- Email: [email protected]
Su1878 Microbial Signatures of Persistent Effects of Chronic Dietary Restriction in Mice
gain and typical features of metabolic syndrome and low-grade inflammation. However, this was not entirely a consequence of dietary fat content as such changes were also exhibited by mice fed a CDD of 10% fat. Altering the protein content, or comprising it exclusively of soy vs. casein protein did not affect the weight gain nor low-grade inflammation induced by this CDD. Nor was the low-grade inflammation and increased adiposity induced by the CDD altered by removing the insoluble (cellulose) fiber typically included in CDD. Under germ free conditions and when given soluble fiber (inulin) in a microbiota independent manner, the fiber did not affect gut mass. Moreover, the inflammatory and obesigenic effects of the CDD were entirely reversed by incorporating of 10% (by weight) soluble (inulin) fiber and slightly halted by incorporation of SCFA. Conclusions: Induction of obesity/ metabolic syndrome by HFD is not only a consequence of its lipid content but rather is strongly promoted by its lack of soluble fiber, which can be metabolized by gut bacteria to produce SCFA that protect against inflammation. These results demonstrate the importance of using properly controlled diets in animal studies. Furthermore, our results suggest that soluble dietary fiber may be important for prevention of obesity and metabolic syndrome.
Japanese Patients With Type 2 Diabetes Can Discriminate From Healthy Subjects by Gut Dysbiosis Makoto Sasaki, Akihiro Shimozato, Kazunori Adachi, Noriko Okaniwa, Shigeki Goji, Atsushi Tanabe, Hisatugu Noda, Kenichiro Yanamoto, Yoshihiro Kondo, Yasuhiro Tamura, Yoshitsugi Ito, Shinya Izawa, Masahide Ebi, Naohiko Kawamura, Akihito Iida, Naotaka Ogasawara, Yasushi Funaki, Kunio Kasugai INTRODUCTION: A lot of evidences indicate that microbiota is important modifier of obesity and diabetes. However, little is known about Japanese diabetic patients whose body mass index is normal compared with that of non-diabetic population. AIMS&METHODS: Using terminal restriction fragment length polymorphism analysis, we determined the composition of fecal gut microbiota in 59 Japanese patients with type 2 diabetes and 10 healthy controls, and its association with various clinical parameters. RESULTS: The ratio of the Clostridium subcluster XIVa and Lactbacillales were significantly lower and higher (p < 0.05), respectively, in fecal samples of diabetic patients than in those of control subjects. The diabetic patients were possible to discriminate from healthy subjects at 85.3% of hit ratio by the following discriminant; 0.023 × [Bacteroides (%)] + 0.184 × [Firmicutes (%)] + 2.238 × [Bacteroides/Firmicutes] - 12.375. The level of fasting blood glucose and triglyceride showed negative and positive correlation to Bacteroides and Clostridium cluster IV, respectively. The blood level of aspartate aminotransferase, alanine aminotransferase, γ-glutamyl transpeptidase, and total cholesterol showed positive correlation to Clostridium subcluster XIVa. CONCLUSION: Japanese patients with type 2 diabetes had gut dysbiosis as well as western patients. Gut microbiota of diabetes patients correlated to metabolism marker of peripheral blood, and discriminated type 2 diabetes from healthy subjects.
