Programmed death ligand 2 regulates arginase induction and modifies Trypanosoma cruzi survival in macrophages during murine experimental infection

Abstract

The programmed death ligands 1 (PD-L1) and 2 (PD-L2) that bind to programmed death 1 (PD-1) have been involved in peripheral tolerance and in the immune escape mechanisms during chronic viral infections and cancer. However, there are no reports about the role of these molecules during Trypanosoma cruzi infection. We have studied the role of PD-L1 and PD-L2 in T. cruzi infection and their importance in arginase/inducible nitric oxide synthase (iNOS) balance in the immunomodulatory properties of macrophages (Mφ). In this work, we have demonstrated that expression of the PD-1/PD-L pathway is modified during T. cruzi infection on Mφs obtained from peritoneal cavity. The Mφs from T. cruzi-infected mice suppressed T-cell proliferation and this was restored when anti-PD-1 and anti-PD-L1 antibodies were added. Nevertheless, anti-PD-L2 antibody treatment did not re-establish T-cell proliferation. PD-L2 blockade on peritoneal cells from infected mice showed an increase in arginase expression and activity and a decrease in iNOS expression and in nitric oxide (NO) production. Additionally, interleukin-10 production increased whereas interferon-γ production was reduced. As a result, this microenvironment enhanced parasite proliferation. In contrast, PD-1 and PD-L1 blockage increased iNOS expression and NO production on peritoneal Mφs from T. cruzi-infected mice. Besides, PD-L2 knockout infected mice showed an increased in parasitaemia as well as in arginase activity, and a reduction in NO production. Taken together, our results demonstrate that PD-L2 is involved in the arginase/iNOS balance during T. cruzi infection having a protective role in the immune response against the parasite.

Introduction

Trypanosoma cruzi is an intracellular protozoan parasite that causes Chagas disease, a debilitating illness that affects Latin-American countries and results in cardiac complications and digestive disorders. During the early stages of infection, this parasite is found within macrophages (Mφs) and they may either inhibit parasite replication or provide a favourable environment in which it can multiply and be disseminated.¹ In addition, Mφs are important effector cells involved in various phases of the immune response, such as phagocytosis, antigen presentation and secretion of bioactive molecules.² The activation of Mφs, by T helper type 1 (Th1) cytokines or bacterial products such as lipopolysaccharide or CpG DNA, induces nitric oxide (NO) production. This provides a key defensive element in various infectious diseases. On the other hand, Mφs differentiated in the presence of Th2 cytokines enhance their capacity for endocytosis but do not exert enhanced killing functions towards microbes.^3–5 Furthermore, NO production is counteracted by the expression of arginase I (Arg I), an enzyme that competes with inducible nitric oxide synthase (iNOS) for l-arginine, leading to the production of l-ornithine and urea.^6–8

In addition, iNOS/Arg I balance is important during T. cruzi infection because a controlled response is necessary to eliminate the parasite and to avoid tissue damage. Cytokines, such as interferon-γ (IFN-γ), interleukin-12 (IL-12) and tumour necrosis factor-α are produced at high levels in response to the infection,^9–11 leading to an increase in iNOS expression in Mφs.^12–14 As a result, NO synthesis is enhanced, contributing to parasite killing and host survival.^13,15,16 However, the excessive production of NO has been proposed as one of the mechanisms that decreases the proliferative ability of T cells from infected mice and it has also been implicated in lymphocyte apoptosis.¹² Several studies have shown that Arg I expression and activity are induced by different parasites or parasite antigens controlling the collateral tissue damage.^17–25 However, Arg I produces polyamines, from l-arginine, which are essential for growth and differentiation of several parasites.^17–25

On the other hand, this enzyme suppresses the T-cell response^26,27 and this suppression might be mediated through different mechanisms. Among them, anti-inflammatory and immunosuppressive action of polyamines^28,29 and depletion of l-arginine in the T-cell environment, which leads to CD3ζ chain down-regulation.^20,27 Furthermore, it is currently recognized that l-arginine metabolism influences the relationship between innate and acquired immune responses.³⁰ Taken together, Mφ activation and Arg I/iNOS balance are essential for the disease outcome.^25,31

In addition, co-stimulatory molecules constitute an important mechanism that determines the T-cell response and they also affect the interplay between innate and acquired immunity.³² The ultimate fate of T cells, and hence of immune responses, appears to be mediated, at least in part, by the interplay between positive and negative T-cell co-stimulatory pathways.^33,34

In addition, new members of the B7 family have been identified. The most relevant are programmed death ligand 1 (PD-L1) and PD-L2,³⁵ which bind to the programmed death 1 (PD-1) receptor, which is expressed on activated T cells, B cells and myeloid cells.³⁶ Their interactions result in down-modulation of the T-cell response.^37,38 Besides, PD-L1 and PD-L2 exhibit distinct expression patterns and they are differentially up-regulated upon stimulation.^39,40 Whereas PD-L1 is expressed more broadly and is strongly induced by IFN-γ, PD-L2 is restricted to dendritic cells and activated Mφs and is induced by IL-4 and IL-13. Expression studies suggest that PD-L1 may have a preferential role in regulating Th1 responses, whereas PD-L2 may regulate Th2 responses.^41,42 Therefore, PD-L1 and PD-L2 functions may depend on the tissue and cytokine microenvironment.

