Abstract
Although miltefosine and paromomycin were only recently introduced to treat visceral leishmaniasis, increasing numbers of miltefosine treatment failures and occasional primary resistance to both drugs have been reported. Understanding alterations in parasite behaviour linked to drug resistance is essential to assess the propensity for emergence and spread of resistant strains, particularly since a positive effect on fitness has been reported for antimony-resistant parasites. This laboratory study compared the fitness of a drug-susceptible parent WT clinical Leishmania infantum isolate (MHOM/FR/96/LEM3323) and derived miltefosine and paromomycin drug-resistant lines that were experimentally selected at the intracellular amastigote level.
Parasite fitness of WT, paromomycin-resistant and miltefosine-resistant strains, in vitro and in vivo parasite growth, metacyclogenesis, infectivity and macrophage stress responses were comparatively evaluated.
No significant differences in promastigote fitness were noted between the WT and paromomycin-resistant strain, while clear benefits could be demonstrated for paromomycin-resistant amastigotes in terms of enhanced in vitro and in vivo growth potential and intracellular stress response. The miltefosine-resistant phenotype showed incomplete promastigote metacyclogenesis, decreased intracellular growth and weakened stress response, revealing a reduced fitness compared with WT parent parasites.
The rapid selection and fitness advantages of paromomycin-resistant amastigotes endorse the current use of paromomycin in combination therapy. Although a reduced fitness of miltefosine-resistant strains may explain the difficulty of miltefosine resistance selection in vitro, the growing number of miltefosine treatment failures in the field still requires further exploratory research.
Introduction
The spread of primary antimony resistance in the Indian subcontinent has enforced the introduction of miltefosine and paromomycin for the treatment of visceral leishmaniasis (VL). Despite the increasing number of miltefosine treatment failures,1 clinical reports on primary miltefosine or paromomycin resistance in the field are still very scarce.2,3 Contrary to Leishmania donovani, miltefosine relapse isolates from Leishmania infantum-infected patients display a decreased susceptibility, which may possibly also be related to its veterinary use to treat canine leishmaniasis and L. infantum-associated HIV coinfections. Since incomplete parasite eradication is the rule in both groups, the large residual parasite reservoir will promote selection of miltefosine resistance upon repeated drug exposure.2–4 To support strategies concerning treatment and emergence of drug resistance, experimental selection of drug resistance can facilitate applied and fundamental ‘drug resistance’ research with the particular advantage that the drug-susceptible WT can be directly compared with matched derived resistant lines. While in the past resistance has mostly been selected in promastigotes, an earlier study by our group demonstrated that selection of drug resistance strongly depends on the selection protocol leading to the recommendation to use intracellular amastigotes whenever possible.5 Although former research mainly focused on unravelling resistance mechanisms, parasite fitness must be considered a relevant factor as well, potentially influencing the spreading potential of resistant strains. Comparison of unmatched antimony-susceptible and -resistant L. donovani strains from the Indian subcontinent indicated an enhanced fitness of antimony-resistant isolates.6–11 Likewise, a large-scale field study on L. donovani miltefosine cure and relapse isolates from Nepal suggested a higher in vitro infectivity of miltefosine relapse isolates.12 Recently, evidence for increased fitness was obtained after selection of paromomycin resistance in L. donovani promastigotes.13
The present laboratory study aimed to evaluate the impact of experimental miltefosine and paromomycin resistance on parasite fitness in an L. infantum strain isolated from an HIV coinfected patient.14 Besides promastigote growth and metacyclogenesis, the in vitro and in vivo amastigote growth pattern and the intracellular stress response upon macrophage stimulation were compared between the drug-susceptible parent WT and the miltefosine-resistant and paromomycin-resistant derived strains. While a decreased fitness could be demonstrated for the miltefosine-resistant strain, the paromomycin-resistant isolate displayed enhanced intracellular amastigote growth and survival.
