Development of Vaccines against Visceral Leishmaniasis

Leishmaniasis is a neglected disease resulting in a global morbidity of 2,090 thousand Disability-Adjusted Life Years and a mortality rate of approximately 60,000 per year. Among the three clinical forms of leishmaniasis (cutaneous, mucosal, and visceral), visceral leishmaniasis (VL) accounts for the majority of mortality, as if left untreated VL is almost always fatal. Caused by infection with Leishmania donovani or L. infantum, VL represents a serious public health problem in endemic regions and is rapidly emerging as an opportunistic infection in HIV patients. To date, no vaccine exists for VL or any other form of leishmaniasis. In endemic areas, the majority of those infected do not develop clinical symptoms and past infection leads to robust immunity against reinfection. Thus the development of vaccine for Leishmania is a realistic public health goal, and this paper summarizes advances in vaccination strategies against VL.


Introduction
There is currently no vaccine available for any form of leishmanias, including visceral leishmaniasis (VL), which if left untreated is almost always fatal. VL results from systemic infection with L. infantum (also known as L. chagasi) [1] which occurs in Europe, North Africa, South and Central America, and L. donovani, which is found throughout East Africa, India, and parts of the Middle East. Infection is initiated by the bite of an infected sand fly vector and parasites disseminate from the site of infection in the skin to reside and multiply within macrophages of the liver, spleen, and bone marrow [2]. VL, also known as kala-zar, is associated with fever, weight loss, enlargement of the spleen and liver, and anaemia. Leishmaniasis has strong links with poverty [3] and is considered one of the most neglected tropical diseases. Each year there are approximately 500,000 new cases of VL with over 90% of cases arising in India, Bangladesh, Nepal, Sudan, and Brazil. VL has become a frequent coinfection in HIV-positive individuals in endemic areas and is associated with enhanced onset of AIDS-related illness and increased VL treatment failure. Current VL therapy is based on the long-term parenteral administration of pentavalent antimonials, which, despite being expensive and highly toxic, has been the standard treatment for over 50 years.
Following L. donovani infection some individuals develop Post-kala-azar Dermal Leishmaniasis (PKDL), a complication that occurs either during or after treatment. During PKDL parasites reappear in the skin resulting in an array of small skin lesions, and patients are considered a significant infection reservoir because of the large number of parasites accessible to sand fly bites. Thus the treatment and control of PKDL are an important public health measure for controlling VL and must be considered in the development of VL vaccine strategies.
Host resistance to Leishmania infection is mediated by cellular immune responses leading to macrophage activation and parasite killing. Antileishmanial immunity is mediated by both innate and adaptive immune responses and requires effective activation of macrophages, dendritic cells (DCs), and antigen-specific CD4 + and CD8 + T cells [4]. Although strong humoral responses are induced by VL infection, antibodies play no role in protection and are often associated with disease exacerbation [5]. Effector CD4 + T cells are responsible for the production of cytokines critical for the activation of macrophages and are required for optimal host response to infection [6]. Cytotoxic CD8 + T cells also play Journal of Tropical Medicine a host protective role, and are required for the effective clearance of parasites [7] and the generation of memory responses [8]. Interestingly, 80 to 90% of human infections are subclinical or asymptomatic, and this asymptomatic infection is associated with strong cell-mediated immunity. Only a small percentage of infected individuals develop severe disease, and patients who recover from VL display resistance to reinfection. This suggests the development of clinical immunity and provides a biological rationale for the development of VL vaccines that impart a strong cellular immunity.
Humans are the only known hosts for L. donovani; however L. infantum is primarily a zoonotic disease and canine species are the main animal reservoir. Canine visceral leishmaniaisis (CVL) affects millions of dogs in Europe, Asia, North Africa, and South America and has been associated with outbreaks of human VL [9]. Both symptomatic and asymptomatic Leishmania-infected dogs act as a source of parasites for VL transmission [10], and CVL represents a significant public health issue. A current approach to breaking the VL transmission cycle is the development of CVL vaccine, which may be crucial for controlling VL infection in human populations.
