarlane, G. T. & 105, 1281–1285 (1993). 16. Strocchi, A. & Levitt, M. D. Factors affecting Cummings, J. H. Occurrence of sulphate- 52. Flourie, B. et al. Site and substrates for methane hydrogen production and consumption by reducing bacteria in human faeces and the production in human colon. Am. J. Physiol. 260, human fecal flora. The critical roles of hydrogen relationship of dissimilatory sulphate reduction G752–G757 (1991). tension and methanogenesis. J. Clin. Invest. 89, to methanogenesis in the large gut. J. Appl. 53. Gibson, G. R. et al. Alternative pathways for 1304–1311 (1992). Bacteriol. 65, 103–111 (1988). hydrogen disposal during fermentation in the 17. Hammer, H. F. Colonic hydrogen absorption: 35. Segal, I., Walker, A. R., Lord, S. & human colon. Gut 31, 679–683 (1990). quantification of its effect on hydrogen Cummings, J. H. Breath methane and large 54. Marteau, P. et al. Comparative study of bacterial accumulation caused by bacterial fermentation bowel cancer risk in contrasting African groups within the human cecal and fecal of carbohydrates. Gut 34, 818–822 (1993). populations. Gut 29, 608–613 (1988). microbiota. Appl. Environ. Microbiol. 67, 18. Le Marchand, L., Wilkens, L. R., Harwood, P. & 36. Wilkens, L. R., Le Marchand, L., Harwood, P. & 4939–4942 (2001). Cooney, R. V. Use of breath hydrogen and Cooney, R. V. Use of breath hydrogen and 55. Macfarlane, G. T. & Gibson, G. R. in methane as markers of colonic fermentation in methane as markers of colonic fermentation in Gastrointestinal Microbiology (eds Mackie, R. I. & epidemiologic studies: circadian patterns of epidemiological studies: variability in excretion. White, B. A.) 269–318 (Chapman and Hall, New excretion. Environ. Health Perspect. 98, Cancer Epidemiol. Biomarkers Prev. 3, 149–153 York, 1997). 199–202 (1992). (1994). 56. Duncan, S. H., Hold, G. L., Barcenilla, A., 19. Strocchi, A., Ellis, C. & Levitt, M. D. 37. O’Keefe, S. J. et al. Why do African Americans Stewart, C. S. & Flint, H. J. Roseburia intestinalis Reproducibility of measurements of trace gas get more colon cancer than Native Africans? sp nov., a novel saccharolytic, butyrate-producing concentrations in expired air. Gastroenterology J. Nutr. 137, 175S–182S (2007). bacterium from human faeces. Int. J. Sys. Evol. 101, 175–179 (1991). 38. Thauer, R. K., Jungermann, K. & Decker, K. Microbiol. 52, 1615–1620 (2002). 20. Calloway, D. H. Respiratory hydrogen and Energy conservation in chemotrophic anaerobic 57. Duncan, S. H. et al. Proposal of Roseburia faecis methane as affected by consumption of gas- bacteria. Bacteriol. Rev. 41, 100–180 (1977). sp. nov., Roseburia hominis sp. nov. and forming foods. Gastroenterology 51, 383–389 39. Nava, G. M., Carbonero, F., Croix, J. A., Roseburia inulinivorans sp. nov., based on (1966). Greenberg, E. & Gaskins, H. R. Abundance and isolates from human faeces. Int. J. Syst. Evol. 21. Calloway, D. H. & Murphy, E. L. The use of diversity of mucosa-associated Microbiol. 56, 2437–2441 (2006). expired air to measure intestinal gas hydrogenotrophic microbes in the healthy 58. Duncan, S. H. & Flint, H. J. Proposal of a neotype formation. Ann. NY Acad. Sci. 150, 82–95 human colon. ISME J. 6, 57–70 (2011). strain (A1–86) for Eubacterium rectale. Request (1968). 40. Fite, A. et al. Identification and quantitation of for an opinion. Int. J. Syst. Evol. Microbiol. 58, 22. Bjorneklett, A. & Jenssen, E. Relationships mucosal and faecal desulfovibrios using real 1735–1736 (2008). between hydrogen (H2) and methane (CH4) time polymerase chain reaction. Gut 53, 59. Chassard, C., Delmas, E., Lawson, P. A. & production in man. Scand. J. Gastroenterol. 17, 523–529 (2004). Bernalier-Donadille, A. Bacteroides xylanisolvens 985–992 (1982). 