Intestinal FFA3 mediates obesogenic effects in mice on a Western diet

Abstract

Free fatty acid receptor 3 (FFA3) is a recently-deorphanized G-protein-coupled receptor. Its ligands are short-chain fatty acids (SCFAs), which are key nutrients derived from the gut microbiome fermentation process that play diverse roles in the regulation of metabolic homeostasis and glycemic control. FFA3 is highly expressed within the intestine, where its role and its effects on physiology and metabolism are unclear. Previous in vivo studies involving this receptor have relied on global knockout mouse models, making it difficult to isolate intestine-specific roles of FFA3. To overcome this challenge, we generated an intestine-specific knockout mouse model for FFA3, Villin-Cre-FFA3 (Vil-FFA3). Model validation and general metabolic assessment of male mice fed a standard chow diet revealed no major congenital defects. Because dietary changes are known to alter gut microbial composition, and thereby SCFA production, an obesogenic challenge was performed on male Vil-FFA3 mice and their littermate controls to probe for a phenotype on a high-fat, high-sugar “Western diet” (WD) compared with a low-fat control diet (CD). Vil-FFA3 mice versus FFA3^fl/fl controls on WD, but not CD, were protected from the development of diet-induced obesity and exhibited significantly less fat mass as well as smaller adipose depositions and adipocytes. Although overall glycemic control was unchanged in the WD-fed Vil-FFA3 group, fasted glucose levels trended lower. Intestinal inflammation was significantly reduced in the WD-fed Vil-FFA3 mice, supporting protection from obesogenic effects. Furthermore, we observed lower levels of gastric inhibitory protein (GIP) in the WD-fed Vil-FFA3 mice, which may contribute to phenotypic changes. Our findings suggest a novel role of intestinal FFA3 in promoting the metabolic consequences of a WD, including the development of obesity and inflammation. Moreover, these data support an intestine-specific role of FFA3 in whole body metabolic homeostasis and in the development of adiposity.

NEW & NOTEWORTHY Here, we generated a novel intestine-specific knockout mouse model for FFA3 (Vil-FFA3) and performed a comprehensive metabolic characterization of mice in response to an obesogenic challenge. We found that Vil-FFA3 mice fed with a Western diet were largely protected from obesity, exhibiting significantly lower levels of fat mass, lower intestinal inflammation, and altered expression of intestinal incretin hormones. Results support an important role of intestinal FFA3 in contributing to metabolism and in the development of diet-induced obesity.

INTRODUCTION

The gut microbiome is a crucial determinant of human health, and alterations in its composition have been linked to a growing number of disease states (1, 2). Comprised of more than 100 trillion microbes living in the digestive tract, the gut microbiome is shaped early in life and is influenced by numerous environmental factors, including dietary intake (3). Recent studies have found that the composition of the gut microbiome can change markedly in response to diet, and that dietary changes lead to alterations in microbial metabolites, which act as major regulators of human metabolism (4, 5). Short-chain fatty acids (SCFAs) are an important group of metabolites that are generated via the fermentation of non-digestible carbohydrates by the gut microbiota and have been suggested to have metabolically beneficial effects (6, 7). The SCFAs that are physiologically relevant include acetate, propionate and butyrate, and they act as ligands for their cognate receptors, which sense and mediate the many actions of SCFAs in their regulation of metabolic homeostasis (8).

Free fatty acid receptor 3 (FFA3) is a recently-deorphanized G-protein-coupled receptor (GPCR) that signals exclusively via a Gα_i/o pathway, producing a predominantly inhibitory tone on intracellular cyclic adenosine monophosphate (cAMP) levels (9). FFA3 is expressed in multiple tissues throughout the body, including the pancreas and nervous system, and is thought to contribute to metabolic homeostasis through the summation of its tissue-specific effects (10). Within the pancreatic beta cell, FFA3 inhibits insulin secretion and modifies the β cell response to conditions of stress (11). In the nervous system, FFA3 is expressed within a variety of neural populations, including the sympathetic ganglia, vagus nerve, and enteric nervous system, and has been suggested to play a role in mediating gut-brain functions including regulation of food intake and energy expenditure (12, 13).

FFA3 is highly expressed within the epithelium of the digestive tract in both humans and mice, specifically within enteroendocrine cells (EECs) and intestinal epithelial cells (IECs) (12, 14, 15). Within EECs, FFA3 has been suggested to mediate the release of postprandial peptide hormones GLP-1 and PYY, which together modulate glucose metabolism, intestinal transit and satiety (16, 17). These roles are still considered controversial, as there is a large body of conflicting data regarding the ability of FFA3 to mediate the secretion of multiple hormones in response to SCFAs. Within IECs, FFA3 has been suggested to play a role in modulating protective immunity and tissue inflammation in both human and mouse models (18, 19). Overall, the role of intestine-specific FFA3 and its effects on physiology and metabolism is still largely unclear, in part due to the use of global knockout mouse models to characterize tissue-specific effects.

Here, we report the first intestine-specific characterization of FFA3 in vivo using a novel knockout mouse model, Villin-Cre-FFA3. We conducted a general metabolic assessment of male Villin-Cre-FFA3 mice on normal chow and then performed an obesogenic challenge using a high-fat, high-sugar “Western diet.” Our strategy reveals novel contributions of intestinal FFA3 to metabolic homeostasis, as well as in the development of obesity and hyperglycemia.

MATERIALS AND METHODS

Animals

Animal studies were conducted in accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health and with the approval of University of Illinois at Chicago Institutional Animal Care and Use Committee (IACUC). Villin-Cre mice were obtained from Jackson Laboratory (Maine). Mice were group-housed under standard temperature and humidity conditions with a 12:12-hr light/dark cycle and had ad libitum access to standard rodent diet (Teklad LM-485, Envigo, Indianapolis, IN). For the obesogenic challenge experiment, mice received either a low-fiber control diet (CD; Custom Product No. D19010907, Research Diets, Inc., New Brunswick, NJ) or a low-fiber Western diet (WD; Custom Product No. D19010908, Research Diets) beginning at 10-wk of age. Both diets were customized with amioca corn starch replacing standard corn starch. The formulation of nutrients in each diet is listed in Supplemental Table S1 for standard rodent diet and Supplemental Table S2 for the experimental diets used during the obesogenic challenge study. Only male mice were used in this study.

Generation of Villin-Cre-FFA3 Mice

Ffar3 floxed mice were generated using the targeted knockout first, reporter-tagged insertion with conditional potential strategy by the International Knockout Mouse Consortium. The critical portion of exon 2 was flanked by loxP sites, with upstream elements including FRT – LacZ – loxP – neo – FRT to generate the targeted allele (Ffar3^tm1a). IVF was performed at the Transgenic and Targeted Mutagenesis Laboratory (Northwestern University, Chicago, IL), and subsequent breeding was carried out at University of Illinois at Chicago. Flp-mediated recombination converted the mutated allele to conditional allele. All subsequent breeding was carried out at the University of Illinois at Chicago. Ffar3 floxed mice of C57BL/6J background were mated with Villin-Cre (+/−) to generate Villin-Cre-FFA3 mice and FFA3^fl/fl littermate controls.

