Official journal of the American College of Gastroenterology


INTRODUCTION

In 2001, the global obesity epidemic earned the moniker “globesity” from the World Health Organization (WHO). By 2016, the WHO reported that more than 1.25 billion adults were overweight (body mass index [BMI] of 25–29.9 kg/m2), and more than 650 million were obese (BMI ≥30 kg/m2), accounting for 25% and 13% of the world’s adult population, respectively (1). If current trends continue, global obesity prevalence will reach 18% in men and 21% in women by 2025 (1–3).

Obesity is associated with increased risks of a wide variety of diseases, including cardiovascular diseases, metabolic syndrome and type 2 diabetes (T2D), and certain cancers (4,5). Worldwide, obesity contributes to an estimated 2.8 million deaths annually (1). In the United States, being obese or overweight is one of the leading causes of preventable death, close behind tobacco use (6,7). Importantly, risk of death increases significantly with increasing weight: A population-based cohort study of 3.6 million adults in the United Kingdom found that, below a BMI of 25 kg/m2, the hazard ratio for all-cause mortality increased by 0.81 (95% confidence interval [CI] 0.80–0.82) per 5-kg/m2 increase in BMI; above a BMI of 25 kg/m2, this ratio increased to 1.21 (1.20–1.22) (8).

The gut microbiome includes the bacteria, archaea, fungi, and viruses that inhabit the human gastrointestinal (GI) tract (9). Imbalances in gut microbial populations have repeatedly been reported in patients with obesity-related diseases (10–13). After initial demonstrations that an obese phenotype was transmissible from genetically obese mice to germ-free mice through cecal transplant (14), early investigations into particular microbes associated with weight gain and obesity generally found an increased Firmicutes-to-Bacteroidetes ratio in both mice (14) and humans (15). However, a number of subsequent studies failed to support this finding (16–18). More recent research has demonstrated that individuals with lower bacterial richness and gene content exhibit a greater overall inflammatory phenotype, as well as adiposity, insulin resistance, and dyslipidemia, suggesting that which specific bacterial species are altered may be less important than their effects on the functional redundancy of the microbiome, particularly for key genes involved in host metabolism (19).

Importantly, most gut microbiome studies conducted to date have used stool samples as a measure of gut microbial content, largely because of the cost and difficulty of sampling the small intestine (20). However, as microbial load, pH, and transit time vary dramatically along the length of the GI tract, stool is an insufficient surrogate for the entire gut microbiome. The small intestine functions as a critical site for nutrient absorption, endocrine regulation, and immune function (21), and the Revealing the Entire Intestinal Microbiota and its Associations with the Genetic, Immunologic, and Neuroendocrine Ecosystem (REIMAGINE) study was created to examine the specific importance of the small intestinal microbiome in human health and disease. To do this, we developed and validated luminal small bowel collection techniques, which are optimized to reduce cross-contamination with saliva and increase DNA recovery from small intestinal microbes (22). Using these techniques, we have shown that small bowel microbial populations are greater than previously thought (23) and differ significantly from those in stool (24).

Although previous studies have explored alterations in the fecal microbiome in obesity, little is known about differences in the small bowel microbiome between obese and nonobese subjects, and even less is known about the shifts that occur in gut microbial ecology in the progression from normal weight to obesity. Understanding this relationship, particularly given the critical role of the small bowel as a regulator of host nutrient absorption and metabolism, is central to uncovering the role of the gut microbiome in obesity and to developing targeted interventions. In this study, we applied the validated techniques described above to characterize how the small bowel microbiome changes with increasing BMI.

METHODS

Study subjects

Subjects were recruited for the REIMAGINE study at Cedars-Sinai, which includes male and female subjects aged 18–80 years undergoing standard-of-care upper endoscopy (esophagogastroduodenoscopy [EGD]) without colon preparation. Details of the REIMAGINE study have been described previously (24–26). Subjects were grouped according to BMI per Centers for Disease Control and Prevention guidelines:

  1. Group 1: normal weight—BMI 18.5–24.9;
  2. Group 2: overweight—BMI 25–29.9;
  3. Group 3: obesity—BMI >30.


To avoid biases that could potentially affect the small bowel microbiome, including small intestinal bacterial overgrowth (SIBO) (27) and antibiotic use (23), a postscreening step was performed to exclude subjects with known diabetes, SIBO, history of gastric bypass surgery, use of weight control medications, or antibiotic use within the 3 months before the procedure.

The REIMAGINE study protocol was approved by the Institutional Review Board at Cedars-Sinai, and all subjects provided informed written consent.

Questionnaires

Before EGD, subjects completed a demographic and family/medical history questionnaire, including GI diseases and conditions, GI symptoms, medication use, use of alcohol and recreational drugs, travel history, and dietary habits. All subject-provided medical information was verified through medical records audits. Data were deidentified before downstream analysis.

Blood collection and analysis

Circulating levels of fasting glucose, total cholesterol, low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides, and total protein were measured on a Cobas Integra 400 (Roche Diagnostics, Rotkreuz, Switzerland), and levels of fasting insulin, C-peptide, gastric inhibitory polypeptide, total glucagon-like peptide-1, leptin, and glucagon were analyzed on a Luminex FlexMap 3D (Luminex, Austin, TX) using a bead-based multiplex assay (EMD Millipore, Billerica, MA). Circulating cytokine and chemokine levels were analyzed on a Luminex FlexMap 3D (Luminex) using a multiplex panel that included granulocyte-macrophage colony-stimulating factor, interferon-gamma, interleukin (IL)10, IL12P70, IL13, IL1β, IL2, IL4, IL5, IL6, and IL8, and tumor necrosis factor-alpha (TNFα) (EMD Millipore).

Fasting blood glucose measurements >100 mg/dL were considered to be prediabetes (28). Total cholesterol levels >200 mg/dL, LDL levels ≥130 mg/dL, and triglycerides levels ≥150 mg/dL were considered elevated (29).

Small intestinal luminal sample collection

During the EGD, duodenal luminal samples were procured using a custom-designed sterile double-lumen aspiration catheter (Hobbs Medical, Stafford Springs, CT), which was pushed through a sterile bone wax cap only after the endoscopist entered the second portion of the duodenum, to reduce contamination from the mouth, esophagus, and stomach (22).

Aspirate processing and microbial culture

Because of the relatively low biomass in the duodenum and high viscosity of duodenal aspirates, an equal volume of sterile 1× dithiothreitol was added to each aspirate (∼1 mL), and samples were vortexed until fully liquified (∼30 seconds). One hundred microliters of each sample were then serially diluted with 900 μL sterile 1× phosphate-buffered saline and plated on MacConkey agar (Becton Dickinson, Franklin Lakes, NJ) and blood agar (Becton Dickinson). Plates were incubated at 37 °C for 16–18 hours under aerobic (MacConkey) or anaerobic (blood agar) conditions. If no colonies were detected, samples were incubated for a further 24 hours. Colony-forming units (CFU) were counted electronically using a Scan 500 (Interscience, Paris, France). Subjects with ≥1,000 CFU/mL in small bowel aspirates were considered positive for SIBO.

