Mice lacking NLRP3 show an altered composition of peritoneal immune cells and are resistant to T. crassiceps infection.

Since T. crassiceps dwells in the peritoneal cavity, waves of both innate and adaptive immune cells are constantly recruited to this site. First, we determined total cell counts and immune cell populations assayed by flow cytometry. Surprisingly, the number of resident peritoneal cells was different between WT and NLRP3−/− mice before infection. We harvested resident total peritoneal cells from WT animals and found an average of 2 × 106 peritoneal exudate cells (PECs), whereas NLRP3−/− mice presented a significant reduction ~ 30% with an average of 1.4 × 106 PECs (Fig. 1a). Moreover, when cells were immuno-typed, flow cytometry assays revealed that reduced populations included macrophages and eosinophils (Fig. 1b), although other immune cell types may also be underrepresented. This suggests an intrinsic defect in peritoneal homeostasis in female NLRP3−/− mice. Next, we proceeded to infect both WT and NLRP3−/− mice with metacestodes harvested from 8-week-infected BALB/c mice.

Fig. 1

NLRP3 is involved in maintaining homeostatic distribution of peritoneal cells. Age- and sex-matched uninfected WT and NLRP3−/− mice were sacrificed, and peritoneal cells were retrieved and counted to characterize immune cell populations. Panel a shows the differences found in terms of absolute numbers of peritoneal exudate cells (PECs) between WT and NLRP3−/− groups, and representative plots are shown, where gates of monocytic (R1) and granular cells (R2) are evident. Next, cells were prepared for staining and representative plots and percentages of macrophages (F480+ CD11b+) and eosinophils (F480lo Siglec F+) are shown in (b), also absolute numbers of both populations are depicted in graphs. Data shown are mean ± SEM from 3 experiments (n = 8 each group) where each point represents an individual mouse, and * indicates p < 0.05 when compared with unpaired student’s t test

The experimental infection with T. crassiceps results in chronic persistence of this parasite in its host peritoneal cavity. To achieve this latter, T. crassiceps progressively creates an environment dominated by a Th2 response [23, 24] and modulates antigen-presenting cells (APCs) like macrophages [25]. Therefore, both the emergence of Th2 cytokines and alternatively activated macrophages (AAMs) dictate the disease outcome. In this study, we observed that infected WT mice presented an increasing number of parasites determined at weeks 4 and 8 post-infection (121 ± 7.8 and 477 ± 14.5 parasites/mouse, respectively) (Fig. 2). Conversely, similarly infected NLRP3−/− mice were highly resistant as evidenced for significantly reduced numbers of parasites at 4 and 8 weeks of infection (30 ± 3.6 and 10 ± 3.5 parasites/mouse, respectively) (Fig. 2). Thus, endogenous NLRP3 turned out to be indispensable for T. crassiceps establishment.

Fig. 2

T. crassiceps parasites require the NLRP3 receptor to successfully colonize its host. Experimental mice from both groups (WT and NLRP3−/−) were infected with 20 metacestodes and susceptibility was assessed at 4 and 8 weeks upon infection. Average of parasites per mouse recovered from WT and NLRP3−/− mice is shown, where lack of NLRP3 turned mice into highly resistant. Values represent mean ± SEM from 3 experiments (n = 11–12 each group) where * is p < 0.05, and ** is p < 0.01 when unpaired student’s t test was performed

Upon infection, peritoneal cells progressively increased in samples obtained from WT animals, in contrast, peritoneal cells in NLRP3−/− individuals peaked at 4 weeks of infection and then declined reaching a substantial reduction on week 8 (Fig. 3a). Flow cytometry assays revealed that based on FSC and SSC parameters, peritoneal cells from infected animals could be gated on regions compatible with lymphocytes (R1), granular cells (R2), and monocytes/macrophages (R3) (Fig. 3a). Interestingly, when lymphocytes were characterized, it was evident that the presence of T. crassiceps in WT mice caused lower numbers of both CD4+ and CD8+ T cells (gated on CD3+ CD19−) (Fig. 3b). In contrast, we observed that concomitant to parasite clearance in NLRP3−/− mice, higher numbers of T lymphocytes remained in the peritoneal cavity (Fig. 3b). This is suggestive that, the well-known hypo-proliferative response caused by T. crassiceps as immune evasion strategy [26] is not occurring in NLRP3−/− mice.

