Site and substrates for methane production BERNARD PHILIPPE
FLOUR& POCHART,
in human colon
PIERRE PELLIER, CHRISTIAN FLORENT, AND JEAN-CLAUDE RAMBAUD
PHILIPPE
MARTEAU,
Unite de Recherches sur les Fonctions Intestinales et la Nutrition, Institut National de la Sante et de la Recherche Medicale Unite 290, Service d’Hepato-Gastroenterologie, H6pital Saint-Lazare, 75475 Paris Cedex 10; and Service d’Hepato-Gastroenterologie, H6pital Saint-Antoine, 75571 Paris Cedex 12, France
FLOURI& BERNARD, PIERRE PELLIER, CHRISTIAN FLORENT, PHILIPPE MARTEAU; PHILIPPE POCHART, AND JEAN-CLAUDE RAMBAUD. Site and substrates for methaneproduction in human colon. Am. J. Physiol. 260 (Gastrointest. Liver Physiol. 23):
G752-G757,1991.-On two occasions separated by seven days, 22 g mucin (hog gastric mucin) was infused into right and left colon of 12 healthy volunteers (6 CH, producers and 6 nonproducers) maintained on a controlled diet. In the six CH, producers, excessvolumes of H2 excreted in breath were 73.4 t 11.9 and 35.1 & 14.1 (SE) ml/8 h (P < 0.05) in response to right and left colonic infusion of mucin, respectively; excess volumes of CH, were, respectively, 6.7 2 1.7 and 38.9 t 11.1 ml/8 h (P < 0.05). In the six CH, nonproducers, excessvolumes of Hz excreted in breath were 76.6 t 17.6 and 30.8 t 6.3 ml/8 h (P < 0.02) in response to right and left colonic infusion of mucin, respectively; excess volumes of CH, were, respectively, 0.0 * 0.0 and 0.1 & 0.1 ml/8 h (not significant). In a further experiment, 17 healthy volunteers (10 CH, producers and 7 nonproducers) were given on 2 consecutive days an oral load and an enema of 10 g lactulose. In the 10 CH, producers, excess volumes of HP excreted in breath were 74.6 t 15.1 and 32.3 t 11.5 ml/6 h (P < 0.001) in response to oral ingestion and lactulose enema, respectively; excess volumes of CH4 were, respectively, 7.7 & 3.0 and 38.2 k 7.2 ml/6 h (P < 0.001). In the seven CH* nonproducers, excess volumes of Hz excreted in breath were 94.0 t 21.8 and 43.0 -f 16.4 ml/6 h (P < 0.01) in response to oral ingestion and lactulose enema, respectively; excess volumes of CH4 were, respectively, 0.0 t 0.0 and 2.1 * 1.2 ml/6 h (not significant). These results show that 1) Hz production occurs in both right and left colon, whereas CHI production occurs in CH, producers mainly in distal large intestine; 2) both exogenous and endogenous substrates promote CH4 production, but in usual dietary situations the latter might provide the bulk of the fuel for methanogenesis. hydrogen; colonic metabolism; mucin; lactulose HYDROGEN AND METHANE are constituents of human breath derived from bacterial fermentation in the large intestine. Most people in the Western world excrete Hz, but only 33-70% of healthy adults excrete CH, in breath in significant amounts (8). Methane production is a matter of growing interest, because in addition to its physiological importance and possible involvement in functional intestinal disorders, it seems to have connections with rectal and colonic diseases, particularly adenocarinoma (23). This paper focuses on the substrates and the site producing CH4 in healthy humans. Any hypothesis concerning these points must take into account several conflicting data. Unlike the H2 concentrations in exhaled gas, the CH4 concentrations, which G752
0193-1857/91
reflect colonic CH4, do not vary during the nychthemeron (at least during daylight) and are therefore not influenced by meals (4). Similarly, ingestion of 10 g lactulose, a nonabsorbable disaccharide, does not increase pulmonary CH, excretion in most CH4 producers (4). These findings led Wolin and Miller (30) to postulate that HP is produced in the human colon in two different ways: by a “pulse” phase that is due to the fast catabolism of carbohydrates from exogenous sources and does not lead to further CH4 formation and by large sustained release from endogenous substrates, which provides a continuous supply of H2 and leads to CH, synthesis by methanogenic bacteria (mainly Methanobrevibacter smithii; Ref, 21) in subjects harboring sufficient counts of these microorganisms. However, this hypothesis does not take into account two facts: 1) larger amounts of nonabsorb-
able exogenous substrates, such as 20, 25, or 33 g lactulose (3, 10, 24), 25 g pentosans (19), or even only 10 g lactulose in certain CH4 producers (7), can significantly increase pulmonary CH, excretion; and 2) regular consumption of dietary fibers rich in xylan or pectin (20) or of a high-starch diet (13) causes a sustained augmentation of-the diurnal concentrations of expired CH4. To overcome these apparently contradictory observations, we formulated an alternative hypothesis, i.e., that CH.4 p reduction is confined to the distal part of the large intestine (8). Three lines of evidence supported this hypothesis: first, M. smithii, like all methanogens, is needs known to be a strict anaerobe whose implantation a very low redox potential and which has an optimal external pH of 7. These conditions exist in the distal but not in the proximal large intestine (2,9). Second, during experiments aimed to measure the intestinal gas production in response to in situ lactose infusion, Bond et al. (4) incidentally found in one subject that most CH4 formation occurred distally to the colonic splenic flexure. Third, in vitro incubation with lactulose of cecal and fecal human homogenates from CH producers under strictly anae robic conditions show:d that CH4 was formed from fecal homogenates only ( 11). According to our hypothes is, on ly the exogenous or endogenous substrates reaching the distal part of the large intestine would induce H2 and CH, production in CH4 producers (8). In nonproducers, only HZ would be formed. The only exogenous substrates that could reach the distal gut are those incompletely or slowly catabolized in the right colon, such as dietary fibers, or those quickly propelled by a large osmotic water flow, such as large amounts of lactose or the nonabsorbable disaccharide lactulose (3,
$1.50 Copyright 0 1991 the American Physiological
Society
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
SITE
AND
SUBSTRATES
FOR
METHANE
10, 24). On the contrary, endogenous substrates of coionic origin, mainly mucins, sloughed epithelial cells, and dead bacteria are obviously available directly and continuously in the distal large intestine. These substrates, together with slowly migrating dietary fibers, might be responsible for the sustained and apparently non-mealrelated CH4 production. Consequently, the aims of the present work were 1) to confirm in healthy humans that CH4 production is confined to the distal large intestine and 2) to show that both endogenous and exogenous substrates are adequate fuels for CH* production, provided they are available in this part of the bowel. Our results also provide interesting data concerning the part played by endogenous substrates in Hz production and hence in fulfilling the metabolic requirement of the colonic flora. SUBJECTS
AND
METHODS
