A new technique to determine hydrogenexcreted by gnotobiotic rats
L. Hartmann, D. Taras, B. Kamlage & M. BlautDepartment of Gastrointestinal Microbiology, German Institute of Human Nutrition,Arthur-Scheunert-Allee 114-116, 14558 Potsdam-Rehbrucke, Germany
SummaryA new system, that allowed the monitoring of hydrogen (H2) excretion by gnotobiotic ratswithout affecting their defined microbial status, was developed. The system consists of anisolator containing a chamber for an experimental animal, and a life-support system (LSS).with a sampling port outside the isolator connected to it. H2 accumulation in the system wasmeasured by analysing a defined volume of gas after removal. H2 concentrations weredetermined with an electrochemical cell or by gas chromatography. To validate thistechnique, H2 excretion by germ-free (GF) and mono-associated rats fed a chemically defineddiet was measured after oral application of lactulose. Mono-associated rats had been obtainedby colonizing GF rats with a Hrproducing Clostridium per!ringens type A strain isolatedfrom human faeces of a healthy volunteer. Application of 50 mg lactulose to the mono-associated rats resulted in a significant increase in H2 excretion. The net H2 excretion was7.82 ± 1.28ml H2 in 12h corresponding to a net maximal rate of 1.1± 0.3 ml H2/h. In contrast,in experiments with GF rats, less than 0.13 ml H2 were detectable within 12h. The techniquepresented is a useful tool for studying bacterial H2 metabolism in vivo under gnotobioticconditions.
Keywords Gnotobiotic ratsj C. perfringensj colonic fermentationsj H2 excretion
Hydrogen production in the colon of humansand animals results from the fermentation ofsubstrates that are not absorbed in the smallintestine. H2 is a major component of theintestinal gas which is partly excreted inbreath and flatus, but to a large extent util-ized by other components of the gut micro-flora (Levitt 1969, Levitt et al. 1974). Accu-mulation of H2 in the large intestine may becaused by an increased H2 production asso-ciated with pathological conditions such ascarbohydrate malabsorption and by ingestionof various foods, such as beans and legumes,generally considered to cause gas formation(Bond & Levitt 1972, Wagner et al. 1977,Cummings & Macfarlane 1991). Impairedremoval of H2 may also result in H2 accu-
Correspondence to: Dr Ludger Hartmann
mulation in the gastrointestinal tract. Cer-tain antibiotics inhibit Hrutilizing intestinalbacteria thus increasing H2 excretion mark-edly (Levitt et al. 1974, Muir et al. 1996). Inthe case of patients with pneumatosiscystoides intestinalis, a disease characterizedby a massive H2 excretion, it was shown thatthe high levels of H2 formed are due to adeficiency in H2-utilizing microflora (Christlet al. 1993). Accordingly, an impaired utili-zation of H2 has been assumed to be the mainreason for flatulence (Strocchi & Levitt 1992).Recent work suggests that an effective H2
disposal during fermentation is essential fornormal large bowel function and thusimportant for human health (Gibson et al.1993). Utilization of H2 occurs pre-dominantly through the processes of metha-nogenesis, acetogenesis, and sulphate
reduction (Gibson et al. 1990, Christl et al.1992a). Human studies revealed that a cer-tain proportion of individuals do not harbourmethanogenic bacteria which catalyse thetransformation of H2 and carbon dioxide(C02) to methane. Individuals devoid ofmethanogens harbour sulphate-reducing orhomoacetogenic bacteria as the major Hrutilizing bacterial population groups (Christlet al. 1992a). Unlike methane and acetate,the product of sulphate reduction, H2S, hasthe potential to damage the colonic epithe-lium and was therefore suggested for playinga role in the pathogenesis of ulcerative colitis(Pitcher & Cummings 1996). Since metha-nogenesis and sulphate reduction are usuallymutually exclusive, a competition exists forthe common substrate H2 (Christl et al.1992b, Gibson et al. 1993). All processes ofmicrobial hydrogen disposal reduce the gasvolume in the intestinal tract and thus pre-vent excessive gas formation, but only a littleis known about the quantitative contributionof these processes and the factors that governtheir establishment or suppression inhumans and animals.