Su1880 Metabolomic Signatures in Human Serum and Roux Limb Tissue After Rouxen-Y Gastric Bypass; Comparison of Patients With Sustained Weight Loss Versus Weight Regain Wasif M. Abidi, Shuba Gopal, Nicholas Stylopoulos, Clary B. Clish, Christopher C. Thompson Introduction: Roux-en-Y gastric bypass surgery (RYGB) is complicated by weight regain in 20% of patients. The precise metabolic changes underlying sustained weight loss or weight regain are unclear. Prior studies have reported the serum metabolomic changes that occur after RYGB, but the metabolite profiling after weight regain has not been explored. In this study, we sought to determine whether patients after RYGB with sustained weight loss (RYGB-S) and weight regain (RYGB-WR) exhibit unique metabolomic signatures in serum and Roux limb tissue. Methods: This is an ongoing prospective study. To date, fasting serum has been obtained from 47 patients: 11 obese controls and 36 RYGB. Of the RYGB patients, 22 had weight regain (defined as >20% regain of the weight lost after surgery). Of the 47 patients 21 patients also had tissue biopsies from Roux limb (RYGB) or proximal jejunum (controls). Serum and tissue samples were analyzed using reversed-phase chromatography/ mass spectrometry (MS) and hydrophilic interaction liquid chromatography/MS. Metabolite sets were compared pairwise with enrichment analysis to evaluate for pathways that were significantly different between groups, as determined using permutation testing. Results: RYGB-S patients were significantly less obese than controls and RYGB-WR (BMI: 29±5 vs 36±7 vs 38±6). In serum samples, 398 metabolites were detected; with 85 metabolites being significantly different in the RYGB-S compared to control, and 49 different in the RYGBWR compared to controls. 88 metabolites were significantly different between RYGB-S and RYGB-WR, possibly serving as markers for weight regain. 19 metabolites were changed in both RYGB-WR and RYGB-S compared to controls. Citrulline, a quantitative marker of functional small bowel mass, was increased in both RYGB groups, consistent with Rouxlimb remodelling. Metabolomic profile from Roux limb samples was then performed to further characterize the intestinal metabolic changes. Of 401 detected metabolites, 116 were significantly different between RYGB-S and control, 82 different between RYGB-WR and control, but only 3 were different between RYGB-S and RYGB-WR. Analysis of the metabolic pathways in the tissue samples showed that purine, pyrimidine, arginine and proline metabolism pathways were enriched in both RYGB groups, consistent with tissue remodeling. Both RYGB groups exhibited elevated glucose utilization through glycolysis. Additionally TCA cycle was enriched only in RYGB-S. Of note, primary bile acid synthesis was depleted only in RYGB-WR. Conclusions: Sustained weight loss and weight regain after RYGB exhibit unique serum metabolomic signatures. The changes in the Roux limb are consistent with tissue remodelling. Most of these changes are not lost despite weight regain and both groups continue to exhibit increased intestinal glucose utilization.
Su1878 Microbial Signatures of Persistent Effects of Chronic Dietary Restriction in Mice Jun Chen, Yujiro Hayashi, Yoshitaka Toyomasu, David R. Linden, Heidi Nelson, Gianrico Farrugia, Purna C. Kashyap, Nicholas Chia, Tamas Ordog Background & Aims: Restoration of body weight and growth remains challenging in conditions involving negative nutritional balance. A recent study revealed microbiota immaturity in malnourished children that was only partially ameliorated by nutritional interventions (Subramanian/Gordon, Nature 2014). Previously we established a murine model of stunting by restricting food access in young adult mice and found that the body weight (bwt) of these animals remained below those of ad-libitum-fed controls even after refeeding (DDW 2014). Here, we investigated whether the lack of bwt restoration by refeeding could be due to persistent changes in the gut microbiota. Methods: 4-wk-old female BALB/c mice (n= 12/group) were fed ad-libitum (AL) or exposed to limited feeding to abolish bwt gain (LF). Food access was restored in a subset of LF mice on day 97 (refed; LF-RF). Freshly expelled fecal pellets were collected aseptically on days 69, 96, 97, 98, 99 and 120. The V3-5 region of the 16S rRNA gene was amplified using primers 357F and 926R and sequenced using the Illumina MiSeq platform. Paired-end reads were processed using IM-TORNADO. Microbiota-host factor associations were analyzed by distance-based PERMANOVA using unweighted, generalized and weighted UniFrac as distance metrics and quantified by the coefficient of determination R2 for age, diet and bwt. Microbial taxa detected on day 120 were compared by Wald test at FDR Q<0.1 following fitting with an overdispersed Poisson regression model. The machine learning algorithm Random Forests was used to predict age based on the microbiota profiles. Results: On day 120, bwt change from day 0 (Δbwt) in AL, LF and LF-RF mice was 52±3, -6±1* and 40±3%* (*, P<0.05 vs. AL). Age and diet but not Δbwt were associated with gut microbiota composition (P<0.05). Alpha diversity increased with age; and age explained a larger proportion of the microbiota variability (6.47.3%) than diet (≤3.8%) or Δbwt (≤1.4%). The Random Forests algorithm was able to predict chronological age based on the microbiota profile (R2=0.7-0.8) and indicated microbiota immaturity in the LF vs. AL mice on day 69 (P<0.05) but not after refeeding. However, on day 120, the community structure of LF-RF microbiota was significantly different from that of both AL (P<0.001) and LF mice (P=0.002). Relative to AL mice, the refed groups had higher abundance of Bacteroidetes and Proteobacteria and lower abundance of Firmicutes. Conclusions: Similarly to bwt, dietary restriction-induced changes in gut microbiota were not restored by refeeding. However, the lack of significant association between Δbwt and microbiota does not support causal relationship between these hysteresis effects. Mice exposed to limited feeding are a good model of moderate human undernutrition. Support: NIH DK68055; Rebecca E. Saunders Fund in Anorexia Nervosa and Bulimia Research.
