Immunotherapy using anti-PD-1 and anti-PD-L1 in Leishmania amazonensis-infected BALB/c mice reduce parasite load

Leishmaniasis is a neglected disease, for which current treatment presents numerous issues. Leishmania amazonensis is the etiological agent of cutaneous and diffuse cutaneous leishmaniasis. The roles of the programmed death-1 (PD-1) receptor on lymphocytes and its ligand (PD-L1) on antigen-presenting cells have been well studied in tumor and other infection models; but little is known about their roles in non-healing cutaneous leishmaniasis. In this study, we observed that L. amazonensis induced PD-1 expression on both CD4+ and CD8+ T cells and PD-L1 on dendritic cells on BALB/c mice. We tested the therapeutic potential of anti-PD-1 and anti-PD-L1 monoclonal antibodies (MoAbs) against a non-healing L. amazonensis infection in BALB/c mice, and that anti-PD-1 and anti-PD-L1 treatment significantly increased IFN-γ-producing CD4+ and CD8+ T cells, respectively. Compared with infection controls, mice treated with anti-PD-1 and anti-PD-L1, but not anti-PD-L2, displayed bigger lesions with significantly lower parasite loads. Treatment did not affect anti-Leishmania antibody (IgM, IgG, IgG1 and IgG2a) or IL-10 production, but anti-PD-1 treatment reduced both IL-4 and TGF-β production. Together, our results highlight the therapeutic potential of an anti-PD-1-based treatment in promoting the reinvigoration of T cells for the control of parasite burden.

parasite load quantification. Parasite load was evaluated by limiting dilution assay as previously described 22 . Briefly, after the mice were euthanized, the infected paws were removed, weighed, individually macerated with a tissue mixer, the homogenate (1 ml/footpad) and submitted to serial dilution (diluted 1:4) into 96-well culture plates (final volume of 200 µl/well) and incubated at 26 °C for 15 days. The presence of promastigote cultures were examined via optical microscope (Olympus, Japan), and the last well containing promastigotes in the limiting dilution assay was recorded to calculate the parasite load.
Cell staining for flow cytometry. Lymph nodes were individually removed and macerated with a tissue mixer.

Dosage of immunoglobulins.
Soluble L. amazonensis antigen (LaAg) was obtained from stationary phase promastigotes, which were washed 3 times in PBS, freeze-thawed for 3 cycles, lyophilized, stored at −20 °C and reconstituted with PBS just prior to use. The 96-well plates were coated with LaAg (1 μg/well) overnight at 4 °C, blocked with PBS/5% milk/0.05% Tween 20 (Sigma-Aldrich) for 2 h, and washed 3 times with PBS/0.05% Tween 20. The mouse serum samples (1:250 diluted in PBS/5% milk/0.05% Tween 20) were added. The plates were incubated at room temperature for 1 h and washed with PBS/0.05% Tween 20. Anti-IgM or anti-IgG-HRP (Southern Biotech) were added (1:2000) for 1 h at room temperature. After washing, the color was developed with a TMB solution (Life Technologies) and stopped with 1 M HCl. Data analysis. Results are expressed as mean ± SEM with confidence level p ≤ 0.05. For lesion development analysis, a two-way ANOVA with a Bonferroni post-test was used. For multiple comparisons, a one-way ANOVA followed by Tukey pairing was performed. Paired t-test analysis was done as indicated in the figure legends. Data analysis was performed using GraphPad Prism ® 5.00 software.

L. amazonensis infection induces PD-1 expression on CD4 + and CD8 + t cells.
To study the induction of PD-1 by Leishmania amazonensis infection in mice as observed in humans 21 , we decided to evaluate the expression of PD-1 in lymph node cells of BALB/c infected mice. Our data showed an increase in the percentage ( Fig. 1(A,C)) and in the absolute numbers ( Fig. 1(B)) of PD-1 + CD4 + and CD8 + T cells in infected compared to uninfected BALB/c mice. We also evaluated the expression of PD-L1 on dendritic cells. We observed an increase in the percentage ( Fig. 2(A,B)) and numbers ( Fig. 2(C)) of PD-L1 + CD11c + cells in the infected compared to uninfected BALB/c mice. These results indicated that infected mice display similar profile observed in cutaneous diffuse leishmaniasis and allowed us to study these molecules using the BALB/c mice model.
