Kala-azar treatment with miltefosine is jeopardized as elevated level of miltefosine relapse has occurred in Nepal; we hypothesize that emergence of drug resistance in L. donovani could be associated with increased rate of relapse and focused to analyse association between miltefosine susceptibility of L. donovani and relapsed patients.
Methods: Confirmed kala-azar patients (n=11) were recruited for treatment (Miltefosine = 9 patients, Amphotericin = 2 patients) and they were followed up till 12 months. In vitro miltefosine susceptibility of parasites (n=12) were determined by promastigote based alamar blue assay, and correlated with treatment outcome of patients.
Results: We did not found significant difference in drug susceptibility of L. donovani in clinical isolates from miltefosine responsive and relapse kala-azar patients.
Conclusion: Relapse kala-azar with miltefosine treatment in Nepal was not driven by alteration in drug susceptibility phenotype of L. donovani, however analysis on large data is essential for validation. Host factors should be considered to unravel the complexity of treatment outcome that would strengthen the treatment efficacy.
INTRODUCTION
Kala-azar (KA) (also known as visceral Leishmaniasis, VL) is a life threatening vector borne parasitic disease caused by Leishmania donovani and transmitted by the bite of female sandflies in Indian subcontinent. More than 60% of global burden of KA are confined among three countries namely Bangladesh, India and Nepal of the Indian subcontinent where it has also been estimated around 200 million people at risk of KA infection (Joshi et al., 2008). In this context, Kala-azar elimination programme (KAEP) has been recently launched in this greatly affected region in order to reduce KA incidence 1 in 10,000 per annum by the year 2015. The main arms of the KAEP are early diagnosis by rapid dipstick test, effective treatment by oral drug miltefosine and vector control by insecticide spraying (WHO, 2005).
Miltefosine (hexadecylphosphocholine) is a phosphocholine analogue that interferes with cell-signalling pathways was initially registered for the antineoplastic activities but fell out due to severe gastrointestinal toxicities. In contrast, the entry of miltefosine into the therapeutic armamentarium of Leishmaniasis is considered as a landmark event as for the first time, an orally effective antileishmanial agent has been identified (Sundar et al., 2005). However, from the very beginning doubtful results were coming from the clinical trial and the field based studies such as the doubling of treatment failure rate in India during clinical trials from phase III (3%) to phase IV (6%) (Sundar et al., 2006) and 11% miltefosine relapse rate from Nepal (Rijal et al., 2005). More recently, more than 20% of miltefosine relapse has been reported in Nepal which definitely threaten the efficacy of miltefosine treatment (Ostyn et al., 2010) and would jeopardize the ongoing KAEP as well.
In addition, laboratory evidence also showed that miltefosine has a long half life of 150 hrs and 28 days long treatment course which also risk the treatment efficacy of drug (Perez-Victoria et al., 2006). Indeed, development of drug resistant phenotype with drug pressure of miltefosine (Seifert et al., 2003) and persistent in vivo infectivity of resistant L. donovani (Seifert et al., 2007), there could be great risk of emergence of miltefosine resistant L. donovani that consequence to miltefosine non-responsive/relapse case. So there is urgent need for the assessment of miltefosine susceptibility in clinical L. donovani isolates to determine the emergence of miltefosine resistance in L. donovani. Therefore, in this work we focus to analyse, the association between drug susceptibility of L. donovani and KA case treated with miltefosine in Nepal.
MATERIALS AND METHODS
Patients
Confirmed KA patients (clinical criteria: fever for more than 2 weeks, hepatomegaly, spleenomegaly, rK39 and bone marrow smear positive) were recruited during 2004 (phase IV miltefosine trail), and May 2009 to July 2010 at B. P. Koirala Institute of Health Sciences (BPKIHS), Dharan, Nepal, the tertiary care medical centre. Out of 11 KA patients, 9 patients were treated with oral miltefosine (Definite cure = 2, relapse = 7), and 2 patients were initial cured but treated with Amphotericin-B i.e. miltefosine control. Written informed consent was obtained from patients and from parents (or guardians) in case of children. Ethical clearance was obtained from the institutional review boards of Nepal Health Research Council, Kathmandu, Nepal and the corresponding body of the Institute of Tropical Medicine, Antwerp, Belgium.
Treatment
Diagnosed kala-azar cases were treated by oral miltefosine for 28 days except pregnant case and other complication as described elsewhere (Rijal et al., 2005). Clinical and laboratory methods were applied in order to evaluate the end of treatment, after 6th and 12th months respectively. Considering the laboratory results, presence or absence of parasites at the end of treatment were used to consider initial cure/responder, non responder, relapse and definite cure. Absence of parasites were considered as initial cure/responder where as the presence of parasites were considered as non-responders. Initial cures were also further evaluated at 12 month and were considered as definite cure. Relapse cases were indicated as the initial cured patient but reappearance of KA symptoms with positive parasitemia during follow up.