Subjects. The study included 41 healthy volunteers comprising 24 males and 17 females, aged 22-40 yr. None had been given antibiotics, laxatives, or enemas during the month preceding the study. Twenty-three were identified as CH4 producers because their breath CH4 concentrations exceeded one part per million (ppm) above room atmosphere (4). Ten CH4 producers and 10 nonproducers participated in the right and left colonic infusion studies, and 13 producers and 8 nonproducers were included in the lactulose ingestion and rectal infusion experiments. All gave informed written consent to the protocol, which was approved by the local ethics committee. Right and left colonic infusion studies. For 3 wk, the 20 volunteers were given a controlled diet that provided 100 g starch and 27 g wheat bran daily, which we have previously shown to induce low pulmonary CH, excretion in CH4 producers (13). On day 6 of the two first weeks, the subjects were intubated with a four-lumen tube attached to a bag that could be filled with or emptied of mercury and/or air. Two lumina ending near the tracting bag and 30 cm proximally could be used to infuse mucin so that the point nearest the colonic segment to be studied could be selected fluoroscopically before the beginning of infusion. One colonic infusion was performed on the right side and one on the left in random order on day 7, after a fast of at least 12 h, when the perfusion point was found by fluoroscopy to be located either in the ascending colon or at the level of the left iliac crest. Twenty-eight grams of hog gastric mucin (Sigma Chemical, St. Louis, MO) containing 22 g pure mucin (i.e., 16.5 g carbohydrate) were dissolved in 170 ml distilled water. After homogenization, the solution was dialyzed for 4 h in distilled water by means of Spectralor membrane tubing with a molecular weight cutoff of 6,000-8,000 (Spectrum Medical Industries, Los Angeles, CA). Its osmolality was then brought up to 300 mosmol/ kgHa0 by adding sodium bicarbonate, sodium, and potassium chloride. Eventually, 10 ml of barium sulfate was added to the solution. The final pH, osmolality, and volume of the solution were 6.8,300 mosmol/kgH20, and 200 ml, respectively. To determine the basal time course of Hz and CH4 excretion in breath in the fasting state,
PRODUCTION
IN
HUMANS
G753
eight subjects (4 CH4 producers and 4 nonproducers) were not perfused with the mucin solution but with a control plasmalike solution containing 10 ml of barium sulfate. All solutions were continuously stirred and infused at 37°C at a constant rate of 1.5 ml/min, i.e., a total perfusion time of 133 min. At the end of each perfusion, an abdominal plain film allowed us to check the absence of migration of the tube and the site reached by barium sulfate. The latter never reached the splenic flexure of the colon when infused into the ascending colon and did not reflux above the descending colon when infused at the level of the left iliac crest. In two subjects, we ascertained that barium sulfate acted as a true marker of water compounds by adding 100 &i “‘In to the solutions infused in the right and left colon; scintigrams were performed at the same time as abdominal plain films. There was a very good concordance between images obtained by the two techniques. Endexpiratory gas samples were obtained via a modified Haldane-Priestley tube in 60-ml plastic syringes equipped with three-way stopcocks (13). Three basal samples were taken at 15min intervals before colonic infusion and at 30-min intervals for 8 h after infusion had been started. During these 8 h, subjects continued to fast, and smoking was not allowed. Any symptom that occurred during this period and throughout the 2 days that followed was recorded. No movement or mucous discharge occurred during any of the infusion studies. On day 7 of the 3rd wk, fasting subjects drank 10 g lactulose in 100 ml water (lactulose syrup; Duphar Laboratories, Villeurbanne, France), and expired gas was collected before and during 8 h after ingestion in the same way as during the mucin infusion studies. Lactulose oral ingestion and rectal infusion. On two consecutive days, 21 fasting subjects who received a lowfiber dinner on the preceding day were randomly given an oral load of 10 g lactulose in 100 ml water and an enema of 10 g lactulose in 45 ml water at 37°C. Four of these subjects (3 CH4 producers and 1 nonproducer) did not retain the enema throughout the three first hours of the experiment and were excluded from the study. Endexpiratory gas samples were obtained as previously (i.e., three basal samples were taken at 15.min intervals before oral ingestion or rectal infusion of lactulose and at 30min intervals for 8 h after oral ingestion and for 6 h after rectal infusion). Last, to ascertain that barium sulfate used in the right and left colonic infusions of mucin could not modify He and CH4 pulmonary excretion, four further CH4 producers received on 2 consecutive days lactulose enemas with or without 10 ml barium sulfate; H2 and CH4 excretion in breath was not significantly affected by barium sulfate. Assays. All assays were performed within 12 h of sampling. HZ concentrations were measured with an electrochemical cell (Exhaled Hydrogen Monitor; GM1 Medical, Renfrew, Scotland). Methane concentrations were determined by gas-liquid chromatography (IGC 121 DFL; Intersmat, Courtry, France). The minimal detectable HZ and CH4 concentrations were 1.0 and 0.5 ppm, respectively. Calm&ions and statistics. Ha and CH4 concentrations were recorded as the change in parts per million over basal values, defined as the arithmetic mean of the three
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
G754
SITE AND SUBSTRATES
FOR METHANE
concentrations obtained before infusion of mucin and before oral ingestion and rectal infusion of lactulose. The excess volumes of Hz and CH, excreted in breath during the sampling period were determined by integrating the areas under the HP and CH4 concentration curves; tidal volumes were determined from the Radford normogram, and data were expressed in milliliters per total sampling period (13). As discussed later, we also determined the mean volume of HP excreted in breath in the eight fasting subjects who received the plasmalike infusions; this value was then subtracted from the volume of H2 excreted by subjects who were perfused with mucin. To compare the volumes of Hz and CH4 produced after lactulose ingestion and after rectal infusion in the second part of the study, we only used, in the case of lactulose ingestion, the volumes exhaled during the 6 h after the first sharp increase in breath HP concentrations (>5 ppm). Statistical analysis for significance was performed by use of Student’s paired and unpaired t tests. Results are expressed as means t SE.
PRODUCTION
IN HUMANS
TIME (hourd 2. Breath CH, excretion during and after infusion of plasmalike solution into right and left colon of 4 CH, producers (means k SE). FIG.
RESULTS
Right and left colonic infusion studies. Breath H2 and CH4 excretions in experiments where plasmalike solutions alone were perfused in the right and left colon are given in Figs. 1 and-2. Fasting breath Hz concentrations were 5 t 1 and 7 t 3 ppm before the right and left colonic infusions, and then Hz concentrations dropped slightly. The mean volumes of HP excreted in breath were 18.3 t 5.4 and 18.5 t 8.3 ml/8 h for the right and left colonic infusions of plasmalike solutions, respectively. Fasting breath CH4 concentrations in the four CH4 producers were 21 t 13 and 14 t 5 ppm before the right and left infusion studies, respectively, and a trend to decrease was also observed after the 5th h. The mean volumes of CH4 excreted in breath were 70.3 t 28.6 and 38.9 t 6.6 ml/8 h for the right and left colonic infusions, respectively. In the six CH, producers, the excess volumes of Hz exhaled during and after the right and left colonic infusions of mucin were 73.4t 11.9and 35.1 t 14.1ml/8 h, respectively (P < 0.05),and the excess volumes of CH4 were 6.7 t 1.7 and 38.9 t 11.1ml/8 h (P < 0.05; Figs. 3 and 4). In the six nonproducers, the excess volumes of HZ exhaled during and after the right and left colonic
0
2
4
6
8
TIME (hours) FIG. 3. Excess volumes of Hz (0) and CH, (0) in breath of 6 CHI producers (means t SE) during and after infusion of 22 g mucin into right colon.
15 1
0
1
2
4
6
8
TIME (hours) FIG. 4. Excess volumes of Hz (0) and CH, (0) in the breath of 6 CH, producers (means t, SE) during and after infusion of 22 g mucin into left colon.
2
4
6
8
TIME (hours) 1. Breath Hz excretion during and after infusion of plasmalike solution into right and left colon of 8 subiects (means t SE). FIG.