The remainder of each duodenal aspirate (the portion not used for microbial culture) was used for sequencing. Samples were centrifuged at maximum speed (>13,000 RPM) for 5 minutes, the supernatant was removed, and sterile Allprotect reagent (Qiagen, Hilden, Germany) was added to the microbial pellet. Pellets under Allprotect were stored at −80 °C before DNA isolation.

DNA isolation and 16S sequencing

DNAs were isolated from duodenal aspirate pellets using the MagAttract PowerSoil DNA KF Kit (Qiagen), as described previously (22), using aliquots of 1× dithiothreitol as negative controls. The concentrations of DNAs obtained from negative controls were undetectable. Libraries for the 16S rRNA gene V3/V4 regions were prepared and amplified as described previously (22), using gene-specific primers (S-D-Bact-0341-b-S-17 and S-D-Bact-0785-a-A-21) (30). Paired-end sequencing (2 × 301 cycles) was performed on a MiSeq system (Illumina, San Diego, CA) using v3 (600 cycles) kits (Illumina), with 10%–15% PhiX (Illumina). Operational taxonomic unit (OTU) clustering and taxonomic analyses were performed using CLC Genomics Workbench v.22.0.2 and CLC Microbial Genomics Module v.20.1.1/22.1 (Qiagen). Sequences were first trimmed using the Trim Reads Tool (Qiagen) and following parameters: quality limit of 0.5, maximum number of ambiguities of 2, and a trim adapter list containing the sequence of Illumina adapters. Trimmed reads were merged using the Merge Overlapping Pairs Tool (mismatch cost of 2, gap cost of 3, and minimal score of 30). Clustering into OTUs was performed using the Amplicon-Based OTU Clustering Tool and 97% sequence similarity. The creation of new OTUs was not allowed. Taxonomy was assigned using CLC Microbial Genomics default values by comparing against the SILVA Database (2019 release). Low-depth samples (<5,000 sequences/sample) were removed, and alpha diversity indexes were calculated. Bray-Curtis and UniFrac metrics were used to calculate intersample (beta) diversity. Abundant OTUs in negative controls were removed from the data set, as previously described (22).

Library preparation and sequencing for metagenomics analysis

Libraries for whole-genome (shotgun) sequencing were prepared using the Illumina DNA Prep kit (Illumina) and IDT for Illumina—DNA/RNA UD Indexes (Illumina) following the manufacturer’s instructions. Library qualities were checked using the High Sensitivity D5000 ScreenTape kit (Agilent) on a 2200 TapeStation (Agilent). Only libraries above 0.48 nM were sequenced. Paired sequencing was performed on a NovaSeq platform (Illumina) using the NovaSeq 6000 S2 Reagent Kit v1.5 (300 cycles). Demultiplexed reads were analyzed using CLC Genomics Workbench 22.0.2/20.0.3 and Microbial Module. Reads were trimmed using the Trim Reads Tool (Qiagen) and following parameters: quality limit of 0.5, maximum number of ambiguities of 2, and a trim adapter list containing the sequence of Illumina adapters. Taxonomic identification was performed using the Taxonomic Profiling Tool (Qiagen) after removal of host reads using the Homo sapiens hg38 genomic taxpro index available with CLC Workbench. A minimum seed length of 30 was used. The Autodetect Paired Distances and Adjust Read Count Abundances options were checked. A curated Microbial Reference Database with selected references from GenBank optimized to 16-GB memory including all annotation tracks (CDS, genes, etc.) for 4,275 species was used as the reference. Paired reads that could not be mapped as an intact pair to the reference database were dismissed. Host-filtered reads that mapped to the reference database were first assigned to the lowest common ancestor of all mapping positions with the highest mapping score. Next, qualification and quantification of the abundance of each qualified taxon was performed, to determine whether a particular microbial taxon was represented in a sample. The qualification step was based on a confidence score that a reference did not receive its reads by pure chance. Any microbial taxon with a confidence score <0.995 was ignored, and the reads assigned to it were reassigned to its closest qualified ancestor (by construction, the confidence score was very close to 1.0). To confirm the identification of bacterial species and strains, a second taxonomic analysis using assembled reads to longer sequences (contigs) was also performed. Host-filtered reads from the previous step were assembled to longer sequences using the De Novo Assembly tool (Qiagen). Reads were mapped back to contigs using a mismatch cost of 2, insertion cost and deletion cost of 3, and length fraction and similarity fraction of 0.8. Contigs were extracted, and another host cleaning step was performed by filtering contigs against the Homo sapiens hg38 sequence. Host-filtered contigs were then mapped to the same microbial database described above.

Data and statistical analysis

OTUs with significantly different abundances between groups were identified following published recommendations (31,32) used when the average library size for each group is approximately equal, and/or the fold difference between groups is lower than 2- to 3-fold on average. OTU tables were not rarefied for downstream analysis if no differences in library size were identified between groups.

Multiple comparisons and statistical analyses for both 16S and shotgun data were performed with CLC Genomics Workbench v.20.0.3/22.0.2 and CLC Microbial Genomics Module v.20.1.1/22.1 (Qiagen). Log-transformed fold change (FC) differences were calculated using a generalized linear model, which also corrects for differences in library size between samples and effects of confounding factors. An unbiased analysis was performed to compare differences in relative abundances (RAs) between groups at each taxonomic level. A negative binomial generalized linear model was used to obtain maximum likelihood estimates for FCs between groups, using the Wald test for determination of significance. P-values were corrected using a false discovery rate (FDR). An FDR-corrected P-value <0.05 was considered significant. A similar pipeline was applied for the metagenomics shotgun data downstream analysis.

Analysis of predicted microbial pathways and functions was performed for the 16S data set, using the Functional Analysis tool at CLC Workbench 22.0.2. Functional profiles were inferred using EC database and Clusters of Orthologous Genes terms, and pathways were identified using the MetaCyc Pathway Database (2022-05). Randomization analyses were performed with 1,000 replicates, and super-pathways were included for data analysis. An FDR-corrected P-value <0.1 was considered significant.

Two-tailed Spearman correlations, statistical tests, and graph construction were performed with normalized OTU tables using GraphPad Prism 7.02 (GraphPad Software, La Jolla, CA) and IBM SPSS Statistics Version 24.

Study data sets are available at the National Center for Biotechnology Information BioProject Repository (https://www.ncbi.nlm.nih.gov/bioproject) under BioProject ID PRJNA1010206.

RESULTS

Subject demographics

A total of 459 subjects whose duodenal microbiomes had been successfully sequenced were prescreened for inclusion in this study. To avoid known and potential confounders that could affect duodenal microbial composition, subjects with known diabetes, SIBO, history of gastric bypass surgery, weight control medications, or antibiotic use within the 3 months before the procedure were excluded.

After these exclusions, 214 subjects remained in the main cohort and were used for analysis of duodenal microbial populations and serum biomarkers (Table 1). One hundred five were of normal weight (mean age: 52.5 ± 17 years, females = 62 [59.6%]), 67 were overweight (mean age: 56.3 ± 16 years, females = 21 [31.3%]), and 42 had obesity (mean age: 51.2 ± 16 years, females = 22 [52.4%], Table 1). Indications for EGD are summarized in Table 2.

