ABSTRACT
Intestinal roundworms cause chronic debilitating disease in animals, including humans. A lack of effective vaccines and the emergence of widespread drug resistance only increase the need to better understand parasite clearance mechanisms within the host. Heligmosomoides polygyrus larvae induce a strong intestinal granuloma response within their murine host, which has been associated with resistance. Immune cells, mostly alternatively activated macrophages and eosinophils, accumulate around the tissue encysted parasites to immobilize and damage/kill developing worms. In a one dose (bolus) experimental infection, infected C57Bl/6 mice are unable to clear parasites which results in chronic infection with high worm burdens. However, using a frequent dose trickle model of infection, we, like others, have found that C57Bl/6 mice can clear infection. We found that the clearance is associated with higher granuloma numbers, but no changes in systemic/intestinal Th2 responses. Within the granulomas, we found that myeloid cells had a different transcriptional profile in each of the infected groups, and that high IgG1, but not IgG2c, IgA or IgE, levels were observed around the larvae of only trickle-infected mice. Our results highlight the importance of the granuloma in the host’s ability to clear H. polygyrus and emphasise the need to study this key tissue in more depth, rather than using correlates such as general intestinal or systemic responses.
AUTHOR’S SUMMARY Despite decades of research on intestinal parasitic worms, we are still unable to clearly point to why so many people (approximately 1.8 billion) and most livestock/wild animals are infected with these parasites. We have made progress in understanding how the immune system responds to parasitic worms, and how these parasites manipulate our immune system. However, identifying effective clearance mechanisms is complex and context dependent. We have used a model of trickle infection (multiple low doses of parasites) to simulate how people/animals get infected in the real world. Using this model, we have identified the host/parasite interface (the granuloma) within the intestinal tissue to be key in determining the host’s ability to clear worms. Specific gene expression signatures in granuloma immune cells and the presence/absence of antibodies within the granuloma are key factors associated with parasite clearance. Surprisingly, more common identifiers of parasitic worm infections (increased serum antibody levels and/or generalized immune markers) did not associate with protection. These novel findings contribute to a better understanding of the mechanisms underlying effective parasitic worm clearance.
INTRODUCTION
Gastrointestinal nematodes are parasites adept at causing chronic recurring infections. Hosts mount a strong immune response to these parasites, essential to control worm burden and host tissue damage. However, the efficacy of the response is dependent on genetics, infection dynamics and environment.
Heligmosomoides polygyrus is an enteric nematode parasite of mice [1],[2]. Ingested larvae encyst in the host intestinal wall and mature into adults that escape into the lumen. Adults remain in the intestinal lumen for the duration of infection. H. polygyrus tissue dwelling stages cause the release of alarmins from epithelial cells as they damage the intestinal wall [3]–[5]. Alarmin-activated innate lymphoid cells and Th2 polarized CD4+ T cells produce Th2 cytokines [6]–[9] which promote innate immune cell influx to the intestine [2],[6],[10]. The accumulation of immune cells is referred to as a granuloma [7],[11],[12]. Increased granuloma size and number are associated with increased resistance to nematodes [13].
Within the granuloma, the host response focuses on damaging or killing the parasitic nematodes, as well as healing the damage caused by the growing worm (reviewed in [14]). Eliminating tissue stage parasites is thought to rely on antibody dependent cell mediated cytotoxicity (ADCC) by macrophages and eosinophils [7],[15],[16], the main cellular players within the granuloma. Antibodies are key to this process. Resistant strains of mice have been shown to develop faster and more intense parasite specific antibody responses following H. polygyrus infections, as compared to susceptible strains, where isotypes IgG1, IgA and IgE have been linked to worm clearance [2],[17],[18]. Passive transfer of serum, and specifically IgG1, from infected mice results in decreased adult worm burden and fecundity [19]–[22]. H. polygyrus adult numbers are increased in infected mice lacking IgA [23]. IgG1 and IgE have also been negatively correlated with worm survival across different strains of mice [24]–[26].
Alternatively activated macrophages (AAMs) isolated from H. polygyrus-induced granulomas have increased surface levels of FcγRs as well as complement receptor CD11b and surface binding of IgG1 and IgG3 [27]. However, the exact mechanisms responsible for parasite adherence and killing remain controversial. In vitro observations have yet to be validated in vivo. For example, the CD11b receptor on bone marrow derived macrophages can directly bind H. polygyrus larvae using complement 3 when cultured with immune serum. In vitro, after adhering to the larvae via the CD11b-C3 interaction [27], FcγR1 on macrophages can interact with parasite bound IgG2a/c (but not IgG1) antibodies to immobilize the larvae [28]. However, this binding does not reduce the infectivity of larvae in vivo [27]. In contrast, mice vaccinated with H. polygyrus excretory/secretory products (HES) are protected from chronic infection as a result of the IgG1 response [5]. IgG1 antibodies are thought to bind and neutralize parasite excretory secretory products, that in unvaccinated mice are able to interfere with the functions of protective innate immune cells [5]. However, passive transfer of purified IgG1 did not induce sterile immunity suggesting that other mechanisms are at play [5]. As well as being involved in parasite damage and death, AAMs and eosinophils are both also involved in host tissue repair [14]. AAMs produce immunoregulatory and wound healing molecules [29],[30] which promote extracellular matrix (ECM) deposition during helminth infections [31]–[34]. H. polygyrus infections induce Ym1 and RELM-α secretion from AAMs [2], both linked with the wound healing phenotype [35]. Eosinophils also produce RELM-α [36], TGF-α, TGF-β and fibroblast growth factors [37]–[39]. In addition, Arginase 1, a marker for AAMs has been found to be essential in both parasite expulsion and wound healing during H. polygyrus infections (9,40). Collagen is a major component of the ECM and excessive collagen deposition leads to fibrosis and scarring during chronic helminth infections [41].
