**2. Vaccine research for leishmaniasis**

Nevertheless, concomitant immunity, *i.e.* efficient protection upon a challenge due to the longtermandsimultaneouspersistenceofthepathogen, seems tobe ahallmarkinleishmaniasis [18].

Dogs are primary reservoir hosts of zoonotic visceral leishmaniasis (ZVL) caused by *Leishmania infantum* and play a key role in the long-term maintenance of the parasite in the endemic areas of Mediterranean countries, the Middle East, Asia and Latin America. Epidemiological surveys estimate that, for example in western Mediterranean countries, seroprevalence ranges from 5 to 37%, varying from region to region depending on ecological aspects. Nevertheless, surveys based on PCR diagnosis demonstrated high infection rates in endemic areas, for example 80% in Marseille, France [19], and 67% in Majorca, Spain [20]. Longitudinal studies in Italy have also shown high incidences (40-92%) during the season of transmission [21]. Importantly, not all infected dogs develop canine leishmaniasis; more than 50% of infected dogs remain asymptomatic after infection, though it has been shown that these asymptomatic carriers are

The high prevalence of infected dogs in endemic areas, their common presence in the domestic surroundings where ZVL transmission occurs, and the high infectiousness of both sympto‐ matic and asymptomatic animals makes that *Leishmania*-infected dogs represent not only a serious veterinary but also an important public health problem. Infected dogs have been associated with the emergence of new *foci* of ZVL, like those in the North of Argentina, where the appearance of human cases is preceded by those of canine leishmaniasis [23], and also with the spread of VL observed in large Brazilian cities [24] and the northward spread of the disease reported in Italy [25]. Therefore, the control of parasite-infected dogs is of prime urgency to

The outcome of *Leishmania* infection in dogs is variable and depends on the persistence and multiplication of the parasite and the immune response of the animal. Not all the infected dogs will develop clinical disease, part of them can control the expansion of the parasite and spontaneously cure the infection; in others, the infection is subclinical for an undefined time (years or even the whole life) during which the animal remains asymptomatic. Few than 50% of infected animals do not have (or have lost) the capacity to control the parasite, in these cases being distributed extensively throughout the organism: spleen, liver, lymph nodes, bone marrow, kidney, skin, etc., (as opposed to what occurs in humans, where the parasite is normally limited to bone marrow, spleen and liver) [26]. In these dogs the disease progresses, the parasite burden and the *Leishmania*-specific antibody levels increase, and after two to four

The natural history of canine leishmaniasis mostly depends on the efficacy of the dog´s immune response to *L. infantum* infection which determines the development of resistance or susceptibility to the disease. In general, resistance is associated with low levels of specific antibodies and presence of a predominant Th1 cell-mediated response against the parasite, with the production of IFNγ that is able to stimulate, in collaboration with TNFα, the leish‐ manicidal activity of macrophages mediated by the induction of iNOS. Absence of symptoms is related with high levels of IFNγ expression in the peripheral blood as detected by quanti‐

reduce the number of cases of human VL by decreasing prevalence in dogs [26].

months of incubation the symptoms of canine leishmaniasis appear [27].

**1.2. Canine leishmaniasis**

280 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

also infective to sandflies [22].

The ideal vaccine for leishmaniasis should be safe, effective, long lasting, transversal to all infective *Leishmania* species and affordable, to be available to the populations most in need.

### **2.1. Animal models for vaccine research for leishmaniasis**

Much of the knowledge generated from leishmaniasis research come from experimental infections in animal models. Differently from human population and natural infections, the most common models of disease employed in leishmaniasis research are based on infection of inbred mice with cloned lines of parasites. These experimental settings reduce the variety of factors that play a role on the disease manifestation, such as host's genetic background and immune competence, concomitant infections with other microorganisms, autoimmune or inflammatory diseases or drug treatments that may affect the fitness of the immune system, diversity of parasite's strains and species, site of infection and inoculum dose, infecting sand fly's species, etc. However, in comparison to the natural transmission and disease, the same limitations are also the major advantages, as in laboratorial settings researchers control all those variables and are able to focus on their specific target to unravel the molecular and immunological mechanisms behind leishmaniasis.

Many models of leishmaniasis have been tested, although none is able to mimic the exact pathology of cutaneous and, principally, visceral human diseases, or to develop the same immune responses. Despite valuable information has come from animal models, careful generalizations must be done when transposing it to the human disease.

The animal species applied on studies of human CL is almost exclusively the mouse (*Mus musculus*). Inbred strains are experimentally infected by subcutaneous or intradermal route with millions of promastigotes cultivated *in vitro* or axenic or tissue-derived amastigotes. The mice's genetic background has a major impact on the severity of the disease. For instance, when infected with a high dose BALB/c strain develops extensive skin lesions that spread away from the inoculation site leading to death of the animal, while C57BL/6 and CBA/N are able to control the infection and skin ulcers self-heal with time [38].

Considering animal models for VL, golden hamsters (*Mesocricetus auratus*) are among the best mimicking model of the human disease. Despite the artificial route (intracardiac or intrave‐ nous) and the high amount of parasites usually inoculated, *L. infantum*- or *L. donovani*-infected hamsters show heterogeneous phenotypes of infection, with animals that are asymptomatic, oligosymptomatic or polysymptomatic, in quite a good correlation with human and also canine epidemiology in endemic areas. Symptomatology comprise weight loss, uncontrollable increase in the splenic, hepatic and bone marrow parasite loads, hepatosplenomegaly, pancytopenia, hypergammaglobulinemia and ultimately death (Carrillo *et al.*, submitted, 2013 and [39]). Due to the lack of specific reagents needed to study the immunological mechanisms associated with *Leishmania* infection, the hamster model has been put apart and neglected over the mice models. Visceral leishmaniasis in mouse do not fully resemble the human nor canine disease, but the availability of numerous strains genetically modified and an endless offer of anti-mouse antibodies make the mouse the most preferred model to understand the hostparasite interactions and the immunological aspects of visceral leishmaniasis. However, in the scope of vaccine development and drug screening, where more than the mechanism behind the most important read-out is efficacy (*i.e.* parasite loads and pathology), golden hamsters are the most appropriate rodent model for the human disease.

The use of dogs (*Canislupusfamiliaris*) as model ofleishmaniasis is an advantage in the way that dogs are themselves natural hosts. Some breeds, like German Shepard, Boxer and Doberman, seem to be more susceptible to natural *Leishmania* infection [40]. However, the most common breed used in laboratorial studies is the Beagle. In addition, many research studies are done with field animals. They can be artificially infected or put in natural contact with sand flies in endemic areas to test the efficacy of vaccines or anti-*Leishmania* drugs for veterinary practice or be used in the scope of model for human VL. Despite the existence of some dog-specific tools thatwouldallowthestudyoftheimmuneresponse,workingwithdogs isnotaseasyashandling mice, due to their size, the unpredictability of the infection rates, the cost of the experiments and the emotional connection that naturally exist between humans and dogs.

Non-human primates are usually confined to pre-clinical trials in humans. Some models based on artificial inoculation of rhesus macaques (*Macaca mulata*) [41], African vervet monkeys (*Chlorocebus* spp.) [42] or langur monkeys (*Presbytis entellus*) [43, 44] have been tested for *Leishmania* vaccines. Due to the close phylogenetic relation with humans and considerably good mimicking of pathology and immune responses generated upon infection (depending on the parasite and animal species), these models are attractive for vaccine research. But the difficulty on the handling, the very expensive costs and the impossibility of exposing the animals to a natural challenge are drawbacks on the use of non-human primates for *Leishma‐ nia* vaccine research.