In addition, several studies demonstrate that PD-L1 and PD-L2 have overlapping functions and support a role for the PD-1/PD-Ls pathway in down-regulating T-cell responses.³² Some reports suggest that PD-L1 and PD-L2 inhibit T-cell proliferation and cytokine production,⁴³ whereas others propose a co-stimulatory role for PD-L2. This molecule would enhance proliferation and effector functions through a PD-1-independent mechanism, suggesting the existence of an as yet unknown receptor.^44–48

In this work we have studied the role of PD-1 and its ligands, PD-L1 and PD-L2, during T. cruzi infection. We have demonstrated that PD-1, PD-L1 and PD-L2 are up-regulated on Mφs during infection. In addition, PD-L1 and PD-L2 modulated immunity to T. cruzi infection in different ways. Blockade of PD-1 and PD-L1, but not PD-L2, reverses the characteristic T-cell suppression seen during T. cruzi infection. However, blocking PD-L2, but not PD-1 or PD-L1, induces Mφs to up-regulate Arg I. This change in Mφ phenotype is associated with an increase in susceptibility to infection following PD-L2 blocking or in PD-L2 knockout (KO) mice.

Materials and methods

Animals

Female BALB/c mice, 6–8 weeks old, were obtained from the Comisión Nacional de Energía Atómica (CNEA; Buenos Aires, Argentina). PD-L2 KO mice were a gift from Dr Frank Housseau and Dr Drew Pardoll (Johns Hopkins University, Baltimore, MD).

Reagents

Antibodies and flow cytometry reagents, FITC-labelled anti-mouse CD3 monoclonal antibody (mAb), FITC-labelled anti-mouse CD11c mAb, FITC-labelled anti-mouse F4/80 mAb, FITC-labelled anti-mouse B220 mAb, and FITC-labelled anti-CD90.2 mAb were purchased from BD PharMingen (Palo Alto, CA). Phycoerythrin (PE) -labelled anti-mouse PD-1 mAb, PE-labelled anti-mouse PD-L1 mAb, PE-labelled anti-mouse PD-L2 mAb, purified anti-mouse PD-1 mAb, purified anti-mouse PD-L1 mAb and purified anti-mouse PD-L2 mAb were purchased from eBioscience (San Diego, CA).

Parasites

BALB/c mice, 6–8 weeks old, were intraperitoneally infected with 1 × 10⁶ blood-derived T. cruzi Trypomastigote (Tp) forms from Tulahuén strain and were maintained through intraperitoneal inoculation every 11 days.

Model of Trypanosoma cruzi infection

Female BALB/c mice 6–8 weeks old were infected intraperitoneally with 500 blood-derived T. cruzi trypomastigote forms (Tulahuén strain) diluted in saline solution as described by Zuniga et al.⁴⁹ After different times post-infection (p.i.), mice were killed by CO₂ asphyxiation and peritoneal cells were obtained. Non-infected control normal littermates were processed in parallel. The studies were approved by the Institutional Review Board and Ethical Committee of the School of Chemical Sciences, National University of Córdoba, Argentina. For in vitro experiments, Tp forms were obtained from blood of acutely infected mice and were enriched. Briefly, mouse blood was centrifuged at 500 g for 10 min and then incubated for 2 hr at 37° in a humidified 5% CO₂ atmosphere to allow parasites rise and concentrate in the plasma. Then, plasma was centrifuged at 15600 g for 7 min. The pellet was washed twice with complete RPMI-1640 medium and parasites were counted. Finally, cells were infected at a 3 : 1 Tp : cell ratio.

For parasitaemia studies, BALB/c wild-type (WT) and PD-L2 KO mice were infected with 1 × 10³ Tps (Tulahuén strain) diluted in saline solution. Parasite number was quantified at different days p.i. in a Neubauer chamber.

Peritoneal cell cultures

Resident peritoneal cells from T. cruzi-infected or non-infected mice were obtained by several peritoneal washouts with completed RPMI-1640 supplemented with 10% fetal bovine serum (FBS), l-glutamine (2 mm) and gentamicin (40 g/ml). The cellular suspension was distributed at 1 ml/well in 24-well tissue culture plates or 500 μl/well in 48-well tissue culture plates and cultured for 48 hr at 37° in a humidified 5% CO₂ atmosphere. Cells were used to assay surface expression of lineage markers, PD-1, PD-L1 and PD-L2, arginase expression and activity and iNOS expression and the supernatants were collected to evaluate NO and cytokine production.

