Fermentable Dietary Fiber Promotes Helminth Infection and Exacerbates Host Inflammatory Responses

This article is featured in In This Issue, p.2865

Lack of dietary fiber, as found in typical Western-style diets in developed countries, can lead to increased incidence of chronic inflammatory diseases, such as Crohn disease and metabolic syndrome (1). Often, chronic intestinal inflammation leads to increased susceptibility to infections as a result of compromised mucosal barrier function and integrity (2–4). Diets rich in soluble fiber, such as inulin (a fermentable fructosaccharide polymer), may therefore alleviate intestinal inflammation through direct modulation of innate immune cell activity (5, 6) and/or by stimulating a healthier gut microbiota composition. This selective propagation of beneficial bacteria (such as Bifidobacterium spp.) is associated with increased production of metabolites of microbial origin, such as short-chain fatty acids (SCFA), which can enhance gut barrier function, and induce tolerogenic dendritic cells (DCs) and T cells (7, 8). Experimental animal models have shown that inulin-induced microbiota compositional changes can reduce the prevalence of obesity and chronic inflammation, subsequently improving colonic health in mice (9, 10) as well as improving resistance to gastrointestinal pathogens such as Salmonella and Giardia (11, 12).

Intestinal parasitic worms (helminths) are highly prevalent in animals and humans, particularly in developing countries. They may cause chronic infections leading to substantial morbidity, but paradoxically, they may also confer health benefits in some contexts. Infection with gastrointestinal helminths has been shown to alter the host microbiota composition and SCFA production, subsequently modulating the host immune response toward a modified Th2/regulatory state and protecting against allergic inflammation (13). Several studies using different experimental animal models have reported the efficacy of intestinal helminths to limit the severity of colitis (14, 15) and type 1 diabetes (16, 17). Consequently the use of deliberate helminth infection (e.g., with the porcine whipworm T. suis) or treatment with helminth-derived products has gained interest as a novel therapy for human inflammatory bowel diseases. Some clinical studies have observed a reduction in Crohn disease activity and an improvement in clinical symptoms of ulcerative colitis as a result of T. suis treatment (18, 19); however, other trials have failed to show an effect (20).

The murine whipworm, T. muris, is a well-characterized model of intestinal helminth infection. In C57BL/6 mice, experimental infections can be manipulated to induce either chronic infection characterized by Th1-driven, colitis-like immunopathology or acute infection characterized by a Th2-driven worm expulsion followed by expression of wound-healing and regulatory mechanisms (21). As well as modulating immune responses, gut microbial composition is also affected by whipworm infection, with chronic T. muris infection shown to reduce bacterial diversity, while at the same time increasing Lactobacillus abundance (22, 23).

Whereas modulation of host immune responses and the gut microbiome by either dietary fiber or helminth infections is well established (24, 25), the interactions between diet and host immune function during concurrent helminth infection are not well understood. Recently, we reported that dietary inulin fibers enhanced the acquisition of local Th2 and mucosal barrier–related immune responses during establishment of T. suis in the porcine host (26). In the current work, we used T. muris infection to explore the ability of dietary inulin to modulate host mucosal immune responses and microbiota composition during helminth infection regimes that induce either Th1- or Th2-driven immunopathology and therefore chronic or acute infection, respectively. Collectively, our data show that inulin, although promoting a healthy gut phenotype in uninfected mice, exacerbates inflammatory responses induced by T. muris infection. Our findings highlight the dynamic interaction between diet and host immune responses elicited by pathogens and demonstrate that the beneficial properties of dietary inulin are context dependent, thus having clear implications for the application of inulin fibers as health-promoting dietary components.

All experimentation was conducted in line with the Danish Animal Experimentation Inspectorate (license number 2015-15-0201-00760) and approved by the Experimental Animal Unit, University of Copenhagen according to The Federation of European Laboratory Animal Science Associations guidelines and recommendations.

T. muris (strain E) was maintained in immune-suppressed C57BL/6 mice (Envigo). Briefly, mice were immune suppressed with dexamethasone supplied in drinking water (1 mg/l; Sigma-Aldrich), then infected with 300 T. muris eggs by oral gavage and maintained for 6 wk. T. muris eggs were isolated from feces and embryonated for 8 wk in distilled water at 21°C. To obtain T. muris excretory/secretory (E/S) product, mice were sacrificed at 8 wk postinfection (p.i.). The cecum and colon were excised, opened longitudinally, and rinsed with sterile Dulbecco’s PBS (Sigma-Aldrich). Adult worms were isolated under a stereo microscope using forceps, washed three to five times in warm, sterile PBS, then incubated for 3 d in sterile RPMI 1640 medium (Life Technologies) supplemented with 500 U/ml penicillin and 500 μg/ml streptomycin at 37°C and 5% CO₂. Culture medium was removed each day and stored at −80°C, and replaced with fresh medium. Pooled culture medium containing the E/S product was concentrated and subsequently dialyzed to PBS by a series of centrifugation steps using Amicon Ultracel Centrifugal Filters (molecular mass cut-off, 10 kDa; Sigma-Aldrich). The resulting E/S product was sterile-filtered (0.22 μm), and protein content was measured by BCA assay (Thermo Fisher Scientific).

