Zoonotic intestinal helminths interact with the canine immune system by modulating T cell responses and preventing dendritic cell maturation

Parasite co-evolution alongside the mammalian immune system gave rise to several modulatory strategies by which they prevent exaggerated pathology and facilitate a longer worm survival. As little is known about the immunoregulatory potential of the zoonotic canine parasites Ancylostoma caninum and Toxocara canis in the natural host, the present study aimed to investigate whether their larval excretory-secretory (ES) products can modulate the canine immune system. We demonstrated TcES to increase the frequency of CD4+ Foxp3high T cells, while both AcES and TcES were associated with elevated Helios expression in Foxp3high lymphocytes. ES products were further capable of inducing IL-10 production by lymphocytes, which was mainly attributed to CD8+ T cells. ES treatment of PBMCs prior to mitogen stimulation inhibited polyclonal proliferation of CD4+ and CD8+ T cells. Moreover, monocyte-derived ES-pulsed dendritic cells reduced upregulation of MHC-II and CD80 in response to lipopolysaccharide. The data showed that regulation of the canine immune system by A. caninum and T. canis larvae comprises the modification of antigen-specific and polyclonal T cell responses and dendritic cell maturation.

AcES and TcES decrease CD25++ Foxp3 high lymphocytes. Analysing the co-expression of Foxp3 and CD25 showed ES products to decrease CD25++ Foxp3 high lymphocytes at 150 µg/mL (means: 150 µg/mL AcES 1.1% with p = 0.01, 150 µg/mL TcES 1.1% with p = 0.005, medium 8.7%; Fig. 1D). The decline in CD25++ Foxp3 high lymphocytes was much lower at 15 µg/mL (means: 15 µg/mL AcES 6.8% with p = 0.32, 15 µg/mL TcES 3.9% with p = 0.04). This effect was less distinct with loss of statistical significance when analysing the CD25+ Although single individuals show increased frequencies of Foxp3+ lymphocytes after treatment with AcES at 150 µg/mL (illustrated by the flow cytometric plots) this effect did not achieve statistical significance. In contrast, AcES at 15 µg/mL was associated with a slight decrease in Foxp3 expression in lymphocytes. (B) Three-colour flow cytometry revealed the TcES-associated increase in Foxp3 high lymphocytes to be associated with CD4+, CD4+ CD8+ double-positive and CD4-CD8-double-negative subsets, while this effect was lower in CD8+ T cells. Compared to TcES, treatment with AcES at 150 µg/mL was associated with a much lower elevation in Foxp3 high expression by lymphocytes and CD4+ CD8+ T cells. (C) Cultivating the cells in the presence of 150 µg/mL AcES or TcES induced marked increase in Helios expression by Foxp3 high lymphocytes. (D) Treatment of canine PBMCs with ES products was associated with decreased frequencies of CD25+ and CD25++ cells within Foxp3 high lymphocytes and this effect was most obvious at 150 µg/ mL. P-values were calculated according to paired t-test. Error bars represent standard error of mean (SEM). FSC = forward scatter, SSC = side scatter. Experiments were repeated six times at two different time points. subset in Foxp3 high lymphocytes (means: 150 µg/mL AcES 7.7% with p = 0.17, 15 µg/mL AcES 21.5% with p = 0.87, 150 µg/mL TcES 5.7% with p = 0.07, 15 µg/mL TcES 12.0% with p = 0.053, medium 20.4%).

ES-induced IL-10 expression is mainly related to CD8+ T cells.
To further characterise the phenotype of IL-10+ lymphocytes, we used three-colour flow cytometry for IL-10, CD4 and CD8 (Fig. 3A) Fig. 3B). Although the data suggest ES products to slightly diminish the frequencies of IL-10+ Foxp3+ cells in the absence of LPS stimulation, this effect did to gain statistical significance. For AcES at 150 µg/mL, however, this effect was more obvious in the presence of LPS (LPS: 15.6%; 150 µg/ mL AcES/LPS: 9.8%, p = 0.02).
