Leishmaniasis, a vector-borne parasitic disease, is caused by dimorphic protozoan flagellates of the genus Leishmania with a worldwide distribution. The disease is characterized by diversity and complexity, presenting a wide spectrum of clinical forms in humans ranging from self-healing cutaneous leishmaniasis (CL) to fatal visceral leishmaniasis (VL). In VL, parasites colonize internal organs, primarily the spleen, liver, and bone marrow. With an estimated 0.5 million cases per year, VL has emerged as an important public-health concern with major clinical and socioeconomic impacts¹. In South Europe, VL is caused almost exclusively by Leishmania (L.) infantum, which is transmitted as a zoonosis with the domestic dog serving as the main reservoir of the parasite (1). Current attempts against leishmaniasis are based on chemotherapy to alleviate disease (2, 3) and on vector control to reduce transmission (4). Toxic side effects and growing resistance to available therapeutic drugs against VL has made the global demand for an effective vaccine capable to elicit a protective immune response, a major public-health priority.

Despite the substantial knowledge regarding the various life stages of the parasite, the considerable inter-specific diversity, the extraordinary host evasive mechanisms of parasite and the heterogeneity of population (5, 6), effective vaccine development against human VL represents an unprecedented challenge. Although a great number of potent vaccine candidates has shown promising results in mice (7, 8) and dogs (9–11), none of them has entered human trials except for LeishF1 with reported phase I and II clinical trials (12, 13).

Antigen identification is considered as a significant barrier in vaccine design, as this is usually achieved through time consuming and labor-intensive in vitro and in vivo experiments. Efforts have thus focused on developing novel strategies for more rational and faster antigen identification among large numbers of pathogen proteins. Furthermore, recent reports support that epitope-based vaccines appear to be capable of inducing more potent responses than whole protein vaccines (14). Until recently, the search of immunodominant peptides relied on the direct testing of overlapping peptides or peptide libraries. T cell epitope prediction by bioinformatic analysis of protein sequences has been proposed as a promising strategy for vaccine development and an increasing number of tools have been developed, based on different algorithms and methods (15, 16). There is great possibility of missing the emergence of the sequence mutants that would potentially escape the vaccine’s protective effect. Moreover, the fact that T cells from genetically distinct populations would recognize and respond to a single peptide epitope, underline the need of identifying one or more epitope(s) that bind to multiple HLA alleles and cover close to 100% of the genetically diverse human population (17). Multi-peptide-based vaccines are designed to generate a diverse immune response to incorporate antigens and to reduce limitations due to MHC restriction into a single entity.

The effectiveness of a vaccine depends on its capacity to ensure long-lasting cell-mediated immunity. In VL, there is evidence that an interplay of T helper cytokines (T_H1/T_H2) is observed, while resistance or resolution of infection is associated with dominant T_H1 response and CD8⁺ T cells (18–20). Furthermore, successful treatment of VL with sodium stibogluconate requires the presence of both CD4⁺ and CD8⁺ T cells (21) accompanied with IL-12 and IFN-γ production (22). In contrast, T_H2 response with IL-4 and IL-10 production results in susceptibility to infection and development of severe disease. Murine models of leishmaniasis have been extensively used to study the pathogenesis of the disease and to test novel therapeutic agents or potent vaccine candidates in pre-clinical studies. One of the most widely studied and commonly used model of VL is the BALB/c strain of mice infected intravenously with L. infantum. Although this strain is considered to be susceptible and the infection progresses during the first month, the infection is then controlled by the host immune response. This mouse model is comparable to self-controlled oligosymptomatic cases and therefore is useful for the study of the protective immune response (23).

