Definitions

A. Treatment outcome visceral leishmaniasis (VL)

• Definite cure: Bone marrow or spleen aspirate is negative at the end of therapy and at 6 months follow-up with no evidence of relapse till 12 months follow-up.
• Non-responder: Bone marrow or spleen aspirate remains positive at the end of therapy.
• Relapse: Bone marrow or spleen aspirate is negative at the end of therapy but positive during 12 month follow-up.
• Treatment failure: Either non-responder or relapse.

B. Treatment outcome cutaneous & mucocutaneous leishmaniasis (CL & MCL)

• Initial cure: Complete scarring of lesions and disappearance of inflammatory signs < 3 months post-treatment.
• Definite cure: Initial cure confirmed 6 to 12 months post-treatment.
• Non-response: Absence of initial cure at 3 months post-treatment or worsening of the lesion during or < 3 months post-treatment.
• Relapse: Reactivation of the lesion(s) after initial cure.
• Treatment failure: Non-response or relapse.

C. SbV sensitivity Leishmania isolates

Protocol:
The SbV sensitivity of all isolates was tested in vitro within seven passages after isolation using the following protocol:

Promastigotes were maintained in M199 medium supplemented with 20% heat-inactivated foetal calf serum, at 25°C. Late stage promastigotes were used to infect primary isolated mouse peritoneal macrophages at a ratio of 7 promastigotes to 1 macrophage in quadruplicate, and kept at 37°C, 5% CO2/air mix. Twenty four hours after infection, the level of infection was assessed. If infection level was higher than 80%, the infected cultures were exposed to pentavalent antimonials (3 ≠ sources, including Pentostam, GSK) over a dose range of 60, 30, 10 and 3 µg/mL. After five days, the % of infected macrophages in each well was determined by microscopy. From a comparison of counts from treated with untreated cultures, the % inhibition was calculated by sigmoidal regression analysis (MS xlfit™) and ED50 (ED90) values were determined.

Reference strains:

For Old World isolates: L. donovani MHOM/ET/67/HU3, a WHO reference strain sensitive to sodium stibogluconate and meglumine antimoniate was included in each assay as a reference.
For New World isolates: L. braziliensis MHOM/BR/75/M2903, a WHO reference strain sensitive to sodium stibogluconate and meglumine antimoniate.

Definition in vitro SbV sensitivity of a Leishmania isolate:

Activity Index (A.I.) = EC50 tested isolate / EC50 reference strain

Isolated Strains & Patient Samples

A. Nepal

Pathology	Isolates (reference cryobank ITMA)	Samples (all patients)
VL:	103 isolates from confirmed VL patients
	(including 21 isolates from unresponsive or relapse patients)	1) Bone marrow for PCR
	Current status characterisation:	use: validation resistance markers
	- 43 isolates species typed: all L. donovani	2) Serum samples
	- 27 isolates in vitro SbV sensitivity tested	use: cytokine profiling (see immunology)
PKDL:	4 isolates	skin biopsies & serum samples
Table: Summary of isolated strains & patient samples collected at BPKHIS, Nepal.

B. Peru/Bolivia

Pathology	Isolates (reference cryobank ITMA)	Samples
CL:	302 isolates from confirmed CL patients
	(including 51 isolates from unresponsive or relapse patients)	1) 272 skin biopsies for PCR
	Current status characterisation:	use: validation resistance markers
	- 99 isolates species typed: 48 L. braziliensis, 27 L. guyanensis, 14 L. peruviana, 3 mixed L. braziliensis/L. peruviana, 6 L. lainsoni, 1 L. mexicana	2) 219 skin biopsies for immunology
	- 39 isolates in vitro SbV sensitivity tested
MCL:	81 isolates from confirmed MCL patients	1) 61 biopsies for PCR
	Current status characterisation:	use: validation resistance markers
	3 isolates species typed: all L. braziliensis	2) 37 biopsies for immunology
	4 isolates in vitro SbV sensitivity tested
Table: Summary of isolated strains & patient samples collected at IMTavH, Peru and CUMETROP, Bolivia