Metabolic Assessment

For the chow diet experiment, mice were monitored weekly for body weight, blood glucose levels, and plasma insulin levels under ad libitum and fasting conditions from 6 wk of age to 20 wk of age. Fasting measurements were taken after a 16-h, overnight fast. Blood for insulin measurements was collected from the tail vein. For the obesogenic challenge, mice were placed on either CD or WD from 10 to 35 wk of age.

Fecal Collection and 16S Sequencing

Bacterial genomic DNA was extracted from approximately 100 mg of mice feces using the Qiagen DNA Stool Mini Kit (Qiagen, Hilden, Germany). The gut microbiota composition was analyzed using the Illumina MiSeq platform (Illumina, San Diego, CA) and 16S rRNA approach along with bioinformatics tools established in our laboratory (20–22). The 16S rDNA V4 hypervariable region was amplified using universal primer pairs 515 F (barcoded) and 806 R (23). The unique barcoded amplicons were purified using AMPure magnetic purification beads (Agencourt, Beckman Coulter, CA) and further quantified using the dsDNA HS assay kit (Life Technologies, Carlsbad, CA) in a Qubit-3 fluorimeter (InVitrogen, Carlsbad, CA) and normalized to prepare the amplicon library (23). The library consisting of each amplicon in equal concentration (8 pM) was used for sequencing on an Illumina MiSeq sequencer using Miseq reagent kit v3. The sequenced bacterial sequences were demultiplexed, quality filtered, clustered, and analyzed by using quantitative insights into microbial ecology 2 (QIIME2) and R-based analytical tools (20, 21).

Nuclear Magnetic Resonance Spectroscopy

Body composition, including lean tissue, fat tissue, and free body fluid, was quantified via Nuclear Magnetic Resonance (NMR) every 4 wk following dietary intervention. Measurements were taken using a benchtop LF50 BCA-Analyzer (Bruker Corporation, Billerica, MA). Animals were not anesthetized and were placed in the provided Bruker measurement chamber, which used a plunger mechanism to restrict excessive animal movement. Measurements were taken in singlicate for each animal.

Metabolic Cages

Following 3 wk of dietary intervention, mice were individually housed in metabolic cages (Mouse Promethion Continuous caging system; Sable Systems, Las Vegas, NV). Cages were maintained at a temperature of 22°C on a standard 12-h light/dark cycle and mice were acclimatized for 24 h before data collection. Carbon dioxide (CO₂) production, oxygen (O₂) consumption, and respiratory exchange ratio (RER) were measured via continuous air-flow sampling. Sensory technology was used to measure metabolic parameters including energy expenditure, body mass, food intake, and water intake. The experiment was repeated following 10 wk of dietary intervention. Cumulative food and water intake during this timepoint was measured by subtracting the remainder of food and water from their respective starting measurements over a period of 96 h in singly housed cages. Data are presented as hourly averages for each metabolic parameter. Energy expenditure was analyzed by ANCOVA using scripts available at the National Mouse Metabolic Phenotyping Centers (MMPC, Nashville, TN) Energy Expenditure Analysis Page (https://www.mmpc.org/shared/regression.aspx) using total body mass as the covariant.

Food Intake

The BioDaq food intake monitoring system (Research Diets) was used to record food intake from each mouse. Mice were individually housed and acclimated to the food intake cages for 5 days before recording. During the 96-h recording period, mice were given ad libitum access to diet. Data was analyzed using the corresponding BioDaq DataViewer software.

Glucose Tolerance Test

All mice were fasted for 16 h overnight then injected intraperitoneally [IP-glucose tolerance test (IP-GTT)] with 2 g/kg dextrose (Hospira, Lake Forest, IL). For oral glucose tolerance test (O-GTT), mice were administered dextrose via oral gavage. Blood glucose measurements were taken via tail bleed using a OneTouch Ultramini glucometer (Lifescan, Malvern, PA) at timepoints 0, 15, 30, 60, and 120 min, and whole blood was also collected in heparinized capillary tubes for insulin measurement at timepoints 0, 15, and 30 min. For the chow diet study, mice underwent both O-GTTs and IP-GTTs between 21 and 23 wk of age. For the obesogenic challenge study, an O-GTT was performed before dietary intervention at 9 wk of age and then again following 14 wk of dietary intervention.

Insulin Tolerance Test

Mice were fasted for 6 h from the beginning of the light cycle. Insulin tolerance tests (ITTs) were performed on mice with insulin (Humalog U-100, Eli Lilly, Indianapolis, IN) administered intraperitoneally at a dose of 1.0 U/kg body wt. Blood glucose measurements were taken via tail bleed using a OneTouch Ultramini glucometer (Lifescan) at timepoints 0, 5, 10, 15, 20, 25, and 30 min. The glucose disappearance rate (K_ITT) was calculated from the established formula K_ITT = (0.693/t_1/2) × 100, where glucose half-life (t_1/2) stands for time for glucose to reach 50% of the basal value (24, 25). Glucose t_1/2 was calculated from the slope of the least square analysis of blood glucose concentration during the linear phase of decline.

Plasma Insulin ELISA

Whole blood was also collected in heparinized capillary tubes, centrifuged at 2,500 g for 15 min at 4°C and used for measurement of insulin by mouse anti-insulin enzyme-linked immunosorbent assay (ALPCO, Salem, NH).

Oral Fat Tolerance Test

Mice were fasted overnight for 16 h and were then administered a 200-µL bolus of olive oil (Sigma Aldrich, St. Louis, MO) via oral gavage. Whole blood from the tail vein was collected at 0, 60, 120, 240, and 300 min after oral gavage and then centrifuged for the extraction of serum portion.

Serum Triglyceride Levels

Serum triglyceride levels were quantified using the Infinity triglycerides liquid stable reagent assay (Thermo Fisher Scientific, Inc., Middletown, VA) from serum collected at 0, 60, 120, 240, and 300 min after olive oil bolus.

Serum Cholesterol Levels

Serum cholesterol levels were quantified using the Infinity cholesterol liquid stable reagent assay (Thermo Fisher Scientific) from serum collected at 0, 60, 120, 240, and 300 min after olive oil bolus.

Gastrointestinal Transit Time and Fecal Measurements

Mice were transferred to individual cages before the experiment and were singly housed until the completion of the study. A 6% solution of a nonabsorbable dye, carmine red (natural red 4; Sigma Aldrich) in 0.5% methylcellulose was prepared and 0.2 mL was administered to each mouse by oral gavage (26). The cage floor was covered with white paper to facilitate detection of the red dye in feces. The time of gavage was set as t₀. Following gavage, mice were left undisturbed without food and water until the first red fecal pellet appeared (t_end). Gastrointestinal transit time (GITT) was calculated as t_end − t₀. During the first 2 h of the study, all fecal pellets were collected to calculate the number of pellets and fecal pellet weight. Mice that did not produce any pellets were excluded from analysis.