Vaccination against VL has received limited attention compared with cutaneous leishmaniasis (CL). Historically CL has been the focus of vaccination attempts, as it has been known for centuries that people who resolve a primary CL skin lesion are protected from further infections. It is generally acknowledged that human VL trials will follow on from any successful CL immunization programme. Ideally a vaccine would provide cross-protection against multiple Leishmania species. The recent comparative genomic analysis of three Leishmania species, which cause distinct disease pathologies, showed that L. major, L. braziliensis, and L. infantum genomes are highly conserved and have very few species-specific genes [11]. The level of amino acid conservation within coding regions is also high between species, suggesting that the major Leishmania antigens are conserved and that a pan species vaccine may be achievable. However there is a high degree of variability in the crossprotective immunity induced by infection with different Leishmania species [12,13] and VL-specific vaccines may provide a more successful intervention.
Experimental infection models are used to screen and evaluate VL vaccines, and several animal species have been used including mice, hamsters, monkeys, and dogs [14]. However no single in vivo model accurately reflects all aspects of human VL disease, which has been a major limitation in the development of VL vaccines. The precise immune mechanisms underlying human VL are still not fully understood, and the responses necessary for protection by vaccination in experimental infection models may not reflect those required for efficacy in endemic areas [15].
The profile of an antileishmanial vaccine would need to incorporate several important features, such as safety, ease of production at a low cost in endemic countries, the induction of robust, long-term T cell responses, and both prophylactic and therapeutic efficacy. Ideally, such vaccine would offer cross-species effectiveness against CL and VL. As this might not be feasible, the development of a VL-specific vaccine remains an important global health priority.

First-Generation Vaccines
The only successful intervention against leishmaniasis is inoculation using virulent parasites, a process known as leishmanization (LZ). This ancient practise involves the administration of cutaneous Leishmania parasites to a discrete skin location, allowing a self-healing lesion to form. Initial immunological exposure then protects the individual from further infection and lesion development. LZ was traditionally practised by directly transferring infectious material from cutaneous lesions to uninfected individuals. However the establishment of an in vitro culture system in the early 20th century led to the large-scale production of promastigote forms of Leishmania for expanded clinical use. LZ induces a controlled, but full, infection and was successfully used as a prophylaxis throughout the Soviet Union, Asia, and the Middle East, with reported efficacy levels up to 100% [16,17]. However LZ was largely abandoned due to safety issues associated with the use of live vaccines. Also, standardisation of the inoculum proved difficult as parasites used for LZ experience a dramatic loss of infectivity when subject to repeated subculturing [18]. Infection with live Leishmania also causes immunosuppression, which resulted in reduced immune responses to childhood vaccines and threatened the efficacy of immunization programmes [19,20]. Currently only one country, Uzbekistan, employs the use of LZ, where a mixture of live and dead L. major is licensed as a vaccine for high-risk populations [21]. As LZ is the only vaccine strategy against Leishmania with proven efficacy in humans, efforts are being made to improve the safety of this practise. The inclusion of killed parasites in the inoculum and the use of adjuvants that promote rapid immune responses reduce the severity of primary lesions and accelerate wound healing during LZ [16,22].
Research into first-generation vaccines based on wholecell, killed Leishmania parasites dates back to the late 1930s, when pioneering work by Brazilian scientists demonstrated that killed parasites showed efficacy as both therapeutic and prophylactic vaccines [23]. Over the ensuing decades numerous preparations of killed parasites were tested, either alone or in combination with a variety of different adjuvants. Although displaying well-tolerated safety profiles, to date no first-generation vaccine using killed parasites has demonstrated sufficient efficacy as a prophylactic vaccine to be used in widespread control programmes [24]. Most vaccine studies focus on CL, and there have been no clinical trials of first-generation vaccines produced from VL Leishmania species. Due to the strongly conserved genomes of the Leishmania species, it is anticipated that human VL trials will follow any successful CL immunization program. Whether the same vaccine can show efficacy against both CL and VL remains to be determined. Interestingly, killed parasite vaccines using an alum-precipitated autoclaved L. major (ALM) given with a BCG adjuvant have shown promise as vaccines for VL and PKDL [25]. When given to patients with persistent PKDL in combination with antimonial therapy this vaccine showed enhanced cure rates and lower incidence of relapse as compared to antimonial treatment alone [26]. Based on these initial studies, recommendations have been put forward for expanded trials to examine the prophylactic and therapeutic effects of the alum-ALM + BCG vaccine for PKDL and VL [27].