Immune responses are not only generated to tissue dwelling parasitic stages but also to the adults found in the intestinal lumen. The cytokines IL-4 and IL-13 enhance smooth muscle contractility of the intestine via STAT6 dependent pathways [42] to help eliminate adult worms [43]–[45]. IL-4, IL-9 and IL-13 also regulate goblet cell hyperplasia and increase mucus production during gastro-intestinal (GI) nematode infections [46],[47], which makes it more difficult for adult parasites to coil around intestinal villi. In addition, RELM-β produced by goblet cells interferes with the ability of adult parasites to feed, thus limiting their numbers [48]. Finally, H. polygyrus infections induce polyclonal and parasite specific serum antibody responses, which function to limit adult female egg production [5],[23],[27],[28].
Most of the murine studies on helminth infection use a bolus model of infection (one large dose), with some groups adopting a drug clearance model (bolus infection, drug clearance, bolus infection) to simulate mass drug administration programs [7],[49].
However, under natural conditions, GI nematodes are ubiquitous in the environment [50] and hosts are constantly coming into contact with them. While many hosts are infected, few have life threatening levels of worms implying immune regulatory mechanisms are at play [5]. Most hosts are unable to clear infection but can limit excessively damaging worm burdens [13]. Hence, we (and others) have set up experimental infection models using trickle infections to study parasite clearance of GI nematodes in a more natural setting [18],[51]–[53]. We use multiple low doses of larvae, given over a specific time period to achieve this.
H. polygyrus trickle infections in genetically resistant and susceptible strains of mice reveal that the frequency of infection is an important determinant of parasite expulsion, where frequently infected mice eliminate worms more rapidly than mice infected with the same total number of larvae but in less frequent doses [53]. The aim of our study was to identify the host protective immune mechanisms underlying these results. Previous studies speculated that improved antibody and innate immune cell responses to tissue dwelling parasites were key elements [18],[53]. We were able to reproduce the data demonstrating that in susceptible mice, trickle infection results in reduced worm burdens. However, we were also able to show that this reduction was associated with increased levels of antibodies bound to tissue larvae and a specific gene expression signature in the granulomas. All other correlates of Th2 immunity measured between the bolus- and trickle-infected mice were similar including systemic Th2 cytokine responses and antibody levels, as well as local physiological, mucosal and immunological responses in the small intestine. Our results highlight the importance of the granuloma in the host’s ability to clear H. polygyrus and emphasise the need to study this key tissue in more depth, rather than using correlates such as general intestinal or systemic responses.
MATERIALS AND METHODS
Mice, parasites and antigen
Female and male C57Bl/6 mice aged 6-8 weeks (bred and maintained at the animal care facility, Department of Biological Sciences, University of Calgary or University of California, Riverside, USA) were used for all experiments. All animal experiments were approved by the University of Calgary’s Life and Environmental Sciences Animal Care Committee (protocols AC17-0083 and AC17-0240) and the University of California, Riverside’s Institutional Animal Care and Use Committee (https://or.ucr.edu/ori/committees/iacuc.aspx; protocol A-20180023). All protocols for animal use and euthanasia were in accordance with either the Canadian Council for Animal Care (Canada) or National Institutes of Health (USA) guidelines. Animal studies are in accordance with the provisions established by the Animal Welfare Act and the Public Health Services (PHS) Policy on the Humane Care and Use of Laboratory Animals.
Female BALB/c mice and Swiss Webster mice aged 6-8 weeks were purchased from Charles River Laboratories (Senneville, Quebec). Infected mice were orally gavaged with 200 third stage Heligmosomoides polygyrus larvae (maintained in house, original stock was a gift from Dr. Allen Shostak, University of Alberta, Canada) and euthanized at either 7, 14, 21 or 28 days post initial infection. Mice were infected according to the bolus or trickle infection regimes (Fig 1A, 1B & 8A). To avoid differences in counts during the trickle infections, on day 0, two identical bolus solutions were made up (200 worms/100ul). One was used to infect the bolus infected mice and one was used for the trickle infected mice and diluted as necessary according to the number of trickle doses.
H. polygyrus antigen was prepared by collecting live adult worms from 14-day infected mice using modified Baerman’s apparatus. Worms were washed multiple times and homogenized in PBS using a glass homogenizer. The resulting solution was centrifuged (13, 000 g, 10 minutes, 40C) and the supernatant filtered (0.2μm filter, Nalgene). The protein concentration was calculated using the Bradford assay. The antigen was stored at 15 mg/ml at −800C.
Adult worm burden and granuloma number
Small intestines of infected mice were harvested and opened longitudinally. The number of adult worms present in the intestinal lumen and of granulomas present along the length of the small intestine were counted using a dissection microscope.
Transit time
Gastrointestinal transit time was measured one day prior to euthanasia. Mice were fasted for 6 hours and 200 μl of 5% Evans blue (Sigma) in 5% gum arabic (ACROS organics) was orally gavaged using a ball tip 20 gauge 1.5’’, 2.25mm curved animal feeding needle. Each mouse was labelled, with the time of dye administration recorded. Mice were transferred to clean empty cages and the time to pass the first blue fecal pellet was recorded. Gastrointestinal transit time was calculated for each mouse.