Fig. 3

Immunophenotyping of peritoneal cells. At 4 and 8 weeks of infection, a peritoneal lavage and flow cytometry were carried out on experimental mice to identify immune cell populations. Collected cells were acquired in ATTUNE® cytometer where singlets and live cells were included in the analysis. Panel a shows the absolute numbers of peritoneal cells found at indicated times of infection and cell distribution based on FSC and SSC parameters. In b, cells selected from lymphocyte gate (R1) were further selected from T cell population CD3+CD19− cells (data not shown) and assayed for CD4 and CD8, where resistant NLRP3−/− mice presented an abundant population of CD4+ T cells. Also, absolute numbers of lymphocytes show a significant increase of T cells in peritoneal cavity of NLRP3−/− mice as compared to WT individuals. Panel c shows representative plots with indicated percentages and absolute numbers of eosinophils (F480lo Siglec F+) gated on R2 (granular cells), where a few eosinophils remained in peritoneal cavity of resistant NLRP3−/− mice. Interestingly, F480 fluorescence is decreased on cells from week 8 as compared to cells from week 4. In panel d, the percentage of macrophages (gated from F480+ cells) expressing both PDL1 and PDL2 in experimental groups is shown. WT mice progressively increased influx of suppressive macrophages, whereas NLRP3−/− mice presented only a few macrophages which mostly express PDL1 only. Representative dot plots and analysis of absolute numbers are from 3 experiments (n = 8–9 per group) mean ± SEM is shown, where * indicates p < 0.05 when unpaired student’s t test was carried out

Additionally, helminthic infections trigger eosinophil expansion, and we sought evidence of any significant changes within this population. Cells gated on R2 (granulocytes) demonstrated to be predominantly eosinophils (> 80% were F480lo Siglec F+) in infected animals as compared to a negligible eosinophil population found in uninfected mice from both experimental groups (Fig. 1b). A massive increase in peritoneal eosinophils was evident as early as 4 weeks of infection, where despite of similar percent of eosinophils (shown in plots), absolute numbers were higher in NLRP3−/− mice than those numbers found in Taenia-infected WT mice (Fig. 3c). Upon 8 weeks of infection an abundant population of eosinophils was observed within peritoneal cavity of WT mice, where parasites were still present, in contrast, in NLRP3−/− mice a significantly reduced population of eosinophils was found (Fig. 3c). An intriguing finding was the fact that eosinophils recovered from WT as well as NLRP3−/− mice (gated from the same region) at 8 weeks of infection displayed a reduced expression of the antigen F480 (Fig. 3c). We assayed these cells for Ly6G marker to rule out neutrophil presence, given neutrophils also express Siglec F [27], and they preserved eosinophil identity as only 1–2% of cells were Siglec F+ neutrophils as gauged by Ly6G co-expression (data not shown). Thus, whereas T. crassiceps infection in WT mice caused a progressive increase of peritoneal eosinophils, similarly infected NLRP3−/− counterparts, which cleared the infection, presented lower numbers of eosinophils.

We have shown that recruitment and re-programming of macrophages is central in allowing T. crassiceps growth in the peritoneal cavity of experimentally infected mice [28]. Thus, we aimed to identify PDL1+ PDL2+ macrophages in peritoneal cavity and observed that the majority of peritoneal macrophages harvested at 4 weeks of infection from infected WT mice were PDL1+ PDL2+ (70% from F480+ cells). This macrophage influx was further increased at 8 weeks of infection, where PDL1+ PDL2+ macrophages reached 85% of total macrophages; interestingly, other macrophage populations (i.e., double negative and PDL1 single positive) were also found, which suggested a steady infiltration of macrophages (Fig. 3d). Interestingly, NLRP3−/− mice showed mostly PDL1 single positive macrophages and a significantly reduced population of PDL1+ PDL2+ peritoneal macrophages (~ 7% from F480+ cells) at 4 weeks of infection. At 8 weeks of infection, NLRP3−/− mice exhibited a nearly absent population of these macrophages (Fig. 3d). Therefore, the inability to recruit and likely re-program macrophages might explain the resistance displayed by NLRP3−/− mice against T. crassiceps infection, which ultimately results in optimal T cell responses and reduced eosinophils resulting from parasite clearance.