Co-staining for IL-10 and IL-17 showed low frequencies of IL-10+ lymphocytes to express IL-17 (mean: 18.7%;  For AcES this effect was more obvious at 15 µg/mL. In contrast, TcES-associated IL-10 secretion was stronger at 150 µg/mL and was only detected when cells were co-stimulated with LPS. (B) For analysis of intracellular IL-10 staining, PBMCs were gated on lymphocytes and monocytes based on side (SSC) and forward scatter (FSC) morphology. This revealed (C) lymphocytes but not (D) monocytes to be responsible for ES-induced IL-10 expression and the effect was independent from stimulation with LPS. P-values were calculated according to Wilcoxon signed rank test (ELISA without LPS stimulation) or paired t-test (ELISA with LPS stimulation, intracellular cytokine staining). Error bars show standard error of mean (SEM). Experiments were repeated six times at two different time points.

Discussion
Modulating and suppressing the immune system is an important skill of many parasites by which they are able to influence the host's immune response 26 . However, little is known about the immunomodulatory pathways adopted by canine parasites during their co-evolution alongside the dog's immune system. We recently demonstrated that Foxp3+ lymphocytes are elevated in the intestinal mucosa of nematode-infected dogs in vivo and hypothesised a Treg-inducing potential for canine intestinal helminths in the natural host 41 . Therefore the present study aimed to investigate whether the zoonotic organisms A. caninum and T. canis mediate immunosuppression in dogs and whether this includes the induction of Tregs. Confirming our hypothesis, the results demonstrated CD4+ Foxp3 high T cells to be elevated in PBMCs after cultivation in the presence of TcES. This finding is in accordance with observations in mice demonstrating Foxp3 to be increased on both the mRNA and protein levels in T. canis-infected animals and interestingly this effect was enhanced when mice obtained TcES prior to infection 37 . Additionally, a Treg-inducing ability is also known for several murine helminths including Brugia malayi 42 , H. polygyrus 43 and Strongyloides ratti 44 . Tregs comprise a small Th subset with anti-inflammatory activity by which the immune system can maintain a homeostatic environment in the steady state and prevent exaggerated inflammation during immune responses on a cellular level 45,46 . This is of particularly high importance in the context of the intestinal mucosa as induction and expansion of Tregs is essential for establishing oral tolerance 46 . Considering the immunoregulatory potential of intestinal parasites, several studies hypothesised their therapeutic potential for chronic inflammatory disorders such as inflammatory bowel disease, which is hallmarked by a breakdown of intestinal homeostasis associated with a decline in mucosal Tregs 19,20,41 . Interestingly, TcES also caused increased Foxp3 expression in CD4+ CD8+ double-positive T cells. In dogs, this subset represents a heterogeneous population of activated T cells in the peripheral blood that can develop from CD4+ and CD8+ T cells 47 . Our observation that these cells express Foxp3 is in accordance to a previous study on canine Tregs in the peripheral blood and popliteal lymph node 48 . Although canine CD4+ CD8+ T cells are suggested to exert a regulatory function 49 , the targets of this mechanism still remain unknown. Therefore, additional studies are needed to further investigate the role of canine CD4+ CD8+ double-positive T cells in helminth infection.
Treatment with ES extracts was not associated with elevation of the CD25++ compartment of Foxp3 high lymphocytes which was in contrast to the effect of Concanavalin A (ConA). CD25 is classically linked to the phenotype of canine Tregs and the degree of CD25 expression in ConA-stimulated CD4+ T cells is correlated with Treg function suggesting CD25++ Foxp3 high T cells to be activated canine Tregs 48 . The results may therefore suggest ES products to induce canine Tregs in a less activated state. However, this finding may otherwise indicate that the phenotype of activated canine Tregs can differ depending on the stimulus (e.g., ConA versus parasitic antigen), which needs to be investigated in further studies.
