Drug-resistance of Trypanosoma b. rhodesiense isolates from Tanzania

Transcription

1 Tropical Medicine and International Health doi: /j x volume 11 no 2 pp february 2006 Drug-resistance of Trypanosoma b. rhodesiense isolates from Tanzania S. N. Kibona 1,2, L. Matemba 1, J. S Kaboya 1 and G. W. Lubega 2 1 National Institute for Medical Research, Tabora, Tanzania 2 Faculty of Veterinary Medicine, Makerere University, Kampala, Uganda Summary objective To determine the drug resistance of Trypanosoma brucei rhodesiense strains isolated from sleeping sickness patients in Tanzania. method We first screened 35 T. b. rhodesiense strains in the mouse model, for sensitivity to melarsoprol (1.8, 3.6 and 7.2 mg/kg), diminazene aceturate (3.5, 7 and 14 mg/kg), suramin (5, 10 and 20 mg/ kg) and isometamidium (0.1, 1.0 and 2 mg/kg). A 13 isolates suspected to be resistant were selected for further testing in vitro and in vivo. From the in vitro testing, IC 50 values were determined by short-term viability assay, and MIC values were calculated by long-term viability assay. For in vivo testing, doses higher than those in the initial screening test were used. results Two T. b rhodesiense stocks expressed resistance in vivo to melarsoprol at 5 mg/kg and at 10 mg/kg. These strains had high IC 50 and MIC values consistent with those of the melarsoprol-resistant reference strain. Another isolate relapsed after treatment with 5 mg/kg of melarsoprol although it did not appear resistant in vitro. One isolate was resistant to diminazene at 14 mg/kg and another was resistant at both 14 and 28 mg/kg of diminazene. These two isolates had high IC 50 values consistent with the diminazene-resistant reference strain. Two isolates relapsed at a dose of 5 mg/kg of suramin, although no isolate appeared resistant in the in vitro tests. Two isolates were resistant to isometamidium at 1.0 mg/kg and had higher IC 50 values. Two isolates were cross-resistant to melarsoprol and diminazene and one isolate was cross-resistant to suramin and isometamidium. conclusion The reduced susceptibility of T. b. rhodesiense isolates to these drugs strongly indicates that drug resistance may be emerging in north western Tanzania. keywords Trypanosoma brucei rhodesiense, treatment, melarsoprol, diminazene aceturate, suramin, isometamidium, relapse, drug resistance Introduction African trypanosomes cause diseases in humans and domestic animals. Trypanosoma brucei gambiense and Trypanosoma brucei rhodesiense, two subspecies of vector-borne hemoflagellate protozoans, cause Human African Trypanosomiasis (HAT) or sleeping sickness. In animals, trypanosomiasis is caused by various Trypanosoma species, the major ones being Trypanosoma vivax, Trypanosoma congolense and Trypanosoma evansi. Human African Trypanosomiasis constitutes a serious public health threat in Africa, particularly in east and central Africa, where approximately 60 million people are at risk of contracting the disease. The new cases reported to the World Health Organization (WHO) in 1999 do not accurately reflect the real situation they merely underscore the inadequacy of the current campaign against this disease. It is estimated that persons are infected with HAT. Both Gambiense and Rhodesiense forms of HAT or sleeping sickness have been reported from Tanzania. Gambiense sleeping sickness was first recorded around Lake Tanganyika, where an active focus persisted until 1958 (Fairbairn 1948; Ormerod 1961). Rhodesiense sleeping sickness, which appeared in the 1920s and 1930s, was endemic in eight regions of Tanzania: Arusha, Kigoma, Lindi, Mbeya, Kagera, Rukwa, Ruvuma and Tabora. The disease tends to be concentrated in Kigoma (Figure 1). In the past 7 years only five regions have reported sleeping sickness cases; a large number originated in Kigoma (Kibondo and Kasulu districts). Over the past 4 years, the number of cases in Tabora (Urambo district) has been rising. Currently between 4 and 5 million people are thought to be at risk of infection in Tanzania, but only 1% of these are under regular surveillance. In the past 30 years, 144 ª 2006 Blackwell Publishing Ltd