Cell isolation and in vitro re-stimulation assay
MLN and SPL were mechanically dissociated into single cell suspensions. Cells were counted using a Beckman-Coulter ViCell XR. MLN and SPL were cultured at 1 x 106 cells/ml for 48 hours in RPMI medium, 10% FCS, 1% L-glutamine, 1% penicillin/ streptomycin (supplemented RPMI 1640) in the presence of 10 μg/ ml H. polygyrus antigen or 2 μg/ml concanavalin A (Sigma) at 37 0C with 5% CO2. Supernatants were collected for cytokine measurements. Measurements for antigen specific production were not included in the analysis unless cytokine production was observed in the wells with concanavalin A stimulation.
Serum
Blood samples were collected using a terminal cardiac bleed. Blood was left to clot for 30 minutes and then centrifuged twice at 11, 000 g at 4oC for 10 minutes. Serum was collected and used either fresh or stored at −80oC.
Intestinal Tissue homogenates
Small intestines were opened longitudinally and washed with PBS to remove luminal content. The mucosal surface was identified under a dissecting microscope. The mucosal surface (with its mucus) was gently scraped using a glass slide. Scrapings were weighed, added to 500 μl lysis buffer (10 μM tris HCl, 0.025% sodium azide, 1% tween 80, 0.02% phenylmethylsulfonyl fluoride) with one complete protease inhibitor tablet (Roche diagnostics GmbH, Germany) and homogenized using a bead beater (40 seconds at speed 6 using the Fast-prep-24 bead beater, MP biomedical). The homogenate was centrifuged at 11, 000 g at 4oC for one hour. Supernatants were collected and used fresh or stored at −80oC.
ELISAs
Cytokines in serum and intestinal tissue homogenates were measured by ELISA according to manufacturer’s guidelines (R & D systems, DY404 (IL-4), DY413 (IL-13), DY594 (IL-21). Total IgE (BD, 555248) and IgA (capture antibody, BD, 556969, detection antibody, BD, 556978) levels were measured by ELISA according to manufacturer’s instructions.
Antigen specific antibody responses were also measured by ELISA. ELISA microplates were coated with 10 μg/ml H. polygyrus antigen in carbonate buffer (0.1mM NaHCO3, pH 9.6), overnight at 4oC. Plates were blocked with 2% BSA in TBS/0.05% tween 20 for 2 hours at 37oC. Sera were diluted in TBS Tween and added to wells overnight at 4oC. Antigen-specific IgG1 was detected with HRP-conjugated corresponding detection antibodies (anti-IgG1 (BD, 553441) with TMB peroxidase substrate (T3405, Sigma). The reaction was stopped using 1M H2SO4 solution and the colour change was read at 450nm.
Antibody detection in the granuloma
Consecutive formalin fixed paraffin embedded mouse small intestinal sections were deparaffinized using two, five-minute xylene washes, rehydrated by washing in 95% ethanol and 70% ethanol for 5 minutes. Slides were incubated in 2% sodium borohydride (VWR, BDH4604) in PBS for 40 minutes at RT to remove auto fluorescence. Antigens were retrieved using 2.5% trypsin (Thermo scientific, 15090046) in 0.1% HEPES buffer, incubated at 37 0C for 25 minutes. Blocking steps were performed following PBS washes. All samples were blocked with starting block (Thermo Scientific, 37578) for 1 hour at RT and rat/rabbit/goat serum for 30 minutes at RT. Following blocking steps, slides were incubated with rat anti-mouse IgG1 (BD, 562026), IgE conjugated to FITC (BD, 553415) or rat anti-mouse IgA (BD, 559354) overnight at 4oC. For IgG2c staining, slides were incubated overnight at 4oC with unconjugated rabbit anti-mouse IgG2c antibody (Invitrogen, SA5-10221) followed by incubation with goat anti-rabbit IgG conjugated to Alexafluor488 (ImmunoResearch Laboratories, 111-545-003) for 2 hours at RT. Slides were washed in PBS for 15 minutes and mounted with Fluoroshield with DAPI (Sigma, F56057). Images were acquired using Thorlabs Tide whole-slide scanning microscope, x20 objective and analysed using Fijji 5.59.05 software.
Nanostring nCounter gene expression assay
Intestinal tissue from naïve mice or dissected pooled granulomas from infected mice were snap frozen in liquid nitrogen and RNA was isolated using phenol-chloroform extraction (TRIZOL, Sigma). RNA was quantified using a nanodrop and 50 ng was used for the Myeloid Innate Immunity V2 panel (NanoString) according to the manufacturer’s guidelines. Gene expression analysis was conducted in R (1). Gene counts obtained via the nanostring hybridization assay were normalized with NanostringNorm (2) using the negative control probes, positive control probes and housekeeping genes Eif2b4, Polr1b, and Edc3. Of the 20 housekeeping genes included in the assay Eif2b4, Polr1b, and Edc3 were the only ones found to have consistent expression among all samples in preliminary comparisons that used all 20 housekeeping genes for normalization. Therefore Eif2b4, Polr1b, and Edc3 were the only housekeeping genes used for normalization in subsequent analyses. The normalized counts were then compared using DESeq2 (3) to find differentially expressed genes in pairwise comparisons between treatment groups. A false discovery rate adjusted p-value cut off of 0.05 and a fold-change cutoff of 2 were used to identify genes that were differentially expressed in each pairwise comparison. The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus (4) and are accessible through GEO Series accession number GSExxx (number will be available upon acceptance).
Statistical analysis (except for nanostring results)
Mann-Whitney and Kruskal Wallis tests with Dunn’s multiple comparisons were used to assess differences between either two or more experimental groups using GraphPad Prism.