Compared to TcES, treatment with AcES caused a lower increase in Foxp3 high cells in canine lymphocytes and CD4+ CD8+ double-positive T cells. Although this effect did not achieve statistical significance, it is partly similar to data from humans as the levels of circulating Tregs are elevated in patients infected with N. americanus, a human hookworm closely related to A. caninum 50 . In addition, we found AcES and TcES to cause marked upregulation of Helios within Foxp3 high lymphocytes. Helios is a member of the Ikaros transcription factor family that was initially thought to discriminate naturally occurring from peripherally induced Tregs 51 . Its expression has been also demonstrated in canine PBMCs and, based on the literature, the authors suggested that Helios might be a suitable marker for thymic-derived Tregs in dogs 48 . However, subsequent studies revised the hypothesis of Thornton et al. 51 by showing that Helios expression can be induced in Foxp3+ T cells 52 and by demonstrating Helios+ and Helios-cells within natural occurring Tregs 53 . Functionally, Helios has been linked to T cell activation and proliferation 54 , and its expression is necessary for a stable inhibitory Treg activity 55 . Although functional studies are needed to confirm the impact of Helios in dogs, its upregulation by canine Foxp3 high lymphocytes after treatment with ES products suggests A. caninum and T. canis larvae to activate canine Tregs and to improve their regulatory activity.
The mechanism by which parasites can raise the Treg frequency may encompass direct and indirect effects. For instance, Treg induction may be facilitated by a parasite-related homeostatic environment, which includes the shift to a pro-regulatory cytokine milieu 56 . Moreover, parasites can support Treg generation by modulating DCs, for example by preventing their maturation, which may then guide naïve T cells to differentiate into a regulatory subset. This has been illustrated for H. polygyrus or Echinococcus granulosus in which treatment of DCs with respective ES antigens elevate their Treg-inducing capacities 57,58 . The results presented here show both AcES and TcES to impair upregulation of MHC-II and CD80 after TLR4 stimulation, thus implicating A. caninum and T. canis to be capable of reducing DC maturation in dogs. However, further studies should verify whether DCs pulsed with AcES or TcES are indeed more competent in modulating canine Tregs. Interestingly, AcES seems to be more potent in reducing the MHC-II expression of moDCs than TcES. Despite the fact that this could implicate different modes of action by which AcES and TcES influence DCs, the effect may further be related to quantitative differences in the fractions of ES antigens responsible for this finding. In addition to the interaction with DCs, parasites can directly drive Treg differentiation by secreting homologues that mimic the function of host molecules. For instance, this has been demonstrated for ES products of H. polygyrus that induce expression of Foxp3 via a TGF-β-like signalling pathway 16 . The identification of Ac-daf-7 and Ac-dbl-1 in larval secretions of A. caninum, both of which are TGF-β-like ligands with homology to proteins of Caenorhabditis elegans, may implicate AcES to be capable of influencing T cells independent from DCs 59, 60 . Therefore, conditions required for Treg induction and/or modulation by AcES and TcES in dogs should be further addressed in future studies to better define whether this is dependent on or independent of the presence of DCs.
Confirming the immunoregulatory potential of AcES and TcES in the definitive host on the cytokine level, we show ES antigens to enhance IL-10 secretion of canine lymphocytes. This is in accordance with previous studies on the cytokine pattern induced by larval secretions of A. caninum and T. canis in humans and/or mice showing them to be dominated by anti-inflammatory molecules including IL-10 17,35,36 . Similarly, PBMCs of dogs infected with T. canis during pregnancy are reported to show higher secretion of IL-10 39 . IL-10 is one of the most important anti-inflammatory cytokines as reviewed by Sabat et al. 61 and its ablation induces severe failure in immune homeostasis causing inflammatory disorders such as colitis 62 . In the context of helminth infection, parasite-induced IL-10 secretion may enhance the homeostatic capacity of the host including elevated generation of Tregs 63 . In addition, IL-10 may also interfere with the maturation of antigen-presenting cells by preventing the transport of peptide-laden MHC-II molecules to the cell surface 64 . This could represent a possible mechanism by which AcES and TcES stop MHC-II upregulation by LPS-stimulated canine moDCs.
