Impaired activation of lesional CD8+ T-cells is associated with enhanced expression of Programmed Death-1 in Indian Post Kala-azar Dermal Leishmaniasis

Post Kala-azar dermal leishmaniasis (PKDL), caused by Leishmania donovani is the dermal sequel of Visceral Leishmaniasis and importantly, is the proposed disease reservoir. The survival of Leishmania parasites within monocytes/macrophages hinges on its ability to effectively nullify immune activation mechanisms. Thus, delineating the disease-promoting immune mechanisms can facilitate development of immunotherapeutic strategies. Accordingly, in the absence of an animal model, this study aimed to delineate the status of CD8+ T-cells in patients with PKDL. At disease presentation, the absence of CD4+ T-cells at lesional sites was concomitant with an overwhelming infiltration of CD8+ T-cells that demonstrated an absence of Perforin, Granzyme and Zap-70, along with an enhanced expression of Programmed Death-1 (PD-1) and the skin-homing CCL17. Additionally, the lesional CCR4+CD8+ population was associated with an enhanced expression of IL-10 and IL-5. In circulation, the enhanced CD8+CCR4+ T-cell population and raised levels of CCL17/22 was associated with an increased frequency of PD-1, while CD127 was decreased. Taken together, in PKDL, the enhanced plasma and lesional CCL17 accounted for the dermal homing of CD8+CCR4+ T-cells, that along with a concomitant upregulation of PD-1 and IL-10 mediated immune inactivation, emphasizing the need for designing immunotherapies capable of reinvigorating T-cell potency.

In PKDL, a disease where no animal model exists, information is derived solely from human studies, and understandably remains limited. Studies have endorsed the presence of a systemic and dermal immunosuppressive milieu and includes the presence of an increased population of antigen-specific IL-10 producing anergic T-cell population in peripheral blood 20 , a decreased presence of dendritic cells at lesional sites 21 , dampening of the CD26 regulated pathways 22 , a huge infiltration of CD68 + alternatively activated macrophages 23 and a dermal pathology dominated by IL-10 and FoxP3 15,17,20 , that individually or more likely collectively contribute towards establishment of a pro-parasitic milieu. In the peripheral blood of polymorphic PKDL as compared to the macular variant, stimulation with Leishmania antigen enhanced levels of activated CD8 + and CD4 + T cells 24 . However, what remains poorly defined in PKDL is the status of chemokines and T-cells at the lesional sites, along with defining their contribution, if any, in supporting disease progression. Accordingly in this study, the activation status of CD4 + and CD8 + T-cells, cytotoxic markers e.g. Perforin, Granzyme and p-Zap-70, inhibitory receptor-Programmed death-1 (PD-1), skin homing chemokine CCL17 and its receptor, Chemokine Receptor 4 (CCR4) along with IL-5 and IL-10 were evaluated in dermal lesions of patients with PKDL. The results demonstrated an increased proportion of CD8 + CCR4 + T-cells and CCL17/CCL22 indicative of dermal homing, while the upregulation of PD-1 and IL-10 suggested impaired activation of CD8 + T-cells. Taken together, this dermal homing of anergic/exhausted CD8 + T-cells supported parasite survival and disease progression in patients with PKDL.