RESULTS
C57Bl/6 mice develop a resistant phenotype when infected using the trickle regimen
Two inbred strains of mice, C57Bl/6 (genetically susceptible) and BALB/c (genetically resistant) [13] were infected with H. polygyrus according to the bolus or trickle adult infection regimes (Fig. 1A & 1B). When given a bolus infection of 200 worms, BALB/c mice, being partially resistant to H. polygyrus, eliminated the majority of their worms by 14 days post-infection (mean worm burden of 53, SD +/- 30, Fig. 1C). Worm burdens declined over time until day 28 post infection, where the average number of worms per mouse was 8.5. SD +/- 13 (Fig. 1C). In contrast, C57Bl/6 mice being susceptible to H. polygyrus harboured high numbers of adult worms (an approximate mean value of 100 worms per mouse, Fig. 1D) in the intestinal lumen at all time points tested (Fig. 1D left). When infected according to the trickle protocol, infection dynamics in BALB/c mice were similar to those infected with the bolus regimen, whereby mice had low burdens at all time points post-infection, with near complete clearance by day 28 post-infection (Fig. 1C). However, C57Bl/6 mice infected according to the trickle protocol also eliminated most of the adult worms by 28 days post infection (mean worm burden 11, SD +/- 22, Fig. 1D). This mirrored the BALB/c mice results (Fig. 1C) but was in stark contrast to the results obtained for C57Bl/6 mice infected according to the bolus protocol despite similar worm burdens at day 14 post-infection (bolus: 129, SD +/- 38, trickle: 106, SD +/- 24, Fig. 1D). Administering H. polygyrus larvae in low frequent doses to C57Bl/6 mice changed their susceptibility to infection (Fig. 1D).
Systemic IL-4 and IL-13 responses do not differ between bolus and trickle infected mice
H. polygyrus clearance has been associated with a strong Th2 response, specifically increases in IL-4 and IL-13 cytokines [54]–[56]. As the Th2 immune response develops in response to H. polygyrus, MLN (mesenteric lymph nodes) and SPL (spleen) cell numbers increase [57]. Despite BALB/c mice having higher numbers of cells than C57Bl/6 mice at all time points in both organs, we found no differences between the trickle and bolus groups in either strain at any of the time points in either of the organs (Fig. 2A-D, left panel).
We measured the levels of the antigen-specific Th2 cytokine production (IL-4 and IL-13) in both the MLN and SPL as well as in the serum of mice by ELISA. Levels of cytokines measured in BALB/c mice were in general higher than in C57Bl/6 mice, as has previously been reported [13],[58],[59]. In BALB/c animals, levels were increased early and remained elevated over the course of infection for IL-4 in the MLN (increased from day 7, Fig. 2A middle panel) and the SPL (increased from day 14, Fig. 2C middle panel) and IL-13 in the MLN (increased from day 14 but decreased by day 28, Fig. 2A right panel) and SPL (increased from day 7, Fig. 2C right panel). In C57Bl/6 animals, levels were increased early but decreased by the later time points for IL-4 in the MLN (decreased from day 21, Fig. 2B middle panel) and the SPL (decreased from day 21, Fig. 2D middle panel) and IL-13 in the MLN (decreased by day 28, Fig. 2B right panel) and SPL (increased over the whole infection, Fig. 2D right panel). Levels of IL-4 and IL-13 were undetectable in the serum, as seen by others [60].
Taken together, we show that despite differences in cytokine levels according to mouse strain, organ and post-infectious time-point, no differences were detected between bolus and trickle infected mice in any of the conditions tested (Fig. 2).
Serum antibody levels do not differ between trickle and bolus infected mice
IgE and parasite specific IgG1 have both been shown to increase during primary and secondary H. polygyrus infection [49]. IgG1 has been associated with parasite clearance [5] while IgE is thought to reduce parasite fecundity [23]. We investigated changes in serum antibody responses following bolus and trickle infections in both BALB/c and C57Bl/6 mice over the course of infection.
Despite an increase in total serum IgE over the course of infection in both mouse strains (BALB/c maximum levels on day 14 post infection with 559 ng/ml, SD +/- 94 compared to naïve levels at 68 ng/ml, SD +/- 58 and C57Bl/6 maximum levels on day 21 post-infection with 624 ng/ml, SD +/- 445 compared to naïve levels of 40 ng/ml, SD +/- 21), there was no difference in levels between bolus- and trickle-infected groups at any post-infection time point (Fig. 3A & B).
We also measured an increase in parasite specific IgG1 over time, as has previously been reported [23]. In BALB/c mice, at both days 21 and 28 post-infection, titers were above 10^4 (Fig. 3C), while in C57Bl/6 mice, despite an increase in titers, these remained below 10^4 at all time points (Fig. 3D). However, again, no differences were observed between trickle and bolus infected animals in either strain, at any time point. Also, no detectable levels of H. polygyrus larval or adult parasite antigen specific IgE, IgG2c or IgA were observed in the serum of bolus- or trickle-infected mice at any post-infection time point.
Local physiological, mucosal and immunological responses in the small intestine are not responsible for the improved protection observed in trickle-infected C57Bl/6 mice
Worm infections result in physiological changes in the small intestine that have been linked to promoting worm expulsion. These include increased intestinal smooth muscle contractility [61], and therefore, decreased intestinal transit time, as well as increased mucus production. We found that transit time (as measured by the time to pass dyed gavaged material) was reduced in infected C57Bl/6 animals by day 7 post infection, in both trickle and bolus infected groups by approximately 20% (Fig. 4A, left panel).