Treatment Outcome

A. Visceral Leishmaniasis (Nepal)

Recruitment & Treatment (January 2002 to September 2003)
If VL was confirmed by parasitological examination of bone marrow or spleen, a 30 day SbV treatment was started as follows:

1) 20 mg/kg/day (intramuscular); drug used: SSG (Albert David, India), all used batches were sent for quality control at the International Dispensary Association (Amsterdam), all batches had sufficient SbV-content.
2) All patients were hospitalised for the full duration of therapy. (If impossible for the patient, a full supply of SSG was given for ambulatory treatment at the nearest health facility.)
3) Direct Observed Therapy (DOT) = a daily record was made of the administered dose of drug, any interruptions in treatment were recorded.

Follow-up (FU) of treated patients at end of treatment and 3, 6 & 12 months after treatment with:

1) Repeated parasitological examination of bone marrow at (i) end of therapy, (ii) 6 months FU and (iii) anytime if suspicion of relapse.
2) Patients were actively traced if they missed a hospital appointment.

Treatment outcome of all treated patients is given in the table below:

301 suspect cases
224 definite VL		67 non VL 10 probable VL
198 SSG treated		24 treated with amphotericin B 2 defaulted before treatment
181 > or 25 days SSG treated		17 < 25 days SSG treated
169 completed 12 months FU		12 patients lost to FU
153 definite cure 90.5%	16 treatment failure 9.5% (14 non-responders, 2 relapse)

B. Cutaneous and Mucocutaneous Leishmaniasis (Peru)

Recruitment & Treatment (November 2001 to August 2004)

If CL or MCL proven (by direct examination, culture and/or PCR), SbV treatment was started as follows:
1) For CL: 20 mg/kg/day for 20 days (intravenous)
2) For MCL: 20 mg/kg/day for 30 days (intravenous)
3) Drug used: Glucantime (Aventis), Generic SSG from Colombia (Viteco SA), India (Albert David Ltd, Calcutta) and Peru (Marfan); all used batches were sent for quality control at the International Dispensary Association (Amerstam), all batches had sufficient SbV-content.

Follow up (FU) of treated patients at 1,3,6 & 12 months after treatment with parasitological evaluation if suspicion of treatment failure.

Treatment outcome of SbV treated patients with 12 month FU complete is presented in the table below:

CL		MCL
113 responders 71.4%	45 non-responders 25.5%	17 responders 73.9%	6 non-responders 26.1%
No statistical difference between 2 clinical presentations.

Risk Factors for Treatment Failure

As factors on the level of treatment and host can potentially cause SbV failure, an analysis was performed using the clinical and epidimiological data collected for all patients, to identify clinical risk factors for SbV treatment failure.

A. Visceral Leishmaniasis (Nepal)

Using univariate analysis, it was demonstrated that treatment failure is not associated (P-value > or = 0.05) with sex, age, nutritional status, spleen size or other clinical parameters, hemoglobin count, presence of concomitant disease or batch of SSG. The following factors were identified as risk factors for VL treatment failure:

• Ambulatory treatment (OR=10.87, 95% CI (1.26, 93.37), P-value=0.0297)

  may be related to:
  1) Inadequate dose of SbV treatment
  2) Possible loss of potency of the drug due to inadequate storage (SbV should be kept in dark, cool place)

• Fever > 12 weeks (OR=9.63, 95% CI (2.45, 37.84), P-value=0.0012)

  Prolonged fever reflects the prolonged duration of illness. Therefore immune status may be more depressed thus risk of increased treatment failure.

• Origin Western districts (OR=6.91, 95% CI (1.61, 29.72), P-value=0.0094)

  The Western districts are located closer to the Bihar focus (India) and thus there is a higher likelihood of harboring SbV resistant strains.