Tissue Isolation and Endpoint Measurements

For the chow diet study, all mice were euthanized at 25 wk of age. For the obesogenic challenge study, all mice were euthanized at 35 wk of age. Tissues were removed, weighed if necessary, and then immediately snap frozen in liquid nitrogen for RNA isolation.

Whole Blood Sampling and ELISA Assays

Following a 4-h fast, mice were administered an oral dextrose bolus and anesthetized via 5% isopropanol inhalation several minutes before euthanasia. Mice were euthanized by cardiac puncture 15 min after administration of oral glucose bolus and whole blood was collected and stored in heparinized blood collection tubes. A cocktail of protease inhibitors including dipeptidyl peptidase-4 inhibitor (Sigma Aldrich) and aprotinin (Phoenix Pharmaceuticals Inc., Burlingame, CA) was immediately added, and samples were stored on ice for 30 min before centrifugation (2,500 rpm) at 4°C for 15 min. Plasma portions were isolated for ELISA assays. Plasma GLP-1 levels and gastric inhibitor protein (GIP) levels were quantified by mouse total GLP-1 ELISA kit (CrystalChem, Elk Grove Village, IL) and mouse total GIP ELISA kit (Millipore, Billerica, MA), respectively. Lipopolysaccharide (LPS) levels were measured with the murine LPS ELISA kit (Cusabio, Wuhan, China).

RNA Isolation and qPCR

RNA was extracted from tissues using TRIzol reagent (Life Technologies) and chloroform for phase separation. RNA was purified using the RNeasy mini kit (Qiagen) and treated with RNase-free DNase (Qiagen). Purified RNA (1 µg) was reversely transcribed using iScript Reverse transcription (Bio-Rad Laboratories, Hercules, CA), and gene expression was measured by qPCR using SYBR Green SuperMix (Bio-Rad Laboratories). Final primer concentrations were 0.250 µM for each reaction, and data were analyzed via the CFX connect real-time PCR detection system (Bio-Rad). The expression of individual genes was normalized to housekeeping gene β-actin and presented as fold-change using the 2^ΔΔCt algorithm. Primer sequences are listed in Supplemental Table S3.

Histology

Tissues were placed in cassettes and fixed in 10% formalin for 48 h following collection. The tissues were then washed twice with PBS and transferred to 70% ethanol at 4°C for storage before paraffin embedding. All tissues were sectioned at 5 µM and transferred to slides at the Research Histology and Tissue Imaging Core of the University of Illinois at Chicago. Tissue sections were stained with hematoxylin and eosin (H&E) using a commercially available staining kit (Vector Laboratories, Burlingame, CA). The slides were then mounted with Poly-mount (Polysciences, Inc., Warrington, PA) and covered with coverslips. Images were acquired with a Leica DMi8 microscope (Leica Biosystems, Wetzlar, Germany).

Histological Assessment of Intestinal Inflammation

All slides were blinded, and pathological scores for intestinal inflammation were assessed according to the following criteria: scores ranging from 0 to 3 were given for each tissue section, where 0 indicates no change and 3 indicates maximal change (27). Parameters scored included architectural changes, chronic inflammatory infiltration, presence of neutrophils, crypt destruction, erosion and ulceration, and edema. Scores for each parameter were totaled to calculate an overall score.

Adipose Droplet Measurements

H&E sections of adipose tissue were prepared as described above and imaged at ×20. ImageJ version 1.53a software (https://imagej.nih.gov/ij) with the adiposoft v. 1.16 plug-in (https://imagej.net/adiposoft) was used to calculate adipocyte size and number of adipocytes per 15 mm² image.

Cecal SCFA Quantification

Cecal content was isolated from the cecal pouch, weighed, and snap frozen at −80°C. The samples were then suspended in 200 µL of 50% MeOH and vortexed thoroughly before sonication for 20 min. Subsequently, the samples were centrifuged at 27.6 g for 10 min, and 30 µL of the supernatant was taken for derivatization. For derivatization, 30 µL of each standard solution or sample supernatant was mixed with 15 µL of 200 mM 3-NPH in 50% aqueous MeOH and 15 µL of 120 mM EDC in the same solution. The reaction was allowed to proceed for 30 min at 40°C. The reaction mix was then diluted with 350 µL of 10% MeOH. A volume of 30 µL of the diluted reaction solution was mixed with 30 µL of premade stable isotope labeled standards for LC/MS analysis using an AB SCIEX 6500 QTRAP coupled with Agilent 1290 UPLC system (Agilent Technologies, Santa Clara, CA). The analysis was performed by the Mass Spectrometry Core in Research Resources Center of University of Illinois at Chicago.

Statistical Analysis

Data are presented as means ± SE. All data were compared by either one-way or two-way ANOVA multiple comparisons (GraphPad Prism 8, La Jolla, CA) followed by a Tukey’s post hoc test when applicable, or Student’s t test. For energy expenditure comparison, ANCOVA was used with total body weight as the covariant. All descriptive statistics are included in figure legends.

Differences in the microbiome’s β diversity were tested by permutational multivariate analysis of variance (PERMANOVA), a permutation-based multivariate analysis of variance to a matrix of pairwise distance to partition the inter-group and intragroup distance. LEfSE [Linear discriminatory analysis (LDA) effect size] was used to identify unique bacterial taxa that drive differences among groups from Galaxy server (https://huttenhower.sph.harvard.edu/galaxy/) (28). The α parameter significance threshold for the Kruskal–Wallis as well as the Wilcoxon signed-rank test, which were implemented among classes, was set to 0.01, the logarithmic LDA score cut-off was set to 3, and the strategy for multi-class analysis was set to “all-against-all.” Comparisons were considered statistically significant when P < 0.05.

RESULTS

No Congenital Defects Were Observed in the Villin-Cre-FFA3 Mice on Standard Chow Diet

The intestine-specific knockout (KO) of Ffar3 within the Villin-Cre-FFA3 (Vil-FFA3) mouse line was validated by measuring gene expression in all intestinal tissues (Supplemental Fig. S1A). We found that expression of Ffar3 was virtually absent in each region of the intestine from the Vil-FFA3 group compared with the FFA3^fl/fl littermate controls. Both groups of mice were characterized on a standard chow diet (Supplemental Fig. S1B) and were metabolically profiled from week 6 to week 25 of age (Supplemental Fig. S1C) to assess for any major congenital defects. Consistent with previous studies performed in global FFA3 KO mice, we observed no differences in body weight between the two groups (Supplemental Fig. S1D) (11). In addition, no change in glucose levels in either ad libitum or fasting states was observed (Supplemental Fig. S1, E and F). Glucose levels during both IP-GTT (Supplemental Fig. S1G) and O-GTT (Supplemental Fig. S1H) did not differ between the Vil-FFA3 group and the FFA3^fl/fl control group, indicating no difference in glucose tolerance, including in response to oral glucose challenge. Insulin levels, measured in both ad libitum and fasting states were the same in both groups (Supplemental Fig. S1I). Taken together, these data indicate that there are no significant congenital metabolic differences between Vil-FFA3 mice and controls fed with standard chow.

