Mice and diet.

All mouse studies were performed in accordance with the guidelines and under approval of the Danish Animal Research authorities issued by the Danish Committee for Animal Research. Transgenic reporter GPR41-RFP mice were kindly provided by Professor Stefan Offermanns and maintained at the Panum Institute, University of Copenhagen. These mice express mouse RFP under the control of the GPR41 promoter and have been characterized before (41). Two experiments were performed in separate cohorts of adult male and female mice, respectively. In all the experiments, mice were provided with food and water ad libitum and maintained on a 12-h light/dark cycle.

In the first experiment, single-housed male GPR41-RFP mice were randomly assigned to one of four diets: 1) chow (standard chow diet containing 4.6% fiber derived from soy hulls, 14 kcal% fat; Altromin 1314, Brogaarden, Denmark); 2) HFD (HFD including 0% fiber, 60 kcal% fat; D12492, Research Diets; New Brunswick, NJ); 3) HFD-Cell [HFD supplemented with 10% (wt/wt) cellulose (BW 200, Sigma-Aldrich, St. Louis, MO)]; and 4) HFD-Flax [HFD supplemented with 10% (wt/wt) flaxseed fibers (Biogin Biochemicals Co. Ltd, China)]. The mice were maintained on their respective diets for 12 wk, and their body weight and fat mass were measured weekly. Mice were transferred to TSE metabolic cages after 8 wk, and at the end of the experiment, cecal contents were collected and snap frozen.

In the second experiment, group-housed female GPR41-RFP mice were randomly assigned to the four diets. The mice were maintained on their respective diets for 12 wk, and their body weight and fat mass were recorded weekly. Ileum and colon tissues were harvested, and single-cell suspensions were isolated for flow cytometry-assisted cell sorting to obtain transcriptomic profiles in RFP-expressing enteroendocrine cells.

Genomic DNA purification, 16S rRNA gene amplification and sequences analysis.

Genomic DNA was extracted from approximately 60 mg of cecal contents collected from single-housed male GPR41-RFP mice using the repeated bead-beating method as described previously (51). The V4 region of the 16S rRNA gene was amplified using 515F and 806R primers designed for dual indexing (34). Each sample was amplified in triplicate in a reaction volume of 25 μl containing 10 μl of Five Prime Hot Master Mix (5 PRIME GmbH, Germany), 0.2 μM of each primer, and 20 ng of genomic DNA. PCR was carried out under the following conditions: initial denaturation for 3 min at 94°C, followed by 25 cycles of denaturation for 45 s at 94°C, annealing for 60 s at 50°C, elongation for 90 s at 72°C, and a final elongation step for 10 min at 72°C and further treated as reported previously (19). The pooled products were sequenced in a pair-end mode (2 × 250 base pair reads) using the MiSeq v2 Reagent Kit (Illumina) and Illumina MiSeq instrument. The forward and reverse reads from the pair-end sequencing were joined by benefiting from the long overlap between both reads using in-house codes. Identical bases in the overlap sequence increased the assurance accuracy of the sequencing, and therefore we assigned the highest possible quality score for those matching bases. The FASTX-Toolkit was used to filter out low-quality reads and reads with a quality Phred score over 20 in at least 98% of their sequences passed through the filter.

Quality filtered Illumina reads were analyzed using the QIIME (Quantitative Insights Into Microbial Ecology) software package (version 1.9.0) as described previously (12). Sequences were clustered into operational taxonomic units (OTUs) at a 97% identity threshold using an open-reference OTU picking approach with UCLUST (25) against the Greengenes reference database (21). Representative sequences for the OTUs were Greengenes reference sequences or cluster seeds and were taxonomically assigned using the Greengenes taxonomy and the Ribosomal Database Project Classifier (60). Representative OTUs were aligned using PyNAST (15) and used to build a phylogenetic tree with FastTree (43), which was used to estimate α- and β-diversity of samples using phylogenetic diversity (27) and unweighted UniFrac (38). Three-dimensional principal coordinates analysis plots were visualized using Emperor (58). Chimeric sequences were identified with ChimeraSlayer (32) and excluded from all downstream analyses. Singletons and sequences with very low abundance (relative abundance <0.005%) were excluded from the analysis. With this approach, a total of 6,754,516 sequences grouped in 698 OTUs were obtained for the 54 samples sequenced, with a median of 125,083 sequences. All samples were rarified to 80,325 sequences to correct for differences in sequencing depth when performing diversity analyses.

Extraction and quantification of SCFAs and organic acids.