Materials and methods
Parasite strains
The L. infantum field isolate (MHOM/FR/96/LEM3323) used for the experimental selection of both miltefosine and paromomycin resistance was obtained from the Centre National de Référence des Leishmania and was isolated from a French HIV patient.14 Promastigotes were grown in HOMEM medium (Gibco™, Life Technologies, Ghent, Belgium) at 25°C and subcultured twice weekly. Resistance was selected in intracellular amastigotes as previously described.5 Promastigote and amastigote susceptibilities were determined as previously described15 and are summarized in Table 1.
Promastigote and intracellular amastigote susceptibilities of the WT, miltefosine-resistant and paromomycin-resistant strains
| Strain | Intracellular amastigote susceptibility | Promastigote susceptibility | ||
|---|---|---|---|---|
| miltefosine (μM) | paromomycin (μM) | miltefosine (μM) | paromomycin (μM) | |
| IC50 ± SEM | IC50 ± SEM | IC50 ± SEM | IC50 ± SEM | |
| WT | 2.3 ± 0.5 | 89.3 ± 4.5 | 5.3 ± 0.3 | 129.8 ± 11.3 |
| Miltefosine resistant | >20 | 78.5 ± 9.2 | >40 | 177.8 ± 29.7 |
| Paromomycin resistant | 0.5 ± 0.1 | 212.6 ± 31.0 | 8.2 ± 1.1 | 138.1 ± 6.0 |
| Strain | Intracellular amastigote susceptibility | Promastigote susceptibility | ||
|---|---|---|---|---|
| miltefosine (μM) | paromomycin (μM) | miltefosine (μM) | paromomycin (μM) | |
| IC50 ± SEM | IC50 ± SEM | IC50 ± SEM | IC50 ± SEM | |
| WT | 2.3 ± 0.5 | 89.3 ± 4.5 | 5.3 ± 0.3 | 129.8 ± 11.3 |
| Miltefosine resistant | >20 | 78.5 ± 9.2 | >40 | 177.8 ± 29.7 |
| Paromomycin resistant | 0.5 ± 0.1 | 212.6 ± 31.0 | 8.2 ± 1.1 | 138.1 ± 6.0 |
For promastigote susceptibility, 105 log-phase promastigotes were exposed to serial dilutions of miltefosine or paromomycin for 72 h at 25°C. Parasite multiplication was assessed after addition of resazurin and fluorimetric reading. For amastigote susceptibility, primary peritoneal mouse macrophages were infected with metacyclic promastigotes at an infection ratio of 15 stages per cell. Total parasite burdens were microscopically assessed on Giemsa-stained wells after 5 days of incubation. The results are expressed as percentage reductions in parasite burden compared with untreated infected controls and IC50s were calculated. Susceptibility values are the result of at least three independent replicates and are expressed as the mean IC50 value ± SEM.
Promastigote and intracellular amastigote susceptibilities of the WT, miltefosine-resistant and paromomycin-resistant strains
| Strain | Intracellular amastigote susceptibility | Promastigote susceptibility | ||
|---|---|---|---|---|
| miltefosine (μM) | paromomycin (μM) | miltefosine (μM) | paromomycin (μM) | |
| IC50 ± SEM | IC50 ± SEM | IC50 ± SEM | IC50 ± SEM | |
| WT | 2.3 ± 0.5 | 89.3 ± 4.5 | 5.3 ± 0.3 | 129.8 ± 11.3 |
| Miltefosine resistant | >20 | 78.5 ± 9.2 | >40 | 177.8 ± 29.7 |
| Paromomycin resistant | 0.5 ± 0.1 | 212.6 ± 31.0 | 8.2 ± 1.1 | 138.1 ± 6.0 |
| Strain | Intracellular amastigote susceptibility | Promastigote susceptibility | ||
|---|---|---|---|---|
| miltefosine (μM) | paromomycin (μM) | miltefosine (μM) | paromomycin (μM) | |
| IC50 ± SEM | IC50 ± SEM | IC50 ± SEM | IC50 ± SEM | |
| WT | 2.3 ± 0.5 | 89.3 ± 4.5 | 5.3 ± 0.3 | 129.8 ± 11.3 |
| Miltefosine resistant | >20 | 78.5 ± 9.2 | >40 | 177.8 ± 29.7 |
| Paromomycin resistant | 0.5 ± 0.1 | 212.6 ± 31.0 | 8.2 ± 1.1 | 138.1 ± 6.0 |
For promastigote susceptibility, 105 log-phase promastigotes were exposed to serial dilutions of miltefosine or paromomycin for 72 h at 25°C. Parasite multiplication was assessed after addition of resazurin and fluorimetric reading. For amastigote susceptibility, primary peritoneal mouse macrophages were infected with metacyclic promastigotes at an infection ratio of 15 stages per cell. Total parasite burdens were microscopically assessed on Giemsa-stained wells after 5 days of incubation. The results are expressed as percentage reductions in parasite burden compared with untreated infected controls and IC50s were calculated. Susceptibility values are the result of at least three independent replicates and are expressed as the mean IC50 value ± SEM.