mucin infusions were 76.6t 17.6and 30.8 t 6.3 ml/8 h (P c 0.02), respectively, and the excess volumes of CH4 were 0.0 t 0.0 and 0.1 t 0.1 ml/8 h [not significant (NS); Fig. 51. Pulmonary H2 and CH4 excretion returned to basal levels before or at the end of the 8 h that followed the beginning of mucin infusion. In addition, as no volunteer passed a movement during this period, it may
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
SITE AND SUBSTRATES
FOR METHANE
be assumed that the entire fermentable fraction of infused mucin was catabolized in the colon. Excess volumes of Hz calculated by subtraction of the breath H2 excretion measured in the eight subjects who received plasmalike solutions were very similar: 78.8 t 12.0 and 40.4 t 14.1 ml/8 h, respectively, for the right and left colonic infusions of mucin in the CH4 producers and 86.5 t 15.9 and 35.6 t 8.9 ml/8 h for the right and left colonic infusions in the nonproducers. In the six CH4 producers, the excess volumes of H2 and CH4 measured after 10 g lactulose ingestion were 98.5 t 14.1 and 12.9 t 8.9 ml/8 h, respectively. However, only two of these producers displayed obvious increases in CH4 excretion above basal values. In the six nonproducers, the excess volumes of Hz and CHI were 105.5 t 21.1 and 0.0 t 0.0 ml/8 h, respectively. In both CH4 producers and nonproducers the excess volumes of Hz exhaled were similar, whether mucin was infused into the right or left colon. During and after the right colonic infusions, pulmonary excretion of Hz in CH4 producers and nonproducers was 3.3 t 0.5 vs. 3.4 t 0.8 ml/8 h, respectively, per gram of infused (and probably degraded) mucin (NS). Similarly, the excess volumes of H2 after oral lactulose ingestion did not differ in CH4 producers and nonproducers (9.9 t 1.4 vs. 10.5 t 2.1 ml/8 h per g of lactulose). Thus, whatever the CH4 status, the. Hz production per gram of substrate that followed the infusion of mucin into the right colon was three times lower than the production that followed oral ingestion of small amounts of lactulose. Clinical tolerance of mucin perfusion was very good. The only symptoms noted were bloating and excess of rectal gas after left colonic infusions in three CH4 producers. Two other CH4 producers complained of the same symptoms during the lactulose breath test. No discharge of mucus was ever observed. Lactulose ingestion and rectal infusion. In the 10 CH, producers, the excess volumes of HP exhaled after oral ingestion and rectal infusion of lactulose were 74.6 t 15.1 and 32.3 t 11.5 ml/6 h (P < O.OOl), respectively, and the excess volumes of CH4 were 7.7 t 3.0 and 38.2 t 7.2 ml/6 h (P < 0.001; Figs. 6 and 7). Three of the producers exhibited obvious increases in CH4 excretion above basal values after oral ingestion. In the seven 15
PRODUCTION
G755
IN HUMANS
nonproducers, the excess volumes of H2 after oral ingestion and rectal infusion of lactulose were 94.0 t 21.8 and 43.0 t 16.4 ml/6 h, respectively (P c 0.01; Fig. 8), and the excess volumes of CH4 were 0.0 t 0.0 and 2.1 t 1.2 ml/6 h (NS). DISCUSSION There are several ways to calculate Hz and CH, excretion in breath in response to a colonic carbohydrate load. We have used the most common method, which consists of subtracting from H2 concentrations the basal value measured in the fasting state, i.e., present before starting the infusions. Alternatively, some authors subtract from the area under the Hz concentration curve the mean area under the H2 concentration curve obtained in another group of subjects fasting for the same duration as that of the carbohydrate challenge (18). When comparing both methods, Rumessen et al. (25) did not find any significant difference. Our results show that the method using the fasting basal value of H2 concentration leads to somewhat smaller values, a shift that does not change the meaning of our data. Probably as a result of the lowstarch diet, the fasting Hz pulmonary concentrations were much lower than those reported by Levitt et al. (18) and hence occurred a smaller further decay. Fasting CH,
0
2
4
6
TIME (hours) FIG. 6. Excess volumes of Ha (0) and CH4 (a) during 6 h after initial rise in breath Hz concentrations following oral ingestion of 10 g lactulose in 10 CH4 producers (means + SE).
15
.@1 ) I 1 h 1I\I &,oy, ) a*&6 0 2 4
8
TIME (hours) FIG. 5. Excess volumes of Hz in breath of 6 CH, nonproducers (means t SE) during and after infusion of 22 g mucin into right colon (0) and left colon (0). Negligible excretion of breath CH, is not shown.
R
0
0
A. ,
1
,
‘
,
2
1 4
,
,
, 6
TIME (hours) FIG. 7. Excess volumes of HP (0) and CH, (0) in breath of 10 CH4 producers (means k SE) after rectal infusion of 10 g lactulose.
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
G756
SITE AND SUBSTRATES
0
0
2
4
FOR METHANE
6
TIME (hours) 8. Excess volumes of Hz during 6 h after sustained rise in breath Hz concentrations after oral ingestion (0) and rectal infusion (0) of 10 g lactulose in 7 CH4 nonproducers (means t, SE). Negligible excretion of breath CH4 is not shown. FIG.
PRODUCTION
IN HUMANS
reduction pathway (l7), and great attention is now being paid to the competition for H2 among the sulfate-reducing bacteria and methanogens (14, 15). The discussion about the metabolic fate of endogenous substrates in the left colon should not mask the fact that exogenous substrates can also contribute to methanogenesis in this part of the large intestine, which contradicts the hypothesis of WoIin and Miller (30). Our present findings clearly show that a 10-g load of lactulose, which in our experiments and those by others (4, 7) did not raise breath CH4 excretion in most CH, producers when given orally, is perfectly able to do so if introduced into the distal large intestine by a retrograde route. The access to this part of the bowel due to colonic migrating pressure waves in response to the large fluid load originating in the small intestine (6, 27) probably explains why larger oral loads of lactulose increased CH, breath excretion in CH, producers (3, 10, 24). Similarly, large loads of fibers such as xylan or pectin might reach the left colon in sufficient amounts to increase CH, production, as a result of their reduced or slow fermentation (19, 20). Alternatively, the H2 formed in the right colon by the fermentation of carbohydrates might reach the distal colon in sufficient amounts to generate CH, production by methanogenic bacteria. This hypothesis is, however, unlikely, since the present mucin infusions into the right colon did not induce any rise in CH, pulmonary excretion in four of our six CH4 producers. However, in two others, a small rise in CH, production was observed. Similarly, the CH, breath level rose in 3 of the 10 CH, producers after the ingestion of 10 g lactulose, and similar findings after a 10-g lactulose load have been reported by others (7). The reasons for these observations are unclear. It could be that a small part of the substrates reached the distal bowel because of rapid colonic transit time. Alternatively, the possibility of CH4 production in the right colon cannot be excluded, as suggested by the fact that a few subjects exhale CH4 but no HP (26), indicating that their large intestine behaves like the stomach of ruminants. Several methods are available for measuring the entry into the colon of exogenous substrates, mainly carbohydrates, but little is known about the supply rates of endogenous substrates. One approach to obtaining this information is to calculate the contribution to total pulmonary H2 excretion made by each substrate delivered to the colon during a period of controlled diet with a low starch content (13). As we know that, in healthy humans consuming the same diet as that used in the present study, 24-h breath H2 excretion averaged 180 ml (13), and that the volume of this gas produced by dietary substrates (starch and wheat bran) was -65 ml (5, 12), we have concluded that endogenous substrates under such a diet could be the main source of Hz in these subjects. A small amount of endogenous substrates enter the colon from the ileum (28). Enzymes and debris of colonic parietal cells or of bacterial origin also help to form these substrates, but mucous secretion by the colon wall probably constitutes the main source of their production (16). As hog gastric mucin resembles human gastrointestinal mucins in structure and composition (l), we can reasonably extrapolate the results of our perfusion studies to the physiological events occurring in the
curves are quite different because breath CH4 excretion in individuals is quite variable, as we also observed in our study. The present results clearly show that in CH4 producers most of the CH4 production from mucin occurs in the distal large intestine and that, although exogenous substrates may also be good fuels for methanogenesis, this endogenous substrate seems to have an important role in this process. Furthermore, Ha production from mucin occurs in both the right and left colon. In humans, AL smithii is mainly responsible for CH, synthesis (2l), and several hypotheses have been formulated to explain why only certain adults harbor this bacteria in significant amounts. The reduction of COa by HP constitutes the single pathway for CH, synthesis by AL smithii, and 4 mol HP are needed to produce 1 mol CH4. During the present mucin perfusions into the left colon of CH4 producers, the mean total excess volume of H2 (including free and CH4-engaged Hz) was -35 + (38.9 x 4) = 191 ml (i.e., 8.5 mmol/8 h). However, in CH4 nonproducers perfused with mucin in the left colon, the mean total excess volume of H2 was only 31 ml (i.e., 1.4 mmol/8 h), which was almost the same as the volume of free Hz exhaled by the CH* producers. An interesting finding was that after mucin infusion into the right colon or after oral lactulose ingestion, the volumes of H2 exhaled in breath were also similar in CH, producers and nonproducers. Thus, in response to infusion of the endogenous substrate mucin into the left colon, total H2 excretion (calculated from free H2 and CHJ was six times higher in CH4 producers than nonproducers. Wolin and Miller (30) had calculated that this production should be 10 times higher in CH4 producers than nonproducers. They suggested that the rate of endogenous substrate supply is several times higher in CH4 producers than nonproducers, which in fact seems very unlikely because our experiments clearly show that in nonproducers the availability of the endogenous substrate mucin is not the rate-limiting step in Hz production. In fact, in CH4 nonproducers, Hz is probably used by the colonic microflora to reduce substrates to compounds other than CH*. In this connection, there is evidence that the colonic microflora produces acetate via a COa Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
SITE AND SUBSTRATES
FOR METHANE
large intestine. The carbohydrate moieties of gastrointestinal mucus have been shown to be degraded in vitro by the colonic flora from normal individuals (22). Using an in vitro fecal incubation system, Perman and Modler (22) showed that hog gastric mucin yielded less H2 than glucose; similarly, in vivo the volume of H2 produced from mucin was smaller in our study than that derived from lactulose. Here the volumes of Hz exhaled by the CH4 nonproducers during the probably complete catabolism of 22 g hog gastric mucin containing 16.5 g carbohydrate averaged 77 ml and 31 ml in the right and left colon, respectively. At this rate of Hz production, a daily colonic secretion of -47 g mucin (i.e., 35 g carbohydrate) would theoretically be needed to fill completely the gap between total pulm .onary excretion of H2 and the H2 from substrates of exogenous origin. This value is in rather good agreement with the estimate by Wolin and Miller (29, 30), who calculated from the fecal bacterial mass that the daily amount of carbohydrates required for colonic bacterial turnover was 66 g, with a carbohydrate supply from endogenous substrates equal to onehalf to two-thirds of this amount. We are grateful to C. Franchisseur, M. Maurel, and M.-C. Morin for expert technical assistance. Address for reprint requests: B. Flour%, INSERM U. 290, Hopital Saint-Lazare, 107 bis, rue du Fbg Saint-Denis, 75475 Paris Cedex 10, France. Received 27 September 1989; accepted in final form 1 January 1991. REFERENCES A. Structure of gastrointestinal mucus glycoproteins and the viscous and gel-forming properties of mucus. Br. Med. Bull. 34:
1. ALLEN,
B-33,1978. 2. BENTLEY, GORBACH.