In rat experiments, H2 was collecteddirectly from the rat's intestine after analligation and after killing (Hedin & Adachi1962, Venkataraman & Jaya 1975). Ostranderet al. (1982) transferred anaesthetized ratsinto H2 collection chambers. A closed systemto collect H2 was used after instillation ofgases or lactulose into different parts of therat intestinal tract (Bond & Levitt 1972,Levitt & Levitt 1973, Levitt et al. 1974).Gumbmann and Williams (1971) described aclosed system equipped with a LSS, thissystem has been used and modified by severalinvestigators (Wagner et al. 1976, Fleming1980, Reddy et al. 1980, Phillips et al. 1988).However} these systems may be hampered bypoor reproducibility due to variations in theintestinal microflora (Fleming 1980, Phillipset al. 1988, Christl et al. 1995). The firstexperiments on H2 metabolism with gnoto-biotic animals were performed by Schulzeet al. (1995), who removed rats from thegnotobiotic isolator to measure H2 excretionin a H2 collection system.
Here we report on a new experimentalstrategy by which the total H2 excretion can
163
be monitored routinely in the gnotobiotic ratfed a chemically defined diet. Total H2
excretion measured included expired air andflatus. After the technical description of theexperimental system, results of a studyinvestigating the effect of lactulose on rats'mono-associated with a Hrproducing C.perfringens type A strain are shown.
Material and methods
Animals and housingRats of the inbred strain AVN/Ipcv/Wistar/Rehbriicke were raised under gnotobioticconditions. For gnotobiotic husbandry ratswere kept in Makrolon type III cages onirradiated wood chips at a regulated tem-perature (22± 2°C), a regulated relativehumidity (55±5%), and on a 12h light/darkcycle (07:00/19:00 h). The Makrolon cageswere placed in positive pressure isolators(Metall & Plastik, Radolfzell, Germanyl.Irradiated food (Altromin fortified® type1314, Altromin, Lage, Germany) and auto-claved distilled water were offered ad libi-tum. The GF state was confirmed at 2-weekintervals by controlling faecal samples of theanimals for the absence of any bacteria(Kunstyr 1992). The absence of rat specificviruses, bacteria, and protozoa was mon-itored by serological methods according tothe FELASA health monitoring recommen-dations (Kraft et al. 1994).
Six GF (8-week-oldl female rats with anaverage initial body weight of 209 ± 7.4 grepresented one experimental group whichunderwent various sequential treatmentsthat affected the diet and the microbialstatus. These rats were individually housedin wire-bottomed metabolism cages (STT200, Ebeco, Castrop-Rauxel, Germany). Thechemically defined basic diet (BD; Walzem& Clifford 1988) used during the experimentscontained 405.6 g sucrose, 210 g glucose,171.4 g L-amino acids (Merck, Darmstadt,Germany) 10g vitamins including choline(STD vitamin mix, Altromin) and 50 gmineral. mix per kilogram of diet. In addition,100 g sunflower seed oil, 50 g cellulose(Vivapur® type 102, Rettenmaier & S6hne,Ellwangen, Germany), and 3 g lactulose
Laboratory Animals (2000) 34
164
(Merckl were added per kilogram BD. Waterwas provided ad libitum to all animals. Lac-tulose (4-0-f3-D galactopyranosyl-D-fruc-tose), which is not absorbed, was added as afermentable carbohydrate that leads to H2
fermentation by intestinal bacteria. Both theBD and desired aliquots of lactulose weresterilized by irradiation prior to entry intothe isolator. Individual body weights wererecorded weekly with a spring balance.
