SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

It is increasingly understood that gastrointestinal (GI) methanogens, including Methanobrevibacter smithii, influence host metabolism.

Objective:

Therefore, we compared M. smithii colonization and weight gain in a rat model under different dietary conditions.

Design and Methods:

Sprague-Dawley rats were inoculated with M. smithii or vehicle (N = 10/group), fed normal chow until day 112 postinoculation, high-fat chow until day 182, then normal chow until day 253. Thereafter, five rats from each group were fed high-fat and normal chow until euthanasia.

Results:

Both groups exhibited M. smithii colonization, which increased following inoculation only for the first 9 days. Change to high-fat chow correlated with significant increases in weight (P < 0.00001) and stool M. smithii (P < 0.01) in all rats, with stool M. smithi decreasing on return to normal chow. Rats switched back to high-fat on day 253 further increased weight (P < 0.001) and stool M. smithii (P = 0.039). Euthanasia revealed all animals had higher M. smithii, but not total bacteria, in the small intestine than in the colon. Rats switched back to high-fat chow had higher M. smithii levels in the duodenum, ileum, and cecum than those fed normal chow; total bacteria did not differ in any bowel segment. Rats which gained more weight had more bowel segments colonized, and the lowest weight recorded was in a rat on high-fat chow which had minimal M. smithii colonization.

Conclusions:

We conclude that M. smithii colonization occurs in the small bowel as well as in the colon, and that the level and extent of M. smithii colonization is predictive of degree of weight gain in this animal model.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The human gastrointestinal (GI) tract is host to a vast number of microorganisms, which include archaea, bacteria, and eukaryotes. To date, at least 70 divisions of bacteria and 13 divisions of archaea have been identified, and their collective genome (the microbiome) is believed to contain 100 times more genes than the human genome (1, 2). Although the composition and number of microbes in the gut depends on many factors (3, 4), by adulthood most humans reach an established, relatively stable balance of type, and numbers of microbes that is unique to a given individual (5). This microbial community is thought to develop with the host by establishing symbiotic relationships which favor their coexistence (3, 6), such as assisting the host in the breakdown of food for absorption and elimination (7). While the full breadth of the impact of these gut microbes on the human host will take years to uncover, the complex and often interdependent relationships between gut microbes and the human host have been of increasing scientific interest this past decade, and this interest continues to grow. In particular, there is ample and growing evidence to suggest potential roles for gut microbes in energy homeostasis, inflammation, and insulin resistance (8, 9, 10), and as a result, gut microbes have been considered as possible causative factors of metabolic conditions and obesity, as well as potential therapeutic targets (11-15).

The methanogenic archaea (methanogens) are important constituents of the human gut microbiota. This distinct group grows primarily under anaerobic conditions, and produces methane (CH4) as a byproduct of fermentation. Methanogens are unique in that their metabolism increases in the presence of products from other gut microbes (16), as they scavenge hydrogen and ammonia as substrates for the generation of methane (17, 18). Once absorbed into systemic circulation, methane is cleared via the lungs. The majority of methanogens found in the human gut are from the genus Methanobrevibacter; predominantly Methanobrevibacter smithii (7). M. smithii is found in 70% of human subjects, and analysis of expiratory methane by lactulose breath testing can serve as an indirect measure of methane production (7, 19). A minority of subjects (15%) produce large quantities of methane early in the breath test, suggesting a greater methane potential (20), and increased methane production on breath test correlates with increased levels of M. smithii in stool, as determined by quantitative PCR (qPCR) (20, 21).

Significantly, studies in animal models have suggested a potential role for M. smithii in the development of obesity. Introduction of both a Bacteroides species (Bacteroides thetaiotaomicron) and M. smithii into germ-free mice resulted in greater body weights than with B. thetaiotaomicron alone (22), and methanogens have been shown to increase the capacity of polysaccharide-metabolizing bacteria to digest polyfructose-containing glycans in the colons of germ-free mice (22), suggesting that methanogens may play a role in caloric harvest. In humans, we have recently found that increased methane on breath test is associated with a higher average BMI, both in normal population and in obese subjects. In the obese population, methane was associated with a remarkable 6.7 kg/m2 greater BMI compared to non-methane controls (P < 0.05) (23). While these data are suggestive of a role for methanogens in caloric harvest and weight gain in humans, this is weakened by the fact that, to date, colonization with methanogens has only been demonstrated in the large bowel (24, 25, 26). Therefore, we tested and compared weight gain and the location and extent of M. smithii colonization in the GI tracts of rats under different dietary conditions.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