NLRP3 is required for optimal Th2 polarization in T. crassiceps infection

We aimed to identify cytokine profiles in plasma samples to determine the influence of NLRP3 on cytokine production. First, we quantified IL-1β and IL-18 as the main products of canonical NLRP3 activation and observed that IL-1β showed an increasing pattern as disease progressed in WT animals reaching significantly higher levels in WT than in their NLRP3−/− counterparts only at 8 weeks of infection (2417 ± 752 pg/ml and 953 ± 242 pg/ml, respectively) (Fig. 4a). Interestingly, both experimental groups presented negligible circulating levels of IL-18 with a decreasing trend, and no differences were found between WT and NLRP3−/− mice infected with T. crassiceps during 8 weeks of follow-up (Fig. 4b). On the other hand, early reports showed that a switch from Th1 to Th2 immune response is imperative during T. crassiceps infection [24], and high systemic IL-4 levels are indicative of successful colonization by T. crassiceps parasites. Thus, we quantified IL-4 in plasma samples and found that T. crassiceps infection in WT mice elicited an increased and sustained IL-4 production as compared to uninfected animals (Fig. 4c), whereas in infected NLRP3−/− mice significantly reduced IL-4 levels were observed when compared to those found in WT animals (Fig. 4c), suggesting that NLRP3 is involved in Th2 polarization in T. crassiceps infection. Accordingly, IL-4 release was also found to be decreased in secondary lymphoid organs from infected NLRP3−/− mice when compared to IL-4 levels produced by mitogen-activated cells from WT animals (Fig. 4e and g).

Fig. 4

Mice lacking NLRP3 presented a diminished Th2 response. Cytokine output was quantified in plasma samples from experimental animals. Panels (a) and (b) show the main products of NLRP3 inflammasome, IL-1β and IL-18. Circulating levels of IL-4, as indicative of Th2 response, are depicted in (c), while panel (d) shows the differences of IL-15 levels, which is a survival factor for lymphocytes. Panels (e), (f), (g), and (h) show cytokine levels in supernatants from spleen and MLN cell culture supernatants as indicators of systemic and local responses. Data are from 3 experiments (n = 8–9 each group) where mean ± SEM is shown and * is p < 0.05

Additionally, it has been reported that IL-15 is a cytokine overproduced by NLRP3−/− mice induced with intestinal inflammation [4] and a likely role for IL-15 in T. crassiceps infection has not been explored. Hence, we determined circulating IL-15 levels and observed that during the chronic stage of disease (8 weeks of infection) resistant NLRP3−/− mice produced higher levels of this lymphocyte growth factor, as compared to those found in WT mice (1540 ± 240 pg/mL versus 853 ± 130 pg/ml, respectively) (Fig. 4d), indeed, IL-15 levels were higher in NLRP3−/− mice starting on week 6 (data not shown). Interestingly, mitogen-activated MLN cells showed a steady increased IL-15 levels starting on week 4 of infection in NLRP3−/− mice, whereas spleen levels were higher only at the chronic stage of infection (Fig. 4). These data suggest that IL-15 may also be a target for T. crassiceps, and its supression is part of the immune-regulatory strategy.