Dietary Changes Alter Expression Levels of SCFA Receptors in the Intestine

Because dietary changes are known to affect the composition of the gut microbiome, and thereby SCFA levels, and because FFA3 has been implicated in conditions of metabolic stress, we performed an obesogenic challenge on a group of male Vil-FFA3 mice and FFA3^fl/fl littermate controls. Mice were either placed on a high-fat, high sugar Western diet (WD) or a low-fat control diet (CD), with both diets containing amioca corn starch as the sole source of fermentable fiber (Fig. 1A). All mice were kept on standard chow until 10 wk of age, and baseline measurements were taken, confirming no differences between Vil-FFA3 mice and their littermate controls (Supplemental Fig. S2, A–E). Mice were then placed on either WD or CD and metabolically profiled until 35 wk of age (Fig. 1B).

Figure 1.

Intestinal SCFA receptor expression is altered by dietary changes. Macronutrient compositions of Western diet (WD) and control diet (CD) (A). Experimental timeline and dietary conditions (B). Expression of Ffar3 in the ileum (C) and distal colon (D). Expression of Ffar2 in the ileum (E) and distal colon (F). Values represent means ± SE, n = 5–9 per group, *P < 0.05, **P <0.01 vs. respective genotypes on CD, and “#” indicates a significant difference between Western diet groups at a #P > 0.05, ##P > 0.01, and ####P > 0.0001 using one-way ANOVA.

We first explored the effects of diet on the mRNA levels of SCFA receptors in the intestine. We found that expression of Ffar3 was significantly elevated in the ileum of the WD-fed FFA3^fl/fl mice and, as expected, was abolished in both Vil-FFA3 mouse groups (Fig. 1C). Although expression of Ffar3 was also significantly abolished in the distal colon of the Vil-FFA3 mouse groups, no changes were observed between the CD-fed and WD-fed FFA3^fl/fl groups (Fig. 1D). We also measured the expression levels of free fatty acid receptor 2 (FFA2), a SCFA receptor that is similar to FFA3 and has known roles in the intestine (16, 29, 30). Although no changes were observed in the CD-fed group, we found that FFA2 expression was significantly elevated in the ileum of the WD-fed FFA3^fl/fl mice, whereas only marginally elevated in the WD-fed Vil-FFA3 group (Fig. 1E). This trend was also observed in the distal colon, where FFA2 was even more elevated in both WD-fed groups (Fig. 1F). Collectively, these data indicate that expression of FFA2 and FFA3 are influenced by diet, and that expression of intestinal FFA2 may be diminished when intestinal FFA3 is absent.

Intestinal FFA3 Mediates Changes in Diet-Induced Adiposity

Body weights in the WD-fed FFA3^fl/fl mice were significantly increased compared with the CD-fed groups, and this effect was lost in the WD-fed Vil-FFA3 group, which did not differ from the CD-fed groups (Fig. 2A). Measurements in body composition taken throughout the study using NMR revealed that the WD-fed Vil-FFA3 group developed significantly less fat mass compared with the WD-fed FFA3^fl/fl control group (Fig. 2B). Consistent with these results, weights of subcutaneous (SAT) and visceral adipose depots (VAT) at euthanasia after 25 wk of the WD were significantly lower in the WD-fed Vil-FFA3 group than those extracted from the WD-fed FFA3 ^fl/fl group (Fig. 2, C and D). These changes were absent in the CD-fed mouse groups.

Figure 2.

Diet-induced adiposity is significantly reduced in Vil-FFA3 mice. Body weight measurements during and at the conclusion of the study (A). Fat mass represented as a percentage of total body weight (B). Raw weight and images of inguinal subcutaneous (SAT) fat pad (C) and epididymal visceral (VAT) fat pad (D) measured at the conclusion of the study. Hematoxylin and eosin (H&E) stain of adipocytes from SAT and VAT (scale bar = 100 µm) (E). Average area of individual adipocytes in SAT (F) and VAT (G). Number of adipocytes per image in SAT (H) and VAT (I). For A–D, n = 7–9 per group. For F–I, n = 9–12 per group. Values represent means ± SE, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 vs. respective genotypes on control diet (CD), and “#” indicates a significant difference between Western diet groups at a #P > 0.05, ##P> 0.01, and ####P > 0.0001. For A and B, significance was calculated using two-way ANOVA and Tukey post hoc test analyses. For C and D and F–I, significance was calculated using one-way ANOVA.

Long-term consumption of high levels of fat and sugar creates a physiological state of energy surplus in which lipid-storing adipocytes expand and become hypertrophic (31). To assess for hypertrophy, fixed and embedded adipose tissues were stained for morphology using hematoxylin and eosin. Extensive adipocyte hypertrophy was observed in the WD-fed FFA3^fl/fl control group but was significantly reduced in the WD-fed Vil-FFA3 group (Fig. 2E), and this effect was absent in the CD-fed groups. A similar trend was observed in both subcutaneous and visceral adipose tissue. Measurements of subcutaneous adipocyte droplet area also revealed a significant increase in adipocyte size in the WD-fed FFA3^fl/fl group when compared with the WD-fed Vil-FFA3 group, which both had significantly larger adipocytes than the CD-fed groups (Fig. 2F) as quantified by droplets per section. The same trend was observed in the visceral adipocyte droplets, however, adipocytes from the WD-fed Vil-FFA3 mice did not increase in size when compared with the CD-fed groups (Fig. 2G). The number of adipocyte droplets per section was significantly lower in the WD-fed FFA3^fl/fl group compared with the WD-fed Vil-FFA3 group, and an overall decrease of adipocyte droplet number was observed in both WD-fed groups when compared with the CD groups in both SAT (Fig. 2H) and VAT (Fig. 2I). Reduction in adipocyte number in the WD-fed Vil-FFA3 group was more prominent in the VAT than the SAT. These data collectively show a clear reduction in obesogenic diet-induced hypertrophy of both SAT and VAT when FFA3 is absent in the intestine. Furthermore, WD-fed Vil-FFA3 were significantly protected from obesity, indicating that intestinal FFA3 mediates changes in adiposity in response to an obesogenic diet.

Loss of Intestinal FFA3 Does Not Overtly Alter Glucose Homeostasis in Response to an Obesogenic Challenge

FFA3 is known to contribute to glucose homeostasis by mediating the secretion of insulin within the pancreatic β cell under conditions of metabolic stress (11). Because pancreatic function is strongly influenced by signaling from the digestive system, we posited that there may be an intestine-specific contribution of FFA3 to glucose homeostasis and probed for glucometabolic differences in the Vil-FFA3 mice. There were no differences in ad libitum glucose levels between the Vil-FFA3 and FFA3^fl/fl mice fed either WD or CD throughout the study (Fig. 3A). After 6 wk on their respective diets, fasting glucose levels in the WD-fed Vil-FFA3 group began trending lower than the WD-fed FFA3^fl/fl group, but only reached significance at week 12 of the study (Fig. 3B). Results from an oral glucose tolerance test performed at week 14 of the study showed that both WD-fed groups developed glucose intolerance when compared with the CD-fed groups, but there was no difference between the Vil-FFA3 and FFA3^fl/fl groups (Fig. 3, C and D). Although fasting insulin levels also measured at week 14 of the study were unchanged (Fig. 3E), insulin levels measured 15 min into an O-GTT at the end of the study (Fig. 3F) were significantly elevated in the WD-fed FFA3^fl/fl group while unchanged in the WD-fed Vil-FFA3 group compared with the CD-fed groups.