Su1881 High Protein Diet Improves Meal Pattern and Response to Cholecystokinin (CCK) in High Fat Diet-Induced Obese Rats Lixin Wang, Joseph R. Reeve, Joseph R. Pisegna, Yvette Tache Background: High protein diet (HPD) is one of the interventions to reduce body weight of obese subjects. We showed that switching rats from high fat diet (HFD) to HPD reduced body weight (Stengel et al. AJP 2012). HFD-induced obese (DIO) rats have impaired response to intraperitoneal (IP) CCK (Duca et al., Horm Behav 2013). Aim: To get insight to the mechanisms of HPD effects in DIO rats by assessing changes in meal pattern and responsiveness to IP CCK. Methods: Male SD rats were fed ad libitum HFD (45% calories from fat with 5.5% from soybean oil and 39.4% from lard) for 10 weeks and control rats remained on normal diet (ND, 12% calories from fat and 29% from protein). DIO rats on HFD and normal rats on ND were placed in an automatic feeding monitoring system (BioDAQ). DIO rats were switched to HPD (52% calories from protein) 5 weeks later. Body composition was determined using rodent MRI (EChoMRI 700). CCK-8 sulfated (1.8 and 5.2 nmol/kg, IP) was injected before the dark phase in DIO rats on HFD after 3 weeks in BioDAQ, and then 6 weeks after being switched to HPD with no access to food 90 min before dark phase. Results: Rats on HFD for 10 weeks were divided to DIO rats with fat increase superior to 50% of that in normal rats on ND, and the others as non-responders. Compared to their initial body and fat weights, the DIO rats showed increases by 73% and 280%; and nonresponders by 93% and 185% respectively. Switching DIO rats in BioDAQ to HPD for 6 weeks induced 9% fat mass loss. ND in normal rats, and HFD and HPD in DIO rats resulted in similar daily calories intake (Table). HFD significantly increased the daily average calories at each meal, while reducing bouts, meal frequency and time spent on meals compared to ND. Similar meal pattern was observed during the dark and light phase, except an increase in meal rate (Table). Switching to HPD reversed HFD-induced increased calories and rate of each meal, and decreased bouts and time on meal, and also increased meal duration assessed for 24 h, dark phase and light phase, but did not alter the HFD-increased each meal calories in light phase (Table). CCK-8 (1.8 and 5.2 nmol/kg) did not significantly
Su1879 Microbiota Metabolism of Soluble Fiber Protects Against Low-Grade Inflammation and Metabolic Syndrome Jennifer P. Miles-Brown, Benoit Chassaing, Andrew T. Gewirtz Background: A common mouse model of metabolic syndrome is utilized wherein mice are fed a compositionally-defined diet high in fat and compared to mice fed standard mouse chow, which although relatively low in fat, has a complex, variable, and relatively undefined, composition. Metabolic syndrome is associated with low-grade inflammation, which can be driven by the gut microbiota, whose composition and activity might be affected by a variety of dietary components. Goal: Define the extent to which high-fat diet (HFD)-induced lowgrade inflammation and metabolic syndrome was a consequence primarily of fat content and define the role of other dietary factors that might distinguish HFD from a normal chow diet (NCD). Methods: Mice, raised on NCD, were, at 4 weeks of age, fed NCD, HFD, or a panel of other compositionally defined diets (CDD) that were high (60%) or low (10%) in fat and supplemented with 150nM short chain fatty acid (SCFA) mix or antibiotics, respectively. Mice were monitored for weight gain, euthanized from 1-12 weeks later, and assayed for a variety of parameters that reflect metabolic syndrome and low-grade inflammation including fat mass, colon length, cecum mass, and spleen size. Results: In accordance with other studies, relative to mice maintained on NCD, mice fed a HFD exhibited marked weight
S-541
AGA Abstracts
AGA Abstracts
Su1877