Results
The study population included patients with PKDL (n = 20) recruited from 2004-2014, whose median age was 27.50 years with a male preponderance (Table 1) 3,25,26 . The majority demonstrated hypopigmented, papular and/ or nodular lesions, termed 'polymorphic' (n = 18, 90.0%), while a minority presented with hypopigmented lesions termed 'macular' (n = 2, 10.0%). The papular/nodular lesions appeared primarily on sun-exposed areas like face, neck and upper limbs and ranged from 10-12 in number, whereas for the macular variant, the distribution was more diffuse (Supplementary Fig. S1). All were ITS-1 PCR positive whereas Leishman-Donovan (LD) bodies were detected only in Giemsa stained smears of the polymorphic variant. Two patients gave no prior history of VL, while in the remaining 18, the mean time interval between cure from VL and onset of PKDL was 2.75 years (Table 1). Following completion of treatment, polymorphic PKDL cases showed complete resolution of papules/nodules whereas in macular PKDL, there was substantial reduction in hypopigmentation ( Supplementary  Fig. S1).
Extensive dermal infiltration and absence of granuloma are features of Indian PKDL. An overall histopathological analysis by H&E staining showed a dense, diffuse inflammatory cell infiltrate involving the entire dermis, consisting mainly of lymphocytes, macrophages and plasma cells, and varied from severe in the polymorphic cases to patchy perivascular and periappendageal with relative sparing of the reticular dermis in the macular variant. Other features included an atrophic epidermis, follicular plugging, hyperkeratosis and papillomatosis. Unlike leprosy, the narrow sub-epidermal Grenz zone was spared and there were no well-formed granuloma. The neural Schwann cells remained unaffected and did not harbour parasites; furthermore, except for one case where perineural infiltration was prominent 27 , none reported any loss of sensation or involvement of the sub-cutis ( Supplementary Fig. S2).
CD8 + T-cells predominated in the dermal infiltrate. In Indian PKDL Rathi et al. 28 reported a preponderance of T-suppressor over T-helper cells but as the cellular phenotype had not been characterized, further studies are warranted, especially to establish their activation status. Immunohistochemical analysis at the lesional site showed a conspicuous absence of CD4 + T-cells, vis-a-vis healthy controls [0 vs. 30 Fig. 1a,b] that remained so even after treatment (Fig. 1a,b). CD4 + staining was confirmed in human lymph node sections ( Supplementary Fig. S3). However, the proportion of CD8 + T-cells was 4.61 fold higher than controls, 50.30 (49.30-52.05) vs. 10.9(9.04-11.81), p < 0.001 ( Fig. 1c-e). Importantly, there was a strong correlation between the frequency of CD8 + T-cells and disease duration, r = 0.70 (CI = −0.15-0.94). With treatment, a dramatic decline occurred, 16.21 (15.37-18.33), p < 0.05 ( Fig. 1c-e), which was evident even on an individual basis, p < 0.01 (Fig. 1f). Flow cytometric analysis of peripheral blood showed that the proportion of CD4 + and CD8 + T-cells remained unchanged and was comparable with healthy controls.
Increased CD8 + CCR4 + cells in dermal lesions and peripheral blood. The raised levels of dermal homing chemokines CCL17 and CCL22 (Fig. 2a-c) suggested that the increased presence of CD8 + T-cells was secondary to their homing to dermal lesions 30 . Accordingly, the status of CCR4, the chemokine receptor for CCL17/22 was examined at dermal sites by immunohistochemistry and flow cytometry in peripheral blood. In dermal lesions, the proportion of CCR4 + cells was significantly up regulated, 69. 56

Decreased expression of granzyme and perforin in lesional CD8 + T-cells.
To assess the cytotoxic potential of lesional CD8 + T-cells, the lesional status of Perforin-H and Granzyme-B along with nuclear factor Zap70 was examined by immunohistochemistry 31 . Despite an overwhelming presence of CD8 + T-cells, PKDL cases demonstrated a conspicuous absence of Perforin, Granzyme and Zap70 in the dermal lesions ( Fig. 4a-c).
As exhaustion is associated with an immunosuppressive milieu comprising IL-4 and IL-10 35 , their plasma levels in PKDL were examined and found to be significantly raised, and with treatment a significant decrease was reported 16 . Furthermore, as the sustenance of CCR4 + T-cells (Fig. 3) is generally supported by an IL-10 and IL-5 rich milieu 36,37 , their expression was examined at lesional sites. During active disease, there was a strongly positive diffuse sub-epidermal expression of IL-10 which receded with treatment (n = 10, Fig. 6a). Concomitantly, an enhanced expression of IL-5 was evident, 18.12(41.35-60.14) vs. 0 (Fig. 6b,c), which also declined with treatment, 4.30 (10.21-31.25), p < 0.05 (Fig. 6b,c); this was mirrored on an individual basis, p < 0.01 (Fig. 6c). As the expression of IL-10 at the lesional sites was diffuse, individual cells could not be counted and hence, a correlation, if any, between PD-1 and IL-10 was not assessed.