Evidence from experimental animal models supports the development of first-generation VL vaccines. Complete soluble antigen (CSA) from an attenuated L. donovani strain was effective as both a therapeutic and prophylactic vaccine in susceptible mice, without the use of an adjuvant [28]. Importantly, CSA immunization was effective against both pentavalent antimony sensitive and resistant strains of L. donovani. Vaccination with purified excreted-secreted antigens from L. infantum promastigotes (LiESAp) fully protected dogs from experimental challenge and induced a long-lasting cell-mediated immunity [29].
A major advantage of first-generation vaccines is that they are conceptually simple and relatively easy to produce in Leishmania endemic countries at low cost. However standardization of vaccines derived from cultured parasites is difficult, and this has hindered commercial development efforts. The route of administration, formulation, and adjuvant are also important considerations in the development of whole-parasite vaccines, and optimisation is essential for the induction of protective immune responses. The most recent clinical trials of first-generation vaccines have demonstrated a good safety profile but have not conferred significant levels of protection for use as prophylactic vaccines. However promising results from trials using therapeutic vaccination in combination with chemotherapy warrant further investigation.

Second-Generation Vaccines
The development of Second-generation vaccines for Leishmania has included recombinant proteins, polyproteins, DNA vaccines, liposomal formulation, and dendritic cell vaccine delivery systems. A variety of different molecules have been tested to date with varying degrees of efficacy (Table 1). In general VL vaccination studies have been hampered by the lack of a suitable animal model. The natural combination of dogs and L. infantum [30] and L. donovani in golden hamsters [31] reproduces many features of human VL. The canine model is particularly useful in evaluating vaccine candidates since successful vaccination of dogs might control the spread of disease to humans in endemic areas where the dog is the reservoir of the parasite [32]. However, both models suffer from lack of immunological reagents and assays needed for the characterisation of immune responses. Therefore, the mouse model of VL has been widely used to assess vaccine candidates. While experimental VL infection in mice does not fully reproduce the disease observed in humans, mice are competent hosts for both L. donovani and L. infantum and exhibit organ-specific pathology in the liver and spleen. Other major advantages of the mouse model are that it is amenable to genetic manipulation to create mutants with specific deficiencies in the immune system and a wide range of immunological reagents is available.
Only a small number of recombinant proteins have been tested against VL in murine models. Early studies showed that promastigote-derived membrane protein dp72 protected mice against L. donovani infection [59,82], but there has been no further advance on the use of this antigen for the development of vaccines. The L. donovani amastigote LCR1 protein containing 67-amino-acid repeats homologous to repeats in a Trypanosoma cruzi flagellar polypeptide was tested for protection in mice. Recombinant protein led to partial protection against L. infantum challenge [44], while immunization with BCG-LCR1 elicited better protection [45]. Vaccine efficacy was influenced by the site of immunization with subcutaneous administration superior to intraperitoneal inoculation [45]. Recombinant hydrophilic acylated surface protein B1 (HASPB1), a member of a family of proteins expressed only in metacyclic and amastigote stages, has shown efficacy in an experimental mouse model of VL [8]. This vaccine did not require the use of adjuvant, and protection was associated with the induction of antigen-specific, IFN-γ producing CD8 + T cells, a mechanism similar to DNA vaccination [8]. Immunization with the L. donovani A2 cysteine proteinase delivered as recombinant protein or as DNA also afforded protection against experimental challenge infection [60,61]. Other antigens tested include amastigote cysteine proteases (CPs) [40], kinetoplastid membrane protein-11 (KMP-11) [65], amastigote LCR1 [45], leishmanial antigen ORFF [63], and NH36, a main component of the fucose-mannose ligand [47].