Specific pathogen–free, 6–7-wk-old C57BL/6JOlaHsd female mice (Envigo) were housed in individually ventilated cages in groups of five to eight mice. On arrival, mice received ad libitum either a purified control diet (13 kJ% fat; ssniff Spezialdiäten) or the same diet containing 10% long-chain inulin (OraftiHP; Beneo) or an equivalent amount of cellulose in place of corn starch, with free access to water (Supplemental Table I). After 2 wk of diet acclimatization, mice were infected with either 20 or 300 eggs to obtain low or high infections, respectively, or served as uninfected controls (day 0). Egg doses were prepared in tap water, and mice were inoculated by oral gavage. Body weight and welfare were continuously monitored. No differences in weight gain between groups were observed in any experiment. Fresh feces were collected from all mice individually on arrival (day −14), day of infection (day 0), and day 21 or 35 p.i. Fecal samples were cooled immediately on ice upon collection and stored at −80°C. Mice were sacrificed by cervical dislocation at either day 21 or day 35 p.i. Luminal contents were collected from the cecum, cooled immediately on ice, and stored at −80°C, for quantification of SCFAs. Where mentioned, 7-wk-old BALB/c female mice (Envigo) were treated as described above but with 1 wk diet adaptation and sacrificed at day 21 p.i.

At day 21 or 35 p.i., the cecum and proximal colon were opened longitudinally to isolate and enumerate worms as detailed above. Worms (n = 2–19 worms collected per mouse) were rinsed with PBS, and length measurements were taken using a Leica DMRB microscope plus DFC480 camera and Leica Application Suite analysis software version 4.7 (Leica Microsystems). For DNA extraction and microbiota sequencing [protocol modified from White et al. (27)], intact adult worms were surface-sterilized using 1% sodium hypochlorite for 2 min and washed several times with sterile water.

Seven-week-old C57BL/6JOlaHsd female mice were treated with either 500 μg of purified anti-mouse IFN-γ Ab (clone XMG1.2; BioLegend) or rat IgG1κ isotype control (clone RTK2071) by i.p. injection every 5 d starting from day 0 (T. muris infection) until day 21 p.i.

Fresh full-thickness proximal colon tissue samples (adjacent to the cecum) were excised at day 21 p.i. and stored in 4% paraformaldehyde. All samples were paraffin-embedded, sectioned, and mounted on glass slides, followed by staining with periodic-acid Schiff for goblet cell enumeration. To visualize mast cells, additional paraffin-embedded sections were de-waxed, and Ag retrieval was performed with citrate buffer. Tissue sections were incubated with rat anti-mouse primary monoclonal mast cell protease-1 (MCPT-1) Ab (1:100, clone RF6.1; Thermo Fisher Scientific), followed by secondary staining with biotinylated rabbit anti-rat IgG (Abcam). All slides were blinded for histological analysis and enumeration of cells.

T. muris E/S–specific Abs were measured from whole blood serum by ELISA, using biotin-conjugated rat anti-mouse IgG2a (clone R19-5; BD Biosciences), followed by anti-mouse IgG conjugated to HRP (Bio-Rad Laboratories). ELISA was performed as described by Dige et al. (28). Serum IgE was also measured using purified rat anti-mouse IgE Ab (R35-92; BD Biosciences) with serum samples diluted 1:20. Absorbance was measured at an OD of 450 nm with a Multiskan FC plate reader (Waltham).

Mesenteric lymph nodes (MLN) were dissected, stripped of surrounding adipose tissue, disrupted, and filtered through a 70-μM cell strainer to achieve a single-cell suspension. Cells were washed in RPMI 1640 medium (Life Technologies) supplemented with 10% FCS (Sigma-Aldrich) and 100 U/ml penicillin plus 100 μg/ml streptomycin (complete medium) and manually counted using a hemocytometer. At day 35 p.i., MLNs were digested with RPMI 1640 supplemented with collagenase IV (0.5 mg/ml; Sigma-Aldrich) and DNase I (30 μg/ml; Sigma-Aldrich) for 40 min at 37°C on a shaking incubator. Single-cell suspensions were prepared using a 70-μM cell strainer and counted using a hemocytometer. CD11c⁺ DCs were magnetically isolated from digested MLN cell suspensions using MACS MS Columns for cell separation and CD11c⁺ Microbeads (both from Miltenyi Biotec), according to the manufacturer’s guidelines. Finally, the enriched MLN CD11c⁺ DCs were counted using a hemocytometer for further flow cytometry analysis (see below).

MLN cells were seeded in triplicate into 96-well culture plates at a density of 5 × 10⁶ cells/ml in complete medium and stimulated with 50 μg/ml T. muris E/S Ag, or PBS (control). After 24 h incubation at 37°C, cell-free supernatants were harvested and stored at −20°C for subsequent analyses. Secreted cytokines were measured using a BD Th1/Th2/Th17 Cytometric Bead Array Kit (BD Biosciences) according to the manufacturer’s instructions. Samples were processed on a BD Accuri C6 flow cytometer (BD Biosciences), with data acquired using Accuri CFlow Plus software (Accuri Cytometers).

Cells were surface stained for 20 min on ice with the following Abs: FITC-conjugated hamster anti-mouse TCRβ (clone H57-597; BD Biosciences); PerCP-Cy5.5-conjugated rat anti-CD4 (RM4-5; BD Biosciences); PE-conjugated mouse anti-CD64 (X54-5/7.1; BD Biosciences); allophycocyanin-conjugated rat anti–I-A/I-E (M5/114.15.2; Thermo Fisher Scientific); FITC-conjugated hamster anti-CD103 (2E7; Thermo Fisher Scientific). Intracellular staining was performed using Foxp3 Transcription Factor Staining Buffer Set (eBioscience) according to the manufacturer’s guidelines. Fixed/permeabilized cells were incubated for 30 min on ice with the following Abs: Alexa Fluor 647–conjugated mouse anti-mouse T-bet (4B10; BD Biosciences), PE-conjugated rat anti-mouse GATA3 (TWAJ; Thermo Fisher Scientific), and FITC-conjugated rat anti-mouse Foxp3 (FJK-16s; Thermo Fisher Scientific). Cells were analyzed on a BD Accuri C6 flow cytometer (BD Biosciences), and data were acquired using Accuri CFlow Plus software (Accuri Cytometers).

