OUP user menu

Experimental pathogenicity of viscerotropic and dermotropic isolates of Leishmania infantum from immunocompromised and immunocompetent patients in a murine model

Annie Sulahian, Yves Jean François Garin, Francine Pratlong, Jean Pierre Dedet, Francis Derouin
DOI: http://dx.doi.org/10.1111/j.1574-695X.1997.tb01005.x 131-138 First published online: 1 March 1997


The pathogenicity of 22 strains of Leishmania infantum from 11 HIV-infected and 11 immunocompetent patients with visceral (VL, n = 16) or cutaneous (CL, n = 6) leishmaniasis, belonging to 3 zymodemes (MON-1, n = 14; MON-29, n = 5; MON-33, n = 3), was studied using a murine model. For each strain 16–20 BALB/c mice were infected at day 0 (d0) by i.v. injection of 107 stationary-phase promastigotes. Parasite burdens were quantified in the spleen and liver of 4–5 mice of each strain at d7, d20, d60 and d90 or d100, using a sensitive culture microtitration technique. A great variability of infection profiles between strains was observed: (i) six strains showed a progressive infection, with a predominance of hepatic parasites at d7 or d20 (104−106 g−1), then a continuous rise of splenic parasites reaching 105−107 g−1 at d90 or d100 contrasting with a stagnation or decrease in the liver; (ii) ten strains gave a controlled infection with hepatic parasite burden reaching 104−105 g−1 at d7 or d20, followed by a more or less rapid decline leading frequently to no detectable parasites; (iii) six strains resulted in other profiles, i.e., undetectable infection (n = 1) or low parasite loads (n = 4), or late occurrence of parasites in the spleen (n = 1). No relationship was observed between profile and growth characteristics in vitro or zymodeme of the strain. Strains originating from CL never gave a visceralizing pattern in mice, but belonged more frequently to the avirulent type compared to VL strains. Strains from HIV-infected patients were not less virulent than those from immunocompetent individuals. These results showed that the course of L. infantum infection varies markedly with intrinsic parasite factors that display striking intraspecific variability.

Key words
  • Visceral leishmaniasis
  • Leishmania infantum
  • Experimental pathogenicity
  • Immunocompromised patient

1 Introduction

Overlapping of AIDS and visceral leishmaniasis is increasing sharply and Leishmania/HIV co-infections are being reported more and more frequently [1]. It is estimated that in southern Europe between 25 and 70% of adult VL cases are related to HIV and 1.5–9% of AIDS cases suffer from newly acquired or reactivated VL [2].

In the Mediterranean basin, Leishmania infantum is responsible for both localized cutaneous and visceral forms with rural and peri-urban distribution. Within this geographic area, this species shows a broad enzymatic polymorphism with 20 different zymodemes, of which 18 have been found in humans. L. infantum MON-1 is the most frequent zymodeme in humans and is usually responsible for VL along the Mediterranean littoral, more rarely for CL [3,4]. Zymodemes like MON-29 and MON-33, thus far isolated only from CL lesions in immunocompetent patients and considered as dermotropic [5], resulted in VL in HIV-infected patients [3]. Visceralization of parasites which are normally confined to the skin provides further arguments in favor of the opportunistic nature of L. infantum in HIV-infected individuals [6].

HIV infection also appears to modify the clinical manifestations of VL, with greater rates of recurrence, atypical course of infection and unusual locations: esophageal, gastric, pancreatic, rectal, pulmonary, pleural, cutaneous and hematologic (for details, see [12]). Mondain-Miton et al. [7] reported an atypical case of leishmaniasis in an HIV-positive patient which progressed from a cutaneous to multi-visceral form. Those peculiarities may be related to the immunodepression in HIV-infected patients, but also to biological characteristics of various strains.

The aim of this study was to explore the variability of the pathological expression of various strains of L. infantum, belonging to three different zymodemes and isolated from immuno-compromised and immunocompetent patients, in a murine model of VL.

