IL-17RA-Signaling Modulates CD8+ T Cell Survival and Exhaustion During Trypanosoma cruzi Infection

The IL-17 family contributes to host defense against many intracellular pathogens by mechanisms that are not fully understood. CD8+ T lymphocytes are key elements against intracellular microbes, and their survival and ability to mount cytotoxic responses are orchestrated by several cytokines. Here, we demonstrated that IL-17RA-signaling cytokines sustain pathogen-specific CD8+ T cell immunity. The absence of IL-17RA and IL-17A/F during Trypanosoma cruzi infection resulted in increased tissue parasitism and reduced frequency of parasite-specific CD8+ T cells. Impaired IL-17RA-signaling in vivo increased apoptosis of parasite-specific CD8+ T cells, while in vitro recombinant IL-17 down-regulated the pro-apoptotic protein BAD and promoted the survival of activated CD8+ T cells. Phenotypic, functional, and transcriptomic profiling showed that T. cruzi-specific CD8+ T cells derived from IL-17RA-deficient mice presented features of cell dysfunction. PD-L1 blockade partially restored the magnitude of CD8+ T cell responses and parasite control in these mice. Adoptive transfer experiments established that IL-17RA-signaling is intrinsically required for the proper maintenance of functional effector CD8+ T cells. Altogether, our results identify IL-17RA and IL-17A as critical factors for sustaining CD8+ T cell immunity to T. cruzi.


INTRODUCTION
IL-17A and IL-17F were initially associated with the pathogenesis of autoinflammatory and autoimmune disorders (1). Nonetheless, the primary function of these cytokines is likely host protection against microbes. Indeed, mice deficient in IL-17A and/or IL-17F are highly susceptible to infection with a wide array of fungi and extracellular bacteria, as well as viruses and parasites (2,3). We previously demonstrated that the IL-17 family of cytokines plays a critical role in host survival by regulating exuberant inflammation and immunopathology during infection with the protozoan Trypanosoma cruzi (4,5). Additional data from our laboratory suggests that IL-17 plays additional protective roles against T. cruzi by modulating adaptive immunity.
The IL-17 family is comprised of 6 members, IL-17A to IL-17F, which have different cellular sources and expression patterns but often show overlapping activities. The cytokines signal through a receptor complex composed of at least two identical or different subunits of the IL-17R family (IL-17RA to IL-17RE). IL-17RA is the common signaling subunit used by at least four ligands: IL-17A, IL-17C, IL-17E, and IL-17F. IL-17 members regulate inflammation by recruiting and activating neutrophils, NK cells, and other cells of the innate immune system and by inducing several pro-inflammatory mediators (i.e., cytokines, chemokines, microbial peptides, and metalloproteinases). In addition, these cytokines play important roles during adaptive immune responses, including the modulation of germinal center reactions, as well as the regulation of Th1 and CD8+ cellular responses (6)(7)(8)(9).
Infection with T. cruzi is known to cause Chagas disease, the third most frequent parasitic disease worldwide that is endemic in Latin America and is increasing globally due to migratory flows. Disease progression, from asymptomatic to severe cardiac and digestive forms, is linked to parasite heterogeneity and variable host immune responses. CD8+ T cell mediated immunity is essential for parasite control throughout the various stages of infection, although it is not sufficient for complete parasite elimination (10,11). CD8+ T cell deficient mice are extremely susceptible to infection (12,13), and strategies that improve specific CD8+ T cell responses result in increased host protection (11,14,15). This finding, coupled with the fact that parasite persistence correlates with disease severity in human and experimental models of T. cruzi infection (16,17), have prompted efforts to better understand the mechanisms that promote anti-parasitic CD8+ T cell immunity.
The general features of protective CD8+ T cell responses (as defined with model viral and bacterial infections) consist of the generation and expansion of short-lived, highly functional effector populations with antimicrobial activities. After the pathogen has been cleared, most of the effector cells die and the few remaining cells differentiate into memory T cells, contributing to long-lived immunological protection (18). Persistently activated CD8+ T cells can become dysfunctional during certain chronic infections, which results in a hierarchical loss of effector functions and proliferation potential, as well as sustained expression of multiple inhibitory receptors. Severe exhaustion leads to pathogen-specific cells that are prone to deletion and have reduced ability to produce cytokines and degranulate. These exhausted CD8+ T cells fail to provide optimal protection and contribute to poor pathogen control (19). In addition to antigen-specific and co-stimulatory signals, which are the driving forces behind CD8+ T cell responses, various cytokines (i.e., IL-21, IL-2, IL-12) are required to provide the "third-signal" essential for protective CD8+ T cell immunity (20,21).
Herein, we demonstrate that IL-17RA signaling is required for the maintenance of robust, specific CD8+ T cell responses that contribute to host resistance to T. cruzi. We show that absence of IL-17RA signaling during T. cruzi infection alters the transcriptional program of effector CD8+ T cells, and consequently affects their survival, effector function, and exhaustion. Our findings are relevant for gaining a better understanding of the role of the IL-17 family in the orchestration of protective immunity against infections.

Parasites and Experimental Infection
Bloodstream trypomastigotes of the Tulahuén strain of T. cruzi were obtained from BALB/c mice infected 10 days earlier. For experimental infection mice were inoculated intraperitoneally with 0.2 ml PBS containing 5 × 10 3 trypomastigotes (usual dose) or doses of 500 and 5 × 10 4 trypomastigotes when indicated. For infection of transferred CD8α knockout mice, parasite dose was 1,000 trypomastigotes.

Quantification of IL-21
The concentration of IL-21 in the plasma of in plasma of infected WT and IL-17RA KO mice was assessed by ELISA using paired specific Abs (eBiosciences) according to standard protocols.

Evaluation of Proliferation and Apoptosis
T. cruzi-infected WT and IL-17RA KO mice were given BrdU (1 mg/ml, Sigma-Aldrich)/1% sucrose in the drinking water (carefully protected from light) ad libitum. After 5 days, mice were sacrificed and spleen mononuclear cells were collected and stained with surface antibodies. Incorporated BrdU was detected with a BrdU Flow kit according to the manufacturer's specifications (BD Biosciences).
Apoptosis was determined by Annexin V and 7AAD staining according to the manufacturer's specifications (BD Biosciences). Mitochondrial depolarization (ψ) was measured by FACS using 50 nM TMRE (Invitrogen) as described (24). When indicated, 7AAD or LIVE/DEAD Fixable Aqua dead cell stain was used in combination to identify live, early apoptotic and late apoptotic/necrotic cells.