B. Cutaneous Leishmaniasis (Peru)

Using univariate analysis, it was shown that CL treatment failure is not associated (P-value > or = 0.05) with nutritional status of the patient, location, size and number of skin lesion(s), total dose of antimonials administered and the status of the lesion at the end of treatment. The following factors were identified as risk factors for CL treatment failure:

• Younger age (< 16 years) (OR=3.75, 95%CI = (1.51,9.38), P-value=0.001) and shorter stay (< 12 months) in transmission area (OR=2.64, 95% CI (0.97, 8.37), P-value=0.04)

  May involve:
  Lower risk of previous exposure(s) to infectious bites.
  Age-related difference in pharmacokinetics of SbV ?

• Duration of disease < 3 months (OR=3.14, 95% CI (1.18, 8.67), P-value=0.01)

  Probably explained by lower contribution of the cellular immunity in younger lesions.

• 1 or more nodule(s) (OR=7.78, 95% CI (0.89, 64.52), P-value = 0.064) or multiple lesions of various types (OR=78.86, 95% CI (8.6, 723.19), P-value= < 0.001)

  Association with infection with L. braziliensis or L. peruviana.
  Possible insufficient penetration in nodular lesions of antimonials ?

SbV Sensitivity of Leishmania Isolates

All in vitro SbV sensitivity tests were performed in an amastigote-macrophage model, as described above. The results for isolates from VL patients and CL patients are given below.

A. VL isolates, Nepal

In vitro SbV sensitivity was so far tested for 27 isolates, all characterised as L. donovani, including 20 isolates from patients with definite cure and 7 isolates from patients with treatment failure. The results are listed in the table below:

Activity Index	Definite Cure (n=20)	Treatment Failure (n=7)
0	3	0
1	3	2
2	2	0
3	2	0
4	0	0
5	0	0
6	10	5
P-value=0.5339 Table: In vitro SbV sensitivity of isolates vs. treatment outcome

We conclude:

1) Natural SbV resistant L. donovani isolates are present in Nepal
2) There is no association between clinical outcome and the parasite’s in vitro SbV sensitivity. Factors not present in the in vitro model could affect treatment outcome e.g. immunological factors, drug pharmacokinetics, parasite adaptation in vivo.

Treatment failure is not a synonym for drug resistance.

B. CL isolates, Peru

Among the isolates for CL, several species were encountered (see above). SbV resistant isolates were found within the species L. braziliensis, L. guyanensis, L. peruviana and L. lainsoni. However, there seems to be a significant association between in vitro SbV sensitivity and the specific Leishmania species (see table below), with L. guyanensis (median activity of 2) as the most sensitive species in vitro.

Species	No.	Median A.I. (range A.I.)
L. braziliensis	26	6 (1-6)
L. guyanensis	5	2 (0-5)
L. peruviana	1	5 (5)
L. lainsoni	4	5 (5)
P-value=0.039 Table: Species vs in vitro SbV sensitivity

For patients with a defined outcome (12 months FU complete), there is a significant association between treatment outcome and parasite species (see table below). Patients infected with L. guyanensis had a 94% likelihood of responding to treatment whereas for patients infected with L. braziliensis or L. peruviana this was reduced to 60%. This supports the association between species and in vitro sensitivity; L. guyanensis is more sensitive to pentavalent antimony than L. braziliensis and L. peruviana.

Species	Definite Cure (n=38)	Treatment Failure (n=13)
L. braziliensis	11 (58%)	8 (42%)
L. guyanensis	17 (94%)	8 (6%)
L. peruviana	6 (60%)	4 (40%)
L. lainsoni	4 (100%)	0 (0%)
P-value=0.029 Table: Species vs. treatment outcome (12 months FU)

Treatment failure is not a synonym for drug resistance.

