Yogurt supplementation attenuates insulin resistance in obese mice by reducing metabolic endotoxemia and inflammation
Data files
Jan 11, 2023 version files 1.19 MB
-
BodyWeight.xlsx
17.27 KB
-
Cytokine.xlsx
22.59 KB
-
DietIntake.xlsx
9.69 KB
-
GlucoseInsulin.xlsx
11.59 KB
-
MetabolicEndotoxemia_gene.xslx.xlsx
24.92 KB
-
MetabolicEndotoxemia_protein.xlsx
12.69 KB
-
Microbiota_abundance.txt
648.67 KB
-
Microbiota_Alpha.xlsx
13.07 KB
-
Microbiota_metadata.xlsx
13.43 KB
-
Microbiota_per.sample.fastq.count.csv
1.67 KB
-
Microbiota_taxonomy.txt
386.22 KB
-
README.txt
14.36 KB
-
TissueWeight.xlsx
14.85 KB
Abstract
Background
Inflammation is an underlying mechanism for development of obesity-related health complications. Yogurt consumption inhibits obesity-associated inflammation, but the tissue-specific mechanisms have not been adequately described.
Objectives
We aimed to determine the tissue-specific responses by which yogurt supplementation inhibits inflammation.
Methods
C57BL/6 male mice (5 weeks old) were fed a Teklad Global 14% Protein Rodent Maintenance diet as a control or a high-fat diet (60% calories from fat) to induce obesity for 11 weeks, followed by feeding a Western diet (WD; 43% carbohydrate & 42% fat) or WD supplemented with 5.6 % lyophilized yogurt powder for 3 weeks to test for the impact of yogurt supplementation. Markers of metabolic endotoxemia and inflammation were assessed in plasma and tissues. Cecal and fecal microbiota were profiled by 16S rRNA sequencing.
Results
In obese mice, relative to the WD control group, yogurt supplementation attenuated HOMA-IR by 57%) (p=0.020), plasma TNF-a by 31% (p<0.05) and colonic IFN-g by 46% (p=0.0034), which were accompanied by a 40% reduction in plasma LBP (p=0.0019) and 45% less colonic Lbp expression (p=0.037), as well as alteration in the beta diversity of cecal microbiota (p=0.0090) and relative abundance of certain cecal microbes (e.g., Lachnospiraceae Dorea longicatena with p=0.049). There were no differences in the LBP, Lbp, and Cd14 levels in the liver and small intestine between obese mice with and without yogurt supplementation (p>0.05).
Conclusions
Yogurt consumption inhibited obesity-induced inflammation in mice by modulating colonic endotoxin detoxification, changing the gut microbiota, and improving glucose metabolism. This work helps to establish the underlying mechanisms by which yogurt consumption affects markers of metabolic and immune health.
Methods
Animals and experimental design
C57BL/6 male mice were obtained from Jackson Laboratory (Bar Harbor, ME) at three weeks old and acclimated for 2 weeks. Mice were housed together with 2 to 5 per cage and maintained by the University of Wisconsin-Madison Research Animal Resource Center and reared under 12 h light-dark cycle. Mice were received in small batches from the breeding center and assigned to the same group and co-housed to avoid fighting. All experiments were approved by the Institutional Animal Care and Use Committee of the University of Wisconsin-Madison (A005914).
Western diets (WDs) induce metabolic stress through the TLR4 pathway. The objective of the feeding protocol was to test the impact of yogurt consumption in obese and non-obese mice in the context of a WD. WDs were selected to more closely mimic dietary patterns in which yogurt reduced inflammation in human participants. Therefore, in this study, a standard high-fat diet (HFD) was used to induce obesity in mice, then obese and non-obese groups consumed Western diets with or without yogurt supplementation.
For 11 weeks, mice were randomized by cage and fed a Teklad Global 14% Protein Rodent Maintenance Diet (2014C: 13, 20, 67 % calories were from fat, protein, carbohydrate; ENVIGO, IN) as a control group (C group: n=33) or a HFD (TD.06414: 60, 18, 21 % calories were from fat, protein, carbohydrate; ENVIGO) to sufficiently induce obesity (O group: n=37). From Week 11, approximately half of both the C group and O group were randomized by cages and fed either a WD (TD.07201: 42, 15, 43 % calories were from fat, protein, carbohydrate; ENVIGO) (LC group: n=18; OC group: n=18) or a WD supplemented with 5.6% lyophilized yogurt in place of corn starch in TD.07201 (TD.190642: 44, 16, 40% calories were from fat, protein, carbohydrate; ENVIGO) (LY group: n =15 OY group: n=19) for another 3 weeks. The yogurt powder was produced as follows: 2 % reduced-fat milk (Kemps, MN) was pasteurized at 80-85 °C for 30 min and fermented by Yo-Mix 495 LYO 750 DCU (Streptococcus thermophilus and Lactobacillus delbrueckii subsp. bulgaricus) and HOWARU® Dophilus LYO 100 DCU-S (Lactobacillus acidophilus NCFM) supplied by Danisco-Dupont at 40.6 °C for 6 h, followed by lyophilization to obtain the dried yogurt powder. No animals died throughout the study.
Body weight and food intake were recorded weekly. At Week 14, after 4 h of fasting, mice were euthanized by isoflurane and the following samples were collected and frozen with liquid nitrogen to be stored at -80 °C: plasma, gonadal, retroperitoneal, and omental adipose tissues, liver, small intestine, colon, and feces and cecal samples. Upon collection, the tissue samples were weighed.