2 Materials and methods

2.1 Leishmania strains

Twenty-two strains collected from various hospitals in France, and belonging to three zymodemes, were selected from HIV-infected or non-immunocompromised VL or CL patients as represented:

View this table:

Strains were cryopreserved at the International Leishmania Cryobank and Identification Center in Montpellier, France. Strains were characterized as zymodemes by determining the isoenzyme profiles of 15 enzymes using starch gel electrophoresis combined where appropriate with isoelectrofocusing [8,9]. Three zymodemes (22 strains) from immunocompromised and immunocompetent patients causing visceral or cutaneous leishmaniasis were used (Table 1). In vitro growth characteristics in RPMI medium (GIBCO-BRL, France) supplemented with 20% foetal calf serum (FCS) (GIBCO-BRL, France) were determined by flow cytometry for each strain after propidium iodide labeling [10]. Results are expressed as the mean number of promastigotes in culture established on three replicates at d4 and d7 (Table 1).

View this table:

Characteristics of 22 Leishmania strains from immunocompromised and immunocompetent patients

GroupSampleWHO codeZM (MON)Virulence patternGrowth in vitro (parasites×106)
VL/HIV+BMMHOM/FR/91/LEM 22591Vn.d.n.d.
BMMHOM/FR/94/LEM 29561V4.05.2
BMMHOM/FR/91/LEM 29591V2.81.3
BMHOM/FR/94/CRE 59/LEM 28121V7.55.3
BMMHOM/FR/92/LEM 238529V12.05.0
BMMHOM/FR/94/LEM 2811*29V5.32.5
BMMHOM/FR/94/LSL 22/LEM 29511R2.93.1
BMMHOM/FR/93/LEM 268029R6.33.5
BMMHOM/FR/91/LEM 217633R7.04.1
BMMHOM/FR/91/LEM 2328*33R5.95.4
BMMHOM/FR/94/LPN/ 05/LEM 28781U6.34.2
VL/HIVBMMHOM/FR/94/LPN 104/LEM 28751R1.30.4
BMMHOM/FR/94/LPN 106/LEM 28821R3.31.9
BMMHOM/FR/95/LPN 107/LEM 29491R2.94.6
BMMHOM/FR/92/LEM 25091R4.82.2
BMHOM/FR/94/LEM 28601U5.34.9
CL/HIVSMHOM/00/93/LEM 269829R3.73.2
SMHOM/FR/94/LEM 278333R4.44.4
SMHOM/FR/91/LEM 22171Un.d.n.d.
SMHOM/FR/94/LEM 28591U3.02.4
SMHOM/FR/91/LEM 22961U9.44.6
SMHOM/FR/93/LEM 272729U9.54.2
  • *, after relapse; B, blood; BM, bone marrow; S, skin.

  • n.d., not done; R, regulated; U, undetermined; V, visceralizing; ZM, zymodeme.

Mass cultures for animal inoculation were done on Schneider's Drosophila medium (GIBCO-BRL, France) supplemented with 20% BSA (GIBCO-BRL, France). For all experiments we used 1-week-old stationary-phase promastigotes. In order to assess whether there was a difference in infectivity between promastigotes in the logarithmic or stationary phase of growth, we measured parasite burdens at d7 and d20 in a preliminary experiment, in mice inoculated with 107 promastigotes of LEM 2259 and LEM 2811 of 3 or 7 days old. No difference was evident in liver and spleen parasite burdens of mice inoculated with either logarithmic (d3) or stationary (d7) phase parasites (data not shown).