Monitoring of H2 excreted by thegnotobiotic ratsThe system used to house the rat and tocollect the H2 is schematically shown inFig 1. In detail, the animal chamber was madeof 12.mm thick Plexiglas and consisted of alower box (inner dimensions 2.7.5x 2.0.1x19em), a Plexiglas lid (6mm thick) and arubber seal between box and lid. The cham-ber could be tightly closed by squeezingthe sealing ring with adjustable fastenersattached at all four edges of both the box and
Hartmann et al.
lid. A Makrolon type II cage within thischamber contained the animal during thetest periods. A metal mesh was placed intothis cage to provide a dry environment for theanimal. Water was placed in a plastic flask onthe wire-lid of the cage. A thermometer waspresent within the chamber to permit thecontrol of the temperature. Both chamberand Makrolon type II cage were transparentso that the animal could be observed. Leak-proof seals were made between the tubingand the chamber with rubber ferrules inSwadgelok fittings threaded through thePlexiglas. Gases were continuously circu-lated through iso-versinic-viton tubings(inner 0= 4.0, outer 0= 6.0mm, Ochs,Bovenden, Germany) or through high-gradesteel pipes (inner 0=4.0, outer 0=6.0mm).To connect the gnotobiotic with the con-ventional area, two high-grade steel pipeswere passed through a perforated plug in theisolator envelope. The air entering and leav-ing the isolator passed through autoclavablesterile filters (Midisart 2.000,Sartorius, Got-
ERILEFILTE
10 TOR
150-
P
1
L1F - PPORTS EM
GPORTRI GEl
MAG HIC L
OXYGCYLI DER
GELI
Fig 1 Scheme depicting the isolator and the life-support system to quantitatively collect and measurehydrogen gas excreted by the gnotobiotic rat. Gassamples were removed through the sampling port with 0.5or 20 ml gas-tight syringes, and analysed for H2 with a gas chromatograph (concentrations for >250 ppm H2) orwith an exhaled hydrogen monitor (concentrations for the range from 0-250 ppm H2), respectively (see'Material and methods' section for details). --+ indicates the direction of airflow
Laboratory Animals (2000) 34
Hydrogen excretion by gnotobiotic rats
tingen, Germany) placed immediately out-side the isolator. All components describedso far were sterilized before assembling.
The LSS,on the conventional side, was amodification of that described by Gumb-mann and Williams (1971).A brief descrip-tion pointing out the differences is given: Theadequate circulation of gases was maintainedby a gas tight and adjustable diaphragm pump(type R 409.1-150 G, Sera, Immenhausen,Germany). C02 was absorbed by passing theair through a gas wash bottle (500ml, screw-cap system, Schott, Mainz, Germany) filledwith 400ml 2.5 M potassium hydroxide(Merck), which was changed after eachexperiment. Moisture was removed by pas-sing the air through another gas wash bottlefilled with silica gel (Merck). Gas sampleswere removed through a sampling port withgas-tight syringes and the volume withdrawnwas replaced by room air. The consumptionof oxygen (02) by the animal and the con-tinuous absorption of respired CO2 created anegative pressure inside the system. In orderto replace the used-up O2, a manometricsensor was connected to the system. A pres-sure-decrease in the system registered by thesensor triggered a valve to open and to permitentry of O2 (grade 5.0, Linde, Berlin, Ger-many). Upon reaching normal pressure thevalve was closed again.
Hydrogen concentrations were determinedwith an electrochemical cell in a GMIExhaled Hydrogen monitor (GMI MedicalLtd, Renfrew, Scotland, UK) for the rangefrom 0-250 ppm H2. Sensitivity of the detec-tor is specified with 2ppm. Calibration wasdone daily with a 96.8ppm H2 standard (Sti-motron, Wendelstein, Germany), and roomair was set to 0ppm. The injection volumewas 20mL H2 levels> 250ppm were mea-sured with a Hewlett Packard gas chromato-graph (series 6890) equipped with a molecularsieve 5 A capillary column (30m by 0.32 mm;12Jlm film thickness). N2 (lml/min) wasused as the carrier gas. The temperatures ofthe oven and the thermal-conductivitydetector were 40°C and 205°C, respectively.The injection volume was 0.5m!.
The total volume of the system wasdetermined by introducing known volumesof pure H2 into the system and calculating it
165
from the H2 concentration measured. Thecomposition of the air circulating throughthe system during operation was controlledby collecting gas into a one litre gas washbottle and analysing it with a multiple gasanalyser (type MGA 1600, Perkin Elmer,Weiterstadt, Germany).