Adoptive Cell Transfer
Competitive adoptive transfer experiments were performed by i.v. injection of recipient CD45.1+/CD45.2+ WT mice (F1 hybrid obtained by crossing CD45.1+ and CD45.2+ WT mice) or CD8α knockout mice with a mixture 1:1 of CD8+ T cells purified from spleen of CD45.1+ WT mice and CD45.2+ IL-17RA KO mice. Total 5 × 10 6 cells and 15 × 10 6 cells were injected in WT and CD8α knockout recipients, respectively. Recipient mice were immediately infected and frequency of the injected CD45.1+ WT and CD45.2+ IL-17RA KO cells within the total T cell population were determined in blood, spleen and liver at different days pi.
To address the relevance of CD8+ T cell-intrinsic IL-17RA signaling for T. cruzi infection progression, CD8α knockout hosts were injected i.v. with the same number (7 × 10 6 ) of CD8+ T cells purified from spleen of WT or IL-17RA KO mice and immediately infected. Transfer of equal number of CD8+ T cells in both mouse groups was evaluated 1 day after injection by analyzing the frequency of CD8+ T cells in blood. The progression of the infection was followed by the quantification of parasitemia and survival. Infected non-transferred CD8α knockout mice were used as control.

Treatment With Neutralizing Anti-IL-17A and PD-L1 Abs
To block the PD-1/PD-L1 pathway, infected WT and IL-17RA KO mice were injected i.p. with 200 µg rat anti-mouse PD-L1 antibody (10F.9G2, BioXcell) or total rat IgG isotype control (Jackson Research) every 3 days from 15 dpi until sacrifice at day 21 dpi. Infection matched WT mice were assayed in parallel as controls.
For the in vivo neutralization of IL-17A, infected WT mice were injected i.p. with 200 µg rat anti-mouse IL-17 antibody (17F3, BioXcell) or rat total IgG1 isotype control (MOPC-21, BioXcell) every 3 days from day 13 dpi until sacrifice at 21 dpi. Infection-matched IL-17RA KO mice were assayed in parallel as controls.

Real-Time Quantitative PCR and DNA Microarray
RNA from total spleen tissue was purified using TRIzol Reagent (Life Technologies) following manufacturer's instructions. Oligo-dT and a MMLV reverse transcriptase kit (Invitrogen) were used for cDNA synthesis. IL-17RA, IL-17RC, and IL-17RD transcripts were quantified by real-time quantitative PCR on an ABI PRISM 7700 Sequence Detector (Perkin-Elmer Applied Biosystems) with Applied Biosystems predesigned TaqMan Gene Expression Assays and reagents according to the manufacturer's instructions. The following probes were used: Mm00434214_m1 for IL-17RA, Mm00506606_m1 for IL-17RC and Mm00460340_m1 for IL-17RD. For each sample, the mRNA abundance was normalized to the amount of 18S rRNA (Mm03928990_g1) and is expressed as arbitrary units (AU).
For gene-expression analysis, CD8+ T cells were purified by FACS from the spleen of non-infected and 22-day infected WT and IL-17RA KO mice and lysed with TRIzol reagent. Total RNA was extracted with the RNAeasy Mini Kit (Qiagen). RNA quality was verified in an Agilent Bioanalyzer and measured with a Nanodrop 1000 (Thermo Scientific). cDNA was hybridized on Affymetrix Mouse Gene 2.1 ST arrays as described elsewhere.

Analysis of Microarray Data
The microarray data from this publication have been deposited to the GEO database (https://www.ncbi.nlm.nih.gov/geo/) and assigned the identifier: GSE104886. Gene expression data was  Bioinformatic analyses were performed with R software environment. Heatmaps were produced using Heatmapper (http://www.heatmapper.ca) and show gene expression of CD8+ T cells from WT and IL-17RA KO normalized by the average of the gene expression in noninfected counterparts, except specified. Gene Set Enrichment Analyses (GSEA) between the different spleen CD8 T cell samples were done using specific gene sets from the Molecular Signatures DB (MSigDB). Genes induced by the infection in WT and IL-17RA KO mice but showing significantly higher expression in IL-17RA KO mice were uploaded to Ingenuity Pahtway Analysis (Ingenuity R Systems, www.ingenuity.com) for the analysis of "Disease and bio-function, " "Canonical pathway, " and "Up-stream regulators." It was considered significantly activated (or inhibited) with an overlap p ≤ 0.05 and an IPA activation Z-score as defined under each specific analysis category on the IPA website.

Statistics
Statistical significance of comparisons of mean values was assessed as indicated by a two-tailed Student's t test, two way ANOVA followed by Bonferroni's posttest and Gehan-Breslow-Wilcoxon Test using GraphPad software. P < 0.05 was considered statistically significant.