Cecal tissue was excised, rinsed with PBS, and stored in RNAlater at −20°C. Tissue was mechanically homogenized in QIAzol lysis buffer using a gentleMACS Dissociator (Miltenyi Biotec). RNA was extracted using a miRNeasy Mini Kit (QIAGEN) in accordance to the manufacturer’s guidelines. Total RNA purity and concentration were measured using a NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies), and RNA integrity was measured using an Agilent 2100 Bioanalyzer (Agilent Technologies). cDNA synthesis was performed using the Quantitect Reverse Transcriptase kit (QIAGEN). Quantitative PCR was performed using PerfeCTa SYBR Green Fastmix (Quantabio) using the following program: 95°C for 2 min followed by 40 cycles of 15 s at 95°C and 20 s at 60°C. The primers used were Ido1 (forward, 5′-GACTGCGACAAGGGCTTCTT-3′ and reverse, 5′-TGCAGTGCCTTTTCCAATGC-3′) and Gapdh (forward, 5′-TATGTCGTGGAGTCTACTGGT-3′ and reverse, 5′-GAGTTGTCATATTTCTCGTGG-3′). Fold changes were calculated using ΔΔCT values.

Gene expression microarray analysis was conducted using the GeneChip WT PLUS Reagent Kit (Thermo Fisher Scientific) and Affymetrix mouse Clariom S HT 24-Array Plate pipeline (Eurofins AROS), with array plate processing carried out on a GeneTitan Instrument (Thermo Fisher Scientific). Transcriptome Analysis Console Software (Thermo Fisher Scientific) was used to analyze data, and additional pathway analysis was performed using Gene Set Enrichment Analysis (GSEA) software (Broad Institute). Microarray data are available at Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE136492.

DNA was extracted from feces at either day 21 or day 35 p.i. and from T. muris worms (only at day 35 p.i.) in a randomized order using the Bead-Beat Micro AX Gravity Kit (A&A Biotechnology) according to the manufacturer’s instructions. Prior to extraction, samples were lysed in lysis buffer supplemented with lysozyme (4000 U) and mutanolysin (50 U) and incubated at 50°C for 20 min. The concentration and purity of extracted DNA were determined using a NanoDrop ND-1000 Spectrophotometer and normalized to 10 ng/μl. High-throughput sequencing-based 16S rRNA gene amplicon (V3 region) sequencing was carried out on an Illumina NextSeq platform as previously described (29).

The raw dataset containing pair-ended reads with corresponding quality scores were merged and trimmed using fastq_mergepairs and fastq_filter scripts implemented in the USEARCH pipeline as described previously (29). Purging the dataset from chimeric reads and constructing zero-radius operational taxonomic units (OTU) was conducted using UNOISE. The Greengenes (13.8) 16S rRNA gene collection was used as a reference database. The Quantitative Insight Into Microbial Ecology open source software package (1.7.0, 1.8.0, 1.9.0) was used for subsequent analysis steps (30). Microbiota data were pooled from the two independent experimental studies. The α-diversity measures included the following: observed species (number of zero-radius OTU) and Shannon diversity indices were computed for rarefied OTU tables (10,000 reads per sample) using the α-rarefaction workflow. Differences in α-diversity were determined using a t test–based approach employing the nonparametric (Monte Carlo) method (999 permutations) implemented in the compare α-diversity workflow. Principal coordinates analysis (PCoA) plots were generated with the jackknifed β-diversity workflow based on 10 distance metrics calculated using 10 subsampled OTU tables. The number of sequences taken for each jackknifed subset was set to 85% of the sequence number within the most indigent sample (∼10,000). Community differences (β-diversity) were revealed by weighted and unweighted distance. Permutational multivariate ANOVA (PERMANOVA) was used to evaluate group differences based on weighted and unweighted UniFrac distance matrices. The datasets accompanying this study can be found in the Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra/) under accession number PRJNA614961.

SCFAs were extracted from cecal luminal digesta solutions prepared with 550 μl of 2.2% formic acid (Thermo Fisher Scientific) and Milli-Q water (containing 1 mM 2-ethylbutyrate as internal standard). All samples were homogenized with two 2.4-mm zirconia beads (BioSpec Products) in a Mini-BeadBeater-96 (BioSpec Products) for three 20-s cycles. The pH was confirmed to be within two to three, and samples were subsequently filtered by centrifugation (10,000 × g, 4°C for 10 min) using a 2-ml tube with a 0.45-μm filter (Costar Spin-X, CLS8170; Corning). The filtrate was transferred to glass vials and kept cold until analysis by gas chromatography–flame ionization detection (HP 6890 GC System; Agilent Technologies) with a CP-FFAP CB Column (Agilent Technologies). Helium was used as the carrier gas, and a 3-μl volume of the filtrate was injected with a 5:1 spit ratio. The concentration of each SCFA was derived from a standard curve from experimental standards of each SCFA (Sigma-Aldrich) ranging from 3.91 to 8000 μM, whereas intersample variation was adjusted by use of an internal 2-ethylbutyrate standard (Sigma-Aldrich).

Data were tested for normality using Shapiro–Wilk coefficient correlation test using GraphPad Prism 7 (GraphPad Software) and were log or square root transformed to obtain normality. Normally distributed data were then analyzed by mixed linear model using IBM SPSS Statistics 24 or by Student t test. The mixed linear model included the fixed factors of diet and infection status and random factors of mouse and cage. If data did not follow a normal distribution, analyses were performed by general linear model, which included diet and infection status as fixed factors, or by Mann–Whitney rank test.

