Gut Microbiota Contribution to Weight-Independent Glycemic Improvements after Gastric Bypass Surgery

The study investigates whether and how the gut microbiota contributes to the weight loss-independent improvements in glycemic control observed after Roux-en-Y gastric bypass (RYGB). While metabolic surgeries like RYGB are the most effective interventions for severe obesity and type 2 diabetes, their benefits extend beyond mechanical restriction and malabsorption. RYGB substantially remodels the gut microbiota. Prior evidence suggests microbiota can influence energy balance and glycemia, but the interdependence of body weight, microbiota composition, and glucose metabolism remains unclear. This work aims to disentangle weight loss from microbiota-mediated effects on glycemic control by comparing RYGB-treated rats with sham-operated and body weight-matched (diet-restricted) counterparts, correlating microbiota features with glycemic indices, and testing causality via fecal microbiota transplantation (FMT) to germfree mice and supplementation with a candidate bacterium.

Prior studies show obesity-associated microbiotas increase energy harvest and promote weight gain. RYGB and sleeve gastrectomy induce pronounced and lasting microbiota shifts, often decreasing Firmicutes and increasing Proteobacteria. Antibiotic depletion of microbiota can blunt the weight loss and glycemic benefits of metabolic surgery, suggesting a causal role. FMT from post-surgery donors to germfree mice or antibiotic-treated recipients has been reported to reduce weight gain or improve glycemia; however, many studies confound weight changes with microbiota effects. Human FMT studies post-bariatric surgery indicate altered intestinal morphology and reduced glucose absorption independent of obesity, and donor characteristics can drive changes in insulin sensitivity. Erysipelotrichaceae have been implicated in metabolic dysfunction. Despite these links, whether specific taxa mediate weight-independent glycemic improvements post-RYGB and whether correlational associations imply causation remain unresolved.

Design and animals:

• Zucker fatty (fa/fa) male rats (n=45, 6 weeks old) were randomized to: Sham (n=12), RYGB (n=20), or sham with caloric restriction to match RYGB body weight (BWM; n=13). Surgeries and perioperative care followed prior protocols. Food intake and body weight were monitored for 28 days; OGTT performed on postoperative day 27. Four RYGB rats were euthanized early due to excessive weight loss.
• Indices: OGTT area under the curve (AUC), Matsuda-DeFronzo insulin sensitivity index (ISI-M), and HOMA-IR. On day 28, fresh feces collected (Sham n=12, RYGB n=16, BWM n=5) and frozen.

Microbiota profiling:

• DNA extraction from 60–80 mg feces using repeated bead beating, lysis buffer, and purification (QIAamp kit).
• 16S rRNA gene V4 amplification (primers 515F/806R), duplicate PCRs, Illumina MiniSeq 250 bp paired-end sequencing (two batches).
• QIIME 2 pipeline: DADA2 for denoising and chimera filtering; taxonomy via RDP classifier against Greengenes v13.8. Total 3,756,593 reads; 1,170 OTUs included; subsampling to 37,000 reads per sample. Relative abundance analyses at genus (L6) and species (L7) levels; ANCOM-BC used for differential absolute abundance.
• Beta-diversity by Bray-Curtis; PERMANOVA with pairwise post hoc tests.

Identification of candidate species:

• Filtered sequences assigned to Erysipelotrichaceae; extracted sequence with low relative abundance in RYGB vs BWM; NCBI BLAST used to identify closest species. The unidentified Erysipelotrichaceae matched Longibaculum muris (97.2% identity). qPCR primers designed targeting L. muris 16S V4 region; qPCR quantified fecal L. muris copies.

FMT experiments in germfree (GF) mice:

• GF Swiss Webster mice (4-hour fasted) gavaged under aseptic conditions with fecal slurries from three RYGB and three BWM rat donors; housed in autoclaved cages with sterilized chow and water. OGTT (2 g/kg) performed 14 days post-FMT; body weights recorded days 1 and 14. Fecal microbiota composition assessed, focusing on Erysipelotrichaceae and the specific unidentified species.

L. muris culture and supplementation:

• L. muris DSM 29487 cultured anaerobically in YCFA at 37°C; heat-killed preparation via autoclaving.
• Experiment 1: GF mice first transplanted with feces from one RYGB rat donor, then randomized after 1 week to receive live or heat-killed L. muris by oral gavage for 3 weeks; OGTT performed.
• Experiment 2: Conventionally raised C57BL/6 mice on chow, gavaged twice weekly with live or heat-killed L. muris for 3 weeks; OGTT performed; subsequently switched to Western-style diet (WSD) with continued gavage for 3 weeks; second OGTT performed.

Statistics:

• GraphPad Prism: Mann-Whitney U test, unpaired t test, one-way ANOVA with Holm-Šidák, two-way ANOVA with Bonferroni. R (v4.1.2): PERMANOVA, Kruskal-Wallis with Benjamini-Hochberg FDR correction, Spearman correlations among OGTT AUC, HOMA-IR, ISI-M, body weight, and food intake. Significance at adjusted or raw P<0.05.