IL-17RA and IL-17A/IL-17F Deficiencies Compromise Parasite Control and Reduce the Magnitude of the Specific CD8+ T Cell Response During T. cruzi Infection
We previously showed that IL-17RA-signaling and IL-17secreting B cells are required for host survival during T. cruzi infection by, at least in part, regulating innate immunity and inflammation (4,5). In addition, histological data from these initial studies suggested that infected IL-17RA knockout (KO) mice exhibited increased parasite levels in tissues, but similar levels of parasitemia, compared with those of their infected wildtype (WT) counterparts (4). To confirm this and to study possible additional protective mechanisms mediated by IL-17, we evaluated tissue parasitism in infected WT and IL-17RA KO mice. As determined by the amounts of parasite DNA quantified in the spleen, we observed that T. cruzi infected IL-17RA KO mice controlled tissue parasitism similar to WT mice during the first 2 weeks of infection ( Figure 1A). However, IL-17RA KO mice failed to reduce the parasite burden at later stages of infection, as attested by the increased parasite load in the spleen at 22 and 130 days post-infection (dpi) ( Figure 1A). Furthermore, increased parasite levels were also detected in other organs, such as liver and heart (Figures 1B,C).
Given that the aforementioned result suggested a deteriorated adaptive immune response, we focused on studying the role of IL-17RA in the development of parasite-specific CD8+ T cells. To this end, we examined the generation of CD8+ T cells specific for the immunodominant epitope TSKB20 (T. cruzi trans-sialidase amino acids 569-576 -ANYKFTLV-) (25) in infected IL-17RA KO and WT mice. We determined that the frequency of TSKB20-specific CD8+ T cells was reduced in the spleen, liver, and blood of infected IL-17RA KO mice at 20 dpi ( Figure 1D). Kinetics studies showed that the frequency and absolute numbers of TSKB20-specific CD8+ T cells in the spleen of infected IL-17RA KO mice was similar to that of infected WT control mice up to 14 dpi, but these values dropped dramatically soon after ( Figure 1E). Remarkably, total CD8+ T cell numbers at 20 dpi were also reduced in the spleen of infected IL-17RA KO mice ( Figure 1F). In agreement with the notion that IL-17A and IL-17F are the major IL-17RA-signaling cytokines that modulate immune responses, IL-17A/IL-17F double knockout (DKO) mice also showed increased tissue parasitism in spleen and liver (Supplementary Figure 1A) and reduced frequency of parasitespecific CD8+ T cells in comparison to WT control mice (Supplementary Figure 1B).
Considering the kinetics of the TSKB20-specific CD8+ T cell immunity observed in infected IL-17RA KO mice ( Figure 1E), along with our results showing that IL-17 production peaked around 14 dpi (4, 5), we speculated that IL-17RA signaling plays a role during the expansion and maintenance of parasitespecific CD8+ T cells after priming. To test this, we performed neutralization experiments in which infected WT mice were injected with anti-IL-17A antibodies (Abs) from 13 to 19 dpi ( Figure 1G). A blockade of IL-17 signaling, specifically during the expansion/maintenance phase, resulted in a significant decrease in the frequency of TSKB20-specific T cells ( Figure 1H). Indeed, mice treated with anti-IL-17 once the CD8+ T cell response was already established (from 13 to 19 dpi), showed a frequency of parasite-specific T cells comparable to that observed in infected IL-17RA KO mice that were devoid of IL-17 signaling throughout the entire infection period ( Figure 1I).