Anti-PD-1 or anti-PD-L1 MoAb treatment reduces parasite loads without affecting lesion growth in mice. As we observed that PD-1 and PD-L1 was upregulated during L. amazonensis infection, we next assessed the effect of treatment with the anti-PD-1 and anti-PD-L1 blocking antibodies on the disease profile. First, L. amazonensis-infected BALB/c mice were treated with the individual MoAbs (100 μg each/mice) once weekly, beginning at 7 days post-infection, receiving a total of 6 doses. The footpad thickness was measured weekly, and the parasite load was determined after 49 days of treatment. We found that this therapy was not effective in modifying the lesion development profile (Supplemental Fig. 1(A) or the parasite load (Supplemental Fig. 1(B)) compared to control mice injected with PBS.
In the second treatment protocol, mice were given individual MoAbs twice weekly, beginning at 7 days post-infection and receiving a total of 12 doses during the 56 day observation period. We found that although the lesion sizes ( Fig. 3(A)) were increased during anti-PD1 and anti-PD-L1 treatment, the parasite loads were significantly decreased after anti-PD-1 or anti-PD-L1 treatment ( Fig. 3 induction of ifn-γ from CD4 + and CD8 + t cells after MoAb treatment. Focusing on the twice a week treatment protocol, we then examined the mechanism underlying parasite load reduction in anti-PD-1 or anti-PD-L1-treated mice by measuring the IFN-γ production of CD8 + and CD4 + T cells from the draining lymph nodes. It is documented in murine models of L. major infection that the expression of cytokines such as IL-12 and IFN-γ by Th1 contributes to host protection, whereas IL-4, IL-5, and IL-13 expression by Th2 contributes to host susceptibility 24,25 . In the murine L. amazonensis infection model, we and others have shown that impaired IFN-γ production and insufficient macrophage activation favor parasite survival and persistence [24][25][26][27] . In our treatment studies, we found more CD3 + CD8 + T cells in the L. amazonensis-infected mice compared to uninfected mice ( Fig. 1(A)), but there were no significant difference in the percentage and absolute number of CD3 + CD8 + T cells between the MoAb-treated and the PBS-injected groups ( Fig. 4(A,B)). However, both, percentage and absolute number of IFN-γ-producing CD3 + CD8 + T cells (IFN-γ + CD8 + T cells) were significantly increased in anti-PD-1 or anti-PD-L1-treated mice when compared to the PBS-injected group ( Fig. 4(C-E).
As the percentage and number of PD-1 + CD8 + T cells were increased in L. amazonensis infection ( Fig. 1), we next examined whether the IFN-γ production could also be affected in the CD8 + T cells lacking PD-1 (PD-1 − CD8 + ) during MoAb treatment. Similar to the total CD3 + CD8 + T cells, the PD-1 − CD8 + T cells of both groups of MoAb-treated mice presented a higher percentage of IFN-γ expression, but there was not a significantly higher number of IFN-γ + PD-1 − CD8 + T cells compared to the PBS-injected group (Fig. 5(A-C)).
Regarding PD-1 + CD8 + T cells not expressing IFN-γ (IFN-γ − PD-1 + CD8 + ), there was a significant decrease in the percentage of these cells in anti-PD-L1-treated mice compared to the PBS-injected group, without any effect on the absolute numbers ( Fig. 5(D,E)). We did observe an increase in the IFN-γ-expressing PD-1 + CD8 + T cells in both the percentage and absolute number of anti-PD-1-treated mice, but there was no significant difference in anti-PD-L1-treated mice compared to the PBS-injected group ( Fig. 5(F,G)).