• Despite matched body weight between RYGB and BWM groups by day 28, RYGB-treated rats exhibited superior glycemic control vs both Sham and BWM: lower OGTT AUC, higher ISI-M, and lower HOMA-IR at postoperative day 27.
• Microbiota shifts after RYGB:
  • Beta-diversity: RYGB fecal microbiota clustered separately from Sham and BWM (Bray-Curtis; PERMANOVA Q<0.01 vs both groups).
  • Phylum level: RYGB increased Proteobacteria and decreased Firmicutes and Verrucomicrobia relative to Sham and BWM (e.g., Q<0.0001 to Q<0.01 vs Sham; P<0.001 to P<0.05 vs BWM).
  • Species level: Multiple differentially abundant taxa; Akkermansia muciniphila, Lactobacillus reuteri, Collinsella aerofaciens were lower in RYGB vs BWM.
  • An unidentified Erysipelotrichaceae species was significantly lower in RYGB vs both Sham and BWM. In RYGB rats, this species correlated positively with OGTT AUC and HOMA-IR and negatively with ISI-M (Q<0.05–0.01), suggesting association with worse glycemic control.
• FMT causality test:
  • GF mice receiving RYGB donor feces vs BWM donor feces had similar body weight gain over 14 days, but exhibited improved oral glucose tolerance (lower blood glucose at 15 min during OGTT; two-way ANOVA, P<0.01) independent of body weight.
  • The depletion pattern of the unidentified Erysipelotrichaceae species in RYGB donors was mirrored in recipients.
• Candidate species identification:
  • The unidentified Erysipelotrichaceae sequence aligned closest to Longibaculum muris (97.2% identity). qPCR confirmed lower fecal L. muris in RYGB vs BWM rats (P<0.001). In RYGB rats, L. muris abundance correlated positively with OGTT AUC (greater abundance associated with worse tolerance).
• L. muris supplementation outcomes:
  • Contrary to expectation, supplementing live L. muris to RYGB-recipient GF mice further reduced peak glucose at 15 min during OGTT compared with heat-killed L. muris, without affecting body weight.
  • In conventionally raised mice, administering live L. muris (vs heat-killed) had minimal effects on body weight and oral glucose tolerance under both chow and WSD conditions.
• Overall, results demonstrate a transmissible, weight-independent microbiota contribution to improved glycemic control after RYGB and illustrate that correlational links between a species (e.g., L. muris) and glycemic traits do not necessarily indicate causation.

The study shows that RYGB reshapes the gut microbiota and that these changes can partially transfer improved oral glucose tolerance to germfree mice without affecting body weight, indicating a weight-independent microbiota-mediated effect on glycemic control. The improvement was most evident at the OGTT peak, aligning with the hypothesis that post-RYGB microbiota may reduce intestinal glucose absorption rather than markedly enhancing systemic insulin sensitivity. Depletion of Erysipelotrichaceae members in RYGB donors and recipients is consistent with literature linking this family to metabolic dysfunction and suggests a role in modulating glucose handling. Identification of Longibaculum muris as the closest match to the RYGB-depleted Erysipelotrichaceae and its positive correlation with glucose intolerance in RYGB rats suggested a potential detrimental role; however, supplementation with live L. muris unexpectedly improved peak glucose in RYGB-recipient mice and had little effect in conventional mice. This highlights the complexity of host-microbiota interactions and cautions against interpreting species-trait correlations as causal. Differences from prior FMT studies that observed weight effects may relate to differences in dosing frequency, duration, donor material, and host context. Collectively, the findings support that multiple taxa and community-level functions, rather than a single species, likely contribute to post-RYGB glycemic benefits, potentially via reduced jejunal absorptive capacity and altered nutrient metabolism.

This work provides causal evidence that gut microbiota changes after RYGB contribute to weight loss-independent improvements in glycemic control. RYGB in Zucker fatty rats decreased Erysipelotrichaceae (including an L. muris-like species) and altered phylum-level composition; FMT from RYGB donors to germfree mice improved OGTT peak glucose without affecting weight. Although L. muris abundance correlated with worse glycemia in RYGB rats, supplementation paradoxically improved peak glucose in RYGB recipients and had minimal effects in conventional mice, underscoring the dissociation between correlation and causation and pointing to community-level effects. Future research should identify the functional bacterial consortia and mechanisms (e.g., intestinal glucose transport, bile acid-receptor signaling, protein metabolism) underlying improved glycemia, define context-dependent interactions among taxa, and evaluate microbiota-based adjunct therapies for type 2 diabetes.

• Group sizes were modest, particularly for BWM rats (n=5 for fecal sampling) and some mouse experiments (e.g., n=4 per L. muris arm in RYGB-recipient GF mice), potentially limiting power and generalizability.
• 16S rRNA sequencing with Greengenes taxonomy provides limited taxonomic resolution; identification of the Erysipelotrichaceae species as Longibaculum muris was based on 97.2% sequence identity and not confirmed by isolation from the rat samples.
• FMT and L. muris supplementation experiments used specific donor(s) and relatively short follow-up; effects on broader metabolic endpoints (insulin, GLP-1, intestinal transporters) were not measured.
• Improvements in OGTT were limited to peak glucose time points in recipients, and mechanisms (e.g., altered glucose absorption vs insulin action) were inferred rather than directly tested.
• Germfree mouse models may not fully reflect conventional or human physiology, and inter-study discrepancies in weight outcomes suggest protocol-dependent effects (dose, frequency, housing).
• Potential dietary and protein metabolism shifts post-RYGB that could influence L. muris abundance were not directly quantified.