#### **2.2. Leishmanization**

most common models of disease employed in leishmaniasis research are based on infection of inbred mice with cloned lines of parasites. These experimental settings reduce the variety of factors that play a role on the disease manifestation, such as host's genetic background and immune competence, concomitant infections with other microorganisms, autoimmune or inflammatory diseases or drug treatments that may affect the fitness of the immune system, diversity of parasite's strains and species, site of infection and inoculum dose, infecting sand fly's species, etc. However, in comparison to the natural transmission and disease, the same limitations are also the major advantages, as in laboratorial settings researchers control all those variables and are able to focus on their specific target to unravel the molecular and

Many models of leishmaniasis have been tested, although none is able to mimic the exact pathology of cutaneous and, principally, visceral human diseases, or to develop the same immune responses. Despite valuable information has come from animal models, careful

The animal species applied on studies of human CL is almost exclusively the mouse (*Mus musculus*). Inbred strains are experimentally infected by subcutaneous or intradermal route with millions of promastigotes cultivated *in vitro* or axenic or tissue-derived amastigotes. The mice's genetic background has a major impact on the severity of the disease. For instance, when infected with a high dose BALB/c strain develops extensive skin lesions that spread away from the inoculation site leading to death of the animal, while C57BL/6 and CBA/N are able to control

Considering animal models for VL, golden hamsters (*Mesocricetus auratus*) are among the best mimicking model of the human disease. Despite the artificial route (intracardiac or intrave‐ nous) and the high amount of parasites usually inoculated, *L. infantum*- or *L. donovani*-infected hamsters show heterogeneous phenotypes of infection, with animals that are asymptomatic, oligosymptomatic or polysymptomatic, in quite a good correlation with human and also canine epidemiology in endemic areas. Symptomatology comprise weight loss, uncontrollable increase in the splenic, hepatic and bone marrow parasite loads, hepatosplenomegaly, pancytopenia, hypergammaglobulinemia and ultimately death (Carrillo *et al.*, submitted, 2013 and [39]). Due to the lack of specific reagents needed to study the immunological mechanisms associated with *Leishmania* infection, the hamster model has been put apart and neglected over the mice models. Visceral leishmaniasis in mouse do not fully resemble the human nor canine disease, but the availability of numerous strains genetically modified and an endless offer of anti-mouse antibodies make the mouse the most preferred model to understand the hostparasite interactions and the immunological aspects of visceral leishmaniasis. However, in the scope of vaccine development and drug screening, where more than the mechanism behind the most important read-out is efficacy (*i.e.* parasite loads and pathology), golden hamsters

The use of dogs (*Canislupusfamiliaris*) as model ofleishmaniasis is an advantage in the way that dogs are themselves natural hosts. Some breeds, like German Shepard, Boxer and Doberman, seem to be more susceptible to natural *Leishmania* infection [40]. However, the most common breed used in laboratorial studies is the Beagle. In addition, many research studies are done with field animals. They can be artificially infected or put in natural contact with sand flies in

generalizations must be done when transposing it to the human disease.

immunological mechanisms behind leishmaniasis.

282 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

the infection and skin ulcers self-heal with time [38].

are the most appropriate rodent model for the human disease.

Until date, the only successful, long-lasting strategy for human immunization against leish‐ maniasis is the leishmanization process. It consists on the inoculation of live virulent parasites in a hidden area of the skin of healthy people with the purpose of development of immunity for protection when the individuals are challenged by a natural infection. Leishmanization showed 100% protection when used as prophylaxis for cutaneous leishmaniasis (CL) through‐ out the ex-Soviet Union, Asia, and the Middle East [45]. Due to risk of complications in healthy people and difficult standardization of the live *L. major* inoculum, this procedure was mostly abandoned. However, this is still a current practice in Uzbekistan [45] and a few years ago it was reported to be applied in the evaluation of the efficacy of new vaccines [46].

A "natural" form of leishmanization may be the reason why in Sri Lanka so many cases of CL by *L. donovani* are reported while VL is rare [47]. McCall *et al*. have recently reproduced this scenario in the BALB/c model, immunizing the mice subcutaneously with a dermotropic *L. donovani* strain from Sri Lanka followed by intravenous challenge with a viscerotropic autochthonous strain, and indeed, partial protection was obtained in the liver of the infected mice [48]. The authors attributed the ability of the cutaneous strain to protect against the challenge with the visceral strain to a probable great similarity between the two *L. donovani* strains; this hypothesis may justify the opposing phenotype observed by others [49]. Also, an epidemiological study in Sudan indicated that only individuals previously negative for leishmanin (Montenegro skin test) developed VL, thus, though without scientific evidences, leishmanin-positive individuals that were possibly formerly infected with *L. major* were protected against the visceral disease [50].

#### **2.3. First generation vaccines**

First generation vaccines comprise whole killed parasites and live attenuated parasites. They were primarily developed to overcome one of the major concerns related to leishmanization: the risk of disease development in immunocompetent persons and the total improperness for immunosuppressed patients for this same reason.

#### *2.3.1. Killed parasites*

With more or less success, some examples of killed vaccines include *L. braziliensis* crude antigens tested in dogs [51] and trivalent (*L. braziliensis* + *L. guayanensis* + *L. amazonensis*) phenolkilled whole *Leishmania* promastigotes with bacille Calmette-Guérin (BCG) as adjuvant in Ecuadorian children [52]. According to a meta-analysis conducted in 2009 by Noazin *et.al.* to evaluate the efficacy of the clinical trials performed with whole killed parasites in endemic areas since 1970's, with the exception of this latter in Ecuador, none of the other eight clinical trials considered (based on autoclaved *L. major* (ALM) with BCG tested against CL in the Old World and *L. amazonensis* or multivalent preparations inactivated with merthiolate used against CL in the New World) showed significant protection against natural infection [53]. A new option was tested recently: a killed but metabolically active (KBMA) *L. infantum*. This vaccine showed partial protection in spleen and liver of BALB/c mice 2 and 8 weeks after challenge triggering a mixed Th1/Th2 response but the authors claim that improved results could be obtained by adding TLR agonists and Th1 adjuvants [54].

#### *2.3.2. Live attenuated parasites*

For the live attenuated parasites many are the works reported whether using physical, chemical or genetic manipulation for reducing the virulence of the strains, or even naturally attenuated strains, like the non-pathogenic *L. tarentolae* [55]. Some of the most successful vaccine candidates for VL based on genetically altered live parasites were *L. donovani* biopterin transporter gene knockout (KO) (*BT1*−/−) [56], *L. donovani*replication deficient centrin gene KO (*Cen*−/−) [57], *L. donovani* cytochrome c oxidase complex component p27 gene KO (*Ldp27*−/−) [58], *L. infantum* silent information regulatory 2 single KO (*SIR2*+/−) [59] and *L. tarentolae* expressing *L. donovani* A2 antigen [60]. Despite showing hopeful efficiency in murine models, the promising candidates that were tested in human and canine diseases failed to protect (re‐ viewed in [61]).

#### **2.4. Second generation vaccines**

A different approach relies on recombinant proteins, polyproteins, DNA vaccines, liposomal formulations and dendritic cell vaccine delivery systems [45]; these constitute the second generation vaccines.

#### *2.4.1. Purified or recombinant Leishmania antigens and engineered polyproteins*

The *Leishmania* antigen that has been more extensively studied in the scope of a vaccine is the gp63 glycoprotein that is expressed on the surface of both the amastigotes and the promasti‐ gotes forms. The recombinant and the native proteins have been inoculated in several strains of mice as models of CL, generally showing a protective phenotype (see [62] for details). Also, an early study using monkeys revealed a partial protection against CL by *L. major* [63]. This gp63 is one of the few recombinant antigens studied in the scope of VL; Bhowmick *et al.* showed that gp63 encapsulated in cationic liposomes induced more than 80% reduction of the parasite loads in spleen and liver of BALB/c mice infected with *L. donovani* [64]. In this group of recombinant antigens, some others of the most successful candidates against VL were the amastigote-specific protein rA2, rHASPB1 (hydrophilic acylated surface protein B1), KMP-11 (kinetoplastid membrane protein-11) and rORFF (open reading frame fragment). *Li*ESA (*L. infantum* promastigotes' excreted/secreted antigens), FML (fucose-mannose ligand) and GRP78 (glucose-regulated protein 78) are the few purified antigens tested in vaccines for VL and all of them revealed at least certain degree of protection (see the reviews from Evans and Kedzierski [45] and Nagill and Kaur [62] for details and references). For CL, other antigens tested by several groups, though with conflicting results, are rLACK (*Leishmania* homologue of receptor for activated C kinase) [65-67] and PSA-2 (promastigote surface antigen 2) [68, 69].