Unlike the PD-1 − CD8 + T cells, MoAb treatment had no effect on the IFN-γ production by PD-1 − CD4 + T cells ( Fig. 7(A-C)). Again, opposite to what was observed in the IFN-γ − PD-1 + CD8 + T cells, there was a significantly higher percentage of IFN-γ − PD-1 + CD4 + T cells in anti-PD-L1-treated mice, with no difference on the absolute number of these cells (Fig. 7(D,E)). Additionally, a significant increase in the percentage of IFN-γ-producing PD-1 + CD4 + T cells (IFN-γ + PD-1 + CD4 + ) was found for both MoAb-treated groups compared to the PBS-infected group, which was not observed in the number of cells (Fig. 7(F,G)). We did not observe any difference in the production of IFN-γ + or expression of PD- Figs. 6 and 7). www.nature.com/scientificreports www.nature.com/scientificreports/ Altogether, our results suggest that anti-PD-1 and anti-PD-L1 MoAb treatment stimulated the production of IFN-γ in both CD4 + and CD8 + T cells, which may be one of the possible mechanisms for the control of parasite load.
Anti-PD-1 treatment reduces IL-4 and TGF-β. As IFN-γ production was induced by the MoAb treatment, we tested the effect of treatment on the modulation of the cytokines, IL-4, IL-10 and TGF-β, at the site of L. amazonensis infection. Only anti-PD-1 treatment significantly decreased IL-4 ( Fig. 8(A)) and TGF-β ( Fig. 8(B)) production in situ compared to the PBS-injected control group. No alteration in the IL-10 levels was found after MoAb treatment (Fig. 8(C)). Finally, we demonstrated that treatment with the MoAbs did not affect the production of anti-Leishmania specific IgM or IgG antibodies as assessed in the blood sera (Supplemental Fig. 4(A,B)). Altogether, our results suggest that anti-PD-1 MoAb treatment modulates the production of IFN-γ in CD4 + T and CD8 + T cells, but anti-PD-L1 only affects CD8 + T cells.  Test (B)).

Discussion
The use of leukocyte receptor blockers in immunotherapy has been extensively studied in oncology 28 . One of the clinical trials concerning the systemic administration of therapeutic antibodies to block PD-1 or PD-L1 has produced promising results for the treatment of several tumors 29,30 . However, the use of this therapy is still relatively limited in non-healing leishmaniasis.
For L. major infection in arginase-deficient mice, deficiency in T cell activation resulted in increased PD-1 expression, impairing the immune response and inducing T cell exhaustion 31 . In dogs injected with L. infantum antigens, PD-L1/PD-1 blockade with specific antibodies recovered the proliferation of CD4 + and CD8 + T cells, in addition to the production of IFN-γ by CD4 + T cells 32 . In studies of L. donovani infection in BALB/c mice, the parasite induced the initial expansion of IFN-γ-producing CD4 + and CD8 + T cells in the acute phase of the disease, the frequency of which was reduced after 21 days post-infection even with a robust parasite presence. In this model, blocking PD-L1 resulted in the restoration of CD4 + and CD8 + T cell responses, leading to a reduction in www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ parasite load 33 . In view of these studies, our results presented herein show that L. amazonensis infection interferes in the production of IFN-γ by both CD4 + and CD8 + T cells.
Recently, the presence of PD-1 and PD-L1 was detected in a patient with diffuse cutaneous leishmaniasis caused by L. amazonensis 21 . Based on the induction of PD-1 and PD-L1 by Leishmania infection, the blocking of these molecules may be a new strategy to treat leishmaniasis.