Several reports demonstrate that different leishmanial antigens elicit desired T_H1 and CTL responses capable to sustain protection against experimental challenges (24). Among these antigens, cysteine peptidase A (CPA), histone H1, kinetoplastid membrane protein 11 (KMP-11), and Leishmania eukaryotic initiation factor (LeIF) are considered important immunogens, as supported by numerous studies. Specifically, CPA induces protection against L. major in the experimental model of CL through development of specific T_H1 immune responses (25–27). Histone H1 and KMP-11, structural highly conserved proteins, are able to trigger specific immune responses (28–33), and immunization with these proteins confer protection against L. major or L. infantum infections in experimental animal models (34–38). Moreover, LeIF, originally described as a T_H1-type natural adjuvant, is capable of stimulating IL-12 mediated T_H1 responses in PBMCs of patients (39). Furthermore, recombinant forms of CPA, histone H1, and KMP-11 act as potent B cell immunogens since they are recognized by sera from either recovered or active cases of CL and VL, as well as by sera from asymptomatic or symptomatic dogs with leishmaniasis (40–45).

In the present study, we applied immunoinformatics using currently available online algorithms in order to identify potentially immunogenic T cell epitopes from the above mentioned L. infantum proteins, and design multi-epitope peptides containing both MHC class I and II-restricted epitopes as possible candidate peptide vaccines for VL. Immunogenicity of the synthetic multi-epitope peptides in terms of T cell activation was validated in immunized BALB/c mice by analyzing peptide-specific proliferative responses and cytokine production by CD8⁺ and CD4⁺ T cells.

In the perspective of second generation vaccines, a variety of different parasite molecules, such as secretory or transmembrane proteins, including enzymes and receptors, has been tested to date as candidate antigens for anti-Leishmania vaccine development (56). Among them, CPA, Histone H1, KMP-11, and LeIF were found to be highly immunogenic as described in murine experimental models, cured VL patients and L. infantum infected dogs and have been considered as potential vaccine candidates (39–41, 44, 45, 57). The induction of an effective T cell response against vaccine antigens requires antigen processing and peptide presentation by antigen-presenting cells (APCs), and it is well-established that T cells recognize the peptide sequence in association to appropriate MHC molecules. The discovery of MHC-binding motifs in proteins has led to the development of several algorithms predicting MHC class I- and II-restricted epitopes for presentation to CD8⁺ or CD4⁺ T cells, respectively, accelerating research related to peptide-based vaccine approach (15).

In the present study, we investigated the use of three algorithms, SYFPEITHI, BIMAS, and NetMHCII to predict sequences in CPA, Histone H1, KMP-11, and LeIF able to bind to MHC class I and II molecules of the H2^d haplotype. Furthermore, combining this approach with experimental validation in MHC compatible BALB/c mice, we determined epitopes in each protein and designed multi-epitope peptides capable to induce peptide-specific T cell proliferation and cytokine production by CD4⁺ and/or CD8⁺ T cells.

The analysis of protein sequences yielded a significant number of possible epitopes from all four proteins, but only few of them were predicted by all algorithms used with binding efficiency to more than one supertypes or alleles. Interestingly, comparison of predicted peptides for each MHC class I and II alleles showed a low overlapping level between the results obtained from different algorithms used in the study, indicating the significant differences existing in the database source of building matrix motifs and different forms of scoring function of each algorithm. Also, it was observed an antigenic region clustering. Based on these findings and to the fact that prediction of MHC class I-restricted epitopes is considered more reliable (>85%) than that of MHC class II-restricted epitopes, we designed eight multi-epitope peptides for all proteins, based predominantly on highly scored MHC class I-restricted epitopes. Adjacent or overlapping MHC class II-restricted epitopes scored high were also nested in each synthetic peptide. These multi-epitope peptides contained epitopes recognized also by HLA class I and II molecules as defined by in silico analysis. Until now, very few vaccine antigens against different pathogens such as viruses, bacteria, and parasites, contain promiscuous T cell epitopes that have the ability to induce T cell-mediated protective immune responses both in mice and human by binding to several alleles of a supertype or between different supertypes (58, 59). Thus, these promiscuous epitope-driven vaccines could have the capacity of increasing the frequency of responders in genetically variable species, such as human populations (60).