Measurement of arginase activity

Arginase activity was measured in cell lysates as previously described.⁵⁰ Peritoneal cells were plated at 0·5 million/well in 48-well tissue culture plates infected and treated with blocking antibodies anti-PD-1, anti-PD-L1 or anti-PD-L2 (5 μg/ml). Briefly, cells were lysed with 50 μl 0·1% Triton X-100 containing protease inhibitor cocktail (Sigma-Aldrich, St Louis, MO, USA). After that, the mixture was stirred for 30 min at room temperature. Then, lysates were incubated with 50 μl 10 mm MnCl₂ and 50 mm Tris–HCl to activate the enzyme by heating for 10 min at 56°. Arginine hydrolysis was carried out in Eppendorf tubes by the addition of 25 μl 0·5 m l-arginine, pH 9·7, at 37° for 45 min. The reaction was stopped by the addition of 400 μl of a mixture containing H₂SO₄ (96%)/H₃PO₄ (85%)/H₂O [1/3/7, volume (v)/v/v]. The urea concentration was measured at 540 nm after the addition of 25 μl α-isonitrosopropiophenone, ISPF, (dissolved in 100% ethanol) and heating at 100° for 45 min. After 10 min in the dark the optical density (OD) was determined in the microplate reader (BioRad, Hercules, CA, USA) using 200-μl aliquots in non-sterile micro-culture plates. A calibration curve was prepared with increasing amounts of urea between 1·5 and 30 μg and 400 μI of the acid mixture and 25 μl ISPF were added to 100 μl urea solution. The results are expressed as an arginase index (fold increase of arginase activity in samples above that of non-infected cells).

NO assay

Nitrite concentration in supernatants of peritoneal cell cultures was determined spectrophotometrically using Griess reagent.⁵¹ Peritoneal cells were plated at 1 million/well in 48-well tissue culture plates infected and treated with blocking antibodies (5 μg/ml). Supernatants were collected after 72 hr, mixed 1 : 1 with Griess reagent, and OD was measured at 540 nm using a microplate reader (BioRad). The nitrite concentration was determined using sodium nitrite as a standard.

Cytokine determination

Production of IL-10 and IFN-γ was determined in culture supernatants after 72 hr by capture ELISA using mAb pairs purchased from eBioscience. Briefly, ELISA half-area plates were coated with 0·5–4 μg/ml anti-cytokine antibodies overnight at 4°. Plates were washed and blocked with 10% BSA for 1–2 hr at room temperature. Supernatants (25 μl) from different groups were added to the plates and incubated overnight at 4°. Plates were washed and incubated further with biotinylated anti-cytokine antibody for 1 hr at room temperature. After washing, avidin-peroxidase was added to the wells and the plates were incubated for a further 30 min. Plates were washed and developed using tetramethylbenzidine as substrate. The reaction was stopped with 0.5 m H₂SO₄. Plates were read at 450 nm in an ELISA plate-reader (BioRad). Standard curves were generated with recombinant cytokines (eBioscience).

Flow cytometry analysis

To examine PD-1/PD-Ls expression, peritoneal cells were removed from infected female BALB/c mice at different days p.i. Cells were washed with saline solution 2% FBS and pre-incubated with anti-mouse CD32/CD16 antibody for 20 min at 4° to block Fc receptors. Then, cells were incubated with FITC-labelled mAb against mouse CD3, CD11c, B220 or F4/80 and with PE-labelled mAb against mouse PD-1, PD-L1 or PD-L2 for 30 min at 4°. Cells were washed twice with saline solution with 2% FBS, and stored at 4° in the dark until analysis. This analysis was carried out in a FACS flow cytometer (FACS Canto II; BD Biosciences, San Jose, CA, USA). Results were analysed using facs-diva software (BD Biosciences).

Immunoblot analysis

To examine Arg I and iNOS expression, peritoneal cells were removed from infected female BALB/c mice at different days p.i. These cells were plated at 1 million/well in 48-well tissue culture plates infected and treated with blocking antibodies (5 μg/ml). Cells were washed after 48 hr and lysed for 30 min at 4° in radioimmunoprecipitation assay buffer [1% Triton X-100 (v/v), 0·5% sodium deoxycholate (w/v), 0·1% SDS] containing protease inhibitor cocktail (Sigma-Aldrich). Cell debris was spun down at 15 600 g for 15 min. Precipitates were removed and aliquots of cell lysates were diluted in SDS sample buffer, boiled at 100° for 3 min, spun down, and applied to precast 10% acrylamide Tris–glycine gels at 40 g protein/lane and run at 150 V for 1 hr. Samples were transferred to nitrocellulose membrane (BioRad) at 100 V for 1 hr. Membranes were probed using rabbit anti-mouse Arg I polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA) and rabbit anti-mouse iNOS (NOS2) polyclonal antibody (BD Biosciences) at a 1 : 500 and 1 : 2000 dilutions, respectively, followed by peroxidase-conjugated anti-rabbit antibody (Sigma-Aldrich) at a 1 : 1000 dilution. Bands were visualized using a chemiluminescence reaction.