Parasites isolation
Parasite promastigotes were isolated in tobie's blood agar medium with locke's overlay (Tobie et al., 2001) with penicillin 200IU/ml and streptomycin 200µg/ml from KA confirmed patient. Once the parasites were fully obtained from the clinical materials, the parasites were dispatched to ITM, Belgium in order to perform further in vitro drug susceptibility testing. Upon their arrival at ITM-A, parasites were cryopreserved in liquid nitrogen with 10% final concentration of sterile glycerol (Evans, 1993) until the start of drug susceptibility testing.
Twelve L. donovani strains were included in this study. Out of them, 2 strains were isolated from miltefosine responsive KA patients i.e. definite cure, 8 strains isolated from relapse KA patients, and 2 strains were isolated from clinical cured patients who were treated with Amphotericin-B i.e. miltefosine control.
In vitro drug-susceptibility testing
Briefly, parasites were taken out from the cryobank and cultured in Modified Eagle's Medium/HOMEM (Berens et al., 1976) (Gibco, USA, Cat. Number 041-94699M) with 20% heat inactivated fetal calf serum (HIFCS) (PAA Laboratories GmbH) with 5 X 105 parasite per/ml then incubate at 26°C. The parasite population dynamics were microscopically observed every 24 hours by Uriglass disposable plastic counting chamber slide (Menarini diagnostics, ITALY) until 7 days to determine growth and differentiation profile of the parasite. The parasites were again sub-cultured until they were growing more feasible with consistent growth, parasites were cultured in 4 flasks of HOMEM+20%HIFCS for the miltefosine susceptibility assay.
Log phase promastigotes from third day of sub-culture were centrifuged at 2061g for 10 minutes; parasites were counted and diluted to 106 parasite/ml. 100µl parasites were plated out in duplicate on sterile 96 well plates (Corning Costar, USA, Cat. Number CLS3596) and 100µl of miltefosine (donated by Aeterna Zentaris Gmbh, Germany) prepared in 10 different concentrations (Final concentration in well 40µM, 25µM, 16µM, 10µM, 6.4µM, 4µM, 2.56µM, 1.6µM, 1.024µM and 0.64µM) were also added serially in wells. One untreated control and one Blank (HOMEM+20%HIFCS) per strain were plated out in duplicate along the treated wells. The plate was sealed and incubated at 26°C. After 3 days, 20µl of alamar blue (12.5mg/ml) was added in each well and incubated for 24 hrs. Fluorescence on each well was determined by Victor X3 Multilabel Reader (PerkinElmer, USA) with excitation at 560nm and measuring emission at 590nm.
Calculation of EC50 and activity index
The observed fluorescence reading for wells with parasite were reduced from the blank, these values were transferred in a GraphPad Prism version 5 and EC50 values (molar concentration of drug that produce 50% of maximal possible response to drug) were analysed for each strain using sigmoidal dose-response model (Non-linear Regression) for miltefosine assay (Motulsky et al., 2004). The Linear method for determining EC50 was also used, which was analysed by entering, above 50% and below 50% inhibition percent signal in cells of excel formula based on the Linear regression method (Maes et al., 2010). In this experiment BPK206/0 clone10 was used as a reference strain for each experiment of miltefosine susceptibility assay. The ratio of EC50 of a test strain to the EC50 of the reference strain (range 2.7 to 6.5 µM, based on experimental series), which we defined the "Activity Index" (AI), was used to express the in vitro susceptibility of the analysed strain that normalize the result obtained from different series of experiments (Yardley et al., 2006).
Data analysis
The detail records of miltefosine treatment outcome of patient, and parasite isolates well documented in MS-Access 2003 Version. Statistical analyses (Unpaired T-test, ANOVA, Correlation) were done at a critical α-level of 0.05 by entering data in GraphPad Prism Version 5 (Motulsky, 1995).
RESULT
Growth curve of parasite
Parasites' growth curves analysis showed that maximum population density were reached at the end of actively growing log phase and decline with early stationary phase. Due to some technical problems (the missing data on growth curve was due to the restricted lab work at weekends according to the policy of ITM) data of population density at late log phase for some parasites were lacking (BPK502/0, BPK505/0). Detail analysis showed that the parasite NEP052/3 was the rapid grower since it had comparatively higher population density at third day of growth curve. In contrast, parasites BPK499/0, BPK455/0, BPK514/0, BPK455/2, had comparatively lower population density at day 3. Parasites isolated from patients at relapse were comparatively rapid growing since they had comparatively higher average population density at day3 (Fig. 1). However, parasite growth rate was different in clinical isolates from various treatment outcomes; no significant differences found in parasite populations at day3 (as confidence interval overlapped).