SCFAs and organic acids in the cecum were measured by gas chromatography. For the extractions, 70–280 mg of frozen cecal content was transferred to a glass tube (16 × 125 mm) fitted with a screw cap and 100 μl of stock solution of internal standard (1 M [1-¹³C]acetate, 0.2 M [²H₆]propionate and [¹³C₄]butyrate, 0.5 M [¹³C]lactate, and 40 mM [¹³C₄]succinic acid) was added. Prior to extraction, samples were freeze-dried at −50°C for 3 h (yield 20–98 mg dry wt). The extraction and quantification of SCFAs and organic acids were performed as previously reported (50). The m/z ratios of monitored ions were as follows: 117 (acetic acid), 131 (propionic acid), 145 (butyric acid), 261 (lactic acid), 289 (succinic acid), 121 ([²H₂]- and [1-¹³C]acetate), 136 ([²H₅]propionate), 149 ([¹³C₄]butyrate), 264 ([¹³C]lactate), and 293 ([¹³C₄]succinic acid).

Metabolic cages.

Whole-body metabolism was assessed for 3 days in male GPR41-RFP mice after 8 wk by indirect calorimetry in metabolic cages (LabMaster, TSE Systems, Bad Homburg, Germany). Specifically, oxygen consumption, CO₂ production, energy expenditure, food and water intake, as well as locomotor activity were monitored.

Glucose tolerance.

Oral and intraperitoneal glucose tolerance tests were performed in male GPR41-RFP mice following 10–11 wk of diet feeding. Mice were fasted for 6 h, and fasting glucose was measured from tail blood with the OneTouch Ultra glucometer (LifeScan, Milpitas, CA). For fasting plasma insulin analyses, 80 μl of blood was sampled and measured using the Mesoscale Discovery platform. Subsequently, a bolus of glucose (2 g/kg) was either delivered into the stomach by a gavage needle or injected into the intraperitoneal cavity. Blood glucose was measured at 15, 30, 60, and 120 min.

Single-cell suspension and FACS purification.

Female GPR41-RFP mice from the second cohort were killed by cervical dislocation. The distal 10 cm of the small intestine (ileum) and colon was extracted, inverted, inflated, and digested for 20 min with 0.26 Wünsch units Liberase (Roche) in DMEM (5 mM glucose, 1 mM pyruvate) while being slowly shaken at 37°C. Every 5 min, the tissue was shaken vigorously for 5 s. This procedure was repeated three times. Before sorting, the cells were slowly shaken for a second period of 20 min at 37°C, passed through a 70-µm pore diameter cell strainer, pelleted at 1,500 revolutions/min for 5 min, and resuspended in DMEM 1885 with 10% fetal bovine serum. RFP-expressing GPR41-positive (GPR41-RFP_pos) cells were purified by FACS using a MoFlo Astros (Beckman Coulter) and stored at −80°C. Total cellular RNA was extracted using NucleoSpin RNA XS kits (Macherey-Nagel) as previously reported (41). RNA concentration and quality were evaluated using capillary electrophoresis on a 2100 Bioanalyzer with RNA 6000 Pico Kit (Agilent Technologies, Santa Clara, CA).

Microarray.

Total RNA (500 pg) extracted from GPR41-RFP_pos cells of the ileum and colon was used to generate amplified and biotinylated sense transcript cDNA from the entire expressed transcriptome, according to the Nugen Technologies (San Carlos, CA) Ovation Pico WTA System V2 (M01224v2) and Encore Biotine Module (M01111v5). GeneChip ST Arrays (GeneChip Mouse Gene 2.0 ST Array) were hybridized for 16 h in a 45°C incubator and rotated at 60 revolutions/min. The arrays were then washed and stained using the Fluidics Station 450 and finally scanned using the GeneChip Scanner 3000 7G according to the GeneChip Expression Wash, Stain, and Scan Manual (PN 702731 rev. 3, Affymetrix, Santa Clara, CA).

The whole-transcript level of the mice genome was measured by MoGene 2.0 ST chips (Affymetrix). The probe set summarization and normalization was done with Affymetrix expression console software. All the downstream analysis was done in R version 3.1.3 software environment (45). The probe sets were annotated to ENSEMBL gene reference using MoGene 2.0 ST probe set mapping provided by Affymetrix MoGene20 annotation data R-package (39). The differential expression of the genes was assessed by the robust method of limma (linear models for microarray and RNA-seq data) R-package (49). Gene ontology (GO) grouping and enrichment analysis was performed by ClusterProfiler R-package (63) and using biological processes database from genome wide annotation for mouse R-package (org.Mm.eg.db) (17). All Pvalues were corrected using the Benjamini-Hochberg method (4). Venn diagrams were plotted using VennDiagram R-package, and heatmaps were plotted using ggplot2 R-package (61).

Statistics.