T1
Table 1.:

Subject demographics and fasting biomarker levels

T2
Table 2.:

Indications for esophagogastroduodenoscopy

Serum: inflammatory markers

Circulating anti-inflammatory and pro-inflammatory cytokines were quantified in fasting serum samples. Although no differences were found between absolute levels in subjects with overweight and obesity compared with subjects of normal weight (P > 0.05), circulating TNFα levels correlated positively with BMI (Spearman R = 0.171, P = 0.029).

Serum: glucose and insulin

Although mean fasting glucose levels were similar between subjects of normal weight (97.5 ± 7.2 mg/dL) and subjects with overweight (98.7 ± 11.9 mg/dL, P = 0.55) and obesity (107.21 ± 27.4 mg/dL, P = 0.14), the incidence of elevated fasting glucose was, as expected, slightly higher in subjects with obesity (48.3%) vs subjects of normal weight (33.8%, χ2P = 0.17, fasting glucose range ≥100 mg/dL, Table 1). Although no glucose challenge was performed, mean fasting C-peptide levels were progressively higher in subjects with overweight (1.33 ± 0.68 ng/mL, P < 0.0001) and obesity (1.84 ± 0.93 ng/mL, P < 0.0001) vs subjects of normal weight (0.88 ± 0.42 ng/mL) (Table 1). In fact, almost half of the subjects with obesity (45%) had fasting C-peptide levels above 2 ng/mL, which may reflect relative insulin resistance in these subjects.

Serum: lipids

Levels of fasting total cholesterol, cholesterol fractions, and triglycerides were also compared between groups. No differences in mean triglycerides or total cholesterol levels were found comparing subjects with overweight (triglycerides = 119.29 ± 77.82 mg/dL, total cholesterol = 187.83 ± 47.85 mg/dL) with subjects of normal weight (triglycerides = 95.44 ± 47.88 mg/dL, P = 0.217; total cholesterol = 175.96 ± 41.51 mg/dL, P = 0.224, Table 1). When cholesterol fractions were analyzed separately, subjects with overweight were found to have higher LDL (P = 0.039) and lower HDL (P = 0.056) when compared with subjects of normal weight (Table 1).

Dyslipidemia patterns according to Fredrickson phenotypes were also determined. Of the subjects with normal weight who exhibited dyslipidemia (N = 12), 75% exhibited type IIa (elevated LDL only), and 25% exhibited type IIb (elevated LDL and very LDL [VLDL]). No other dyslipidemias were identified in this group. Dyslipidemias in subjects with overweight included type IIa (41%, odds ratio [OR] 2.4, 95% CI 0.88–6.75, P = 0.08), type IIb (41%, OR 7.3, 95% CI 2.05–26.17, P = 0.0017), and type IV (elevated VLDL only, 18%, P = 0.003).

Although no differences in mean total cholesterol levels were identified in subjects with obesity (181.59 ± 55.61 mg/dL) vs subjects of normal weight (175.96 ± 41.51 mg/dL, P = 0.75, Table 1), when fractions were analyzed separately, HDL levels were much lower in subjects with obesity (P = 0.02, Table 1). In addition, mean triglycerides levels were higher in subjects with obesity (176.63 ± 184.57 mg/dL) vs subjects of normal weight (95.44 ± 47.88 mg/dL, P = 0.014, Table 1), and the dyslipidemia pattern in subjects with obesity shifted toward type IV (31%, P = 0.0002), and fewer subjects had type IIa (31%, OR 1.8, 95% CI 0.55–6.07, P = 0.36) or type IIb (38%, OR 6.77, 95% CI 1.46–27.02, P = 0.0069) when compared with subjects with normal weight and overweight.

Small bowel: microbiome changes related to overweight and obesity—phylum level

Small bowel (duodenal) 16S microbial profiles were analyzed in 214 subjects without SIBO. No differences in library size or distribution were found between groups (mean number of reads classified under an OTU—normal weight: 216,540; overweight: 202,431; and obesity: 178,190, P = 0.18). Therefore, a Wald test and adjusted P-value (significance <0.05) were used to determine both shared and condition-specific duodenal microbial features in each group. Duodenal microbial diversity was higher in subjects with overweight vs normal weight (P = 0.04), but no differences were observed in subjects with obesity vs overweight (P = 0.3, see Supplementary Figure 1, Supplementary Digital Content 1, https://links.lww.com/AJG/D239).

Both the taxonomy and the functional potential of the duodenal microbiome differed significantly in individuals with overweight and obesity vs normal weight. The RA of Firmicutes, the most dominant phylum in the duodenum (24), was not different among groups (see Supplementary Figure 2, Supplementary Digital Content 1, https://links.lww.com/AJG/D239). Despite being a small player in the whole duodenal microbiome architecture, a gradual loss of photosynthetic organisms from phylum Cyanobacteria was associated with overweight and obesity (FC overweight vs normal weight = −2.38, Adj-P-value = 3.04E-4; FC obesity vs normal weight = −2.35, Adj-P-value = 5.31E-3). It is often assumed that sequences assigned to this phylum represent genomic material derived from ingestion of chloroplasts or Cyanobacterial cells (9,33); however, the differences in this study were identified after filtering out ASVs associated with chloroplast sequences from plants. Therefore, this difference is not diet-related. The RA of another small player in the duodenal microbiome, phylum Synergistetes, also seemed to be associated with changes in BMI (Spearman = −0.127, P = 0.063). Interestingly, the RA of this phylum was lower only in subjects with overweight compared with normal weight (FC = −1.15, P-value, 9.89E-3, Adj-P-value = 0.05), with less significant differences between subjects with obesity compared with normal weight (FC = −1.2, P-value = 0.02, Adj-P-value = 0.12).

Small bowel: microbiome changes related to overweight and obesity—family level

At the family level, using an FC cutoff of 1.5 and FDR P-value cutoff of 0.05, the RA of 33 and 28 families was statistically different in subjects with overweight and obesity when compared with normal weight (see Supplementary Figure 3A, Supplementary Digital Content 1, https://links.lww.com/AJG/D239). Within the top 15 most widespread bacterial families in the duodenum, the RA of 2 families—Family XI and Fusobacteriaceae—was slightly lower in subjects with overweight (FC = −0.73, Adj-P-value = 0.03, FC = −0.78, Adj-P-value = 0.03), but not in subjects with obesity, when compared with normal weight (see Supplementary Figure 3B, Supplementary Digital Content 1, https://links.lww.com/AJG/D239). In subjects with obesity, 2 of the 15 most widespread families were different, with higher RA of Lactobacillaceae when compared with normal weight (FC = 3.18, Adj-P-value = 7.96E-5) and lower RA of Leptotrichiaceae (FC = −0.95, Adj-P-value = 0.03, see Supplementary Figure 3B, Supplementary Digital Content 1, https://links.lww.com/AJG/D239). Neither of these differences were seen in subjects with overweight.

