Chronic Helminth Infection Perturbs the Gut-Brain Axis, Promotes Neuropathology, and Alters Behavior

Abstract

Helminth infections in children are associated with impaired cognitive development; however, the biological mechanisms for this remain unclear. Using a murine model of gastrointestinal helminth infection, we demonstrate that early-life exposure to helminths promotes local and systemic inflammatory responses and transient changes in the gastrointestinal microbiome. Behavioral and cognitive analyses performed 9-months postinfection revealed deficits in spatial recognition memory and an anxiety-like behavioral phenotype in worm-infected mice, which was associated with neuropathology and increased microglial activation within the brain. This study demonstrates a previously unrecognized mechanism through which helminth infections may influence cognitive function, via perturbations in the gut-immune-brain axis.

cognition, helminth, inflammation, microbiome

Soil-transmitted helminth (STH) infections, such as Ascaris lumbricoides, Trichuris trichiura, and hookworms, cause significant morbidity, particularly in children, including anaemia, reduced growth [1], impaired cognitive function, and reduced educational advancement [2]. Multiple studies have looked at the effects of STHs on cognitive function, with some conflicting evidence regarding their causal role that likely reflects the complexity in human population-based studies [3, 4]. Hence, more carefully controlled studies in human populations and animal models are required. The mechanisms by which STHs influence cognitive development also remain unclear, and they are likely to involve both direct effects of nutritional deficits on the brain and indirect effects of pathophysiological events occurring in the gut environment. Recent research has highlighted the crucial role for the intestinal microbiome in regulating central nervous system (CNS) development, especially synaptogenesis and myelination [5]. Disruption of gut microbial homeostasis (ie, dysbiosis) can affect emotional behavior and related brain systems, leading to a range of mental health disorders including autism spectrum disorder, anxiety, depression, and chronic pain [6]. There is a bidirectional functional communication between gut microbiota, the gastrointestinal tract, and the CNS; these relationships have been recognized as the microbiota-gut-brain axis, which operates through a variety of physiological mechanisms, including neural, hormonal, and immunological pathways [7].

Recent evidence from human and animal studies indicates that helminth infections can influence the gut microbiota [8]. These interactions may be beneficial for some inflammatory conditions [8], but they may be detrimental if they occur during a critical window of childhood development and if worm burdens are high. In a recent study, we hypothesized that helminth-induced changes in the gut microbiome are associated with intellectual development at a number of critical time windows, including prenatal, postnatal, during adolescence, and into adulthood [9]. In the present study, we extend on these findings by using a murine model of helminth infection. We describe a previously unrecognized pathway by which helminths may negatively impact brain structure, behavior, and cognitive function, by disturbing the homeostasis of the gut microbiota and the immune system.

METHODS

Study Design

Male wild-type C57BL/6 mice were maintained at James Cook University (JCU). All protocols were approved by the JCU Animal Ethics Committee. Trichuris muris was maintained in genetically susceptible mice, and excretory/secretory (ES) antigens were derived from adult worms. At 8 weeks of age, groups of 12 mice received either 30 T muris eggs, 6 T muris eggs, or saline placebo (naive) by oral gavage. Researchers performing behavioral, microbiome, and histological analyses were blinded as to the treatment group. Blood was collected by cardiac puncture or submandibular bleeding and serum was stored at −80°C. Colon tissue (1-cm sections) were sliced longitudinally and cultured overnight at 37°C (organ culture). Trichuris ES-specific cytokine responses were assessed by restimulation of mesenteric lymph node (mLN) cells with 50 μg/mL ES. Serum and culture supernatants were assessed for cytokine levels using enzyme-linked immunosorbent assay (eBioscience).