Concerning the recombinant polyproteins, rLeish-111f (or LEISH-F1, 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)) and its non His-tag form rLeish-110f are undoubtedly the best studied and the most promising candidates for a vaccine against leishmaniasis. After having proved to protect mice with CL [70] and VL [71], rLeish-111f with MPL-SE adjuvant has also demon‐ strated to be safe and well tolerated in humans [72] as well as immunogenic in healthy subjects of endemic areas with or without previous contact to *L. donovani* [73]. Clinical trials in dogs have resulted in disparate conclusions about the efficacy of the vaccine in the prophylaxis of ZVL [74, 75], though survival of infected dogs was increased after vaccination and treatment with glucantime [76]. rLeish-110f with MPL-SE was shown to be immunogenic and protective in BALB/c mice after *L. major* and *L. infantum* challenges [77] (see [78] for complete information about the clinical trials run with rLeish-f111 and rLeish-f110).

Another polyprotein named Protein Q, composed of the fusion of four fragments of the acidic ribosomal protein Lip2a, Lip2b, P0 and histone 2A, has shown 90% protection as measured by parasite clearance in vaccinated dogs using BCG as adjuvant [79]. After testing other adjuvants in mice, 99% protection was achieved against *L. infantum* when administering Protein Q with CpG-ODN [80].

#### *2.4.2. DNA vaccines*

the risk of disease development in immunocompetent persons and the total improperness for

With more or less success, some examples of killed vaccines include *L. braziliensis* crude antigens tested in dogs [51] and trivalent (*L. braziliensis* + *L. guayanensis* + *L. amazonensis*) phenolkilled whole *Leishmania* promastigotes with bacille Calmette-Guérin (BCG) as adjuvant in Ecuadorian children [52]. According to a meta-analysis conducted in 2009 by Noazin *et.al.* to evaluate the efficacy of the clinical trials performed with whole killed parasites in endemic areas since 1970's, with the exception of this latter in Ecuador, none of the other eight clinical trials considered (based on autoclaved *L. major* (ALM) with BCG tested against CL in the Old World and *L. amazonensis* or multivalent preparations inactivated with merthiolate used against CL in the New World) showed significant protection against natural infection [53]. A new option was tested recently: a killed but metabolically active (KBMA) *L. infantum*. This vaccine showed partial protection in spleen and liver of BALB/c mice 2 and 8 weeks after challenge triggering a mixed Th1/Th2 response but the authors claim that improved results

For the live attenuated parasites many are the works reported whether using physical, chemical or genetic manipulation for reducing the virulence of the strains, or even naturally attenuated strains, like the non-pathogenic *L. tarentolae* [55]. Some of the most successful vaccine candidates for VL based on genetically altered live parasites were *L. donovani* biopterin transporter gene knockout (KO) (*BT1*−/−) [56], *L. donovani*replication deficient centrin gene KO (*Cen*−/−) [57], *L. donovani* cytochrome c oxidase complex component p27 gene KO (*Ldp27*−/−) [58], *L. infantum* silent information regulatory 2 single KO (*SIR2*+/−) [59] and *L. tarentolae* expressing *L. donovani* A2 antigen [60]. Despite showing hopeful efficiency in murine models, the promising candidates that were tested in human and canine diseases failed to protect (re‐

A different approach relies on recombinant proteins, polyproteins, DNA vaccines, liposomal formulations and dendritic cell vaccine delivery systems [45]; these constitute the second

The *Leishmania* antigen that has been more extensively studied in the scope of a vaccine is the gp63 glycoprotein that is expressed on the surface of both the amastigotes and the promasti‐ gotes forms. The recombinant and the native proteins have been inoculated in several strains of mice as models of CL, generally showing a protective phenotype (see [62] for details). Also, an early study using monkeys revealed a partial protection against CL by *L. major* [63]. This

*2.4.1. Purified or recombinant Leishmania antigens and engineered polyproteins*

immunosuppressed patients for this same reason.

284 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

could be obtained by adding TLR agonists and Th1 adjuvants [54].

*2.3.1. Killed parasites*

*2.3.2. Live attenuated parasites*

viewed in [61]).

generation vaccines.

**2.4. Second generation vaccines**

DNA vaccines are able to activate both CD4+ and CD8+ T cells through the engagement of MHC class II and MHC class I, respectively [38]. In addition, co-administration of cytokines and CpG oligonucleotides allows the modulation of the cellular immune response [81]. Besides being relatively easy to prepare and stable, another unique advantage is the appropriate folding of the intracellularly synthetized peptide on its native structure [38].

The first DNA vaccines to be studied were the classical candidates that have been tested as proteins. As single plasmids or in multicomponent DNA vaccines, there are successful examples that have shown to protect from some *Leishmania* species but not others, or were effective in some animal models but not others (see [62] for an extensive description of DNA vaccines). Among the most investigated are gp63 for CL in mice, LACK and KMP-11 for both CL and VL tested in mice, hamsters and dogs. Some reports have shown the use of the strategy of heterologous prime-boost using LACK DNA followed by administration of rLACK protein with positive results [82-86].

#### *2.4.3. Dendritic cell vaccine and liposomal formulations delivery systems*

The unique capacity of DCs in amplifying the innate defense mechanisms and providing the link between these and the acquired immune responses makes them ideal candidates for anti-*Leishmania* vaccines [87]. In the recent years, DCs pulsed with gp63 or gp63-derived peptides [88, 89], histone H1 [90] or a mixture of histones [91] delivered to mice challenged with *L. major* or DCs pulsed with KMP-11(12-31aa) peptide + CpG ODN [92] against *L. infantum* have shown to decrease lesion size and parasite loads through the production of antigen-specific IFNγ.

On another approach, the concept behind the use of liposomes to deliver *Leishmania* antigens is that they can modulate antigen presentation, enhancing antigen-specific T cell proliferation and humoral responses. Conventional liposomes are presented by MHC class II molecules, whereas the presentation *via* MHC class I requires pH-sensitive liposomes [93]. The encapsu‐ lation of rgp63 or rLmSTI1 in liposomes has proven to develop a Th1 response that protected BALB/c mice from *L. major*[94, 95] or *L. donovani* infection [64]. A different strategy using polar phospholipids from *Escherichia coli* to encapsulate *L. donovani* SLA protected hamsters from *L. donovani* infection by the production of CD4+ and CD8+ T cell-specific responses [96]. Impor‐ tantly, the route of administration of the liposomes may have a crucial role on the generation of the protective response. For example, BALB/c mice that were immunized by intravenous or intraperitoneal routes with liposomal *L. donovani* membrane antigens were protected from a *L. donovani* challenge, whereas the intramuscular or subcutaneous immunizations failed to protect [97].

#### **2.5. Adjuvants**

Adjuvants are synthetic or natural highly immunogenic components that are combined with the specific immunizing antigen with the purpose of efficiently stimulate the immune cells to mount a strong response against that antigen. Adjuvants are usually categorized in two classes. Immunostimulatory or non-particulate adjuvants are agonists of the pathogen-recognition receptors (PRRs) that localize at the surface or inside intracellular vesicles of innate immune cells [93]. These are activated by the binding of the cognate pathogen associated molecular patterns (PAMPs) (or their agonists) and signal a complex cascade of events that triggers the secretion of cytokines, chemokines and type I interferons [98]. The other class comprise the particulate adjuvants which are mineral-, lipid- or polymer-based delivery systems that, along with being transporters of the *Leishmania* antigen, are themselves immunostimulators due to their size, charge and composition; their properties can even be further improved by the decoration with other PAMP-like adjuvants [93].