Absence of IL-17RA Signaling During T. cruzi Infection Reduces the Survival of Total and Specific CD8+ T Cells
Considering the conventional kinetics of T cell responses (18), it was conceivable that the abortive CD8+ T cell response observed in infected IL-17RA KO mice could be a consequence of reduced cell expansion and/or increased contraction. To address the role of IL-17RA in CD8+ T cell proliferation, we assessed the in vivo incorporation of BrdU between 15 and 20 dpi, as these time points directly preceded the dramatic decrease in the numbers of specific CD8+ T cells in infected IL-17RA KO mice. The results depicted in Figure 2A Figure 2A). We then tested a possible role of IL-17RA signaling in supporting CD8+ T cell survival. In comparison to their infected WT counterparts, total and TSKB20-specific CD8+ T cells from infected IL-17RA KO mice showed significantly higher frequencies of apoptotic cells as early as 10 dpi and also at later time points (20 dpi) ( Figure 2B). Similar results were obtained by analyzing Annexin V+/7AADcells (Supplementary Figure 2B). In the same direction, total and parasite-specific CD8+ T cells from infected WT mice treated with anti-IL-17A Abs ( Figure 1H) showed higher percentage of apoptotic cells than their counterparts from WT mice, and exhibited levels of apoptosis similar to that of infected IL-17RA KO mice (Supplementary Figure 2C).
Given the results described above, we evaluated whether cytokines that signal through IL-17RA were able to directly increase survival of activated CD8+ T cells. CD8+ T cells purified from the spleen of non-infected WT mice were activated for 24 h with plastic-coated anti-CD3 and anti-CD28 Abs in the presence of recombinant cytokines. The addition of IL-17A, but not IL-17F, IL-17C, or IL-17E, lead to a modest but significant reduction in the percentage of apoptotic activated CD8+ T cells ( Figure 2C). Of note, the direct anti-apoptotic effect of IL-17 in CD8+ T cells was comparable to that of IL-21, a well-recognized survival factor for this population (26). We also determined that CD8+ T cells purified from the spleen of IL-17RA KO mice and activated in vitro were more susceptible to cell death (apoptosis and necrosis) induced by camptothecin (CPT) challenge than CD8+ T cells from WT mice (Supplementary Figure 2D). As expected based on the lack of IL-17RA expression, IL-17 was not able to reduce activationor CPT-induced cell death in CD8+ T cells from IL-17RA KO mice. However, IL-17 also failed to reduce the cell death triggered by CPT in CD8+ T cells from WT mice. The differential effect of IL-17 in protecting CD8+ T cells from TCR-but not CPTinduced cell death may result from differences in the apoptotic signaling pathways triggered by these two treatments (27,28).
In an attempt to elucidate the mechanisms underlying the prosurvival effects of IL-17A against activation-induced apoptosis, we evaluated the expression of pro-and anti-apoptotic members of the Bcl-2 superfamily that are critically involved in the regulation of T cell death (29). We determined that IL-17A significantly downregulated the expression of Bad ( Figure 2D). Little to no changes were observed in other members such as Bcl-xL, Bcl-2, Bim, and Bax (Supplementary Figure 2E). Altogether, these results indicate that IL-17A may promote CD8+ T cell survival by down-regulating pro-apoptotic factors in the Bcl-2 family.
Absence of IL-17RA Perturbs the Transcriptional Program of CD8+ T Cells Activated During T. cruzi Infection To gain further insights into the role of IL-17RA signaling during T. cruzi infection, we determined the gene expression profiles of purified CD8+ T cells using Affymetrix microarrays. Using gene expression data normalized to the corresponding non-infected samples, we compared the transcriptional changes induced by infection in CD8+ T cells from IL-17RA KO mice vs. WT mice in a fold change/fold change plot. We identified 4287 genes that showed significant expression differences (greater than 1.2 fold; p < 0.05) and grouped these genes into different sets according to the differential expression patterns ( Figure 3A). Remarkably, 1,647 genes (pink, orange and red sets) were expressed at higher level in CD8+ T cells from IL-17RA KO infected mice in comparison to WT counterparts, whereas 2,640 genes (light green, green and dark sets) were expressed at lower levels. To further evaluate the biological characteristics of the transcriptional landscape of CD8+ T cells from infected IL-17RA KO vs. WT mice, we performed gene-enrichment set analysis (GSEA) for the genes that were differentially expressed. After conducting a broad analysis using gene ontology, as well as immunological and curated databases, we decided to focus on specific gene sets (Molecular Signatures DB-MSigDB) related to CD8+ T cell function and differentiation, such as GSEA26495 and GSEA41867. This supervised analysis showed that IL-17RA KO CD8+ T cells were significantly enriched in six gene sets in which the most enriched CD8+ T cell gene sets in the significance order (size of FWER P values) were PD1 hi vs. PD1 low (p < 0.001) and Naïve vs. PD1 low (p < 0.001) ( Figure 3B). These results denoted that CD8+ T cells derived from IL-17RA KO mice infected with T. cruzi showed a phenotype with characteristics of dysfunctional and, surprisingly, naïve cells.
We then analyzed the expression patterns of genes within the categories relevant for CD8+ T cell biology and IL-17RA signaling. Direct comparison of the gene expression profiles of CD8+ T cells from infected WT and IL-17RA KO mice after normalization of the date to gene expression in cell from non-infected mice revealed that absence of IL-17 expression resulted in important differencesin the transcriptional profile of CD8+ T cells ( Figure 3C). Indeed, genes encoding activation and effector cell markers, such as the KLRG family molecules CD11c and CD69, showed a completely opposite pattern of expression in CD8+ T cells from infected IL-17RA KO vs. WT mice. Furthermore, CD8+ T cells from infected IL-17RA KO mice showed clearly increased expression of genes encoding inhibitory and death receptors, including Ctla4, Havcr2 (Tim3), Pdcd1 (PD1), Tigit, and Fas. We also analyzed how the absence of IL-17RA signaling affected the expression of genes encoding transcription factors (TFs) that regulate CD8+ T cell fate. Of note, there was a group of TF-encoding genes whose expression was significantly higher in CD8+ T cells from infected IL-17RA KO mice than in those of their WT counterparts. This group included genes such as Foxo1, Irf4, and Foxo3 that are critical for the generation of effector CD8+ T cells (30,31). Also, we determined that CD8+ T cells from infected WT and IL-17RA mice showed differences in the expression of many genes encoding effector molecules, cytokine receptors, homing and migration molecules, and apoptosis mediators, thus further supporting the fact that absence of IL-17RA influences CD8+ T cell fate. Finally, analysis of the gene expression of molecules associated to IL-17RA signaling (6) evidenced significant differences between CD8+ T cells from infected WT and IL-17RA KO mice. Thus, CD8+ T cells from WT mice showed increased amounts of transcripts encoding MMP-13, CCL-20, and IL-6 among others while CD8+ T cells from IL-17RA KO mice showed higher amounts of transcripts encoding TNF, IKBKb, CXCL-2, S100a8, MMP-8, and others ( Figure 3D).
To broadly analyze the significance of the transcriptional differences between CD8+ T cells from infected IL-17RA KO mice and their WT counterparts, we performed an Ingenuity Pathway Analysis (IPA) on genes marked in red in Figure 3A.
The genes were up-regulated in infected CD8+ T cells from both groups of mice, but were expressed at significantly higher levels in CD8+ T cells from IL-17RA KO mice. The top 20 "Diseases and Bio-functions, " "Canonical pathways, " and "Upstream regulators" that were most significantly altered in CD8+ T cells from IL-17RA KO mice infected with T. cruzi are shown in Figures 3E-G. Interestingly, "Cell death, " "Apoptosis of Lymphocytes" bio-functions ( Figure 3E) and the "p53 Signaling" (Figure 3F) canonical pathway emerged as significantly upregulated. These results support the notion that the lack of IL-17RA signaling during T. cruzi infection significantly affected the survival and senescence status of CD8+ T cells, as well as various biological features. Finally, to elucidate which molecule/s were involved in the observed differences in gene expression, we performed an "upstream regulator analysis" (Figure 3G). Many modulators of the inflammatory responses, including the cytokines IL-2 and IL-6 and the NF-κB regulators CHUK and IKBKG, were significantly up-regulated in CD8+ T cells from infected IL-17RA KO mice. Conversely, the genes for MKL-2 and MKL-1, which interact with transcriptional regulator serum response factor, were down-regulated. Notably, IRF4, and to a lesser extent other TFs such as BATF, STAT1, and AHR, emerged as potential upstream regulators of the transcriptional program induced by T. cruzi infection in CD8+ T cells from IL-17RA KO mice.