Small bowel: microbiome changes related to overweight and obesity—genus and species levels

At the genus and species levels, the RA of 137 features was significantly different in subjects with overweight vs normal weight (Figure 1). A total of 113 features were significantly different in subjects with obesity vs normal weight, and 65 features were significantly different in subjects with overweight vs obesity (Figure 1). Considering that most of these microbial changes, with a few exceptions, were in less prevalent families in the small bowel (perhaps a reflection of the heterogeneity of the groups), and considering that obesity is a progressive and multifactorial disease, another approach was applied for analysis at the genus and species level (for the 16S data set) as an attempt to narrow down relevant features in the context of disease progression. A multigroup comparison was performed to identify both condition-specific and shared microbial features in subjects with overweight and obesity vs subjects with normal weight (34). Condition-specific features were either obesity-specific or overweight-specific (Figure 1). Shared microbial features were either escalation features (going in the same direction from normal weight to overweight and from overweight to obesity) or de-escalation (stabilization) features (going in opposite directions) (Figure 1). Whole-genome (shotgun) sequencing was applied to a subset of subjects from each group for species confirmation only (normal weight N = 26, overweight N = 12, and obesity N = 14).

F1
Figure 1.:

Small bowel microbial differences at the genus level among subjects with normal weight, overweight, and obesity. FDR, false discovery rate.

Small bowel: specific microbial features in subjects with overweight

A total of 103 microbial features were identified as overweight-specific, most of which are rare genera in the small bowel (∼96%, Figure 2). Changes in the RA of 6 of the most widespread genera (see Supplementary Figure 4, Supplementary Digital Content 1, https://links.lww.com/AJG/D239) were specific to subjects with overweight, including lower RA of Paraprevotella (FC = −9.20, Adj-P-value = 5.67E-32), Escherichia-Shigella (FC = −6.89, Adj-P-value = 4.44E-27), an unknown species from genus Escherichia further identified as Escherichia coli (FC = −5.20, Adj-P-value = 9.43E-17), and Pseudomonas (FC = −1.45, Adj-P-value = 0.05), as well as higher RA of Prevotella (FC = 1.05, Adj-P-value = 0.03) and an unknown species from genus Lactobacillus (FC = 2.96, Adj-P-value = 1.63E-5, Figure 2). Interestingly, the RA of Lactobacillus spp. seemed to be higher in subjects with overweight and type IIa dyslipidemia (N = 9) when compared with subjects with overweight but without dyslipidemia (N = 9, FC = 3.77, P = 0.04, Figure 2), and higher RA of Lactobacillus was associated with lower systemic HDL (Spearman R = −0.273, P = 0.002), despite the fact that species from this genus have previously been associated with better lipid profile outcomes in studies using stool samples (35).

F2
Figure 2.:

Overweight-specific features and microbial metabolic pathways.

The whole-genome (shotgun) sequencing performed in a subset of subjects did not identify species from the genus Paraprevotella but did reveal a lower RA of a strain of E. coli sharing genomic similarities with the strain K-12 in subjects with overweight vs those with normal weight (FC = −15, Adj-P-value = 2.2E-14). A total of 12 potential species and several strains of Pseudomonas were identified by shotgun, and except for 3 species including Pseudomonas aeruginosa, the RA of most of these species was confirmed to be lower in subjects with overweight vs normal weight (Adj-P-value <0.05). The RA of 7 of these species was also lower in subjects with overweight vs obesity but was not different between normal weight and obesity. Therefore, these species were reclassified as potential de-escalation features (see section “Small bowel: microbial de-escalation features”). By contrast, 2 of these Pseudomonas species—P. parafulva and P. psychrotolerans—were confirmed to be specific to subjects with overweight.

Most of the identified Prevotella species (95%) did not differ between subjects with overweight and obesity when compared with normal weight, except for Prevotella loescheii, the RA of which was confirmed to be higher in subjects with overweight (FC = 5.01, Adj-P-value = 0.01). A total of 31 Lactobacillus species were identified by shotgun sequencing, and the RA of 4 of these was higher in the duodenum of subjects with overweight: L. acidophilus (FC = 20.63, Adj-P-value = 2.51E-33), L. hominis (FC = 14.71, Adj-P-value = 5.07E-23), L. intestinalis (FC = 14.13, Adj-P-value = 7.47E-16), and L. johnsonii (FC = 5.65, Adj-P-value = 3.91E-3). The RA of L. intestinalis and L. johnsonii was also higher in subjects with obesity vs normal weight (FC = 14.71, Adj-P-value = 1.91E-18; FC = 6.19, Adj-P-value = 7.69E-4, respectively) but was not significantly different between overweight and obesity, confirming them to be specific to subjects with overweight. By contrast, from the shotgun sequencing, the RA of L. acidophilus and L. hominis was lower subjects with obesity vs overweight but was not different between subjects with obesity vs normal weight. Therefore, L. acidophilus and L. hominis were reclassified as potential de-escalation features (see section “Small bowel: microbial de-escalation features”).

Although not part of the top 25 most widespread genera in the duodenum, the RA of genus Bifidobacterium, an extensively studied genus related to a healthier microbiome and potentially associated with weight loss, was higher in the duodenum of subjects with normal weight vs those with overweight (FC = 4.91, Adj-P-value = 1.79E-18) and obesity (FC = 4.03, FDR P-value = 1.33E-9). A total of 15 Bifidobacterium species were identified by shotgun sequencing. The RA of 11 of these species was higher in the duodenum of subjects with normal weight vs those with overweight and obesity but was not different between subjects with overweight and obesity, confirming them as overweight-specific.

A total of 65 microbial pathways were significantly different between subjects with overweight and normal weight, so to isolate the most important pathways related to duodenal microbial changes, a second analysis approach was applied where only highly prevalent and abundant pathways (frequency per sample of 100%) were analyzed. The duodenal microbial metabolic potential in subjects with overweight was characterized by reductions in pathways involved in polymyxin resistance (FC = −2.06, P-value = 7.36E-5, Adj-P-value = 0.08), sulfoquinovose degradation I (FC = −1.79, P-value = 3.05E-4, Adj-P-value = 0.09), ursodeoxycholate biosynthesis (bacteria) (FC = −1.95, P-value = 3.68E-4, Adj-P-value = 0.09), bile acid epimerization (FC = −1.95, P-value = 3.64E-4, Adj-P-value = 0.09), and several pathways involved in biogenic amines metabolism (P-value = 1.43E-3, Adj-P-value = 0.1, Figure 2).