Three-colour flow cytometry revealed CD8+ T cells to be mainly responsible for ES-induced IL-10 expression. Although their marked increase in IL-10 suggests a regulatory function, it remains speculative whether they represent CD8+ Tregs. Being present in much lower frequencies than their CD4+ counterpart, CD8+ Foxp3+ Tregs have been described in humans, mice and dogs 48,65 . In the context of helminth infection, Cuéllar et al. showed murine splenic CD8+ cells, which are further characterised by the expression of Foxp3, CD25 and remarkable amounts of IL-10, to be responsible for suppressive effects of the recombinant metalloproteinase Ac-TMP-2 derived from adult A. caninum 15 . This implicates an important role of IL-10 in murine splenic CD8+ Tregs, which has been substantiated by others who found IL-10 to be required for the suppressive capacity of this cell population 66  Besides the analysis of antigen-specific effects, proliferation assays have been used to investigate whether parasites are capable of reducing polyclonal T cell activation. While this can be indirectly achieved by parasite-induced Tregs 15,16 , several studies provide evidence of worm-related factors by which they can directly interfere with lymphocyte proliferation in response to mitogens as demonstrated for Ascaris lumbricoides, H. polygyrus, N. americanus and Trichuris trichiura 50,68,69 . Similar effects have also been reported for A. caninum when adult worm extracts or living infective larvae were added to human mitogen-stimulated PBMCs 17 . Interestingly, the anti-proliferative effect was elevated in those patients infected with N. americanus 17 . In the present study, we demonstrate that larval secretions of A. caninum and T. canis inhibit the proliferation of canine PHA-stimulated T cells. With respect to the mechanism responsible for this effect, the anti-proliferative potential of adult worm secretions of N. americanus on human PBMCs is directly dependent on the presence of CD4+ CD25+ Foxp3+ Tregs 50 . Moreover, this effect may be further attributed to CD8+ Tregs as they are shown to outrun the suppressive capacity of CD4+ Tregs on lymphocyte proliferation 15,67 . To better define the circumstances essential for the inhibition of polyclonal immune cell activation during canine helminth infection, additional studies should investigate the contribution of soluble and cellular factors.
In summary, we demonstrated several effects by which A. caninum and T. canis can interfere with the canine immune system. These include the induction and/or modulation of Foxp3 high T cells, an increase in CD8+ IL-10+ T cells, inhibition of polyclonal T cell proliferation and the prevention of DC maturation. These effects are mediated by secretions of infective larvae, suggesting that the anti-inflammatory capacities already evolve early in the parasitic life-cycles, i.e. in the pre-adult stages. Remarkably, the present study further shows many of the immunomodulatory effects to be mediated by both nematodes. Considering the differences between A. caninum and T. canis, which are not only obvious on a phylogenetic level but also illustrated by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) of ES antigens or proteomic studies 29,31 , this may raise the question whether the findings presented here are related to common molecules shared by both organisms or attributed to distinct mechanisms that differ between both species. Therefore, future studies may focus on distinct ES fractions or single molecules to further increase the understanding of parasite-related immune regulation.

Material and Methods
Ethics statement. All  Animals. For generating nematode larvae, dogs were experimentally infected with T. canis or A. caninum, respectively, at the Institute for Parasitology (University of Veterinary Medicine Hannover). For isolating PBMCs, peripheral blood was obtained from dogs of various breeds, which were presented to the Unit of Reproductive Medicine (University of Veterinary Medicine Hannover) due to reasons unrelated to the present study and samples were only collected after the owner's agreement. In addition, peripheral blood was also obtained from Beagle dogs permanently housed in the Unit of Reproductive Medicine for reasons unrelated to the present study. All dogs were free from apparent disorders according to the clinical examination. Owners' dogs were routinely treated with anthelminthic drugs before blood sample collection. In the case of Beagles, absence of intestinal parasites was checked by parasitological examination of faecal samples.
Preparation of parasitic antigens. To obtain T. canis and A. caninum L3 for ES culture, eggs were obtained from faecal samples of experimentally infected dogs, these then being processed via the routine sedimentation-flotation technique. T. canis eggs were washed and embryonated in tap water at 25 °C for about 4 weeks. Larval hatching was carried out in vitro as described elsewhere 70 . For A. caninum, obtained eggs were incubated in a coproculture for about one week until development of L3. Larvae were exsheathed by incubation with sodium hypochlorite for 15 min at 37 °C.