2.2 Experimental animal infection

Female BALB/c mice 8 weeks old (IFFA-CREDO, France) were used. Animals housed in standard conditions were infected at d0 by i.v. injection of 107 stationary phase promastigotes of different strains. Animals were sacrificed at d7, d21, d60 and d90 postinfection and the parasite burdens were determined in liver and spleen homogenates using a culture microtitration assay [11]. Briefly, organs were excised, weighed then homogenised using a tissue grinder (ULTRATURAX, Staunfen, Germany), in 4 ml of Schneider's Drosophila medium supplemented with 20% heat inactivated FCS, penicillin (100 U/ml) and streptomycin (50 µg/ml) (BIOMERIEUX, France). Blood was withdrawn by retroorbital puncture. Serial fourfold dilutions of blood or organ homogenates, ranging from 1 to 4×106, were made in duplicate under sterile conditions in 96 wells microtitration plates containing 225 µl of culture medium. After 7–14 days of incubation at 26–28°C, plates were examined using an inverted microscope for the presence of mobile promastigotes. The final titre was the last dilution containing at least one parasite. Results were expressed as mean log (parasites g−1s+1) established with 4–5 mice for each control group.

2.3 Statistical analysis

The relationship of in vivo and in vitro growth characteristics of the strains was assessed using the Spearman non parametric rank correlation test between parasite burdens and number of promastigotes observed in culture at d3 and d7. Significance of the relationship between the profile of infection in mice and the zymodeme of the parasite strain or the patient's clinical status from which the strain originated, was calculated using Fisher's exact test.

3 Results

No mortality was observed in infected mice with any strain throughout the experiment. Evaluation of parasite burdens showed a great variability in the course of infection within the 22 strains. Three clearly distinct profiles of infection could be identified (Fig. 1a–c).

Figure 1

Evolution of parasite burdens in spleen (○) and liver (■) of mice inoculated i.v. at day 0 with 107 promastigotes of 22 strains of Leishmania infantum: a, visceralizing (‘V’), b, regulating (‘R’), and c, undetermined (‘U’) profiles.

(i) Visceralization (‘V’) (Fig. 1a), observed with six strains originating from HIV-infected patients. At the early phase of infection parasites predominated in the liver (104–106 g−1). After d20, a stagnation or a decrease of hepatic burdens occurred, whereas parasite numbers rose continuously in spleen, reaching 105−107 at the end of the experiment at d90–100.

(ii) Regulation (‘R’) (Fig. 1b), observed with four strains from HIV-infected patients, and 6 from immunocompetent patients (4 VL, 2 CL). Parasites were first observed in the liver especially at d7 and d20 (104−106 g−1), then parasite burdens rapidly declined in both liver and spleen, leading frequently to no detectable parasites at d90–100.

(iii) Undetermined (‘U’) (Fig. 1c), observed with one strain from an HIV-infected patient, and 5 from immunocompetent patients (1 VL, 4 CL). These profiles consist of low (n=4) or even undetectable infection (n = 1), or late occurrence of parasites in the spleen (LEM 2296).

The results obtained with the 22 strains used are summarised in Table 1. No significant correlation was found between in vitro growth characteristics of strains (i.e., number of promastigotes at d4 or d7 of culture), and parasite burdens observed in liver or spleen at any day of examination (−0.01 ≤ r ≤ 0.42, 0.99 ≥ P ≥ 0.07). No significant relationship was observed between parasite zymodemes and virulence patterns in mice (P = 1). The relationship between the clinical status of the patient from which the strain was isolated, and the infection pattern in mice was analysed:

(i) The most virulent ‘V’ profile of infection was more frequently observed in mice infected with strains from VL patients (6/16, 37%), compared to CL (0/6), but this difference was not significant (P=0.13, NS).

(ii) The ‘V’ profiles all originated from HIV positive VL patients (6/11, 54%), and none from 5 immunocompetent VL patients. However, the difference was not significant (P=0.09).

(iii) No ‘V’ profile originated from the 6 non-immunocompromized patients with CL.

(iv) The ‘U’ pattern (low or undetectable infection), was significantly more frequent with strains originating from CL (4/6, 66.7%), than with VL strains (2/16, 12.5%) (P=0.025).