Lymphocyte proliferation assay

Splenocytes were prepared from naive mice, and enriched for CD90.2⁺ cells (90% by FACS analysis) using anti-FITC-coated magnetic beads (MACS; Miltenyi Biotec, Bergisch Gladbach, Germany) after incubation with FITC-conjugated anti-CD90.2 mAb. Peritoneal cells were enriched for F4/80⁺ Mφs (85% by FACS analysis) using anti-FITC-coated magnetic beads (MACS; Miltenyi Biotec) after incubation with FITC-conjugated anti-F4/80 mAb.

The T-cell proliferative response was evaluated after co-culturing CD90.2⁺ spleen cells (2 × 10⁵ cells/200 μl/well) with F4/80⁺ peritoneal cells in flat-bottom microwell tissue culture plates at different T-cell/ Mφ ratios (2 : 1, 5 : 1, 10 : 1 and 20 : 1) in the presence of 2.5 μg/ml concanavalin A (Con A; Sigma-Aldrich). The presence of naive Mφs increased proliferation of CD90.2⁺ T cells and the most effective Mφ-to-T-cell ratio was 1 : 10 (data not shown).

The PD-1/PD-Ls pathway blockade on the proliferative response was evaluated after co-culturing CD90.2⁺ spleen cells (2 × 10⁵ cells/200 μl/well) with infected or non-infected F4/80⁺ peritoneal cells in flat-bottom microwell tissue culture plates treated with 5 μg/ml isotype control, anti-PD-1, anti-PD-L1 or anti-PD-L2 in the presence of 2.5 μg/ml Con A (Sigma-Aldrich). Cultures were maintained at 37° in a humidified 5% CO₂ atmosphere for 3 days and 0·5 μCi/well [methyl-³H] thymidine (Amersham, Chicago, IL) was added for the last 18 hr of culture. Cells were collected with a cell harvester (Cambridge Technology, Watertown, MA) and processed for standard liquid scintillation counting using a counter from Beckman Instruments (Fullerton, CA). Values are represented as counts per minute from triplicate wells.

Blockade of PD-1, PD-L1 and PD-L2 in peritoneal cell cultures

The T. cruzi-infected and non-infected peritoneal cells were obtained and single cell suspensions were prepared in RPMI-1640 supplemented as above. One million viable cells/well were placed into 24-well or 48-well culture plates (Greiner Bio-One, Frickenhausen, Germany) and 5 μg/ml isotype control, PD-1, PD-L1 or PD-L2 blocking antibodies was added. For in vitro experiments, mouse peritoneal cells were treated with blocking antibodies for 24 hr and then infected with Tp forms of T. cruzi at a 3 : 1 Tp : cell ratio. Cell cultures were maintained at 37° and 5% CO₂ for 72 hr.

Parasite growth

Peritoneal cells from female BALB/c mice (1 × 10⁶ to 1·5 × 10⁶) were cultured on slides in 24-well tissue culture plates and treated with isotype control, anti-PD-1, anti-PD-L1 and anti-PD-L2 blocking antibodies for 24 hr. Then, cells were infected with Tp at a 3 : 1 Tp : cell ratio and were cultured for 48 hr at 37° in a humidified 5% CO₂ atmosphere. After 24 hr, cells were washed to remove extracellular parasites. The number of parasites within Mφs, amastigotes, was determined by indirect immunofluorescence (IFI).²² The slides were taken 72 hr later; washed three times with PBS and fixed in 4% formol–PBS for 45 min. Then, they were treated with 1% Triton X-100 for 15 min. After washing with PBS, the slides were blocked with 1% PBS–BSA for 15 min. Subsequently, the slides were incubated overnight at 4° with positive Chagas serum diluted 1 : 50 to 1 : 100 with PBS. Slides were washed and FITC-labelled anti-human IgG was added in a 1 : 100 dilution in 1% PBS–BSA. After 1 hr, the slides were washed three times with PBS and were mounted on PBS-Glycerin. In addition, Tp that were released, 5 days p.i., in culture supernatants were quantified in a Neubauer chamber.