The physiological data collected over time (such as body weight and fat mass) were analyzed using two-way repeated measures ANOVA with Holm-Sidak’s post hoc analysis to determine pairwise significance between the groups. The significance between the four groups of mice was determined using one-way ANOVA followed by Holm-Sidak’s multiple comparisons test. Significance was established at P < 0.05.

Flaxseed fiber fermentation alters the gut microbiota composition and SCFA production.

Soluble dietary fibers are known to modulate the composition of the gut microbiota and are fermented to SCFAs that affect host metabolism (31). To assess how flaxseed fibers affect the gut microbiota and the SCFA production, we fed mice HFDs enriched with either flaxseed (HFD-Flax) or nonfermentable fiber cellulose (HFD-Cell) and compared them with mice fed an HFD only or chow for 12 wk. We characterized the cecal microbiota through 16S rRNA gene profiling and observed that the phylogenetic diversity was reduced in all three HFD-fed groups compared with mice on chow diet (Fig. 1A) but did not differ between the three HFD-fed groups (Fig. 1A). In contrast, β-diversity based on unweighted UniFrac analysis (sensitive to the phylogenetic relatedness of taxa) revealed that flaxseed fibers profoundly altered microbiota composition compared with cellulose (Fig. 1B). At the highest taxonomic level (phylum), we observed that an HFD reduced the abundance of Bacteroidetes and increased the abundance of Firmicutes (Fig. 1, C and D), the two major phyla identified in the mouse and human gut (3). Supplementation of either cellulose or flaxseed fibers reduced Firmicutes and increased Bacteroidetes compared with HFD (Fig. 1C). Interestingly, supplementation of flaxseed fibers specifically increased the proportions of the lower abundant phyla, in particular Actinobacteria and Verrucomicrobia, compared with cellulose (Fig. 1C), suggesting specific effects of fermentable dietary fibers.

Next, we identified discriminative bacterial genera that may account for phyla differences. Specifically, we found that the abundances of Prevotella and S24_7 (both belonging to Bacteroidetes) were significantly lower in all three HFD-fed groups compared with the chow-fed mice (Fig. 1, D and E). Consistent with higher abundance of Firmicutes, HFD-fed mice had higher abundance of genera Allobaculum (Fig. 1F) and Lactobacillus (Fig. 1G), but unidentified taxa from order Clostridiales were significantly reduced compared with chow (Fig. 1H). This is in contrast to previous studies that suggest reduction of Allobaculum in diet-induced obesity (DIO) mouse models (44, 46). The addition of cellulose to an HFD restored levels of both Allobaculum and Lactobacillus similar to chow (Fig. 1, Fand G). Flaxseed fiber supplementation increased  Lactobacillus  compared with cellulose (Fig. 1G), which is previously associated with fermentation of dietary fibers in the gut (1). The levels of unidentified species from order Clostridiales increased following cellulose or flaxseed fiber supplementation compared with HFD (Fig. 1H). In addition, flaxseed fiber supplementation increased abundance of Bifidobacterium  compared with HFD and HFD-Cell (Fig. 1I) and Akkermansiacompared with HFD (Fig. 1J).

As expected, mice fed an HFD alone had reduced levels of total SCFAs in the cecal contents (Fig. 2A), as well as lower levels of individual SCFAs and the organic acids lactate and succinate (Fig. 2, B–F) compared with mice on chow diet. In contrast to cellulose, flaxseed fiber supplementation restored the level of butyrate to the level observed in the chow-fed mice (Fig. 2D). Similarly, flaxseed fibers also normalized the level of organic acid lactate similar to the chow group (Fig. 2E). Succinate followed a similar pattern, with higher levels in flaxseed fiber-supplemented mice compared with HFD and HFD-Cell groups (Fig. 2, E and F).

Taken together, flaxseed fibers are fermentable dietary fibers that remodel the gut microbiome with specific effects on potentially beneficial genera such as Bifidobacterium and Akkermansia in the cecum, which is associated with increased total SCFAs in cecal contents, and fully restored the HFD suppression of butyrate and lactate.

Flaxseed fiber supplementation alters transcriptome of GPR41_poscells.

Based on the observation that flaxseed fibers promoted SCFA production, we next evaluated its effects on the molecular functions of enteroendocrine cells. To this end, we collected an FACS-purified GPR41-RFP-expressing enteroendocrine cell population (referred to as GPR41-RFP_pos) from the ileum and colon of all four groups of mice and subjected them to global transcriptomic analysis.