The presence of NLRP3 enables transcriptional activity of suppressive ligands in BMMs

Macrophages and DCs represent the most efficient type of APCs and helminth parasites have co-evolved to manipulate these innate cells in their mammalian hosts which ultimately also shape the adaptive immune response. In this context, T. crassiceps parasites create a peritoneal environment where macrophages are induced to express membrane-bound suppressive molecules such as PDL1 and PDL2 [25]. However, the signaling pathways controlling the expression of PDLs are not known, and we hypothesized that NLRP3 may be contributing to their expression. To test this, we incubated BMMs, given these are a major source of macrophages to replenish peritoneal cavity in response to infection [29], in the presence of TcES ± recombinant IL-4 for 48 h. We noticed a constitutive expression of pdl1 in macrophages growth from WT mice, whereas in the absence of NLRP3, no signal was detected (Fig. 5a). Furthermore, pdl1 expression was strongly induced by simultaneous exposure to TcES and IL-4 at 24 h of incubation, in contrast, a weak induction was observed in macrophages derived from NLRP3−/− animals (Fig. 5a). Interestingly, upon 24 h of incubation, pdl2 was only transcriptionally induced when using combination of TcES + IL-4 but not separately (Fig. 5a). At 48 h of incubation, pdl2 was significantly induced in WT macrophages co-treated with TcES + IL-4, but we did not observe transcripts of pdl2 when BMMs from NLRP3−/− mice were used. Further, resistin-like molecule alpha (relm α) expression is indicative of alternative activation in macrophages, since it is induced by IL-4 signaling [29], thus we determined whether relm α expression is dependent on NLRP3 as well. We found that BMMs exposed to TcES and TcES/IL-4 for 24 h presented a strong induction of relm α gene; however, in BMMs lacking NLRP3, the expression of relm α was poorly induced (Fig. 5a). Also, IL-4 alone induced relm α expression, and this was not dependent on NLRP3. At 48 h of incubation, similar to pdl1, relm α expression decreased as compared to 24 h, and only the combination TcES/IL-4 sustained its expression (Fig. 5a). Interestingly, when macrophages were exposed to IL-4, we observed transcriptional activity of pdl1 gene. Although this finding requires a deeper understanding, others have reported that IL-4 induced PDL1 in a particular subtype of AAMs, the multinucleated giant cells induced by biomaterials [30] as well as in CSF-1 primed bone marrow precursors from C57BL/6 [31]. Thus, in our in vitro system, NLRP3 turned out to be required to induce transcription of markers associated with suppressive/alternative programs in BMMs.

Fig. 5

NLRP3 transcriptionally activates a suppressor phenotype in macrophages. Mature macrophages were obtained by incubating bone marrow precursors in the presence of M-CSF, and these cells seeded at density of 0.5 × 106/well were incubated with TcES (50 µg) ± IL-4 (20 ng). RNA extraction was carried out at 24 and 48 h of incubation and transcripts for markers of suppressive ability (pdl1 and pdl2) and alternative activation (relm α) were determined, representative agarose gel image is shown in panel (a) (n = 4, from 1 experiment). Additionally, WT and NLRP3−/− mice were ip. injected with TcES (50 µg) plus IL-4 (20 ng), and peritoneal cells retrieved and stained with fluorochrome conjugated PDL1 and PDL2. Panel (b) shows representative plots with percentages of flow cytometry assays. Additionally, absolute numbers of indicated cells were retrieved from singlets and live cells analyzed from respective gates. Data are mean ± SEM from 2 experiments (n = 6) with similar results, where absolute numbers of indicated cells from WT and NLRP3−/− mice were compared with unpaired student’s t test

We also aimed to determine mRNA transcripts of the NLRP3 inflammasome components and interestingly, at 24-h post-incubation a low transcriptional activity was found for nlrp3 gene in response TcES (Fig. 5a). Whereas il-1b gene was virtually non-activated at time points assayed by either stimulus. Of note, transcription of il-18 gene was triggered by the stimuli used and required NLRP3, in contrast, caspase 1 (casp1) was induced when BMMs were exposed to TcES and NLRP3 was not required, whereas IL-4 strongly induced casp1 which intriguingly required NLRP3.