Japanese Patients With Type 2 Diabetes Can Discriminate From Healthy Subjects by Gut Dysbiosis Makoto Sasaki, Akihiro Shimozato, Kazunori Adachi, Noriko Okaniwa, Shigeki Goji, Atsushi Tanabe, Hisatugu Noda, Kenichiro Yanamoto, Yoshihiro Kondo, Yasuhiro Tamura, Yoshitsugi Ito, Shinya Izawa, Masahide Ebi, Naohiko Kawamura, Akihito Iida, Naotaka Ogasawara, Yasushi Funaki, Kunio Kasugai INTRODUCTION: A lot of evidences indicate that microbiota is important modifier of obesity and diabetes. However, little is known about Japanese diabetic patients whose body mass index is normal compared with that of non-diabetic population. AIMS&METHODS: Using terminal restriction fragment length polymorphism analysis, we determined the composition of fecal gut microbiota in 59 Japanese patients with type 2 diabetes and 10 healthy controls, and its association with various clinical parameters. RESULTS: The ratio of the Clostridium subcluster XIVa and Lactbacillales were significantly lower and higher (p < 0.05), respectively, in fecal samples of diabetic patients than in those of control subjects. The diabetic patients were possible to discriminate from healthy subjects at 85.3% of hit ratio by the following discriminant; 0.023 × [Bacteroides (%)] + 0.184 × [Firmicutes (%)] + 2.238 × [Bacteroides/Firmicutes] - 12.375. The level of fasting blood glucose and triglyceride showed negative and positive correlation to Bacteroides and Clostridium cluster IV, respectively. The blood level of aspartate aminotransferase, alanine aminotransferase, γ-glutamyl transpeptidase, and total cholesterol showed positive correlation to Clostridium subcluster XIVa. CONCLUSION: Japanese patients with type 2 diabetes had gut dysbiosis as well as western patients. Gut microbiota of diabetes patients correlated to metabolism marker of peripheral blood, and discriminated type 2 diabetes from healthy subjects.
Su1880 Metabolomic Signatures in Human Serum and Roux Limb Tissue After Rouxen-Y Gastric Bypass; Comparison of Patients With Sustained Weight Loss Versus Weight Regain Wasif M. Abidi, Shuba Gopal, Nicholas Stylopoulos, Clary B. Clish, Christopher C. Thompson Introduction: Roux-en-Y gastric bypass surgery (RYGB) is complicated by weight regain in 20% of patients. The precise metabolic changes underlying sustained weight loss or weight regain are unclear. Prior studies have reported the serum metabolomic changes that occur after RYGB, but the metabolite profiling after weight regain has not been explored. In this study, we sought to determine whether patients after RYGB with sustained weight loss (RYGB-S) and weight regain (RYGB-WR) exhibit unique metabolomic signatures in serum and Roux limb tissue. Methods: This is an ongoing prospective study. To date, fasting serum has been obtained from 47 patients: 11 obese controls and 36 RYGB. Of the RYGB patients, 22 had weight regain (defined as >20% regain of the weight lost after surgery). Of the 47 patients 21 patients also had tissue biopsies from Roux limb (RYGB) or proximal jejunum (controls). Serum and tissue samples were analyzed using reversed-phase chromatography/ mass spectrometry (MS) and hydrophilic interaction liquid chromatography/MS. Metabolite sets were compared pairwise with enrichment analysis to evaluate for pathways that were significantly different between groups, as determined using permutation testing. Results: RYGB-S patients were significantly less obese than controls and RYGB-WR (BMI: 29±5 vs 36±7 vs 38±6). In serum samples, 398 metabolites were detected; with 85 metabolites being significantly different in the RYGB-S compared to control, and 49 different in the RYGBWR compared to controls. 88 metabolites were significantly different between RYGB-S and RYGB-WR, possibly serving as markers for weight regain. 