Statistical analysis

Statistical analyses were performed by a statistical one-way analysis of variance test to compare infected cells with non-infected and infected treated cells. Student's t-test was performed to compare WT and PD-L2 KO infected mice. The differences between data were considered statistically significant when P < 0·05.

Results

T. cruzi infection regulates PD-1, PD-L1 and PD-L2 expression on peritoneal macrophages, in vivo and in vitro

Recent studies indicate that the PD-1/PD-Ls pathway not only has an important role in the regulation of peripheral tolerance, but also in the control of the immune response against microorganisms that cause acute and chronic diseases. Given that its function during T. cruzi infection has not been explored, we evaluated PD-1, PD-L1 and PD-L2 expression on peritoneal Mφs of acute infected BALB/c mice by flow cytometry. We observed an increase in expression of PD-1 and its ligands on peritoneal Mφs as infection progressed as well as during in vitro infection (Fig. 1a,b). PD-L1 was also up-regulated on T cells but PD-1 and PD-L2 expression was not modified on T. cruzi-infected peritoneal T cells (Fig. 1c). Expression of PD-L1 was also increased on B cells and dendritic cells (data not shown).

Figure 1.

Programmed death 1 (PD-1), programmed death ligand 1 (PD-L1) and PD-L2 are up-regulated on peritoneal macrophages during Trypanosoma cruzi infection. (a) FACS analysis of PD-1, PD-L1 and PD-L2 expression on F4/80⁺ gated cells from non-infected (grey line) or T. cruzi-infected (dark line) mice. (b) Expression of PD-1/PD-Ls on F4/80⁺ gated cells in peritoneal cells exposed to Tps in vitro. Medium-treated cells (grey line) and Tps-treated cells (dark line) are shown. (c) FACS analysis of PD-1, PD-L1 and PD-L2 expression on CD3⁺ gated cells from non-infected (grey line) or T. cruzi-infected (dark line) mice. All data represented are the geometric mean of mean fluorescence intensities (MFI) from duplicate wells (above, non-infected; below, infected). Data are representative of at least three separate experiments.

Macrophages from T. cruzi-infected mice mediate T-cell suppression via PD-1/PD-L1 interaction

During the acute phase of T. cruzi infection, mice exhibit a suppressed response to parasite antigens and to mitogens.^52,53 Some studies have attributed to Mφs a decreased ability to proliferate observed in T cells from infected mice.⁵⁴ In addition, given that the PD-1/PD-Ls pathway generates an inhibitory signal, we evaluate whether these molecules are involved in immunosuppression.

Therefore, peritoneal Mφs from naive or T. cruzi-infected mice were co-cultured with naive CD90.2⁺ T cells purified from spleens of BALB/c mice. Antibodies specific for PD-1/PD-Ls were added to the co-cultures for 72 hr and proliferation was determined before the addition of [³H]thymidine. F4/80⁺ Mφs from naive mice favour Con A-stimulated naive mouse T-cell proliferation. However, F4/80⁺ Mφs from T. cruzi-infected mice suppress naive CD90.2⁺ T-cell proliferation (Fig. 2) as was shown previously.⁵⁴ T-cell proliferation was restored when anti-PD-1 or anti-PD-L1 antibodies were added. Nevertheless, PD-L2 blocking antibody treatment did not re-establish T-cell proliferation.

Figure 2.

Macrophages from Trypanosoma cruzi-infected mice inhibit T-cell proliferation through programmed death 1 (PD-1) and programmed death ligand 1 PD-L1 interaction. In vitro proliferation of CD90.2⁺ T cells co-cultured with macrophages from non-infected and T. cruzi-infected mice and treated with blocking anti-PD-1 (αPD-1), anti-PD-L1 (αPD-L1), anti-PD-L2 (αPD-L2) antibodies or isotype control (Rat IgG2a) in the presence of 2.5 μg/ml concanavalin A. Mean values and SEM are shown. *P < 0.05; **P < 0.01 versus Li T. Experiments shown are representative of two independent tests.

These data suggest that T. cruzi induces a suppressive phenotype of Mφs through the up-regulation of PD-L1, which inhibits activated CD90.2⁺ T cells.

PD-L2 blockade increases Arg I expression and activity

Several studies have shown that Arg I-mediated depletion of l-arginine leads to T-cell suppression.^26,27 To discover whether Arg I is involved in the immunosuppression observed in Fig. 2, we determined Arg I expression and activity in peritoneal cells treated with PD1 and PD-L blocking antibodies and infected in vitro with T. cruzi. Arg I expression and activity were up-regulated in infected cells compared with uninfected cells. Interestingly, Arg I expression and activity were enhanced in infected cells treated with anti-PD-L2 blocking antibody compared with infected cells. However, anti-PD-1 and anti-PD-L1 blocking antibodies did not modify Arg I in infected cells (Fig. 3a,b). Therefore, the increase in Arg I activity and expression observed in anti-PD-L2-treated cells might explain why anti-PD-L2 blocking antibody was not able to re-establish T-cell proliferation (Fig. 2).