41. Zinkevich, V. V. & Beech, I. B. Screening of sp. nov., a xylan-degrading bacterium isolated 23. Cloarec, D. et al. Breath hydrogen response to sulfate-reducing bacteria in colonoscopy from human faeces. Int. J. Syst. Evol. Microbiol. lactulose in healthy subjects: relationship to samples from healthy and colitic human gut 58, 1008–1013 (2008). methane producing status. Gut 31, 300–304 mucosa. FEMS Microbiol. Ecol. 34, 147–155 60. Zoetendal, E. G., Plugge, C. M., Akkermans, A. D. (1990). (2000). & de Vos, W. M. Victivallis vadensis gen. nov., sp. 24. Vernia, P., Camillo, M. D., Marinaro, V. & 42. Chassard, C. et al. Assessment of metabolic nov., a sugar-fermenting anaerobe from human Caprilli, R. Effect of predominant methanogenic diversity within the intestinal microbiota from faeces. Int. J. Syst. Evol. Microbiol. 53, 211–215 flora on the outcome of lactose breath test in healthy humans using combined molecular and (2003). irritable bowel syndrome patients. Eur. J. Clin. cultural approaches. FEMS Microbiol. Ecol. 66, 61. Derrien, M., Vaughan, E. E., Plugge, C. M. & Nutr. 57, 1116–1119 (2003). 496–504 (2008). de Vos, W. M. Akkermansia muciniphila gen. 25. Calloway, D. H., Murphy, E. L. & Bauer, D. 43. Harding, G. K., Sutter, V. L., Finegold, S. M. & nov., sp. nov., a human intestinal mucin- Determination of lactose intolerance by breath Bricknell, K. S. Characterization of Bacteroides degrading bacterium. Int. J. Syst. Evol. Microbiol. analysis. Am. J. Dig. Dis. 14, 811–815 (1969). melaninogenicus. J. Clin. Microbiol. 4, 354–359 54, 1469–1476 (2004). 26. Levitt, M. D. & Donaldson, R. M. Use of (1976). 62. Duncan, S. H., Hold, G. L., Harmsen, H. J., respiratory hydrogen (H2) excretion to detect 44. Simmering, R. et al. Ruminococcus luti sp. nov., Stewart, C. S. & Flint, H. J. Growth requirements carbohydrate malabsorption. J. Lab. Clin. Med. isolated from a human faecal sample. Syst. and fermentation products of Fusobacterium 75, 937–945 (1970). Appl. Microbiol. 25, 189–193 (2002). prausnitzii, and a proposal to reclassify it as 12  |  ADVANCE ONLINE PUBLICATION www.nature.com/nrgastro © 2012 Macmillan Publishers Limited. All rights reserved  REVIEWS Faecalibacterium prausnitzii gen. nov., comb. the presence or absence of methanogens. FEMS 98. Willis, C. L., Cummings, J. H., Neale, G. & nov. Int. J. Syst. Evol. Microbiol. 52, 2141–2146 Microbiol. Ecol. 74, 205–213 (2010). Gibson, G. R. Nutritional aspects of dissimilatory (2002). 82. Hansen, E. E. et al. Pan-genome of the dominant sulfate reduction in the human large intestine. 63. Tolvanen, K. E., Mangayil, R. K., Karp, M. T. & human gut-associated archaeon, Curr. Microbiol. 35, 294–298 (1997). Santala, V. P. Simple enrichment system for Methanobrevibacter smithii, studied in twins. 99. Zverlov, V. et al. Lateral gene transfer of hydrogen producers. Appl. Environ. Microbiol. 77, Proc. Natl Acad. Sci. USA 108, 4599–4606 dissimilatory (bi)sulfite reductase revisited. 4246–4248 (2011). (2011). J. Bacteriol. 187, 2203–2208 (2005). 64. Cammack, R. Hydrogenase sophistication. 83. Miller, T. L. & Wolin, M. J. Methanosphaera 100. Klein, M. et al. Multiple lateral transfers of Nature 397, 214–215 (1999). stadtmaniae gen. nov., sp. nov.: a species that dissimilatory sulfite reductase genes between 65. Schmidt, O., Drake, H. L. & Horn, M. A. Hitherto forms methane by reducing methanol with major lineages of sulfate-reducing prokaryotes. unknown [Fe-Fe]-hydrogenase gene diversity in hydrogen. Arch. Microbiol. 141, 116–122 J. Bacteriol. 183, 6028–6035 (2001). anaerobes and anoxic enrichments from a (1985). 101. Wagner, M., Roger, A. J., Flax, J. L., moderately acidic fen. Appl. Environ. Microbiol. 84. Fricke, W. F. et al. The genome sequence of Brusseau, G. A. & Stahl, D. A. Phylogeny of 76, 2027–2031 (2010). Methanosphaera stadtmanae reveals why this dissimilatory sulfite reductases supports an 66. Schmidt, O. et al. Novel [NiFe]- and [FeFe]- human intestinal archaeon is restricted to early origin of sulfate respiration. J. Bacteriol. hydrogenase gene transcripts indicative of active methanol and H2 for methane formation and ATP 180, 2975–2982 (1998). facultative aerobes and obligate anaerobes in synthesis. J. Bacteriol. 188, 642–658 (2006). 102. Meyer, B. & Kuever, J. Phylogeny of the alpha and earthworm gut contents. Appl. Environ. Microbiol. 