INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Parasitic infection with Leishmania spp. results in a broad spectrum of clinical diseases in humans. Leishmania infantum is responsible for most cases of human leishmaniasis in southern Europe and reflects this diversity with proteiform clinical expressions ranging from cutaneous leishmaniasis (CL) to either asymptomatic or fatal visceral leishmaniasis (VL).

The varying clinical expression observed in VL depends on complex relationships in which not only the genetic potential (3, 4) and/or immunological status of the host (10, 15, 20) but also the proper parasite biodiversity (15, 16, 19, 26) appear as determinant factors. Characterization of Leishmania polymorphism at the species and subspecies levels is currently based on isoenzyme and biomolecular analysis. However, identification of biological markers like parasite tropism or virulence remains crucial as a complementary tool to account for the considerable biodiversity of the parasite.

In this context, characterization of parasite virulence, i.e., the ability to develop and multiply in vitro and/or in vivo appears particularly relevant for analysis of parasitic factors that can affect the pathology of VL. Using 21 Mediterranean strains of L. infantum isolated from humans, we previously showed that there is a remarkable intraspecific heterogeneity in experimental virulence expression of the parasite in BALB/c mice (29). Three major infection profiles were characterized: a visceralization (V) profile with persistent heavy liver parasite burdens and marked progressive spleen involvement, a regulation (R) profile in which the infection is finally contained after an initial phase involving both organs, and an undetermined (U) profile with undetectable, low-level, or poorly characterized infection. Both the V and R profile types were maintained in C57BL/6 mice (cure haplotype) (17) and also in CB-17 congenic immunodeficient scid mice, although heavier parasite loads were observed in scid mice than in BALB/c mice in the late phase of infection (12). Thus, as these infection profiles are observed in various genetic and immunological host contexts, they actually characterize parasite virulence phenotypes in vivo.

Because the population structure of Leishmania spp. and other kinetoplastids is mainly clonal (8, 18, 30) and because strains isolated from naturally infected hosts are likely composed of multiclonal parasite associations (7, 25), such virulence profiles have to be investigated at the clonal level before any conclusion about the genotypic nature of parasite virulence can be drawn. Thus, this work was undertaken to explore the polymorphism of L. infantum at the intrastrain level and its possible impact on the basis of virulence phenotyping. Using a mouse model of VL, we studied the virulence biodiversity of L. infantum clonal populations issuing from strains with various levels of virulence. We also experimentally investigated the interaction between strain and clone populations with diverse levels of virulence when simultaneously or successively present in the same host.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Parasites. The strains and clones used in this study were from the International Leishmania Cryobank and Identification Center in Montpellier, France. Three L. infantum strains were selected for their different infection profiles, as defined in previous experimental studies (29): (i) strain MHOM/FR/91/LEM 2259 zymodeme MON-1 expressing a virulence (V) profile, in which parasite burdens rise continuously in the spleen (10⁶ to10⁷ ml¹ at day 100) while persisting at high levels in the liver (LEM 2259/V); (ii) strain MHOM/FR/91/LEM 2176 (MON-33), previously known as having a regulation (R) profile in which liver and spleen involvement is observed at the early phase of infection and followed by regulation with parasite burdens decreasing to low or undetectable levels in the liver while remaining moderate in the spleen (LEM 2176/R); and (iii) strain MHOM/FR/94/LEM 2859 (MON-1), expressing an uncharacterized (U) phenotype consisting of low-level or even undetectable infection (LEM 2859/U). Both strains LEM 2259/V and LEM 2176/R were isolated from the bone marrow of human immunodeficiency virus-infected patients with VL. Strain LEM 2859/U originated from the skin of an immunocompetent patient.

In vitro culture. For in vitro growth characteristic study, parasites were cultivated at 27°C in Schneider's drosophila medium (SDM) (GIBCO BRL) and HOSMEM liquid medium (2) supplemented with hemin 10 µM (Sigma) and 20% heat-inactivated fetal calf serum (GIBCO). Parasites were inoculated into 25-ml culture flasks at day 0 at a final concentration of 10⁵ ml¹. Parasite concentrations were evaluated by means of nucleoside hydrolase activity determination using p-nitrophenyl--D-ribofuranoside substrate. Nucleoside hydrolase activity was measured at day 4 and day 7 in duplicate samples of culture supernatants in 96-well microplates as previously described (13). Parasite concentrations (per milliliter) in samples were calculated from a standard curve established with a serial twofold dilution of a promastigote suspension made in duplicate from 2 × 10⁷ to 5 × 10³ parasites ml¹. For animal inoculation, mass cultures were produced in SDM.