Apart from defined single molecules, multicomponent vaccines have been shown to protect against VL in experimental infection systems. Recombinant Q protein formed by fusion of antigenic determinants from four cytoplasmic proteins from L. infantum (Lip2a, Lip2b, P0, and histone H2A) coadministered with live BCG protected 90% of immunised dogs by enhancing parasite clearance [49]. To date, only one multicomponent vaccine, Leish-111f, has been assessed in clinical trials [83]. Leish-111f is a single polyprotein composed of three molecules fused in tandem: the L. major homologue of eukaryotic thiol-specific antioxidant (TSA), the L. major stress-inducible protein-1 (LmSTI1), and the L. braziliensis elongation and initiation factor (LeIF) [83]. There is some evidence that the Leish-111f vaccine can also induce partial protection against VL in animal models [56]; however, it failed to protect dogs against infection and did not prevent disease development in a Phase III vaccine trial in dogs [55]. An optimized version, known as Leish-110f, has recently demonstrated strong immunogenicity and some protective efficacy against L. infantum in mice [84]. The Leish-111f vaccine is moving forward into clinical trials as LeishF1 and is being trialled in combination with the MPL-SE adjuvant. This adjuvant consists of monophosphoryl lipid A, a potent TLR4 agonist, formulated with the antigen as a stable emulsion. A recent small-scale clinical trial in a L. donovani endemic area showed Leish-F1-MPL-SE was safe and well tolerated in people with and without prior VL exposure and induced strong antigen-specific T cell responses [85]. In addition to recombinant proteins, DNA has been extensively tested as means of vaccine delivery. The induction of Th1 responses leading to strong cytotoxic T cell immunity is a general property of DNA vaccines [86], and a growing body of evidence implicates CD8 + T cells in protective antileishmanial responses [87]. The PapLe22 antigen, which is recognised by T cells from VL patients [88], was administered as a DNA vaccine and led to a marked decrease in parasite burden in immunised hamsters [46]. However stimulation of peripheral blood mononuclear cells from VL-infected individuals with recombinant PapLe22 induced IL-10 production [88], which is associated with VL pathogenesis in humans [89]. The Leishmania homologue for receptors of activated C kinase (LACK) is the most extensively studied DNA vaccine against both cutaneous and visceral leishmaniasis, but has shown inconsistent outcomes. DNA vaccination with a plasmid harbouring the LACK gene coadministered with, or without, IL-12 induced robust, long-lasting protection against L. major challenge in mice, which was dependent on CD8 + T cells [90][91][92]. In a heterologous system, priming with L. infantum LACK followed by a vaccinia booster afforded protection against L. major infection [93]. The prime-boost regimen was also employed to immunise dogs against L. infantum infection and elicited protective responses in 60% of vaccinated animals [38], but this positive outcome has been overshadowed by studies where immunisation with LACK offered no protection. In an experimental mouse model the LACK DNA vaccine induced strong Th1 responses, but failed to protect against L. donovani challenge [58]. Other studies in the L. infantum mouse model confirmed that LACK DNA vaccination does not confer protection against VL despite the presence of Th1 responses [36]; however, a strategy using a heterologous prime-boost vaccination using DNA and vaccinia viruses has shown some efficacy [35]. The heterologous DNA-prime protein-boost approach has also shown success for other VL vaccine antigens such as ORFF [63] and cysteine proteinases [94]. Heterologous prime-boost with gp63 antigen with CpG-ODN as adjuvant provided durable protection against L. donovani challenge in an experimental mouse model and was associated with robust cellular immune responses [95]. As gp63 is a major surface protein present on both amastigote and promastigote forms and shows a high homology between VL species [96] it is an attractive target for further vaccine development.
Insights into the role of the innate immune system, in particular dendritic cells (DCs), have provided the impetus for the use of DCs as a delivery system for Leishmania antigens [97][98][99]. DCs loaded with L. donovani soluble extract and expressing high levels of IL-12 induced protection in the mouse model of VL when used as a therapeutic vaccine [100]. Moreover, coadministration of DCs with antimonial therapy resulted in complete clearance of parasites from the liver and spleen, unlike DC immunisation alone which was not able to clear the infection from these organs [101].
Liposome formulations have been adopted as Leishmania drug delivery systems, and liposomal Amphotericin B is the current preferred drug treatment for VL in resource-rich settings [102]. Vesicle delivery systems are also being considered for VL vaccines and have been shown to adjuvant protein antigens and induce sustained Th1 immune responses [103]. These delivery systems have shown some protection against L. donovani infection in experimental mouse models [104] and provide a new approach to the development of VL vaccines.