85. Eckburg, P. B. et al. Diversity of the human beta subunits of the dissimilatory adenosine‑5'- 77, 5842–5850 (2011). intestinal microbial flora. Science 308, phosphosulfate (APS) reductase from sulfate- 67. Drake, H. L., Gossner, A. S. & Daniel, S. L. Old 1635–1638 (2005). reducing prokaryotes—origin and evolution of acetogens, new light. Ann. NY Acad. Sci. 1125, 86. Zhang, H. et al. Human gut microbiota in obesity the dissimilatory sulfate-reduction pathway. 100–128 (2008). and after gastric bypass. Proc. Natl Acad. Sci Microbiology 153, 2026–2044 (2007). 68. Lajoie, S. F., Bank, S., Miller, T. L. & Wolin, M. J. USA 106, 2365–2370 (2009). 103. Hungin, A. P., Chang, L., Locke, G. R., Acetate production from hydrogen and [13C] 87. Weaver, G. A., Krause, J. A., Miller, T. L. & Dennis, E. H. & Barghout, V. Irritable bowel carbon dioxide by the microflora of human feces. Wolin, M. J. Incidence of methanogenic bacteria syndrome in the United States: prevalence, Appl. Environ. Microbiol. 54, 2723–2727 (1988). in a sigmoidoscopy population: an association symptom patterns and impact. Aliment. 69. Bernalier, A. et al. Acetogenesis from H2 and CO2 of methanogenic bacteria and diverticulosis. Pharmacol. Ther. 21, 1365–1375 (2005). by methane- and non‑methane‑producing human Gut 27, 698–704 (1986). 104. Everhart, J. E. in The Burden of Digestive colonic bacterial communities. FEMS Microbiol. 88. Miller, T. L. & Wolin, M. J. Enumeration of Diseases in the United States (ed. Everhart, J. E.) Ecol. 19, 193–202 (1996). Methanobrevibacter smithii from human feces. 77–87 (US Government Printing Office. US 70. Dore, J. et al. Enumeration of H2-utilizing Arch. Microbiol. 45, 317 (1982). Department of Health and Human Services, methanogenic archaea, acetogenic and sulfate- 89. Abell, G. C. J., Conlon, M. A. & McOrist, A. L. Public Health Service, National Institutes of reducing bacteria from human feces. FEMS Methanogenic archaea in adult human faecal Health, National Institute of Diabetes and Microbiol. Ecol. 17, 279–284 (1995). samples are inversely related to butyrate Digestive and Kidney Diseases, Washington, DC, 71. Bernalier, A., Rochet, V., Leclerc, M., Dore, J. & concentration. Microb. Ecol. Health Dis. 18, 2008). Pochart, P. Diversity of H2/CO2-utilizing 154–160 (2006). 105. Drossman, D. A. et al. A prospective assessment acetogenic bacteria from feces of 90. Scanlan, P. D., Shanahan, F. & Marchesi, J. R. of bowel habit in irritable bowel syndrome in non‑methane‑producing humans. Curr. Microbiol. Human methanogen diversity and incidence in women: defining an alternator. Gastroenterology 33, 94–99 (1996). healthy and diseased colonic groups using mcrA 128, 580–589 (2005). 72. Bernalier, A., Willems, A., Leclerc, M., Rochet, V. gene analysis. BMC Microbiol. 8, 79 (2008). 106. Drossman, D. A. et al. International survey of & Collins, M. D. Ruminococcus 91. Mihajlovski, A., Alric, M. & Brugere, J. F. patients with IBS: symptom features and their hydrogenotrophicus sp. nov., a new H2/CO2- A putative new order of methanogenic Archaea severity, health status, treatments, and risk utilizing acetogenic bacterium isolated from inhabiting the human gut, as revealed by taking to achieve clinical benefit. J. Clin. human feces. Arch. Microbiol. 166, 176–183 molecular analyses of the mcrA gene. Res. Gastroenterol. 43, 541–550 (2009). (1996). Microbiol. 159, 516–521 (2008). 107. Hasler, W. L. Irritable bowel syndrome and 73. Lovell, C. R. & Leaphart, A. B. Community-level 92. Marquet, P., Duncan, S. H., Chassard, C., bloating. Best Pract. Res. Clin. Gastroenterol. 21, analysis: key genes of CO2-reductive Bernalier-Donadille, A. & Flint, H. J. Lactate has 689–707 (2007). acetogenesis. Methods Enzymol. 397, 454–469 the potential to promote hydrogen sulphide 108. Serra, J., Azpiroz, F. & Malagelada, J. R. Impaired (2005). formation in the human colon. FEMS Microbiol. transit and tolerance of intestinal gas in the 74. Wolin, M. J. & Miller, T. L. Bacterial strains from Lett. 299, 128–134 (2009). irritable bowel syndrome. Gut 48, 14–19 (2001). human feces that reduce CO2 to acetic acid. 93. Kleessen, B., Kroesen, A. J., Buhr, H. J. & 109. Salonen, A., de Vos, W. M. & Palva, A. Appl. Environ. Microbiol. 