Small bowel: specific microbial features in subjects with obesity

A total of 39 duodenal microbial features were identified as specific to obesity (Figure 3). The RA of most of these taxa (60%) was higher in the duodenum of subjects with obesity vs those with normal weight. Within the 25 most widespread genera in the duodenum (see Supplementary Figure 4, Supplementary Digital Content 1, https://links.lww.com/AJG/D239), changes in the RA of 3 features were specific to obesity, including higher RA of an unknown Lactobacillus species (Adj-P-value <0.05), further identified by shotgun as L. gasseri (FC = 4.5, Adj-P-value = 8.54E-3, Figure 3). By contrast, the RA of 1 specific strain of L. reuteri (subspecies rodentium) was lower in subjects with obesity (FC = −6.75, Adj-P-value = 1.67E-16), confirmed by shotgun sequencing (FC = −7.03, Adj-P-value = 1.05E-4, Figure 3). The RA of genus Leptotrichia was slightly lower in subjects with obesity (FC = −0.95, FDR P-value = 0.03), and shotgun sequencing confirmed a slightly lower RA of Leptotrichia trevisanii (FC = −2.61, Adj-P-value = 0.1), but also revealed that the RA of an unidentified species from this genus was 12.67-fold lower in the duodenum of subjects with obesity vs those of normal weight (Adj-P-value = 2.6E-12). The RA of genus Alloprevotella, further identified by shotgun sequencing as A. rava, was also lower in subjects with obesity (FC = −2.93, Adj-P-value = 0.02, Figure 3). Interestingly, lower RA of Alloprevotella seemed to be associated with subjects with dyslipidemia related to VLDL and triglycerides (IIb and IV), independent of weight. Lower RA of this genus was identified in subjects with normal weight and type IIb dyslipidemia vs subjects with normal weight and normal ranges of LDL, VLDL, and triglycerides (FC = −5.35, P-value = 8.85E-3), in subjects with overweight with type IIb dyslipidemia vs subjects with overweight and normal ranges of LDL, VLDL, and triglycerides (FC = −5.26, P-value = 6.62E-3, Figure 1), and in subjects with obesity and type IV dislipidemia (FC = −6.38, P-value = 1.32E-3, Figure 3). Finally, the duodenal microbial metabolic potential of subjects with obesity was characterized by reductions in pathways involved in biogenic amines metabolism (FC = −4.62, P-value = 1E-13, Adj-P-value = 1.27E-10) and sulfoacetaldehyde degradation (FC = −5.02, P-value = 2.54E-13, Adj-P-value = 1.61E-10, Figure 3).

F3
Figure 3.:

Obesity-specific features and microbial metabolic pathways. FDR, false discovery rate.

Small bowel: microbial escalation features

A total of 5 features were classified as escalation features (Figure 4). We defined an escalation feature as one for which the direction of change (increase or decrease) persisted throughout the spectrum of normal weight to overweight to obesity. The RA of 3 features gradually decreased with increasing weight, including unknown species from the genera Faecalibacterium (obesity vs normal weight FC = −11.97, Adj-P-value = 4.75E-32; obesity vs overweight FC = −2.81, Adj-P-value = 0.05; overweight vs normal weight FC = −9.16, Adj-P-value = 5.98E-40), Bacteroides (obesity vs normal weight FC = −9.50, Adj-P-value = 6.48E-29; obesity vs overweight FC = −2.56, Adj-P-value = 0.03; overweight vs normal weight FC = −6.94, Adj-P-value = 6.72E-24), and Staphylococcus (obesity vs normal weight FC = −6.29, Adj-P-value = 2.67E-12; obesity vs overweight FC = −2.92, Adj-P-value = 0.01; overweight vs normal weight FC = −3.37, Adj-P-value = 2.65E-9, Figure 4). The RA of 2 features gradually increased with increasing weight, including an unknown species from genus Mycobacterium (obesity vs normal weight FC = 3.34, Adj-P-value = 2.39E-8; obesity vs overweight FC = 1.82, Adj-P-value = 6.32E-3; overweight vs normal weight FC = 1.52, Adj-P-value = 0.02, Figure 4), and an unknown Lactobacillus species (obesity vs normal weight FC = 4.69, Adj-P-value = 6.17E-14; obesity vs overweight FC = 2, Adj-P-value = 8.73E-3; overweight vs normal weight FC = 2.69, Adj-P-value = 3.21E-6, Figure 4).

F4
Figure 4.:

Escalation features (going in the same direction from normal weight to overweight and from overweight to obesity) in the small bowel of subjects with overweight and obesity. FDR, false discovery rate.

Analysis by whole-genome sequencing identified B. pyogenes and S. hominis as 2 of the species that gradually decreased with increasing weight, but no information was available for species identification within the genus Faecalibacterium, nor was the genus itself detected by shotgun sequencing. The unknown Mycobacterium species could not be identified by shotgun, despite being part of the data set as an unidentified species, the RA of which was also higher in subjects with obesity (FC = 14.47, Adj-P-value = 1.16E-16) and overweight (FC = 15.36, Adj-P-value = 3.09E-17) when compared with normal weight. In addition, as previously noted, the RA of several Lactobacillus species was higher in subjects with obesity and overweight vs normal weight, but none of these could be classified as escalation features using the shotgun data set derived from a smaller subset of subjects.

Small bowel: microbial de-escalation features

A total of 29 features were classified as de-escalation features (Figure 5). We defined a de-escalation feature as a microbial change for which the direction of change (increase or decrease) seen going from normal weight to overweight is reversed going from overweight to obesity. Three of these, unknown species from Lactobacillus, Prevotella, and Pseudomonas, were from the top 25 most widespread genera in the duodenum.

F5
Figure 5.:

De-escalation (stabilization) features (going in the opposite direction) in the small bowel of subjects with overweight and obesity. FDR, false discovery rate.

Shotgun analysis identified 3 Lactobacillus species as de-escalation features (Figure 5). The RA of L. acidophilus and L. hominis was higher in subjects with overweight vs those with normal weight (FC = 20.63, Adj-P-value = 2.51E-33; FC = 14.71, Adj-P-value = 5.07E-23, respectively) and obesity (FC = 20.34, Adj-P-value = 3.38E-23; FC = 14.44, Adj-P-value = 1.48E-15, respectively) but was not significantly different between subjects with obesity and normal weight. The RA of the third species, L. iners, was lower in subjects with overweight vs normal weight (FC = −15.06, Adj-P-value = 1E-11) and obesity (FC = −14.26, Adj-P-value = 4.64E-9) but was not significantly different between subjects with obesity and normal weight.

As previously stated, by shotgun sequencing, most of the identified species of Prevotella (95%) were not different between groups, except for Prevotella loescheii, which had higher RA in subjects with overweight. The RA of 2 species (which remained unidentified by shotgun) was higher in subjects with overweight vs normal weight and obesity but was not significantly different between subjects with obesity and normal weight (P-value <0.05, Adj-P-value <0.05).

The RA of 7 Pseudomonas species, including P. putida and P. fluorescens, was lower in subjects with overweight vs those with normal weight and obesity but was not significantly different between subjects with obesity and normal weight.

One Bifidobacterium species, B. dentium, was identified as a de-escalation feature with a lower RA in the duodenum of subjects with overweight vs normal weight (FC = −19.23, Adj-P-value = 3.42E-11) and obesity (FC = −17.55, Adj-P-value = 5.04E-7) but was not significantly different between subjects with obesity and normal weight (FC = 1.68, Adj-P-value = 0.98).

DISCUSSION

In this paper, we present the largest study to date linking specific microbes in the metabolically active small intestine to overweight and obesity. Obesity is a progressive disease, and we also show that specific small bowel microbes can be further classified as escalation or de-escalation features, which may be related to aggravation or attenuation/stabilization of the progression from overweight to obesity. Interestingly, although most findings are unique to the small bowel, some genera previously associated with obesity in studies using stool samples, such as Bifidobacterium and Lactobacillus, were also identified as key players in the small bowel microbiome. For example, Bifidobacterium dentium is classified as a de-escalation feature and is enriched in small bowel microbiome of subjects with normal weight. Findings for the genus Lactobacillus are highly species-dependent, with some species classified as de-escalation features (e.g., L. acidophilus), and others enriched in subjects with obesity (e.g., L. gasseri).