Recently, Peters et al. [105] demonstrated that sand fly transmission of parasites abrogates vaccine-induced protective immunity. While mice vaccinated with killed parasites were refractory to a needle challenge, they were susceptible to the sand fly inoculum implying that the protective responses in vaccinated mice were either not generated or not maintained. These data provide a rationale for the inclusion of sand fly saliva components, which are specific to natural infection, in vaccine design. The sand fly injects Leishmania parasites in the presence of saliva, which contains a range of pharmacologically active molecules that can modulate host's immune and inflammatory responses and facilitate establishment of infection. For a number of years salivary gland antigens have been targeted as potential candidates for antileishmanial vaccine development, primarily against L. major. Nevertheless, it has been shown that children who underwent anti-VL delayed-type hypersensitivity (DTH) conversion also had increased titers of antibodies directed to sand fly saliva suggesting that mounting an effective antileishmanial response might be linked to neutralization of saliva components [106].

Development of Canine Vaccines
Eliminating animal reservoirs has been an essential public health tool for the control of many zoonotic diseases, such as rabies [107] and brucellosis [108]. Canines, particularly domestic dogs, are the main reservoir for VL species and are considered the main source of zoonotic transmission to humans. The development of an effective canine visceral leishmaniasis (CVLs) vaccine represents a cost-effective tool for interrupting the transmission cycle and controlling zoonotic VL infection in humans.
CVL is widespread throughout South America [9] and the Mediterranean [109] where L.infantum is the most significant causative agent of disease. L. donovani is considered to be zoonotic, but as yet there has been no clear identification of the reservoir host animal [110]. Asymptomatic infection is common in dogs, and as a large reservoir of parasites are present in the skin, asymptomatic animals are a major source of infection for vector transmission [10]. Human VL is an emerging disease in many areas of the world, including Northern Europe [111] and North America [112], and the spread of VL into nonendemic areas is often preceded by increased incidence of canine infection. There is concern that increased mobility of dogs and changes in vector habitat will result in increased transmission of human VL in previously nonendemic areas [113].
Treatment of CVL shows low efficacy with drugs successfully used for human VL chemotherapy, and drug treatment of dogs rarely results in cure [114]. Control programmes for CVL have a demonstrated capacity to reduce the prevalence of human VL disease following interventions that target dog populations in endemic regions [115]. However these public health campaigns are often complex and expensive to maintain, leading to varying degrees of efficacy. The use of insecticide-impregnated collars can reduce the risk of contracting CVL [116], but is costly and difficult to implement at a national level. The culling of seropositive dogs has long been recommended in Brazil; however this approach has not lead to a reduction in the number of human VL cases and may be of limited value [117]. Therefore the development of vaccines against CVL is an attractive approach to controlling infection in dogs, reducing the parasite reservoir and thus reducing the risk of transmission of VL to human populations.
Immunological characterisation of CVL reveals cellular and humoral immune responses comparable to human infection, including immune dysregulation and increased IL-10 which is associated with disease manifestation and progression [118]. Disease resistance is associated with strong Th1-type immune responses, including IFN-γ expression by antigen-specific T cells. Thus, analogous to a human VL vaccine, an effective CVL vaccine needs to induce strong and long-lasting cell-mediated immunity. Adjuvant choice must be carefully considered for CVL interventions, as live BCG is not appropriate for use in dogs and the identification of appropriate and effective adjuvants will be essential for safe and effective CVL vaccines [119]. In addition, sand fly components are being considered for inclusion in CVL vaccine. Reactive antibodies to two sand fly saliva components (LuLo-D7 and LuLo YELLOW) were identified in infected dogs and proposed as possible vaccine candidates against CVL [120]. Evaluation of a killed Leishmania vaccine containing sand fly saliva extract indicated that the vaccine is highly immunogenic and provided support for further development of saliva components as candidates for anti-VL vaccine [121]. This is supported by vaccination studies using the hamster VL model, showing that salivary protein LJM19 was able to protect hamsters from fatal infection with L. infantum [122]. In addition, immunization with salivary proteins LJM17 and LJL143 induced strong cellular and humoral responses in dogs and might be an advantageous addition to anti-CVL vaccine [123].
Currently there are two commercially available CVL vaccines, Leishmune and Leishtec, and new vaccines under development include recombinant antigen vaccines and both live and killed whole-cell vaccines.