Promastigote growth
The growth profile of WT, miltefosine-resistant and paromomycin-resistant promastigotes was assessed by flow cytometry (FCM). Promastigotes were diluted in PBS (Gibco™, Life Technologies) for FCM counting, using a FACSCalibur® flow cytometer (BD Biosciences, NJ, USA) with addition of CountBright absolute counting beads (Molecular Probes™, OR, USA) as internal standard for quantification of the exact volume analysed. To generate growth curves, promastigotes of each strain were inoculated into 5 mL of HOMEM at exactly 5 × 105 promastigotes/mL. Every 24 h, three biological replicates were quantified in duplo for up to 240 h and analysed using BD CellquestPro® software. The average promastigote density at each timepoint was calculated and used to draw the final growth curves.
Promastigote metacyclogenesis
Promastigote morphology was evaluated microscopically to assess metacyclogenesis. The promastigote flagellum/cell body length ratio was determined and promastigotes were considered metacyclic when this ratio was >2.9 Starting from 96-h-old cultures and complementary to FCM assessment of promastigote density, a drop of promastigote suspension was Giemsa-stained every 24 h and visualized with bright field microscopy (Axiovert 200m®, Carl Zeiss) using a Zeiss Axiocam MRm®. The flagellum/cell body length ratio of ≥50 promastigotes was determined using Axiovision® software.
Promastigote infectivity
In vitro intracellular amastigote growth
In vivo amastigote multiplication
To evaluate the in vivo infectivity and growth of each strain, 12 female BALB/c mice were infected intracardially with 2 × 107 metacyclic promastigotes. Up to 28 days post-infection, three animals per group were sacrificed at weekly intervals to determine the parasite burdens in liver, spleen and bone marrow. Amastigote burdens are expressed as Leishman–Donovan units (LDU) after microscopic quantification of the Stauber index19 and by SYBR Green-based real-time PCR targeting the cysteine protease b (cpb) gene. Both the forward primer (5′-ATG TCT TAC CAG AGC GGC G-3′) and the reverse primer (5′-TCA CCC CAC GAG TTC TTG AT-3′) were purchased from Integrated DNA Technologies (Leuven, Belgium). To assess amastigote viability, a small piece of organ was placed in HOMEM medium and incubated at 25°C for 2 weeks to assess promastigote back-transformation.
Intracellular amastigote stress resistance
To evaluate the capacity of intracellular amastigotes to cope with intracellular stress, infected macrophages were exposed to either Escherichia coli-derived LPS and IFN-γ at concentrations ranging from 0.05 to 100 ng/mL (Sigma–Aldrich, Diegem, Belgium) or S-nitroso-N-acetyl-dl-penicillamine (SNAP) (Sigma–Aldrich, Diegem, Belgium) at concentrations ranging from 0 to 800 μM for 48 h, as described previously.7,13,20,21 Macrophage stress responses were determined by microscopic assessment of the percentage amastigote burden reduction compared with unstimulated infected control cells. Since previous research on WT and paromomycin-resistant strains suggested differences in host cell IL-10 production upon stimulation,13 the production of endogenous IL-10 in the supernatant upon stimulation with 100 ng/mL LPS and 5 ng/mL IFN-γ21 was measured using ELISA (eBioscience, Vienna, Austria).13 To correct for the variable infection ratio between strains and the associated differences in IL-10 production, the ratio of IL-10 production of stimulated infected cells was compared with unstimulated infected cells with correction for possible background IL-10 production.