Experimental animal infection. Female 8-week-old BALB/c mice (IFFA CREDO) housed under standard conditions were used. Groups of mice were inoculated intravenously (i.v.) with 6-day-old promastigotes of the different strains or clones. The kinetics of infection was monitored throughout the experiment by determining the parasite burdens in the livers and spleens of three to five mice for each group at each measurement point.

Parasite burdens. Parasites in organs were quantified by means of a culture microtitration assay as previously described (5). Briefly, organs were excised, weighed, and then homogenized with a tissue grinder (ULTRATURRAX, Stauffen, Germany) in 4 ml of SDM supplemented with 20% heat-inactivated fetal calf serum, penicillin (100 U/ml), and streptomycin (50 µg/ml) (Bio-Merieux). Serial fourfold dilutions of organ homogenates, ranging from 1/1 to 1/4 × 10⁶, were made in duplicate under sterile conditions in 96-well microtitration plates containing 225 µl of culture medium. Plates were examined for the presence of mobile promastigotes under an inverted microscope after 10 days of incubation at 26°C. The final titer was the last dilution containing at least one parasite. Results were expressed as the logarithmic mean number of parasites per gram established from three to five mice for each group. The detection threshold of the method is 5 × 10² parasites g¹.

Experimental design. Three sets of experiments were carried out (Fig 1). (i) The first set of experiments was carried out in order to explore intrastrain virulence biodiversity. For that study, the virulence phenotypes of strains LEM 2259/V, LEM 2176/R, and LEM 2859/U and 11 clones were characterized. Groups of 16 mice were infected at day 0 with each strain or clone, and the kinetics of infection was monitored from day 7 to day 98 as shown in Fig 1a. (ii) A second set of experiments was designed to determine the kinetics of infection in mice injected simultaneously with clones with various virulence phenotypes. Infections were done with mixtures of clones that expressed different virulence profiles in the first experiment. Three clones originating from strain LEM 2259/V were selected, virulent clone 3511/V and less pathogenic clones 3518/R and 3512/R. An equal mixture of two or three clones (2 × 10⁶ promastigotes each) was injected into mice at day 0, and the evolution of the infection was monitored in their livers and spleens at day 21, day 40, and day 90 (Fig 1b). (iii) The third set of experiments was carried out to study the interactions between successive primary and challenge infections with strains or clones expressing various pathogenic phenotypes.

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Experimental design of single or concurrent animal infections. Panels: a, single infections with strains or clones; b, single infections with mixed clones; c, successive infections with strains; d, successive infections with clones.

Serological studies. Serum antibody levels were monitored in cross-infections (experiment iii) before the challenge infection and at the end of the experiment.

1

Statistical analysis. Data were analyzed with GraphPad Prism 3.0 statistical software (GraphPad Software). For better comparison of data from separate experiments in which there were mild variations in the dates of examination, we chose to present data extrapolated from the regression curves calculated by using Pearson's correlation analysis. Thus, for in vitro studies, parasite concentrations (per milliliter) at day 4 and day 7 were calculated from the regression equation of the promastigote growth curve. Similarly, for in vivo experiments, liver and spleen parasite burdens (per gram) were extrapolated from the corresponding regression lines of parasite burdens. Thus, parasite loads were calculated at day 7 and day 100 for single infections and at day 7 and day 65 or day 70 (before reinfection) and day 75 or day 80 and then day 150 in challenge experiments. Comparisons were made by one-way analysis of variance (ANOVA) with Bonferroni's posttests or by two-way ANOVA. For two-way-ANOVA comparisons, the area under the curve was chosen as a global estimate of parasite growth in vitro or virulence in animals (24).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of clones. The in vitro growth characteristics of the three strains and their respective clones are summarized in Table 1. All of the strains and clones grew readily, with a 1- to nearly 3-log increase in parasite concentration from day 0 to day 7, except for strains LEM 2259 and LEM 2859 in SDM. Promastigote growth rates in the two media were not correlated (r² = 0.296, P = 0.32). Promastigote growth was significantly better in HOS culture medium than in SDM (P = 0.005).

View larger version (25K): [in this window] [in a new window]   FIG. 2.   Experiment i. Calculated parasite loads ± the standard errors at day 7 (column 1) and day 100 (column 2). Mice were inoculated i.v. at day 0 with 107 promastigotes of L. infantum LEM 2259/V, LEM 2176/R, and LEM 2859/U () or their respective clones ().