Discussion
CD8 + naive T-cells, being mediators of pathogen control, are targets for microbial modification. To achieve this, they orchestrate differentiation via induction of cytotoxic/effector cells that in turn release IFNγ, activate macrophages and dendritic cells, as also promote differentiation of CD4 + T-cells 38 . In dermal Leishmaniasis, CD8 + T-cells show a diverse disease repertoire with varying intensities of inflammatory infiltrate, and depending on the parasite species, can be protective or deleterious 39 . In patients with localised self healing CL (LCL) caused by L. major, an increased proportion of CXCR3/CD4 + and/or CD8 + T-cells provided a protective role 40 , whereas, in severe forms of CL caused by L. braziliensis, increased disease severity occurred secondary to dysregulated CD8 + T-cell cytotoxicity 41 . Studies have shown skin-resident, IFN-γ producing CD4 + T-cells protect against L. major, via recruitment and activation of inflammatory monocytes that mediate parasite elimination by production of reactive oxygen species and nitric oxide 42 . Conversely, in patients with diffuse CL (DCL) caused by L. mexicana, the uncontrolled spread of the parasite was associated with an enhanced presence of exhausted CD8 + T-cells 43 . In a head on comparison of LCL vs. DCL, it was demonstrated that CD8 + T-cells sourced from DCL showed a significant reduction in their effector responses vis-à-vis their self healing variant 43 . In self healing PKDL in East Africa/ Sudan, a preponderance of CD4 + T-cells vis-a-vis its CD8 + variant was demonstrated 44 . In contrast, in Indian non self-healing PKDL, although there was a massive influx of CD3 + T-cells 20 , it comprised primarily of CD8 + , with a near total absence of CD4 + T-cells (Fig. 1) that possibly facilitated development of a pro-parasitic milieu and disease sustenance. This absence of CD4 + T-cells can be attributed to the enhanced presence of alternatively activated macrophages, that via IL-4/IL-13 mediated activation of Stat-6 inhibit T-cell proliferation and survival [45][46][47] .
In Leishmaniasis, the inhibition of IL-12 along with an induction of IL-10 and TGFβ underlies the bias towards a Th2 response 11 . In PKDL, the enhanced secretion of IL-10 and TGF-β by antigen-stimulated PBMCs has been associated with disease severity 18 . Cellular infiltration is common to Sudanese and South Asian PKDL ( Supplementary Fig. S2), but the absence of granuloma in the latter can be attributed to the IL-10 rich milieu (Fig. 6), which can inhibit granuloma formation as demonstrated in tuberculosis 48 . Furthermore, the resultant microenvironment of IL-4/IL-13 and IL-10 facilitated the emergence of alternatively activated M2 monocytes/ macrophages 23 , that in turn supported the expression of Th2 chemoattractants, CCL17 and CCL22 (Fig. 2). These chemokines upon interaction with their receptor CCR4, facilitate migration of skin homing T-cells to the pathologic site, and was endorsed in PKDL by the increased presence of IL-5 and IL-10 ( Fig. 6) 37 . Post treatment in a few cases, the levels of CCL17 and CCL22 remained increased or unchanged, that may be attributed to their longer duration of disease (6-12 years) vis-a-vis the median disease duration of 2.71 years. The enhanced frequency of CD8 + CCR4 + T-cells in circulation and dermal lesions promoted the dermal extravasation of CD8 + T-cells (Fig. 3).  In CL, the higher expression of CXCR3 vis-a-vis CCR4 accounted for their self healing propensity 40 . This was corroborated in CL caused by L. braziliensis, wherein during the early phase of the disease, there is a raised expression of CXCR3 49 . However, in peripheral blood and dermal lesions of PKDL cases, the frequency of CXCR3 + remained unaltered as compared to healthy individuals ( Supplementary Fig. S4). Subsequently, in the late stage of the disease, a higher expression of CCL17 and CCR4 has been reported, suggestive of a preferential recruitment of regulatory T-cells that facilitated disease progression 49,50 .
A blockade of costimulation can result in the induction of T helper cell anergy and subsequent differentiation of antigen-specific CD8 + T suppressor/regulatory cells. Accordingly, it may be envisaged that in PKDL, the enhanced frequency of IL-10 producing antigen-specific circulating CD8 + T-cells 15 and CD8 + CD28 − T-cells 20 , along with impairment of antigen specific proliferative responses 20 caused attenuation of the effector CD8 + T-cell responses in the dermal lesions, as evident by the conspicuous absence of Perforin and/or Granzyme positivity 51 (Fig. 4) and absence of tissue damage. This was corroborated by absence of the intracellular signaling molecule, Zap-70 which is essential for TCR mediated CD8 + T-cell activation and cytolytic activity 31 . Additionally, as the decreased presence of CD127 or the IL7 receptor is associated with an increased rate of disease progression and T-cell exhaustion 52 , its reduced expression (Fig. 4) endorsed the impaired activation status of CD8 + T-cells in PKDL.
The activation of CD8 + and CD4 + T-cells is tightly regulated by an inhibitory receptor, Programmed death PD-1 53 . In their efforts to survive and proliferate, the Leishmania protein gp63 employs immune-escape markers, such as activation of SHP-1 phosphatases, increased PD-1 functions and collectively suppresses T-cell responses 54 . In an experimental model of VL, Joshi et al. 55 demonstrated that L. donovani parasites can evade CD8 + T-cell responses via induction of functional exhaustion. This was substantiated in human VL wherein splenic CD8 + T-cells displayed an anergic phenotype, leading to an immunological imbalance and a breakdown of protective responses 56 . In patients with non self healing DCL, the enhanced expression of PD-1 was associated with a pronounced non-responsiveness of CTLs and disease progression 43 . The scenario was similar in PKDL as CD8 + T-cells showed an enhanced presence of PD-1 in peripheral blood and dermal lesions (Fig. 5), thereby creating an environment conducive for parasite survival. Ideally, functional assays would provide confirmatory evidence of the non-responsiveness of these PD-1 expressing CD8 + T-cells, but obtaining the requisite number of cells would pose a logistic challenge. Taken together, in PKDL, the increased presence of CCL17 and CCL22 in circulation accounted for the dermal homing of the CD8 + CCR4 + T-cells to the lesional sites, while the upregulation of PD-1 and IL-10 suggested exhaustion, collectively promoting disease progression (Fig. 7). Thus, Leishmania donovani parasites have deviously evolved immune escape mechanisms emphasizing the importance of designing immunotherapeutic strategies aimed at restoring effector responses.