The Leishmune vaccine was the first commercially licensed vaccine for CVL, produced by Fort Dodge Animal Health and has been available in Brazil since 2004 [124]. This vaccine consists of the fucose mannose ligand (FML) isolated from L. donovani plus a saponin adjuvant. FML is a glycoprotein mixture, and the surface glycoconjugate GP36 is the major immunogen component [125]. This vaccine induced a significant and strong protective effect during phase III trials in dogs living in a VL-endemic area in Brazil with a vaccine efficacy as high as 80% [77,126]. This protection lasted up to 3.5 years following vaccination, indicating induction of a long-lasting immunity [77]. As Leishmune-vaccinated dogs showed a complete absence of parasites, this renders them noninfectious and contributes to the breakdown of the zoonotic VL transmission cycle [79]. During phase III trials of Leishmune there was a concomitant reduction in human VL cases in districts where dogs were vaccinated [126] demonstrating that CVL vaccination interrupts the transmission of disease to humans. FML antigens are present on the surface of Leishmania parasites throughout the life cycle, and antibodies raised in vaccinated dogs prevented the binding of procyclic promastigotes to the sand fly midgut [78]. Thus Leishmune acts as a transmission blocking vaccine by clearing parasites from the animal reservoir and preventing survival of the parasite in the sand fly vector. Currently the Leishmune vaccine is used as a prophylactic and is recommended for asymptomatic noninfected dogs. However studies show that Leishmune is effective as a therapeutic vaccine for naturally infected dogs [127], particularly when given in combination with chemotherapy [128]. Emerging wide-scale field studies reveal that Leishmune decreases the incidence of both human and canine visceral leishmaniasis after dog vaccination with Leishmune [129].
A second vaccine, known as Leish-Tec, is being commercially developed by the Hertape Calier Saúde Animal and consists of adenovirus expressing the L. donovani A2 antigen. Whilst the results from phase-III trials of Leish-Tec are yet to be published it is known that immunization with a recombinant A2 protein elicits protection against the onset of clinical VL in experimental dog infections [51]. The recombinant adenovirus encoding the A2 gene was capable of inducing strong Th1-type immune responses in vaccinated mice and reduced parasite burdens following challenge with VL parasites [130]. Together these studies indicate that A2 is an important candidate antigen for the development of CVL vaccines, and future studies should report on the impact of this intervention on both canine and human VL infection.
As many of the clinical and immunological features of CVL are similar to those observed in human VL, experimental challenge in dogs represents a useful system for evaluating the efficacy of vaccine candidates. The Leish-111f + MPL-SE vaccine is a leading vaccine candidate from human VL and has shown therapeutic efficacy in recent CVL trials [57]. Live attenuated parasites vaccines are also being explored in canine models, including a drug-attenuated line of L. infantum established by culturing promastigotes under gentamicin pressure. The attenuated L. infantum vaccine strain did not induce clinical symptoms of VL in dogs and provided protection from subsequent challenge with live virulent L. infantum [131].
The elimination of human VL will be difficult to achieve in the presence of persisting animal reservoirs, and veterinary intervention is an important tool for reducing the global burden of human VL disease. The identification of measurable and reliable biomarkers of immunogenicity and protection induced by CVL vaccines may also be informative for human VL vaccine efforts.

Live Attenuated Vaccines
Historically the most successful vaccines against intracellular pathogens have been based on live attenuated organisms. Vaccination strategies using live attenuated Leishmania parasites are attractive as they closely mimic the natural course of infection and may elicit clinically protective immune responses. A live attenuated vaccine strain would present a full complement of Leishmania antigens to the host immune system along with appropriate pattern-recognition molecules for the parasite. Live vaccines also deliver antigens to the correct cellular and tissue compartments for appropriate processing and presentation to the host immune system. Together, this enhances the capacity of live attenuated vaccines to promote antigen-specific effector and memory immune responses that confer long-lasting protective immunity.