2 RWANDA UGANDA KENYA BURUNDI L.Eyasi Kibondo Tabora L.Manyara Kasulu Kigoma Tabora L.Tanganyika ZAIRE Mpanda Rukna L.Rukna ZAMBIA Present foci of Rhodesian sleeping sickness Areas sampled L.Victoria Old focus of Gambian sleeping sickness (ceased 1954) MALAWl Dar Es Salaam TANZANIA MOZAMBIQUE Figure 1 Map of Tanzania showing the main foci and sampled areas of rhodesiense sleeping sickness. the number of new cases reported annually has risen above 500 (Annual Sleeping Sickness Reports for Tanzania ), although this is likely to be an underestimate. In Tanzania, as in many other countries, the control of HAT is based primarily on chemotherapy with only a few active drugs available for treatment, a situation that has remained virtually unchanged for more than 40 years. All these drugs have adverse side effects, are expensive or sometimes fail to cure. Recent progress in HAT research suggests that a vaccine against the disease is far from being successful (Atouguia & Costa 1999; Van Nieuwenhove 1999). Early-stage HAT is successfully treated with pentamidine and suramin; but treatment of late-stage HAT, where the trypanosomes have invaded the cerebrospinal fluid (CSF), depends exclusively on the arsenical compound melarsoprol. The one new drug marketed in the past 40 years, Eflornithine (DFMO, Ornidyl Ò ), is only effective against T. b. gambiense and is very expensive. Chemotherapeutic intervention is facing the problem of emergent resistance, as current drugs have been in continuous use for decades. Consequently, resistant T. brucei species have been reported from various parts of Africa (Matovu et al. 2001). Much less information is available regarding human trypanosomiasis. Relapses after treatment with the first-stage drugs pentamidine and suramin are rare; some early second-stage infections are misdiagnosed as resistance. Drug treatment failures at the second stage are L. Malawi N INDIAN OCEAN mainly reported for melarsoprol. T. b. rhodesiense is innately refractory to eflornithine while T. b. gambiense is susceptible based on the selective toxicity as a result of different turnover rates for ODC in the two sub-species (Iten et al. 1997). Robertson and Hawking (1960) observed T. b. rhodesiense patients in Uganda who relapsed after two or more courses of melarsoprol; Ruppol and Burke (1977) reported up to 40% melarsoprol relapses in T. b. gambiense patients in Zaire; and Perez Martin et al. (1991) describe melarsoprol-resistant T. b. rhodesiense patients in Mozambique. The overall melarsoprol relapse rate over the past 50 years has been 5% 8% (Pepin et al. 1994), but in recent years this has increased alarmingly. In epidemic T. b. gambiense areas, in Uganda for example, it is 30% (Legros et al. 1999). Regarding animal trypanosomiasis, there are persistent reports of resistant trypanosomes from all regions of Africa and to all drugs in use (Matovu et al. 2001). Resistance to the major drugs (diminazene, isometamidium, homidium, suramin and quinapyramine) has been reported in T. congolense, T. vivax and T. evansi (Anene et al. 2001); even multiple-drug resistance in T. congolense (Afewerk et al. 2000). Between 2000 and 2002 we isolated T. b. rhodesiense stocks from sleeping sickness patients to determine whether trypanosome strains with reduced drug sensitivity occurred in Kasulu, Kibondo and Urambo districts in Tanzania, where the number of relapses is increasing. Sensitivity was tested both in vitro and in vivo. Materials and methods Study population and area The study proposal was approved by the National Institute for Medical Research and National Ethics Committee of the Ministry of Health in Tanzania. All participating patients or their legal guardians gave free and informed advance consent. The study area included Kibondo and Kasulu districts in Kigoma region, north-western Tanzania and Urambo district in Tabora region, central Tanzania. These three districs were selected due to the high number of reported HAT cases in the past 5 years. Patients at dispensaries, health centres and hospitals and during active sleeping sickness surveys at home were selected regardless of sex, age or treatment history. Patients were separated into two groups: early stage without central nervous system (CNS) involvement, and late stage with CNS involvement. Persons suspected to be infected with trypanosomes based on clinical observations were examined by blood ª 2006 Blackwell Publishing Ltd 145