Figure 3.

Fasting glucose and insulin levels are modestly lower in Vil-FFA3 mice with no change in glucose tolerance or insulin tolerance. Ad lib (A) and fasting (B) glucose levels measured throughout the study. O-GTT measured after 14 wk on diet (C) and corresponding AUC (D). Fasting plasma insulin levels measured after 18 wk on diet (E). Insulin levels from plasma collected 15 min into an O-GTT at study endpoint (F). ITT (G), glucose clearance (H), and insulin half-life (I) measured after 12 wk on diet. HOMA-IR calculated after 14 wk on diet (J). For A–F and J, n = 5–10 per group. For G–I, n = 3–4 per group. Values represent means ± SE, *P > 0.05 vs. respective genotypes on CD, and “#” indicates a significant difference between Western diet groups at a #P > 0.05 using two-way ANOVA and Tukey’s post hoc test analyses in B and one-way ANOVA in F. AUC, area under the curve; ITT, insulin tolerance test; O-GTT, oral-glucose tolerance test.

Consumption of an obesogenic diet can adversely affect insulin sensitivity (32). We performed an insulin tolerance test (ITT) on the mice on week 12 of the study and observed no changes in insulin sensitivity in any of the groups (Fig. 3G). No changes in glucose clearance rate (Fig. 3H) or insulin half-life (Fig. 3I) were observed. HOMA-IR, a measurement of insulin resistance, trended higher in the WD-fed FFA3^fl/fl group while unchanged in the WD-fed Vil-FFA3 group, though measurements within each group were largely varied (Fig. 3J). Although fasting glucose levels and glucose-stimulated insulin levels trended lower in the WD-fed Vil-FFA3 mice, no major changes in glucose homeostasis, including glucose tolerance and insulin sensitivity were observed. This suggests that the absence of intestinal FFA3 does not overtly affect glucose homeostasis in response to an obesogenic challenge but may contribute to a modest improvement in fasting glucose and insulin levels.

Phenotypic Changes Are Not Driven by Alterations in Food Intake, Energy Expenditure or Respiratory Exchange Ratio

Changes in adiposity can be reflective of alterations in metabolic processes including energy expenditure, respiratory exchange ratio (RER), and food intake. Alterations in these parameters were assessed using indirect calorimetry on mice using metabolic chambers at different timepoints during the study. Metabolic parameters were first measured 3 wk following dietary invention, before a divergence in body weight, and then again at 10 wk following dietary intervention, in which clear differences in adiposity were observed. O₂ consumption and CO₂ production were unaffected by either diet or genotype (Fig. 4, A and B). RER, which is derived from O₂ and CO₂, shifts to lower values during fat consumption and higher during carbohydrate consumption or extreme exercise. In agreement with this, mice fed the WD had noticeably reduced RER levels, indicating the metabolism of fat. However, there were no differences in RER between the Vil-FFA3 mice and control mice for either the WD or CD groups (Fig. 4C). Energy expenditure, as normalized by total body weight, was unaltered by diet or genotype (Fig. 4D). ANCOVA analysis was performed to adjust energy expenditure variables for the group differences in body weight. Averaged ANCOVA-adjusted energy expenditure totals separated by light cycle (Fig. 4E) and dark cycle (Fig. 4F) did not differ between the Vil-FFA3 and FFA3^fl/fl groups on either diet. Finally, cumulative food and water intake measurements were unchanged in both WD-fed and CD-fed groups (Fig. 4, G and H).

Figure 4.

Phenotypic changes are not driven by alterations in food intake, energy expenditure, or respiratory exchange ratio. Oxygen consumption (A), carbon dioxide production (B), and respiratory exchange ratio (RER) (C) measured over a 48-h period in singly housed metabolic cages following 3 wk of dietary intervention. Energy expenditure (EE) (D) normalized by total body weight and average ANCOVA-adjusted EE during the light cycle (E) and dark cycle (F). Cumulative amounts of food intake (G) and water intake (H) during the study. Oxygen consumption (I), carbon dioxide production (J), and respiratory exchange ratio (RER) (K) measured following 10-wk dietary intervention. Energy expenditure (EE) (L) normalized by total body weight and average ANCOVA-adjusted EE during the light cycle (M) and dark cycle (N). Cumulative food intake (O) and cumulative water intake (P) totals as measured over a 96-h period. For A–F, n = 3–6 per group. For G–N, n = 6–7 per group. For O and P, n = 8–10 per group. Values represent means ± SE. Significance was calculated using two-way ANOVA and Tukey post hoc test analyses in A–D and G–L, and one-way ANOVA in O and P. For E and F and M and N, total body weight was selected as the covariant and data were analyzed using analysis of covariance (ANCOVA) and presented as ANCOVA-adjusted values separated by light cycle.

Metabolic parameters were reassessed following 10 wk of dietary intervention, after the development of differences in adiposity. Similarly, no differences in O₂ consumption (Fig. 4I), CO₂ production (Fig. 4J) or RER (Fig. 4K) were observed. Energy expenditure, as normalized by total body weight, was unaltered by diet or genotype (Fig. 4L) at this timepoint. Similarly, averaged ANCOVA-adjusted energy expenditure totals separated by light cycle (Fig. 4M) and dark cycle (Fig. 4N) did not differ. Cumulative food intake (Fig. 4O) and water intake (Fig. 4P) totals measured over a period of 96-h also showed no differences between groups. Taken together, these data indicate that the resistance to obesity in the WD-fed Vil-FFA3 mice is not significantly impacted by alterations in food intake, energy expenditure or RER.

Intestinal Inflammation Is Reduced in WD-Fed Vil-FFA3 Mice

Inflammation plays a key role in the pathogenesis of diet-induced obesity (33, 34). While FFA3 has been linked to a pro-inflammatory role within IECs, its effects on both intestinal and systemic immunity have yet to be elucidated (18). To assess whether changes in adiposity and glucose homeostasis are influenced by changes in intestinal inflammation, we examined the morphology of H&E stained sections of both the ileum and distal colon. We observed significant changes in the WD-fed FFA3^fl/fl mice, including loss of epithelial morphology and immune cell infiltration, that were notably reduced in the WD-fed Vil-VFFA3 mice (Fig. 5A). These changes were more markedly reduced in the ileum than in the distal colon. Blinded histopathological scoring revealed severe mucosal damage and inflammation both in the ileum and distal colon of the WD-fed FFA3^fl/fl group compared with CD group (Fig. 5B). Mild inflammation was also observed in the CD-fed groups, which may be due to the reduction of dietary fibers within the experimental diets. Overall levels of inflammation were much higher in the distal colons of both WD-fed groups, with significant reduction in the Vil-FFA3 group.

Figure 5.