The development of robust in vitro culture systems for growth and differentiation of Leishmania promastigote and amastigote life cycle stages has enabled the production of attenuated vaccine strains. It should be noted that most research in this area has utilized CL strains, such as L. major; however the attenuation techniques are broadly transferrable to VL causing species. It is has been known for some time that long-term in vitro culture of promastigote parasites leads to a loss of virulence in vivo. Studies in experimental mouse models of CL have shown that infection with cloned avirulent lines provides clear protection against 8 Journal of Tropical Medicine a virulent challenge infection [132]. Avirulent strains of the VL species L. donovani and L. infantum have been generated by repeated in vitro subculture of promastigotes in the presence of gentamicin [133]. These drug-attenuated promastigotes were able to invade macrophages but could not survive as intracellular amastigote forms [133]. Drugattenuated L. infantum was avirulent in an experimental canine model, induced strong cellular immunity production and protection against challenge with live virulent L. infantum [131]. Early experiments showed that γirradiation rendered Leishmania parasites nonpathogenic and infection protected against challenge in a cutaneous Leishmania model. Protection depended on the presence of viable irradiated parasites, suggesting that transformation into amastigote forms is required for efficacy. Interestingly the underlying mechanism of protection may relate more to the induction of tolerization rather than immunization in this system [134]. Other approaches to the generation of attenuated parasites include chemical mutagenesis screens selecting for temperature sensitive CL strains [135] which are avirulent during infection and significantly protect against subsequent challenge.
The major concern regarding these approaches to attenuation is that the underlying genetic mechanisms are not defined. This creates safety concerns as the stability of parasite attenuation is uncertain and parasites could revert to a virulent form. Conversely, a progressive loss of virulence may occur, resulting in parasite lines that are incapable of establishing infection or inducing protective host responses. A loss of parasite virulence due to long-term in vitro culture has been demonstrated in both human patients undergoing leishmanization and experimental mouse models [19]. Thus in the absence of a clear genetic profile, nonspecific parasite attenuation is not acceptable for the development of a human VL vaccine.
Over the last few decades the development of a powerful "genetic toolkit" for Leishmania species has enabled research involving transgenic parasites [136]. L. donovani and L. infantum parasites can be stably transfected using integrating expression constructs that target genes for disruption by homologous recombination. As Leishmania organisms are diploid throughout their lifecycle, the production of null mutants requires each allele of a gene to be targeted individually with genetic constructs containing two different and independent selectable markers. The recent availability of Leishmania genome sequences has facilitated the identification and in-depth analysis of parasite genes crucial for infection and virulence. Comparative genomics studies of Leishmania species have shown a highly similar gene content and gene order and annotation studies have revealed only a few species-specific genes [11]. Increased knowledge of potential parasite virulence factors and a greater understanding of the antigens involved in the acquisition of immunity have generated much interest in the development of genetically attenuated parasite vaccines.
To date, there have been two general approaches to the genetic attenuation of Leishmania parasites. First, by deletion of genes encoding virulence factors or the enzymes responsible for their synthesis and second, by targeting genes essential for intracellular survival. Gene targeting aims to produce parasites that are capable of being produced and manipulated in vitro, usually in promastigote form, but incapable of sustaining virulent infection in the host, in amastigote form. The first genetically attenuated parasite vaccine was the L. major dihydrofolate reductase-thymidylate synthase (dhfrts) knockout, which targeted an essential metabolic gene [137]. This null mutant was able to establish a persistent infection in experimental mouse models, but remained avirulent with respect to disease. Importantly vaccination with dhfr-ts knockout parasites elicited substantial protective immunity, as mice were resistant to subsequent challenge with virulent L. major. Although further experiments in nonhuman primate models failed to show protection these initial studies provided proof of principle for the safety and immunogenicity of live attenuated Leishmania vaccines [138]. Drug-sensitive Leishmania mutants containing suicide genes [139][140][141] are also being developed for use during leishmanization, and inducible suicide mutants in L. amazonensis have shown protective efficacy in an experimental hamster model [142].