The success of many vaccines is dependent on IFNγ-secreting CD4⁺ T cells recruitment for long term protection. This accounts for better immunologic memory leading to sustained immunity after healing of live infections (61, 62). CD4⁺ T cells are activated in terms of recognition of peptides-MHC class II complexes in the surface of APCs after protein processing in cells’ endocytic compartment. Activation of IFN-γ-producing CD4⁺ T cells plays a pivotal role in protective immune responses against leishmaniasis. Specifically, IFN-γ mediates macrophage activation against both the promastigote and amastigote forms in H₂O₂-dependent manner (63, 64) and nitric oxide production for parasite killing (65). According to our results, CPA_p2, CPA_p3, LeIF_p3, and LeIF_p6 induced peptide-specific IFN-γ production from CD4⁺ T cells in immunized mice indicating the processing and recognition of MHC class II-specific epitopes by CD4⁺ T cells. CPA and LeIF are considered significant candidate proteins for vaccine design against leishmaniasis. In the case of CPA, it has been shown that administration of plasmid encoding CPA induced specific T_H1 immune responses resulting to partial protection against L. major in the experimental model of CL. However, protection was significantly enhanced when co-administered with CPB or as CPA/B hybrid protein (25–27), indicating the need of CPA co-administration with another protein or adjuvant. On the other hand, LeIF was originally described as a T_H1-type natural adjuvant and as an antigen inducing an IL-12 mediated T_H1 response in the PBMCs of leishmaniasis patients (39). LeIF is also capable of inducing the secretion of cytokines IL-12, IL-10, and TNF-α by APCs from healthy individuals (52, 66, 67). Furthermore, recombinant trifusion vaccines (leish111; leish110f) were developed by incorporating the amino-terminal region of LeIF antigen. These vaccines were shown to be efficient in experimental or clinical trials for vaccination or immunotherapy (68).

Existing data suggest that secretory and surface exposed proteins strongly induce specific CD8⁺ T cell responses (69, 70). A previous study applying in silico analysis revealed that a high number of peptides derived from L. major secretome could bind to H2 BALB/c molecules (71). Several studies have shown the great role played by CD8⁺ T cells in protective immune responses against parasite in the susceptible BALB/c strain (72–74). Specifically, CD8⁺ T cells either contributed in the destruction of Leishmania-infected cells by activating macrophages to oxidative burst via cytokines produced upon antigen stimulation (75, 76), or regulating CD4⁺ T cell-mediated immune responses (77, 78). In our study, both synthetic multi-epitope peptides of CPA, CPA_p2, and CPA_p3, except from CD4⁺IFN-γ⁺ T cells activation, induced significant IFN-γ production by CD8⁺ T cells. CPA_p2 and CPA_p3 belong to the secreted region of CPA as SignalP analysis showed (30, 31). Interestingly, another study applying in silico analysis in CPA sequence with MULTIPRED algorithm indicated the existence of four highly immunogenic regions recognized by the HLA-A2 supertype (79), which harbored parts from our CPA multi-epitope peptides.

As for KMP-11, KMP-11_p1 belonged to the amino-terminal region of the protein and in contrast to previous observations this synthetic peptide was proved to be poorly immunogenic, indicated by the absence of peptide-specific proliferative response and cytokine secretion in immunized mice. Previous studies concerning the identification of T cell epitopes using infected macrophages or DCs as APCs, revealed the existence of potential HLA class I- and II-restricted T cell epitopes in the amino-terminal region, characterizing a dominant cluster between position 1 and 33 of KMP-11 sequence that could trigger specific cellular immune responses in L. donovani- or L. panamensis-infected volunteers (80, 81). Furthermore, hybrid-cell, DNA-based or heterologous KMP-11-DNA/rVV based vaccination exhibited immunoprotective capacity in susceptible VL murine models. Protection was accompanied with generation of antigen specific CD4⁺ and CD8⁺ T cells that produced effector cytokines such as IFN-γ, IL-2, and TNF-α (36–38, 82, 83). Also in a previous work, we demonstrated that vaccination with ex vivo pulsed bone marrow-derived dendritic cells with KMP-11_12–31aa peptide and CpG as adjuvant induced strong Th1 and Th17 protective immune responses in murine model of VL (50). However, in the present study it is noteworthy that secretion of IL-10 was also abrogated. These results together suggest that KMP-11_p1 may be consisted from natural epitopes contributing in parasite host immunomodulation, allowing parasite dissemination rather than stimulate protective immune responses.