In a vaccine for leishmaniasis, it is expected that adjuvants modulate the immune system towards a Th1 response, with high amounts of secreted IL-12 and IFNγ. Indeed, recombinant IL-12 has been successfully tested in animal models as a potent adjuvant. However, stimulation with IL-12 was unable to induce a strong memory response to the immunizing antigen in BALB/c mice [99]. Nevertheless, when administered as IL-12 DNA it induced long-lasting protection against *L. major* [100].

CL and VL tested in mice, hamsters and dogs. Some reports have shown the use of the strategy of heterologous prime-boost using LACK DNA followed by administration of rLACK protein

The unique capacity of DCs in amplifying the innate defense mechanisms and providing the link between these and the acquired immune responses makes them ideal candidates for anti-*Leishmania* vaccines [87]. In the recent years, DCs pulsed with gp63 or gp63-derived peptides [88, 89], histone H1 [90] or a mixture of histones [91] delivered to mice challenged with *L. major* or DCs pulsed with KMP-11(12-31aa) peptide + CpG ODN [92] against *L. infantum* have shown to decrease lesion size and parasite loads through the production of antigen-specific

On another approach, the concept behind the use of liposomes to deliver *Leishmania* antigens is that they can modulate antigen presentation, enhancing antigen-specific T cell proliferation and humoral responses. Conventional liposomes are presented by MHC class II molecules, whereas the presentation *via* MHC class I requires pH-sensitive liposomes [93]. The encapsu‐ lation of rgp63 or rLmSTI1 in liposomes has proven to develop a Th1 response that protected BALB/c mice from *L. major*[94, 95] or *L. donovani* infection [64]. A different strategy using polar phospholipids from *Escherichia coli* to encapsulate *L. donovani* SLA protected hamsters from *L.*

tantly, the route of administration of the liposomes may have a crucial role on the generation of the protective response. For example, BALB/c mice that were immunized by intravenous or intraperitoneal routes with liposomal *L. donovani* membrane antigens were protected from a *L. donovani* challenge, whereas the intramuscular or subcutaneous immunizations failed to

Adjuvants are synthetic or natural highly immunogenic components that are combined with the specific immunizing antigen with the purpose of efficiently stimulate the immune cells to mount a strong response against that antigen. Adjuvants are usually categorized in two classes. Immunostimulatory or non-particulate adjuvants are agonists of the pathogen-recognition receptors (PRRs) that localize at the surface or inside intracellular vesicles of innate immune cells [93]. These are activated by the binding of the cognate pathogen associated molecular patterns (PAMPs) (or their agonists) and signal a complex cascade of events that triggers the secretion of cytokines, chemokines and type I interferons [98]. The other class comprise the particulate adjuvants which are mineral-, lipid- or polymer-based delivery systems that, along with being transporters of the *Leishmania* antigen, are themselves immunostimulators due to their size, charge and composition; their properties can even be further improved by the

In a vaccine for leishmaniasis, it is expected that adjuvants modulate the immune system towards a Th1 response, with high amounts of secreted IL-12 and IFNγ. Indeed, recombinant

and CD8+ T cell-specific responses [96]. Impor‐

*2.4.3. Dendritic cell vaccine and liposomal formulations delivery systems*

with positive results [82-86].

286 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

*donovani* infection by the production of CD4+

decoration with other PAMP-like adjuvants [93].

IFNγ.

protect [97].

**2.5. Adjuvants**

MPL® is a purified derivative of the monophosphoryl lipid A hydrophobic moiety of *Salmonella minnesota*'s lipopolysaccharide (LPS). As LPS, MPL® is a potent TLR4 activator, though without the pyrogenicity of the parent molecule [101]. To even increase its efficacy, MPL® has been formulated in an oil-in-water stable emulsion in squalene (MPL-SE) which rendered high levels of IFNγ and low amounts of IL-4 and IL-10 [102]. A similar deriva‐ tive, GLA-SE (glucopyranosyl lipid adjuvant) has been chemically synthetized in order to obtain a pure molecule, free of biological components, but still maintaining the same multifunctional immunomodulatory activity as the naturally-derived MPL® [103]. Current‐ ly, MPL-SE and GLA-SE are undergoing clinical trials with the antigen LEISH-F3 for the first vaccine for human VL (see section 2.6.2).

Other TLR agonists are CpG-containing oligonucleotides (CpG ODNs) and imiquimod, which are ligands for the TLR9 and TLR7/8, respectively. CpG ODNs are strong immunostimulators by the upregulation on macrophages and DCs of CD40 and MHC class II costimulatory receptors [104] and the induction of IFNα, IFNβ and IFNγ, IL-12, IL-18 and TNFα secretion by lymphocytes [105]. In the same direction, imiquimod, a synthetic imidazoquinoline, is a Th1 activator. But noteworthy, imiquimod has itself anti-leishmanial activity through the activation of macrophages leading to the secretion of IL-12 and IFNγ [106]. Also, signal transduction directed to NO production was detected on *L. donovani*-infected macrophages treated with imiquimod [107]. Indeed, a recent report showed the effective application of topic imiquimod on the cutaneous lesions of a child infected with *L. infantum* unresponsive to liposomal amphotericin B [108].

Bacillus Calmette-Guérin (BCG), besides being the most widely administered vaccine in the world, it is also commonly used as adjuvant in numerous vaccine candidates for infectious diseases. In anti-*Leishmania* treatment [109] and vaccine research it has been tested in murine [110-112], hamster [113], canine [114, 115] and non-human primate models [42] (just to mention the most recent works). Its mechanism of immunostimulation relies on the activation of TLR2, TLR 4 and TLR9 [116, 117] in addition to its anti-leishmanial properties revealed in early studies [118, 119].

Saponins are natural products from the *Quillaja saponaria* tree chemically modified in order to increase their adjuvant properties [120]. QuilA is the heterogeneous mixture of saponins obtained from the aqueous extract of the *Quillaja* bark. Due to its high toxicity, purification by HPLC and chemical modifications have originated several saponins which display different toxicity and immunogenicity [121]. Saponins are common adjuvants used in vaccines for *Leishmania*. Indeed, the three approved vaccines for ZVL include saponins in their formulation (see section 2.6.1).

Particulate adjuvants have many properties that can be designed to bias the immune system in the desirable way which make them very versatile adjuvants. They serve as carriers for antigens and non-particulate adjuvants, targeting both vaccine components to the same APC and controlling their release. They can be used to increase the stability of antigens, like proteins, peptides or oligonucleotides, to improve the solubility of hydrophobic compounds or to bypass gastric degradation [93].

Aluminum salts are common in human and veterinary vaccines, though they are not proper adjuvants *per se* to be used in vaccines for leishmaniasis because their immunostimulatory properties drive a Th2 response. However, they have been used as carriers for other adjuvants, like IL-12 [122] or BCG [111], in combination with ALM antigen. Lipid-based vesicles (lipo‐ somes and niosomes) have been tested to carry ALM antigen with or without BCG in C57BL/ 6 [123] and BALB/c mice [124]. Similarly, virosomes are spheres formed by a phospholipid bilayer but that also contain viral glycoproteins (hemagglutinin and neuraminidase from influenza virus) which confer structural stability and enhance the adjuvanticity of these particles [93].