However, this decrease was only apparent in trickle infected animals by day 21 post-infection (approximate 12% difference between the trickle- and bolus-infected groups). At day 28 post-infection, the difference between bolus- and trickle-infected groups disappeared (approximate 3% difference between the trickle and bolus infected groups). We also measured mucus production indirectly through intestinal tissue weights. Infected mice had significantly higher intestinal tissue weights at 7 days (approximately 1.5 times naïve weights in both trickle- and bolus-infected mice) and 21 days (approximately 2.5 times naïve weights in both trickle- and bolus-infected mice) compared to naïve animals. Levels had returned to naive levels by 28 days post-infection (Fig. 4A, right panel). There was no difference between the bolus- and trickle-infected groups.
Next, we measured the levels of the Th2 cytokines IL-4 and IL-13 by ELISA in intestinal tissue, since both these cytokines have been associated with stimulating increased mucus production in the small intestine [62]. While levels of IL-13 were increased at day 21 and 28 post-infection, they were not different between the trickle- and bolus-infected groups (naïve: mean level 125 pg/ml, SD +/- 141, day 21: bolus mean level 558 pg/ml, SD +/- 268, trickle mean level 309 pg/ml, SD +/- 231, day 28: bolus mean level 363 pg/ml, SD +/- 254, trickle mean level 366 pg/ml, SD +/- 216, Fig. 4B, right panel).
Intestinal IL-4 levels were not increased at any point (approximately 40pg/ml) apart from at day 21, where they were significantly increased in the bolus-infected group (132 pg/ml, SD +/- 85) and different to the trickle-infected group. No levels of IL-5, IL-9 or IL-10 were detectable in intestinal tissue from control or infected mice.
Mucosal IgA levels, regulated by the cytokine IL-21 [63], are also increased in the presence of intestinal dwelling parasites [64]. We therefore measured IgA and IL-21 levels in the small intestine (Fig. 4C). Intestinal IgA levels did not differ between bolus- or trickle-infected mice over the course of infection (Fig. 4C, right panel). Interestingly, levels in both infection groups were similar to naïve levels (1292 pg/ml, SD +/- 924) at day 7 and 21 post-infection, but decreased at day 28 post-infection (bolus: 832 pg/ml, SD +/- 754; trickle: 555 pg/ml, SD +/- 93). Contrastingly, IL-21 levels were increased at day 21 and 28 post-infection (from 98 pg/ml, SD +/- 84 in naïve animals to approximately 950 pg/ml in infected animals, Fig. 4C, left panel). Like IgA levels, they were not different between trickle- and bolus-infected animals.
In summary, at the level of the small intestine, very few differences were observed between bolus- and trickle-infected animals, and these differences were not associated with reduced worm burden.
Unlike bolus infection, trickle infection results in stable granuloma numbers over the course of infection in C56BL/6 mice
Since few differences were recorded between trickle- and bolus-infected mice in the small intestine, we focused on the host/parasite interface, the intestinal granuloma.
Granulomas are a characteristic response to intestinal roundworms [14]. In response to H. polygyrus., they are round, opaque hard structures protruding from the small intestinal wall, easily identifiable and quantifiable using a dissection microscope (Fig. 5A). We measured these structures in both C56Bl/6 and BALB/c mice over the course of infection (days 7, 14, 21 and 28 days post-infection). We found that, BALB/c mice, both trickle- and bolus-infected, had consistently high granuloma numbers over the first three time points (granuloma numbers >45, Fig. 5B). At day 28 post-infection, numbers dropped below this. In C57Bl/6 mice, a similar pattern was observed for the bolus-infected mice: at days 7 and 14, granuloma numbers were high, and >45, at days 21 and 28, granuloma numbers were reduced to ∼17 (Fig. 5C). However, in trickle-infected C57Bl/6 mice, granulomas remained > 45 for the entire time course (Fig. 5C).
In C56BL/6 mice, granulomas from bolus and trickle-infected mice have different patterns of gene expression
Since protective immune responses to H. polygyrus infection are thought to be localized in the granuloma (92), we isolated all granulomas along the small intestine of bolus- and trickle-infected C57Bl/6 mice at days 7 and 21 post-infection. To identify the granuloma transcriptional profiles, we extracted the mRNA and quantified transcript levels using the nanostring myeloid innate immunity V2 panel. For naïve mice that do not have any granulomas along their small intestine, we harvested intestinal tissue from similar areas to those harvested in infected mice.
Using principal component (PC) analysis, we found that infection is responsible for the largest differences observed between the three groups (naïve, bolus-infected and trickle-infected) at both post-infection time points. Naïve and infected groups clustered distantly from each other, with infection explaining 62.1% of the variation (PC1) at day 7 post-infection and 90.5% of the variation (PC1) at day 21 (Fig. 6). Along the second principal component, granulomas from day 21 (but not day 7) infected mice clustered separately according to their infection mode (trickle or bolus, Fig. 6).
We found a number of genes that were differentially expressed between naïve intestinal tissue (naive group) and granuloma tissue (infected groups trickle and/or bolus) at day 7 and day 21 post-infection. At day 7, a total of 40 genes were upregulated in infected (NvT and NvB) vs. naïve animals: 9 are involved in cell migration, 11 in chemokine signalling, 6 in ECM and 5 in lymphocyte activation (Fig. 7A, Fig. S1). Over half of the upregulated genes (57%, 23/40) were also found to be upregulated at day 21 post-infection. Many of the identified differentially expressed (DE) genes have already been implicated in the immune response to helminths, such as Retnla, Arg1, Chil3, Chil4 and Retlnb [48],[65]. Fifteen of the 40 genes were highly upregulated (more than 16-fold); they are mostly involved in chemokine signaling, likely attracting macrophages and eosinophils to the granuloma. A total of 46 genes were downregulated in infected vs. naïve animals (NvT and NvB) at day 7 post-infection: 9 involved in cytokine signaling, 11 in growth factors, 8 in ECM, 6 in metabolism and 9 in lymphocyte activation. Seven of these genes (∼15%) were also downregulated at day 21 post-infection. Twenty four of the 46 genes were highly downregulated, and were associated with all functional categories. At day 21 post-infection, 75 genes were upregulated in infected vs. naïve (NvT and NvB) (Fig. 8A, Fig. S2): 12 involved in cell migration, 12 in chemokine signalling, 17 in ECM, 14 as growth factors, 10 in metabolism and 15 in lymphocyte activation.