19 metabolites were changed in both RYGB-WR and RYGB-S compared to controls. Citrulline, a quantitative marker of functional small bowel mass, was increased in both RYGB groups, consistent with Rouxlimb remodelling. Metabolomic profile from Roux limb samples was then performed to further characterize the intestinal metabolic changes. Of 401 detected metabolites, 116 were significantly different between RYGB-S and control, 82 different between RYGB-WR and control, but only 3 were different between RYGB-S and RYGB-WR. Analysis of the metabolic pathways in the tissue samples showed that purine, pyrimidine, arginine and proline metabolism pathways were enriched in both RYGB groups, consistent with tissue remodeling. Both RYGB groups exhibited elevated glucose utilization through glycolysis. Additionally TCA cycle was enriched only in RYGB-S. Of note, primary bile acid synthesis was depleted only in RYGB-WR. Conclusions: Sustained weight loss and weight regain after RYGB exhibit unique serum metabolomic signatures. The changes in the Roux limb are consistent with tissue remodelling. Most of these changes are not lost despite weight regain and both groups continue to exhibit increased intestinal glucose utilization.
Su1878 Microbial Signatures of Persistent Effects of Chronic Dietary Restriction in Mice Jun Chen, Yujiro Hayashi, Yoshitaka Toyomasu, David R. Linden, Heidi Nelson, Gianrico Farrugia, Purna C. Kashyap, Nicholas Chia, Tamas Ordog Background & Aims: Restoration of body weight and growth remains challenging in conditions involving negative nutritional balance. A recent study revealed microbiota immaturity in malnourished children that was only partially ameliorated by nutritional interventions (Subramanian/Gordon, Nature 2014). Previously we established a murine model of stunting by restricting food access in young adult mice and found that the body weight (bwt) of these animals remained below those of ad-libitum-fed controls even after refeeding (DDW 2014). Here, we investigated whether the lack of bwt restoration by refeeding could be due to persistent changes in the gut microbiota. Methods: 4-wk-old female BALB/c mice (n= 12/group) were fed ad-libitum (AL) or exposed to limited feeding to abolish bwt gain (LF). Food access was restored in a subset of LF mice on day 97 (refed; LF-RF). Freshly expelled fecal pellets were collected aseptically on days 69, 96, 97, 98, 99 and 120. The V3-5 region of the 16S rRNA gene was amplified using primers 357F and 926R and sequenced using the Illumina MiSeq platform. Paired-end reads were processed using IM-TORNADO. Microbiota-host factor associations were analyzed by distance-based PERMANOVA using unweighted, generalized and weighted UniFrac as distance metrics and quantified by the coefficient of determination R2 for age, diet and bwt. Microbial taxa detected on day 120 were compared by Wald test at FDR Q<0.1 following fitting with an overdispersed Poisson regression model. The machine learning algorithm Random Forests was used to predict age based on the microbiota profiles. Results: On day 120, bwt change from day 0 (Δbwt) in AL, LF and LF-RF mice was 52±3, -6±1* and 40±3%* (*, P<0.05 vs. AL). Age and diet but not Δbwt were associated with gut microbiota composition (P<0.05). Alpha diversity increased with age; and age explained a larger proportion of the microbiota variability (6.47.3%) than diet (≤3.8%) or Δbwt (≤1.4%). The Random Forests algorithm was able to predict chronological age based on the microbiota profile (R2=0.7-0.8) and indicated microbiota immaturity in the LF vs. AL mice on day 69 (P<0.05) but not after refeeding. However, on day 120, the community structure of LF-RF microbiota was significantly different from that of both AL (P<0.001) and LF mice (P=0.002). Relative to AL mice, the refed groups had higher abundance of Bacteroidetes and Proteobacteria and lower abundance of Firmicutes. Conclusions: Similarly to bwt, dietary restriction-induced changes in gut microbiota were not restored by refeeding. However, the lack of significant association between Δbwt and microbiota does not support causal relationship between these hysteresis effects. Mice exposed to limited feeding are a good model of moderate human undernutrition. Support: NIH DK68055; Rebecca E. Saunders Fund in Anorexia Nervosa and Bulimia Research.