However, not only external or secreted Leishmania antigens are able to be presented in the context of MHC class I molecules but also intracellular proteins (84, 85). As such, in our study H1_p1 and H1_p3 induced a T cell response characterized mainly by CD8⁺ T cell priming and production of IFN-γ in immunized mice. Although, this way of cell activation in leishmaniasis remains controversial and it is not clear how non-secretory parasite antigens such as histone H1 can be presented endogenously in the context of MHC class I molecules, a number of studies supports the induction of specific CD8⁺ T cell responses against structural parasite proteins in animal models and VL patients (82, 84, 86, 87). Previous results from our group supported that ex vivo pulsed bone marrow-derived dendritic cells with the Leishmania histone H1 elicited significant protection in the experimental model of VL, with a pronounced enhancement of parasite-specific IFNγ-producing CD8⁺ T cells (88). The protective effect of Leishmania histone H1 against L. major or L. infantum infections was also shown in different experimental animal models (34, 35) suggesting that it is also a promising vaccine candidate against leishmaniasis. In contrast, none of the LeIF peptides tested could evoke specific CD8⁺ T cell responses. This finding was in agreement with the study of Rafati et al. showing that PBMCs from patients recovered from L. major failed to elicit HLA-A2-restricted CD8⁺ T cell responses against three synthetic nonamer peptides of LeIF, suggesting that these peptides are not able to induce a CD8⁺ T cell-induced protective immunity (86).

The relative low concentrations of IL-10 detected in the supernatants of immune lymphocytes compared to IFN-γ after peptide re-stimulation were consistent with the suggestion of a dynamic reciprocal relationship between these two cytokines. IL-10 primarily down-modulates innate as well as acquired immunity leading to parasite establishment or disease progression. In experimental model of VL, IL-10 prevents DCs migration in spleen to activate T cells (89, 90) and suppresses both T_H1 and T_H2 cells (91). Also, the CD4⁺IL-4⁺ T lymphocytes detected in the presence of all peptides may be attributed to BALB/c intrinsic feature to induce the production of type-2 cytokines, such as IL-4 (53), since there was not any significant difference of IL-4 levels between peptide-immunized mice and control mice receiving the adjuvant alone. Furthermore, IFA adjuvant has a propensity to induce preferentially T_H2 cytokines (92, 93). Similar study for the evaluation of immunoreactivity of in silico predicted T_H1 epitopes of Schistosoma japonicum showed that high levels of IL-4 were attributed to Freund’s adjuvant and BALB/c strain used (94). Therefore, in terms of proportion of intracellular cytokine production of recall CD8⁺ and CD4⁺ T cells, it is concluded that CPA_p3, H1_p1, H1_p3, CPA_p2, LeIF_p3, and LeIF_p6 synthetic multi-epitope peptides are likely to include potential epitopes for the induction of protective cytotoxic (CTL) and T_H1-type immune responses. Taken into account that the sequences of these synthetic peptides are highly conserved and bind in a promiscuous manner to murine or human MHC molecules according to in silico analysis and structure-based techniques, make them candidate vaccines against leishmaniasis. Based on these results, it would be worthwhile conducting future investigations for the verification of peptides’ possible ability to induce protection in common or humanized mouse models of leishmaniasis. The incorporation of alternative and/or additional epitopes, the use of modern adjuvants and new antigen delivery systems should be combined. Conclusively, these findings give complementary data on epitope mapping for Leishmania proteins and demonstrate that combination of immunoinformatic approaches with experimental validation enables peptide identification with greater accuracy contributing to rational epitope-based vaccine development.

Evita Athanasiou and Olga Koutsoni contributed equally to this work. Conceived and designed the experiments: Evdokia Karagouni. Performed computational analysis: Maria Agallou, Evita Athanasiou, Olga Koutsoni, Evdokia Karagouni. Performed the experiments: Maria Agallou, Olga Koutsoni, Evita Athanasiou. Analyzed the data: Maria Agallou, Evita Athanasiou, Olga Koutsoni, Evdokia Karagouni, Eleni Dotsika. Wrote the paper: Evdokia Karagouni, Maria Agallou.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.