A. caninum and T. canis L3 were washed approximately 10 times in sterile phosphate-buffered saline (PBS, pH 7.4) containing antibiotics (100 U/mL penicillin, 100 µg/mL streptomycin) followed by cultivation under standard conditions in RPMI-1640 supplemented with 1% glucose and antibiotics as follows: 100 U/mL penicillin, 100 µg/ mL streptomycin, 50 µg/mL gentamycin, 0.5 µg/mL amphotericin B. Parasites were cultured for up to 5 weeks and culture dishes were monitored daily for larval viability (at least 80% viable larvae to obtain ES products) and absence of bacterial contamination. Tissue culture supernatants were collected weekly and concentrated using a centrifugal filtration unit with a 3 kDa cut-off (Vivaspin ® Centrifugal Concentrators, Sartorius AG, Göttingen, Germany) according to the manufacturer's recommendations. Extracts were then filtered through a low-binding disposable 0.22 µm filter and stored at -80 °C for up to 6 months until use. Protein concentrations of ES extracts were estimated by measuring the absorbance at 280 nm with a NanoDrop ND-1000 Spectrophotometer and results were calculated using logistic regression analysis based on a bovine gamma globulin standard curve (see Supplementary Method S1). SDS-PAGE showed no major differences between batches of ES extracts. ES preparations were further found to have a mean endotoxin content of 0.52 EU/mL as determined by using the Pierce ™ LAL Chromogenic Endotoxin Quantitation Kit (Thermo Fischer Scientific, Waltham, Massachusetts, USA).
Cell culture. To analyse whether ES antigens are able to induce Tregs, PBMCs were cultured in 96 well microplates at 0.5 × 10 6 cells/well and stimulated with ES antigens at 15 and 150 µg/mL, respectively, or medium under standard culture conditions. ConA, which is known to induce canine Tregs in vitro 48 , was further used at 5 µg/mL as control. After 72 hours, cells were harvested and analysed by flow cytometry regarding their expression of Foxp3, CD4 and CD8, Foxp3 and Helios, as well as Foxp3 and CD25.
For investigating the potential of larval ES antigens to induce IL-10 production, PBMCs were cultured in 96 well microplates at 0.5 × 10 6 cells/well and pulsed with ES products at 15 and 150 µg/mL, respectively, following stimulation with LPS at 1 µg/mL (LPS-B5 Ultrapure, InvivoGen, Toulouse, France) or medium after 3 hours. ConA, LPS and medium were used as controls. After 24 hours, culture supernatants were collected and stored at -80 °C until further use.
To investigate IL-10-expressing cells on the single cell level, intracellular cytokine staining was used. For this purpose, PBMCs cultured in 96 well microplates at 0.5 × 10 6 cells/well were pulsed with AcES or TcES at 15 and 150 µg/mL, respectively. After 3 hours, cells were supplemented with 1 µg/mL LPS or medium and cultured for a further 18 hours. Brefeldin A was added during the final 12 hours at 5 µg/mL. Finally, PBMCs were stained for IL-10, CD4 and CD8, IL-10 and Foxp3, as well as IL-10 and IL-17, respectively.
To further analyse the potential of ES products to suppress mitogen-induced lymphocyte proliferation, PBMCs labelled with CFSE were pulsed with AcES or TcES at 15 and 150 µg/mL, respectively, for 3 hours following stimulation with PHA-L at 2.5 µg/mL in 96 well microplates at 0.5 × 10 6 cells/well. BSA at 15 and 150 µg/mL, respectively, was used as additional controls. After 4 days, cells were stained for CD4 and CD8 and analysed by flow cytometry.

MoDC maturation.