We fed mice an inulin-enriched or a control diet prior to inoculation with either 20 (low dose) or 300 (high dose) infective T. muris eggs, as previously described (21). Mice fed the control diet had worm burdens characteristic of susceptibility or resistance elicited by low- or high-dose T. muris. Control-fed, low-dosed mice had approximate burdens of 4.2 ± 0.8 worms (mean ± SEM) and 1.3 ± 0.6 worms at day 21 and 35 p.i., respectively (Fig. 1A). Retention of worms at both time points is indicative of an established T. muris infection (subsequently confirmed by serum T. muris E/S–specific IgG2a, Fig. 1B). Similarly, inulin-fed, low-dosed mice retained worms at both time points; however, inulin resulted in significantly higher infection levels with worm burdens of 8.6 ± 1.1 and 9.5 ± 0.8 worms at day 21 and 35 p.i., respectively, compared with control-fed, low-dosed mice (p < 0.01; Fig. 1A).

High-dosed mice fed the control diet had low worm burdens present at day 21 p.i. (3.1 ± 1.6 worms), with all worms expelled by day 35 p.i. Remarkably, inulin-fed, high-dosed mice retained significantly high burdens at both day 21 (124.7 ± 14.1 worms, p < 0.001) and day 35 p.i (129.1 ± 14.2 worms, p < 0.001) compared with control-fed, high-dosed mice (Fig. 1A). In addition to higher worm burdens, mice fed inulin also had significantly larger worms compared with those from control-fed mice (Fig. 1C).

To determine whether the increased worm burdens in inulin-fed mice were due to the fermentation of inulin in the cecum or solely to the increased fiber content, an additional group of mice received high-dose T. muris infections and were fed a nonfermentable fiber (cellulose)-supplemented control diet. This cellulose control diet matched the crude fiber content present in the inulin-supplemented diet and served as a control for the potential nonfermentable fiber effects on T. muris expulsion (Supplemental Table I). At day 21 p.i., inulin-fed mice retained high worm burdens as observed in previous experiments (169.8 ± 16.2 worms, p < 0.001; Fig. 1D). Yet, cellulose-fed mice efficiently expelled the infection (worm burden of 5.6 ± 1.7), and tended to have lower worm burdens than control-fed mice (35.0 ± 8.8; Fig. 1D). This was mirrored in the serum IgG2a Ab levels with inulin-fed mice having increased IgG2a levels compared with both control- and cellulose-fed mice; however, this was not statistically significant (p > 0.05, Fig. 1E).

The levels of infection were also reflected in the morphology of the proximal colon and the number of mucin-producing goblet cells observed at day 21 p.i. (Fig. 2A, 2B). Uninfected controls and low-dosed mice displayed a normal mucosa with no differences observed in goblet cell numbers or intestinal morphology between the different diets or infection treatments. Control-fed, high-dosed mice, however, had significantly higher numbers of goblet cells per square millimeter (316.8 ± 29.8) compared with uninfected controls (222.7 ± 34.4, p < 0.05), which is indicative of an active host immune-mediated worm expulsion mechanism, resulting in the declining worm burdens observed at day 21 p.i. In contrast, inulin-fed, high-dosed mice had significantly reduced goblet cell number as compared with uninfected controls (128.7 ± 16.1, p < 0.05). Moreover, these mice had substantial inflammatory cell infiltrates of the lamina propria and, in parts, a degenerated epithelial barrier compared with uninfected controls (Fig. 2B, Supplemental Fig. 1). Thus, fermentable fiber in the form of dietary inulin prevented the expulsion of T. muris in normally resistant mice and resulted in significant parasite-induced inflammation.

The total number of MLN cells harvested from each mouse was significantly influenced by infection status, with both low- (p < 0.001) and high-dosed mice (p < 0.001) having greater MLN cell numbers compared with uninfected controls (Fig. 3A). Analysis of T cell phenotypes showed that infection resulted in a higher percentage of TCRβ⁺CD4⁺T-bet⁺ cells in the MLN compared with uninfected controls (p < 0.001 low dose, p < 0.05 high dose; Fig. 3B). Within the infected mice, a significantly higher percentage of TCRβ⁺CD4⁺T-bet⁺ cells was also observed in mice fed inulin compared with the control diet (p < 0.01 low dose, p < 0.001 high dose; Fig. 3B); with the control-fed, high-dosed group having the lowest proportion of T-bet⁺ cells. In contrast, TCRβ⁺CD4⁺GATA3⁺ T cells were mainly increased in the control-fed, high-dosed mice, consistent with a pronounced Th2 response in this group, with lower levels of these cells present in inulin-fed mice in both infected groups (Fig. 3C). No significant difference in Foxp3⁺ T cells was observed, although inulin did tend to reduce the proportion of Foxp3⁺ cells in infected groups only, compared with the respective controls (Fig. 3D). At day 35 p.i., MLN CD103⁺ DC populations were reduced in inulin-fed infected mice compared with controls (p = 0.059; Fig. 3E). CD103⁺ DCs are commonly depleted during persistent inflammation, and moreover, reduced numbers are associated with reduced epithelial responsiveness and delayed T. muris expulsion in Nod2^−/− mice (31). Altogether, these observations suggest that inulin-mediated impaired worm expulsion is not due to increased regulatory immune responses but instead derives from a Th1-skewed immune profile, which is normally associated with chronicity in this mouse strain.