To date, only a small number of studies have focused on generating attenuated forms of the VL species L. infantum and L. donovani as a route to the production of an attenuated VL vaccine. One approach has targeted the transporters for the metabolic precursors of the folate pathway, as Leishmania parasites are auxotrophic for folate and pterin [143]. L. donovani parasites lacking the main biopterin transporter (BT1) showed a marked reduction in infectivity in an experimental mouse model, and this attenuated strain conferred protection to subsequent challenge with wildtype L. donovani [144]. Parasites incapable of intracellular reproduction were produced by targeting centrin, a calcium binding cytoskeletal protein. A loss of centrin from L. donovani parasites did not affect the growth of promastigote forms, but null mutants were unable to survive as axenic amastigotes or in human macrophages in vitro [145]. Immunization of mice and hamsters by infection with centrin deficient L. donovani protected against virulent homologous challenge. Importantly the centrin null vaccine strain elicited parasite-specific Th1 responses which strongly correlated with sustained protection and also induced a level of cross protection against L. braziliensis infection [146].
The genetic attenuation of Leishmania does not necessarily require the production of null mutants. Deletion of one allele of the L. infantum silent information regulatory 2 (SIR2) locus was sufficient to prevent amastigotes from undergoing intracellular replication in macrophages. Immunization with L. infantum lacking one SIR2 gene copy elicited strong parasite-specific T cell responses and conferred complete protection against virulent challenge in a VL mouse model [147].
Other approaches to developing live attenuated parasites as VL vaccines have utilised nonpathogenic Leishmania species, an approach comparable to the use of BCG as a vaccine against Mycobacterium tuberculosis infection. The lizard protozoan parasite L. tarentolae has never been found to be associated with any human leishmaniasis and is considered nonpathogenic. Whilst L. tarentolae is capable of infecting mammalian cells and transforming into amastigotes, the parasite does not cause clinical symptoms of disease in either mouse or hamster models [148]. In experimental vaccine trials L. tarentolae elicited a strong Th1-driven protective immune response and conferred protection against infectious challenge with L. donovani in a susceptible mouse strain [149]. The use of L. tarentolae as a vaccine vector to deliver specific Leishmania antigens in the context of a live infection has also been explored. The L. donovani A2 antigen was expressed in L. tarentolae, which normally lacks this protein [150] and used as a vaccine strain in an experimental mouse model. Vaccination protected susceptible mice against L. infantum challenge and was associated with the production of high levels of IFN-γ production [151].
The use of live attenuated vaccines provides a promising vaccination strategy for VL; however safety issues regarding the use of genetically attenuated parasites as vaccines still need to be addressed. Many of the proposed live attenuated vaccines induce long-lasting immunity to reinfection by maintaining a low level asymptomatic infection. The establishment of subclinical infection is particularly valuable as the persistence of antigen is thought essential for the generation of effective memory responses to Leishmania. However reactivation of Leishmania has been observed in patients who are immunocompromised, such as following HIV infection, thus the safety of attenuated parasites that induce a subclinical infection will need to be carefully assessed.
Transgenic parasites provide an enticing lead for vaccine development. A continuing synergy between molecular and immunological approaches to the development of VL vaccines will accelerate development of the next generation of therapeutics. In addition, transgenic parasites are invaluable tools for understanding host-parasite interactions [152] and inform vaccine design by providing insight into immunity and pathogenesis during VL.

Concluding Remarks
Preventive vaccines are recognized as the best and most costeffective protection measure against pathogens and save millions of lives across the globe each year. Leishmania vaccine development has proven to be a difficult and challenging task and is hampered by an inadequate knowledge of disease pathogenesis, the complexity of immune responses needed for protection, and the cost of vaccine development. The burden of VL is concentrated in resource poor nations, and a lack of political will and philanthropic investment further aggravates the situation. However, the rise of biotechnology industries in endemic countries, such as India, may provide an impetus for VL vaccine development and investment. A recent clinical trial in India assessed the safety and immunogenicity of the LEISH-F1+MPL-SE vaccine [85] which is the only Second-generation vaccine currently in clinical development for human VL. There are currently several new European-based VL vaccine efforts including a synthetic vaccine RAPSODI (http://www.fp7-rapsodi.eu/) [153], a DNA-based LEISHDNAVAX (http://www.leishdnavax.org/) [154], and an adenovirus vectored therapeutic vaccine (Paul Kaye, personal communication). New adjuvants are also being developed, and there are several clinical vaccine trials in progress and in planning [18]. Given the rapid progress in the fields of parasite immunology and genomics, a successful anti-Leishmania vaccine should be achievable sooner rather than later. There is a clear need for greater investment in research and development to move promising vaccine leads along the development pathway toward an effective, affordable VL vaccine.