M. smithii hypercolonization of rats

Twenty adult Sprague-Dawley rats were obtained as 21-day-old weanlings (Harlan Labs, Indianapolis, IN). After 3 days of quarantine, all rats were weighed, and then received a 1 ml aliquot of 5% sodium bicarbonate by oral gavage using a ball-tipped inoculating needle, in order to neutralize the gastric acid. After ∼20 min, one group of rats (n = 10) received a 0.5 ml gavage of M. smithii in liquid growth media. A second group of rats (n = 10) received a 0.5 ml gavage of liquid growth media. After another 20 min, rats gavaged with M. smithii were given 0.2 ml enemas of the same inoculate following isoflurane anesthesia in a dessicator jar. The gavage and enemas were performed to determine whether M. smithii levels in intestine could be enhanced through hypercolonization.

Tracking colonization by M. smithii

After inoculation, all rats were housed two per microisolater cage under standard vivarium procedures and maintained on normal rodent chow (5.7% fat) (Lab Rodent Diet 5001; Newco Distributors, Rancho Cucamonga, CA). Fresh stool samples were collected daily for the first week, and then approximately every 2 weeks thereafter. The stool specimens on specific weeks were tested for the levels of M. smithii and of total bacteria by performing qPCR as previously described (27). M. smithii levels were quantitated using primers for the RpoB gene (5′-AAG GGATTTGCACCCAACAC-3′ (forward) and 5′-GACCACAGTTAGGACCCTCTGG-3′ (reverse)) and total bacteria were quantitated using 16S recombinant DNA (5′-TCCTACGGGAGGCAGCAGT-3′ (forward) and 5′-GGACTACCAGGGTATCTAATCCTGTT-3′ (reverse)). Animal weights were also obtained once a week.

Diet manipulation

Rats were observed until three consecutive weights were obtained within 10 g to suggest an end of growth curve and arrival at adult weight (corresponded to day 112). On day 112, all rats were then switched to a high-fat diet (34.3% fat) (Teklad high fat diet TD.06414; Harlan Laboratories, Madison, WI) and maintained on this diet for 10 weeks until day 182. Fresh stool samples and animal weights were collected from all animals on a weekly basis. On day 182, all rats were returned to normal chow. Finally, on day 253, five rats from each group were again fed high-fat chow. The rats were maintained on their respective diets while stool samples and weekly weights continued to be obtained for 5 weeks until euthanasia at day 287. This last phase was to guarantee that 10 rats were on high-fat and 10 on normal chow for a period of time before euthanasia.

Euthanasia and bowel sampling

On day 287 postinoculation, all rats were euthanized by CO2 asphyxiation and pneumothorax. Laparotomy was performed and sections of the left colon, cecum, ileum, jejunum, and duodenum were resected from each rat as previously described (27). DNA was extracted from luminal contents of each segment as previously described (27), and qPCR with M. smithii-specific and universal bacterial primers was performed to determine the levels of M. smithii and total bacteria, respectively in each segment. The study protocol was approved by the Cedars-Sinai Institutional Animal Care Utilization Committee (IACUC).

Statistical analysis

The levels of M. smithii in stool by qPCR between inoculated and noninoculated rats were compared by Mann-Whitney U-test. Comparisons of body weight before and after diet changes were compared by paired t-test. Levels of M. smithii in bowel segments or stool between groups were again compared by Mann-Whitney U-test. For comparison of M. smithii levels before and after an intervention, Wilcoxon signed-rank test was used. For weights, data were expressed as mean ± s.d. and data for M. smithii levels were expressed as mean ± s.e. Statistical significance was determined by P < 0.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Colonization of rats with M. smithii

At baseline, all rats demonstrated the presence of M. smithii in the stool which were not different between groups (Figure 1). After inoculation with M. smithii, rats demonstrated an increased detection of stool M. smithii than control animals. However, this did not persist as levels returned to control levels by day 9 (Figure 1). Since hypercolonization did not occur, in the remaining experiments all of the rats were examined as a single group.

thumbnail image

Figure 1. Effect of Methanobrevibacter smithii gavage on stool quantity of this species over time (before diet manipulation). *P < 0.01.