Consistent with this, T. muris E/S–stimulated MLN cells from low-dosed mice (both control- and inulin-fed) secreted higher levels of proinflammatory TNF-α (p < 0.001), IL-6 (p = 0.056), and IFN-γ (p < 0.001) compared with uninfected controls, clearly associating with the presence of a chronic, established T. muris infection (Fig. 3F–H). In contrast, the control-fed, high-dosed mice had baseline proinflammatory cytokine secretion similar to that of uninfected controls but with increased levels of IL-2 (p < 0.001), IL-4 (p < 0.01), and regulatory IL-10 (p < 0.01), confirming the presence of an active Th2-mediated immune response (Fig. 3I–K). However, production of these Th2 cytokines was completely abrogated in inulin-fed, high-dosed mice, which instead produced a Th1 cytokine profile similar to that of low-dosed mice. Interestingly, inulin treatment tended to restrain IL-2 and IL-4 cytokine secretion (and to a lesser extent IL-10) in both low and high-dosed mice, with cytokine levels resembling that of uninfected controls. IL-17A cytokine secretion was significantly influenced by infection, with both low- and high-dosed mice secreting higher levels (p < 0.001, p < 0.05, respectively) compared with uninfected controls. In addition, inulin appeared to increase IL-17A levels further in both infected groups; however, this was not statistically significant (Fig. 3L). Inulin alone had no effect on T. muris E/S–induced MLN cytokine secretion from uninfected control animals.

To gain insight into the mechanisms underlying the inulin-induced worm persistence, we performed microarray analysis on cecal tissue taken at day 21 p.i. from control-fed uninfected mice, inulin-fed uninfected mice, and high-dose T. muris–infected mice fed either inulin or control diet.

Visualization of the treatment groups by principal component analysis highlighted the strong influence of treatment on cecal tissue gene expression. Treatment with either inulin or a high T. muris dose alone induced significant changes to differential gene expression. However, the combined treatment of inulin and helminth infection induced a distinct gene expression profile, indicating that the interaction between treatments further modified the distinct immune responses elicited by each treatment in isolation (Fig. 4A). Inulin or high T. muris infection alone resulted in 224 and 1307 differentially expressed genes, respectively (p < 0.05, fold change >2), compared with control-fed uninfected mice. However, combined treatment of inulin and high T. muris dose drastically altered transcriptional profiles, with a total of 3693 genes differentially expressed compared with control-fed uninfected controls (Fig. 4B). Differential gene expression profiles for each of the treatment groups had limited homology, with the largest portion of shared genes (23.6%) observed between the high-dosed group and the inulin plus high-dosed group.

Dietary inulin in the absence of T. muris infection resulted in the regulation of a variety of immune-related genes, such as Pla2g3, Ccl8, Retnla, and Ccl12, as well as the SCFA-associated, sodium-coupled monocarboxylate transporter Slc5a8, compared with control-fed uninfected controls (Supplemental Fig. 2). Furthermore, genes involved with microbial sensing (Saa1, Duoxa2) were also significantly upregulated. Interestingly, inulin also downregulated several gut barrier function genes, such as small proline-rich protein 1a (Sprr1a), tight-junction claudin 4 (Cldn4), and also inflammatory related vanin-1 (Vnn1). This indicates that inulin contributes to the modulation of mucosal barrier immune and structural function, consistent with in vitro and in vivo mouse studies (32–35).

As expected, high-dosed mice displayed characteristic helminth-induced gene expression profiles, with upregulation of many genes involved in eosinophilia, mastocytosis, and epithelial cell hyperplasia. For instance, mast cell enzyme–related genes Pla2g4c, Cma2, Cpa3, Mcpt-1, -2, -4, and -9, and the IgER gene Fcer1a were significantly upregulated by the presence of T. muris compared with uninfected controls (Supplemental Fig. 2). In addition, helminth infection associated with the upregulation of eosinophil-related Ccr3 and goblet cell–derived angiogenin Ang4 and intelectin Itnl1, both of which have been implicated as key genes involved in helminth expulsion (36). Interestingly, expression of these helminth-induced, mast cell–related genes and eosinophil Rnase2a, were markedly downregulated in inulin-fed, T. muris–infected mice compared with infected mice fed the control diet. Simultaneously, inflammatory related genes (Ifng, Cxcl9, Nos2) and antimicrobial, peptide-related genes (Reg3b, Reg3g) were significantly upregulated (Fig. 4C, Supplemental Fig. 2). Notably, the epithelial cell–related genes Ido1 and Cxcl10, both of which have previously been shown to contribute to chronic T. muris infection, were also significantly upregulated (37, 38). Together, these data confirm the polarization of a Th1 immune environment as a result of inulin supplementation in high-dosed mice, which likely contributed to the persistence of T. muris infection. This observation was explored further by GSEA of microarray data from high-dosed mice fed either the control diet or the inulin diet. GSEA revealed significant upregulation of IL-12–related pathways, including the IL-12, IL-23, and IL-27 pathways, and IFN signaling pathways (p < 0.001, false discovery rate q value <0.001; Fig. 4D). IL-12–related cytokines (IL-12, IL-23, and IL-27) are secreted from activated macrophages and DCs and interact with either naive or activated CD4⁺ T cells to promote inflammatory immune responses (39, 40).

Given the substantial suppression of IgER signaling and mast cell–related genes in infected mice fed inulin, we further examined serum IgE levels and intestinal mast cell numbers. At day 21 p.i., circulating T. muris, E/S–specific IgE levels were significantly elevated for low-dosed mice (p < 0.05) and control-fed, high-dosed mice (p < 0.001) compared with uninfected controls (Fig. 5A). However, an IgE response was completely absent at day 21 p.i. in the high-dosed mice fed inulin. At day 35 p.i., IgE levels were reduced in the high-dosed mice compared with day 21 p.i., which is indicative of a resolution phase after effective worm expulsion. Interestingly, IgE levels were significantly increased in the inulin-fed, high-dosed mice compared with the control-fed, high-dosed group (day 35 p.i.). Similarly, the low-dosed group also had detectable circulating IgE, although this was not significant. Elevated levels of IgE in these two groups at day 35 p.i. could signify the initiation of host effector mechanisms in response to the presence of T. muris (Fig. 5A). Moreover, mast cell numbers (Mcpt1⁺ cells) in the proximal colon were significantly elevated in the control-fed, high-dosed mice at day 21 p.i. (p < 0.001) compared with all other treatment groups (Fig. 5B). The lack of mastocytosis in the inulin-fed, high-dosed mice is consistent with the absence of an active Th2-driven helminth expulsion mechanism.