59, 3551–3556 (1993). Blaut, M. Mucosal and invading bacteria in Gastrointestinal microbiota in irritable bowel 75. Ohashi, Y., Igarashi, T., Kumazawa, F. & patients with inflammatory bowel disease syndrome: present state and perspectives. Fujisawa, T. Analysis of acetogenic bacteria in compared with controls. Scand. J. Gastroenterol. Microbiology 156, 3205–3215 (2010). human feces with formyltetrahydrofolate 37, 1034–1041 (2002). 110. Kunkel, D. et al. Methane on breath testing is synthetase sequences. Biosci. Microflora 26, 94. Gibson, G. R., Macfarlane, S. & associated with constipation: a systematic 37–40 (2007). Macfarlane, G. T. Metabolic interactions review and meta-analysis. Dig. Dis. Sci. 56, 76. Cord-Ruwisch, R., Seitz, H. & Conrad, R. The involving sulphate-reducing and methanogenic 1612–1618 (2011). capacity of hydrogenotrophic anaerobic bacteria bacteria in the human large intestine. FEMS 111. Spiegel, B. M. R. Questioning the bacterial to compete for traces of hydrogen depends on Microbiol. Ecol. 12, 117–125 (1993). overgrowth hypothesis of irritable bowel the redox potential of the terminal electron 95. Newton, D. F., Cummings, J. H., Macfarlane, S. syndrome: An epidemiologic and evolutionary acceptor. Arch. Microbiol. 149, 350–357 (1988). & Macfarlane, G. T. Growth of a human perspective. Clin. Gastroenterol. Hepatol. 9, 77. Rey, F. E. et al. Dissecting the in vivo metabolic intestinal Desulfovibrio desulfuricans in 461–469 (2011). potential of two human gut acetogens. J. Biol. continuous cultures containing defined 112. Chatterjee, S., Park, S., Low, K., Kong, Y. & Chem. 285, 22082–22090 (2010). populations of saccharolytic and amino acid Pimentel, M. The degree of breath methane 78. Miller, T. L. & Wolin, M. J. Pathways of acetate, fermenting bacteria. J. Appl. Microbiol. 85, production in IBS correlates with the severity of propionate, and butyrate formation by the human 372–380 (1998). constipation. Am. J. Gastroenterol. 102, fecal microbial flora. Appl. Environ. Microbiol. 62, 96. Hopkins, M. J., Macfarlane, G. T., Furrie, E., 837–841 (2007). 1589–1592 (1996). Fite, A. & Macfarlane, S. Characterisation of 113. King, T. S., Elia, M. & Hunter, J. O. Abnormal 79. Hedderich, R. & Whitman, W. B. in The intestinal bacteria in infant stools using real- colonic fermentation in irritable bowel syndrome. Prokaryotes (ed. Dworkin, M.) 1050–1079 time PCR and northern hybridisation analyses. Lancet 352, 1187–1189 (1998). (Springer, New York, 2006). FEMS Microbiol. Ecol. 54, 77–85 (2005). 114. Simren, M. & Stotzer, P. O. Use and abuse of 80. Miller, T. L. & Wolin, M. J. Methanogens in human 97. Gibson, G. R., Cummings, J. H. & hydrogen breath tests. Gut 55, 297–303 (2006). and animal intestinal tracts. Syst. Appl. Microbiol. Macfarlane, G. T. Growth and activities of 115. Pimentel, M., Chow, E. J. & Lin, H. C. 7, 223–229 (1986). sulphate-reducing bacteria in gut contents of Comparison of peak breath hydrogen production 81. Chassard, C., Delmas, E., Robert, C. & Bernalier- healthy subjects and patients with ulcerative in patients with irritable bowel syndrome, chronic Donadille, A. The cellulose-degrading microbial colitis. FEMS Microbiol. Lett. 86, 103–111 fatigue syndrome and fibromyalgia. community of the human gut varies according to (1991). Gastroenterology 118, A413–A413 (2000). NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY ADVANCE ONLINE PUBLICATION  |  13 © 2012 Macmillan Publishers Limited. All rights reserved REVIEWS 116. Pimentel, M., Chow, E. J. & Lin, H. C. 135. Strocchi, A., Ellis, C. J. & Levitt, M. D. Use of 154. Perman, J. A. Methane and colorectal cancer. Eradication of small intestinal bacterial metabolic inhibitors to study H2 consumption by Gastroenterology 87, 728–730 (1984). overgrowth reduces symptoms of irritable bowel human feces: evidence for a pathway other than 155. Karlin, D. A., Jones, R. D., Stroehlein, J. R., syndrome. Am. J. Gastroenterol. 