Download figure to PowerPoint

M. smithii levels and weight after initial transition to high-fat diet

All rats were initially fed normal rat chow until three consecutive weights were obtained within a 10 g plateau to suggest the rats had reached adult weight. This plateau occurred in the 2 weeks preceding day 112 (Figure 2a). During three consecutive measurements obtained between days 98 and 112, there was only a mean change in weight of 5.5 ± 5.8 g. After switching to high-fat chow on day 112, a sudden increase in rat weights was observed (Figure 2a). The average weight increased from 268 ± 13 g on day 112 to 292 ± 16 g on day 119 (P < 0.00001). This resulted in a 1-week increase in weight of 23.2 ± 9.5 g from day 112 to 119, as compared to 5.1 ± 5.4 g in the preceding week (P < 0.00001). Despite continuing on this high-fat diet, by day 182 the rats weighed 296 ± 22 g, which was not statistically different from their weights 1 week after starting on high-fat chow (P = 0.39).

thumbnail image

Figure 2. Effects of dietary fat content on rat weights and stool Methanobrevibacter smithii levels. (a) Rat weights over time starting from the adult weight plateau. *P < 0.00001 change in weight after 1 week on high-fat diet. P < 0.001 change in weight after return to high-fat diet. (b) Methanobrevibacter smithii levels over time. *P < 0.01 for increase in stool M. smithii after starting on high-fat chow. P < 0.001 for decrease in stool M. smithii after return to normal chow. P = 0.039 for increase in stool M. smithii after return to high-fat chow.

Download figure to PowerPoint

In addition to the weight change seen after switching to high-fat chow, stool M. smithii levels also increased suddenly after the high-fat diet was implemented (Figure 2b). M. smithii levels were 5.6 × 104 ± 2.8 × 103 cfu/ml which increased by nearly 1 log to 3.0 × 105 ± 7.0 × 103 cfu/ml after 1 week of high-fat diet (P < 0.01) (Figure 3a). Like the change in body weight, the change in M. smithii occurred in 1 week and did not further increase with additional weeks on high-fat chow (Figure 3a).

thumbnail image

Figure 3. Effect of high-fat diet on stool Methanobrevibacter smithii levels. (a) Methanobrevibacter smithii levels before, 1 week after, and 5 weeks after high-fat diet. (b) Stool M. smithii levels and the degree of weight gain. Comparing weight gain from day 98 to day 154. (c) Effect of returning to high-fat chow on stool M. smithii levels.

Download figure to PowerPoint

In another analysis, rats were divided into groups based on those that gained more or less weight with high-fat. In this analysis, rats which had >10% weight gain with high-fat had higher stool levels of M. smithii than rats which gained less weight (<10% weight gain) (P = 0.08, Figure 3b).

M. smithii and body weight on returning to normal diet

On returning to normal chow after 10 weeks of high-fat diet on day 182, rats did not experience a reduction in body weight (Figure 2a). As depicted in this figure, the rat weights remained at a plateau. However, the return to normal chow resulted in a gradual reduction in stool M. smithii levels over time (Figure 2b). On day 189, 1 week after cessation of fat and resumption of normal chow, M. smithii levels were 3.4 × 103 ± 8.1 × 102 cfu/ml, which was significantly reduced from day 154 (P < 0.001) (Figure 2b). Stool M. smithii levels continued to decline in rats continued on normal chow to the end of the study (2.0 × 102 ± 2.0 × 102 cfu/ml) (P < 0.05) (Figure 2b).

Randomizing back to high-fat diet a second time

In the final phase of the study, rats were randomized into two groups (10 rats returned to high-fat chow and the other 10 continued on normal chow). While Figure 2a suggests that returning to high-fat chow did further increase body weight compared to continuing on normal chow, the differences in weights between the two groups (high-fat vs. normal chow) did not reach statistical significance for any timepoint. However, the 10 rats returned to high-fat chow exhibited an increase in average weight from 292 ± 16 g to 319 ± 26 g, which was significant (P < 0.001, Figure 2a). The return to high-fat chow also resulted in an increase in M. smithii levels in these animals (P = 0.039, Figure 3c).