EC50 of parasite
EC50 values of 12 parasites calculated by GraphPad prism were ranged from 1.8 to 8.1µM
(mean= 4.4µM, SEM= 0.6µM) and linear method determined the EC50 values range from 1.8 to 8.1 µM (mean= 4.4µM, SEM= 0.6). Both methods showed similar results and found significant correlation (r=0.9, P<0.05), as shown in Fig. 2. Most of EC50 analysed by Linear method was equal or lesser than analysed by GraphPad, except the EC50 of three parasites (BPK455/2, NEP098/6 NEP107/6) was found higher than the analysis done by GraphPad (Fig.3), EC50 values were not significantly different, as 95% confidence intervals overlap. The reference strain BPK206/0 clone10, showed average EC50 value 4.3µM ± SEM 0.9µM miltefosine (Range from 2.7µM to 6.4µM) when analysed by GraphPad, and average of 4.1µM. ± SEM 0.9µM (Range from 2.8µM to 6.5µM) EC50 was determined with Linear method. Since, the Linear method was based on two values of above and below 50% inhibition signals, the slope of regression curve for Linear method was slightly different from GraphPad method (non-linear regression), and found the minor differences in EC50 values for same parasite.
Activity index of parasite
Activity indices of parasites, analysed by both GraphPad (mean= 1.3 ± SEM= 0.2, range= 0.3 to 2.7) and Linear method (mean= 1.4 ± SEM= 0.2, range= 0.3 to 2.7) were not significantly different (r= 0.9, P-value<0.05), except a few parasites which showed minor difference. Highest activity index was observed in BPK514/0 (GraphPad AI= 2.7, Linear AI= 2.6), and least activity index observed in BPK499/0 (GraphPad AI= 0.3, Linear AI= 0.3). Activity index of reference strain BPK206/0 clone10 (AI= 1) was included as a baseline value for the comparison of the studied parasites. The detail analysis revealed that 6 parasites (BPK455/2, BPK509/0, BPK514/0, NEP052/3, NEP098/6 and NEP107/6) had activity indices 2.1, 1.4, 2.7, 1.4, 1.7, and 2.0 respectively, which was greater than reference strain, and remaining parasites had either equal (AI = 1) or lower activity indices (AI < 1) than reference strain.
The average parasite population had no relation with parasite activity indices of parasites even though parasites were isolated from different treatment outcome (Fig. 4). However, growth phage of parasite had affected the miltefosine activity index, as least activity index of BPK499/0 found in relation with extended log phase, and comparatively short log phase in parasite from relapse cases showed higher activity indices.
In vitro drug susceptibility and treatment outcome
The results of the drug susceptibility of parasites isolated from KA patients with different treatment outcomes were depicted in Fig. 5. Briefly, the activity indices of parasites from miltefosine responders were comparatively lower (mean= 0.6 ± SEM= 0.3µM) than the parasite from miltefosine control (mean= 1.0 ± SEM= 0.4µM). Further analysis showed that the activity indices were 1.5 to 2 fold higher in parasite isolates from miltefosine relapse cases than miltefosine control, both before treatment (mean= 1.9 ± SEM= 0.8µM) and at relapse (mean= 1.5 ± SEM= 0.2µM). Similarly, the paired parasite isolates (parasites isolated from same patient, before and after treatment) showed 2 fold increased activity index at relapse than before treatment (BPK455/0= 1.0, BPK455/2= 2.1).
However, the activity indices of parasites from miltefosine relapsed case was higher than parasite from the miltefosine responsive and miltefosine control, statistically the difference was not significant (ANOVA, Dunnett's test, P-value >0.05, 95% confidence interval overlapped). We did not also found significantly higher miltefosine activity index in parasite isolated from miltefosine relapsed KA.
DISCUSSION
Miltefosine has been recently introduced to treat KA in Indian subcontinent; therefore, extensive reports on drug susceptibility testing among the clinical isolates of L. donovani are scant. However, recent alarming report of >20% miltefosine relapse in Nepal (Ostyn et al., 2010) indicated that there would be unpredictable threat on efficacy of miltefosine at Indian sub-continent. In this context, our preliminary study on drug susceptibility testing with parasite level showed that the EC50 of clinical parasite isolates range from 1.8 to 8.1µM (mean= 4.4 ± SEM= 0.6µM), which was lower than the Seifert et al. 2003 (mean 5.1 ±
SEM= 0.02µM), but not significantly different as 95% confidence interval overlapped. In contrast, Escobar et al. 2002 reported very low EC50 of L. donovani promastigote to miltefosine (0.4 ± SEM= 0.04 µM), which was inconsistent with our result. The contrasting results of EC50 were due to the different protocol of drug susceptibility used, parasites counted to measure drug mediated inhibition in third study, but the fluorescence method used in first and second study. Similarly, Yardley et al. 2005 reported EC50 (amastigote-macrophage model) of L. donovani isolated from Nepal was 6.9 ± SEM= 1.2µM which was higher than our results but not significantly different (95% confidence interval overlapped), however the methods of drug susceptibility assessment were different between two experiments. Furthermore, parasites strains from different geographical regions could possess different drug susceptibility phenotype, and hence inconsistent drug susceptibility found in aforementioned experiments.