Figure 3.

Programmed death ligand 2 (PD-L2) blockage induces an increase in Arg I expression and activity during Trypanosoma cruzi infection. (a) Arginase activity determined in extracts of peritoneal cell lysates from uninfected cells (−), cells infected in vitro (+) or cells infected in vitro and treated (5 μg/ml rat IgG2a, αPD-1, αPD-L1 or αPD-L2). (b) Western blot analysis for Arg I detection in extracts of the same peritoneal cells indicated in (a). Bars represent the fold increase value of activity related with normal non-infected cells. Experiments shown are representative of four independent tests. **P < 0.01 versus infected (+).

PD-1/PD-Ls pathway blockade modifies iNOS expression and NO production

Because l-arginine is the substrate for Arg I as well as for iNOS, we evaluated iNOS expression and NO production in peritoneal cells from infected mice or cells infected in vitro treated with blocking antibodies. Peritoneal cells from infected mice produce large amounts of NO compared with uninfected cells (Fig. 4a). In addition, the same effect was observed in peritoneal cells infected in vitro (Fig. 4c). Anti-PD-L2 blocking antibody treatment reduced NO production and iNOS expression in cells from infected mice (Fig. 4a,b) as well as in cells infected in vitro (Fig. 4c,d). On the other hand, we observed a slight increment in NO production in cells from infected mice treated with anti-PD-1 or anti-PD-L1.

Figure 4.

Programmed death ligand 2 (PD-L2) and PD-L1 have opposite effects in inducible nitric oxide synthase (iNOS) expression and nitric oxide (NO) production during Trypanosoma cruzi infection. (a) NO production was determined by Griess reaction in culture supernatants of peritoneal cells obtained from uninfected (−) or infected (+) mice. One group of infected cells was treated with blocking antibodies (rat IgG2a, αPD-1, αPD-L1 or αPD-L2). (b) Western blot analysis for iNOS detection in the same peritoneal cells indicated in (a). (c) NO production was determined by Griess reaction in culture supernatants from uninfected cells (−), cells infected in vitro (+) or cells infected in vitro and treated with blocking antibodies (rat IgG2a, αPD-1, αPD-L1 or αPD-L2). (d) Western blot analysis for iNOS detection in peritoneal cell lysates from the same peritoneal cells indicated in (c). Experiments shown are representative of three independent tests. *P < 0.05, **P < 0.01, ***P < 0.001 versus Infected (+).

Therefore, anti-PD-L2 blocking antibody shifts the Arg I/iNOS balance towards Arg I in T. cruzi-infected cells (Figs 3 and 4).

PD-L2 blockade modifies cytokine production

It has been demonstrated that T2-type cytokines shift l-arginine metabolism in Mφs towards Arg I, leading to polyamine biosynthesis. To investigate the influence of the PD-1/PD-Ls pathway in the cytokine profile, IL-10 and IFN-γ production were determined in infected cells treated with PD-1/PD-Ls blocking antibodies. PD-L2 treatment produced an increase in IL-10 production in peritoneal cell cultures from infected mice as well as in peritoneal cells infected in vitro (Fig 5a,b). However, IL-10 production did not change when anti-PD-1 and anti-PD-L1 antibodies were added (Fig. 5a,b). In addition, there was a decrease in IFN-γ levels in peritoneal cell cultures from infected mice when PD-L2 was blockaded (Fig. 5c). Therefore, PD-L2 blockade shifts the IL-10/IFN-γ balance to IL-10 production. However, no changes were observed in IFN-γ levels when peritoneal cells were treated with anti-PD-1 and anti-PD-L1 antibodies (Fig. 5c).

Figure 5.

Programmed death 1 (PD-1)/programmed death ligands (PD-Ls) regulate cytokine production during Trypanosoma cruzi infection. Interleukin-10 (IL-10) and interferon-γ (IFN-γ) production was assessed in culture supernatants by ELISA. (a) IL-10 levels were measured in peritoneal cell supernatants from non-infected and T. cruzi-infected mice treated with blocking antibodies. (b) IL-10 levels from non-infected and in vitro T. cruzi-infected cells treated with blocking antibodies. (c) IFN-γ was measured in cell supernatants from non-infected and T. cruzi-infected mice treated with blocking antibodies. Experiments shown are representative of three independent tests. *P < 0.05 versus Infected (+).