Absence of IL-17RA Signaling During T. cruzi Infection Results in CD8+ T Cell Exhaustion
Given the results obtained from the microarray analysis, we performed a phenotypic characterization of the CD8+ T cells from the different experimental mouse groups. Analysis  A). Top 20 "Diseases and Bio-functions" (E), "Canonical pathways" (F), and "Up-stream regulators" (G) that were most significantly altered (P < 0.05 with Fischer's exact test) in KO mice are shown. The activation Z-score was calculated to predict activation (red), inhibition (blue), categories where no predictions can be made and unknown results (with Z-score close to 0). Categories significantly activated or inhibited according to Z-score are marked with a star. of CD62L vs. CD44 expression showed that upon T. cruzi infection, more than 80% of the total CD8+ T cells from the spleen of infected WT and IL-17RA KO mice acquired an activated/effector phenotype characterized by CD62L low CD44 hi expression ( Figure 4A). Within the TSKB20-specific population, the effector cells constituted more than 95% of the total number of cells. These results highlighted that CD8+ T cell activation triggered by T. cruzi infection is not compromised in the absence of IL-17RA. Indeed, total and parasite-specific CD8+ T cells from the infected IL-17RA KO mice had significantly elevated expression of prototypical activation markers such as CD69 and CD25 in comparison to that of CD8+ T cells from WT mice (Figure 4B), which corroborated the results from our microarray analysis.
We next determined the expression of inhibitory and death receptors in CD8+ T cells from WT and IL-17RA KO mice infected with T. cruzi. The total and TSKB20-specific CD8+ T cells from infected WT mice had an increased frequency of PD-1+ and CTLA-4+ cells, slightly higher TIGIT expression, and reduced frequency of BTLA+ cells compared to that of non-infected mice ( Figure 4C). Notably, total and TSKB20specific CD8+ T cells from infected IL-17RA KO mice showed significantly increased frequency of PD-1+, CTLA-4+ and BTLA+ cells as well as higher expression of TIGIT. Furthermore, these cells presented an apoptotic phenotype characterized by the expression of high levels of CD95/Fas and CD120b/TNF-R2 (Figure 4D). Total and TSKB20-specific CD8+ T cells from infected IL-17A/IL-17F DKO mice showed a profile of expression of inhibitory (PD-1) and death receptors (Fas/CD95) that was comparable to that of total and TSKB20specific CD8+ T cells from infected IL-17RA KO mice (Supplementary Figures 1C,D).
Further phenotypic characterization included effector molecules involved in cytotoxicity such as Granzyme A (GzmA), Granzyme B (GzmB), and Perforin-1 (Prf). As GzmB and Prf were barely detectable ex vivo (data not shown), the expression of the three cytotoxic molecules was determined in splenocytes from non-infected and infected WT and IL-17RA KO mice after 5 h of stimulation with PMA/ionomycin (Supplementary Figure 4A). In agreement with the microarray data, we determined that CD8+ T cells from infected IL-17RA KO mice had significantly reduced expression of GzmA in comparison to their counterparts from infected WT mice, whereas the expression levels of GzmB and Prf were similar between both groups ( Figure 4E).
To address whether the phenotypic features of CD8+ T cells elicited in absence of IL-17RA correlated with altered functionality, we compared the effector function of CD8+ T cells from infected WT and IL-17RA KO mice. To this end, we analyzed the mobilization of CD107a and secretion of effector cytokines upon specific and polyclonal activation in vitro. Stimulation of splenocytes from infected IL-17RA KO mice with the TSKB20 peptide resulted in a significant reduction of the frequency of CD8+ T cells with polyfunctional characteristics (i.e., surface CD107a expression and IFNγ and/or TNF production) in comparison to WT control mice (Figure 4F, left panel and Supplementary Figure 3). Although a poor response was anticipated due to the low frequency of TSKB20specific CD8+ T cells in infected IL-17RA KO mice, the antigenspecific effector response of CD8+ T cells was not significantly different from background, suggesting the presence of both quantitative and qualitative defects. Polyclonal stimulation with PMA/ionomycin increased the percentage of CD8+ T cells from infected IL-17RA KO mice that had a polyfunctional response, although the magnitude of this effector response was significantly lower than that of WT mice (Figure 4F, right panel and Supplementary Figure 3).
To determine if the higher parasite burden in infected IL-17RA KO mice may underlie the upregulation of inhibitory receptors in CD8+ T cells from these mice, we evaluated the phenotype of CD8+ T cells from WT mice infected with increasing parasite doses. Interestingly, increasing the load of T. cruzi did not diminish the frequency of parasite-specific CD8+ T cells, nor did it promote the upregulation of the inhibitory and death receptors evaluated (Supplementary Figures 4B-D).

Checkpoint Blockade Partially Restores
Parasite-Specific CD8+ T Cell Immunity and Enhances Parasite Control in Infected IL-17RA KO Mice CD8+ T cell exhaustion has been associated with poor microbial control during infection with various pathogens, as well as with tumor progression in cancer. In these settings, a checkpoint blockade has been shown to efficiently restore the magnitude and functionality of effector T cells and minimize damage from the insult (32). Furthermore, a PD-L1 blockade has been effective in preventing T cell apoptosis in murine and human sepsis models (33,34). Our results highlight that the absence of IL-17RA signaling during T. cruzi infection leads to poor CD8+ T cell effector function and parasite persistence. This prompted us to evaluate whether a checkpoint blockade could restore a resistance phenotype to infected IL-17RA KO mice. As PD-1 is highly expressed on parasite-specific CD8+ T cells from infected IL-17RA KO mice, we targeted this checkpoint receptor by injecting blocking anti-PD-L1 Abs at 15 and 18 dpi, before the contraction of a parasite-specific CD8+ T cell response ( Figure 5A). We determined that PD-L1 blockade in infected IL-17RA KO significantly augmented the magnitude of the TSKB20specific CD8+ T cell response by 21 dpi, particularly in the spleen and liver (Figure 5B). A similar response to treatment was observed in infected WT mice. Notably, the enhanced T. cruzi-specific CD8+ T cell immunity was correlated with a marked reduction in the levels of tissue parasitism ( Figure 5C). In addition, inhibition of PD-L1 also lead to a milder infection, as highlighted by the significantly reduced levels of several damage biomarkers, such as the transaminase aspartate aminotransferase (AST), creatine kinase total (CK), and myocardial band (CK-MB) ( Figure 5D). The rescue effects of the PD-L1 checkpoint blockade during T. cruzi infection, namely parasite persistence and tissue damage, were more impressive in IL-17RA KO mice than in WT mice, as they showed extremely high susceptibility to T. cruzi infection in the absence of treatment.