SbV Resistance Markers & Mechanism

Comparative studies on transcriptomic level of SbV resistant and SbV sensitive isolates (as shown by in vitro SbV sensitivity testing) has resulted so far in the identification of 4 putative molecular markers, summarised in the table below:

Marker	Features	Species *	Hypothetical Mechanism ?
DPF	= Differential Polyadenylation Factor Precursor ribosomal RNA is polyadenylated in Leishmania, the degree of polyadenlyation varies in all SbV resistant and SbV sensitive isolates tested so far. The DPF reflects the characteristic polyadenylation of a particular strain; SbV resistant strains have a DPF < or = 1, while the SbV sensitive isolates have a DPF > 1. (More Info...)	L. braziliensis L. donovani	Unknown
GCS	= gene coding for enzyme gamma–glutamylcysteine synthetase, catalyses the rate limiting step in the synthesis of glutathione (thiol moiety of trypanothione). GCS is lower expressed in SbV resistant compared to SbV sensitive isolates.	L. donovani	decrease thiol biosynthesis, leading to inhibition activation SbV
ODC	= gene coding for enzyme ornithine decarboxylase, catalyses the rate limiting step in the synthesis of the spermidine moiety of trypanothione. ODC is lower expressed in SbV resistant compared to SbV sensitive isolates.	L. donovani	decrease thiol biosynthesis, leading to inhibition activation SbV
AQP1	= gene coding for the protein aquaglyceroporin 1; catalyses the uptake of SbIII (the active form of SbV). AQP1 is lower expressed in SbV resistant compared to SbV sensitive isolates.	L. donovani	decrease uptake of active form SbIII (resulting from activation in macrophage)
* = Species tested so far: for L. braziliensis = 2 isolates (1 SbV sensitive, 1 SbV resistant); for L. donovani = 4 isolates (2 SbV sensitive, 2 SbV resistant).

Conclusion: Based on the markers GCS, ODC & AQP1, we hypothesise that the mechanism of natural SbV resistance is multifactorial involving (i) a changed thiol metabolism leading to inhibition of the activation of SbV inside the amastigote and (ii) a decreased uptake of SbIII resulting from SbV activation in the macrophage. (compare to SbIII resistance model)

Immunology

Basic immunological parameters were determined and compared in patients with definite cure and patients with treatment failure as to verify the role of the host immune response in treatment outcome. Based on the mouse model for VL, the following hypothesis was formulated for association between cytokine profile and clinical outcome:

Cytokine	Th1-profile	Th2-profile
IFN-gamma	+ +	-
IL-10	-	+ +
IL-4	-	+
TNF-alpha	+	-
	Resistance to infection Response to treatment?	Susceptibility to infection Non-response to treatment?

A. VL patients, Nepal

Serum samples from 61 patients taken before treatment (Day 0) were used to determine the cytokines IFN-gamma, IL-10, IL-4 and TNF-alpha using ELISA at HUG. The obtained cytokine profiles were compared for the following groups:

1)	Patients with definite cure (n=47)	vs	Patients with treatment failure (n=10)
	Result: No significant difference in cytokine profile
2)	Patients with definite cure (n=47)	vs	Patients with failure to previous SbV treatment (n=4)
	Result: IL-10 significantly higher in group of patients with failure to previous SbV treatment

We conclude that:

• Persisting high level of IL-10 during therapy might play an important role in non-response to treatment.
• Levels of IL-10 measured during treatment might predict treatment outcome.
• Activation of the immune response appears similar before treatment in patients with subsequent cure and treatment failure.

B. CL patients, Peru

Skin biopsies taken before treatment (Day 0) and during treatment (Day 10) from 48 patients were used to determine the cytokines IFN-gamma, IL-4, IL-10, IL-13 and TNF-alpha using real-time quantitative PCR at HUG. The obtained cytokine profiles were compared before treatment and during treatment for the following 2 groups:

1)	Before treatment (D0):	Patients with definite cure (n=37)	vs	Patients with treatment failure (n=11)
	Result: No significant differences in cytokine profiles in 2 groups.
2)	Changes during treatment (D10-D0):	Patients with definite cure (n=21)	vs	Patients with treatment failure (n=7)
	Result: - All cytokines decrease during treatment in both groups. - IFN-gamma and IL-13 decrease significantly less in the group of patients with treatment failure compared to the group of patients with definite cure.