Micelles and emulsions likewise fall in this category of particulate adjuvants as, for example, MPL® and GLA formulated in stable emulsions (MPL-SE and GLA-SE). The oil-in-water emulsion formed with squalene (SE) is itself an adjuvant that has been included in the ongoing clinical trials run by IDRI (see section 2.6.2), though immunization with Leish-110f antigen plus SE led to the development of a Th2 response in BALB/c mice [77].

Finally, thought without great expression in *Leishmania* vaccine research, the most complex particulate adjuvants are the immune stimulating complexes. ISCOMATRIX™ are cage-like structures composed of cholesterol, lipids and QuilA bond together by hydrophobic interac‐ tions; they allow the inclusion of several antigens forming ISCOMATRIX™ vaccines [102]. Similar structures that include a hydrophobic antigen are called ISCOMs, while hydrophilic antigens must be held in cationic ISCOMs-like structures named PLUSCOMs [93]. The inclusion of QuilA in these systems allows the reduction of its amount and the bonding to cholesterol, therefore leaving no free QuilA to interact with cell membranes, which decrease its toxicity [93].

#### **2.6. Current status of vaccine research**

#### *2.6.1. Vaccines for zoonotic visceral leishmaniasis*

In canine vaccinology three authorized vaccine options are available.

Leishmune® was the first vaccine licensed for the prevention of ZVL but is authorized only in Brazil. It consists of *L. donovani* purified fucose-mannose ligand (FML antigen) in combination with a saponin adjuvant. Clinical trials have showed that Leishmune® reduces the risk of infection but also prevents disease progression in already infected dogs, though the manufac‐ turer does not recommend the vaccine as immunotherapy. A transmission-blocking activity was also attributed to this vaccine, making it highly appealing for the control of the zoonosis [125]. After 5 years of spread use among veterinary clinics in all the Brazilian territory, the manufacturer reports an efficacy of 97.3% in 8393 vaccinated *Leishmania*-seronegative dogs exposed to the natural challenge [126]. Strong cellular response (determined as *Leishmania*- specific lymphoproliferation with high levels of IFNγ in the absence of IL-10 and positive Montenegro skin reaction test) and favorable humoral response (with high titers of *Leishma‐ nia*-specific IgG2) are behind this protective response in vaccinated animals [126].

Some years later, Leish-Tec® was released, also only in Brazil. The recombinant A2 protein is the antigen that constitutes the vaccine along with saponin adjuvant. Protection was found to be related to high levels of anti-A2 IgG and IgG2, without the presence of IgG1, and high amounts of specific IFNγ with low levels of IL-10 [127]. However, there is no updated information about the efficacy of the vaccine in the field.

Recently, a new vaccine, CaniLeish®, the only authorized in Europe, has entered the market for the prophylaxis of ZVL. The manufacturer claims that vaccinated dogs have a 4-fold reduced risk of developing the disease compared to non-vaccinated animals [128]. The use of *L. infantum* excreted/secreted proteins associated to QA-21 adjuvant (*Li*ESP/QA-21) leads to the increase of IgG2 specific antibodies, stronger *Leishmania*-specific lymphoproliferation with an increased IFNγ-producing T cell population that is able to activate a significant leishmani‐ cidal macrophage ability *in vitro* due to NO production [129].

### *2.6.2. Ongoing clinical trials for a vaccine for VL*

antigens and non-particulate adjuvants, targeting both vaccine components to the same APC and controlling their release. They can be used to increase the stability of antigens, like proteins, peptides or oligonucleotides, to improve the solubility of hydrophobic compounds or to

Aluminum salts are common in human and veterinary vaccines, though they are not proper adjuvants *per se* to be used in vaccines for leishmaniasis because their immunostimulatory properties drive a Th2 response. However, they have been used as carriers for other adjuvants, like IL-12 [122] or BCG [111], in combination with ALM antigen. Lipid-based vesicles (lipo‐ somes and niosomes) have been tested to carry ALM antigen with or without BCG in C57BL/ 6 [123] and BALB/c mice [124]. Similarly, virosomes are spheres formed by a phospholipid bilayer but that also contain viral glycoproteins (hemagglutinin and neuraminidase from influenza virus) which confer structural stability and enhance the adjuvanticity of these

Micelles and emulsions likewise fall in this category of particulate adjuvants as, for example, MPL® and GLA formulated in stable emulsions (MPL-SE and GLA-SE). The oil-in-water emulsion formed with squalene (SE) is itself an adjuvant that has been included in the ongoing clinical trials run by IDRI (see section 2.6.2), though immunization with Leish-110f antigen

Finally, thought without great expression in *Leishmania* vaccine research, the most complex particulate adjuvants are the immune stimulating complexes. ISCOMATRIX™ are cage-like structures composed of cholesterol, lipids and QuilA bond together by hydrophobic interac‐ tions; they allow the inclusion of several antigens forming ISCOMATRIX™ vaccines [102]. Similar structures that include a hydrophobic antigen are called ISCOMs, while hydrophilic antigens must be held in cationic ISCOMs-like structures named PLUSCOMs [93]. The inclusion of QuilA in these systems allows the reduction of its amount and the bonding to cholesterol, therefore leaving no free QuilA to interact with cell membranes, which decrease

Leishmune® was the first vaccine licensed for the prevention of ZVL but is authorized only in Brazil. It consists of *L. donovani* purified fucose-mannose ligand (FML antigen) in combination with a saponin adjuvant. Clinical trials have showed that Leishmune® reduces the risk of infection but also prevents disease progression in already infected dogs, though the manufac‐ turer does not recommend the vaccine as immunotherapy. A transmission-blocking activity was also attributed to this vaccine, making it highly appealing for the control of the zoonosis [125]. After 5 years of spread use among veterinary clinics in all the Brazilian territory, the manufacturer reports an efficacy of 97.3% in 8393 vaccinated *Leishmania*-seronegative dogs exposed to the natural challenge [126]. Strong cellular response (determined as *Leishmania*-

plus SE led to the development of a Th2 response in BALB/c mice [77].

In canine vaccinology three authorized vaccine options are available.

bypass gastric degradation [93].

288 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

particles [93].

its toxicity [93].

**2.6. Current status of vaccine research**

*2.6.1. Vaccines for zoonotic visceral leishmaniasis*

On February 2012 the Infectious Disease Research Institute (IDRI) has launched a phase 1 clinical trial for the first vaccine against VL [130]. Thirty six healthy adult American volunteers were recruited to evaluate the safety, tolerability and immunogenicity of the LEISH-F3 recombinant antigen (composed of two fused proteins) with GLA-SE, MPL-SE or SE adjuvants [131]. About one year and a half later, this first clinical trial was completed and the vaccine was shown to be safe and to induce potent immune responses in healthy volunteers [132]. Later, IDRI partnered with the Indian pharmaceutical company Zydus Cadila to develop, register and market the three vaccine candidates to ensure that the possible future vaccine is affordable and accessible by the people that really need it. Also, in July 2013 this partnership has started phase 1 clinical trials in India to evaluate the effectiveness of the vaccine on individuals from endemic regions [132].

#### **3. Experimental data: Highly infective** *Leishmania infantum* **strain induces strong central and effector memory CD4+ and CD8+ immunity required for partial protection against re-infection**

### **3.1. Aim of the study**

It is well accepted that the broad clinical manifestations described in leishmaniasis are associated with the different cytokine milieu developed in response to the infection, which is highly dependent on the parasite itself. Accordingly, a diversity of immune responses have been described for *L. major* substrains [133] and *L. infantum* strains from the MON-1 zymodeme [134]. These immune responses may have a pivotal importance if the host faces a *de novo Leishmania* infection. In fact, data from endemic countries put on evidence the reality of resistance to re-infection in VL. In the one hand, it is evident the predominance of *L. infan‐ tum* infections in children compared to adults [6], which may result from acquired resistance to re-infection in adulthood, and, on the other hand, there are the examples of fully recovered patients that showed resistance to re-infection by the same *Leishmania* species [61].