3 smear and Haematocrit Centrifugation Technique to confirm the infection. All confirmed cases underwent lumbar puncture and CSF examination to determine the stage of illness, lymphocyte counts and protein levels. If the number of lymphocytes was > /l and the protein level was >25 mg per 100 ml, the patient was considered to be latestage. We collected 2 ml bloodsamples from confirmed sleeping sickness cases by venipuncture and then referred the patients to hospital for treatment according to the stage of their illness. Each blood sample was divided into two and cryopreserved in liquid nitrogen. One portion was used for propagation of the isolates and the other was kept for future reference. Isolates were coded Tabora Medical Research Station (TMRS) followed by the village or location number and the identity number of the patient in brackets. Isolation of trypanosomes For propagation of the isolates, a portion of blood was thawed in the laboratory at Tabora, Tanzania and inoculated intraperitoneally into Albino Swiss mice. The infected mice were monitored for parasitaemia by examining tail blood. Highly parasitaemic mice were anaesthetized and heart blood collected, with heparin as an anticoagulant. The blood was then cryopreserved in 1:1 volume with phosphate saline glucose (PSG) plus 20% glycerol until required for drug sensitivity tests and/or molecular characterization. A total of 35 isolates of T. b. rhodesiense from sleeping sickness patients were collected and expanded; the stabilates were transported to the of Faculty of Veterinary Medicine, Makerere University, Kampala, Uganda for drug sensitivity assays and molecular characterization. In Kampala, the stabilates from Tanzania were recovered and injected intraperitoneally into mice; parasitaemia was monitored by tail blood. On days 5 9 of infection the mice were anaesthetized with trichloromethane and exsanguinated by cardiac puncture. The blood was collected in heparinised tubes and suspended in PSG buffer. The number of trypanosomes per ml in the pooled blood (from a group of mice infected with same isolate) was quantified with a Neubauer (improved) hemocytometre; the resultant trypanosome population was used to infect another group of mice intraperitoneally (i.p.) at a dose of 10 7 trypanosomes per mouse. These mice came from the breeding colony at the Faculty of Veterinary Medicine in Kampala and were 5 8-week-old weighing g. Reference resistant isolates, drugs and preliminary sensitivity assays The T. b. rhodesiense resistant isolates coded KETRI 1989 and KETRI 2653 were obtained from the Kenya Trypanosomiasis Research Institute, Kikuyu, Kenya and pre-tested in vivo to confirm their status. KETRI 1989 was confirmed resistant to suramin and diminazene and KETRI 2653 as resistant to melarsoprol. We used the following drugs in this study: melarsoprol (Mel B or Arsobal, Aventis, France) and suramin (Germanin, Bayer, Germany), commonly used for treating human trypanosomiasis; and diminazene aceturate (Berenil, Hoechst, Germany) and isometamidium chloride (Samorin or Trypamidium, Rhône Merieux, USA), which are commonly used to treat animal trypanosomiasis in Tanzania. Isolates infected in mice were first screened by treatment with doses slightly lower or higher than normal for infected natural hosts. Those isolates, which apparently relapsed upon treatment with low doses were then subjected to in vitro drug sensitivity screening tests (short- and longterm viability assays) followed by in vivo drug sensitivity confirmatory tests using high doses. All 35 isolates were screened in vivo as follows: Three groups of five mice each were treated with different low doses of the appropriate trypanocidal drug but the fourth group remained untreated as a control. This procedure was used for all four drugs. Drugs freshly prepared in sterile distilled water were administered intraperitoneally (i.p.). The different doses of drugs used were adjusted such that the required amounts of drug were contained in a volume of 0.2 ml of sterile distilled water. Controls received the same volume without drug. All mice were treated 36 h after