Forty-one of these 75 genes were highly upregulated; these are involved in attracting macrophages and eosinophils (e.g. chil4, chil3, Serpine 1, ccl7, cxcl3, cxcr4), as well as ECM remodelling (e.g. mmp12, col1a2, cma1, Arg1). Twelve genes were downregulated in infected vs. naïve (NvT and NvB): 2 involved in cytokine signalling, 2 in growth factors, 2 in pathogen response and 3 in metabolism. None were highly downregulated: the most downregulated gene was nos2 at ∼7 times.
When looking at differentially expressed genes between trickle- and bolus-infected animals, we also found differences at both time points (day 7 and day 21). At day 7 post-infection, 10 genes were upregulated in trickle (vs. bolus) infected animals (Fig. 7A & 7B): 3 are TLR related, 3 are growth factors and 3 are involved in cytokine signalling. Of the 10 DE genes, 6 are highly upregulated: Selp, Tlr6, Bcl2, Tlr12, Flrt2, Il3ra and one is also upregulated at day 21: ccl3. At day 21 post-infection, 24 genes were upregulated in trickle (vs. bolus, Fig. 8A & 8B): 4 are TLR related, 7 are growth factors and 8 are involved in cytokine signaling. Of the 24 DE genes, 3 were highly upregulated: Adamts4 involved in ECM, Osm involved in cytokine production and Ndc80 involved in cell division. Twelve genes were upregulated in bolus (vs. trickle) at day 21 post-infection compared to none at day 7. Three of these genes were growth factors and three involved in cytokine signaling. No genes were highly upregulated, all were expressed below 7-fold and were associated with many different functions. The three most highly upregulated in the bolus infected animals (cxcr3, cxcr4, ptgdr) are involved in eosinophil recruitment and Th1 immunity.
Despite having many genes commonly expressed between the granulomas of trickle- and bolus-infected mice, different gene expression signatures were identified between these two modes of infection. Of the genes previously associated with immune responses to helminths, granulomas from day 21 trickle-infected animals had higher levels of Il13 (associated with worm expulsion [47],[54]), S100A8 and S100A9 (associated with neutrophil recruitment [66]), and Retlnb (associated with worm death [48]), as well as ccl3 and cxcl5 (inflammatory chemokines) expression. Surprisingly, we found no differences between the trickle and bolus infections in any genes linked to Fc receptor signalling, despite the importance of antibody-mediated worm killing by macrophages and/or eosinophils within granulomas [14].
IgG1 is the only antibody subtype to accumulate around encysted larvae in trickle infected C57Bl/6 mice
The ability to immobilize and/or kill parasitic worm larvae has been linked to antibody-mediated binding by myeloid cells in the granuloma [14]. IgG1, IgG2c, IgE and IgA antibody subtypes have all been linked to larval binding and/or damage to varying degrees. Since we found no differences in the expression of Fc receptor signalling genes within the granuloma between bolus and trickle-infected groups, we measured the accumulation of IgG1, IgG2c, IgE and IgA antibodies using immunofluorescence, both within the intestine and focusing on the host parasite interface within the granulomas (Fig. 9-10). First, we found that high levels of IgG1 in the serum at day 21 post-infection (Fig. 3D) correlated with the presence of IgG1 in granulomas in both bolus- and trickle-infected mice (Fig. S3A). At day 7, where antigen specific serum IgG1 could not be detected (Fig. 3D), IgG1 levels were minimal and/or absent (Fig. S3B).
To study the host/parasite interface, we set up a trickle (larvae) model (Fig. 9A), in which granulomas containing larvae (acute granulomas) and granulomas where larvae had either escaped or been killed (chronic granulomas) could be observed. Using this model, we found a high concentration of IgG1 at the host parasite interface at both day 14 and day 21 (Fig. 9B) post-infection in acute granulomas. However, in both trickle and bolus infected animals, levels of IgG1 were similar in chronic granulomas (Fig. 9C). We could not detect any IgG2c, IgE or IgA within acute granulomas of trickle (larvae) infected mice (Fig. 10A). This was despite observing IgG2c in intestinal tissue infected by C. rodentium (Fig. S4), IgE in the lamina propria of infected mice (Fig. 10B) and IgA in the lamina propria and Peyer’s patches of infected mice (Fig. 10C).
Overall, our data show that bolus and trickle infection result in similar systemic and tissue-wide immune responses. However, granuloma formation is distinct between the two types of infection, and correlate with different resistant phenotypes.
DISCUSSION
The formation of granulomas around tissue encysted H. polygyrus worms has been associated with resistance to infection [13]. The innate cells of the granuloma are thought to damage/kill worms in conjunction with antibodies and complement components [18],[27],[28]. We and others [13],[67] have shown that resistant BALB/c mice have more granulomas (Fig. 5) as well as higher levels of Th2 cytokines (Fig. 2) and parasite specific antibodies (Fig. 3) compared to susceptible C57Bl/6 mice. These differences are thought to contribute to the BALB/c resistance phenotype.