Some studies on re-infection have been performed in mice as model for visceral leishmaniasis. Streit *et al*. described a partial level of protection against *L. chagasi* when mice were first infected with a high-dose inoculum since it was able to stimulate the immune system towards a Th1 response for counteracting a subsequent infection. On the contrary, an infection with a low dose suppressed IFNγ production and elicited high levels of TGFβ. Also, protective immunity was not achieved if an attenuated *dhfr-ts* knockout strain was used instead for immunization [49]. However, Oliveira *et al*. published opposing results as when they infected mice with a low dose of *L. chagasi* a protective immune response was generated, while a high dose contributed to the development of visceral disease [135].

To our knowledge, there is no previous literature about the concomitant immunity developed with live virulent *L. infantum* infection followed by homologous or heterologous re-infection. Since the severity of the infection and the progression of visceral leishmaniasis are strongly determined by the elicited immune response, in this work we analyzed the ability of two *L. infantum* virulent strains, which have presented different infectivity and immunomodulation, in the generation of an effective adaptive immunity in the context of experimental chronic infection and in the induction of a recall response after re-infection in BALB/c mouse model.

#### **3.2. Development of protection needs highly infective** *Leishmania*

Many efforts have been made to understand how *Leishmania*-specific immunity is generated and maintained over time. Nowadays, it is of scientific consensus that early activation of the innate immune system is essential for the production of a reliable adaptive response that leans on CD4+ and CD8+ specific cellular immunity.

To understand the strain-specific immunomodulation mechanisms that lead to protection to re-infection we used two strains of *L. infantum*, one dermotropic (HL) and the other viscero‐ tropic (ST), which presented differential onset and progression of VL in mice. As previously shown [136], HL was able to colonize the spleen, liver and bone marrow in higher extent than ST parasites 6 weeks after infection (Figure 1, Infection bars). We hypothesized that these differences in infectivity could lead to distinct levels of protection. Thus, we re-infected the mice with homologous or heterologous strains.

In our model, mice that were previously imprinted with HL strain and then challenged with the same highly infective strain (Figure 1, Re-inf HL bars) were able to sustain the splenic parasite load and to decrease in about 1 logarithm the number of parasites colonizing the liver and bone marrow. On the contrary, HL re-infection after ST imprinting led to a significant increase of about 1000 times in all the target tissues. Concomitant immunity was more pronounced when the animals were infected with the highly infective HL strain and then challenged with ST due to its lower infectivity (Figure 1, Re-inf ST bars). As such, the infections in the spleen and liver of HL imprinted mice suffered a significant reduction of ~1000-fold in the parasite loads to levels close to the quantification limit, and in the bone marrow parasitic presence was detected but not quantifiable. Accordingly, ST imprinting and consecutive challenge resulted in a ~10-fold increase in the splenic and hepatic parasite burden compared to the primary infection numbers, though no changes were noticed in the parasite load of the bone marrow.

resistance to re-infection in VL. In the one hand, it is evident the predominance of *L. infan‐ tum* infections in children compared to adults [6], which may result from acquired resistance to re-infection in adulthood, and, on the other hand, there are the examples of fully recovered

Some studies on re-infection have been performed in mice as model for visceral leishmaniasis. Streit *et al*. described a partial level of protection against *L. chagasi* when mice were first infected with a high-dose inoculum since it was able to stimulate the immune system towards a Th1 response for counteracting a subsequent infection. On the contrary, an infection with a low dose suppressed IFNγ production and elicited high levels of TGFβ. Also, protective immunity was not achieved if an attenuated *dhfr-ts* knockout strain was used instead for immunization [49]. However, Oliveira *et al*. published opposing results as when they infected mice with a low dose of *L. chagasi* a protective immune response was generated, while a high dose

To our knowledge, there is no previous literature about the concomitant immunity developed with live virulent *L. infantum* infection followed by homologous or heterologous re-infection. Since the severity of the infection and the progression of visceral leishmaniasis are strongly determined by the elicited immune response, in this work we analyzed the ability of two *L. infantum* virulent strains, which have presented different infectivity and immunomodulation, in the generation of an effective adaptive immunity in the context of experimental chronic infection and in the induction of a recall response after re-infection in BALB/c mouse model.

Many efforts have been made to understand how *Leishmania*-specific immunity is generated and maintained over time. Nowadays, it is of scientific consensus that early activation of the innate immune system is essential for the production of a reliable adaptive response that leans

To understand the strain-specific immunomodulation mechanisms that lead to protection to re-infection we used two strains of *L. infantum*, one dermotropic (HL) and the other viscero‐ tropic (ST), which presented differential onset and progression of VL in mice. As previously shown [136], HL was able to colonize the spleen, liver and bone marrow in higher extent than ST parasites 6 weeks after infection (Figure 1, Infection bars). We hypothesized that these differences in infectivity could lead to distinct levels of protection. Thus, we re-infected the

In our model, mice that were previously imprinted with HL strain and then challenged with the same highly infective strain (Figure 1, Re-inf HL bars) were able to sustain the splenic parasite load and to decrease in about 1 logarithm the number of parasites colonizing the liver and bone marrow. On the contrary, HL re-infection after ST imprinting led to a significant increase of about 1000 times in all the target tissues. Concomitant immunity was more pronounced when the animals were infected with the highly infective HL strain and then challenged with ST due to its lower infectivity (Figure 1, Re-inf ST bars). As such, the infections

patients that showed resistance to re-infection by the same *Leishmania* species [61].

contributed to the development of visceral disease [135].

290 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

**3.2. Development of protection needs highly infective** *Leishmania*

specific cellular immunity.

mice with homologous or heterologous strains.

on CD4+

and CD8+

7-8 week-old BALB/c mice were infected by intraperitoneal route with 108 HL (grey bars) or ST (white bars) *L. infantum* strains cultivated for 4 days in Novy-MacNeal-Nicolle (NNN) medium at 26 ºC. After 6 weeks of infection mice were anesthetized with isoflurane and sacrificed by cervical dislocation (Infection bars). In the re-infection experiments, ani‐ mals were infected for 6 weeks with HL or ST strains as before and challenged intraperitoneally with 108 promasti‐ gotes of the same or the other strain; 6 weeks after challenge they were sacrificed (Re-inf HL and Re-inf ST bars). **(A)** Spleen, **(B)** liver and **(C)** femoral bone marrow were recovered for quantification of the parasite load by real time PCR [136]. Bars represent means ± SD of 5 to 9 animals of one experiment representative of two independent. Statistically significant differences between HL and ST infections were calculated with Mann-Whitney test and are signed with + . Kruskall-Wallis test followed by Dunn's multiple comparison test were used to calculate differences before and after challenge and are depicted with \*. Statistical analysis was done in GraphPad Prism 5 (GraphPad Software). Dashed line indicates the limit of detection for quantification for each tissue.

**Figure 1.** Parasite load after infection and challenge with *L. infantum* strains presenting different infectivities

Based on the data exposed above, in terms of parasitological analysis we established that the onset of pathology (set as hepatosplenomegaly (data not shown; see [136]) and high parasite loads) by an infective *L. infantum* strain confers a degree of protection over a re-infection episode which correlates with the infectivity of both the imprinting and the challenging strains that are inoculated in the host. Similar findings were reported previously, when a high-dose of *L. chagasi* promastigotes was required for the development of protection against re-infection, whereas a low-dose immunization either had no effect or slightly exacerbated disease [49].

#### **3.3. Infectivity may influence downstream adaptive response-triggering events**

To understand the immune response behind this protective phenotype, we analyzed the splenic populations and the T cells functionality. We observed that infection with HL produced a significant increase in the total cellularity and major leukocyte populations when compared to naïve animals, which was not noticed when mice were infected with ST strain (Figure 2A-E).

