Suppression of Obesity by an Intestinal Helminth through Interactions with Intestinal Microbiota

Obesity is increasingly causing lifestyle diseases in developed countries where helminthic infections are rarely seen. Here, we investigated whether an intestinal nematode, Heligmosomoides polygyrus, has a suppressive role in diet-induced obesity in mice.

O besity causes lifestyle diseases such as hypertension, diabetes, and dyslipidemia (1). Excess energy consumption is an important cause of obesity. The overall body energy balance is maintained by adjusting excess or deficient energy in conjunction with energy intake and output. One such metabolic control is present in adipose tissues. Adipose tissue is classified as white or brown (2), with each having different characteristics. White adipose tissue accumulates excessive lipids as triglycerides, while brown adipose tissue consumes fatty acids as heat (3). Brown adipose tissues contain two distinct types of thermogenic adipocytes: classical brown adipocytes and beige or bright adipocytes. A major goal of research on obesity is to understand the factors that activate thermogenic adipocytes, which activate energy consumption, thus preventing obesity.
Uncoupling protein 1 (UCP1) is an integral membrane protein expressed in the mitochondria of brown (4) and beige (5) adipocytes, and it uncouples oxidative phosphorylation. When UCP1 is activated, energy generated by the lipolysis of fatty acids and glucose is directly converted to heat, without being directed to ATP synthesis, and the heat then dissipates (6). Thus, this molecule is crucial for energy expenditure. Heat production by UCP1 is controlled by norepinephrine (NE), which is released from sympathetic nerves that are densely distributed in brown fat. Fatty acids are generated through lipolysis that is induced by NE acting on its receptor and driving activation of adipocyte lipases, including adipose triglyceride lipase and hormonesensitive lipase. The fatty acid produced is oxidatively decomposed and activates UCP1. When NE acts on white adipocytes, fat decomposition similarly occurs, but the fatty acid produced is released into the blood and is consumed by brown adipocytes and muscle. A characteristic of beige adipocytes is the dynamic regulation of UCP1 by external stimuli. ␤3-Adrenergic receptor (␤3AdR) agonists induced the marked expression of UCP1 in beige adipocytes (5).
Recent studies demonstrated a close relationship between intestinal microbiota composition and several diseases, including metabolic, gastrointestinal, and inflammatory diseases (7,8). Additionally, obesity is associated with low gut microbiota diversity, and it may alter the composition of certain bacteria in humans and animal models (9). Many factors that affect the onset of obesity are associated with modulation of the microbiota (10,11). The intestinal microbiota is associated with psychiatric diseases including depression and autism, and previous studies reported that intestinal bacteria communicated with the central nervous system to stimulate the production of neurotransmitters and hormones, including serotonin, dopamine, and ␥-amino butyric acid (12)(13)(14).
The incidence of obesity has increased, especially in developed countries where helminthic infections have almost been eliminated (15). Several lines of evidence indicate an inverse correlation between helminthic infections and obesity as well as inflammation-mediated disorders (16,17), suggesting that helminths may have suppressive effects on these diseases. However, protective mechanisms involved in how helminths suppress obesity are largely unknown. Given that intestinal helminths modulate gut microbiota, the current study investigated the effects of an intestinal nematode on obesity in mice fed a high-fat diet (HFD) by focusing on the gut microbiota. We found that helminthic infection affected gut bacteria, resulting in increased NE production that upregulated UCP1 in adipose tissues.