D. W., R. L. NICHOLS, R. E. CONDON, AND S. L. The microflora of the human ileum and intra-abdominal colon: results of direct needle aspiration at surgery and evaluation of the technique. J. Lab. Clin. Med. 79: 421-429, 1972. 3. BJORNEKLETT, A., AND E. JENSSEN. Relationships between hydrogen (Hz) and methane (CH,) production in man. Stand. J. Gastroenterol. 17: 985-992, 1982. 4. BOND, J. H,, R. R. ENGEL, AND M. D. LEVITT. Factors influencing pulmonary methane excretion in man. J. Exp. Med. 133: 572-588, 197/l. 5. BOND, J. H., AND M. D. LEVITT. Effect of dietary fiber on intestinal gas production and small bowel transit time. Am. J. Clin. Nutr. 31, Suppl.: S169-S174,1978. 6. CHAUVE, A., G. DEVROEDE,
AND E. BASTIN. Intraluminal pressures during perfusion of the human colon in situ. Gastroenterology 70:
336-340,1976. 7. CLOAREC, D., F. BORNET, S. GOUILLOUD, AND J. P. GALMICHE. Breath hydrogen
healthy subjects: relationship
J. L. BARRY, B. SALIM, response to lactulose in to methane producing status. Gut 31:
300-304,199o. 8. COLOMBEL, J. F., B. FLOURIE, C. NEUT, C. FLORENT, A. LEBLOND, AND J. C. RAMBAUD. La methanogenese chez l’homme. Gastroenterol. Clin. Biol. 11: 694-700, 1987. 9. EVANS, D. F,, G. PYE, R. BRAMLEY, A. G. CLARK, T. J. DYSON, AND J. D. HARDCASTLE. Measurement of gastrointestinal pH profiles in normal ambulant human subjects. Gut 29: 195-216, 1988. 10. FLATZ, G., A. CZEIZEL, J. METNEKI, S. D. FLATZ, W. KUHNAU, AND D. JAHN. Pulmonarv hvdroeen and methane excretion follow-
PRODUCTION
G757
IN HUMANS
ing ingestion of an unabsorbable carbohydrate: a study of twins. J. Pediczt. Gustroenterol. Nutr. 4: 936-941, 1985. 11. FLOURIE, B., C. FLORENT, P. PELLIER, Y. BOUHNIK, AND J. C. RAMBAUD. Comparative study of hydrogen and methane production in the human colon by means of caecal and faecal homogenates. Gut 31: 684-685, 1990. 12. FLOURIE, B., C. FLORENT, F. ETANCHAUD, D. EVARD, C. FRANCHISSEUR, AND J. C. RAMBAUD. Starch absorption by healthy man evaluated by lactulose hydrogen breath test. Am. J. C&n. Nutr. 47: 61-66,1988. 13. FLOURIE, ALLI, AND
B., A. LEBLOND, C. FLORENT, M. RAUTUREAU, A. BISJ. C. RAMBAUD. Starch malabsorption and breath gas excretion in healthy humans consuming low- and high-starch diets.
Gastroenterology 14. GIBSON, G.
95: 356-363,
1988.
R., J. H. CUMMINGS, AND G. T. MACFARLANE. Use of a three-stage continuous culture system to study the effect of mucin on dissimilatory sulfate reduction and methanogenesis by mixed populations of human gut bacteria. Appl. Enoiron. Microbial. 54:
2750-2755,1988. 15. GIBSON, G.
R., G. T. MACFARLANE, AND J. H. CUMMINGS. Occurrence of sulphate-reducing bacteria in human faeces and the relationship of dissimilatory sulphate reduction to methanogenesis in the large gut. J. Appl. Bacterial. 65: 103-111, 1988. 16. HOSKINS, L. C., AND E. T. BOULDING. Mucin degradation in human colon ecosystems. Evidence for the existence and the role of bacterial subpopulations producing glycosidases as extracellular enzymes. J. Clin. Inuest. 67: 163-172, 1981. 17. LAJOIE, S. F., S. BANK, T. L. MILLER, AND M. J. WOLIN. Acetate production from hydrogen and [13C]carbon dioxide by the microflora of human feces. Appl. Environ. Microbial. 54: 2723-2727, 1988. 18. LEVITT, M. D., P. HIRSCH, C. A. FETZER, M. SHEAHAN, AND A. S. LEVINE. HZ excretion after ingestion of complex carbohydrates. Gastroenterology 92: 383-389, 1987. 19. MCKAY, L. F., W. G. BRYDON, M. A. EASTWOOD, AND J. H. SMITH. The influence of pentose on breath methane. Am. J. Clin. Nutr. 34: 2728-2733,198l. 20. MARTHINSEN, D., AND S. E. FLEMING. Excretion of breath and flatus gasses by humans consuming high-fiber diets. J. Nutr. 112: 1133-1143,1982. 21. MILLER, T. L., AND M. J. WOLIN. Enumeration of Methunobreuibatter smithii in human feces. Arch. Microbial. 131: 14-18, 1982. 22. PERMAN, J. A., AND S. MOLDER. Glycoproteins as substrates for
production 23.
24.
25.
of hydrogen and methane by colonic bacterial
flora.
Gastroenterology 83: 388-393, 1982. PIQUE, J. M., M. PALLARES, E. Cuso, J. VILAR-BONET, AND M. A. GASULL. Methane production and colon cancer. Gastroenterology 87: 601-605,1984. PITT, P., K. M. DE BRUIJN, M. F. BEECHING, E. GOLDBERG, AND L. M. BLENDIS. Studies on breath methane: the effect of ethnic origins and lactulose. Gut 21: 951-959, 1980. RUMESSEN, J. J., 0. HAMBERG, AND E. GUDMAND-HOYER. Inter-
val sampling of end-expiratory hydrogen (Hz) concentrations to quantify carbohydrate malabsorption by means of lactulose standards. Gut 31: 37-42,199O. 26. SALTZBERG, D. M., G. M. LEVINE, AND C. LUBAR. Impact of age, sex, race and functional complaints on hydrogen (Hz) production. Dig. Dis. Sci. 33: 308-313, 1988. 27. SAUNDERS, D. R., AND H. S. WIGGINS.
Conservation of mannitol, lactulose and raffinose by the human colon. Am. J. Physiol. 241
(Gastrointest. 28. STEPHEN,
Liver
Physiol.
4): G397-G402,
1981.
A. M., A. C. HADDAD, AND S. F. PHILLIPS. carbohydrate into the colon: direct measurements
Gastroenterology 29. WOLIN, M.
85: 589-595,
J. Fermentation
1983.
in the rumen and human large intes-
tine. Science Wash. DC 213: 1463-1468,198l. M. J., AND T. L. MILLER. Carbohydrate
30. WOLIN, Human
Intestinal
Microflora
Passage of in humans.
in Health
J. Hentees. New York: Academic. 1983.
and D.
fermentation. In: edited by D.
Disease, 147-165.