Bacterial strain and microbiologicaltechniquesRandomly picked bacterial isolates obtainedfrom human faecal samples were screened formaximal H2 production. H2 formation of theisolates on WCA medium jWilkins-Chal-gren-Anaerobe Broth, CM 643, Unipath Ltd,Basingstoke, England) was determined after24h of growth at 37°C. The cells were grownin rubber-stoppered tubes under strictlyanoxic conditions. Aliquots of the head spacewere taken by syringe and injected into thegas chromatograph. One of these strains wasidentified as C. perfringens type A strain. Thebiochemical features of C. perfringens type Awere determined with an automatic identi-fication system for bacteria according to theinstructions of the manufacturer (Vitcksystem, bioMerieux, Niirtingen, Germany).C. perfringens type A was able to grow ondissolved BD and to utilize lactulose, but notcellulose. Based on the toxin specification,as done by B. Kohler at the State Veterinaryand Food Investigation Office Potsdam(Germany), the isolate was identified as atype A strain of C. perfringens.
Colonization of the rats with C. perfrin-gens type A was monitored weekly by enu-merafion of viable bacteria in the faeces.Following homogenization of the samples, aseries of lO-folddilutions (10-5 to 10-10
) wasmade in an anoxic chamber (MK3 Anaerobicwork-station, dW Scientific, England) con-taining a gas atmosphere of 80% N2, 10%CO2, and 10% H2 (by volume). One hundredmicro litre aliquots of each dilution wereplated on SPSPerfringens-Selective agar(according to Angelotti, Merck). The inocu-lated plates were transferred to anaerobic jarscontaining 80% N2, 10% CO2, and 10% H2(by volume) and subsequently incubated at37°C for 24h. Bacterial counts are expressed
Laboratory Animals (2000) 34
166
as colony-forming units (CFU) per gram offaeces dry weight.
Experimental designThe interventions were done sequentiallywith one group of six female rats, first underGF and subsequently under mono-associatedconditions. H2 measuring tests were per-formed by placing one animal at a time intothe chamber of the H2 determination system.Before starting the experiment, the systemwas checked for the absence of H2 (Oppm).No BD was offered within the chamber. Oraladministration of lactulose or H20 was donewith a probang directly before placing theanimal into the H2 chamber and starting theexperiment. All experiments were started atthe same time of day (07:00 h) H2 measure-ments were done at hourly intervals over12h.
Time schedule: After an acclimatizationperiod which included the conditioning ofthe GF rats to the BD (days 1-5) all six ani-mals were tested for H2 formation followingthe administration of I ml water (days 6-11)and, subsequently, of 100mg lactulose dis-solved in 1ml water (days 15-20). The colo-nization of GF rats was done by offering BDsoaked with C. perfringens type A in medium(approximately 108 viable bacteria per ml) tothe 12h-fasted rats to ensure rapid intake(days 21-23). As soon as the faecal bacterialcell numbers were stable (day 26), the ratswere tested as described above by adminis-tering I ml water (days 27-32) and 50mglactulose dissolved in I ml water (days 36-41/·
System control tests: Leak tests (i.e. themaintenance of a defined H2 concentrationover time) were carried out four times(1 x 96 h, 3 x 48 h), the O2 determinationthree times (3 x 25 h), and the detection ofpossible other H2 sources once (48hI.
Data analysesResults are expressed as arithmetic meansand variations as standard deviation (± SD).The statistical analyses were done with theFriedman two-way ANOVA test for each ofboth treatments of the group colonized withC. perfringens. For comparison of both of
Laboratory Animals (2000) 34
Hartmann et a/.
these treatments the t-test for dependentsamples was performed Statistical analyseswere conducted with SPSS 8.0 software(SPSSInc. 1998).
ResultsEvaluation of the H2 monitoring systemPrior to the animal experiments, the systemwas tested for its ability to maintain definedH2 concentrations. Known aliquots of pureH2 were introduced into the system and theH2 concentration was determined over time.Operation of the system for up to 96 h did notresult in any changes in the H2 concentra-tion. These tests demonstrated that thesystem was gas-tight and that H2 was notgenerated by the components of the system.
The total volume of the system was 11litres. To ensure adequate gas circulation theflow rate was adjusted to 100 I/h, corre-sponding to approximately 9 volume changesper hour. Circulation was judged to be ade-quate since injection of pure H2 aliquots ledto an equal H2 distribution within 10min. Inaddition, water vapour produced by the ani-mals did not condense on the walls of thePlexiglas chamber during the experiments.