**(A-E)** After infection and consequent challenge with both HL and ST strains, splenocytes were recovered and surfacestained for identification of major leukocytes populations. **(A)** Total cells were counted and **(B)** CD4+ T cells (CD3+CD4+), **(C)** CD8+ T cells (CD3+CD8+), **(D)** B cells (CD19+) and **(E)** macrophages (CD11b+Ly6C+) were evaluated by

#### Can Attenuated *Leishmania* Induce Equally Effective Protection as Virulent Strains in Visceral Leishmaniasis? http://dx.doi.org/10.5772/57214 293

flow cytometry in a FACSCanto (BD Bioscences). Cell numbers from infected mice were normalized with respective val‐ ues from age-matched naïve mice and results are presented as log2 of the fold change relative to naïve animals, with dashed and solid lines indicating 2- and 4-fold difference. Boxes and whiskers with 5-95 percentile and mean (showed with +) of 5 to 9 animals of one experiment representative of two independent. Mann-Whitney test was run to calcu‐ late statistically significant differences between mice infected with HL or ST and results are depicted with +. Differences before and after challenge are indicated with \* for p<0.05 or \*\* for p<0.01 and were calculated with Kruskall-Wallis test followed by Dunn's multiple comparison test in GraphPad Prism 5 (GraphPad Software). **(F-H)** Number of **(F)** in‐ flammatory macrophages (CD11b+Ly6C+ CCR2+), **(G)** inflammatory neutrophils (CD11b+Ly6G+CCR2+) and **(H)** activated dendritic cells (CD11c+CCR2+) in infected mice before (Infection bars) and after homologous challenge (Re-inf bars). Bars represent means ± SD of 5 to 9 animals of one experiment representative of two independent. Statistically signifi‐ cant differences were calculated in GraphPad Prism 5 (GraphPad Software) with Mann-Whitney test between naïve and infected or challenged animals and show \* p<0.05, \*\* p<0.01 and \*\*\* p<0.001.

**Figure 2.** Splenic cellular populations after infection and challenge with highly and low infective *L. infantum* strains

Interestingly, when the animals were subjected to a secondary infection by HL, regardless of the infectivity of the imprinting strain, we detected the same increase in the number of splenocytes, while after challenge with ST there was no change in the cellularity.

Inflammatory macrophages/monocytes and neutrophils, besides its recognition as host cells [137, 138], have been implicated in the remodeling of the spleen during splenomegaly in leishmaniasis [139, 140], as well as in the modulation of the specific CD4+ T cells re‐ sponse in late phases of infection, at least with *L. major* [141]. Infiltration of neutrophils [142], DCs [143] and macrophages [144] in inflamed tissues is tightly regulated by the CC chemokine receptor 2 (CCR2) that also participates in important processes related to anti-*Leishmania* defense [143, 144].

As these are the first cells that need to be committed, we determined the number of inflam‐ matory macrophages, DCs and neutrophils by the expression of CCR2 (Figure 2F-H). Infection and challenge with HL led to the significant increase of these inflammatory cells in the spleen. Similarly, infection with ST also significantly increased the inflammatory DCs and neutrophils, but only with a second wave of parasites the CCR2+ macrophages arisen in numbers signifi‐ cantly higher than in uninfected animals. However, this difference in the number of CCR2+ macrophages relates with the total macrophages present in the spleen, as the relative percen‐ tages were similar between HL and ST (data not shown). These CCR2+ macrophages exert an important role in the defense against *Leishmania*, since it has been previously described that optimal parasite killing require the recruitment of CCR2+ macrophages, followed by stimula‐ tion with combined monocyte chemotactic protein 1 (MCP-1) and IFNγ [144].

Thus, monocyte and neutrophil activation showed no major differences between HL and ST strains, similarly to the findings of Meddeb-Garnaoui *et al*. that compared the cytokine profile of human monocytes infected with dermotropic and viscerotropic *L. infantum* strains which presented respectively high and low infectivity *in vitro* [145]. In their *in vitro* setup, no differences were found in the ability of those two strains in the modulation of monocytesecreted cytokines [145], indicating that the infectivity of a *Leishmania* strain not always produces a direct effect on the innate immune response. Nonetheless, *in vivo*, where other factors that influence macrophage function are present, the effect of the infectivity was not evaluated. We hypothesize that despite monocyte and neutrophil activation were similar, HLand ST-activated cells should present divergent efficiencies when triggering the adaptive

**(A-E)** After infection and consequent challenge with both HL and ST strains, splenocytes were recovered and surfacestained for identification of major leukocytes populations. **(A)** Total cells were counted and **(B)** CD4+ T cells (CD3+CD4+), **(C)** CD8+ T cells (CD3+CD8+), **(D)** B cells (CD19+) and **(E)** macrophages (CD11b+Ly6C+) were evaluated by

292 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

immune response, which may be indicative of intrinsic characteristics of the strains in modulating downstream events.

#### **3.4. Highly infective** *L. infantum* **triggers memory and effector CD4+ and CD8+ T cells**

We have studied the generation of CD4+ and CD8+ memory T cells 6 weeks post-infection and upon challenge with the same strain by the surface expression of CD44 and CD62L (Figure 3).

**(A, B)** CD4+ and **(C, D)** CD8+ T cells were analyzed by flow cytometry in a FACSCanto (BD Bioscences) according to their surface expression of CD44 and CD62L. Naïve (CD44loCD62L+ ), central memory (TCM; CD44hiCD62L+ ) and effector mem‐ ory (TEM; CD44hiCD62L- ) subpopulations were quantified before **(A, C)** and after **(B, D)** challenge. Bars show means ± SD of 5 to 9 animals of one experiment representative of two independent and statistically significant differences be‐ tween naïve and infected mice are depicted with \* for p<0.05, as calculated by two-tailed Mann Whitney test run in GraphPad Prism 5 (GraphPad Software).

**Figure 3.** T cell memory repertoire of mice subjected to infection and homologous re-infection with HL and ST L. infan‐ *tum* strains

HL infection potentiated the expansion of central memory CD8+ (Figure 3C, TCM bars) and especially CD4+ T cells (Figure 3A, TCM bars) that doubled in percentage compared to uninfected mice. These memory populations are probably an important factor in the control of the parasite load in the spleen, as presented before (Figure 1A), when the animals were subjected to re-infection. Memory cells constitute a source of experienced-antigen cells that are able to rapidly respond to face a similar challenge. While TEM cells display protective effector functions, TCM are thought to replenish the TEM pool [146].

immune response, which may be indicative of intrinsic characteristics of the strains in

and CD8+

upon challenge with the same strain by the surface expression of CD44 and CD62L (Figure 3).

T cells were analyzed by flow cytometry in a FACSCanto (BD Bioscences) according to their

) subpopulations were quantified before **(A, C)** and after **(B, D)** challenge. Bars show means ±

T cells (Figure 3A, TCM bars) that doubled in percentage compared to

), central memory (TCM; CD44hi

CD62L+

(Figure 3C, TCM bars) and

) and effector mem‐

CD62L+

SD of 5 to 9 animals of one experiment representative of two independent and statistically significant differences be‐ tween naïve and infected mice are depicted with \* for p<0.05, as calculated by two-tailed Mann Whitney test run in

**Figure 3.** T cell memory repertoire of mice subjected to infection and homologous re-infection with HL and ST L. infan‐

uninfected mice. These memory populations are probably an important factor in the control of the parasite load in the spleen, as presented before (Figure 1A), when the animals were subjected to re-infection. Memory cells constitute a source of experienced-antigen cells that are

HL infection potentiated the expansion of central memory CD8+

 **and CD8+**

memory T cells 6 weeks post-infection and

 **T cells**

**3.4. Highly infective** *L. infantum* **triggers memory and effector CD4+**

modulating downstream events.