PD-L2 but not PD-1 and PD-L1 blockade favours T. cruzi growth

To evaluate if the PD-1/PD-Ls pathway could affect parasite survival we removed peritoneal cells from mice and treated them with anti-PD-1, anti-PD-L1 and anti-PD-L2 blocking antibodies. The growth of parasites in Mφs was evaluated by counting intracellular amastigotes by IFI. Cells were fixed, permeabilized and then blocked. After that, they were stained with Chagas disease patient serum and the secondary staining was then performing with FITC-labelled anti-human IgG.

The IFI assay showed an increase in parasite growth when cells from infected mice were treated with anti-PD-L2 antibodies (Fig. 6a). Moreover, the number of parasites released in culture supernatants when cultures remain for a longer period increased when PD-L2 was blockaded (Fig. 6b).These data correlate with the IFI assay.

Figure 6.

Programmed death ligand 2 (PD-L2) impairs Trypanosoma cruzi growth within macrophages. (a) Peritoneal cells from infected (+) mice 13 days post-infection (p.i.) were cultured for 72 hr. The T. cruzi-infected cells were treated with blocking antibodies and isotype control. Intracellular parasites were counted by indirect immunofluorescence assay. (b) The number of Tp was counted at the 5th day after infection in culture supernatants in a Neubauer chamber (c) In addition, normal non-infected peritoneal cells were treated with blocking antibodies and 24 hr later were infected with Tps at a Tp : cell 3 : 1 ratio. After 24 hr, cells were washed to remove the non-internalized parasites. After 48 hr, intracellular parasites were counted by indirect immunofluorescence assay. (d) A representative field for each group is shown. *P < 0.05 comparing treatment with anti-PD-1, anti-PD-L1, anti-PD-L2 blocking antibodies treatment or with isotype antibodies (rat IgG2a) versus T. cruzi-infected cells. Data are representative of two experiments and show mean of duplicate cultures.

Parasite growth was also favoured when peritoneal cells from non-infected mice were infected with T. cruzi in vitro and treated with anti-PD-L2 antibodies (Fig. 6c,d). Therefore, PD-L2 might be an important molecule involved in T. cruzi growth in Mφs.

PD-L2 lack conditions the immune response in favour of parasite

To confirm the relevance of PD-L2 in the immune response against T. cruzi, BALB/c WT and PD-L2^−/− KO mice were infected with 1 × 10³ Tps intraperitoneally. At different days p.i. the parasitaemia was measured; we observed an increase in parasitaemia over time in PD-L2 KO mice compared with WT mice (Fig. 7a). In addition, peritoneal cells from non-infected BALB/c WT and PD-L2 KO mice were removed and infected in vitro with Tps at a 1 : 3 peritoneal cell-to-parasite ratio. Interestingly, Arg I activity was enhanced and NO was diminished in infected peritoneal cell culture from PD-L2 KO mice (Fig. 7b,c). In addition, there was an increase in IL-10 and a decrease in IFN-γ in peritoneal cell cultures from PD-L2 KO infected mice compared with WT infected mice (Fig. 7d,e). These results confirm the importance of PD-L2 in the immune response against T. cruzi.

Figure 7.

Programmed death ligand 2 (PD-L2) lack conditions the immune response in favour of parasite. (a) Wild-type (WT; grey curve) and PD-L2 knockout (KO; black curve) mice were infected intraperitoneally with 1 × 10³ blood-derived Tps of Trypanosoma cruzi Tulahuén strain and parasitaemia was evaluated on the indicated days post-infection (p.i.). Standard deviations are shown. Six animals per group were analysed. (b) Arginase activity was determined in extracts of in vitro infected peritoneal cells from WT (grey bars) and PD-L2 KO (black bars) mice. (c) Nitric oxide (NO) production was determined by Griess reaction in culture supernatants of the same peritoneal cells indicated in (b). (d) Interleukin-10 (IL-10) and (e) interferon-γ (IFN-γ) production in culture supernatants of T. cruzi-infected cells from PD-L2 KO and WT mice. Experiments shown are representative of two independent tests. *P < 0.05, **P < 0.01, ***P < 0.001 versus WT mice.