1

1

1

1

1

1

1

Mixed infections with clones. Evolution of infection in mice injected with mixtures of clones (experiment ii) is represented in Fig. 3. Inoculation with a mixture of clones 3512/R and 3518/R resulted in a controlled R type of infection. The parasite burden decreased in the liver from 4.7 × 10³ at day 7 to 1.7 × 10³ g¹ at day 100 while remaining at a low level in the spleen (3.3 × 10³ g¹ at day 100). This contrasted strongly with the evolution of infection in mice inoculated simultaneously with clone 3511/V and clone 3512/R, 3518/R, or both (P <0.001). Despite a reduction in parasite burdens in the liver, a straightforward (2- to 2.5-log) increase occurred constantly in the spleen. At day 100, parasite loads reached 5.5 × 10⁵ to 1.6 × 10⁶ g¹. Thus, the virulence phenotype of clone 3511/V was expressed in the three groups of mice, whatever the R clone coinjected.

View larger version (25K): [in this window] [in a new window]   FIG. 3.   Experiment ii. Calculated parasite burdens ± the standard errors at day 7 (column 1) and day 100 (column 2). Mice were inoculated i.v. simultaneously at day 0 with 106 promastigotes of each of two or three clones of L. infantum strain LEM 2259/V. The three clones used expressed either a V (clone 3511) or an R (clones 3512 and 3518) phenotype when inoculated separately. Combinations of clones 3212 and 3218 (R and R phenotypes, respectively; ), 3511 and 3512 or 3518 (V and R phenotypes, respectively; ), and 3511, 3512, and 3518 (V, R, and R phenotypes, respectively; ) were used.

Cross-infections. The kinetics of infection in mice inoculated with strains with various levels of virulence (experiment iiia) are shown in Fig. 4. Single inoculations with strains LEM 2259, LEM 2176, and LEM 2811 resulted in V, R, and U infection profiles, respectively, as already described. In mice primarily infected with strain LEM 2259/V (Fig. 4A), the infection profile was not modified after a challenge at day 70 with the same strain or heterologous strain LEM 2176/R, except for a weak increase in the liver burden at day 80. By contrast, profound changes in the kinetics of infection were observed after reinfection in mice previously infected with strain LEM 2176/R (Fig. 4B) or LEM 2859/U (Fig. 4C). Similar modifications occurred in both groups of mice. A challenge infection with strain LEM 2259/V resulted in massive organ involvement, with parasite burdens reaching levels of >10⁶ g¹ in the liver and >10⁷ g¹ in the spleen at day 150. By contrast, a new infection at day 70 with strain LEM 2176/R resulted in only a weak increase in the hepatic and splenic burdens at day 80, followed by control of the challenge infection (Fig. 4B and C). Thus, the virulence phenotype of strain LEM 2259/V was constantly expressed in both primary and challenge infections. All groups of mice primarily infected and/or challenged with this strain showed very heavy parasite burdens at day 150 in the liver (10⁵ to 10⁶ g¹) and especially in the spleen (>10⁷ g¹).

View larger version (24K): [in this window] [in a new window]   FIG. 4.   Experiment iii. Kinetics of parasite burdens in the livers () and spleens () of mice successively inoculated i.v. at day 0 and then at day 70 with 107 promastigotes of three strains of L. infantum. The primary inoculation was done with strain LEM 2259/V (A), LEM 2176/R (B), or LEM 2859/U (C) expressing a V, R, or U phenotype, respectively. The challenge inoculation at day 70 was done with either the primary infecting strain or a heterologous strain.

1

1

1

1

View larger version (20K): [in this window] [in a new window]   FIG. 5.   Experiment iii. Evolution of parasite burdens in the livers () and spleens () of mice successively inoculated i.v. at day 0 and then at day 66 with 107 promastigotes of three clones of L. infantum strain LEM 2259/V. The clones used expressed V (clone 3511), R (clone 3514), and U (clone 3515) phenotypes.

Serological studies. In uninfected mice, the mean titer was 158 ± 8.6 AU, determining a positive cutoff of 175 AU. On this basis, a significant antibody response was observed in all groups of infected mice (Table 3).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Genetic diversity is a crucial parameter that must be taken into account in human VL. In particular, polymorphism of parasite virulence and tropism is likely to have a profound impact on disease transmission and/or pathology, as suggested by experimental studies on L. guyanensis CL in hamsters and Trypanosoma cruzi trypanosomiasis in mice (11, 22).