Obesity is a complex multifactorial disease and often associated with conditions such as metabolic syndrome, hyperlipidemia, and T2D (4,5). The pathophysiology of obesity is now understood to involve the interplay of intrinsic (genetics, epigenetics, gut microbiome, gut hormones, etc), environmental, socioeconomic, and psychosocial factors that influence food intake and energy expenditure (36). Gut microbes represent another factor influencing metabolism and food absorption, and the development of microbiome-based therapeutics has escalated rapidly. However, most microbiome-based therapeutics for obesity were developed based on studies using stool as a surrogate for the whole gut microbiome, but we and others have shown that small bowel microbial populations differ significantly from those in stool (24).

Why focus on the small bowel microbiome? Many functions related to food metabolism and nutrient absorption converge in the small bowel, as well as endocrine and immune functions that may also influence the development of obesity (37), and as a result, disturbing the balances of small bowel microbial populations may have a greater impact on weight gain. Furthermore, as noted above, small bowel microbial populations differ from those in stool—e.g., phylum Bacteroidetes is much less prevalent in the small bowel, and there is no evidence that the Firmicutes:Bacteroidetes ratio often cited as a marker of obesity has any relevance in the small bowel. In fact, we did not find changes in the RAs of any of the most prevalent phyla in the duodenum, including Firmicutes, in subjects with overweight or obesity. Within phylum Firmicutes, genus Lactobacillus has been strongly linked to obesity in stool studies. However, Lactobacillus populations in the duodenum exhibit differing and highly species- and strain-specific associations with overweight and obesity—L. intestinalis and L. johnsonii are specific to subjects with overweight, whereas L. reuteri and L. gasseri seem specific to subjects with obesity. Previous interventional studies using different strains of L. reuteri and L. gasseri yielded conflicting outcomes related to weight management, possibly because of important differences in the carbohydrate and lipid metabolic capabilities of specific strains (38–41). These strain-specific associations with weight gain/loss are of particular importance when we consider the extensive consumption of various Lactobacillus-containing preparations worldwide.

Furthermore, the weight gain–associated Lactobacillus strains identified here harbor enzymes involved in lipid metabolism, such as thiolases, which may ultimately lead to more efficient fat digestion and absorption in the small bowel (39). Fat metabolism and absorption are facilitated by bile acids, and perturbations of the diversity and composition of the bile acid pool—primarily modulated by gut microbes (42)—are also closely related to dysregulation of lipid metabolism and overweight/obesity (43,44). In this study, we identified several links between specific duodenal microbes and different dyslipidemias, e.g., Alloprevotella rava is associated with type IIb and IV dyslipidemias, independent of weight, whereas an unidentified Lactobacillus species is associated with type IIa dyslipidemia specifically in subjects with overweight, and microbial pathways related to the epimerization of bile acids in the small bowel also seem to be impaired in these subjects.

Three Lactobacillus species are further classified as de-escalation features in our study, including L. acidophilus. This species is associated with more consistent outcomes in human interventional studies, and monotherapy with a single strain of L. acidophilus was associated with weight gain in an animal study (45). De-escalation features may be markers of stabilization and thus represent potential targets for weight management therapeutics. A species of Bifidobacterium sharing genomic similarities with B. dentium is also identified as a de-escalation feature, and the RAs of 11 other Bifidobacterium species are higher in the duodenum of subjects with normal weight when compared with those with overweight and obesity. These findings are consistent with several studies demonstrating that Bifidobacterium species such as B. longum, B. bifidum, B. breve, and B. infantis have antiobesity effects and are inversely associated with visceral adipose tissue, BMI, waist circumstance, blood triglycerides, and fatty liver (46,47).

Chronic low-grade systemic and local inflammation are key components in the pathogenesis of multiple obesity-related metabolic diseases, including increased insulin resistance and dyslipidemia (48). Bacterial strains modulate these effects in various ways, including reducing proinflammatory cytokines (e.g., L. rhamnosus and Parabacteroides distasonis) and producing and metabolizing biogenic amines (e.g., B. animalis). Biogenic amines play central roles in regulating systemic and local gut inflammation, preventing oxidative stress by scavenging reactive oxygen species, inhibiting proinflammatory cytokines release and the formation of inflammasomes, and regulating adipogenesis (49). In this study, microbial potential for biogenic amines metabolism seems to be impaired in the duodenum of subjects with overweight and obesity, suggesting a potential for increased inflammation. Consistent with this, we find an association between higher BMI and higher systemic TNFα levels.

Although this is the largest study of the duodenal microbiome in overweight and obesity to date, some smaller studies have also explored the small bowel microbiome in obesity and related metabolic conditions (reviewed in Steinbach et al (50)). Unfortunately, as noted by the review authors, differences in sampling techniques and locations, study populations, and the conditions under study make it almost impossible to compare the results. Study cohorts ranged from 10 to 66 participants, and compared hyperglycemic vs normoglycemic subjects (51), subjects with low or high Homeostatic Model Assessment for Insulin Resistance and subjects with T2D taking metformin (52), as well as subjects with obesity (53–56), or obesity and T2D (53,54). All but one of these studies used jejunal samples or duodenal biopsies rather than duodenal luminal samples, in several cases obtained during bariatric or gastric bypass surgeries after subjects had been after a very low-calorie diet that could potentially alter the microbiome (50). The only study that used duodenal luminal samples collected them under test meal conditions, which the authors acknowledged were not designed for studying gut microbiota (56). Unsurprisingly, there were very few consistencies between our findings and findings from these previous studies, with the exception of findings of decreased Bifidobacterium species in subjects of normal weight (54), and increased TNFα levels in subjects with hyperglycemia (51). However, the various study authors did recognize the importance of studying the small bowel, as well as the limitations of previous aspirate sampling techniques, in particular the significant risks of contamination with oral and esophageal microbes. Our study is the first performed using a sterile double-lumen catheter as well as a mucolytic to release microbes in the small bowel mucus, and as such addresses a previously unmet need in the field.

This study has some limitations. First, our subject population was heterogeneous, with different indications for upper endoscopy. Diet was not standardized across the cohort, but this may not be a significant limitation for the duodenum (57). In addition, shotgun sequencing was only performed in a subset of subjects, although these were carefully chosen to be as representative as possible of the larger cohort. Some species could not be confirmed by the shotgun sequencing, and it is possible that deeper sequencing could have revealed these. Finally, samples were obtained at a single timepoint only, so distinct cause-and-effect relationships cannot be established.