The O2 concentration in the system ana-lysed in the presence of an experimentalanimal after 5, 10, 15,20, and 25 h was 17-19.5%, indicating the proper functioning ofthe LSS. The humidity within the systemranged from 40 to 55%} and the temperaturein the animal chamber was constant at22±2°C during all experiments.
The perforated plug and the sterile filtersafforded the circulation of the gases betweenthe conventional and the gnotobiotic com-ponents of the system (Fig 1). Regularmicrobiological controls assured that theanimals were either GF or mono-associatedwith C. perfringens.
H2 excretion studiesGerm-free rats and rats colonized with C.perfringens were treated with either H20(control) or lactulose, and H2 formation wasdetermined. The excretion rates of thesetreatment groups were plotted versus time(Fig 2). Intragastric application of 50 mg
Hydrogen excretion by gnotobiotic rats 167
200
180gE 1600-S:C 140I/J~ 120+.2!~ 100c:0 80~()
~ 60N
:r: 40c:."(l)
20::E0
o 2 3 4 5 6 7 8 9 10 11 12
lime (h)
Fig 2 Hydrogen excretion rates of germ-free and mono-associated rats with time. GFor mono-associated rats(n =6) were treated with either 1ml H20 or lactulose as indicated. In the case of mono-associated rats, theincrease in the H2 excretion rate with time was significant at P < 0.01 (Iactulose experiment), the decrease in thecontrol experiment was significant at P < 0.05. Differences between these two treatment groups weresignificant at P < 0.01 for all time points ~ 5h. A = mono-associated + SOmg lactulose; .6. = mono-associa-ted + H20; 0 = GF+ H20; • = GF+ 100 mg lactulose
lactulose to the mono-associated rats led to asignificant increase in the H2 excretion rateas compared with the H20 controls: Twohours after lactulose treatment, the H2excretion rate increased steadily and peaked8h after start of the experiment at165.3±32.9ppm H2/h (1.82±O.36ml H2/h).Until end of the experiment, the ratedecreased to 111.2± 17.5 ppm H2/h(1.22±O.19ml H2/h). The initial H2 excre-tion rates of the H20-treated mono-asso-ciated rats were similar to those of thelactulose group (81.7± 5ppm H2/hj
0.89 ± 0.05 ml H2/hl, but declined sub-sequently to 59.5± 15.1ppm H2/h(0.65±0.16ml H2/h) at 12h. The total H2concentration measured after 12h for thelactulose-treated mono-associated rats was1476±81.9ppm H2 (16.23±0.90ml H2), and766.2± 102.7 ppm H2(8.42± 1.12ml H2) forthe H20-treated mono-associated rats. Thus,the volume of H2 proquced from 50 mg lac-tulose was 7.82± 1.28ml H2. In contrast, theGF rats accumulated less than 12ppm(0.13ml H2) within 12h, no matter whetherthe GF rats had been treated with H20 orwith lactulose (100mg).
Bacterial counts in the faecal samples ofthe rats mono-associated with C. perfringensranged from i08 to 109 CFU/g of dry weight.The average body weight of the rats was209.2± 7.4g on the first day and 233.3±6.1 gon day 42. During all experiments there wasno evidence of indisposition or any outwardsigns of stress observed in i:he animals.
DiscussionHydrogen production in the rat is a predictivemodel for flatulence in man (Wagner et al.1977). The study of H2 excretion affords aninsight into the in vivo metabolic process ofH2 formation by the intestinal flora and pro-vides information of both nutritional andclinical relevance. Therefore, the objective ofthe present work was to develop an animalmodel which enables the determination ofexcreted H2 when microbial and nutritionalcomponents are exactly defined. In parti-cular, the complexity of the intestinalmicroflora has often led to unsatisfactoryresults in earlier human or animal studies(Phillips et al. 1988, Christl et al. 1995).Gumbmann and Williams (1971) showed that
Laboratory Animals (2000) 34
168
H2 excretion by rats on a certain diet variedconsiderably both within a rat and betweenrats. This variability in conventional animalscan be reduced when gnotobiotic animals areemployed. Additional complications may arisefrom the use of diets of complex composition;we therefore used a chemically-defined diet inour study. This diet had to be enriched with anon-digestible carbohydrate that could sup-port growth of C. perfringens Lactulose waschosen because it passes unchanged into thecolon where it undergoes bacterial fermenta-tion accompanied by H2 formation. The colo-nic fermentation of more complex carbo-hydrates such as starch or pectin generates lessH2 per mol of hexose moiety than that of lac-tulose (Christl et al. 1992a). Gas formation inresponse to lactulose intake is also of interest,since it is used as a drug in the treatment ofchronic constipation and portal systemicencephalopathy (Ballongue et al. 1997).