RESULTS
Infection with Heligmosomoides polygyrus reduces established obesity. To investigate whether infection with Heligmosomoides polygyrus has therapeutic effects on established obesity, we used mice fed an HFD for 4 weeks as infection hosts (Fig. 1A). This diet resulted in an increase in body weight up to 20% and an increase in fat mass and dyslipidemia (Fig. 1B to E), and these mice were considered obese. Before analysis, we confirmed that feeding with an HFD had no adverse effect on H. polygyrus infection. Obese mice harbored adult worms and produced eggs at levels comparable to those of mice fed a normal diet (ND) (Fig. 1F and G). Continuously feeding obese mice an HFD H. polygyrus-mediated reduction in obesity is associated with UCP1 expression induced via NE. Because food intake in obese mice was not influenced by infection with H. polygyrus ( Fig. 2A), weight loss might be a result of excess calorie consumption rather than of a reduction of calorie intake. We postulated that, in addition to a marked decrease in fat mass (Fig. 1C), H. polygyrus infection might promote energy generation in adipose tissues, which is controlled by UCP1 expression in adipocyte mitochondria. Thus, UCP1 mRNA expression in adipocytes obtained from H. polygyrus-infected obese mice was quantified (Fig. 2B). Although UCP1 was substantially upregulated even in the absence of H. polygyrus as obesity proceeded, infection with H. polygyrus induced significantly more UCP1 expression. UCP1 expression in adipocytes is regulated by sympathetic innervation through the production of NE (18). In parallel with UCP1 expression, obese mice infected with H. polygyrus showed the highest concentration of NE in the blood (Fig. 2C). Released NE acts on ␤3AdR expressed on white adipocytes. We also analyzed the mRNA expression of ␤3AdR in adipocytes, and the same trends as for UCP1 and NE were observed: obesity increased ␤3AdR expression, and infection with H. polygyrus had additional effects (Fig. 2D). Infection of nonobese ND-fed mice with H. polygyrus resulted in a slight increase in NE levels and the expression of ␤3AdR, but it had no effect on UCP1 expression ( Fig. 2B to D). These results suggest that UCP1 expression is dependent on NE-␤3AdR interactions, which may be responsible for the reduction of obesity in obese mice infected with H. polygyrus. To test this possibility, a selective antagonist of ␤3AdR, SR59230 (19), was administered to H. polygyrus-infected obese mice. Blockade of ␤3AdR did not alter the ability to produce eggs in either obese or nonobese mice (Fig. 3A). Mice treated with SR59230 failed to suppress weight gain even in the presence of H. polygyrus (Fig. 3B). Although SR59230 did not affect the NE concentration, it markedly suppressed the increased UCP1 mRNA expression ( Fig. 3C and D). In mice fed an ND, there was no change in body weight, NE concentration, or UCP1 expression ( Fig. 3B to D). Thus, the protective role of H. polygyrus infection against obesity is dependent on NE, presumably by inducing UCP1 for energy expenditure in adipocytes.
Neurogenic NE contributes to the H. polygyrus-mediated reduction of obesity. We next analyzed the mechanisms involved in the production of NE during H. polygyrus infection. Infection of obese mice with H. polygyrus resulted in an increase in NE concentrations in the serum and adipose tissues (Fig. 4A). Because sympathetic innervation is the major source of NE in adipose tissues (20), H. polygyrus-infected obese mice were subjected to chemical denervation by treatment with reserpine. This agent blocks the intake of catecholamine into presynaptic vesicles and therefore allows The intestinal microbiota is involved in the H. polygyrus-mediated reduction of obesity. Although treatment of uninfected mice with reserpine completely suppressed NE production in adipose tissues and serum, obese mice infected with H. polygyrus still contained substantial amounts of NE after treatment with reserpine (Fig. 4A), suggesting the existence of other sources of NE besides sympathetic nerves. Thus, we next addressed how H. polygyrus induces NE production by focusing on the intestinal microbiota (i) because H. polygyrus parasitizes small intestines and affects the microbiota (16,17) and (ii) because several neurotransmitters, including NE, originate from the microbiota (21,22). In addition, infection of obese mice with H. polygyrus increased NE levels in the feces (Fig. 5A). To this end, obese mice infected with H. polygyrus were orally treated with antibiotics to reduce the intestinal microbiota (Fig. 5B). This treatment did not affect the parasitism of H. polygyrus as assessed by egg production (Fig.  5C). Treatment with antibiotics decreased both the NE concentration and expression of UCP1 in H. polygyrus-infected obese mice, resulting in the attenuated suppression of weight gain (Fig. 5D to G). Thus, the microbiota has a crucial role in the H. polygyrusmediated prevention of obesity.
Infection with H. polygyrus changes the intestinal microbiota in obese mice. To investigate changes in the microbiota responsible for the increase in NE production during infection with H. polygyrus, the fecal microbiota of H. polygyrus-infected mice was comprehensively analyzed. Feeding with the HFD alone altered the composition of microbiota, characterized by a decrease in Proteobacteria and an increase in unassigned bacteria (Fig. 6A). Mice infected with H. polygyrus contained more representatives of the Firmicutes and Proteobacteria than uninfected mice independent of diet they were fed  Fig. 2. All values are presented as the means Ϯ standard deviations of five mice in one representative experiment. Symbols in scatter graphs with bars represent data from an individual animal. *, P Ͻ 0.05; N.S., not significant (using a two-way ANOVA). uninf, uninfected; inf, infected.