infection, which is early enough to minimize the chances of parasites crossing the blood-brain barrier. We used the following dosages for the initial drug sensitivity screening: melarsoprol 1.8, 3.6 and 7.2 mg/ kg; diminazene 3.5, 7 and 14 mg/kg; suramin 5, 10 and 20 mg/kg; isometamidium 0.1, 1.0 and 2 mg/kg. After treatment, progress of parasitaemia in all mice was monitored for 60 days by examining wet films of tail blood. If more than two mice remained parasitaemic after 60 days at any dose, the isolate was suspected resistant and selected for further drug resistance tests. Thus, 13 isolates were selected for more rigorous tests in vitro and in vivo. In vitro drug sensitivity assays Trypanosome cultures were started in a feeder layer of embryo fibroblast cells. For this purpose mouse embryo fibroblast cells were isolated following a slight modification of Brun s procedure (Brun et al. 1981). Briefly, a uterus was removed aseptically from a pregnant mouse 15 days after mating and washed with Earle s Balanced Salt Solution (EBSS). The embryos were removed from the 146 ª 2006 Blackwell Publishing Ltd

4 uterus, heads and livers were discarded, and the remaining parts washed three times with EBSS before mincing them with a sterile blade. The resulting small pieces were transferred into 100 ml bottles containing 20 ml of 0.25% trypsin in EBSS (containing 25 mm Hepes at ph 7.5). The suspension was stirred at 37 C for 10 min and left to settle for 10 min. The supernatant was them discarded, fresh trypsin solution was added and the suspension stirred again for 10 min and left to settle for 10 min. This procedure was repeated once more. The cells were then pelleted by centrifugation for 20 min at 200 g at 4 C. Cells were then resuspended in minimum essential medium (MEM), counted and approximately cells transferred to T-25 flasks containing 5 ml MEM, 15% inactivated fetal bovine serum (FBS) and 20 lg/ml gentamycin. The cells were then incubated at 37 C, the medium was changed after 24 h. After 3 days the cells were confluent and trypsinized to prepare feeder layer stabilates, which were cryopreserved until needed. To initiate the culture, feeder layer cell stabilates were thawed and resuspended in MEM in a culture flask and incubated at 37 C in5%co 2 atmosphere. The optimum age for the feeder layer to be inoculated with trypanosomes was h, when the cells had just attained confluency. Trypanosomes from the mice infected with the 13 isolates suspected to be resistant (from the initial screening test) were harvested aseptically by cardiac puncture and transferred to culture flasks containing feeder layer cells (mouse embryo fibroblast cells) in MEM (MEM GIBCO) with Earle s salts, l-glutamine and non-essential amino acids and supplemented with 25 mm HEPES, 1 g/l additional glucose, 2.2 g/l sodium bicarbonate, and 15% heatinactivated horse serum from GIBCO (Kaminsky & Zweygarth 1989). The medium was further supplemented according to Baltz et al. (1985) with 0.2 mm 2-mercaptoethanol, 2 mm sodium pyruvate and 0.1 mm hypoxanthine. The cultures were incubated at 37 C in a humid atmosphere containing 5% CO 2, and monitored daily with appropriate medium change so as to maintain the trypanosomes in the growth phase. Stocks which grew exponentially for >14 days were considered culture adapted. After this period trypanosomes were transferred into 24-well culture plates (Costar) with the usual culture medium minus feeder layer cells. After 7 days the trypanosomes were completely adapted to culture system without feeder layer cells and were showing constant growth throughout the experiment. Three of the isolates that were cured by all drugs during initial screening tests were also adapted to in vitro culture and used as sensitive reference controls. Bloodstream forms of the culture-adapted T. b. rhodesiense were seeded at 10 5 /100 ll density and incubated in 96-well microtiter plates (Costar) as described above with serial dilutions (seven twofold dilutions) ranging from highest to lowest (100 to 1.6 ng/ml for Melarsoprol, 1000 to 15.3 ng/ml for Diminazene, 1000 to 15.3 ng for suramin and 100 to 1.6 ng/ml for isometamidium). Parasites were then grown for 24 h in the presence of these test drugs. In these