Statistical analysis
All statistical analyses were performed using GraphPad Prism version 4.00 software. Statistical differences between WT and resistant parasites and between the different timepoints within one group were determined using two-way ANOVA with Bonferroni post hoc comparisons for parasite growth, parasite morphology and infection indices. Morphological and infection indices intergroup comparisons were done using the non-parametric Friedman test followed by Dunn's post hoc comparisons. Tests were considered statistically significant if P < 0.05.
Ethics statement
The use of laboratory rodents was carried out in strict accordance to all mandatory guidelines (EU directives, including the Revised Directive 2010/63/EU on the Protection of Animals used for Scientific Purposes that came into force on 1 January 2013, and the Declaration of Helsinki in its latest version) and was approved by the Ethics Committee of the University of Antwerp, Belgium (UA-ECD 2010-17, 18 August 2010).
Results
Promastigote growth
No statistical differences could be demonstrated between paromomycin-resistant and WT parasites (Figure 1a), whereas the miltefosine-resistant strain showed a significantly decreased growth pattern starting after 96 h of cultivation (Figure 1b). All strains reached the stationary phase leading to metacyclogenesis at ∼144 h.
Promastigote growth curves of WT, paromomycin-resistant (PMM-R) and miltefosine-resistant (MIL-R) matched isolates. (a) WT and PMM-R parasites show comparable growth. (b) MIL-R parasites show significantly decreased growth compared with WT starting at 96 h in culture (P < 0.001). Results are expressed as mean ± SEM and are based on three independent replicates run in duplicate.
Promastigote metacyclogenesis
Although some significant differences could be observed for WT and paromomycin-resistant parasites at 144 and 168 h in favour of paromomycin-resistant parasites, the predetermined metacyclogenesis cut-off value for both strains was reached after 192 h. At that timepoint, no differences in percentage metacyclics (∼64%) were observed between WT and paromomycin-resistant parasites (Figure 2a). The metacyclogenesis process in miltefosine-resistant parasites was much less evident (Figure 2b) and did not exceed 16% (data not shown). Despite several efforts to enhance metacyclogenesis and thus increase infectivity of the miltefosine-resistant strain by adaptations to the culture medium and enforced metacyclogenesis by promastigote preconditioning,18 no increase in percentage metacyclic promastigotes could be obtained (data not shown).
Metacyclogenesis of WT, paromomycin-resistant (PMM-R) and miltefosine-resistant (MIL-R) matched isolates. Promastigotes were labelled metacyclic when the flagellum/cell body length ratio exceeded the preset cut-off of 2 (grey broken line). (a) WT and PMM-R parasites demonstrated significant changes at 144 and 168 h. (b) WT and MIL-R parasites revealed significant morphological differences at all selected timepoints apart from 96 h. Results are expressed as mean ± SEM and are based on three independent replicates measuring the flagellum/cell body length ratio of ≥50 promastigotes. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Promastigote infectivity
All strains reached their highest infectivity after 144 h of cultivation. At that timepoint, the infection indices of WT and paromomycin-resistant parasites were not statistically different (Figure 3a), whereas the infection index of the miltefosine-resistant strain was markedly lower (Figure 3b), corresponding to its lower level of metacyclogenesis.
Promastigote infectivity for WT, paromomycin-resistant (PMM-R) and miltefosine-resistant (MIL-R) matched isolates. For all strains, the highest infection ratio was obtained at 144 h. (a) Infection indices of WT and PMM-R parasites differ significantly at 168 h. (b) Although the mean infection index of MIL-R promastigotes was maximal at 144 h, infectivity of MIL-R promastigotes was significantly lower at 144 and 168 h compared with WT parasites. Results are expressed as mean ± SEM and are based on three independent replicates run in duplicate. *P < 0.05; **P < 0.01.
In vitro intracellular amastigote growth
Although an initial increase in intracellular amastigote burden was observed for all strains, paromomycin-resistant amastigotes did show a notable advantage over WT and miltefosine-resistant parasites (Figure 4).