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
FLOUR& POCHART,
in human colon
PIERRE PELLIER, CHRISTIAN FLORENT, AND JEAN-CLAUDE RAMBAUD
PHILIPPE
MARTEAU,
Unite de Recherches sur les Fonctions Intestinales et la Nutrition, Institut National de la Sante et de la Recherche Medicale Unite 290, Service d’Hepato-Gastroenterologie, H6pital Saint-Lazare, 75475 Paris Cedex 10; and Service d’Hepato-Gastroenterologie, H6pital Saint-Antoine, 75571 Paris Cedex 12, France
FLOURI& BERNARD, PIERRE PELLIER, CHRISTIAN FLORENT, PHILIPPE MARTEAU; PHILIPPE POCHART, AND JEAN-CLAUDE RAMBAUD. Site and substrates for methaneproduction in human colon. Am. J. Physiol. 260 (Gastrointest. Liver Physiol. 23):
G752-G757,1991.-On two occasions separated by seven days, 22 g mucin (hog gastric mucin) was infused into right and left colon of 12 healthy volunteers (6 CH, producers and 6 nonproducers) maintained on a controlled diet. In the six CH, producers, excessvolumes of H2 excreted in breath were 73.4 t 11.9 and 35.1 & 14.1 (SE) ml/8 h (P < 0.05) in response to right and left colonic infusion of mucin, respectively; excess volumes of CH, were, respectively, 6.7 2 1.7 and 38.9 t 11.1 ml/8 h (P < 0.05). In the six CH, nonproducers, excessvolumes of Hz excreted in breath were 76.6 t 17.6 and 30.8 t 6.3 ml/8 h (P < 0.02) in response to right and left colonic infusion of mucin, respectively; excess volumes of CH, were, respectively, 0.0 * 0.0 and 0.1 & 0.1 ml/8 h (not significant). In a further experiment, 17 healthy volunteers (10 CH, producers and 7 nonproducers) were given on 2 consecutive days an oral load and an enema of 10 g lactulose. In the 10 CH, producers, excess volumes of HP excreted in breath were 74.6 t 15.1 and 32.3 t 11.5 ml/6 h (P < 0.001) in response to oral ingestion and lactulose enema, respectively; excess volumes of CH4 were, respectively, 7.7 & 3.0 and 38.2 k 7.2 ml/6 h (P < 0.001). In the seven CH* nonproducers, excess volumes of Hz excreted in breath were 94.0 t 21.8 and 43.0 -f 16.4 ml/6 h (P < 0.01) in response to oral ingestion and lactulose enema, respectively; excess volumes of CH4 were, respectively, 0.0 t 0.0 and 2.1 * 1.2 ml/6 h (not significant). These results show that 1) Hz production occurs in both right and left colon, whereas CHI production occurs in CH, producers mainly in distal large intestine; 2) both exogenous and endogenous substrates promote CH4 production, but in usual dietary situations the latter might provide the bulk of the fuel for methanogenesis. hydrogen; colonic metabolism; mucin; lactulose HYDROGEN AND METHANE are constituents of human breath derived from bacterial fermentation in the large intestine. Most people in the Western world excrete Hz, but only 33-70% of healthy adults excrete CH, in breath in significant amounts (8). Methane production is a matter of growing interest, because in addition to its physiological importance and possible involvement in functional intestinal disorders, it seems to have connections with rectal and colonic diseases, particularly adenocarinoma (23). This paper focuses on the substrates and the site producing CH4 in healthy humans. Any hypothesis concerning these points must take into account several conflicting data. Unlike the H2 concentrations in exhaled gas, the CH4 concentrations, which G752
0193-1857/91
reflect colonic CH4, do not vary during the nychthemeron (at least during daylight) and are therefore not influenced by meals (4). Similarly, ingestion of 10 g lactulose, a nonabsorbable disaccharide, does not increase pulmonary CH, excretion in most CH4 producers (4). These findings led Wolin and Miller (30) to postulate that HP is produced in the human colon in two different ways: by a “pulse” phase that is due to the fast catabolism of carbohydrates from exogenous sources and does not lead to further CH4 formation and by large sustained release from endogenous substrates, which provides a continuous supply of H2 and leads to CH, synthesis by methanogenic bacteria (mainly Methanobrevibacter smithii; Ref, 21) in subjects harboring sufficient counts of these microorganisms. However, this hypothesis does not take into account two facts: 1) larger amounts of nonabsorb-
able exogenous substrates, such as 20, 25, or 33 g lactulose (3, 10, 24), 25 g pentosans (19), or even only 10 g lactulose in certain CH4 producers (7), can significantly increase pulmonary CH, excretion; and 2) regular consumption of dietary fibers rich in xylan or pectin (20) or of a high-starch diet (13) causes a sustained augmentation of-the diurnal concentrations of expired CH4. To overcome these apparently contradictory observations, we formulated an alternative hypothesis, i.e., that CH.4 p reduction is confined to the distal part of the large intestine (8). Three lines of evidence supported this hypothesis: first, M. smithii, like all methanogens, is needs known to be a strict anaerobe whose implantation a very low redox potential and which has an optimal external pH of 7. These conditions exist in the distal but not in the proximal large intestine (2,9). Second, during experiments aimed to measure the intestinal gas production in response to in situ lactose infusion, Bond et al. (4) incidentally found in one subject that most CH4 formation occurred distally to the colonic splenic flexure. Third, in vitro incubation with lactulose of cecal and fecal human homogenates from CH producers under strictly anae robic conditions show:d that CH4 was formed from fecal homogenates only ( 11). According to our hypothes is, on ly the exogenous or endogenous substrates reaching the distal part of the large intestine would induce H2 and CH, production in CH4 producers (8). In nonproducers, only HZ would be formed. The only exogenous substrates that could reach the distal gut are those incompletely or slowly catabolized in the right colon, such as dietary fibers, or those quickly propelled by a large osmotic water flow, such as large amounts of lactose or the nonabsorbable disaccharide lactulose (3,
$1.50 Copyright 0 1991 the American Physiological
Society
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
SITE
AND
SUBSTRATES
FOR
METHANE
10, 24). On the contrary, endogenous substrates of coionic origin, mainly mucins, sloughed epithelial cells, and dead bacteria are obviously available directly and continuously in the distal large intestine. These substrates, together with slowly migrating dietary fibers, might be responsible for the sustained and apparently non-mealrelated CH4 production. Consequently, the aims of the present work were 1) to confirm in healthy humans that CH4 production is confined to the distal large intestine and 2) to show that both endogenous and exogenous substrates are adequate fuels for CH* production, provided they are available in this part of the bowel. Our results also provide interesting data concerning the part played by endogenous substrates in Hz production and hence in fulfilling the metabolic requirement of the colonic flora. SUBJECTS
AND
METHODS