Bacteria and M. smithii levels by bowel segment post-mortem

Following euthanasia on day 287 postinoculation, sections of the left colon, cecum, ileum, jejunum, and duodenum were resected from each rat, and DNA was extracted from luminal contents of each segment. qPCR with M. smithii-specific and universal bacterial primers was used to determine the levels of M. smithii and total bacteria, respectively. Surprisingly, the highest levels of M. smithii were found in the small intestine, and were most elevated in the ileum (Figure 4a). In contrast, total bacterial levels were lowest in the small intestine, and highest in the cecum and left colon (Figure 4b). When the levels of M. smithii in each bowel segment were compared for rats switched to a high-fat diet in the final phase of the study vs. those maintained on normal chow, higher M. smithii levels were identified in all bowel segments of rats switched to a high-fat diet (Figure 5a). However, only the duodenum, ileum, and cecum reached statistical significance. In contrast, no significant differences in total bacterial levels were identified between rats switched to a high-fat diet compared to those maintained on normal chow in any bowel segment (Figure 5b).

thumbnail image

Figure 4. Methanobrevibacter smithii and total bacterial levels by segment of bowel. (a) Methanobrevibacter smithii by segment of bowel post-mortem. P < 0.001 between ileum and cecum and left colon; P = 0.03 comparing ileum to jejunum and P = 0.07 comparing ileum to duodenum. (b) Total bacteria by segment of bowel post-mortem.

Download figure to PowerPoint

thumbnail image

Figure 5. Effects of dietary fat content on Methanobrevibacter smithii and total bacterial levels in the bowel. (a) Methanobrevibacter smithii throughout the bowel by diet. *P < 0.05. (b) Total bacteria throughout bowel by diet. None of the comparisons were significant.

Download figure to PowerPoint

Correlation between extent of bowel colonization with M. smithii and weight in rats

The final comparison was to examine the distribution of M. smithii in the GI tract as a determinant of body weight. Although not statistically significant, rats with the greatest extent of M. smithii colonization (i.e., those with no uncolonized bowel segments) had higher weights than those with less widespread M. smithii colonization (i.e., those with one or more uncolonized bowel segments), irrespective of whether or not they were on high-fat chow in the final phase of the study (Figure 6). The lowest body weight of all rats was recorded for a rat on high-fat chow that had three bowel segments out of five lacking M. smithii colonization.

thumbnail image

Figure 6. Number of segments with no Methanobrevibacter smithii colonization and weight. Trend is not statistically significant.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In this study, we demonstrate for the first time that colonization of the rat gut with the methanogen M. smithii is not limited to the large bowel, but rather extends to the small bowel, including the ileum, jejunum, and duodenum. In fact, we found that the levels of M. smithii were higher in the small bowel than in the large bowel, with the most elevated levels seen in the ileum. Moreover, we found that switching rats to a high-fat diet resulted both in increased levels of M. smithii in stool, and in increased levels of M. smithii in all bowel segments tested. Most significantly, we found that rats which gained more weight had higher stool levels of M. smithii than rats which gained less weight, and that the extent of colonization of the bowel with M. smithii colonization also corresponded with weight gain in these rats, irrespective of diet. Taken together, these findings support that the level and extent of colonization of the intestinal tract with M. smithii is predictive of the degree of weight gain in this animal model.

It is becoming increasingly understood that gut microbes play roles in and affect host metabolism and energy homeostasis, and may thus contribute to the development of metabolic disorders and obesity. Through the production of enzymes, gut microbes assist the host in: utilizing non-digestible carbohydrates and host-derived glycoconjugates, resulting in increased short-chain fatty acid (SCFA) production; deconjugating and dehydroxylating bile acids, which alters the solubilization and absorption of dietary lipids; and cholesterol reduction and biosynthesis of vitamins from the K and B group, isoprenoids and amino acids such as lysine and threonine (1, 11, 28, 29). Gut microbes have also been suggested to affect intestinal transit times, and to contribute to the chronic low-grade inflammation and insulin resistance that are associated with obesity via effects on the endotoxin toll-like receptor 4 axis and intestinal barrier function (14, 30). That gut microbes directly affect host metabolism and weight gain is demonstrated by the fact that obesity is associated with increases in the relative abundance of Firmicutes and reductions in Bacteroidetes in both mice (31) and humans (12), and that while germ-free mice weigh less (and have less body fat) than mice with normal gut flora, colonization of these animals with gut microbes from lean mice resulted in dramatic increases in body weight (11), an effect which was even more pronounced when gut microbes from genetically obese (ob/ob) mice were used (9). These data support a role for gut microbes in contributing to weight gain by the host, which validates our finding that weight gain in our animal model was more dependent on the degree and extent of M. smithii colonization of the gut than on dietary fat content.