Deoxyribonucleic Acid Isolation, 16S Recombinant Ribonucleic Acid Gene Amplicon Sequencing, and Analysis

Fecal samples were collected and pooled from each cage of 6 mice and frozen at −80°C before analysis. Samples were put into a 5% potassium dichromate solution before deoxyribonucleic acid (DNA) extraction using the PowerSoil DNA extraction kit (MoBio). Illumina paired-end sequencing (MiSeq) was performed on the V4 region of the 16S ribosomal ribonucleic acid (rRNA) gene as amplified using primers based on the Earth Microbiome Project primers 515F and 806R. Raw data were quality filtered and trimmed using trimmomatic version 0.32 (with parameters PE, leading 15, trailing 15, minlen 180). Paired reads were then assembled with PEAR allowing a minimum overlap of 20 base pairs. Individual read pairs were converted to fasta and qual files. Sequence data analysis was done in QIIME version 1.9.1. Operational taxonomic units (OTUs) were assigned from sequence reads using a closed-reference OTU picking protocol against the Greengenes reference database (version 13_8) clustered at 97% similarity threshold with uclust. Taxonomy was assigned with the Ribosomal Database Project (RDP) classifier accepting the Greengenes taxonomy string of the best matching Greengenes sequence. Statistical analyses including were done in R, including functions from the vegan and indicspecies packages. Shifts in bacterial community composition were visualized with nonmetric multidimensional scaling ([nMDS] function metaMDS in vegan package), and the effects of sampling time and infection treatment were tested with permutational multivariate analysis of variance ([PERMANOVA] function adonis in vegan package). The OTUs that were significantly overrepresented (P > .05, adjusted for multiple testing) in either infection treatment were identified with the function signassoc (indicspecies package). Only these “indicator OTUs” were displayed in the nMDS ordination.

Behavioral Assessments

Animal behavior was tested at 10 months of age using tests and procedures described in detail in Kraeuter et al [10]. In brief, to assess learned helplessness/depression-like behavior, the forced swim test was used; this test was conducted in a 2000-mL beaker filled with 1400 mL water. Mice were placed into the water, and the activity (mobility, immobility, latency to immobility) was video recorded for 6 minutes. The open field test was used to assess psychomotor activity and “emotionality”. The test mouse was placed into the middle of the apparatus (420 × 420 × 420 mm; 205 × 205 mm center zone) for 15 minutes. Behavior in the open field was recorded by a video camera above the arena. A spatial reference memory Y-maze test was used to assess spatial reference memory. During the training session, 1 arm of the Y-maze was closed. The animal was placed into the start arm facing the center and was allowed to freely explore the “start” and the “other” arm for 15 minutes during the acquisition phase. One hour later, during the testing session, the animal was placed back to the apparatus, and the barrier was removed to allow the animal to explore the entire arena (start arm, “other arm”, novel arm) for 5 minutes. Behavior was recorded by a video camera placed above the arena. Video recordings were analyzed using TopScan Light (Cleaver Sys Inc., Reston, VA).

Pathophysiological Studies

Histological studies were performed on tissue sections from 4 mice from each group. Mice were perfused with 4% paraformaldehyde, and brain tissue was fixed overnight in paraformaldehyde and transferred to 30% sucrose solution before tissue sectioning and immunohistochemical staining. Microglial activation was assessed by staining with a rabbit polyclonal IbaI antibody (Wako Cat. No. 019-19741). The degree of immunolabeling across serial sections was ascertained as a measure of activation and degeneration. Neuronal degeneration was assessed using highly sensitive Fluoro-Jade C staining (AG325 Merck) according to the manufacturer’s instructions.

Statistics

Statistical analysis was performed using the SPSS version 23 software package (IBM SPSS Statistics). P < .05 will be considered significant throughout all data. All data are depicted as mean ± standard error of mean.

RESULTS AND DISCUSSION

As expected, infection of mice with 30 T muris eggs resulted in inflammatory pathology in the large intestine after 1 month (Figure 1A) and antigen-specific interferon (IFN)γ responses in mLN (Figure 1B). Our primary study involved examining the effect of different burdens of helminth infections on systemic inflammation, the microbiome, and behavior (Figure 1C). Hence, we treated groups of 12 mice with either 30 T muris, 6 T muris, or saline placebo (naive) by gavage. Trichuris infection resulted in lower mouse weight at 4 months (naive, 36.4 ± 2.4 g; 6 Trichuris, 35.1 ± 2.9 g; and 30 Trichuris, 34.4 ± 1.7 g; mean ± standard deviation); however, these differences were not statistically significant when using a Mann-Whitney t test. Analysis of systemic inflammatory responses revealed significant elevations in the proinflammatory cytokine IFNγ in the serum at 1 month, which returned to naive levels by 4 months (Figure 1D). Levels of other proinflammatory cytokines interleukin (IL)-1β and IL-17A were not elevated after infection (Figure 1E–F).