We did not find correlation between parasites population at day 7 and drug susceptibility of parasite, the result was consistent with Natera et al. 2007 as there was no such predictive value between growth rate and drug susceptibility of parasite. However, the comparison of growth pattern of parasites indicated that parasites from relapse patient were different from others treatment responsive parasites as they reach the stationary phase with high population rapidly than others. However, log phase parasite is more susceptible to drug than stationary phase and they would have lower activity index than parasite at stationary phase. Hence, the least activity index of BPK499/0 was found in relation with the extended log phase. Parasite growth is not always in same nature, the growth could be affected by several factors such as nutrient, pH, high density sensing behavior of parasites (Vassella et al., 1997). Therefore, we used the reference strain (BPK206/0 clone10) to standardize the growth status of parasite for drug susceptibility assay. We found average EC50 of reference parasite (BPK206/0 clone10) 4.2 ± SEM= 0.9µM, which was lower than unpublished report
(mean= 6.5µM ± SEM= 0.9µM) by Salotra 2010, but not significantly different. However, experimental error is most frequent among the different experiments, the analysis of drug susceptibility assessment between different experiments are normalized with use of activity index, in which EC50 of reference strain is considered as baseline (da Luz et al., 2009).
In this experiment, we did not determine the significant difference in miltefosine susceptibility between the clinical parasites from relapse KA and drug responsive KA. There was no association determined between treatment outcome of KA patients and in vitro drug susceptibility results of parasites. Hence, none of clinical parasites had significantly higher activity index among 12 parasites; it could be due to small sample size. We did not have cut-off value of activity index for drug susceptibility that discriminate parasites based on drug susceptibility phenotype. Therefore, we were unable to categorize parasites into miltefosine resistant and sensitive. Furthermore, our result was consistent with Indian strain of L. donovani (Salotra, 2010). Treatment outcome of non-responder/relapse is complex multi-factorial in nature, since efficacy of drug depends on several factors such as host immunity, drug resistance phenotype of parasite, incomplete treatment course etc. The previous study on the relation between antimony in vitro susceptibility and corresponding treatment outcome in Nepal was also ambiguous (Rijal et al., 2007); although Indian study was straightforward between in vitro susceptibility and corresponding treatment outcome (Lira et al., 1999). Since miltefosine adapted L. donovani developed with in vitro drug pressure (Salotra, 2010, Seifert et al., 2003), had similar infectivity to develop KA infection in mice (Seifert et al., 2007), there could be great risk of emergence of resistance phenotype in L. donovani parasite. In vitro induced resistant strain had average 32 fold increased resistant than wild type strain (Salotra 2010) when there was stepwise increased drug exposure. The plasma levels of miltefosine in human body are dose proportional and urine excretion is negligible. Due to the long half-life, a sub-therapeutic level of miltefosine could remain for some weeks after a treatment course. This in vivo remaining of miltefosine might encourage the emergence of resistance in parasite (Bryceson, 2001). Hence, the clinical parasite isolates might develop similar degree of increased drug resistance as found with induced resistant; however, host immunity may affect the in vivo drug resistance development in parasite. The association of drug susceptibility of parasite and relapsed KA would be more precise with analysing paired parasites, as difference in drug susceptibility of parasite before and after treatment, would describe the strong evidence on role of drug for relapsed KA; however, it was not possible in this experiment due to lack of paired samples.
In vitro promastigote assay for drug susceptibility is not entirely predictive to study association between parasite phenotype and treatment outcome of KA, as it lacks the immunity factor and in vivo interaction could be different than in vitro assay (Yardley et al., 2005). However, miltefosine do not affect the macrophage immunity and direct toxic effect to the parasite (Griewank et al., 2010); the efficacy of miltefosine could be dependent on the biological nature of host cell (Wadhone et al., 2009). There could be some host factor, that have a significant role to reduced the host's anti-leishmania immunity to develop relapse/non-responsive KA, which was the beyond the level of this experiment. To understand the relationship between clinical relapse and parasitic phenomenon, more works are urgently required and to apply the developed tools on the large samples coming from KA endemic region which would only determine the efficacy of the drug to control KA in Indian subcontinent.