The use of anti-PD-1 and anti-PD-L1 antibodies in clinical cancer treatment studies, such as in pancreatic tumor, were administered at doses between 10 and 200 μg every three days [34][35][36] . In our study, we tested a lower therapeutic dose of 100 µg of the anti-PD-L1, anti-PD-L2 and anti-PD-1 antibodies once a week. However, our data revealed that this was insufficient in reducing the lesion size or parasite load. Therefore, the dosage of the treatment, in terms of the concentration and frequency of administration are important considerations. Thus,  www.nature.com/scientificreports www.nature.com/scientificreports/ we increased the administration to twice a week, and observed that although the lesion size was increased, the parasite load in the infected footpads, in the spleen and in the draining lymph nodes that received therapy with anti-PD-1 and anti-PD-L1 were controlled.

IFN-γ + IFN-γ + IFN-γ + IFN-γ +
In L. major infection, only treatment with 1 mg/dose of the anti-PD-1 antibody weekly in infected arginase-deficient mice led to complete resolution of the chronic skin lesion and resistance to infection 31 . A similar result was found when using the anti-PD-L1 antibody in mice challenged with L. donovani amastigotes, as these mice showed a reduction of up to 87% of the parasite load in the spleen 37 . The susceptibility of the C57BL/10 and C57BL/6 mice to infection by L. amazonensis is related to the absence of Th1 type cellular immune response and not controlled exclusively by Th2 cells 38,39 . In BALB/c mice, this susceptibility is related to Th2 type cellular immune response 40,41 .
In this study, we observed that the production of IFN-γ was higher in CD8 + T lymphocytes, both in percentage and number of cells after treatment with the anti-PD-1 and anti-PD-L1 MoAbs. We also detected an increase in the percentage of IFN-γ-producing CD4 + T lymphocytes after both treatments. However, when looking at the number of IFN-γ-producing CD4 + T cells only in mice treated with anti-PD-1 a significant increase was observed. These results suggest that treatment with anti-PD-1, acting directly on the lymphocytes, is more competent in invigorating CD4 + T lymphocytes than the anti-PD-L1 therapy, the target of which is in the antigen-presenting cells. Our data indicate that the increase of IFN-γ is one of the possible mechanisms of the therapeutic efficacy of these monoclonal antibody therapies, mainly by CD8 + T and partially by CD4 + T lymphocytes.
The PD-1/PD-L1 ratio is very important in suppressing the CD8 + T cell response during L. donovani infection 37 . In a study of dogs with symptomatic visceral leishmaniasis by L. infantum, these animals were shown to have a five-fold reduction in the proliferative capacity of CD8 + T cells and a reduction of up to three-fold in the ability of these cells to produce IFN-γ. After administration of specific monoclonal antibody therapy, PD-1 blockade significantly increased the proliferative capacity of CD4 + and CD8 + T cell populations, and recovered IFN-γ production in the CD4 + population. In addition, T cell depletion during visceral leishmaniasis was associated with elevated expression of PD-1, which could be identified before the onset of the disease and is considered a determining factor for symptomatic onset 32 . In another study of canine visceral leishmaniasis, it was observed that as the disease progressed, there was a decrease in CD4 + T cell proliferation and also a reduction of IFN-γ production in response to L. infantum antigens 40 .
These findings are reinforced by studies in which anti-PD-1 and anti-PD-L1 therapy reverts the ability of CD8 + T lymphocytes to produce IFN-γ as in the treatment of thyroid cancer 41 , in chronic hepatitis B 42 , and in HIV infection 43 . Further studies should be performed to confirm whether T cell exhaustion occurs in L. amazonensis infection. However, other studies that have assessed how chronic infections can induce exhaustion, in addition to our results reporting the increased capacity of IFN-γ production after MoAbs therapy, this suggests that T cell exhaustion occurs in L. amazonensis infection.
The participation of IFN-γ has been reported as playing a dual effect in L. amazonensis infection when discussing inflammatory response/lesion versus parasite load. IFN-γ participates in the control of parasite load enhancing the killing; however, it is also related to cell recruitment that increases the inflammatory response contributing to lesion development in WT mice 44 . In our experiments, in BALB/c mice, we demonstrated a similar mechanism, together with the increase of IFN-γ, we observed a decreased of parasite load, however, a small increase of the lesion associated to increase of inflammatory response that is also accompanying by the parasite load control.