In conclusion, we identify small bowel microbial species associated with overweight and obesity, as well as escalation and de-escalation features that could potentially be selected as therapeutic targets. Lactobacillus species seem to play distinct and important roles, including 3 species identified as de-escalation features, L. acidophilus, L. hominis, and L. iners. B. dentium was also identified as a de-escalation feature, which is interesting given the known roles of Bifidobacterium species in short-chain fatty acid production and their potential anti-obesity effects. These findings illustrate that, although stool studies can and have provided very valuable data, direct analysis of the small bowel has yielded specific targets for further study. Next steps will include testing the identified microbes features in animal models, in order to develop potential therapeutic strategies.

CONFLICTS OF INTEREST

Guarantor of the article: Ruchi Mathur, MD.

Specific author contributions: R.M. and M.P.: conceptualization. A.H. and M.R.: resources. G.L., G.B., G.P., S.W., and W.M.: investigation. G.L., A.R., and M.P.: formal analysis. G.B., M.P., and R.M.: project administration. G.L., G.B., D.C., and R.M.: writing—original draft. G.L., G.B., A.R., M.P., and R.M.: writing—review and editing.

Financial support: This study was supported in part by funds from The Monica Lester Charitable Trust, and The Elias, Genevieve, and Georgianna Charitable Trust.

Potential competing interests: None to report.

Study Highlights

WHAT IS KNOWN

  • ✓ Stool studies indicate that gut microbial populations are altered in overweight and obesity.
  • ✓ Stool studies have linked alterations in Akkermansia, Bifidobacteria, and Lactobacillus species to obesity.
  • ✓ The composition of the small bowel microbiome is significantly different from that of stool.
  • ✓ The roles of small bowel microbes in overweight and obesity are poorly understood.


WHAT IS NEW HERE

  • ✓ The small bowel (duodenal) microbiome is significantly altered in subjects with overweight and obesity vs normal weight.
  • ✓ Specific microbial alterations are overweight-specific or obesity-specific; others are escalation or de-escalation features.
  • Bifidobacterium dentium is a de-escalation feature, consistent with known anti-obesity effects.
  • ✓ Changes in Lactobacillus gasseri and decreased L. reuteri are obesity-specific, but L. acidophilus and L. hominis are de-escalation features.
  • ✓ Specific Lactobacillus species are linked to type IIa dyslipidemia, and Alloprevotella rava is linked to type IIb and IV dyslipidemias.

ACKNOWLEDGEMENTS

The authors thank the REIMAGINE Study Group for their assistance in obtaining samples. The REIMAGINE Study Group includes Christopher Almario MD, FACG, Benjamin Basseri MD, Yin Chan MD, Bianca Chang MD, Derek Cheng MD, Pedram Enayati MD, Srinivas Gaddam MD, Laith Jamil MD, FACG, Quin Liu MD, Simon Lo MD, Marc Makhani MD, Deena Midani MD, Mazen Noureddin MD, FACG, Kenneth Park MD, Shirley Paski MD, Nipaporn Pichetshote MD, Shervin Rabizadeh MD, Soraya Ross MD, Omid Shaye MD, Rabindra Watson MD, Ali Rezaie MD, and Mark Pimentel MD, FACG. The authors also thank Maria Jesus Villanueva-Milan, PhD, and Maritza Sanchez for assisting with sample processing and analysis. Finally, we thank Frank Lee, the Monica Lester Charitable Trust, and the Elias, Genevieve, and Georgianna Charitable Trust for their generous support of the MAST program.

REFERENCES

1. Fact Sheet—Obesity and Overweight. World Health Organization: Geneva, Switzerland, 2020.

2. Centers for Disease Control and Prevention. Adult Obesity Facts. 2020. (https://www.cdc.gov/obesity/data/adult.html). Accessed July 12, 2023.

3. Centers for Disease Control and Prevention. Overweight & Obesity. 2022. (https://www.cdc.gov/obesity/index.html). Accessed July 12, 2023.

4. Centers for Disease Control and Prevention. Adult Obesity Causes & Consequences. 2021. (https://www.cdc.gov/obesity/adult/causes.html). Accessed July 12, 2023.

5. National Cancer Institute. Obesity and Cancer. 2017. (https://www.cancer.gov/about-cancer/causes-prevention/risk/obesity/obesity-fact-sheet). Accessed July 12, 2023.

6. GBD 2015 Obesity Collaborators, Afshin A, Forouzanfar MH, et al. Health effects of overweight and obesity in 195 countries over 25 years. N Engl J Med 2017;377(1):13–27.

7. Heymsfield SB, Wadden TA. Mechanisms, pathophysiology, and management of obesity. N Engl J Med 2017;376(3):254–66.

8. Bhaskaran K, Dos-Santos-Silva I, Leon DA, et al. Association of BMI with overall and cause-specific mortality: A population-based cohort study of 3·6 million adults in the UK. Lancet Diabetes Endocrinol 2018;6(12):944–53.

9. Human Microbiome Project Consortium. Structure, function and diversity of the healthy human microbiome. Nature 2012;486(7402):207–14.

10. Belizario JE, Faintuch J, Garay-Malpartida M. Gut microbiome dysbiosis and immunometabolism: New frontiers for treatment of metabolic diseases. Mediators Inflamm 2018;2018:2037838.

11. Sharma S, Tripathi P. Gut microbiome and type 2 diabetes: Where we are and where to go? J Nutr Biochem 2019;63:101–8.

12. Hur KY, Lee MS. Gut microbiota and metabolic disorders. Diabetes Metab J 2015;39(3):198–203.

13. Larsen N, Vogensen FK, van den Berg FW, et al. Gut microbiota in human adults with type 2 diabetes differs from non-diabetic adults. PLoS One 2010;5:e9085.

14. Turnbaugh PJ, Ley RE, Mahowald MA, et al. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature 2006;444(7122):1027–31.

15. Ley RE, Turnbaugh PJ, Klein S, et al. Microbial ecology: Human gut microbes associated with obesity. Nature 2006;444(7122):1022–3.

16. Schwiertz A, Taras D, Schafer K, et al. Microbiota and SCFA in lean and overweight healthy subjects. Obesity (Silver Spring) 2010;18(1):190–5.

17. Zhang C, Zhang M, Wang S, et al. Interactions between gut microbiota, host genetics and diet relevant to development of metabolic syndromes in mice. ISME J 2010;4(2):232–41.

18. Duncan SH, Lobley GE, Holtrop G, et al. Human colonic microbiota associated with diet, obesity and weight loss. Int J Obes (Lond) 2008;32(11):1720–4.

19. Le Chatelier E, Nielsen T, Qin J, et al. Richness of human gut microbiome correlates with metabolic markers. Nature 2013;500(7464):541–6.

20. El Aidy S, van den Bogert B, Kleerebezem M. The small intestine microbiota, nutritional modulation and relevance for health. Curr Opin Biotechnol 2015;32:14–20.

21. Leser TD, Mølbak L. Better living through microbial action: The benefits of the mammalian gastrointestinal microbiota on the host. Environ Microbiol 2009;11(9):2194–206.

22. Leite GGS, Morales W, Weitsman S, et al. Optimizing microbiome sequencing for small intestinal aspirates: Validation of novel techniques through the REIMAGINE study. BMC Microbiol 2019;19(1):239.

23. Barlow JT, Leite G, Romano AE, et al. Quantitative sequencing clarifies the role of disruptor taxa, oral microbiota, and strict anaerobes in the human small-intestine microbiome. Microbiome 2021;9(1):214.