Unweighted UniFrac distance metrics analysis demonstrated alterations in microbial diversity at day 21 p.i. as a result of diet, with distinct clustering evident between control-fed and inulin-fed mice (PERMANOVA, p < 0.05). Furthermore, there was a striking shift in microbial β-diversity as a result of the interaction between diet and helminth infection in the inulin-fed, high-dosed mice compared with control-fed, high-dosed mice (p < 0.05; Fig. 6A). Similarly, infection and diet also influenced α-diversity, with the highest number of observed OTUs found in control-fed uninfected mice (1370 ± 55, mean ± SD) and the lowest diversity present as a result of high-dose T. muris infection and dietary inulin treatment (431 ± 210, p < 0.05; Fig. 6B).

Dietary inulin beneficially modulated fecal microbial composition in uninfected mice at day 21 p.i., with increased Bacteroidetes (18–28%, p < 0.05) and reduced Firmicutes (75–46%, p < 0.01) relative abundance compared with control-fed uninfected mice (Fig. 6C). Similar effects of dietary inulin were also observed for low-dosed mice compared with control-fed, low-dosed mice. Dietary inulin also significantly increased the relative abundance of Verrucomicrobia in low-dosed (3–8%, p < 0.01) and uninfected mice (0–15%, p < 0.01) compared with controls, with this change attributed to a single OTU corresponding to Akkermansia muciniphila (Fig. 6D). Furthermore, inulin-fed, high-dosed mice displayed a unique composition compared with the other treatment groups, with significant expansion of Proteobacteria, mainly Enterobacteriaceae (p < 0.001; Fig. 6E), accounting for more than 30% of fecal-relative microbial abundance compared with <3% for controls. In addition, inulin-fed, high-dosed mice had significantly diminished Bacteroidetes (29–4%, p < 0.01) and Verrucomicrobia (A. muciniphila) abundance (16–0.2%, p < 0.001) compared with control-fed, high-dosed controls (Fig. 6C).

The influence of dietary inulin on host microbial composition was the result of significant changes in the relative abundance of several bacterial taxa (Supplemental Fig. 3A). For example, in the order Clostridiales (Firmicutes), the relative abundance was higher in control-fed mice compared with inulin-fed controls, with the highest relative abundance observed in control-fed, T. muris–infected mice. This promotion of Clostridiales as a result of T. muris infection has been reported previously (41). Interestingly, dietary inulin appeared to suppress T. muris–induced Clostridiales expansion in both low- and high-dosed mice (both p = 0.06; Supplemental Fig. 3B). Conversely, dietary inulin increased the relative abundance of OTUs of the genus Sutterella (Proteobacteria), the family Coriobacteriaceae, and an OTU most closely related to the species Bifidobacterium pseudolongum (Actinobacteria) (Supplemental Fig. 3C–E), with increases in abundance of these taxa previously associated with dietary fiber intake (42–44). However, T. muris infection (low and high dose) significantly reduced B. pseudolongum OTU relative abundance in inulin-fed mice compared with inulin-fed uninfected controls (p < 0.01 and p < 0.05, respectively) (Supplemental Fig. 3E). Notably, oral administration of B. pseudolongum has been associated with reduced IFN-γ secretion from cervical lymph nodes and contributed to reduced swelling in a chemically induced skin allergy mouse model (45).

Next, we explored whether the significant increase in Proteobacteria in infected mice fed inulin may be attributed to bacteria derived directly from the large numbers of worms present, as T. muris harbor a distinct endogenous microbiota composition that may interact with the host microbiome (27). We found that endogenous T. muris microbiota composition had considerable homology, irrespective of a host diet based on the weighted UniFrac distance matrix (PERMANOVA, p > 0.05; Fig. 7A), with a significant difference observed between T. muris and host cecal microbial composition based on the unweighted UniFrac distance matrix (PERMANOVA, p < 0.01; Fig. 7B). Generally, T. muris were predominantly colonized with Proteobacteria or Firmicutes, with lower abundances of Actinobacteria and Bacteroidetes also present (Fig. 7C). Thus, the T. muris microbiota appears to be remarkably compartmentalized and unperturbed by the significant diet-induced changes in the host microbiota. The observed Proteobacteria bloom in the host cecum therefore appears to derive from expansion of endogenous host bacteria rather than being colonized by worm-derived bacteria.

Because dietary inulin and T. muris infection brought about significant shifts in microbiota composition, we hypothesized that shifts in production of cecal microbial fermentation products may also occur. At day 21 p.i., we found that inulin-fed, low-dosed mice consistently had the highest concentration of the major SCFAs (acetate, p < 0.05; propionate, p < 0.001; and butyrate, p = 0.074) compared with controls (Fig. 8). In contrast, inulin-fed, high-dosed mice had the lowest cecal concentration of acetate (p < 0.01) compared with controls, with a significant reduction in total SCFA (p < 0.05) production present in the cecum. Reduction in SCFA production is typically a consequence of reduced fermentable substrates present in the gut (46); however, in this study, a reduction in total SCFAs was likely a consequence of shifted gut microbial composition, thus affecting overall fermentation-induced SCFA output.

To confirm that the inulin-mediated Th1 polarization was the cause of the persistent worm burdens and not simply a consequence of the chronic infection and accompaying dysbiosis, C57BL/6 mice were repeatedly treated with neutralizing anti–IFN-γ Ab during the course of high-dose T. muris infection.