( Fig. 6A and B). Previous studies reported that two bacterial genera, Bacillus and Escherichia belonging to Firmicutes and Proteobacteria, respectively, produced NE in the intestines (23). Quantitative real-time PCR analyses of microbiota revealed that infection of obese mice with H. polygyrus significantly increased Bacillus and Escherichia species (Fig. 6C). In addition, the NE concentration in H. polygyrus-infected obese mice was closely correlated with the amounts of those bacteria (Fig. 6D). These results suggest that infection with H. polygyrus modulates intestinal bacteria to produce NE, which is responsible for limiting weight gain.

DISCUSSION
In this study, we demonstrated that H. polygyrus suppressed HFD-induced obesity. Mechanistically, H. polygyrus affected the composition of the intestinal microbiota to increase NE, resulting in enhanced UCP1 expression in adipose tissues. Because stimulation with NE was reported to induce beige adipocytes from adipocyte progenitors associated with UCP1 expression (5), these thermogenic adipocytes might be induced during H. polygyrus infection.
Recently, Su et al. reported the preventive effects of H. polygyrus infection on HFD-induced obesity, focusing on distinct immune responses (24). Alternative activated macrophages (AAMs) induced during H. polygyrus infection suppressed insulin resistance and inflammation associated with obesity and enhanced UCP1 expression in adipose tissues. We also think that the decisive effector that prevents obesity is increased UCP1 expression. However, we revealed different mechanisms from those of Su et al. involving AAMs that upregulated UCP1 expression. This difference may be due to difference in experimental system, and our system elicited mild obesity without causing hyperglycemia or insulin resistance. Although we did not investigate immune

Intestinal Helminths Suppress Obesity in Mice
Infection and Immunity responses, mild obesity was not associated with inflammation, indicating that it may be regulated in ways other than by immune responses.
Our results using a ␤3AdR antagonist demonstrated that NE plays a crucial role in UCP1 expression. Chemical denervation using reserpine increased body weight even in the absence of H. polygyrus infection, indicating that neurogenic NE primarily regulates the induction of UCP1 under physiological conditions. In addition to neurogenic NE,  Fig. 2 and 4. All values are presented as the means Ϯ standard deviations of more than five mice in one representative experiment. Symbols in scatter graphs with bars represent data from an individual animal. *, P Ͻ 0.05, using a two-way ANOVA. DW, distilled water. intestinal microbiota contributed to increased NE levels in obese mice infected with H. polygyrus. Because H. polygyrus resides in the small intestine, it is thought to affect the intestinal microbiota. Several studies demonstrated that the intestinal flora composition was altered after H. polygyrus infection (25,26). Here, we observed that H. polygyrus infection altered the gut microbiota composition, and, specifically, more Bacillus and Escherichia species were detected, both of which are known to generate NE (14,27). Other bacterial products such as short-chain fatty acids are known to have anti-obesity effects, and some bacteria are associated with obesity (28); therefore, additional effects caused by changes in microbiota may contribute to the reduction of obesity. Thus, comprehensive analyses of the intestinal microbiota are required to understand metabolic homeostasis better during H. polygyrus infection.
The question arises as to whether the induction of NE is favorable to H. polygyrus parasitism. The expulsion of intestinal worms depends on the production of mucin and peristaltic movements controlled by parasympathetic nerves. Thus, the activation of sympathetic nerves may help intestinal parasites to settle by suppressing intestinal movement. However, this seems unlikely because egg production was not decreased by treatment with the ␤3AdR agonist or antibiotics ( Fig. 3A and 5C). We did not address how H. polygyrus modulates the microbiota at the molecular level. Further analyses to determine which molecules are involved will be valuable to treat or prevent obesity, such as by using prebiotics and probiotics.

MATERIALS AND METHODS
Mice. Male C57BL/6J mice were purchased from Japan SLC (Hamamatsu, Japan), maintained under specific pathogen-free conditions, and used for experiments at 10 to 12 weeks of age. For experimental feeding, an HFD containing 60% fat (HFD-60; Oriental Yeast Corporation, Tokyo, Japan) and a control normal diet (AIN-93M; Oriental Yeast Corporation) were used. All animal experiments were reviewed and approved by the Committee for Ethics on Animal Experiments at the Graduate School of Gunma University (approval number 16-041) and were conducted under the control of the Guidelines for Animal Experiments in the Graduate School of Gunma University and in accordance with Law No. 105 and Notification No. 6 of the Japanese Government.