95, methanogenesis and sulfate reduction. J. Lab. Mastromarino, A. J. & Potter, G. D. Breath 3503–3506 (2000). Clin. Med. 121, 320–327 (1993). methane excretion in patients with unresected 117. Yu, D., Cheeseman, F. & Vanner, S. Combined 136. Roediger, W. E., Moore, J. & Babidge, W. Colonic colorectal cancer. J. Natl Cancer Inst. 69, oro-caecal scintigraphy and lactulose hydrogen sulfide in pathogenesis and treatment of 573–576 (1982). breath testing demonstrate that breath testing ulcerative colitis. Dig. Dis. Sci. 42, 1571–1579 156. Haines, A., Metz, G., Dilawari, J., Blendis, L. & detects oro-caecal transit, not small intestinal (1997). Wiggins, H. Breath-methane in patients with bacterial overgrowth in patients with IBS. Gut 137. Tragnone, A. et al. Dietary habits as risk factors cancer of the large bowel. Lancet 2, 481–483 60, 334–340 (2011). for inflammatory bowel disease. Eur. J. (1977). 118. Quigley, E. M. Germs, gas and the gut; the Gastroenterol. Hepatol. 7, 47–51 (1995). 157. Pique, J. M., Pallares, M., Cuso, E., Vilar- evolving role of the enteric flora in IBS. Am. J. 138. Truelove, S. C. Ulcerative colitis provoked by Bonet, J. & Gassull, M. A. Methane production Gastroenterol. 101, 334–335 (2006). milk. Br. Med. J. 1, 154–160 (1961). and colon cancer. Gastroenterology 87, 601–605 119. Quigley, E. M. M. The enteric microbiota in the 139. Pitcher, M. C., Beatty, E. R. & Cummings, J. H. (1984). pathogenesis and management of The contribution of sulphate reducing bacteria 158. Karlin, D. A., Mastromarino, A. J., Jones, R. D., constipation. Best Pract. Res. Clin. and 5‑aminosalicylic acid to faecal sulphide in Stroehlein, J. R. & Lorentz, O. Fecal skatole and Gastroenterol. 25, 119–126 (2011). patients with ulcerative colitis. Gut 46, 64–72 indole and breath methane and hydrogen in 120. Pimentel, M., Park, S., Kong, Y., Low, K. & (2000). patients with large bowel polyps or cancer. Chatterjee, S. Methane gas is associated with 140. Edmond, L. M., Hopkins, M. J., Magee, E. A. & J. Cancer Res. Clin. Oncol. 109, 135–141 (1985). constipation predominant symptoms in IBS: Cummings, J. H. The effect of 5‑aminosalicylic 159. Peled, Y. Methane production and colon cancer. Results from a double-blind controlled study. acid-containing drugs on sulfide production by Gastroenterology 88, 1294 (1985). Gastroenterology 130, A514 (2006). sulfate-reducing and amino acid-fermenting 160. Kashtan, H., Rabau, M., Peled, Y., Milstein, A. & 121. Pimentel, M. et al. Methane, a gas produced by bacteria. Inflamm. Bowel Dis. 9, 10–17 (2003). Wiznitzer, T. Methane production in patients with enteric bacteria, slows intestinal transit and 141. Moore, J., Babidge, W., Millard, S. & Roediger, W. colorectal carcinoma. Isr. J. Med. Sci. 25, augments small intestinal contractile activity. Colonic luminal hydrogen sulfide is not elevated 614–616 (1989). Am. J. Physiol. Gastrointest. Liver Physiol. 290, in ulcerative colitis. Dig. Dis. Sci. 43, 162–165 161. Sivertsen, S. M., Bjorneklett, A., Gullestad, H. P. G1089–G1095 (2006). (1998). & Nygaard, K. Breath methane and colorectal- 122. Rajilic-Stojanovic, M. et al. Global and deep 142. Duffy, M. et al. Sulfate-reducing bacteria colonize cancer. Scand. J. Gastroenterol. 27, 25–28 molecular analysis of microbiota signatures in pouches formed for ulcerative colitis but not for (1992).
 fecal samples from patients with irritable bowel familial adenomatous polyposis. Dis. Colon 162. Holma, R. et al. Colonic methanogenesis in vivo syndrome. Gastroenterology 141, 1792–1801 Rectum 45, 384–388 (2002). and in vitro and fecal pH after resection of (2011). 143. Ohge, H. et al. Association between fecal colorectal cancer and in healthy intact colon. Int. 123. Quigley, E. M. & Quera, R. Small intestinal hydrogen sulfide production and pouchitis. Dis. J. Colorectal Dis. 27, 171–178 (2011). bacterial overgrowth: roles of antibiotics, Colon Rectum 48, 469–475 (2005). 163. Babidge, W., Millard, S. & Roediger, W. Sulfides prebiotics, and probiotics. Gastroenterology 144. Gibson, G. R., Cummings, J. H. & impair short chain fatty acid beta-oxidation at 130, S78–S90 (2006). Macfarlane, G. T. Growth and activities of acyl-CoA dehydrogenase level in colonocytes: 124. Evans, B. W., Clark, W. K., Moore, D. J. & sulphate-reducing bacteria in gut contents of implications for ulcerative colitis. Mol. Cell. Whorwell, P. J. Tegaserod for the treatment of healthy subjects and patients with ulcerative Biochem. 