Phenotypic characterization is a prerequisite for further identification of genetic loci and/or mechanisms involved in parasite tropism and virulence. Phenotypic analysis of Leishmania isolates is currently based largely on isoenzyme polymorphism (27). Thus, dermotropic and viscerotropic human L. infantum zymodemes have been identified (15). However, we have previously shown that this method cannot discriminate between parasites with various levels of virulence since strains of L. infantum belonging to the same zymodeme exhibit a large heterogeneity of virulence profiles in mice (29).

In a first attempt to investigate the biodiversity of virulence expression of L. infantum parasites at the intrastrain level, we analyzed the in vitro and in vivo growth characteristics of three strains (isolated from humans) that exhibit different virulence patterns in mice, as well as those of 11 clones originating from these strains. Analysis of parasite growth characteristics in vitro showed no differences among clones from the different parental strains. By contrast, a marked intrastrain heterogeneity was evidenced by in vivo virulence phenotype analysis. One out of five clones obtained from the V strain generated a typical V profile, with marked and prolonged involvement of the liver and spleen until the end of the experiment, while the four remaining clones showed low-virulence infection profiles (three R and one U). No virulent clone was obtained from the two strains selected for their nonvirulence or undetermined phenotype. Clones obtained from these strains gave four R and two U types of infection. Thus, we can conclude that strains are composed of multiclonal parasite populations that are heterogeneous on the basis of experimental virulence in mice.

As a consequence, we examined the course of infection in cases of concurrent infections with clones in order to investigate the character of dominance of the V or R phenotype. First, we used several combinations of two or three clones issuing from the same parental strain, LEM 2259/V, but displaying various virulence profiles when injected alone (one V and two R clones). In this experiment, the presence of the V clone in the inoculum consistently resulted in a V type of infection whatever the R clone(s) coinjected. This shows that virulence is expressed as a dominant character in experimental multiclonal infections without any apparent interaction with the associated low-virulence clonal populations.

Second, we examined the outcome of infection in cases of sequential inoculations with strains or clones with different virulence phenotypes. We found that the expression of strain or clone virulence phenotypes was not modified in successive infections of the same host. The course of infection in previously infected animals was identical to that observed in naive mice, whatever the combination of virulence phenotypes used for the primary and secondary inoculations.

To our knowledge, this is the first data concerning the interactions of parasites varying in virulence in successive experimental VL infections. Using the same L. infantum strain for primary and challenge infections, Rousseau et al. (28) found that a primary infection induced a protective effect that was organ dependent, protection being achieved in the liver but not in the spleen. In CL, previous experiments on the protective role of a primary infection against a challenge have led to contradictory results. Li et al. (21) observed that infection of BALB/c mice with a low-virulence clone of L. major protected mice against a challenge infection with a highly virulent clone. Our results are more in agreement with those of Da Fonseca et al. (9), who showed that a primary infection with a totally avirulent strain of L. amazonensis did not protect mice against a virulent strain.

Serological studies showed that the type of infection significantly influenced the humoral response in mice. Indeed, a serological response was detected in all infected mice, including those inoculated with low-virulence strains or clones. However, mice inoculated with a virulent strain or clone showed higher antibody levels than did animals infected with less virulent parasite populations. Similarly, a challenge infection with a V strain or clone was followed by a marked increase in antibody levels while no changes were observed after a challenge with an R strain or clone. This shows that progression of infection is accompanied by a humoral response in both primarily infected or superinfected animals. This correlates with observations made in human VL, where patent infections are frequently associated with a strong specific antibody response and T-cell anergy to Leishmania antigen (14), while a cure is accompanied by an antibody level decrease and restoration of the cellular response (6, 23).

Our results clearly show that (i) L. infantum strains isolated from humans are composed of clonal populations displaying a marked polymorphism of experimental virulence to mice, (ii) populations with various levels of virulence do not interact in vivo in experimental multiclonal infections and strain and/or clone virulence phenotypes are conserved in simultaneous or successive infections, and (iii) the virulence phenotype is expressed as a dominant character in such experimental infections with different parasite populations. This may have a profound impact on the epidemiology and immunopathology of L. infantum VL. Characterization of parasite virulence is a crucial step that must be done in parallel with immunological or biomolecular research studies on VL.