Using our trickle model, we observed a resistance phenotype in the C57Bl/6 mice as opposed to the susceptible phenotype of bolus infected animals (Fig. 1). The improved immune response we observed is likely due to the continuous stimulation of the host immune system by multiple low doses of larvae resulting in a greater number of more effective granulomas (better responsive myeloid cells and the presence of IgG1 antibodies, Fig. 8 & 9) and ultimately fewer adult worms (Fig. 1). Unlike differences observed between resistant BALB/c and susceptible C57Bl/6 mice, we found no significant differences between the systemic (serum antibody and spleen/MLN cytokine response, Fig. 2 & 3) or tissue-wide (intestinal tissue physiological, cytokine and antibody responses, Fig. 4) immune responses of bolus and trickle infected C57Bl/6 mice. Only when studying the granulomas themselves, at the host parasite interface, were important differences observed (Fig. 8 & 9).
Granulomas are made up of myeloid cells (mainly alternatively activated macrophages and eosinophils) and CD4+ T cells [7],[12] that accumulate around tissue encysted worms and immobilize them. As expected [14], we identified increased gene expression linked to myeloid cell recruitment, Th2 immunity and ECM deposition in granulomas from both trickle and bolus infected groups, at both time points (Fig. 7 & 8).
Overall, the increased levels of myeloid cell chemotactic gene expression within granulomas at day 7 in both bolus and trickle infections highlights the strong response to the tissue dwelling phase of the parasite. Ten of the highly upregulated genes at day 7 post-infection (NvT and NvB) have been previously associated with immune responses to helminth infections. And at day 21 post-infection, many of these same genes remained upregulated, with an increase in the expression of genes related to ECM remodelling (Fig. S1 & S2). Granulomas are novel structures created around tissue dwelling worms. They are not observed in naïve animals (Fig. 5A). As such, we observed the increased expression of genes associated with collagen production. At day 7 post-infection, col1a2 was upregulated with infection (NvT and NvB) as seen in other helminth infections [68],[69]. At day 21 post-infection, Col3a1, Col4a1 and Col15a1 were also upregulated. This is the first time a direct involvement of type III, type IV and type XV alpha 1 collagens have been linked to wound healing in an intestinal parasitic infection.
In the bolus-infected mice at day 7 post-infection, a number of genes were downregulated compared to trickle-infected animals (Fig. 7B). These genes are associated with growth factors, as well as TLR and cytokine signalling, in dendritic cells. This could be attributed to the initial parasite dose. Bolus-infected animals were given a dose of 200 worms, known to strongly downregulate the immune response as early as day 7 in the small intestine [70]. The trickle-infected animals received the same total dose, but split over 3 different time points which may have reduced the impact of the parasite. Interestingly, one of the downregulated genes on day 7 post-infection, Bcl2 (decreased by 33 fold compared to trickle-infected mice), has previously been associated with H. polygyrus infection. It was found to be increased in CD4+ T cells in the MLN, 2 weeks post-infection [71]. Similarly, we also found it was increased ∼4 fold at day 21 post-infection in the granulomas of both bolus- and trickle-infected animals (Fig. S2).
At day 21, the two most highly expressed genes in the trickle-infected mice are involved in tissue remodelling (Adamts4: 124 fold, Osm: 49 fold, [72]). Adamts12 (7 fold increase, (5)), Adamts3 (6 fold increase, (6)), Hdac5 (3 fold increase, [73]), Smad2 (2 fold increase [74]), and Socs 3 (2 fold increase, [75]), were also upregulated and are involved in tissue remodelling. The expression for all these genes is significantly higher compared to both day 21 and day 7 bolus-infected mice. As such, the difference observed is not due to the difference in ‘age’ of the granulomas, with the bolus day 21 granulomas being ‘older’ than the trickle day 21 granulomas. Interestingly, to our knowledge, the 2 most highly expressed genes (Adamts4 and Osm) have not been associated with helminth infection before. However, Osm is thought to stimulate adamts4 mediated degradation of the ECM [72], which may be an important process in the regulation of the granuloma structure.
We also found genes involved in the Th2 response upregulated in the day 21 post-infection trickle-infected animals. Il-13 was increased 10-fold in the trickle-vs. bolus-infected animals. While Igf1 was increased 4-fold. Igf1 is secreted by alternatively activated macrophages, and has been shown to promote worm expulsion. In Nippostrongylus brasiliensis infected mice, animals lacking Igf1 had higher worm counts than their wildtype counterparts [76].
Granulomas from trickle-infected animals have a different transcriptomic signature to bolus-infected animals, with a stronger Th2 and wound healing response. However, worm killing depends not only on activated myeloid cells, but also antibodies.
Antibodies are thought to be the bridge that allows granuloma granulocytes to kill/damage trapped parasites [14]. Antibodies are known to be required for protective immunity against H. polygyrus as B cell deficient mice [80] and mice lacking antibody production (JH or AID deficient, [23]), fail to eliminate adult worms. Of all the antibody isotypes, only IgG1 from immune mice has been shown to reduce adult worm burdens when administered alone [20], and protect mice from re-infection [22]. Passive transfer of IgG1 resulted in stunted worms, and in vitro assays demonstrated that IgG1 promotes peritoneal exudate cell attachment to the larval surface [20]. We found that IgG1 was the only detectable antibody isotype observed within the granulomas of infected mice (Fig. 9 & 10). No IgE or IgA accumulation was detected (Fig. 10A), despite their presence in the intestinal tissue (Fig. 10B & 10C), serum (Fig. 3B) and intestinal scrapings (Fig. 4C). Interestingly, others have shown that while IgE was found to play no role in parasite clearance in H. polygyrus infected animals, IgA contributes to limiting worm development [23]. This was done using IgA deficient animals, and, unlike in our study, levels of antibodies within granulomas were not assessed.