experiments, which were performed in duplicate and repeated three times, the densities of untreated control cultures increased from 10 5 to cells/ml. After determination of cell densities (cell counts per well) at each drug concentration with a hemocytometer, drug sensitivity was expressed as a percentage of the growth of the control cells and the drug concentration, which caused 50% inhibition (IC 50 ) established by interpolation (Huber & Koella 1993) on the graph of percentage growth inhibition vs. drug concentration. A further statistical test was carried out using univariate analysis (Prism GRAPHPAD). A long-term viability assay according to Brun et al. (2001) was used as a second method to assess drug sensitivity in vitro. Briefly, a suspension of culture adapted feeder layer cells from mouse embryo was added to each well of 24 well tissue culture plate (Costar) and incubated at 37 C in5%co 2 atmosphere until ready for addition of trypanosome suspension. Bloodstream forms were then seeded at 10 5 /ml onto feeder layers with serial dilutions of melarsoprol ranging from 72 to ng/ml, of diminazene ng/ml, of suramin ng/ml, and of isometamidium ng/ml (seven two-fold dilutions). The cultures were incubated at 37 C in a humid atmosphere containing 5% CO 2. Every other day the cultures were monitored and the medium replaced with fresh medium containing the appropriate drug concentration. After 10 days we microscopically determined the minimum inhibitory concentration (MIC), the lowest drug concentration at which no trypanosomes with normal morphology or motility could be observed. In vivo drug sensitivity assays The 13 isolates suspected to be drug resistant by initial screening tests were each inoculated into 20 mice randomized into four groups of five mice each. Thirty-six hours later, three of the groups were treated with different high drug doses including melarsoprol (5, 10 and 20 mg/kg), diminazene (14, 28 and 56 mg/kg), suramin (5, 10 and 20 mg/kg), and isometamidium (1, 2 and 4 mg/kg). The fourth group (control) was treated with vehicle containing no drug. Thereafter, the mice were monitored for parasitaemia three times per week for 60 days. The dose (CD 90 ) that completely cured 90% of the animals was determined. ª 2006 Blackwell Publishing Ltd 147

5 Results Initial drug resistance screening tests Isolates treated with 1.8 mg/kg of melarsoprol and 0.1 mg/kg of isometamidium, which are half or less than half of the normal doses used in the natural host, were not considered for further analysis. Other isolates, after treatment with normal or double doses used in the natural host, in vivo, which relapsed in at least three of five mice, were suspected resistant (Table 1) and selected for in vitro and in vivo drug sensitivity testing using higher doses. Thus, 13 isolates were picked for further resistance tests against melarsoprol, diminazene, suramin and isometamidium. Some isolates were suspected resistant to more than one drug; and most of these isolates came from patients with a history of relapse, although two isolates came from new cases. Nevertheless, the cure rate was high: 82.8% 91.4% of the isolates were susceptible. Melarsoprol cured 77% at a dose of 3.6 mg/kg (maximum dose for sleeping sickness patients). All untreated mice used as control died between days 7 and 12. In vitro resistance tests Thirteen isolates of T. b. rhodesiense suspected to be resistant by the initial in vivo screening test were first evaluated in vitro for resistance to melarsoprol, diminazene, suramin and isometamidium (Table 2). Two isolates, KETRI 1989, known to be resistant to diminazene and suramin, and KETRI 2653, known to be resistant to melarsoprol, were used as resistant reference controls in the experiment. Three isolates TMRS 15(6), TMRS 3(7) and TMRS 10(3), which were sensitive to all drugs by the initial in vivo screening test, were used as sensitive reference control. Using melarsoprol in the short term viability assay, test isolates TMRS 11(2) and TMRS 12(4) had IC 50 values between 10.8 and 15.8 ng/ml which are higher (approximately seven-fold) than the values ng/ml for the reference sensitive isolates TMRS 15(6), TMRS 3(7) and TMRS 10(3). Univariate analysis showed no