We decided to confirm these findings and injected (ip.) the most effective combination of stimuli (TcES/IL-4) to determine the presence of PDL1 and PDL2 on cell membrane of peritoneal macrophages harvested at 72 h. As seen in Fig. 5 b, a small fraction of resident F480+ cells were PDL1+, and no significant differences were observed between control WT (3% from F480+ cells) and NLRP3−/− mice (5% from F480+ cells). In response to TcES/IL-4 injection, increased numbers of F480+ cells were found in both groups as compared to control animals; interestingly, NLRP3−/− mice recruited more macrophages than WT animals (Fig. 5b). However, peritoneal macrophages expressed PDL1 but not PDL2, and comparable levels of PDL1+ macrophages were found in both WT and NLRP3−/− mice (18 ± 4 versus 20 ± 3, respectively), suggesting that, unlike BMMs where the activity of only one cell type is being tested, peritoneal cavity is highly complex and interactions with other cells might be affecting the induction of PDLs.

Protection against T. crassiceps can partly be acquired by co-housing with resistant mice

Intestinal microbiota shapes the immune response throughout lifetime and is therefore involved in maintaining homeostasis as well as in promoting pathological conditions when dysbiosis is developed. We tested whether co-housing of WT and NLRP3−/− could phenocopy the resistance against T. crassiceps. Interestingly, WT mice co-housed with NLRP3−/− mice displayed an enhanced resistance when compared to WT mice maintained in separated cages harboring their native microbiota. Parasite counts revealed that WT control animals harbored, at 8 weeks of infection, in average 271 ± 31 metacestodes, whereas in WT mice co-housed with NLRP3−/− animals, an average of 94 ± 33 metacestodes per mouse were recovered (Fig. 6a), showing that WT mice co-housed with NLRP3−/− mice for 4 weeks enhanced their resistance against T. crassiceps infection. On the other hand, control NLRP3−/− mice were consistently resistant as evidenced with low numbers of parasites found (18 ± 8 metacestodes), and we found a non-significant increase in number of parasites (34 ± 15 metacestodes) in NLRP3−/− mice co-housed with WT mice (Fig. 6a). These observations support a likely role for intestinal microbiota in contributing to protection against T. crassiceps infection.

Fig. 6

Co-housing reproduces the resistant profile found in NLRP3−/− mice. In order to explore whether intestinal microbiota played a role in immune response against T. crassiceps, mice were co-housed prior to infection. We compared parasite numbers at 8 weeks post-infection between control groups (animals with native microbiota maintained in separated cages) and similarly infected co-housed animals. As shown in panel a, parasite counts evidenced that resistance to T. crassiceps might be transferred via intestinal microbiota, where influx of peritoneal ells was also altered by cohousing, as shown by decreased absolute numbers of peritoneal cells in WT (co-housed) group. Moreover, changes in peritoneal populations PDL1+ and PDL2+ macrophages (b) as well as eosinophils (c) resulted from co-housing experiments. Interestingly, when IL-15 levels were determined in plasma samples (d), WT (co-housed) animals, which displayed enhanced resistance to T. crasscieps infection, when compared to WT control mice, also showed increased IL-15 production. Data are mean ± SEM from 2 experiments with similar results (n = 6 each group), where significant differences (p < 0.05) are indicated as * when groups were compared with ANOVA and multiple comparisons

We next looked for changes in the distribution of peritoneal cells (macrophages and eosinophils) resulting from co-housing experiments. WT control mice had an abundant population of macrophages PDL1+ PDL2+ relative to WT mice co-housed with NLRP3−/− littermates (Fig. 6b). As expected, control NLRP3−/− mice presented a remnant population of PDL1+ PDL2+ macrophages, which was negligibly affected in NLRP3−/− mice co-housed with WT individuals (Fig. 6b). On the other hand, infected WT control mice presented evident peritoneal eosinophilia and surprisingly, this population was found to be significantly reduced in WT animals co-housed with resistant NLRP3−/− mice (Fig. 6c). Both NLRP3−/− control as well as NLRP3−/− co-housed mice showed no significant changes in parasite numbers and suppressive macrophages; however, NLRP3−/− mice co-housed with WT animals presented increased numbers of eosinophils (Fig. 6c). These data suggest that early-life co-housing for 4 weeks is enough to partly transfer resistance against this non-intestinal tapeworm; interestingly, a dramatic impact of microbiota exchange is only observed in WT mice when compared to changes acquired by NLRP3−/− mice.