24. Leite GGS, Weitsman S, Parodi G, et al. Mapping the segmental microbiomes in the human small bowel in comparison with stool: A REIMAGINE study. Dig Dis Sci 2020;65(9):2595–604.

25. Leite G, Pimentel M, Barlow GM, et al. Age and the aging process significantly alter the small bowel microbiome. Cell Rep 2021;36(13):109765.

26. Leite G, Barlow GM, Parodi G, et al. Duodenal microbiome changes in postmenopausal women: Effects of hormone therapy and implications for cardiovascular risk. Menopause 2022;29(3):264–75.

27. Leite G, Rezaie A, Mathur R, et al. Defining small intestinal bacterial overgrowth by culture and high throughput sequencing. Clin Gastroenterol Hepatol 2024;22(2):259–70.

28. American Diabetes Association. 2. Classification and diagnosis of diabetes: Standards of medical care in diabetes—2021. Diabetes Care 2021;44(Suppl 1):S15–33.

29. LaRosa JC. At what levels of total low- or high-density lipoprotein cholesterol should diet/drug therapy be initiated? United States guidelines. Am J Cardiol 1990;65(12):7f–10f.

30. Klindworth A, Pruesse E, Schweer T, et al. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res 2013;41(1):e1.

31. McMurdie PJ, Holmes S. Waste not, want not: Why rarefying microbiome data is inadmissible. PLoS Comput Biol 2014;10(4):e1003531.

32. Weiss S, Xu ZZ, Peddada S, et al. Normalization and microbial differential abundance strategies depend upon data characteristics. Microbiome 2017;5(1):27.

33. Turnbaugh PJ, Ridaura VK, Faith JJ, et al. The effect of diet on the human gut microbiome: A metagenomic analysis in humanized gnotobiotic mice. Sci Transl Med 2009;1(6):6ra14.

34. Fromentin S, Forslund SK, Chechi K, et al. Microbiome and metabolome features of the cardiometabolic disease spectrum. Nat Med 2022;28(2):303–14.

35. Wu Y, Zhang Q, Ren Y, et al. Effect of probiotic Lactobacillus on lipid profile: A systematic review and meta-analysis of randomized, controlled trials. PLoS One 2017;12(6):e0178868.

36. Barlow GM, Lin EA, Mathur R. An overview of the roles of the gut microbiome in obesity and diabetes. In: Bagchi D, Nair S (eds.). Nutritional and Therapeutic Interventions for Diabetes and Metabolic Syndrome. 2nd edn. Academic Press, 2018, p 65–91.

37. Kastl AJ, Terry NA, Wu GD, et al. The structure and function of the human small intestinal microbiota: Current understanding and future directions. Cell Mol Gastroenterol Hepatol. 2020;9(1):33–45.

38. Rouxinol-Dias AL, Pinto AR, Janeiro C, et al. Probiotics for the control of obesity: Its effect on weight change. Porto Biomed J 2016;1:12–24.

39. Drissi F, Merhej V, Angelakis E, et al. Comparative genomics analysis of Lactobacillus species associated with weight gain or weight protection. Nutr Diabetes 2014;4(2):e109.

40. Jung SP, Lee KM, Kang JH, et al. Effect of Lactobacillus gasseri BNR17 on overweight and obese adults: A randomized, double-blind clinical trial. Korean J Fam Med 2013;34(2):80–9.

41. Crovesy L, Ostrowski M, Ferreira DMTP, et al. Effect of Lactobacillus on body weight and body fat in overweight subjects: A systematic review of randomized controlled clinical trials. Int J Obes 2017;41(11):1607–14.

42. Guzior DV, Quinn RA. Review: Microbial transformations of human bile acids. Microbiome 2021;9(1):140.

43. Wei M, Huang F, Zhao L, et al. A dysregulated bile acid-gut microbiota axis contributes to obesity susceptibility. EBioMedicine 2020;55:102766.

44. Larabi AB, Masson HLP, Bäumler AJ. Bile acids as modulators of gut microbiota composition and function. Gut Microbes 2023;15(1):2172671.

45. Million M, Angelakis E, Paul M, et al. Comparative meta-analysis of the effect of Lactobacillus species on weight gain in humans and animals. Microb Pathog 2012;53(2):100–8.

46. Kim G, Yoon Y, Park JH, et al. Bifidobacterial carbohydrate/nucleoside metabolism enhances oxidative phosphorylation in white adipose tissue to protect against diet-induced obesity. Microbiome 2022;10(1):188.

47. Ma L, Zheng A, Ni L, et al. Bifidobacterium animalis subsp. lactis lkm512 attenuates obesity-associated inflammation and insulin resistance through the modification of gut microbiota in high-fat diet-induced obese mice. Mol Nutr Food Res 2022;66(3):e2100639.

48. Hildebrandt X, Ibrahim M, Peltzer N. Cell death and inflammation during obesity: “Know my methods, WAT(son)”. Cell Death Different 2023;30(2):279–92.

49. Ramos-Molina B, Queipo-Ortuno MI, Lambertos A, et al. Dietary and gut microbiota polyamines in obesity- and age-related diseases. Front Nutr 2019;6:24.

50. Steinbach E, Masi D, Ribeiro A, et al. Upper small intestine microbiome in obesity and related metabolic disorders: A new field of investigation. Metabolism 2024;150:155712.

51. Darra A, Singh V, Jena A, et al. Hyperglycemia is associated with duodenal dysbiosis and altered duodenal microenvironment. Sci Rep 2023;13(1):11038.

52. Gutierrez-Repiso C, Moreno-Indias I, Martin-Nunez GM, et al. Mucosa-associated microbiota in the jejunum of patients with morbid obesity: Alterations in states of insulin resistance and metformin treatment. Surg Obes Relat Dis 2020;16(10):1575–85.

53. Villmones HC, Svanevik M, Ulvestad E, et al. Investigating the human jejunal microbiota. Sci Rep 2022;12(1):1682.

54. Sroka-Oleksiak A, Mlodzinska A, Bulanda M, et al. Metagenomic analysis of duodenal microbiota reveals a potential biomarker of dysbiosis in the course of obesity and type 2 diabetes: A pilot study. J Clin Med 2020;9:369.

55. Granata I, Nardelli C, D’Argenio V, et al. Duodenal metatranscriptomics to define human and microbial functional alterations associated with severe obesity: A pilot study. Microorganisms 2020;8(11):1811.

56. Angelakis E, Armougom F, Carriere F, et al. A metagenomic investigation of the duodenal microbiota reveals links with obesity. PLoS One 2015;10(9):e0137784.

57. Shanahan ER, Kang S, Staudacher H, et al. Alterations to the duodenal microbiota are linked to gastric emptying and symptoms in functional dyspepsia. Gut 2023;72(5):929–38.

0 0 votes
Article Rating
Subscribe
Notify of
guest
0 Comments
Inline Feedbacks
View all comments
0
Your Cart is empty!

It looks like you haven't added any items to your cart yet.

Browse Products
Powered by Caddy
Som2ny Network
Logo