Figure 1.

Chronic helminth infection elicits intestinal pathology, systemic inflammation, and alterations in gut microbiome composition and abundance. Wild-type C57BL/6 mice were infected with 30 embryonated Trichuris muris eggs, or they were uninfected (naive). (A) Representative photomicrographs of caecal histopathology at 1 month postinfection (p.i.). (B) Levels of interferon (IFN)γ in colon organ culture supernatants or Trichuris antigen-restimulated mesenteric lymph node supernatants at 1 month p.i. (C) Main study examining effect of infection with 6 Trichuris eggs or 30 Trichuris eggs on inflammatory, microbiota, and behavioral parameters. Serum levels of (D) IFNγ, (E) interleukin (IL)-17A, and (F) IL-1β at various time points p.i. Data are expressed as mean + standard error of mean; *, P < .05 compared with naive control. (G) The composition of mouse gut microbiome visualized using nonmetric multidimensional scaling (NMDS). Each point refers to a pooled fecal sample consisting of fecal pellets from 2 cages of 6 mice/group at each time point. Points are color-coded according to infection treatment and shaded according to sampling time. Ellipses have been drawn around healthy and infected treatment groups to aid interpretation (H) operational taxonomic units (OTUs) that contribute significantly to the differences in microbiome composition in response to infection treatment (indicator OTUs). Each cross (+) refers to 1 OTU, color-coded by taxonomic assignment (bacterial order level). The OTUs within or close to a treatment group ellipse are overrepresented in that treatment group.

These acute changes in inflammatory parameters were accompanied by changes in the composition of the fecal microbiome. Fecal pellets taken at 1, 4, and 9 months p.i., as well as naive mice were subjected to 16S rRNA sequencing. Our results revealed a Trichuris-dependent shift in microbiota composition 1 month p.i. in infected mice compared with naive mice (Figure 1G). At 4 and 9 months p.i., the microbiota became increasingly similar to the microbiota of naive mice (Figure 1G). Infection dose explained 29% of variation in microbiome community composition (PERMANOVA P = .001), whereas sampling time explained 18% (PERMANOVA P = .04). We observed a generally higher relative abundance of Bacteroidetes and lower abundance of Firmicutes in mice infected for 1–4 months compared with naive mice (Figure 1H). This was confirmed by indicator species analysis, which associated many OTUs of Firmicutes (Clostridiales) with a healthy murine gut microbiota, whereas Bacteroidales were associated with infected mice. The analysis revealed that OTUs affiliated with Enterobacteriales are indicator taxa for the severity of infection during the acute phase, consistent with previous reports [11, 12], which is indicative of a dysbiotic “inflammatory” microbiota.