Intrinsic IL-17RA Signaling Modulates the Maintenance and Phenotype of CD8+ T Cells Activated During T. cruzi Infection
To understand if IL-17RA plays a direct role in the modulation of CD8+ T cell responses to T. cruzi, we first quantified the serum concentration of IL-21, a cytokine known to modulate CD8+ T cell immunity (35). Of note, we determined that in comparison to infected WT mice, infected IL-17RA KO mice displayed higher expression of IL-21 ( Figure 6A). This result highlights that increased expression of IL-21 is not sufficient to compensate for the lack of IL-17RA signaling via restoring CD8+ T cell responses during T. cruzi infection.
We next evaluated the possibility of a direct effect of IL-17RA signaling in the modulation of CD8+ T cell responses, given that IL-17RA is upregulated in response to cytokines that boost CD8+ T cell responses such as IL-21 (26,36). The IL-17RA subunit interacts with IL-17RC to form the functional receptor for both IL-17A and IL-17F [reviewed in Gaffen (6)]. Recently, also IL-17RD was shown to bind to IL-17RA to differentially regulate IL-17A-inducing signaling pathways (37,38). Therefore, we evaluated the mRNA levels of IL-17RA, IL-17RC, and IL-17RD by quantitative PCR. In agreement with published data, mRNA encoding IL-17RA were detected in splenic CD8+ T cells (Figure 6B). Transcripts encoding Il17rc and Il17rd were also detected in these cells, but at remarkably lower levels in comparison to Il17ra. Studies by flow cytometry confirmed the expression of IL-17RA on CD8+ T cells. Notably, IL-17RA expression varied depending on the CD8+ T cell subset, and was higher in memory cells defined as CD44 hi CD62L hi than in the effector (CD44 hi CD62L lo ) and naïve (CD44 lo CD62L hi ) subsets ( Figure 6C). T. cruzi infection upregulated IL-17RA expression in CD8+ T cells, to levels and kinetics specific for each subset (Figures 6C,D). We observed that IL-17RA expression is substantially upregulated in bulk memory CD8+ T cells during T. cruzi infection, but is only transiently increased in bulk effector CD8+ T cells at early time points ( Figure 6D). As expected given their effector nature, TSKB20-specific CD8+ T cells expressed IL-17RA at a level that was comparable to bulk effector cells.
To directly address the impact of intrinsic IL-17RA signaling in the development of CD8+ T cell responses during T. cruzi infection, we performed a series of adoptive transfer experiments. First, we transferred an equal number of congenically marked IL-17RA-deficient (CD45.2+) and WT (CD45.1+) CD8+ T cells into WT (CD45.1+/CD45.2+) recipient mice. Recipient mice were infected, and the presence of injected cells in the total and TSKB20-specific CD8+ T cell pools was examined at 20 dpi. As expected, no CD45.2+ KO or CD45.1+ WT CD8+ T cells were detected in infected F1 (CD45.1+/CD45.2+) WT hosts that were not injected (control mice). When analyzing the competitive adoptive transfer experiment, we observed that within the population of injected cells, IL-17RA KO CD8+ T cells were outcompeted by their WT counterparts in both total and parasite-specific CD8+ T cell subsets in all the organs examined, including blood and spleen ( Figure 6E). To overcome the possible limitations derived from the presence of endogenous WT CD8+ T cells in lymphoreplete hosts, we repeated the competitive adoptive transfer experiment in CD8α-/-mice. At 17 dpi, the frequency of IL-17RA KO cells was double that of WT cells within the total population of CD8+ T cells (Figure 6F), likely as result of increased homeostatic proliferation of polyclonal IL-17RA KO CD8+ T cells (Figure 2A). In contrast, WT cells significantly outnumbered IL-17RA KO cells within the TSKB20-specific CD8+ T cell pool, indicating that intrinsic expression of IL-17RA provides a maintenance advantage for parasite-specific CD8+ T cells. Indeed, the ratio between the percentage of WT CD8+ T cells within TSKB20-specific gate and the polyclonal total CD8+ T cells was significantly higher than that of IL-17RA KO CD8+ T cells (Figure 6F). Phenotypic evaluation revealed that IL-17RA KO CD8+ T cells exhibited higher expression of the inhibitory receptor PD-1 ( Figure 6G). Upon antigen specific stimulation with TSKB20, WT CD8+ T cells exhibited a significantly higher polyfunctional effector response in comparison to IL-17RA KO CD8+ T cell from the same host ( Figure 6H, left panel and Supplementary Figure 5). An identical result was observed after polyclonal stimulation (Figure 6H, right panel and Supplementary Figure 5).
Finally, we aimed to address the relevance of CD8+ T cell-intrinsic IL-17RA signaling for host protection during T. cruzi infection. To this end, we evaluated the ability to control parasite replication in CD8α-/-mice transferred with IL-17RA KO CD8+ T cells in comparison to CD8α-/-mice transferred with WT CD8+ T cells (Figure 6I). As a control, we determined that the frequency of circulating CD8+ T cells at day one post-injection was the same between both groups of mice ( Figure 6J). CD8α-/-mice were extremely susceptible to T. cruzi infection and displayed uncontrolled parasitemia at day 21, even with an infection dose of 100 trypomastigotes, which was 50 times less than the dose used in previous experiments. Furthermore, all these mice succumbed by day 25 post-infection (data not shown). Remarkably, the transfer of IL-17RA-sufficient WT CD8+ T cells, but not of IL-17RA-deficient KO CD8+ T cells, was able to enhance parasite control and significantly reduce parasitemia ( Figure 6K).