Materials and Methods
Reagents. Materials

Study population.
Patients clinically diagnosed as PKDL (n = 20) were recruited from the Dermatology outpatient department, School of Tropical Medicine, Kolkata, India. The initial diagnosis was based on clinical features, a prior history of VL, rK-39 positivity, or if resident in an area endemic for VL. Diagnosis was confirmed by ITS-1 PCR and/or Giemsa staining in dermal biopsies. Patients were randomly allocated to receive SAG (20 mg/kg bw/day i.m., 4 months, n = 8) or Miltefosine (100 mg/day p.o., 4 months, n = 12). Age and gender-matched healthy volunteers (n = 10) were recruited from non-endemic and endemic areas and were negative for anti-leishmanial antibodies and ITS-1 PCR 57 . A dermal biopsy (4 mm) was obtained using a punch biopsy by anaesthetizing the area with lignocaine and venous heparinised blood (5 ml) was collected at disease presentation and completion of treatment. Due to limited availability of biological material, all markers were not evaluated in every patient, and were randomly selected ensuring that at least 5 individuals were studied per assay.
Flow cytometry. Lymphocytes were gated on their characteristic forward vs. side scatter followed by fluorescence; 5000 lymphocytes were acquired and analyzed using CellQuest Pro software (BD Biosciences, San Jose, CA, USA). Fluorescence was evaluated in terms of frequency (% positivity) and expression (geometric mean fluorescence channel, GMFC) in the FL-1 channel for FITC and Alexa 488, for PE in the FL-2 channel, for PerCP in the FL-3 channel and for APC in the FL-4 channel. The lymphocytes were first gated based on their side and forward scatter, the CD8 population was then identified using CD8-PerCP and the frequency of CD127-APC, CD69-PE, Perforin-PE, Granzyme-PE, CCR4-PE, CXCR3-APC and PD1-APC examined within the CD8 + T-cells. Frequency was calculated in the entire CD8 + population by dividing the percentages of upper right quadrant by the sum of upper and lower right quadrant. Analysis was done for fluorescence by Cell-Quest pro software (BD Biosciences, San Jose, CA, USA).

Statistical analysis.
Results were expressed as median (Interquartile range, IQR) and data analyzed between groups by Kruskal-Wallis Post Dunn test for non-parametric data, while paired data were analyzed using Student's t-test (for parametric data). Correlation was done using Pearsons correlation for parametric data using GraphPad Prism software (version 5.0, GraphPad software Inc., La Jolla, CA, USA), p < 0.05 being significant.  gave a written informed consent. Informed consent was also obtained to publish after maintaining patient confidentiality information/images in an online open-access publication.

Data Availability
All data generated or analysed during this study are included in this article and its Supplementary Information.