Markers of intestinal inflammation are reduced in Vil-FFA3 mice. Representative micrographs of hematoxylin and eosin (H&E)-stained distal colonic tissues (scale bar = 100 μm) (A). Graphical representation of the average histopathological score for inflammation (B). Ileal gene expression levels of inflammatory markers F4/80 (C), Cxcl1 (D), Lcn2 (E), and Tnfα (F). Colonic gene expression levels of inflammatory markers F4/80 (G), Cxcl1 (H), Lcn2 (I), and Tnfα (J). Plasma LPS levels (K), colon length (L), and cecum weight (M) measured at the conclusion of the study. GI-transit time as measured via carmine red dye assay after 15 wk on respective diets (N). For B, n = 3–6 per group. For K–N, n = 5–9 per group. For C–J, n= 5–8 per group. Values represent means ± SE, **P < 0.01, ***P < 0.001, and ****P < 0.0001 vs. respective genotypes on control diet (CD), and “#” indicates a significant difference between Western diet groups at a #P > 0.05, ##P> 0.01, and ####P > 0.0001 calculated using one-way ANOVA.

Consistent with this, we found that WD-fed FFA3^fl/fl mice had significantly increased expression of pro-inflammatory markers in the ileum, including macrophage-derived F4/80 (Fig. 5C) and chemokine Cxcl1 (Fig. 5D), as compared with WD-fed Vil-FFA3 mice. Major mediators of intestinal inflammation, including Lcn2 (Fig. 5E) and TNFα (Fig. 5F) also trended higher in the WD-fed FFA3^fl/fl mice compared with the WD-fed Vil-FFA3 group. Similarly, F4/80 expression was also significantly increased (Fig. 5G) in distal colon of WD-fed FFA3^fl/fl mice while Cxcl1 and Lcn2 expression trended high (Fig. 5, H and I). No alterations of TNFα were observed in the distal colon (Fig. 5J). Collectively, these data reveal that WD-induced inflammation is significantly reduced in mice lacking intestinal FFA3.

WD-induced intestinal inflammation is often accompanied by mild endotoxemia, which can be observed as an increase in plasma LPS levels. However, neither group showed any alteration in plasma LPS levels (Fig. 5K). Other features of severe inflammation, including shortening of colon and alteration in GI-TT, as observed in colitis (35), were also absent in our models (Fig. 5, L and N). In addition, cecum content weight representing microbial content (36) was slightly lower in the WD-fed groups but did not differ between the FFA3^fl/fl and the Vil-FFA3 groups (Fig. 5M).

To determine if intestinal fat absorption is altered due to the loss of intestinal FFA3, we performed an oral fat tolerance test on both WD-fed mouse groups and subsequently measured serum triglyceride levels and cholesterol levels. As expected, serum triglyceride levels peaked at 120 min after fat bolus in both WD-fed groups (Supplemental Fig. S3A), but levels were similar between the two groups at each timepoint. Similarly, serum cholesterol levels measured throughout the experiment did not differ between the WD-fed FFA3^fl/fl and the Vil-FFA3 groups (Supplemental Fig. S3B).

Expression Levels of Gut Peptide Hormones Are Altered by the Loss of Intestinal FFA3

FFA3 is hypothesized to mediate the release of postprandial gut hormones, though data are scarce and rely on global KO models of the receptor (15, 16, 29, 37). To determine if intestinal FFA3 deletion leads to chronic effects that change major gut peptide hormone expression, we measured their mRNA levels within the intestine. In both the ileum (Fig. 6A) and distal colon (Fig. 6B), we found that mRNA levels of Gcg, and Pyy trended higher on the WD and were significantly elevated in the distal colon but did not differ between genotypes. Levels of Cck mRNA, encoding for an important postprandial hormone that regulates digestion and appetite, was induced by WD in the colon (Fig. 6B), but not ileum (Fig. 6A) and this effect was lost in the absence of intestinal FFA3. Levels of Gip mRNA trended higher in the WD-fed groups and were significantly reduced in the WD-fed Vil-FFA3 group. These data indicate roles of FFA3 in mediating levels of CCK and GIP expression in the gut, but not for GLP1 or PYY.

Figure 6.

Expression levels of gut peptide hormones are altered by loss of intestinal FFA3. Ileal gene expression levels of incretin hormones Gcg, Pyy, Cck, and Gip (A). Colonic gene expression levels of incretin hormones Gcg, Pyy, Cck, and Gip (B). Plasma levels of GLP-1 (C) and GIP (D) measured 15 min after an oral bolus of glucose. For A and B, n = 4–7 per group. For C and D, n = 6–10 per group. Values represent means ± SE, *P < 0.05 vs. respective genotypes on control diet (CD), and “#” indicates a significant difference between Western diet groups at a #P > 0.05 calculated using one-way ANOVA.

To determine if in vivo incretin secretion is altered in response to an oral glucose challenge, we measured incretin levels in the plasma from mice taken 15 min into an O-GTT at the end of the study. We observed no change in GLP-1 levels (Fig. 6C). In agreement with mRNA expression levels, plasma GIP levels were mildly elevated in the WD-fed FFA3^fl/fl group, but there was high variability between samples within the group (Fig. 6D).

Collectively, these data reveal that while modest changes in gene expression of major postprandial hormones occurs in both the small and large intestine in response to WD feeding, only Gip expression significantly differed in the absence of intestinal FFA3. In agreement with this, plasma GLP-1 levels were unaltered while plasma GIP levels trended higher in the WD-fed FFA3^fl/fl group. While causality cannot be determined from these studies, these data indicate that alterations in adiposity, fasting glucose levels, and inflammation may be driven in part by changes in circulating GIP levels mediated by FFA3 in response to chronic intake of Western diet (38).