Trickle and bolus infected mice had similar levels of serum IgG1 (Fig. 3) and of IgG1 accumulation in their chronic granulomas (containing no tissue dwelling phases) at day 21 post-infection (SupFig. 3A). At this time point, tissue encysted worms had either escaped or been killed and digested within the granulomas in both infected groups. To detect whether IgG1 was playing a role in damaging/killing incoming worms in the trickle model (as has been suggested by [53]), we examined granulomas earlier (3/4 days as opposed to 10 days) after the last trickle dose in order to observe tissue dwelling worms within granulomas (Fig. 9A). In the granulomas containing worms, IgG1 was concentrated around the parasites (Fig. 8C). This observation helps explain how antibodies play a role in worm clearance. Two weeks following infection, both bolus- and trickle-infected mice produce a strong parasite specific IgG1 response (high serum titers, Fig. 3D). However, IgG1 antibodies can only play a role in ADCC and direct parasite killing/damage in trickle-infected mice when tissue encysted worms and IgG1 are present simultaneously in the granuloma. In bolus-infected animals, when high parasite specific IgG1 titers develop after two weeks of infection, larvae have already developed into adults and escaped from the granulomas into the intestinal lumen.
While the nanostring data obtained offer exciting new avenues of research into the regulation of granulomas, our study has some limitations. It is difficult to compare data from day 7 and day 21 post-infection with accuracy since they were obtained from mice of different sexes. Also, variation within groups was observed. Mouse 2 from the trickle group had much reduced gene expression than the other trickle-infected animals, which could in part be due to the low number of granulomas obtained from this animal.
Finally, it would be interesting to study gene expression profiles of acute/chronic granulomas with/without tissue encysted worms from trickle/bolus infected mice using single cell RNA sequencing to provide new and valuable details at a cellular level. This would allow the differentiation of genes that are important during acute and chronic stages of infection, and help identify other mediators involved in antibody accumulation and ADCC as well as wound healing/fibrosis within the granulomas during infection.
Figure S1: Eighty six genes are differentially expressed between granulomas from naïve and infected mice (NvT and NvB) 7 days post-infection. 6-8 week old female (7 days post-infection) C57Bl/6 mice were infected with 200 H. polygyrus according to the bolus and trickle infection regimes. We used the nanostring nCounter mouse myeloid innate immunity V2 panel to measure the expression profiles of 754 gene encoding mRNAs within the granulomas. Heatmap showing the relative expression of differentially expressed genes associated with the regulation of myeloid immune responses between naïve and infected mice (bolus and trickle).
Figure S2: Eighty eight genes are differentially expressed between granulomas from naïve and infected mice (NvT and NvB) 21 days post-infection. 6-8 week old male (21 days post-infection) C57Bl/6 mice were infected with 200 H. polygyrus according to the bolus and trickle infection regimes. We used the nanostring nCounter mouse myeloid innate immunity V2 panel to measure the expression profiles of 754 gene encoding mRNAs within the granulomas. Heatmap showing the relative expression of differentially expressed genes associated with the regulation of myeloid immune responses between naïve and infected mice (bolus and trickle).
Figure S3: IgG1 antibodies are present in the chronic granulomas of both bolus- and trickle-infected mice at day 21 post-infection. 6-8 week old female C57Bl/6 mice were infected with 200 H. polygyrus according to bolus and/or trickle regimes. Formalin fixed, paraffin-embedded 6 μm sections were obtained from small intestine swiss rolls. These sections were co-stained with anti-mouse IgG1 and DAPI. Whole sections were studied using the Thorlabs Tide whole-slide scanning microscope, (20x objective). Representative granulomas (white dashed line) from bolus-(top) and trickle-(bottom) infected mice at 21 days (A) and 7 days (B) post-infection. Antibody stain (left), isotype control stain (middle), and DAPI stain (right). For both groups, at day 7 post-infection, granulomas contained worms (acute granulomas) while at 21 days post-infection they did not (chronic granulomas). Experiments were performed at least twice (minimum of 5 mice per group) and are representative of a total of 5 (bolus, day 21 post-infection), 8 (trickle, day 21 post-infection), 8 (bolus, day 7 post-infection), and17 (trickle, day 7 post-infection), granulomas. Scale bar=100μm.
Figure S4: IgG2c is detectable in the intestinal tissue of C. rodentium infected mice. 6-8 week old female C57Bl/6 mice were infected with 5 x 10^8 cfu/ mouse C. rodentium DBS100 strain by gavage as a positive control for IgG2c staining. Formalin fixed, paraffin-embedded 6 μm sections were obtained from small intestinal swiss rolls. These sections were co-stained with antibodies to the IgG2c (left) as well as DAPI (right). Whole sections were visualized using the Thorlabs Tide whole-slide scanning microscope, (20x objective). Experiment was performed once (minimum of 5 mice per group). White arrows point to IgG2c staining within the lamina propria. Scale bar=100μm.
ACKNOWLEDGEMENTS
We wish to thank Drs. Vuk Cerovic and Simon Hirota for insightful comments. This works was funded through Dr Finney’s grants from the Canadian Foundation for Innovation and the Natural Sciences and Engineering Research Council of Canada (NSERC), Dr Nair’s grant from the National Institutes of Health/NIAID (NIH R01AI153195) as well as scholarships for Drs Anupama Ariyaratne (NSERC Create in Host Parasite Interactions, UCalgary FGS International Research Excellence Award and the Burroughs Wellcome Fund Travel Award), Stephen Pollo (NSERC), Mayara Luzzi (Mitacs Globalink), Joel Bowron (NSERC), Edina Szabo (UCalgary Eyes High).