Interestingly, our results indicate the potential of MoAb to modulate the cytokines present at the lesion site in the mice, as the production of IL-4 and TGF-β were reduced in the group treated with the anti-PD-1 MoAb. The group treated with anti-PD-L1 MoAb showed a tendency in the reduction of TGF-β.
It is already known that patients who progress to the disease have an increase in the production of immunosuppressive responses by the increase of TGF-β in Leishmania infection 45,46 . TGF-β also plays a role in susceptibility of murine cutaneous leishmaniasis caused by L. amazonensis 47 and L. chagasi infection 48 . TGF-β was www.nature.com/scientificreports www.nature.com/scientificreports/ shown to upregulate the PD-L1 expression in dendritic cells, leading to T-cell anergy and diminished anti-tumor response 49 . Thus, our data support the hypothesis that IFN-γ produced by CD4 + T cells reduces TGF-β production. Moreover, it is important to point out that TGF-β is associated with increased parasite load 45 and its reduction is directly related to parasite control. Hence, in our results, parasitic control may also be related to the decrease of TGF-β.
We also evaluated the production of IL-10 and IL-4 in this model. IL-10 has been reported to be associated to pathogenesis in L. mexicana, L. amazonensis 50 , L. donovani 51 and L infantum 40 infections. The participation of IL-10 was not seen in our results, since the concentration of IL-10 showed only a slight reduction in the MoAb-treated groups relative to the PBS-injected control, thus the reduction in parasite load is likely not associated to the levels of IL-10 in our model. We observed the reduction of IL-4 in anti-PD-1 treated mice. It is possible that the IFN-γ produced by CD4 + T cells may inhibit the development of a Th2 response reducing IL-4 production. IL-4 has been associated linked to lesion pathogenesis in murine cutaneous leishmaniasis caused by L. amazonensis, but does not directly affect the parasite load 52 . In human diffuse cutaneous leishmaniasis, there is no persistence of Th2 response in the lesions 53 . These data suggested that the reduction of IL-4 perhaps is not related to control of parasite load in our model.
We also evaluated if the humoral response was altered during MoAb treatment. Kima et al. showed that antibodies play a critical role in the pathogenesis and in the development of more significant lesions due to L. amazonensis infection, since the maintenance of infection by these parasites was impaired by the absence of circulating antibodies in the BALB/c model 54 . Recently, we demonstrated that L. amazonensis infection in XID mice displayed smaller lesions and a decrease in IL-10 and total antibodies in comparison to WT mice suggesting the pathogenic role of B cells in L. amazonensis infection 55 . Here, we demonstrated that treatment with both MoAbs did not affect the anti-Leishmania IgM, IgG, IgG1 and IgG2a antibodies levels.
In summary, our study suggests a potential use of monoclonal antibodies against PD-1 and PD-L1 in the treatment of cutaneous leishmaniasis caused by L. amazonensis. Our model, which employed a low dose treatment of anti-PD-1 or anti-PD-L1 showed therapeutic efficacy to control the parasite load in infected mice. We have also shown that this control is related to CD8 + T lymphocytes and, partially, to CD4 + T lymphocytes, producing IFN-γ. These findings could potentiate a combined therapy using anti-PD-1 or anti-PD-L1 antibodies and the current standard therapies against leishmaniasis, which could be particularly important for diffuse cutaneous leishmaniasis treatment, a disease that is refractory to conventional treatment.

conclusion
The present work analyzed the use of monoclonal antibodies as a potential treatment of cutaneous leishmaniasis caused by L. amazonensis. Our L. amazonensis mouse infection model presents a susceptible immune response and our treatment protocol with the MoAbs against PD-1 and PD-L1 favored the control of parasitic burden. These data together indicate that the monoclonal antibody therapy targeting PD-1 or PD-L1 could be used as a possible treatment for leishmaniasis..