In vitro amastigote growth curves of WT, paromomycin-resistant (PMM-R) and miltefosine-resistant (MIL-R) matched isolates. (a) Comparison of WT and PMM-R parasites reveals a significantly enhanced amastigote growth for PMM-R parasites between 48 and 96 h post-infection (****P < 0.0001). (b) No significant differences could be observed for WT and MIL-R parasites. Results are expressed as mean ± SEM and are based on three independent replicates run in duplicate.
In vivo amastigote multiplication
Female BALB/c mice were infected with WT, miltefosine-resistant and paromomycin-resistant promastigotes for comparative monitoring of the amastigote multiplication ratio in the target organs. Every 7 days, three mice per strain were sacrificed and intracellular amastigote burdens in liver and spleen were determined by microscopic counting and real-time PCR. As bone marrow yields were insufficient to allow PCR, only microscopic counting was performed. Both in liver (Figure 5a) and spleen (Figure 5b), paromomycin-resistant parasites reached the highest burdens while no significant differences were noted between WT and paromomycin-resistant parasites in the bone marrow (Figure 5c). As could be expected based on their in vitro metacyclogenesis profile, miltefosine-resistant parasites displayed significantly lower burdens in all target organs.
In vivo amastigote growth of WT, paromomycin-resistant (PMM-R) and miltefosine-resistant (MIL-R) matched isolates in liver (a), spleen (b) and bone marrow (c) of infected BALB/c mice. Results are expressed as mean ± SD and are based on three independent replicates. The bar graphs represent the infection indices based on microscopic counting of Giemsa-stained smears and are expressed either as LDU for the liver or as the average number of amastigotes per nucleus when infection was limited in spleen and bone marrow. The broken line graphs represent the amastigote burden as determined by RT–PCR.
Intracellular amastigote stress resistance
While paromomycin-resistant parasites showed an enhanced tolerance towards nitrosative stress, no significant differences could be detected between miltefosine-resistant and WT parasites (Figure 6a). Measurement of the endogenous IL-10 production by infected macrophages upon stimulation with LPS and IFN-γ revealed significantly lower IL-10 production of miltefosine-resistant infected macrophages, whereas no significant differences could be observed for paromomycin-resistant parasites compared with paromomycin-susceptible WT parasites (Figure 6b).
Tolerance of amastigotes of WT, paromomycin-resistant (PMM-R) and miltefosine-resistant (MIL-R) matched isolates to cellular stress responses. (a) Tolerance to nitrosative stress upon SNAP exposure. No significant differences could be observed between WT and MIL-R parasites, while PMM-R parasites showed a significantly enhanced tolerance towards nitrosative stress inside primary peritoneal mouse macrophages (*P < 0.05). (b) Endogenous IL-10 production of infected cells stimulated with LPS and IFN-γ. Statistical differences between MIL-R and WT parasites were noted (*P < 0.05), whereas IL-10 production of WT and PMM-R strains did not statistically differ. Results are expressed as mean ± SEM and are based on four independent replicates run in duplicate. **P < 0.01.