Several lines of evidence support that methanogens may play a specific role in host metabolism and energy homeostasis. Methanogens such as M. smithii, which is the most common methanogen in the human gut, produce methane through anaerobic fermentation (17, 18), and remove hydrogen atoms and accelerate the fermentation of polysaccharides and carbohydrates (22). This increases the production of SCFAs, which are subsequently absorbed in the intestines and serve as an additional energy source for the host (11). This more efficient energy extraction may lead to weight gain and ultimately contribute to obesity (32). This is supported by evidence generated by Samuel and Gordon (22) using a germ-free animal model, which showed that introduction of a single Bacteroides species (B. thetaiotaomicron) with M. smithii into these mice resulted in greater body weight than introduction of B. thetaiotaomicron alone (22). One potential mechanism for this is through effects of SCFAs on G protein-coupled receptors, for which they act as ligands. The G protein-coupled receptor Gpr41 is expressed in the intestine, colon, and adipocytes, and stimulates the expression of the adipokine leptin and the peptide tyrosine-tyrosine (peptide-YY), which both influence energy metabolism, and also affect appetite levels/satiety. In addition, modulation of plasma SCFAs has been linked to decreases of inflammatory markers in insulin-resistant human subjects (33, 34), suggesting a potential effect on the chronic low-grade inflammation associated with obesity. Interestingly, in a recent human study we found that during a 75 g oral glucose tolerance test, methane-producing subjects (i.e., those with increased methane on breath test) had greater serum glucose area-under-the-curve than non-methane subjects, despite having comparable BMIs and baseline insulin resistance (homeostatic model assessment-insulin resistance), suggesting that intestinal methane-producing subjects may have impaired glucose tolerance when challenged with a high carbohydrate load, and thus a higher susceptibility to hyperglycemia, than non-methane subjects (R. Mathur et al., unpublished data). A final potential mechanism whereby methanogens may affect energy extraction by the host is by slowing gut motility. Among human irritable bowel syndrome patients, we found that those with methane on breath test are more likely to have constipation as a predominant symptom subtype (19, 35), and that the amount of methane produced is related to the degree of constipation, as measured by Bristol Stool Score, and frequency of bowel movements (35). Methane is also associated with other constipation disorders (36, 37). In an in vivo animal study, we demonstrated that infusion of methane into the small intestine resulted in slowing of small intestinal transit by 59% (38). That slowing of intestinal transit may be associated with greater BMI is demonstrated by a study by the gastroparesis consortium, which showed that subjects with extreme slow motility (gastroparesis) had higher BMIs (39), and by a study of ultrashort bowel patients, in which we found that slowing the gut with exenatide resulted in resolution of diarrhea, nutritional deficiencies and the need for chronic parenteral nutrition, and was accompanied by weight gain (40). Taken together, these represent several potential mechanisms by which the increased M. smithii colonization could contribute to the concomitant weight gain we observed in these rats.

Despite all of this evidence supporting that (increased) colonization with methanogens, and specifically M. smithii, has the potential to result in weight gain by the host, an important missing evidentiary link has been demonstrable colonization of the small intestine. To date, methanogens have been identified primarily in the left colon (24-26), and it has been argued that alterations to a gut microbial population not known to occur outside of the large bowel is unlikely to be a significant direct cause of weight gain. Our results demonstrate for the first time that in the rat, not only does colonization with M. smithii occur in the small bowel, but that M. smithii levels in the duodenum, jejunum, and ileum are in fact higher than in the cecum or left colon, with highest levels in the ileum. Moreover, the degree of weight gain in these animals corresponded with the number of bowel segments colonized, suggesting that the extent of colonization of the intestine with M. smithii may be predictive of, and possibly contribute to, weight gain. Whether this may also be true in humans remains to be determined, although our previous data demonstrating that increased methane on breath test is associated with higher average BMI (23), suggest that increased levels of this methanogen may be among the factors contributing to levels of obesity in some human subjects.

In conclusion, our results demonstrate for the first time that colonization with the methanogen M. smithii is not confined to the large intestine, but also occurs in the small bowel. Moreover, in this rat model, we found that the levels and extent of small intestinal colonization with M. smithii correlated with, and were predictive of, the degree of weight gain, irrespective of dietary fat content. These findings further support the growing body of evidence suggesting that methanogens have an important influence on host metabolism and energy harvest, and may contribute to weight gain and obesity.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study was conducted with support from the Beatrice and Samuel A. Seaver Foundation and the Shoolman Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References