To evaluate whether ES antigens are capable of modulating TLR4-mediated maturation of antigen presenting cells, moDCs were generated as previously described with few modifications 71 . Briefly, PBMCs were cultured at 2 × 10 6 cells/mL in medium under standard culture conditions. After 24 hours, non-adherent cells were removed and adherent monocytes were cultured in the presence of 20 ng/mL canine recombinant IL-4 and 10 ng/mL canine recombinant GM-CSF for an additional 6 days. During this period, fully supplemented medium was changed after 3 days. Subsequently, moDCs were harvested by collecting non-adherent cells and pulsed with 150 µg/mL AcES or TcES for 24 hours in 96 well microplates (at least 20,000 cells/well). MoDCs were then stimulated with LPS at 1 µg/mL or medium for a further 24 hours. LPS was used as control. MoDCs were subsequently analysed regarding the expression of MHC-II and CD80 by flow cytometry. (clone FJK-16s, eBiosciences) and hamster anti-mouse/human Helios:AF488 (clone 22F6, BioLegend). For cell surface staining, cells were washed twice in PBS containing 1% BSA and 0.01% sodium azide (PBS-BSA) followed by incubation with primary antibodies on ice for 45 min. For staining of moDCs, PBS supplemented with 3% FCS, 1 mM EDTA and 0.01% sodium azide was used instead of PBS-BSA and Fc receptors were blocked by pre-incubation in 10% heat-inactivated dog serum for 15 min before adding primary antibodies. For staining canine CD11c, after washing steps, R-PE-labelled F(ab') 2 -fragments of goat anti-mouse IgG (Dianova, Hamburg, Germany) were applied for 30 min on ice. Cells were then washed twice in PBS-BSA and either resuspended in BD FACSFlow ™ Sheath Fluid (Becton Dickinson, Heidelberg, Germany) or further used for intracellular staining procedures.
For detecting intracellular antigens, cells were fixed and permeabilised using the Foxp3 / Transcription Factor Staining Buffer Set (eBiosciences) according to the manufacturer's instructions. Cells were pre-treated with heat-inactivated mouse, rat and dog serum (each 15%) to block Fc receptors followed by incubation with antibodies for 45 min at ambient temperature. Cells were then washed and resuspended in FACSFlow ™ Sheath Fluid.
Samples were analysed with a FACSCalibur TM flow cytometer (Becton Dickinson) within 2 hours by collecting at least 60,000 events if available. Isotype-matched control antibodies were used as negative controls in equal concentrations. Data were analysed with FCS Express (De Novo Software) and FlowJo (Tree Star). Foxp3-stained lymphocytes were differentiated into a Foxp3+ and Foxp3 high population according to isotype controls (0.5% and 0.1%, respectively). CD25+ and CD25++ gates were set based on the 5.0% and 2.0% region of medium control lymphocyte cultures, respectively, which is in accordance to a previous investigation of canine Tregs 48 . IL-10+ lymphocytes were defined based on isotype controls (0.5% gate). For CD4, CD8 and IL-17 staining, positive cells were identified based on morphology as they were obviously detectable in flow cytometry plots. IL-10 ELISA. The amount of IL-10 in culture supernatants of stimulated PBMCs was determined using sandwich ELISA (Canine IL-10 Quantikine ® ELISA kit, R&D Systems) according to the manufacturer's instructions.
Absorbance was measured at 450 nm and IL-10 concentrations were calculated using the 4-parameter logistic nonlinear regression model. Statistical analysis. Statistical analyses were performed using R version 3.3.0 (https://www.R-project.org/). Initially, data were checked for normal distribution by means of the Shapiro-Wilk test. Accordingly, the two-tailed Wilcoxon signed-rank test was used to evaluate differences in the amounts of IL-10 in supernatants of PBMC cultures without LPS stimulation (n = 6 dogs) and in expression of CD80 by moDCs (n = 6 dogs). The two-tailed paired t-test was performed to calculated differences in the percentage of Foxp3+/++ cells in the Treg stimulation assay (n = 6 dogs), to analyse differences in intracellular IL-10 expression (n = 6 dogs), to check for changes in IL-10 in supernatants of LPS-stimulated PBMCs (n = 4 dogs) and to evaluate differences in the expression of MHC-II (n = 6 dogs). One-way ANOVA followed by the Tukey's honest significant difference post-hoc test was applied to evaluate differences in CFSE MFI for the proliferation assay (n = 5 dogs). P-values ≤ 0.05 were considered as statistically significant.