DISCUSSION
During the last few decades, many reports have delineated the immunological mechanisms underlying IL-17-mediated roles in host protection against infection as well as in inflammatory diseases (39,40). In this study, we uncover a novel mechanism by which IL-17RA signaling regulates CD8+ T cell responses and promotes protection against a protozoan infection. Our findings provide new elements to elucidate the cytokines and pathways leading to robust CD8+ T cell responses against infections with intracellular microbes and consequently, may have important implications for vaccine and therapy design.
The integral role of the IL-17 family in host defense against extracellular bacterial and fungal pathogens has been well established, and recent studies have implicated the family of genes in potentiating immunity against intracellular bacteria, viruses, and parasites (3). So far, the IL-17-dependent protective mechanisms were believed to be dependent on the activation of both non-immune and innate immune cells to sustain inflammation (8). Notably, the IL-17 family of proteins only indirectly promotes adaptive cellular immune responses (41,42). In this regard, there is evidence to suggest the role of IL-17 cytokines in sustaining specific CD8+ T cell responses that promote host resistance to Listeria infection and melanoma (43,44). In these settings, the IL-17-mediated induction of protective CD8+ T cells depended on enhanced recruitment of cross-presenting dendritic cells, although direct effects on CD8+ T cells were not ruled out. Very recently, IL-17 was shown to directly potentiate CD8+ T cell cytotoxicity against West Nile Virus infection (9). In line with this, we demonstrate that mice deficient in IL-17RA showed an abortive CD8+ T cell response during T. cruzi infection characterized by an early reduction in the number of parasite-specific CD8+ T cells. Concomitantly, infected IL-17RA KO mice showed poor control of the parasite in target tissues such as spleen, liver, and heart. By using different experimental approaches, including detailed phenotypic, functional and transcriptional profiling of CD8+ T cells from IL-17RA KO mice infected with T. cruzi, we demonstrated that CD8+ T cell-intrinsic IL-17RA signaling is required to sustain CD8+ T cell maintenance and pathogen-specific CD8+ T cell immunity, particularly during the expansion phase.
Whether one or more of the four cytokines that signal through IL-17RA is involved in these events is currently unknown. However, our results led us to postulate that IL-17A is the main cytokine involved in the IL-17RA-mediated effect of CD8+ T cells. This postulate is based on the marked similarities in the phenotype of the infected IL-17A/F DKO and IL-17RA KO mice and the significant consequences of the IL-17A neutralization in mice and IL-17A supplementation in CD8+ T cell cultures. Still, a possible role of other members of this family are not ruled out given that at least two other the IL-17RA signaling cytokines besides IL-17A (i.e., IL-17F and IL-17E/IL-25) are increased during T. cruzi infection (4). Regarding receptor expression, we determined that CD8+ T cells express the transcripts encoding not only IL-17RA but also IL-17RC and IL-17RD. Of note, we determined that Il17ra transcripts are present in significantly higher amounts than Il17rc and Il17rd. These findings are noteworthy as high levels of IL-17RA seem to be required for effective responses to IL-17 (45), and also because they also pose the question of whether IL-17RC or IL-17RD are effectively expressed on CD8+ T cells and required for IL-17 signal transduction. Although, the receptor complex formed by IL-17RA and IL-17RC is considered essential for IL-17 signal transduction (46), some studies reported that hematopoietic cell types (including T cells) show IL-17-dependent responses even in the absence of IL-17RC (47)(48)(49). It should be further established whether IL-17RD may interact with IL-17RA and replace IL-17RC to mediate the IL-17 signaling in these populations as previously reported (37). In the same direction, to dissect the precise signaling requirement for IL-17 signal transduction in CD8+ T cells may provide important information in terms of the IL-17 biology.
Besides a direct role of IL-17 cytokines, it is possible that a crosstalk of IL-17 with other cytokines that regulate CD8+ T cell fate such as IL-21 might be indirectly involved in the observed qualitative defects. In this regard, Hoft and colleagues described that TCR-transgenic CD4+ T cells specific for an immunodominant peptide of T. cruzi that were polarized in vitro into Th17 cells could potentiate CD8+ T cell immunity through a mechanism that is IL-21-dependent, but IL-17-independent (50). It remains to be established whether the in vitro generated Th17 cells are equivalent to those differentiated in vivo within the host during T. cruzi infection. Indeed, we determined that infected IL-17RA KO mice displayed higher concentrations of serum IL-21 than infected WT control mice, suggesting that the altered CD8+ T cell response observed in absence of IL-17RA signaling is unlikely to be a result of deficient IL-21 production. CD8+ T cells from infected IL-17RA KO mice showed increased apoptosis and a transcriptional landscape associated with canonical cell death and senescence pathways. In addition, these cells exhibited high expression of multiple inhibitory and death receptors that correlated with reduced ex vivo effector responses and a gene signature that positively correlated with those defined for dysfunctional PD-1 high CD8+ T cells (51,52). Altogether, these features partially resembled those of dysfunctional CD8+ T cells described in many chronic infections and cancer that are characterized by the progressive loss of T cell function, sustained expression of inhibitory receptors, and high susceptibility to deletion (53). Interestingly, we found that the functional CD8+ T cell response observed in T. cruzi-infected WT mice [this manuscript and (54)] was significantly impaired in the absence of IL-17RA signaling. In this regard, although it is conceivable that the increased tissue parasitism in IL-17RA KO mice favored CD8+ T cell exhaustion, infection of WT mice with increasing parasite loads does not induce expression of inhibitory receptors or deletion of parasite-specific CD8+ T cells, nor does it result in impaired CD8+ T cell cytotoxic effector function (55). As cytokines are essential for the regulation of exhaustion (20), we performed adoptive transfer experiments that supported the notion that IL-17 signaling influenced CD8+ T cell survival and function partially through cell intrinsic mechanisms. Interestingly, we determined that IL-17RA is expressed at different levels according to each particular CD8+ T cell subset. Specifically, IL-17RA expression was increased and sustained in memory cells, but was transiently up-regulated in activated effector cells at early stages of infection. These results, together with those showing that IL17RA downregulation is included in the gene signature of exhausted T cells (56,57), suggest that the IL-17/IL-17RA pathway modulates CD8+ T cell fate, thus preventing cell exhaustion in activated cells and sustaining long-lasting cytotoxic responses. In this regard, adoptive transfer of WT, but not of IL-17RA KO CD8+ T cells, promoted parasite control in immuno-deficient CD8α-/-mice. In line with this, we demonstrated that anti-PD-L1 Abs during the CD8+ T cell expansion phase reinvigorated the parasitespecific CD8+ T cell responses in infected IL-17RA KO mice, thus restoring robust parasite control and reducing pathology to levels comparable to infected WT controls. Remarkably, we observed that a short regimen of PD-L1 neutralization increased the frequency of parasite-specific CD8+ T cells by two-fold in both treated groups, although the magnitude of the response in infected IL-17RA KO mice remained lower than in WT control mice. Given the high levels of inhibitory receptors in CD8+ T cells elicited by T. cruzi infection in absence of IL-17RA, it is likely that the combination of different checkpoint Abs may have a stronger effect in these settings, as reported in cancer models (58).
The role of checkpoint inhibitors during T. cruzi infection, particularly the PD-1/PD-L1 axis, has been understudied thus far. One report demonstrated that genetic disruption of this inhibitory pathway increased the effector immune response to favor parasite control, but also promoted cardiac pathology and compromised host resistance to the infection (59). In this context, our data provide evidence that a PD-L1 checkpoint blockade at the CD8+ T cell expansion phase not only restored resistance to T. cruzi absence of IL-17RA signaling, but also boosted protective CD8+ T cell immunity during the natural infection. These findings are relevant to a growing area of research aimed at investigating the therapeutic potential of immune checkpoint pathways during chronic infections (60). In recent years, several reports have shown that immune cell exhaustion or dysfunction is a common finding in human Chagas disease, particularly in patients with severe symptoms (61)(62)(63)(64). There is certain consensus that even very low parasite loads in tissues may drive, in the long term of the human infection (two decades), the expression of inhibitory receptors and cell dysfunction in T cells. Consequently, this likely drives an immune modulatory mechanism and reduced tissue damage in the later stages of infection. Interestingly, many reports have described that patients with moderate to severe chronic chagasic cardiomyopathy have reduced production of IL-17, supporting a protective role of the cytokine against cardiac damage (65,66). Altogether, previous findings and those reported here support the hypothesis that elevated IL-17 production in the context of persistent parasite levels may prevent T cell dysfunction and the associated immune imbalance that cause chronic myocarditis. Given the relevance for the understanding of the immunopathology during human Chagas disease, this hypothesis merits further investigation in future studies.
The comparison of the transcriptome of CD8+ T cells from infected IL-17RA KO and WT mice supported the notion that IL-17RA signaling influences the transcriptional program of effector CD8+ T cells. These conclusions were based on the transcriptome of bulk CD8+ T cells and, consequently, may present some caveats as result of the heterogeneity of the sample and the "dilution" of the parasite-specific response. However, we determined that during T. cruzi infection most CD8+ T cells acquire an activated phenotype. Furthermore, total and TSKB20-specific CD8+ T cells from infected WT and IL-17RA KO mice exhibit similar expression of activation and differentiation markers and inhibitory and death receptors. Moreover, despite the fact that TSKB20-specific CD8+ T cells were the immunodominant population within effector cells, CD8+ T cells specific for other parasite peptides (i.e., TSKB18) were also detectable in infected WT and IL-17RA KO mice (data not shown). Accordingly, we consider that transcriptomic profiling of bulk CD8+ T cells activated during T. cruzi infection, though with some limitations, provides meaningful evidences about the role of IL-17RA signaling in dictating the transcriptional program of effector CD8+ T cells. Interestingly, the IPA analysis of the transcriptome of CD8+ T cells arising IL-17RA KO mice infected with T. cruzi highlighted IRF4 and, to a minor extent, BATF as potential upstream regulators of the IL-17RA phenotype. Notably, both TFs have been reported to be required for survival and differentiation of early effector CD8+ T cells (67)(68)(69)(70)(71)(72). IRF4 promotes CD8+ T cell expansion and differentiation (69), but it also induces the expression of molecules associated with an exhausted status such as Blimp-1, CTLA-4, and IL-10 (73)(74)(75). Furthermore, BATF expression may be triggered by PD-1 ligation and associated with CD8+ T cell exhaustion during chronic viral infection in humans and mice (76). Very recently, IRF4 and BATF have been shown to form a TCR-responsive transcriptional circuit that establishes and sustains T cell exhaustion during chronic viral infection (77). Therefore, it is conceivable that sustained elevated levels of both TFs could result in the induction of cell death and exhaustion program observed in CD8+ T cells from infected IL-17RA KO mice. The link between IL-17RA signaling, expression of BATF and IRF-4, and CD8+ T cell exhaustion deserves further investigation using genetic dissection approaches.
In conclusion, our results provide evidence for a novel IL-17RA-mediated mechanism that potentiates immunity to intracellular pathogens such as T. cruzi by improving adaptive CD8+ T cell responses. These findings, taken together with the report showing that IL-17 controls functional competence of NK cells during a fungal infection (78), suggest that the IL-17 family is likely important for both the innate and adaptive arms of the cytotoxic response against pathogens. Although it remains to be determined whether this mechanism can be applied to host resistance mediated by IL-17 cytokines in the context of other infections, these findings have two fundamental repercussions in the field of IL-17-mediated immune responses. The first is that the prolonged targeting of the IL-17 pathways as part of an anti-inflammatory treatment during autoimmune diseases or cancer may potentially have harmful effects due to defective cytotoxic responses. The second is the notion that promoting the production of IL-17 cytokines during infection or vaccination may help to elicit stronger cytotoxic responses to fight against microbes, and eventually, tumors.