H. polygyrus infection. Infectious stage III H. polygyrus larvae (L3) were prepared as previously described (29) and stored at 4°C until use. Mice were inoculated orally with 200 L3 larvae using gastric intubation. Eggs in feces were detected using a microscope to confirm successful infection.
Sample collection. Blood was taken from the mice via cardiac puncture under anesthesia, and mice were sacrificed by cervical dislocation. Epididymal adipose tissue was aseptically removed, and adipocytes were purified as previously reported (30). The adipose tissue weight was then measured. Serum samples were separated from the collected blood for analyses. In some experiments, all adult worms recovered from the small intestine of mice infected with H. polygyrus were counted. Serum analysis. Serum samples were analyzed for triglyceride and nonesterified fatty acid (NEFA) using LabAssay (Wako, Tokyo, Japan) and for NE using a standard enzyme-linked immunosorbent assay (ELISA), in accordance with the manufacturer's instructions (ISM, Tokyo, Japan).
Reserpine injection. Reserpine (0.5 mg/kg body weight) was intraperitoneally injected 1 day before and 3 and 5 days after H. polygyrus infection.
Gut microbiota analysis by 16S rRNA sequencing. Fecal samples collected from mice were immediately frozen in liquid nitrogen and stored at Ϫ80°C. Fecal DNA extraction was performed according a previous study with minor modifications (32). A grain of mouse feces was suspended with sterilized sticks in 475 l of TE10 buffer containing 10 mM Tris-HCl (pH 8.0) and 10 mM EDTA. The fecal suspension was incubated with 15 mg/ml lysozyme (Wako) at 37°C for 1 h. A final concentration of 2,000 units/ml of purified achromopeptidase (Wako) was added and then incubated at 37°C for 30 min. We added 1% (wt/vol) sodium dodecyl sulfate and 1 mg/ml proteinase K (Merck Japan, Tokyo, Japan) to the suspension and incubated it at 55°C for 1 h. After centrifugation of the suspension, bacterial DNA was purified using phenol-chloroform-isoamyl alcohol (25:24:1) solution. DNA was precipitated by adding ethanol and sodium acetate. RNase A (Wako) was added to bacterial DNA in TE buffer to a final concentration 1 mg/ml. To remove fragmented low-molecular-weight DNA, precipitation with polyethylene glycol (PEG) 6000 was performed after RNase treatment.
The V4 variable region (residues 515F to 806R) was sequenced on an Illumina MiSeq, according to the method of Kozich et al. (33). Each reaction mixture contained 15 pmol of each primer, 0.2 mM deoxyribonucleoside triphosphates, 5 l of 10ϫ Ex Taq HS buffer, 1.25 U Ex Taq HS polymerase (TaKaRa Bio), 50 ng of extracted DNA, and sterilized water to a final volume of 50 l. PCR conditions were as follows: 95°C for 2 min, 25 cycles of 95°C for 20 s, 55°C for 15 s, and 72°C for 5 min, followed by 72°C for 10 min. The PCR product was purified by AMPure XP (Beckman Coulter, Brea, CA, USA) and quantified using a Quant-iT PicoGreen double-stranded DNA (dsDNA) assay kit (Life Technologies Japan, Tokyo, Japan). Mixed samples were prepared by pooling approximately equal amounts of PCR amplicons from each sample. The pooled library was analyzed with an Agilent High Sensitivity DNA kit on an Agilent 2100 Bioanalyzer (Agilent Technologies, CA, USA). Real-time PCR for quantification was performed on a pooled library using a KAPA Library Quantification kit for Illumina according to the manufacturer's protocols. Based on the quantification, the sample library was denatured and diluted. A sample library with 20% denatured PhiX spike-in was sequenced by MiSeq using a 500-cycle kit. We obtained 2-by 250-bp paired-end reads. Taxonomic assignments and the estimation of relative abundances from sequencing data were performed using the analysis pipeline of the QIIME software package (34).
Statistical analysis. Group means were compared by two-way analysis of variance (ANOVA), followed by Tukey's post hoc test or two-tailed Student's t test. Probability values below 0.05 were considered statistically significant.