Even though C. perfringens type A wasof human origin, it colonized the intestinaltract of GF rats easily as evident from thebacterial counts in the faeces of all six rats(108 to 109 CFU/g of faeces dry weight).Other indigenous species of the humanintestinal microflora such as Bacteroidesvulgatus might occur in higher numbers thanC. perfringens type A, but most of theseproduce no or considerably less H2.
The mammalian cellular metabolism doesnot liberate or utilize H2 due to the absenceof any hydrogenases. In contrast to this,traces of H2 could be detected in GF controlswithout any significant difference betweenH20 and lactulose-treated rats. These tracesof H2 may have resulted from the commer-cially available O2, which enters the systemwhen an animal is present. This explanationagrees with the absence of H2 production inGF rats (Levitt et al. 1968).
The rats associated with C. perfringens typeA and fed BD excreted 0.89± 0.05 ml H2/h atthe start of the experiment. The H2 resultedprobably from the fermentation of the lactu-lose present in BD (3mg/g). The estimatedintake of 5-10 g BD, which corresponds to anintake of 15-30 mg lactulose per day and per ratoccurred at night. The rate of H2 excretion ofH20-treated mono-associated rats declinedslightly during the 12h tests (0.65± 0.16 ml
Laboratory Animals (2000) 34
Hartmann et al.
H2/h) because the rats were not fed during thattime. The gavage of 50mg (0.15mmol) lactu-lose to the mono-associated rats increased theformation of H2 significantly. The net volumeof H2 excreted within 12h was 7.82± 1.28 ml(0.322± 0.06 mmol, 22°C). Variability of H2excretion was probably due to the fact thatpulmonal H2 excretion occurs continuouslywhile flatus does not.
From these data, a minimal yield of 156mlH2/g lactulose can be calculated. From athermodynamic point of view, Clostridiumspecies, in pure culture, can liberate H2 dur-ing carbohydrate metabolism only fromreduced ferredoxin produced in the pyruvate:ferredoxin-oxidoreductase reaction. Thiscorresponds to 4 mol H2/mollactulose or283 ml H2 (at 22°C)/g lactulose, respectively.Taking into consideration that the bacteriado not ferment the lactulose completely, butassimilate a considerable proportion for ana-bolic purposes, our data are in reasonableagreement with the theory.
Christl et al. 1995 estimated that thehuman colonic fermentation is accompaniedby the formation of 50-250ml H2/g fer-mented sugar. Another study revealed that H2production in non-methanogenic humans isabout 20ml H2/g lactulose (Christl et al.1992a). Intracaecal instillation of 250 mg lac-tulose in rats resulted in a H2 excretion rate of0.06 ml H2/h (Bond & Levitt 1972). Levitt etal. (19741showed that the H2 excretion rate ofrats fed a rat chow diet increased by a factor ofmore than 12 upon addition of neomycinindicating that bacteria involved in H2 utili-zation were inhibited primarily. These stu-dies indicate that the intestinal H2
metabolism is difRcult to investigate due tothe complexity of the ecosystem. How muchthe various H2-utilizing bacteria contributedto H2 removal could only be roughly esti-mated. In contrast, the system presented hereallows us to analyse in detail between Hz-producing and H2-consuming bacterial popu-lation groups. Rats may also be associatedwith a human normal micro flora from dif-ferent subjects in order to assess H2 excretionin response to different substrates. However,conclusions from such investigations may beof limited value because important speciesmay be lost during association.