4 Discussion and conclusion

VL due to Leishmania infantum is presently estimated to be an important opportunistic infection in HIV-infected patients [12]. In southern Europe, L. infantum is responsible for both CL and VL in immunocompetent subjects, where both dermotropic and viscerotropic zymodemes have been reported [13]. However, dermotropic zymodemes tend to visceralize in immunocompromized patients [3,6]. In this study we evaluated the pathogenicity in BALB/c mice of various strains of L. infantum belonging to different zymodemes, which were isolated from both HIV-infected or immunocompetent individuals. Our results showed that the course of infection observed in mice varied markedly according to the strain used, including strains belonging to the same zymodeme. The observed variability in the course of infection was unrelated to variations in growth characteristics of the strains in vitro. Three clearly distinct types of infection were well identified: ‘V’, ‘R’ and ‘U’.

With the ‘V’ strains, the course of infection markedly differed in liver and spleen. Whatever the strain used, the liver was the primary target after i.v. inoculation of L. infantum. This had been previously noted in experimental VL due to L. donovani in mice [14]. With the ‘V’ strains, however, the hepatic parasite burdens progressively decreased, or remained constant, after d21. On the other hand, involvement of the spleen, which was at least 10-fold less important at the onset of infection at d7, subsequently tended to rise continuously until the end of the experiment. At this time, parasite burdens in spleen were about 10–100-fold higher than in the liver. This discrepancy could be related to heterogeneity in the ability of resident macrophages to inhibit the growth of Leishmania, depending on the anatomical site they were derived from, as was observed with Candida albicans [15]. This clear-cut difference in the course of liver and spleen infections observed in our study suggests the importance of specific local macrophage and/or lymphocyte-macrophage interaction factors in the progression of L. infantum VL.

The clinico-pathological relationship between the experimental murine and the natural original human infections is unclear. Dermotropic isolates were expected to visceralize less frequently than viscerotropic ones. Indeed, 4–6 CL strains exhibited the less virulent ‘U’ phenotype, and none exhibited a ‘V’ profile. On the other hand, we could expect that strains isolated from HIV-infected patients would prove less frequently virulent to mice, since these patients display severe visceralization of parasites usually showing low virulence in immunocompetent patients [4,16]. However, all of the six most virulent (‘V’) strains in our study originated from HIV positive VL patients.

Host-related factors affecting VL have been extensively studied using L. donovani in various strains of mice. The course of infection has been clearly shown to be under host genetic control. The Lsh gene affects natural innate resistance (Lshr) or susceptibility (Lshs) [17]. Acquired immunological resistance/susceptibility, resulting in a ‘cure’ or ‘non-cure’ phenotype, is controlled by at least three loci (H-2, H-11 and Ir-2) in Lshs susceptible mice [1820]. However, the proper role of the parasite itself in the pathology of VL infection had not been defined. Our results showed that, in a given susceptible host, the course of L. infantum infection varies greatly with intrinsic parasite factors, and that these factors display marked intraspecific variation. Moreover the three main types of infection profiles identified in our study mimicked closely those phenotypes expressed with L. donovani in varying strains of mice: (i) the ‘U’ profile, consisting essentially of hardly detectable low infection if any, paralleling the innately resistant Lshr phenotype; and (ii) the ‘R’ and ‘V’ profiles corresponding to the immunologically acquired ‘cure’ and ‘non-cure’ phenotypes, respectively, evident in innately susceptible strains of mice.

A question of interest is to know what is the result when host-factors or parasite-related virulence phenotypes are varied independently from each other. Studies are in progress in our laboratory to assess if a given ‘U’, ‘R’ or ‘V’ phenotype expressed in BALB/c mice, will persist in mice of different genetic backgrounds. In a preliminary study, the ‘V’ profiles obtained with the LEM 2259 and LEM 2811 strains of L. infantum in BALB/c mice (which are H-2b, i.e., persisting infection haplotype with L. donovani), were both maintained in C57BL/6 mice (which are H-2b, i.e., recovery ‘cure’ haplotype with L. donovani), although parasite burdens were decreased by a 10-fold factor with the two strains in C57BL/6 mice [21].


This study was supported by a grant from the Agence National de la Recherche sur le SIDA, 66 bis Avenue Jean Moulin, 75014 Paris, France.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
View Abstract