We conclude that:

• Persisting high level of IFN-gamma and IL-13 during therapy might play an important role in non-response to treatment.
• Levels of IFN-gamma and IL-13 measured during treatment might predict treatment outcome.
• Activation of the immune response appears similar before treatment in patients with subsequent cure and treatment failure.

Population Genetics

A. Nepal

Fingerprinting allows evaluation of the dynamics of the spreading of drug resistance in natural Leishmania populations. A population genetics study was done on Nepalese L. donovani isolates using kDNA/PCR-RFLP, which is currently the most discriminatory fingerprinting method. 25 isolates were analyzed, including isolates sampled from the same patients before and after treatment failure (interconnected isolates on tree) : (i) 1 month follow up (pairs BPK173/0 & BPK173/1; BPK275/0 & BPK275/1, (ii) 1 year follow up (pair BPK181/0 & BPK181/12).

Phenetic analysis: The UPGMA tree (see figure left) combines data from three restriction enzymes (HaeIII, HpalI and RsaI restriction of the 700 bp minicircle PCR-product). As a reproducibility control the strains BPK275/0, BPK085/0 and BPK091/0 were tested twice (inboxed). This allowed the definition of the background of variation associated with PCR, electrophoresis and pattern reading (right of dashed line). There are 3 major clusters present (shaded blue, red & green). The resistant (R) isolates are spread over all 3 clusters. It also clear that the follow-up pairs (interconnected isolates) are not identical, but branch together. Geographical mapping: All the isolates presented in the tree were mapped geographically. The isolates from cluster A en B were widespread throughout the whole endemic region, while the isolates of cluster C all originated from the two most Eastern districts. The resistant and sensitive isolates were spread over the West-East axis of the whole endemic region.

We conclude that:

• The population of resistant strains has a polyclonal structure, which could be explained by the following hypothesis:
  (i) resistance in the different clusters may have merged from independent events, (ii) the kDNA mini-circle populations of a Leishmania clone segregate randomly during mitosis in contrast to putative resistance markers, or (iii) although Leishmania essentially reproduces clonally, resistance may have been acquired via horizontal transfer.
• The geographical spreading of some kDNA genotypes, as well as resistant strains, supports the concept of a dynamic epidemiological situation in which human activities (e.g. seasonal migration) probably plays a major role.
• Isolates originating from the same patient before treatment and after treatment failure (follow-up pairs) branch together suggesting persistent infection.

B. Peru

No results available

Guidelines

Based on the results and conclusions within this project, the following guidelines were formulated and could contribute to policies to overcome drug resistance:

1) Rapid & improved diagnosis:

• Strengthing peripheral health systems to aid early diagnosis of cases, as late stage detection is a risk factor for treatment failure (see risk factors Nepal).
• Diagnosis allied to speices typing (as treatment outcome can depend on species, see NW).

2) Individual patient managment:

The identification of clinical risk factors of treatment failure is useful for individual patient’s management. For patients at higher risk of Sbv failure:

• Clear information should be given
• Follow-up after treatment should be intensified
• Increase dosage and/or duration of treatment might be considered?
• Other drug could be considered?

3) Access to full treatment for ALL:

• Treatment shoud be at free or at minimal cost for patient (helps to keep resistance at bay where not yet emerged).
• Improving treatment compliance and ensuring full dosage e.g. implementing Direct Observed Therapy treatment.

4) Drug policy & CONTROL including:

• Individual endimic regions should have proper recommendations for treatment regimens.
• Implementation, strict monitoring & control of treatment regimens to ensure proper distrubuiton and use of drug
• Implementation of a supranational integrated sentinel surveillance system as to monitor (i) treatment failure, (ii) emergence of drug resistance (using efficient tools based on molecular markers), (iii) distribution of species. Such a system could also provide good baseline data and materials (biopsies, isolates,...) for further research on emerging and spreading of drug resistance.

5) Promotion of the evaluation of combination therapy.

Combination therapy using the currently available drugs would minimise the risk for emergence of resistance against individual drugs.