Neutralization of IFN-γ restored normal helminth expulsion at day 21 p.i. in inulin-fed mice and abrogated both the T. muris E/S–induced IFN-γ production from MLN cells and the marked upregulation of Ido1 in the cecum observed in inulin-fed isotype controls (Fig. 9A–C). Despite this, inulin-fed mice treated with anti–IFN-γ still had higher levels of T. muris E/S–specific IgG2a (p < 0.05) and lower levels of IgE compared with control-fed mice (p < 0.05; Fig. 9D, 9E). Analysis of T cell phenotypes in the MLN showed that anti–IFN-γ treatment significantly reduced TCR⁺CD4⁺T-bet⁺ T cells compared with isotype controls (p < 0.001), restoring the positive GATA3⁺/T-bet⁺ ratio characteristic of the Th2 response induced by high-dose T. muris infection (Fig. 9F, 9G). This suggests that Th2-polarizing signals that are strong enough can overcome the effects of inulin and restore protective immunity. To test this further, we used BALB/c mice, which are considered more resistant and mount faster Th2 responses to T. muris compared with C57BL/6 mice (47). BALB/c mice fed inulin displayed normal worm expulsion after 21 d p.i.; however, they also had higher amounts of specific IgG2a Abs than mice fed a control diet (Fig. 9H, 9I). Thus, the capacity of inulin to effect the development of Th2 immunity appears to be regulated by the strength of intrinsic and extrinsic Th2-polarizing factors.

In this study, we have demonstrated that supplementation with the soluble, fermentable dietary fiber inulin caused significant alteration in prototypical immune responses toward T. muris in typically resistant C57BL/6 mice. We observed polarized Th1 immune responses, as well as suppression of mucosal mast cell–mediated responses, accompanied by a degenerated epithelial barrier in the proximal colon. Furthermore, dietary fiber caused significant expansion of Proteobacteria and simultaneous depletion of beneficial Verrucomicrobia and Bacteroidetes phyla. Thus, whereas inulin is normally recognized for its anti-inflammatory and gut health-promoting effects, in the context of T. muris infection, inulin significantly exacerbated parasite-induced inflammation.

In mice fed a control diet, treatment with either a low or high dose of T. muris eggs resulted in persistent infection or expulsion at day 35 p.i., respectively. However, C57BL/6 mice treated with a high dose of T. muris eggs and dietary inulin exhibited sporadic colitis-like morphology in the proximal colon and a Th1-dominated immune phenotype, resulting in persistence of worms at day 21 and 35 p.i. The lack of efficient expulsion mechanisms was reflected by the absence of mastocytosis in the proximal colon and the downregulation of cecal mast cell gene expression in the inulin-fed, high-dosed, T. muris–infected mice at day 21 p.i. Mast cells, as well as epithelial-derived factors, such as TSLP, IL-25, and IL-33, are critical initiators of host Th2 immune responses during T. muris infection (48, 49). Accordingly, the suppression of mast cells and associated downstream immune responses was likely related to the significant 100-fold increase of Ifng expression in the cecum and the subsequent polarized inflammatory mucosal immune environment. In addition, lack of IL-10 secretion in inulin-fed, high-dosed mice may also have influenced the inability of the host to mount effective type 2 responses and may also have contributed to the observed intestinal pathology and dysbiosis (50, 51). Further downstream, effects of enhanced Ifng expression in inulin-fed, high-dose–infected mice include significant upregulation of Ido1 and Cxcl10 gene expression. The enzyme IDO has previously been reported to influence epithelial cell turnover, with IDO inhibition resulting in T. muris expulsion in susceptible SCID mice (37). Similarly, neutralization of the chemokine CXCL10 in susceptible mouse strains (SCID, AKR) resulted in an elevated epithelial turnover rate and helminth expulsion (38). Thus, our data suggests that IFN-γ–driven defects in epithelial cell proliferation may be potentially important effector mechanisms limiting Th2 immunity and preventing helminth expulsion in this model. Yet, further experimentation is needed to investigate this hypothesis and the role of inulin-mediated IFN-γ responses.

Inulin supplementation without T. muris infection (a total of 6 wk) created highly distinct fecal microbial profiles compared with control-fed mice, with marked expansion of Verrucomicrobia (Akkermansia) and Actinobacteria (Bifidobacteria), as shown previously (10, 34, 52). A. muciniphila is a mucin-degrading bacteria present in the intestinal epithelial mucus layer and is a major producer of propionate (53). Additionally, A. muciniphila regulates gut barrier function through modulation of mucin production (54). Inulin-type fructans are known to promote expansion of A. muciniphila, both of which have been associated with enhancement of gut barrier defenses and protection against development of inflammatory conditions, obesity, and associated metabolic diseases (35, 54, 55). Similarly, we found that T. muris infection (in mice fed a control diet) also increased Akkermansia abundance and increased the Bacteroidetes/Firmicutes ratio, thus suggesting that in isolation T. muris tended to favor the growth of potentially beneficial bacterial taxa, which is consistent with previous observations with T. suis infection in pigs (26). However, despite the seemingly positive changes induced by these factors in isolation, the combination of dietary inulin and high-dose T. muris infection induced a profound expansion of Proteobacteria (Enterobacteriaceae), accompanied by a striking depletion of Akkermansia, which correlated with a reduction in cecal propionate and presence of colitis-like degeneration of the proximal colon epithelial barrier.