Discussion

Immunosuppression during T. cruzi infection has been broadly documented in humans as well as in mice. Several studies have explored the molecular mechanism(s) involved: immunosuppressor cells,^54–58 immunosuppressor factors released by the parasite, decreased IL-2 production, an increase in NO production, or apoptosis^12,52,53,59 among others. However, the mechanism involved is still not clear. In the present study, we evaluated the role of new members of the B7 family, PD-L1 and PD-L2, during T. cruzi infection and their importance in Arg I/iNOS balance in the immunomodulatory properties of activated Mφs. We found that T. cruzi infection led to increased expression of PD-1 and its ligands on peritoneal Mφs as well as during in vitro infection. On the other hand, F4/80⁺ Mφs from T. cruzi-infected mice suppressed the proliferative response of naive CD90.2⁺ T cells to primary stimulation with Con A. The PD-L1 or PD-1 blockade significantly reduced the suppressive activity of T. cruzi-infected-Mφs, indicating that PD-L1 is directly involved in their suppressive activity. However, PD-L2 blockade was not able to restore the T-cell proliferation suppressed by T. cruzi-infected Mφs. Given that it has been demonstrated that PD-L2 KO mice show an increase in Th2 response,^47,48 we decided to evaluate if PD-L2 blockade was able to induce Arg I. Involvement of Arg I in the suppressive capacity of Mφs has been broadly demonstrated.^26,27,60 Our data showed that PD-L2 blockade in T. cruzi-infected Mφs induced Arg I activity and expression that might explain the immunosuppressive capacity of these Mφs. However, we did not see changes in Arg I expression and activity in cell cultures treated with PD-1 or with PD-L1 blocking antibodies. Therefore, in our T. cruzi infection model the immunosuppression may be directly mediated by PD-1/PD-L1 and indirectly by PD-L2 through Arg I regulation.

Moreover, supporting data demonstrated that Arg I plays a key role in T-cell suppression in non-healing leishmaniasis lesions.²⁶ Arg I, through the local depletion of l-arginine, impairs at the site of lesions the capacity of T cells to proliferate and to produce IFN-γ, which is required for parasite elimination.²⁶ However, during Schistosoma mansoni infection, Arg I from Mφs favours the recovery of the infection by inhibiting T CD4⁺ cells and the production of cytokines. The authors demonstrated that the primary suppressor mechanism was the depletion of arginine by Arg I from Mφs.⁶⁰ Here, we show that PD-L2 blockade as well as PD-L2 deficiency enhances Arg, leading to an increase in parasite proliferation.

In addition, Terrazas et al.³⁸ have shown that Taenia crassiceps-induced Mφs were able to suppress T-cell proliferation through PD-L1 and PD-L2 up-regulation on Mφs in an IL-10, IFN-γ, NO independent and cell-to-cell contact dependent manner. In addition, Schistosoma mansoni worms selectively up-regulate PD-L1 to reduce T-cell activation during early acute stages of infection before the subsequent emergence of egg-induced T-cell suppression in the chronic stages of infection.⁶¹ Recently, It was shown that IL-4-stimulated Mφs up-regulate PD-L2 and the T-cell suppression induced by these Mφs was restored by adding anti-PD-L2 blocking antibodies.⁶² Therefore, T-cell suppression could be mediated by PD-L1 or PD-L2, depending on the manner in which Mφs are stimulated.

On the other hand, the increase in Arg I in Mφs during infections as well as during cancer could be detrimental to the host⁶³ because Arg I shifts l-arginine metabolism towards polyamine synthesis, reducing NO production.³ The NO is necessary to control the replication and survival of T. cruzi as well as Leishmania parasites in Mφs.^{9,13,16,64,65} Here, we showed a reduction in NO production in T. cruzi-infected Mφs treated with anti-PD-L2 blocking antibody. In addition, this result correlates with cytokine production, as we observed an enhancement in IL-10 and a decrease in IFN-γ levels, shifting the balance to Arg I. As a result, the microenvironment favours T. cruzi growth when cells were treated with anti-PD-L2 mAb. Moreover, peritoneal cell cultures from PD-L2 KO mice exhibit enhanced Arg activity and IL-10 levels. In contrast, a decrease in nitrites and in IFN-γ production was observed. Therefore, PD-L2 KO infected mice showed a higher parasitaemia than WT-infected mice. Our work shows for the first time that PD-L2 modifies Arg/iNOS balance in favour of iNOS, consequently, it is a key element in the control of T. cruzi replication in Mφ.

According to our data, Huber et al.⁶² recently demonstrated that in vivo blockade of PD-L2 during Nippostrongylus brasiliensis infection caused an enhanced Th2 response in the lung. Therefore, because Arg I favours parasite growth, it might be possible that PD-L2 interacts with another unknown receptor, modulating Arg I and T. cruzi replication within Mφs. Moreover, Liang et al. showed that PD-L1 and PD-L2 present different roles in regulating the immune response to Leishmania mexicana. In the absence of PD-L1, parasitic load and the development of injuries are sharply reduced. By contrast, PD-L2 KO mice exhibit more severe disease.⁶⁶ To explain these findings, several studies propose that PD-L2 interacts with another, unknown, receptor different from PD-1, with stimulatory functions.^45–48 This would explain why PD-L2 blockade increased Arg I and IL-10 and decreased NO and IFN-γ levels.

Taken together, this work contributes to the knowledge of a new cellular mechanism involved in the control of T. cruzi infection. PD-L2 has a protective role by controlling Arg I/iNOS balance, regulating cytokine production and controlling parasite survival.

Disclosures

The authors have no financial conflict of interest.