Conclusions

At 8 months postinfection, we subjected mice to a range of validated and well characterized behavioral tests, which were performed by an investigator blinded to the treatment. A forced swim test revealed that mice that received 30 Trichuris displayed reduced mobility (Figure 2A) and increased immobility (Figure 2B) in the water, suggestive of depression-like behavior. The open field test demonstrated that 30 Trichuris-treated mice had reduced total distance traveled (Figure 2C) and lower number of entries into center area (Figure 2D) compared with controls, indicating a decreased psychomotor activity and a potential depression/anxiety like-behavioral phenotype. The reference Y-maze test revealed that although naive control and 6 Trichuris-infected mice spent more time in the novel arm of the maze than in the other arm, which indicates good spatial memory, 30 Trichuris-infected mice spent approximately the same amount of time in each arm (Figure 2E), consistent with impaired spatial reference memory. Thirty Trichuris-treated mice did not display significantly decreased activity as defined by total distance traveled in the Y-maze during testing (Figure 2F), suggesting that impaired spatial memory was not due to motor impairment. Together, these data demonstrate that chronic, higher-intensity helminth infection may be associated with significant changes in behavior, chiefly a depression-like phenotype and impaired memory. Critically, histopathological analysis of brain tissue sections (amygdala and adjacent cortical region) revealed clear increases in presence of Fluro-Jade C-positive degenerating neurons in 30 Trichuris-treated mice compared with naive mice (Figure 2G). Furthermore, increased Iba1 staining within cortical tissue was observed, indicating increased microglial activation (Figure 2H). Analyses of brain pathology from mice treated with 6 Trichuris did not reveal obvious differences in Iba1 staining compared with naive control mice (data not shown). These data suggest that exposure to heavier helminth burdens earlier in life were more strongly associated with persistent pathological changes in the brain, potentially accounting for the changes in behavior. To our knowledge, this is the first report of neuropathology associated with a helminth infection. Further studies are required to assess potential causal relationships of the worm infection or the relative importance of the changes in the microbiome and inflammatory responses. Regardless, our findings fit in with the broader paradigm shift in neurobiology and neurodegeneration where increasingly the gut microbiome (and perhaps the macrobiome, ie, worms) and interrelated factors that disturb normal homeostasis such as stress, diet, malnutrition and infections, are a major pathological trigger that can cause similar phenotypes in animal models of neurological disease [13]. Several observational studies reported that exposure to STH infections during pregnancy or in school aged children can impair the efficiency of cognitive processes including memory, learning, motor development, verbal fluency, and nonverbal intelligence [14, 15]. Our study demonstrates that helminth infections may affect the gut-brain axis, altering behavior and causing neuropathology. The helminth-gut microbiota-CNS hypothesis needs support from further animal model and human epidemiological studies that would more definitively determine the potential causal relationship between helminth infections in early life (prenatal exposure and <3 years of age), impaired cognitive development, and the microbiome. This could involve the undertaking of deworming studies using the animal model described herein, to test whether differential timings of anthelminthic treatment could reverse this behavioral phenotype. Such experiments could inform the most appropriate age to implement community-deworming procedures in humans (currently aimed at school-age children), which could limit the development of CNS and behavioral disorders in these at-risk populations.

Figure 2.

Chronic Trichuris infection induces a depression-like behavior phenotype, reduced spatial memory, and neuropathology. Behavioral tests were performed at 8 months postinfection (p.i.). Percentage (A) mobility and (B) immobility in a forced swim test. (C) Total distance traveled (meters [m]) and (D) number of center entries in the open field test. (E) Relative amount of time spent in “Start” (S) arm, “Novel” (N) arm, or “Other” (O) arm assessed using the spatial reference memory Y-maze test. (F) Total distance traveled in the reference Y-maze test. Histopathological analysis of brain pathology was performed at necropsy (9 months p.i.). (G) Photomicrographs of Fluoro-Jade C staining in amygdala and adjacent cortical region (the cortex-amygdala transition zone) tissue sections from naive control or high-dose Trichuris-infected mice. (H) Photomicrographs of IBa1 staining in cortical tissue. Histological studies were performed on tissue sections from 4 mice from each group, and representative images are shown. Data are expressed as mean + standard error of mean; *, P < .05.

Notes

Acknowledgments. We thank Stacey Llewellyn (QIMR Berghofer Medical Research Institute) for performing stool deoxyribonucleic acid extractions.

Disclaimer. The funding sources had no involvement in the study design, data collection, analysis and interpretation, writing of the report, and decision to submit the paper for publication.

Financial support. The study was funded by a University of Queensland Early Career Researchers grant (to R. J. S. M.), the Queensland Department of Science, Information Technology and Innovation (to P. R. G.), and the JCU Center for Biodiscovery and Molecular Development of Therapeutics seed grant (to P. R. G. and Z. S.). T. M. W. is supported by a National Health and Medical Research Council Career Development Fellowship (grant APP1105420).

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.