AUTHOR CONTRIBUTIONS
JT designed and performed most of the experiments, analyzed data, and wrote, commented on the manuscript. CA, FF, and CR performed experiments, analyzed data and commented on the manuscript. MR, MA, and MG performed experiments and commented on the manuscript. NN performed microarray experiment. WR analyzed microarray data. EP participated in microarray design and analysis and provided funding. CM and AG participated in data analysis, commented on the manuscript and provided funding. EA supervised the research, designed experiments, wrote the manuscript, and provided funding.

ACKNOWLEDGMENTS
We thank MP. Abadie, MP. Crespo, F. Navarro, D. Lutti, V. Blanco, A. Romero, L. Gatica and G. Furlán for their excellent technical assistance. We thank Dr. Immo Prinz for providing IL-17A/IL-17F double knockout mice. We acknowledge the NIH Tetramer Core Facility for provision of the APClabeled TSKB20/Kb and TSKB18/Kb tetramers. We thank A. Rapinat and D. Gentien from Genomic Platform, from the translational research department from Institut Curie, Paris, for the transcriptome experiments.
This manuscript has been released as a pre-print at BioRxiv (79). and Annexin V+ (B) cells from the corresponding colored group. Bar graphs represent data as mean ± SD, N = 4 mice. P values calculated with two-tailed T test. (C) Representative histograms of TMRE staining in total (left) and TSKB20/Kb+ (right) CD8+ T cells from the spleen of infected WT mice treated with isotype control or anti-IL-17 as described in Figure 1G. Gray tinted histogram show staining in CD8+ T cells from non-infected WT mice.
Numbers indicate the frequency of TMRElow (apoptotic) cells from the corresponding colored group. Histograms are representative of one out of seven mice. Bar  Representative plots and analysis strategy of the frequency of CD45.1+ WT and CD45.2+ IL-17RA KO CD8+ T cells from the spleen of CD8α-/-mice adoptively transferred and infected as indicated for Figure 6E. The plots show a combination of three and two effector function including expression of CD107a, IFNγ, and/or TNF upon 5 h of culture with the indicated stimulation. Plots are representative of one out of four mice. Data are representative of two independent experiments.