Su1881 High Protein Diet Improves Meal Pattern and Response to Cholecystokinin (CCK) in High Fat Diet-Induced Obese Rats Lixin Wang, Joseph R. Reeve, Joseph R. Pisegna, Yvette Tache Background: High protein diet (HPD) is one of the interventions to reduce body weight of obese subjects. We showed that switching rats from high fat diet (HFD) to HPD reduced body weight (Stengel et al. AJP 2012). HFD-induced obese (DIO) rats have impaired response to intraperitoneal (IP) CCK (Duca et al., Horm Behav 2013). Aim: To get insight to the mechanisms of HPD effects in DIO rats by assessing changes in meal pattern and responsiveness to IP CCK. Methods: Male SD rats were fed ad libitum HFD (45% calories from fat with 5.5% from soybean oil and 39.4% from lard) for 10 weeks and control rats remained on normal diet (ND, 12% calories from fat and 29% from protein). DIO rats on HFD and normal rats on ND were placed in an automatic feeding monitoring system (BioDAQ). DIO rats were switched to HPD (52% calories from protein) 5 weeks later. Body composition was determined using rodent MRI (EChoMRI 700). CCK-8 sulfated (1.8 and 5.2 nmol/kg, IP) was injected before the dark phase in DIO rats on HFD after 3 weeks in BioDAQ, and then 6 weeks after being switched to HPD with no access to food 90 min before dark phase. Results: Rats on HFD for 10 weeks were divided to DIO rats with fat increase superior to 50% of that in normal rats on ND, and the others as non-responders. Compared to their initial body and fat weights, the DIO rats showed increases by 73% and 280%; and nonresponders by 93% and 185% respectively. Switching DIO rats in BioDAQ to HPD for 6 weeks induced 9% fat mass loss. ND in normal rats, and HFD and HPD in DIO rats resulted in similar daily calories intake (Table). HFD significantly increased the daily average calories at each meal, while reducing bouts, meal frequency and time spent on meals compared to ND. Similar meal pattern was observed during the dark and light phase, except an increase in meal rate (Table). Switching to HPD reversed HFD-induced increased calories and rate of each meal, and decreased bouts and time on meal, and also increased meal duration assessed for 24 h, dark phase and light phase, but did not alter the HFD-increased each meal calories in light phase (Table). CCK-8 (1.8 and 5.2 nmol/kg) did not significantly
Su1879 Microbiota Metabolism of Soluble Fiber Protects Against Low-Grade Inflammation and Metabolic Syndrome Jennifer P. Miles-Brown, Benoit Chassaing, Andrew T. Gewirtz Background: A common mouse model of metabolic syndrome is utilized wherein mice are fed a compositionally-defined diet high in fat and compared to mice fed standard mouse chow, which although relatively low in fat, has a complex, variable, and relatively undefined, composition. Metabolic syndrome is associated with low-grade inflammation, which can be driven by the gut microbiota, whose composition and activity might be affected by a variety of dietary components. Goal: Define the extent to which high-fat diet (HFD)-induced lowgrade inflammation and metabolic syndrome was a consequence primarily of fat content and define the role of other dietary factors that might distinguish HFD from a normal chow diet (NCD). Methods: Mice, raised on NCD, were, at 4 weeks of age, fed NCD, HFD, or a panel of other compositionally defined diets (CDD) that were high (60%) or low (10%) in fat and supplemented with 150nM short chain fatty acid (SCFA) mix or antibiotics, respectively. Mice were monitored for weight gain, euthanized from 1-12 weeks later, and assayed for a variety of parameters that reflect metabolic syndrome and low-grade inflammation including fat mass, colon length, cecum mass, and spleen size. Results: In accordance with other studies, relative to mice maintained on NCD, mice fed a HFD exhibited marked weight
S-541
AGA Abstracts
AGA Abstracts
Su1877