181, 117–124 (1998). irritable bowel syndrome and chronic colitis. FEMS Microbiol. Lett. 86, 103–111 164. Cai, W. J., Wang, M. J., Ju, L. H., Wang, C. & constipation. Cochrane Database Syst. Rev. (1991). Zhu, Y. C. Hydrogen sulfide induces human colon Issue 4. Art. No.: CD003960 (2007). 145. Pitcher, M. C. L., Beatty, E. R., Gibson, G. R. & cancer cell proliferation: role of Akt, ERK and 125. Kim, D. Y. & Camilleri, M. Serotonin: a mediator Cummings, J. H. Incidence and activities of p21. Cell. Biol. Int. 34, 565–572 (2010). of the brain-gut connection. Am. J. sulphate-reducing bacteria in patients with 165. Deplancke, B. & Gaskins, H. R. Hydrogen sulfide Gastroenterol. 95, 2698–2709 (2000). ulcerative colitis. Gut 36, A63 (1995). induces serum-independent cell cycle entry in 126. De Ponti, F. & Tonini, M. Irritable bowel 146. Levine, J., Ellis, C. J., Furne, J. K., Springfield, J. nontransformed rat intestinal epithelial cells. syndrome: new agents targeting serotonin & Levitt, M. D. Fecal hydrogen sulfide production FASEB J. 17, 1310–1312 (2003). receptor subtypes. Drugs 61, 317–332 (2001). in ulcerative colitis. Am. J. Gastroenterol. 93, 166. Leschelle, X. et al. Adaptative metabolic 127. Ford, A. C. et al. Efficacy of 5‑HT3 antagonists 83–87 (1998). response of human colonic epithelial cells to the and 5‑HT4 agonists in irritable bowel syndrome: 147. Loubinoux, J., Bronowicki, J. P., Pereira, I. A., adverse effects of the luminal compound systematic review and meta-analysis. Am. J. Mougenel, J. L. & Faou, A. E. Sulfate-reducing sulfide. Biochim. Biophys. Acta 1725, 201–212 Gastroenterol. 104, 1831–1843 (2009). bacteria in human feces and their association (2005). 128. Kawabata, A. et al. Hydrogen sulfide as a novel with inflammatory bowel diseases. FEMS 167. Christl, S. U., Eisner, H. D., Dusel, G., Kasper, H. nociceptive messenger. Pain 132, 74–81 Microbiol. Ecol. 40, 107–112 (2002). & Scheppach, W. Antagonistic effects of sulfide (2007). 148. Tsai, H. H., Sunderland, D., Gibson, G. R., and butyrate on proliferation of colonic mucosa: 129. Matsunami, M. et al. Luminal hydrogen sulfide Hart, C. A. & Rhodes, J. M. A novel mucin a potential role for these agents in the plays a pronociceptive role in mouse colon. Gut sulphatase from human faeces: its pathogenesis of ulcerative colitis. Dig. Dis. Sci. 58, 751–761 (2009). identification, purification and characterization. 41, 2477–2481 (1996). 130. Schemann, M. & Grundy, D. Role of hydrogen Clin. Sci. (Lond.) 82, 447–454 (1992). 168. Attene-Ramos, M. S., Wagner, E. D., Plewa, M. J. sulfide in visceral nociception. Gut 58, 149. Croix, J. A. et al. On the relationship between & Gaskins, H. R. Evidence that hydrogen sulfide 744–747 (2009). sialomucin and sulfomucin expression and is a genotoxic agent. Mol. Cancer Res. 4, 9–14 131. Distrutti, E. et al. A nitro-arginine derivative of hydrogenotrophic microbes in the human colonic (2006). trimebutine (NO2‑Arg‑Trim) attenuates pain mucosa. PLoS ONE 6, e24447 (2011). 169. Attene-Ramos, M. S., Wagner, E. D., induced by colorectal distension in conscious 150. Bjorneklett, A., Fausa, O. & Midtvedt, T. Bacterial Gaskins, H. R. & Plewa, M. J. Hydrogen sulfide rats. Pharmacol. Res. 59, 319–329 (2009). overgrowth in jejunal and ileal disease. Scand. J. induces direct radical-associated DNA damage. 132. Roediger, W. E., Duncan, A., Kapaniris, O. & Gastroenterol. 18, 289–298 (1983). Mol. Cancer Res. 5, 455–459 (2007). Millard, S. Reducing sulfur compounds of the 151. McKay, L. F., Eastwood, M. A. & Brydon, W. G. 170. Attene-Ramos, M. S. et al. DNA damage and colon impair colonocyte nutrition: Implications Methane excretion in man—a study of breath, toxicogenomic analyses of hydrogen sulfide in for ulcerative colitis. Gastroenterology 104, flatus, and faeces. Gut 26, 69–74 (1985). human intestinal epithelial FHs 74 Int cells. 802–809 (1993). 152. Peled, Y., Weinberg, D., Hallak, A. & Gilat, T. Environ. Mol. Mutagen. 51, 304–314 (2010). 