Subjects. The study included 41 healthy volunteers comprising 24 males and 17 females, aged 22-40 yr. None had been given antibiotics, laxatives, or enemas during the month preceding the study. Twenty-three were identified as CH4 producers because their breath CH4 concentrations exceeded one part per million (ppm) above room atmosphere (4). Ten CH4 producers and 10 nonproducers participated in the right and left colonic infusion studies, and 13 producers and 8 nonproducers were included in the lactulose ingestion and rectal infusion experiments. All gave informed written consent to the protocol, which was approved by the local ethics committee. Right and left colonic infusion studies. For 3 wk, the 20 volunteers were given a controlled diet that provided 100 g starch and 27 g wheat bran daily, which we have previously shown to induce low pulmonary CH, excretion in CH4 producers (13). On day 6 of the two first weeks, the subjects were intubated with a four-lumen tube attached to a bag that could be filled with or emptied of mercury and/or air. Two lumina ending near the tracting bag and 30 cm proximally could be used to infuse mucin so that the point nearest the colonic segment to be studied could be selected fluoroscopically before the beginning of infusion. One colonic infusion was performed on the right side and one on the left in random order on day 7, after a fast of at least 12 h, when the perfusion point was found by fluoroscopy to be located either in the ascending colon or at the level of the left iliac crest. Twenty-eight grams of hog gastric mucin (Sigma Chemical, St. Louis, MO) containing 22 g pure mucin (i.e., 16.5 g carbohydrate) were dissolved in 170 ml distilled water. After homogenization, the solution was dialyzed for 4 h in distilled water by means of Spectralor membrane tubing with a molecular weight cutoff of 6,000-8,000 (Spectrum Medical Industries, Los Angeles, CA). Its osmolality was then brought up to 300 mosmol/ kgHa0 by adding sodium bicarbonate, sodium, and potassium chloride. Eventually, 10 ml of barium sulfate was added to the solution. The final pH, osmolality, and volume of the solution were 6.8,300 mosmol/kgH20, and 200 ml, respectively. To determine the basal time course of Hz and CH4 excretion in breath in the fasting state,
PRODUCTION
IN
HUMANS
G753
eight subjects (4 CH4 producers and 4 nonproducers) were not perfused with the mucin solution but with a control plasmalike solution containing 10 ml of barium sulfate. All solutions were continuously stirred and infused at 37°C at a constant rate of 1.5 ml/min, i.e., a total perfusion time of 133 min. At the end of each perfusion, an abdominal plain film allowed us to check the absence of migration of the tube and the site reached by barium sulfate. The latter never reached the splenic flexure of the colon when infused into the ascending colon and did not reflux above the descending colon when infused at the level of the left iliac crest. In two subjects, we ascertained that barium sulfate acted as a true marker of water compounds by adding 100 &i “‘In to the solutions infused in the right and left colon; scintigrams were performed at the same time as abdominal plain films. There was a very good concordance between images obtained by the two techniques. Endexpiratory gas samples were obtained via a modified Haldane-Priestley tube in 60-ml plastic syringes equipped with three-way stopcocks (13). Three basal samples were taken at 15min intervals before colonic infusion and at 30-min intervals for 8 h after infusion had been started. During these 8 h, subjects continued to fast, and smoking was not allowed. Any symptom that occurred during this period and throughout the 2 days that followed was recorded. No movement or mucous discharge occurred during any of the infusion studies. On day 7 of the 3rd wk, fasting subjects drank 10 g lactulose in 100 ml water (lactulose syrup; Duphar Laboratories, Villeurbanne, France), and expired gas was collected before and during 8 h after ingestion in the same way as during the mucin infusion studies. Lactulose oral ingestion and rectal infusion. On two consecutive days, 21 fasting subjects who received a lowfiber dinner on the preceding day were randomly given an oral load of 10 g lactulose in 100 ml water and an enema of 10 g lactulose in 45 ml water at 37°C. Four of these subjects (3 CH4 producers and 1 nonproducer) did not retain the enema throughout the three first hours of the experiment and were excluded from the study. Endexpiratory gas samples were obtained as previously (i.e., three basal samples were taken at 15.min intervals before oral ingestion or rectal infusion of lactulose and at 30min intervals for 8 h after oral ingestion and for 6 h after rectal infusion). Last, to ascertain that barium sulfate used in the right and left colonic infusions of mucin could not modify He and CH4 pulmonary excretion, four further CH4 producers received on 2 consecutive days lactulose enemas with or without 10 ml barium sulfate; H2 and CH4 excretion in breath was not significantly affected by barium sulfate. Assays. All assays were performed within 12 h of sampling. HZ concentrations were measured with an electrochemical cell (Exhaled Hydrogen Monitor; GM1 Medical, Renfrew, Scotland). Methane concentrations were determined by gas-liquid chromatography (IGC 121 DFL; Intersmat, Courtry, France). The minimal detectable HZ and CH4 concentrations were 1.0 and 0.5 ppm, respectively. Calm&ions and statistics. Ha and CH4 concentrations were recorded as the change in parts per million over basal values, defined as the arithmetic mean of the three
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
G754
SITE AND SUBSTRATES
FOR METHANE
concentrations obtained before infusion of mucin and before oral ingestion and rectal infusion of lactulose. The excess volumes of Hz and CH, excreted in breath during the sampling period were determined by integrating the areas under the HP and CH4 concentration curves; tidal volumes were determined from the Radford normogram, and data were expressed in milliliters per total sampling period (13). As discussed later, we also determined the mean volume of HP excreted in breath in the eight fasting subjects who received the plasmalike infusions; this value was then subtracted from the volume of H2 excreted by subjects who were perfused with mucin. To compare the volumes of Hz and CH4 produced after lactulose ingestion and after rectal infusion in the second part of the study, we only used, in the case of lactulose ingestion, the volumes exhaled during the 6 h after the first sharp increase in breath HP concentrations (>5 ppm). Statistical analysis for significance was performed by use of Student’s paired and unpaired t tests. Results are expressed as means t SE.
PRODUCTION
IN HUMANS
TIME (hourd 2. Breath CH, excretion during and after infusion of plasmalike solution into right and left colon of 4 CH, producers (means k SE). FIG.
RESULTS
Right and left colonic infusion studies. Breath H2 and CH4 excretions in experiments where plasmalike solutions alone were perfused in the right and left colon are given in Figs. 1 and-2. Fasting breath Hz concentrations were 5 t 1 and 7 t 3 ppm before the right and left colonic infusions, and then Hz concentrations dropped slightly. The mean volumes of HP excreted in breath were 18.3 t 5.4 and 18.5 t 8.3 ml/8 h for the right and left colonic infusions of plasmalike solutions, respectively. Fasting breath CH4 concentrations in the four CH4 producers were 21 t 13 and 14 t 5 ppm before the right and left infusion studies, respectively, and a trend to decrease was also observed after the 5th h. The mean volumes of CH4 excreted in breath were 70.3 t 28.6 and 38.9 t 6.6 ml/8 h for the right and left colonic infusions, respectively. In the six CH, producers, the excess volumes of Hz exhaled during and after the right and left colonic infusions of mucin were 73.4t 11.9and 35.1 t 14.1ml/8 h, respectively (P < 0.05),and the excess volumes of CH4 were 6.7 t 1.7 and 38.9 t 11.1ml/8 h (P < 0.05; Figs. 3 and 4). In the six nonproducers, the excess volumes of HZ exhaled during and after the right and left colonic
0
2
4
6
8
TIME (hours) FIG. 3. Excess volumes of Hz (0) and CH, (0) in breath of 6 CHI producers (means t SE) during and after infusion of 22 g mucin into right colon.
15 1
0
1
2
4
6
8
TIME (hours) FIG. 4. Excess volumes of Hz (0) and CH, (0) in the breath of 6 CH, producers (means t, SE) during and after infusion of 22 g mucin into left colon.
2
4
6
8
TIME (hours) 1. Breath Hz excretion during and after infusion of plasmalike solution into right and left colon of 8 subiects (means t SE). FIG.