**(A, B)** CD4+

*tum* strains

especially CD4+

ory (TEM; CD44hi

and **(C, D)** CD8+

CD62L-

GraphPad Prism 5 (GraphPad Software).

surface expression of CD44 and CD62L. Naïve (CD44lo

We have studied the generation of CD4+

294 Leishmaniasis - Trends in Epidemiology, Diagnosis and Treatment

In fact, after challenge with HL, both CD4+ (Figure 3B) and CD8+ (Figure 3D) TCM pools remained high and TEM cells also significantly increased compared to naïve mice. Moreover, taking into account that the total numbers of T lymphocytes in the infected animals were significantly increased in relation to naïve mice (Figure 2B and C), the number of memory (CD44hi) T cells was even more expressive in the spleens of those HL re-infected animals. On the contrary, ST strain showed no potential in clonal expansion of memory populations or at least in their high number maintenance in order to bring advantage upon re-infection. Neither in the imprinting infection nor after challenge could we detect CD4+ or CD8+ central or effector memory T cells in a percentage higher than that of the naïve animals. The decrease in the CD8+ TCM cells 6 weeks after ST infection (Figure 3C) was considered not to have any biological meaning since, when adjusted to total number of cells, both naïve and infected mice have similar amounts of that subpopulation.

From the data exposed, we justified the partial protection that a primary infection with HL *L. infantum* strain can generate upon an homologous re-infection. This strain has the ability to activate the innate defenders (DCs, macrophages and neutrophils) for mobilization to the spleen where they can drive an effective generation and expansion of memory CD4+ and CD8+ T cell subsets.

#### **3.5. Double producers CD4+ IFNγ<sup>+</sup> IL-10+ and CD8+ IFNγ<sup>+</sup> TNFα<sup>+</sup> T cells arise after re-infection**

To appreciate the mechanisms underlying the protection observed after re-infection with a highly infective strain, we analyzed the magnitude of the developed T cell response in infected and re-infected mice with HL strain. After infection, we detected high levels of IFNγ-producing CD4+ and CD8+ T cells (Figures 4A and C, respectively). This finding was suspected after having noticed the massive cellular infiltrate of leukocytes in the spleen (Figure 2) and also the existence of approximately 15 % of effector memory lymphocytes (combined CD4+ and CD8+ ) that classically secrete this cytokine [147]. Upon re-infection (Figures 4B and D), however, a more interesting panel of effector cells has emerged. Along with the same IFNγ<sup>+</sup> cells, detected in both CD4+ and CD8+ lymphocytes, we identified IL-10+ in ~1.5 % and IFNγ<sup>+</sup> IL-10+ double producers in ~0.75 % of the CD4+ T cells, which represent an increment of ~1.7 and ~3.1, respectively, compared to uninfected animals.

CD4+ T-bet+ IFNγ<sup>+</sup> IL-10+ cells were recently described by us and others upon infection of BALB/ c mice with *L. infantum* [148] or *L. donovani* [149]. This Th1 population is driven by CD4+ T cells activation by the infected DCs and leads to an unprotective phenotype that accentuates the infection. However, a protective role was previously attributed to CD4+ CD25- Foxp3- IFNγ + IL-10+ cells in a vaccination study with *L. donovani* LdCen1-/- [57] and in a non-healing model of CL with *L. major* [150], which were claimed to arise after a strong inflammatory stimulus as a feedback control of Th1 responses to avoid tissue damage.

In CD8+ T cells, conversely, cytokine double producing cells were found for IFNγ<sup>+</sup> TNFα<sup>+</sup> , in a representation of ~0.86 %, meaning an increase of ~3.4 fold compared to naïve mice. IFNγ and

IFNγ, IL-10 and TNFα production was analyzed by flow cytometry in CD4<sup>+</sup> **(A, B)** and CD8<sup>+</sup> **(C, D)** lymphocytes. Spleno‐ cytes were stimulated *ex-vivo* with phorbol 12-myristate 13-acetate (PMA), ionomycin and brefeldin A, stained for sur‐ face and intracellular molecules and analyzed in a FACSCanto flow cytometer (BD Bioscences). Cytokine single and double producers in each lymphocyte population are depicted from naïve, infected **(A, C)** or challenged **(B, D)** mice. Bars represent means ± SD of 4 to 9 animals of one experiment with statistically significant differences between naïve and infected mice indicated with \* when p<0.05, as calculated by two-tailed Mann Whitney test run in GraphPad Prism 5 (GraphPad Software).

**Figure 4.** Intracellular cytokines of CD4<sup>+</sup> and CD8<sup>+</sup> lymphocytes of HL infected and re-infected animals

TNFα concomitant production by Th1 and CD8<sup>+</sup> T cells has for long proven to be more efficient in the killing of *L. major* [151, 152] and other unrelated microorganisms (*e.g. Mycobacterium tuberculosis* [153]) than the production of IFNγ or TNFα alone. More recently, IFNγ<sup>+</sup> TNFα<sup>+</sup> high quality CD4<sup>+</sup> and CD8<sup>+</sup> T cells were described to be generated after several vaccination protocols against *L. major* and correlate with prognosis of protection much better than IFNγ single producers [154]. Moreover, those double producers CD4<sup>+</sup> T cells, which can also be IL-2<sup>+</sup> , were determined to belong to the central memory subset, providing long-term protection [154, 155]. As for CD8<sup>+</sup> IFNγ<sup>+</sup> TNFα<sup>+</sup> T cells, they were described to have enhanced cytolytic activity compared to IFNγ<sup>+</sup> single producer cells in HIV-infected patients [156]. However, in our study, we could not detect any difference in the cytotoxicity mediated by CD8<sup>+</sup> T cells from HL infected and challenged mice compared to that from naïve animals (data not shown), which may indicate that cytolytic activity of those cells was not required in the containment of the parasites in the spleen or, instead, the persistence of the splenic parasite load is due to an incomplete effector function of the CD8+ T cells.

### **3.6. Conclusions**

Taken together, our results show that HL *L. infantum* strain promotes a robust activation of the immune system upon infection initiated by a strong recruitment of leukocytes to the spleen which stimulates the development of an effective adaptive response. This is a mixed response as considered by the detection of single producers IFNγ<sup>+</sup> and IL-10+ CD4+ T cells that become more evident when the antigen is re-loaded (*i. e.* re-infection). CD8+ T cells also exert their effector function by the production of IFNγ. After re-infection, double producers CD8<sup>+</sup> IFNγ + TNFα<sup>+</sup> and CD4+ IFNγ<sup>+</sup> IL-10+ T cells arise, probably from the expansion of the central and effector memory subsets, to contain the parasites that colonized the spleen and to efficiently resolve the infection in the liver and bone marrow, controlling tissue damage by IL-10 production. To confirm this hypothesis, adoptive transfer of these memory cells produced after re-infection with our highly infective *L. infantum* strain could be performed to evaluate the protective phenotype of such pools of CD4+ and CD8+ T cells in naïve animals challenged with a subsequent *L. infantum* infection.

Taking the fact that HL is a dermotropic strain that caused CL in a human patient, its tropism is possibly justified by the inflammatory potential of the strain that impedes a silent entry into the host. A protective response may immediately be mounted in the skin, abrogating any chance of the parasite to reach internal organs and visceralize [157]. Concerning the ST strain, an agent of human VL, the initial activation of the innate immune system does not translate into efficient adaptive immunity as no memory cells were detected. With this, a primary infection does not serve as imprinting, since a re-infection with the same strain led to the increase of the parasite load in the spleen and liver.

With this work we contributed to the better understanding of the complex modulation that *Leishmania* parasites do to surmount the protective strategies developed by the host's immune system. Much of the knowledge acquired so far by the scientific community was based on *L. major*-infection models that have a clear Th1/Th2 dichotomy on protection/progression of the disease, and more studies with VL models are needed to clarify the intriguing modulation that viscerotropic *Leishmania* strains provide to take advantage of their host.