Hydrogen excretion by gnotobiotic rats
The exact stoichiometry of gas productionor gas utilization in the colon of mono-gastricmammals is not known at present, our modelappears to be suitable to accurately deter-mine these parameters.
The body weights of the rats during theexperiments agreed with the findings of Wal-zem and Clifford (1988),who reported a max-imal growth rate of rats receiving a similar diet.There was no indication that the mono-asso-ciation or other experimental manipulationsaffected the weight of the rats. The influence ofthe animal weight on H2 excretion may beneglected, because H2 formation by the mono-associated rats was determined during the last2 weeks of the experiment. During this timethe increase in body weight was n'egligible.Moreover, no correlation exists between H2
excretion and the body mass of adult femalerats (Phillips et a1. 1988).
Taken together, the system described isnon-invasive and safe to operate, allowingthe accurate investigation of the H2 metabo-lism of both gnotobiotic and conventionalrats. The measurements can be extended toother gases such as methane. The system canalso be applied to other gnotobiotic animalssuch as minipigs after modification of thesystem to the specific requirements. Thetechnique presented is most useful forstudying bacterial gas metabolism in vivoand its modification by various nutrients ordrugs, the digestibility of different food .materials, and the causative factors of flatu-lence. Above all it is a useful tool for gainingdetailed in vivo information about the rela-tive contribution of H2'producing and H2-
oxidizing bacteria to the H2 metabolism inthe intestine of small animals.
Future studies will investigate microbialcommunities of known composition includ-ing the recently described homoacetogenic C.coccoides (Kamlage et a1. 1997), methano-genic and sulphate-reducing microorganisms.Improvement of the system, by introducing adevice that allows continuous registration ofthe H2 concentration, is in progress.
Acknowledgments The authors are indebted toS. Dietrich, R. Herzog, and U. Lehmann for excellenttechnical assistance, and to T Roeder for contribut-ing to the construction of scientific apparatus, We
169
also thank U, Frenz for work with the multiplegas analyser. This work was supported by grantBL 257/6-1 from the Deutsche Forschungsgemeinschaft.
References
Ballongue J, Schumann C, Quignon P (1997) Effects oflactulose and lactitol on colonic microflora andenzymatic activity. Scandinavian Journal ofGastroenterology 32, (Suppl 222), 41-4
Bond TH, Levitt MD (1972) Use of pulmonaryhydrogen (HzI measurements to quantitate carbo-hydrate absorption. Journal of Clinical Investiga-tion 51, 1219-25
Christl SU, Murgatroyd PR, Gibson GR, CummingsTH (1992al Production, metabolism, and excretionof hydrogen in the large intestine. Gastroenterol-ogy 102, 1269-77
Christl SU, Gibson GR, Cummings JH (1992bl Role of'dietary sulphate in the regulation of methanogen-esis in the human large intestine. Gut 33, 1234-8
Christl SU, Gibson GR, Murgatroyd PR, ScheppachW, Cummings JH (1993) Impaired hydrogen meta-bolism in pneumatosis cystoides intestinalis.Gastroenterology 104, 392-7
Christl SU, Scheppach W, Kasper H (1995) Wasser-stoffmetabolismus im Dickdarm-Physiologie undklinische Bedeutung. Zeitschrift fUr Gastroenter-ologie 33, 408-13
Cummings JH, Macfarlane GT (1991) A review: Thecontrol and consequences of bacterial fermentationin the human colon. Journal of Applied Bacteriol-ogy 70, 443-59
Fleming SE 11980) Measurement of hydrogen produc-tion in the rat as an indicator of flatulence activity.Journal of Food Science 45, 1012-15
Gibson GR, Cummings JH, Macfarlane CT, AllisonC, Segal I, Vorster HH, Walker ARP 119901Alter-native pathways for the hydrogen disposal duringfermentation in the human colon. Gut 31, 679-83