Plasma glucose and insulin assessment
Plasma glucose level was assessed using Glucose Colorimetric Assay Kit (Cayman Chemical, MI). Plasma insulin level was assessed by Mouse Insulin ELISA kit (Mercodia, NC). Homeostatic model assessment for insulin resistance (HOMA-IR) was calculated as follows: Glucose (mg/dL) * Insulin (mU/L) / 405.
Cytokine assessment
Interleukin (IL)-1b, IL-6, IL-10, tumor necrosis factor (TNF)-a, and leptin in the plasma, as well as keratinocytes-derived chemokine (KC), TNF-a, IL-1b, IL-4, IL-6, IL-10, IL-17A, and interferon (IFN)-g levels in the omental adipose tissue and distal colon samples measured by a customized multiplex cytokine panel using QuickPlex SQ 120 imager (Meso Scale Discovery, Rockville, MD). Some samples were removed from statistical analysis because the concentrations were below detection limit, suspected of a technical error, or of insufficient sample volume to complete the analysis.
LBP and (s)CD14 assessment
Plasma and liver LBP were measured by Mouse LBP ELISA kit (Abcam, Cambridge, UK) and plasma sCD14 by Mouse CD14 Quantikine ELISA kit (R&D Systems, Inc., MN). Total protein was extracted from liver samples as follows: 30 mg tissue was mixed with 1.5 mL of ice-cold lysis buffer (25 mM Tris-HCl at pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 % NP-40, and 5 % glycerol), homogenized by bead beating for 45 sec at 5 m/s, incubated for 2 h with constant agitation at 4 °C, centrifuged for 20 min at 14,000 ×g at 4 °C, and the supernatant was collected for the ELISA assay. Lbp and Cd14 expression levels in the liver, small intestine, and colon samples were assessed by RT-qPCR as described by others. Briefly, RNA was extracted using TRIzol (Invitrogen, Waltham, MA) and cleaned up using RNeasy Mini Kit and DNase set (Qiagen, Valencia, CA). cDNA was synthesized by iScriptTM cDNA Synthesis Kit (Bio-Rad, Hercules, CA) and qPCR was run with Fast SYBRTM Green Master Mix (Thermo Fisher) using Bio-Rad CFX96 system (Bio-Rad). The mRNA expressions were normalized using that of ribosomal protein large P0 since its gene expression level showed stable expression in the tissues of interest in HFD-fed C57BL/6J male mice.
16S rRNA sequencing analysis
DNA was extracted from the cecal and fecal samples using DNeasy PowerSoil Kit according to the manufacturer’s protocol. 16S Metagenomic Sequencing Library Preparation Protocol (Part # 15044223 Rev. B, Illumina Inc., San Diego, CA) was used with the following modifications. The 16S rRNA gene V3/V4 variable region was amplified with fusion primers and cleaned with a 0.7x volume of AxyPrep Mag PCR clean-up beads (Axygen Biosciences, Union City, CA). Then, Illumina dual indexes and sequencing adapters were added using the primers shown in Supplementary Table 5, which was then cleaned using the same kit written above. Quality and quantity of the constructed libraries were assessed by an Agilent 4200 TapeStation DNA 1000 kit (Agilent Technologies, Santa Clara, CA) and Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific), respectively. Libraries were pooled in an equimolar fashion and appropriately diluted prior to sequencing. Paired end 300 bp sequencing was performed using the Illumina MiSeq Sequencer and a MiSeq 600 bp (v3) sequencing cartridge in the University of Wisconsin-Madison Biotechnology Center. Data was analyzed using the standard Illumina DATA2 Pipeline, version 1.8.2 to generate amplicon sequence variant table for further data analyses.
Statistical analyses
The sample sizes among groups varied by assays and tissue availability. Prior to applying the following statistics, the Shapiro test was used to assess data distribution, and Levene’s test was used to test for data homogeneity. Datasets that had non-normal distribution (datasets except for weight, diet intake, and individual bacterial abundance datasets), log10 transformation was applied to establish normal distribution.
For measurements obtained over multiple time points, a repeated-measure three-way analysis of variance (ANOVA) with a mixed-effects model was used due to the unbalanced sample size among groups. The statistics were applied for “Obese” (body condition at Week 11; L or O groups), “Yogurt” (with or without yogurt supplementation in WD; C or Y groups), “Time,” and the interaction terms, where the resulting p-values are summarized in Tables 1-2 or directly in Figures (e.g., pobese, pyogurt). As post-hoc test, two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli was used, and p-values are summarized in figures. A two-way ANOVA was applied for the rest of the measurements collected at single time point, except for the microbiota dataset, for the effects of Obese, Yogurt, and the interaction terms. Sidak’s multiple comparison test was used for post-hoc tests. Pearson correlation test was used to assess the correlation between plasma LBP, sCD14, cytokines, or leptin, and HOMA-IR. The interpretation of effect size (correlation coefficient, R) is as follows: small when R is less than 0.30, moderate when R is between 0.3–0.5, high when R is above 0.5. Statistical results with p<0.05 were defined to be significant and false discovery rate corrected p-values were used for multiple comparison analyses.
Bacterial abundance with significant group difference was identified by likelihood-ratio test in DESeq2 (version 1.32.0). P-values from the statistics were adjusted by the Benjamini-Hochberg method, and features with adjusted p-values less than 0.05 were considered to have significant group difference. Those with identified genera were reported in the manuscript. Rarefied amplicon sequence variant counts were used to assess the alpha and beta diversity. Alpha diversity was measured using Shannon index. For beta diversity, the data dispersion was assessed by betadisper (vegan package, version 2.5.7), which showed that the homogeneity of data dispersion between groups was significantly different in fecal samples. Therefore, the group difference in the centroids was assessed by analysis of similarities (ANOSIM). Beta diversity was visualized by non-metric multi-dimensional scaling plots based on Bray Curtis distance.