mucin infusions were 76.6t 17.6and 30.8 t 6.3 ml/8 h (P c 0.02), respectively, and the excess volumes of CH4 were 0.0 t 0.0 and 0.1 t 0.1 ml/8 h [not significant (NS); Fig. 51. Pulmonary H2 and CH4 excretion returned to basal levels before or at the end of the 8 h that followed the beginning of mucin infusion. In addition, as no volunteer passed a movement during this period, it may
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
SITE AND SUBSTRATES
FOR METHANE
be assumed that the entire fermentable fraction of infused mucin was catabolized in the colon. Excess volumes of Hz calculated by subtraction of the breath H2 excretion measured in the eight subjects who received plasmalike solutions were very similar: 78.8 t 12.0 and 40.4 t 14.1 ml/8 h, respectively, for the right and left colonic infusions of mucin in the CH4 producers and 86.5 t 15.9 and 35.6 t 8.9 ml/8 h for the right and left colonic infusions in the nonproducers. In the six CH4 producers, the excess volumes of H2 and CH4 measured after 10 g lactulose ingestion were 98.5 t 14.1 and 12.9 t 8.9 ml/8 h, respectively. However, only two of these producers displayed obvious increases in CH4 excretion above basal values. In the six nonproducers, the excess volumes of Hz and CHI were 105.5 t 21.1 and 0.0 t 0.0 ml/8 h, respectively. In both CH4 producers and nonproducers the excess volumes of Hz exhaled were similar, whether mucin was infused into the right or left colon. During and after the right colonic infusions, pulmonary excretion of Hz in CH4 producers and nonproducers was 3.3 t 0.5 vs. 3.4 t 0.8 ml/8 h, respectively, per gram of infused (and probably degraded) mucin (NS). Similarly, the excess volumes of H2 after oral lactulose ingestion did not differ in CH4 producers and nonproducers (9.9 t 1.4 vs. 10.5 t 2.1 ml/8 h per g of lactulose). Thus, whatever the CH4 status, the. Hz production per gram of substrate that followed the infusion of mucin into the right colon was three times lower than the production that followed oral ingestion of small amounts of lactulose. Clinical tolerance of mucin perfusion was very good. The only symptoms noted were bloating and excess of rectal gas after left colonic infusions in three CH4 producers. Two other CH4 producers complained of the same symptoms during the lactulose breath test. No discharge of mucus was ever observed. Lactulose ingestion and rectal infusion. In the 10 CH, producers, the excess volumes of HP exhaled after oral ingestion and rectal infusion of lactulose were 74.6 t 15.1 and 32.3 t 11.5 ml/6 h (P < O.OOl), respectively, and the excess volumes of CH4 were 7.7 t 3.0 and 38.2 t 7.2 ml/6 h (P < 0.001; Figs. 6 and 7). Three of the producers exhibited obvious increases in CH4 excretion above basal values after oral ingestion. In the seven 15
PRODUCTION
G755
IN HUMANS
nonproducers, the excess volumes of H2 after oral ingestion and rectal infusion of lactulose were 94.0 t 21.8 and 43.0 t 16.4 ml/6 h, respectively (P c 0.01; Fig. 8), and the excess volumes of CH4 were 0.0 t 0.0 and 2.1 t 1.2 ml/6 h (NS). DISCUSSION There are several ways to calculate Hz and CH, excretion in breath in response to a colonic carbohydrate load. We have used the most common method, which consists of subtracting from H2 concentrations the basal value measured in the fasting state, i.e., present before starting the infusions. Alternatively, some authors subtract from the area under the Hz concentration curve the mean area under the H2 concentration curve obtained in another group of subjects fasting for the same duration as that of the carbohydrate challenge (18). When comparing both methods, Rumessen et al. (25) did not find any significant difference. Our results show that the method using the fasting basal value of H2 concentration leads to somewhat smaller values, a shift that does not change the meaning of our data. Probably as a result of the lowstarch diet, the fasting Hz pulmonary concentrations were much lower than those reported by Levitt et al. (18) and hence occurred a smaller further decay. Fasting CH,
0
2
4
6
TIME (hours) FIG. 6. Excess volumes of Ha (0) and CH4 (a) during 6 h after initial rise in breath Hz concentrations following oral ingestion of 10 g lactulose in 10 CH4 producers (means + SE).
15
.@1 ) I 1 h 1I\I &,oy, ) a*&6 0 2 4
8
TIME (hours) FIG. 5. Excess volumes of Hz in breath of 6 CH, nonproducers (means t SE) during and after infusion of 22 g mucin into right colon (0) and left colon (0). Negligible excretion of breath CH, is not shown.
R
0
0
A. ,
1
,
‘
,
2
1 4
,
,
, 6
TIME (hours) FIG. 7. Excess volumes of HP (0) and CH, (0) in breath of 10 CH4 producers (means k SE) after rectal infusion of 10 g lactulose.
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
G756
SITE AND SUBSTRATES
0
0
2
4
FOR METHANE
6
TIME (hours) 8. Excess volumes of Hz during 6 h after sustained rise in breath Hz concentrations after oral ingestion (0) and rectal infusion (0) of 10 g lactulose in 7 CH4 nonproducers (means t, SE). Negligible excretion of breath CH4 is not shown. FIG.
PRODUCTION
IN HUMANS
reduction pathway (l7), and great attention is now being paid to the competition for H2 among the sulfate-reducing bacteria and methanogens (14, 15). The discussion about the metabolic fate of endogenous substrates in the left colon should not mask the fact that exogenous substrates can also contribute to methanogenesis in this part of the large intestine, which contradicts the hypothesis of WoIin and Miller (30). Our present findings clearly show that a 10-g load of lactulose, which in our experiments and those by others (4, 7) did not raise breath CH4 excretion in most CH, producers when given orally, is perfectly able to do so if introduced into the distal large intestine by a retrograde route. The access to this part of the bowel due to colonic migrating pressure waves in response to the large fluid load originating in the small intestine (6, 27) probably explains why larger oral loads of lactulose increased CH, breath excretion in CH, producers (3, 10, 24). Similarly, large loads of fibers such as xylan or pectin might reach the left colon in sufficient amounts to increase CH, production, as a result of their reduced or slow fermentation (19, 20). Alternatively, the H2 formed in the right colon by the fermentation of carbohydrates might reach the distal colon in sufficient amounts to generate CH, production by methanogenic bacteria. This hypothesis is, however, unlikely, since the present mucin infusions into the right colon did not induce any rise in CH, pulmonary excretion in four of our six CH4 producers. However, in two others, a small rise in CH, production was observed. Similarly, the CH, breath level rose in 3 of the 10 CH, producers after the ingestion of 10 g lactulose, and similar findings after a 10-g lactulose load have been reported by others (7). The reasons for these observations are unclear. It could be that a small part of the substrates reached the distal bowel because of rapid colonic transit time. Alternatively, the possibility of CH4 production in the right colon cannot be excluded, as suggested by the fact that a few subjects exhale CH4 but no HP (26), indicating that their large intestine behaves like the stomach of ruminants. Several methods are available for measuring the entry into the colon of exogenous substrates, mainly carbohydrates, but little is known about the supply rates of endogenous substrates. One approach to obtaining this information is to calculate the contribution to total pulmonary H2 excretion made by each substrate delivered to the colon during a period of controlled diet with a low starch content (13). As we know that, in healthy humans consuming the same diet as that used in the present study, 24-h breath H2 excretion averaged 180 ml (13), and that the volume of this gas produced by dietary substrates (starch and wheat bran) was -65 ml (5, 12), we have concluded that endogenous substrates under such a diet could be the main source of Hz in these subjects. A small amount of endogenous substrates enter the colon from the ileum (28). Enzymes and debris of colonic parietal cells or of bacterial origin also help to form these substrates, but mucous secretion by the colon wall probably constitutes the main source of their production (16). As hog gastric mucin resembles human gastrointestinal mucins in structure and composition (l), we can reasonably extrapolate the results of our perfusion studies to the physiological events occurring in the
curves are quite different because breath CH4 excretion in individuals is quite variable, as we also observed in our study. The present results clearly show that in CH4 producers most of the CH4 production from mucin occurs in the distal large intestine and that, although exogenous substrates may also be good fuels for methanogenesis, this endogenous substrate seems to have an important role in this process. Furthermore, Ha production from mucin occurs in both the right and left colon. In humans, AL smithii is mainly responsible for CH, synthesis (2l), and several hypotheses have been formulated to explain why only certain adults harbor this bacteria in significant amounts. The reduction of COa by HP constitutes the single pathway for CH, synthesis by AL smithii, and 4 mol HP are needed to produce 1 mol CH4. During the present mucin perfusions into the left colon of CH4 producers, the mean total excess volume of H2 (including free and CH4-engaged Hz) was -35 + (38.9 x 4) = 191 ml (i.e., 8.5 mmol/8 h). However, in CH4 nonproducers perfused with mucin in the left colon, the mean total excess volume of H2 was only 31 ml (i.e., 1.4 mmol/8 h), which was almost the same as the volume of free Hz exhaled by the CH* producers. An interesting finding was that after mucin infusion into the right colon or after oral lactulose ingestion, the volumes of H2 exhaled in breath were also similar in CH, producers and nonproducers. Thus, in response to infusion of the endogenous substrate mucin into the left colon, total H2 excretion (calculated from free H2 and CHJ was six times higher in CH4 producers than nonproducers. Wolin and Miller (30) had calculated that this production should be 10 times higher in CH4 producers than nonproducers. They suggested that the rate of endogenous substrate supply is several times higher in CH4 producers than nonproducers, which in fact seems very unlikely because our experiments clearly show that in nonproducers the availability of the endogenous substrate mucin is not the rate-limiting step in Hz production. In fact, in CH4 nonproducers, Hz is probably used by the colonic microflora to reduce substrates to compounds other than CH*. In this connection, there is evidence that the colonic microflora produces acetate via a COa Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
SITE AND SUBSTRATES
FOR METHANE
large intestine. The carbohydrate moieties of gastrointestinal mucus have been shown to be degraded in vitro by the colonic flora from normal individuals (22). Using an in vitro fecal incubation system, Perman and Modler (22) showed that hog gastric mucin yielded less H2 than glucose; similarly, in vivo the volume of H2 produced from mucin was smaller in our study than that derived from lactulose. Here the volumes of Hz exhaled by the CH4 nonproducers during the probably complete catabolism of 22 g hog gastric mucin containing 16.5 g carbohydrate averaged 77 ml and 31 ml in the right and left colon, respectively. At this rate of Hz production, a daily colonic secretion of -47 g mucin (i.e., 35 g carbohydrate) would theoretically be needed to fill completely the gap between total pulm .onary excretion of H2 and the H2 from substrates of exogenous origin. This value is in rather good agreement with the estimate by Wolin and Miller (29, 30), who calculated from the fecal bacterial mass that the daily amount of carbohydrates required for colonic bacterial turnover was 66 g, with a carbohydrate supply from endogenous substrates equal to onehalf to two-thirds of this amount. We are grateful to C. Franchisseur, M. Maurel, and M.-C. Morin for expert technical assistance. Address for reprint requests: B. Flour%, INSERM U. 290, Hopital Saint-Lazare, 107 bis, rue du Fbg Saint-Denis, 75475 Paris Cedex 10, France. Received 27 September 1989; accepted in final form 1 January 1991. REFERENCES A. Structure of gastrointestinal mucus glycoproteins and the viscous and gel-forming properties of mucus. Br. Med. Bull. 34:
1. ALLEN,
B-33,1978. 2. BENTLEY, GORBACH.