Intestine-Specific Deletion of FFA3 Alters the Gut Microbiome

The gut microbiome is major source of SCFAs, the endogenous ligands of FFA3, and dietary changes are known to influence its composition. To understand how changes in the gut microbiota influence phenotypic differences in the absence of intestinal FFA3, 16S sequencing was performed on fecal samples from all experimental mouse groups. As expected, the composition of the gut microbiome, as measured in terms of β diversity, was significantly different between the CD and WD groups irrespective of FFA3 expression (Fig. 7A). The microbiome composition between the CD-fed FFA3^fl/fl group and Vil-FFA3 group were not distinct, however, significant changes occurred when fed with WD. Similarly, the Shannon Index, a measurement of alpha-diversity, did not differ between the CD-fed groups, but changed significantly between the WD-fed FFA3^fl/fl group and Vil-FFA3 group (Fig. 7B). Differences were observed in the abundances of major phyla, including Firmicutes, and the ratio of Firmicutes to Bacteroidetes was increased in the CD-fed Vil-FFA3 group compared with the CD-fed FFA3^fl/fl group while modestly lower in the WD-fed groups (Fig. 7, C and D). This indicates that in the absence of intestinal FFA3, the gut harbors a lower amount of bacteria implicated to negatively affect metabolic homeostasis including Firmicutes. Furthermore, major alterations at both the genus and species level were observed, including a lower abundance of Faecalitalea in the WD-fed Vil-FFA3 group compared with the WD-fed FFA3^fl/fl controls, which was unchanged in the CD-groups (Fig. 7, E and F). The abundance of Bifidobacterium was significantly reduced in both WD-fed groups independent of genotype, while the abundance of Lactobacillus was increased in the CD-fed Vil-FFA3 group compared with CD-fed FFA3^fl/fl controls (Fig. 7, E and G). Interestingly, the abundance of Lactococcus decreased in the CD-fed Vil-FFA3 group and increased in the WD-fed Vil-FFA3 group compared with FFA3^fl/fl controls, further reflecting diet-specific changes (Fig. 7, E and G). Linear discriminant analysis effect size (LefSe) analyses also revealed the genotype versus diet effects, demonstrating that the increased abundance of Lactobacillus was unique for CD-fed Vil-FFA3 mice as compared with uniquely increased abundance of Bifidobacteria, Rikenellaceae, and Turibacter of CD-fed FFA3^fl/fl mice. However, WD-fed Vil-FFA3 mice show uniquely increased abundance of multiple bacteria including Mucispirillum, Defferribacteraceae, Enterococcus, Lactococcus, Dehalobacterium, Coprococcus, Dorea, Ruminicoccus, Peptococcaceae, Peptostreptococcaceae, Butyricoccus, Oscillispria, Bilophila, Desulfovibriaceae, and Enterobacteriaceae, whereas few bacteria such as Lachnospiraceae, Allobaculum, and Erysipeoltrichale were uniquely abundant in WD-fed FFA3^fl/fl mice (Fig. 7H). Furthermore, the abundance of bacterial species including Faecalibacullum rodentium and Muribacullum interstinale was increased, whereas Tunicibacter sanguinis, Olsennella_sp., Porphyromonadaceae bacterium, and Alistipes finegoldii decreased in the CD-fed Vil-FFA3 group compared with their FFA3^fl/fl counterparts (Fig. 7F). The abundance of Porphyromonadaceae bacterium, Alistipes finegoldii, Romboutsia timonensis, Muribacullum interstinale, and Clostidrium scindens was increased, whereas the abundance of Faecalibacullum rodentium and Clostridium aminophilum was reduced in WD-fed Vil-FFA3 compared with FFA3^fl/fl littermates. Collectively, these results reveal that the gut microbial signatures were more strongly influenced by diet rather than genotype. However, the WD-fed Vil-FFA3 group showed a significantly different microbial signature when compared with WD-fed FFA3^fl/fl group, presenting lower abundance of metabolically detrimental bacteria.

Figure 7.

Intestine-specific deletion of FFA3 alters the gut microbiome. β diversity as indicated by principal component analysis (PCA) (A) and α diversity as indicated by Shannon Index (B). Relative abundance of major phyla (C), ratio of Firmicutes to Bacteroidetes (D), relative abundance of major genera (E) and species (F). Relative abundance of select genera (G) and Linear discriminant analysis effect size (LefSe) cladogram (H). Values represent means ± SE, n = 7–9 per group, *P < 0.05, **P < 0.01, and ***P > 0.01 vs. the FFA3^fl/fl control group calculated using one-way ANOVA.

Cecal SCFA Levels Are Altered When FFA3 Is Absent in the Intestine

SCFAs are the endogenous ligands for FFA3 and are predominantly generated in the distal colon before absorption into the intestinal epithelium (39). Therefore, we assessed how diet, in addition to the absence of intestinal FFA3, would affect cecal SCFA levels. Consistent with literature reporting reduced SCFA levels in a state of gut dysbiosis (40, 41), we observed reduced levels of acetate, propionate and butyrate in both WD-fed mouse groups, with no differences between the Vil-FFA3 and FFA3^fl/fl groups (Supplemental Fig. S4, A–C). Interestingly, SCFA levels were elevated in the CD-fed Vil-FFA3 group, indicating a potential compensatory effect in the absence of FFA3 that is abolished in an obese state. These data also suggest that differences in SCFA production are not driving the phenotypic changes observed in the WD-fed Vil-FFA3 mice.

DISCUSSION

FFA3 plays a critical role in metabolism by mediating effects of SCFAs throughout the body through a summation of its tissue-specific effects. FFA3 is highly expressed in the intestinal epithelium; however, its role is unclear. Furthermore, there is conflicting data regarding its function, underscoring the need for tissue-specific models to resolve various discrepancies. In this study, we report the first in vivo characterization of FFA3 in the intestine and reveal novel insights into receptor function based on chronic consumption of different diets. We found that WD-fed Vil-FFA3 mice were largely protected from the development of obesity and its secondary effects, thereby revealing novel contributions of intestinal FFA3 to the development of obesity and in the regulation of whole body metabolism and intestinal inflammation in the obesogenic state.

Although we found no differences in metabolic parameters in the absence of intestinal FFA3 on a standard chow diet, significant changes were observed when Vil-FFA3 mice were challenged with an obesogenic diet. WD-fed Vil-FFA3 mice were partially protected from the development of obesity and exhibited significantly decreased fat mass and decreased hypertrophy of adipocytes when compared with the WD-fed FFA3^fl/fl controls. Global ablation of FFA3 has yielded conflicting results in previous work. Although some studies using global knockout mouse models for FFA3 reported an increase in adiposity in the absence of the receptor (42, 43), other studies show that, in accordance with our results, ablation of global FFA3 results in protection from gain in body weight (15, 44). These conflicting data support the idea that FFA3 has numerous, tissue-specific roles that act synchronously to regulate body weight. Our work reveals that intestinal FFA3 is a crucial contributor to this process, and that loss of intestinal FFA3 disrupts the body’s ability to appropriately mediate the metabolic consequences of an obesogenic diet.

There were no overt changes in glucose metabolism in the WD-fed Vil-FFA3 group, however, fasting glucose levels and insulin levels trended lower. Although it is possible that intestinal FFA3 may contribute directly to the regulation of glucose homeostasis in response to obesity, these modest trends could also be secondary effects driven by significant changes in fat mass. FFA3 is known to alter glucose-stimulated insulin secretion (GSIS) within the pancreatic beta cell (11, 45, 46), and it is possible that its role in gut-pancreas axis via mediation of incretin secretion could also be involved in glucose regulation. Evidence for alterations in GSIS could be seen in the lowering of insulin levels in response to an OGTT in the WD-fed Vil-FFA3 mice, but these effects were modest and highly variable within each mouse group. Nonetheless, it is possible that alterations within the gut-pancreas axis, which relies on signaling of gut-produced incretin hormones including GLP-1 and GIP, could be contributing to the effects upon insulin and glucose control (47, 48).

We also observed a drastic reduction in intestinal inflammatory markers in both the upper and lower intestines of WD-fed Vil-FFA3 compared with the WD-fed FFA3^fl/fl mice, including improved morphology and histopathological scoring for inflammation. Interestingly, gut barrier and intestinal transit did not appear to be overtly affected by the absence of intestinal FFA3. These data agree with several previous studies that identified a pro-inflammatory role for FFA3 in the development of intestinal inflammation in IECs through a MEK/ERK 1/2 signaling pathway (18, 49). Moreover, our results support that these changes are mediated by FFA3 specifically by intestinal epithelial cells rather than in circulating immune cells such as macrophages and monocytes, where FFA3 expression is low (50).