133. Roediger, W. E., Duncan, A., Kapaniris, O. & Factors affecting methane production in 171. Kanazawa, K. et al. Factors influencing the Millard, S. Sulphide impairment of substrate humans. Gastrointestinal diseases and development of sigmoid colon cancer— oxidation in rat colonocytes: a biochemical alterations of colonic flora. Dig. Dis. Sci. 32, bacteriologic and biochemical studies. Cancer basis for ulcerative colitis? Clin. Sci. (Lond.) 85, 267–271 (1987). 77, 1701–1706 (1996). 623–627 (1993). 153. Pimentel, M. et al. Methane production during 172. Ramasamy, S., Singh, S., Taniere, P., 134. O’Neil, M. J. The Merck Index. An Encyclopedia lactulose breath test is associated with Langman, M. J. & Eggo, M. C. Sulfide-detoxifying of Chemicals, Drugs, and Biologicals (Merck & gastrointestinal disease presentation. Dig. Dis. enzymes in the human colon are decreased in Co, Whitehouse Station, NJ, 2001). Sci. 48, 86–92 (2003). cancer and upregulated in differentiation. Am. J. 14  |  ADVANCE ONLINE PUBLICATION www.nature.com/nrgastro © 2012 Macmillan Publishers Limited. All rights reserved  REVIEWS Physiol. Gastrointest. Liver Physiol. 291, community of human gut microbiota reveals an but not for familial adenomatous polyposis. Dis. G288–G296 (2006). increase in Lactobacillus in obese patients and Colon Rectum 45, 384–388 (2002). 173. Balamurugan, R., Rajendiran, E., George, S., methanogens in anorexic patients. PLoS ONE 4, 201. Bullock, N. R., Booth, J. C. & Gibson, G. R. Samuel, G. V. & Ramakrishna, B. S. Real-time e7125 (2009). Comparative composition of bacteria in the polymerase chain reaction quantification of 188. Zheng, X. F., Sun, X. J. & Xia, Z. F. Hydrogen human intestinal microflora during remission specific butyrate-producing bacteria, Desulfovibrio resuscitation, a new cytoprotective approach. and active ulcerative colitis. Curr. Issues Intest. and Enterococcus faecalis in the feces of Clin. Exp. Pharmacol. Physiol. 38, 155–163 Microbiol. 5, 59–64 (2004). patients with colorectal cancer. J. Gastroenterol. (2011). 202. Smith, F. M. et al. A characterization of anaerobic Hepatol. 23, 1298–1303 (2008). 189. Ohsawa, I. et al. Hydrogen acts as a therapeutic colonization and associated mucosal 174. Marchesi, J. R. et al. Towards the human antioxidant by selectively reducing cytotoxic adaptations in the undiseased illeal pouch. colorectal cancer microbiome. PLoS ONE 6 oxygen radicals. Nat. Med. 13, 688–694 Colorectal Dis. 7, 563–570 (2005). e20447 (2011). (2007). 203. Bambury, N., Coffey, J. C., Burke, J., 175. Kostic, A. D. et al. Genomic analysis identifies 190. Hong, Y., Chen, S. & Zhang, J. M. Hydrogen as a Redmond, H. P. & Kirwan, W. O. Sulphomucin association of Fusobacterium with colorectal selective antioxidant: a review of clinical and expression in ileal pouches: emerging carcinoma. Genome Res. 22, 292–298 (2011). experimental studies. J. Int. Med. Res. 38, differences between ulcerative colitis and 176. Castellarin, M. et al. Fusobacterium nucleatum 1893–1903 (2010). familial adenomatous polyposis pouches. Dis. infection is prevalent in human colorectal 191. Huang, C. S., Kawamura, T., Toyoda, Y. & Colon Rectum 51, 561–567 (2008). carcinoma. Genome Res. 22, 299–306 (2011). Nakao, A. Recent advances in hydrogen 204. Coffey, J. C. et al. Pathogenesis of and unifying 177. Fiorucci, S., Distrutti, E., Cirino, G. & research as a therapeutic medical gas. Free hypothesis for idiopathic pouchitis. Am. J. Wallace, J. L. The emerging roles of hydrogen Rad. Res. 44, 971–982 (2010). Gastroenterol. 104, 1013–1023 (2009). sulfide in the gastrointestinal tract and liver. 192. Kajiya, M., Silva, M. J. B., Sato, K., Ouhara, K. & 205. Lim, M. et al. An assessment of bacterial Gastroenterology 131, 259–271 (2006). Kawai, T. Hydrogen mediates suppression of dysbiosis in pouchitis using terminal restriction 178. Wang, R. Two’s company, three’s a crowd: colon inflammation induced by dextran sodium fragment length polymorphisms of 16S can H2S be the third endogenous gaseous sulfate. Biochem. Biophys. Res. Commun. 386, ribosomal DNA from pouch effluent microbiota. transmitter? FASEB J. 16, 1792–1798 (2002). 11–15 (2009). Dis. Colon Rectum 52, 1492–1500 (2009). 