D. W., R. L. NICHOLS, R. E. CONDON, AND S. L. The microflora of the human ileum and intra-abdominal colon: results of direct needle aspiration at surgery and evaluation of the technique. J. Lab. Clin. Med. 79: 421-429, 1972. 3. BJORNEKLETT, A., AND E. JENSSEN. Relationships between hydrogen (Hz) and methane (CH,) production in man. Stand. J. Gastroenterol. 17: 985-992, 1982. 4. BOND, J. H,, R. R. ENGEL, AND M. D. LEVITT. Factors influencing pulmonary methane excretion in man. J. Exp. Med. 133: 572-588, 197/l. 5. BOND, J. H., AND M. D. LEVITT. Effect of dietary fiber on intestinal gas production and small bowel transit time. Am. J. Clin. Nutr. 31, Suppl.: S169-S174,1978. 6. CHAUVE, A., G. DEVROEDE,
AND E. BASTIN. Intraluminal pressures during perfusion of the human colon in situ. Gastroenterology 70:
336-340,1976. 7. CLOAREC, D., F. BORNET, S. GOUILLOUD, AND J. P. GALMICHE. Breath hydrogen
healthy subjects: relationship
J. L. BARRY, B. SALIM, response to lactulose in to methane producing status. Gut 31:
300-304,199o. 8. COLOMBEL, J. F., B. FLOURIE, C. NEUT, C. FLORENT, A. LEBLOND, AND J. C. RAMBAUD. La methanogenese chez l’homme. Gastroenterol. Clin. Biol. 11: 694-700, 1987. 9. EVANS, D. F,, G. PYE, R. BRAMLEY, A. G. CLARK, T. J. DYSON, AND J. D. HARDCASTLE. Measurement of gastrointestinal pH profiles in normal ambulant human subjects. Gut 29: 195-216, 1988. 10. FLATZ, G., A. CZEIZEL, J. METNEKI, S. D. FLATZ, W. KUHNAU, AND D. JAHN. Pulmonarv hvdroeen and methane excretion follow-
PRODUCTION
G757
IN HUMANS
ing ingestion of an unabsorbable carbohydrate: a study of twins. J. Pediczt. Gustroenterol. Nutr. 4: 936-941, 1985. 11. FLOURIE, B., C. FLORENT, P. PELLIER, Y. BOUHNIK, AND J. C. RAMBAUD. Comparative study of hydrogen and methane production in the human colon by means of caecal and faecal homogenates. Gut 31: 684-685, 1990. 12. FLOURIE, B., C. FLORENT, F. ETANCHAUD, D. EVARD, C. FRANCHISSEUR, AND J. C. RAMBAUD. Starch absorption by healthy man evaluated by lactulose hydrogen breath test. Am. J. C&n. Nutr. 47: 61-66,1988. 13. FLOURIE, ALLI, AND
B., A. LEBLOND, C. FLORENT, M. RAUTUREAU, A. BISJ. C. RAMBAUD. Starch malabsorption and breath gas excretion in healthy humans consuming low- and high-starch diets.
Gastroenterology 14. GIBSON, G.
95: 356-363,
1988.
R., J. H. CUMMINGS, AND G. T. MACFARLANE. Use of a three-stage continuous culture system to study the effect of mucin on dissimilatory sulfate reduction and methanogenesis by mixed populations of human gut bacteria. Appl. Enoiron. Microbial. 54:
2750-2755,1988. 15. GIBSON, G.
R., G. T. MACFARLANE, AND J. H. CUMMINGS. Occurrence of sulphate-reducing bacteria in human faeces and the relationship of dissimilatory sulphate reduction to methanogenesis in the large gut. J. Appl. Bacterial. 65: 103-111, 1988. 16. HOSKINS, L. C., AND E. T. BOULDING. Mucin degradation in human colon ecosystems. Evidence for the existence and the role of bacterial subpopulations producing glycosidases as extracellular enzymes. J. Clin. Inuest. 67: 163-172, 1981. 17. LAJOIE, S. F., S. BANK, T. L. MILLER, AND M. J. WOLIN. Acetate production from hydrogen and [13C]carbon dioxide by the microflora of human feces. Appl. Environ. Microbial. 54: 2723-2727, 1988. 18. LEVITT, M. D., P. HIRSCH, C. A. FETZER, M. SHEAHAN, AND A. S. LEVINE. HZ excretion after ingestion of complex carbohydrates. Gastroenterology 92: 383-389, 1987. 19. MCKAY, L. F., W. G. BRYDON, M. A. EASTWOOD, AND J. H. SMITH. The influence of pentose on breath methane. Am. J. Clin. Nutr. 34: 2728-2733,198l. 20. MARTHINSEN, D., AND S. E. FLEMING. Excretion of breath and flatus gasses by humans consuming high-fiber diets. J. Nutr. 112: 1133-1143,1982. 21. MILLER, T. L., AND M. J. WOLIN. Enumeration of Methunobreuibatter smithii in human feces. Arch. Microbial. 131: 14-18, 1982. 22. PERMAN, J. A., AND S. MOLDER. Glycoproteins as substrates for
production 23.
24.
25.
of hydrogen and methane by colonic bacterial
flora.
Gastroenterology 83: 388-393, 1982. PIQUE, J. M., M. PALLARES, E. Cuso, J. VILAR-BONET, AND M. A. GASULL. Methane production and colon cancer. Gastroenterology 87: 601-605,1984. PITT, P., K. M. DE BRUIJN, M. F. BEECHING, E. GOLDBERG, AND L. M. BLENDIS. Studies on breath methane: the effect of ethnic origins and lactulose. Gut 21: 951-959, 1980. RUMESSEN, J. J., 0. HAMBERG, AND E. GUDMAND-HOYER. Inter-
val sampling of end-expiratory hydrogen (Hz) concentrations to quantify carbohydrate malabsorption by means of lactulose standards. Gut 31: 37-42,199O. 26. SALTZBERG, D. M., G. M. LEVINE, AND C. LUBAR. Impact of age, sex, race and functional complaints on hydrogen (Hz) production. Dig. Dis. Sci. 33: 308-313, 1988. 27. SAUNDERS, D. R., AND H. S. WIGGINS.
Conservation of mannitol, lactulose and raffinose by the human colon. Am. J. Physiol. 241
(Gastrointest. 28. STEPHEN,
Liver
Physiol.
4): G397-G402,
1981.
A. M., A. C. HADDAD, AND S. F. PHILLIPS. carbohydrate into the colon: direct measurements
Gastroenterology 29. WOLIN, M.
85: 589-595,
J. Fermentation
1983.
in the rumen and human large intes-
tine. Science Wash. DC 213: 1463-1468,198l. M. J., AND T. L. MILLER. Carbohydrate
30. WOLIN, Human
Intestinal
Microflora
Passage of in humans.
in Health
J. Hentees. New York: Academic. 1983.
and D.
fermentation. In: edited by D.
Disease, 147-165.
Downloaded from www.physiology.org/journal/ajpgi by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 15, 2019.