FFA3 has been shown to co-localize with EEC markers in the intestine and has been suggested to regulate the release of gut peptide hormones, including GLP-1, PYY, and GIP (12, 15, 16, 37). In this study, we did not find any evidence of altered GLP-1 or PYY levels, which we measured both directly and by probing for changes in physiological functions associated with incretin hormones. Of note, we observed reduced Gip mRNA expression in both the ileum and distal colon of the WD-fed Vil-FFA3 mice, and levels of plasma GIP trended modestly lower. GIP is known to increase considerably in response to chronic intake of diets high in fat and promotes obesogenic effects in the adipose tissue (51–53). In addition, GIP has been found to promote inflammation during obesity (54, 55). Interestingly, mouse knockout studies of GIP, GIPR, or ablation of GIP-secreting K cells all demonstrate protection from diet-induced obesity (56–58). This raises the possibility of FFA3 mediating GIP levels during obesity, an effect which is blunted when the receptor is absent in the intestine. One study observed an opposite effect, and it found that GIP levels increased in the absence of the receptor (37). However, that study was carried out in global KO mice under different experimental conditions. Future studies should further examine the intestine-specific contribution of FFA3 to the release of gut peptide hormones, including GIP, to examine causality.

In accordance with other published literature, we found that the gut microbiome is significantly altered in response to chronic consumption of a Western-style diet (59). Furthermore, the absence of intestinal FFA3 resulted in a significantly different microbial signature, particularly between the WD-fed mouse groups. Intestinal FFA3 deletion favored the reduction of metabolically detrimental bacteria, Firmicutes, over the increased abundance of metabolically beneficial Bacteroidetes, with an increased microbiome diversity as measured by the Shannon Index. The enrichment of Firmicutes with reduced Bacteroidetes and their ratio are common manifestations in the intestines of patients with obesity and diabetes, as well as in rodent models consuming high-fat diets (60–62). Bifidobacteria are the most common probiotics and are known to promote several beneficial effects in the gut and beyond, including metabolic effects, by promoting production of beneficial metabolites like SCFAs (63, 64). However, the abundance of these bacteria was significantly reduced in WD-fed groups, suggesting WD consumption caused a reduction in beneficial bacteria that favor of production of ligands SCFAs, FFA3’s endogenous ligands. Cecal levels of SCFAs were reduced in both WD-fed mouse groups but did not differ between the Vil-FFA3 and FFA3^fl/fl mice, indicating that changes in SCFA levels are not driving the phenotypic differences between the two WD-fed groups.

Other SCFA receptors, including FFA2, are highly expressed in the intestinal epithelium. We found that intestinal knockout of FFA3 did not affect the expression levels of intestinal FFA2. Furthermore, expression levels of Ffar2 were significantly elevated in WD-fed FFA3^fl/fl mice, but this effect was lost in the absence of intestinal FFA3. These data suggest FFA2 does not compensate for the loss of intestinal FFA3, and instead raises the possibility that the expression of FFA2 and FFA3 may be influenced by one another (65).

We found that intestine-specific ablation of FFA3 in mice fed an obesogenic diet resulted in resistance to the development of obesity, however, the underlying physiological contributions to this phenotype remain unclear. No major differences in core metabolic processes, including food intake, energy expenditure, and RER were observed in the WD-fed Vil-FFA3 mice. Similarly, no overt differences were observed in intestinal functionality, including GI transport time and intestinal permeability. Although intestinal inflammation was significantly reduced in the WD-fed Vil-FFA3 mice, this could be a downstream effect of reduced obesity. Nevertheless, an immunogenic role of intestinal FFA3 in IECs has been previously described, and loss of the receptor may impair the development of systemic inflammatory responses that mediate changes in adipose tissue function (18, 49, 50). We observed significantly smaller adipocytes and reduced adipocyte hypertrophy in the WD-fed Vil-FFA3 mice, which could be due to alterations in immune cell-mediated cross talk between the intestine and adipose tissues during obesity. The relationship between inflammation, obesity, and intestinal FFA3 is a critical avenue to be explored. Likewise, it is important to consider the impact of other physiological processes not measured in this study. Impaired intestinal nutrient absorption in the WD-fed Vil-FFA3 mice may be a potential contributor to the lack of weight gain. Differences in lipid metabolism, including alterations in lipolysis and lipogenesis, as well as alterations in plasma metabolites may also contribute reduced adiposity.

A limitation of this study is that the receptor was characterized under conditions in which its endogenous ligands, SCFAs, were low. The effects we observed in this study underscore the importance of FFA3 in an obesogenic state, but that may not be the best condition to assess the mechanisms driving phenotypic differences. Future studies may involve characterization of the Vil-FFA3 mice in response to a diet high in fermentable fiber, which may reveal further roles of the receptor in the intestine. In addition, mouse organoids may be utilized for in vitro studies to elucidate the cellular mechanisms mediated by intestinal FFA3, including incretin secretion from EECs and inflammatory responses in IECs.

FFA3 has been identified as a key therapeutic target for metabolic disorders based on its ability to mediate the ubiquitous effects of SCFAs and has important implications in health and disease (66). Here, we present the first in vivo characterization of FFA3 exclusively within the intestine and identify a novel role of intestinal FFA3 in mediating the obesogenic consequences of a Western diet, which will be useful in designing pharmacological agents that target FFA3. Moreover, these data support an intestine-specific role of FFA3 in the regulation of whole body metabolic homeostasis and in the development of adiposity.

SUPPLEMENTAL DATA

Supplemental Tables S1–S3 and Supplemental Figs. S1–S4: https://doi.org/10.6084/m9.figshare.18792083.v2.

GRANTS

K.R.L. is supported by the John and Kathy Solaro Graduate Fellowship through the Physiology and Biophysics Department at the University of Illinois College of Medicine. B.T.L. is supported by National Institutes of Health under Award Number R01DK104927 and P30DK020595; and Department of Veterans’ Affairs, Veterans Health Administration, Office of Research and Development, and VA merit (Grant No. 1I01BX003382). H.Y. is supported by funding from National Institutes of Health Grants R21AG072379, RF1AG071762, and R56AG064075, and Department of Defense—W81XWH-18-PRARP-NIRA.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.R.L., M.P., H.Y., and B.T.L. conceived and designed research; K.R.L., C.N., S.G., M.P., K.X., B.W., S.M., S.J., and J.L.Z. performed experiments; K.R.L., S.M., and S.J. analyzed data; K.R.L. interpreted results of experiments; K.R.L., S.M., and S.J. prepared figures; K.R.L. drafted manuscript; K.R.L., C.N., M.P., B.W., J.L.Z., H.Y., and B.T.L. edited and revised manuscript; K.R.L., C.N., S.G., M.P., K.X., B.W., S.M., S.J., J.L.Z., H.Y., and B.T.L. approved final version of manuscript.