Gut inflammation is correlated with increased intestinal oxygenation and dysbiosis (56). Similarly to observations in our study, this altered intestinal environment often facilitates expansion of facultative anaerobes, such as Enterobacteriaceae, and simultaneously inhibits growth of obligate anaerobes, Bacteroidia and Clostridia (57). Furthermore, during active ulcerative colitis, intestinal oxygenation can result from the production of reactive oxygen species, such as hydrogen peroxide, through the action of epithelial cell–derived enzymes, DUOX and DUOXA2 (58). It is interesting to note that dietary inulin in the absence of helminth infection significantly upregulated cecal Duoxa2 expression. The induction of reactive oxygen species by inulin alone may not have resulted in intense intestinal inflammation. Yet, when coupled with the high number of T. muris larvae disrupting the epithelial layer and the associated increased LPS leakage (59), inulin may have primed the intestinal environment, thus intensifying the innate Th1 polarization, leading to an IFN-γ–induced inflammatory response to early helminth infection (60) and subsequently facilitating growth of opportunistic Enterobacteriaceae. Although we cannot state unequivocally whether these events are the cause or consequence of the observed inflammatory intestinal state in inulin-fed, high-dose–infected mice, they do provide an interesting avenue for further investigation into the capacity of dietary inulin to modify intestinal conditions and subsequently alter helminth-induced host immune responses.

Consistent with the hypothesis that inulin-mediated disturbances of the Th1/Th2 balance during immune priming lead to impaired immunity to T. muris, manipulation toward Th2 polarization (either through the use of a more resistant mouse strain or IFN-γ neutralization) resulted in normal worm expulsion and no increase in Ido1 expression in C57BL/6 mice. BALB/c mice typically show greater resistance to T. muris infection and thus mount effective Th2 responses and expel worms faster than C57BL/6 mice (47). This suggests that restricting the inulin-mediated Th1 priming allowed the normal progression of Th2 effector mechanisms required for immunity. Also, BALB/c mice have been shown to harbor a different intestinal microbial composition compared with C57BL/6 mice (61), which may have influenced the ability of inulin to elicit host inflammatory responses during high-dose infection.

It is interesting to note that some associations between dietary inulin and inflammation have been observed previously in other models of acute inflammatory responses. Inulin can exacerbate proinflammatory responses in a DSS-induced colitis mouse model (62) and also worsen the effects of IL-10R deficiency in mice (63). Thus, whereas inulin appears to promote gut health during low-grade inflammation, detrimental effects may be apparent in acute infectious or inflammatory challenges. Adding to this complexity, the effects of dietary inulin may also differ dependent on the experimental animal model. For instance, in contrast to the results in this study, we have previously shown that pigs fed inulin and infected with the porcine whipworm T. suis (26) had enhanced Th2 mucosal immune responses and harbored a more diverse intestinal microbiota composition. It is interesting to note the differing effect of inulin during infection with two phylogenetically similar helminth species. Despite the relatedness, it is well-known that chronic T. muris infection induces inflammation and Th1 activation in mice, whereas small doses of embryonated T. suis eggs have been shown to reduce inflammation in inflammatory bowel disease patients (18, 19). Thus, the differing pathogenicity of each helminth species may play a role in the host-dependent effects observed. Moreover, the form and proportion of crude fiber is also an important consideration when determining the potential beneficial effects of dietary additives. The intake of purified dietary fiber has previously been associated with development of cancer and severe colitis in mice (62, 64). The diet composition reported in Myhill et al. (26) included complex insoluble fiber (derived from whole grains) as well as purified inulin, whereas the current study focused on the use of a highly refined, purified diet. Insoluble fibers, such as cellulose, are known to act as a bulking agent and promote intestinal transit (65), which in the context of intestinal helminth infection could be beneficial for expulsion mechanisms. This is supported by our observation that an insoluble fiber (cellulose)–based diet may accelerate T. muris expulsion, resulting in faster elimination of worm burdens compared with control-fed mice. Conversely, the abundance of soluble fibers (inulin) in an otherwise refined diet with little complex plant fiber could be a key factor in the development of intestinal inflammation and dysbiosis, resulting in the helminth persistence observed in this study. Clearly, the mechanisms underlying this complex diet–pathogen–microbiota interaction need further investigation, given both the intense interest of inulin and related fibers as functional food components and the abundance of refined, low-fiber Western style diets in modern society.

In conclusion, we found that dietary inulin modulated host microbial composition and mucosal immune phenotype as well as enhancing T. muris establishment and growth in C57BL/6 mice. Inulin significantly interacted with high-dose T. muris infection, resulting in the depletion of mucosal mast cell proliferation, which consequently led to abrogated IL-4/IL-13–mediated Th2 immune responses. This coupled with the expansion of Proteobacteria (Enterobacteriaceae) likely intensified polarized Th1 inflammatory responses observed in these mice, overcoming the beneficial modulatory effects of dietary inulin and allowing persistence of T. muris in typically resistant mice. Yet this suppression of Th2 immune response and subsequent increase in inflammatory Th1 in some circumstances may be beneficial. For example, hyperactive Th2 immune responses can result in sensitivity to benign food allergens. Supplementation of pregnant mice with galacto-oligosaccharides and inulin resulted in enhanced regulatory and suppressed Th2 responses in wheat-sensitized offspring (66). Thus, it should be noted that the use of dietary inulin as a potential therapeutic for treating inflammatory intestinal conditions and enhancing gut health is context-dependent, with further investigation needed to elucidate the mechanism responsible for the interaction between inulin and T. muris infection. Taken together, this study clearly demonstrates the dynamic interplay between diet and helminth infection, which in turn influences host gut homeostasis and immune function. Because this diet–helminth interaction modulated both local and peripheral immune cells, it would also be interesting to investigate the effects on other mucosal surfaces and associated disease models such as allergic inflammatory responses in the lung.

We thank M. Schjelde, L. Christensen, and A. Valente (Department of Veterinary and Animal Sciences, University of Copenhagen) for practical support throughout the duration of the study. We also thank D. Liboriussen (DTU Bioengineering Department of Biotechnology and Biomedicine, Technical University of Denmark) for assistance with gas chromatography–flame ionization detection for SCFA quantification.

The authors have no financial conflicts of interest.