179. Wallace, J. L., Ferraz, J. & Muscara, M. Hydrogen 193. Nakao, A., Toyoda, Y., Sharma, P., Evans, M. & 206. Rowan, F. et al. Desulfovibrio bacterial species sulfide: an endogenous mediator of resolution of Guthrie, N. Effectiveness of hydrogen rich water are increased in ulcerative colitis. Dis. Colon inflammation and injury. Antioxid. Redox Signal on antioxidant status of subjects with potential Rectum 53, 1530–1536 (2010). http://dx.doi.org/10.1089/ars.2011.4351. metabolic syndrome—an open label pilot study. 207. Verma, R., Verma, A. K., Ahuja, V. & Paul, J. Real- 180. Furne, J., Saeed, A. & Levitt, M. D. Whole tissue J. Clin. Biochem. Nutr. 46, 140–149 (2010). time analysis of mucosal flora in patients with hydrogen sulfide concentrations are orders of 194. Metz, G. L., Blendis, L. M. & Jenkins, J. A. inflammatory bowel disease in India. J. Clin. magnitude lower than presently accepted values. Proceedings: alveolar H2 in the diagnosis of Microbiol. 48, 4279–4282 (2010). Am. J. Physiol. Regul. Integr. Comp. Physiol. 295, carbohydrate malabsorption. Gut 16, 398 208. Strauss, J. et al. Invasive potential of gut R1479–R1485 (2008). (1975). mucosa-derived Fusobacterium nucleatum 181. Brock, T. D. & Od’ea, K. Amorphous ferrous 195. Metz, G., Gassull, M. A., Drasar, B. S., positively correlates with IBD status of the host. sulfide as a reducing agent for culture of Jenkins, D. J. & Blendis, L. M. Breath-hydrogen Inflamm. Bowel Dis. 17, 1971–1978 (2011). anaerobes. Appl. Environ. Microbiol. 33, 254–256 test for small-intestinal bacterial colonisation. 209. Scanlan, P. D., Shanahan, F. & Marchesi, J. R. (1977). Lancet 1, 668–669 (1976). Culture-independent analysis of desulfovibrios in 182. Hungate, R. E. in Methods in Microbiology (eds 196. El Oufir, L. et al. Relations between transit time, the human distal colon of healthy, colorectal Norris, J. R. & Ribbons, D. W.) 117–132 fermentation products, and hydrogen cancer and polypectomized individuals. FEMS (Academic Press, London, 1969). consuming flora in healthy humans. Gut 38, Microbiol. Ecol. 69, 213–221 (2009). 183. Turnbaugh, P. J. et al. An obesity-associated gut 870–877 (1996). microbiome with increased capacity for energy 197. Pimentel, M., Kong, Y. & Park, S. IBS subjects Acknowledgements harvest. Nature 444, 1027–1031 (2006). with methane on lactulose breath test have Related research was supported by grants from the 184. Haines, A. P., Imeson, J. D. & Wiggins, H. S. lower postprandial serotonin levels than NIH (RO1 CA135379) and Carle Foundation-University Relation of breath methane with obesity and subjects with hydrogen. Dig. Dis. Sci. 49, 84–87 of Illinois Translational Research Program. The other factors. Int. J. Obes. 8, 675–680 (1984). (2004). authors thank Matthew T. Leslie for help in the 185. Million, M. et al. Obesity-associated gut 198. Dear, K. L., Elia, M. & Hunter, J. O. Do bibliographical search. This Review is dedicated to microbiota is enriched in Lactobacillus reuteri interventions which reduce colonic bacterial the scientific legacies of Dr Michael D. Levitt and Dr and depleted in Bifidobacterium animalis and fermentation improve symptoms of irritable Meyer J. Wolin, both of whom consistently contributed Methanobrevibacter smithii. Int. J. Obes. (Lond.) bowel syndrome? Dig. Dis. Sci. 50, 758–766 key studies over many years relating to the http://dx.doi.org/10.1038/ijo.2011.153. (2005). importance of the microbial hydrogen economy on 186. Schwiertz, A. et al. Microbiota and SCFA in lean 199. Levine, J., Furne, J. K. & Levitt, M. D. Ashkenazi colonic homeostasis. and overweight healthy subjects. Obesity (Silver Jews, sulfur gases, and ulcerative colitis. J. Clin. Spring) 18, 190–195 (2010). Gastroenterol. 22, 288–291 (1996). Author contributions 187. Armougom, F., Henry, M., Vialettes, B., 200. Duffy, M. et al. Sulfate-reducing bacteria All authors contributed equally to all aspects of the Raccah, D. & Raoult, D. Monitoring bacterial colonize pouches formed for ulcerative colitis manuscript. NATURE REVIEWS | GASTROENTEROLOGY & HEPATOLOGY ADVANCE ONLINE PUBLICATION  |  15 © 2012 Macmillan Publishers Limited. All rights reserved