Discussion
Miltefosine and paromomycin were approved for the treatment of VL to combat the expanding antimony resistance in the Indian subcontinent, where their use in combination therapy is now being explored as first-line option to avoid development of miltefosine or paromomycin resistance.22,23 Experimental selection of paromomycin-resistant L. donovani and L. infantum strains was shown to occur fairly rapidly both in vitro and in vivo5,24 and resistant clinical isolates have already been described.2 The number of miltefosine treatment failures in the Indian subcontinent has increased up to 20% within only one decade of use,1,25 without any link between relapse and susceptibility to miltefosine.1,26 In L. infantum, a noticeable decrease in miltefosine susceptibility was demonstrated in vitro for relapse isolates in Brazil4 and isolation of miltefosine-resistant isolates has been reported.2,3 Studying modifications associated with the resistant phenotype should provide valuable insights to monitor the emergence of drug resistance in the field. One of the phenotypic traits that recently gained a lot of interest is parasite fitness, defined as the complex interaction of numerous factors guaranteeing survival, reproduction and transmission between hosts in a given environment.27–29 While for most organisms the acquisition of drug resistance associates with several drawbacks, the effect of drug resistance in Leishmania on fitness remains fairly debatable and appears to be dependent on the respective drug and parasite species.30–34 For antimony-resistant L. donovani, various reports suggested a trend towards increased fitness enhancing the spreading potential of resistant parasites.7–9 Indian miltefosine relapse isolates displayed an enhanced in vitro metacyclogenesis and infectivity,12 a promastigote-selected miltefosine-resistant L. donovani strain was shown to cope better with oxidative stress35 and an experimentally promastigote-selected Leishmania major isolate demonstrated an increased metacyclogenesis profile.36 However, most of these studies relied on large groups of non-matched isolates and such findings should be interpreted with caution as the in vitro comparison of dissimilar Leishmania species or strains can be obscured by various species/strain-dependent traits.37 Ideally, resistance characteristics should be compared between the same pre- and post-resistance strain, but as few resistant clinical isolates are currently available, obtaining such couples may prove to be extremely challenging. Hence, experimental selection of drug resistance currently remains the only way to accommodate this need. Our research group developed in vitro2,5 and in vivo24 selection protocols on the intracellular amastigote stage. Compared with paromomycin,5 selection for miltefosine resistance proved to be much more complex with only one L. infantum isolate (LEM3323) gaining a definite miltefosine-resistant phenotype on both promastigote and amastigote levels.14 To assess the possible impact on fitness, LEM3323 and its miltefosine- and paromomycin-resistant derived counterparts were subjected to the same battery of ‘virulence’ assays. Although increased parasite fitness for a promastigote-selected paromomycin-resistant strain has been described,13 no apparent benefit of paromomycin resistance could be demonstrated for the promastigote stage. When focusing on the intracellular amastigote stage, the paromomycin-resistant strain did show marked benefits over WT parasites, e.g. enhanced in vitro and in vivo amastigote replication and resistance towards macrophage-induced stress responses.
After comparison of the WT and miltefosine-resistant strain, all results suggested a decrease in parasite fitness upon acquisition of miltefosine resistance under the stated experimental conditions. This decline was evident at both the promastigote and intracellular amastigote stages and was reflected by a reduced replication potential, metacyclogenesis and stress tolerance. Although it was hypothesized that miltefosine treatment outcome could be related to the enhanced infectivity and metacyclogenesis of miltefosine relapse isolates,12 the present study does not support this. Indeed, if relapse isolates did in fact display an enhanced infectivity, experimental selection of drug resistance would be more straightforward. The association between high infectivity and facilitated resistance development may explain why our selection protocol failed for most L. infantum and L. donovani isolates, and was successful only for the LEM3323 whose promastigotes were shown to cause massive macrophage infection and intracellular replication, even at an infection ratio of 1 promastigote/macrophage (S. Hendrickx, unpublished results). Another natural miltefosine-resistant clinical isolate (LEM5159)2 shared almost identical in vitro infectivity, though its amastigote replication profile was less pronounced (S. Hendrickx, unpublished results).
Although the fitness decrease in miltefosine-resistant L. infantum should be confirmed in other miltefosine-resistant isolates, additional research by our group has already suggested a decline in fitness upon repeated miltefosine exposure of amastigotes in vivo as a reduced amastigote-to-promastigote transformation ability.24 Despite the fact miltefosine resistance will certainly not be the sole factor contributing to the increasing levels of miltefosine treatment failures worldwide,1,4 it is encouraging to know that actual full-blown miltefosine resistance comes with a fitness cost for the parasite and may to some extent restrict extensive spread of primary resistance. With regard to the latter, parasite fitness in the vector should be explored as well. Finally, some pharmacokinetic and pharmacodynamic properties of miltefosine should encourage revision of the current miltefosine treatment regimen as drug exposure appears inadequate in some patients.38
Funding
This work was funded by the Research Fund Flanders (FWO: project G051812N). LMPH is a partner of the Antwerp Drug Discovery Network (ADDN; www.addn.be).
Transparency declarations
None to declare.
Acknowledgements
We want to acknowledge Dr Gaëlle Boulet for the skilful design of the PCR primers and the statistical analysis. Mandy Vermont and Pim-Bart Feijens are acknowledged for running the in vivo experiments.