Gibson GR, Macfarlane GT, Cummings JH 119931Sulphate reducing bacteria and hydrogen metabo-lism in the human large intestine. Gut 34, 437-9
Gumbmann MR, Williams SN (19711 The quantita-tive collection and determination of hydrogen gasfrom the rat and factors affecting its production.Proceedings of the Society for Experimental Biolo-gical Medicine 137, 1171-5
Hedin PA, Adachi RA (1962) Effect of diet and time offeeding on gastrointestinal gas production in rats.Tournal of Nutrition 77, 229-36
Kamlage B, Gruhl B, Blaut M (1997) Isolation andcharacterization of two new hom 0aceto genichydrogen-utilizing bacteria from the humanintestinal tract that are closely related to Clostri-dium coccoides. Applied and EnvironmentalMicrobiology 63, 1732-8
Kraft V, Deeny AA, Blanchet HM, Boot R, Hansen AK,Hem A, van Herck H, KunstyrI, Milite G,
Laboratory Animals (2000) 34
170
Needham JR, Nicklas W, Perrot A, Rehbinder C,Richard Y, de Vroey G /1994/ Recommendations forthe health monitoring of mouse, rat, hamster,guinea pig, and rabbit breeding colonies. LaboratoryAnimals 28, 1-12
Kunstyr I (1992) In: Diagnostic Microbiology forLaboratory Animals, Vol II. Stuttgart: GustavFischer Verlag, GV-SOLAS, pp 42-4
Levitt MD, French P, Donalldson RM (19681 Use ofhydrogen and methane excretion in the study of theintestinal flora (Abstractl. TournaI of Laboratoryand Clinical Medicine 72, 988
Levitt MD (19691 Production and excretion of hydro-gen gas in man. New England TournaI of Medicine281, 122-7
Levitt MD, Levitt DG (1973) Use of inert gas to studythe interaction of blood flow and diffusion duringpassive absorption from the gastrointestinal tract ofthe rat. TournaI of Clinical Investigation 52, 1852-62
Levitt MD, Berggren T, Hastings 1, Bond JH (1974)Hydrogen (H2) catabolism in the colon of the rat.TournaI of Laboratory and Clinical Medicine 84,163-7
Muir P, Gruffydd-Jones T1, Cripps PJ, Brown PJ 119961Breath hydrogen excretion by healthy cats after oraladffiinistration of oxytetracycline and metronida-ziole. Veterinary Record 138, 635-9
Ostrander CR, Stevenson DK, Neu J, Kerner JA,Moses SW 119821A sensitive analytical apparatusfor measuring hydrogen production rates. I. Appli-cation to studies in small animals. Evidence of theeffects of an o::-glucosidehydrolase inhibitor in therat. Analytical Biochemistry 119, 378-86
Laboratory Animals (2000) 34
Hartmann et a/.
Phillips RD, Abbey BW, Nnanna 1A (1988) A systemfor determining flatus (hydrogen) production inlaboratory rats. Nutrition Reports International37,623-8
Pitcher MCL, Cummings JH 119961Hydrogensulphide: a bacterial toxin in ulcerative colitis?Gut 39, 1-4
Reddy NR, Salunkhe DK, Sharma RP (1980) Flatu-lence in rats following ingestion of cooked andgerminated black gram and a fermented product ofblack gram and rice blend. TournaI of Food Science45, 1161-4
Schulze 1, Schmidt E, Lorenz A, Zunft H-J (1995)Hydrogen output of gnotobiotic and conventionalrats. Microecology and Therapy 21, 70-81
Strocchi A, Levitt MD (19921 Maintaining intestinalH2 balance: credit the colonic bacteria. 'Gastroen-terology 102, 1424-6
Venkataraman LV,Jaya TV (19751 Gastrointestinal gasproduction in rats fed on diets containing germi-nated legumes. Nutrition Reports International12,387-98
WagnerJR, BeckerR, GumhmannMR, OlsonAC (1976)Hydrogen production in the rat following ingestion ofraffinose, stachyose and oligosaccharide-free beanresidue. TournaI of Nutrition 106, 466-70
Wagner JR, Carson JF, Becker R, Gumbmann MR,Danhof IE (1977) Comparative flatulence activity ofbeans and bean fraction for man and the rat. Journalof Nutrition 107, 680-9
Walzem RL, Clifford AT (1988) Folate deficiency inrats fed diets containing free amino acids or intactproteins. TournaI of Nutrition 118, 1089-96