Principal component analysis of gene expression profiles demonstrated that cellulose or flaxseed fiber supplementation to an HFD did not result in separate transcriptomic profiles in GPR41_pos cells from the ileum (Fig. 3A) and altered the expression of <100 genes (P_adj < 0.05) compared with HFD (Fig. 3B). In contrast, we observed separate clustering of colonic GPR41_poscells from mice fed chow or HFD-Flax than mice fed HFD or HFD-Cell, where the profile of HFD-Flax was similar to chow-fed mice (Fig. 3, C and D). GO enrichment analysis revealed that 62 biological functions related to the cellular processes, such as “transcription,” “mRNA transport,” “protein transport,” “autophagy,” “cell cycle,” and “intracellular vesicular transport,” were underrepresented in GPR41-RFP_pos cells from the colons of mice on an HFD compared with chow (Fig. 4). Flaxseed fiber supplementation to the diet could restore the expression of nine biological processes, including mRNA transport, protein transport, and cell cycle, compared with HFD in GPR41-RFP_pos cells from the colon (Fig. 4). No biological processes were significantly altered in HFD-Cell versus HFD comparison.

We observed that 999 differentially expressed genes were regulated by both chow and flaxseed fiber supplementation compared with HFD in GPR41-RFP_pos cells from the colon (Fig. 3D). Most of these genes (992 out of 999) were altered in the same direction (Supplementary Table S1; Supplemental Material for this article is available online at the Journal website). The common genes regulated by both chow and flaxseed fibers were annotated to top GO functions related to signaling pathways. Genes that were only found in chow versus HFD comparison (n = 2,281) were mostly upregulated (Supplementary Table S1) and were associated with GO functions related to developmental processes. In contrast, genes specific for HFD-Flax versus HFD comparison (n = 573) were annotated to GO functions, such as response to reactive oxygen species and negative regulation of translation. Taken together, our findings demonstrate that fermentable fibers in the diet affect the enteroendocrine cell transcription in the colon and that supplementation of flaxseed fibers can replicate some of the effects observed with chow diet containing several fiber types.

Flaxseed fibers protect against obesity and improve glucose tolerance in DIO mice.

Next, we assessed if fermentability of flaxseed fibers is associated with protection against DIO. We observed that dietary treatments had a significant effect on body weight (diet: P < 0.0001; time: P < 0.0003; interaction: P < 0.0001) and fat mass (diet: P < 0.0001; time: P < 0.0001; interaction: P < 0.0001) of mice over time. Mice that received an HFD supplemented with flaxseed fibers had lower body weight gain (Fig. 5A) and fat mass (Fig. 5B) compared with mice fed only an HFD starting from week 2. Fat mass was significantly lower in the HFD-Flax group compared with the HFD and HFD-Cell groups starting from week 2 (Fig. 5B), indicating that the fermentable fibers has a more pronounced effect in lowering fat mass compared with the nonfermentable fibers. Lean mass did not differ among the HFD-fed groups but was significantly lower in all three groups compared with chow-fed mice (Fig. 5C).

As expected, food intake (measured in the metabolic cages) and the respiratory exchange ratio did not change among the HFD-fed groups of mice but was significantly lower in all three HFD-fed groups of mice compared with chow-fed mice (Fig. 5, D and E), in accordance with the fact that mice used the energy-dense dietary fat as their main energy source. Energy expenditure was reduced in the HFD and HFD-Cell groups compared with the chow group (Fig. 5, Fand G). Flaxseed fiber supplementation increased the energy expenditure compared with the HFD group in both the dark and light periods (Fig. 5, F and G) and with the HFD-Cell group in the dark period (Fig. 5G). The increased energy expenditure is in accordance with the reduced fat mass gain observed in the flaxseed fiber-supplemented group (Fig. 5A). There was no significant difference in the activity as determined by beam breaks per minute between the different groups (Fig. 5H), suggesting that the flaxseed fiber-supplemented mice are protected against DIO, primarily by reverting the decrease in energy expenditure induced by an HFD.

The increase in fasting glucose induced by an HFD was completely restored to normal levels by supplementation of flaxseed fibers (Fig. 5I). Similarly, fasting insulin levels were significantly lower in the HFD-Flax group compared with the HFD group (Fig. 5J), and flaxseed fibers also prevented the detrimental effect of an HFD on glucose tolerance following an oral glucose tolerance test (Fig. 5K). Interestingly, insulin levels in response to oral glucose were the highest in the flaxseed fiber-supplemented group (Fig. 5L), which might suggest an incretin response. We also observed partial restoration of intraperitoneal glucose tolerance (Fig. 5M), which likely reflects partial prevention of body weight gain by the flaxseed fibers (Fig. 5, A and B). However, there was no significant difference between cellulose and flaxseed supplementation on glucose tolerance in response to both oral and intraperitoneal glucose dose, indicating that viscosity of fibers might also be important. Taken together, flaxseed supplementation improves both oral glucose tolerance, potentially because of profound effects on transcriptome of enteroendocrine cells, as well as intraperitoneal glucose tolerance, potentially by reducing weight gain.