significant difference (P > 0.05) in sensitivity between these isolates and the resistant reference isolate KETRI 2653 which had an IC 50 of 18.8 ng/ml. The long-term viability assay (MIC values) also exhibited a similar pattern. Therefore the in vitro test appeared to confirm isolates TMRS 11(2) and TMRS 12(4) as resistant to melarsoprol. TMRS 12(2) presented an IC 50 value of 14.1 ng/ml, which placed it among the melarsoprol resistant isolates, but this resistance was not repeated in the long viability assay (MIC value of 5.6 ng/ml, which is far below the 10 ng/ml of the resistant control isolate). Therefore, TMRS 12(2) could not be classified as resistant to melarsoprol at this stage. For diminazene the same two isolates TMRS 11(2) and TMRS 12(4) had significantly (P < 0.01) higher IC 50 values (31.6 and 39.8 ng/ml, respectively) than the sensitive isolates TMRS 15(6), TMRS 3(7) and TMRS 10(3) (IC 50 values ng/ml). The IC 50 values of these isolates were not significantly different (P > 0.05) from that of the control (KETRI 1989) resistant isolate. Thus, these isolates are apparently resistant to both melarsoprol and diminazene. But the resistance of these isolates could not be repeated in the long-term viability assay (see the MIC values), which left their status unconfirmed. IC 50 values for all isolates tested with suramin were significantly (P < 0.05) less than that of the resistant control isolate (KETRI 1989). Therefore, there was no isolate, which showed reduced susceptibility to suramin by in vitro drug resistance tests. Using isometamidium in the short term viability assay, three isolates TMRS 1(13), TMRS 2 (11) and TMRS 10 (6), had significantly (P < 0.01) higher IC 50 values (8.8, 8.9 and 8.4 ng/ml, respectively) than the IC 50 values ( ng/ml) of reference sensitive isolates. Again, the resistance of these isolates was not reflected in the longterm viability assay (MIC values). Probably, the short term in vitro test indicates the reduced susceptibility of these three isolates to isometamidium. In vivo resistance tests After the in vitro drug resistance test the 13 isolates were tested in vivo using high doses in new groups of mice. The sensitive reference isolates TMRS 15(6), TMRS 3(7) and TMRS 10(3) and the resistant reference isolates, KETRI 1989 and KETRI 2653 were included. All treated mice were aparasitaemic 12 days post-treatment but some treated mice relapsed (Tables 3 and 4). An isolate was suspected to be resistant if more than two mice relapsed at any dose (Table 3). Melarsoprol cured all mice including the controls at a dosage of 20 mg/kg, which is approximately the highest tolerated dose. Two of the treated isolates (TMRS 11(2), TMRS 3(11)) could not be cured by the medium single dose of 5 mg/kg. TMRS 12(4) also could not be cured at 5 mg/kg and relapsed on day 38 post-treatment at the higher dose of 10 mg/kg. The positive control (KETRI 2653) also relapsed at medium and maximum doses of 5 and 10 mg/kg. Isolate TMRS 12(4), which had apparent high IC 50 values in vitro, was also resistant to high dose in vivo. In contrast, isolate TMRS 3(11), which did not present as resistant in vitro, was resistant at 5 mg/kg. Therefore, three isolates were confirmed to be resistant to melarsoprol: TMRS 11(2) and TMRS 3(11) at 5 mg and TMRS 12(4) at both 5 and 10 mg/kg doses. 148 ª 2006 Blackwell Publishing Ltd

6 Table 1 Initial in vivo drug resistance screening tests No. of mice still infected after 60 days post-treatment with the following drug dosages Isolate Melarsoprol (M) Diminazene (D) Suramin (S) Isometamidium (I) 1.8à 3.6à 7.2à 3.5à 7à 14à 5à 10à 20à 0.1à 1à 2à Apparent sensitivity status (drug) TMRS 10(7) R (D) TMRS 10(6) R (I) TMRS 11(2) R (M andd) TMRS 12(4) R (M andd) TMRS 13(2) R (M) TMRS 15(6) S TMRS 15(10) R (M) TMRS 7(2) TMRS 3(2) TMRS 3(3) TMRS 3(4) TMRS 3(11) R (M) TMRS 3(6) TMRS 4(1) R (M ands) TMRS 2(1) TMRS 2(2) TMRS 1(6) TMRS 1(13) R (M) TMRS 5(1) TMRS 3(7) S TMRS 9(5) TMRS 2(11) R (M,S and D and I) TMRS 2(12) R (M andd) TMRS 8(13) TMRS 8(14) R (D) TMRS 3(9) TMRS 2(7) TMRS 10(3) S TMRS 10(4) TMRS 11(1) TMRS 11(3) TMRS 11(4) TMRS 11(5) TMRS 12(2) R (M andd) TMRS 12(1) No. of isolates cured Percentage of isolates cured For each dose five infected mice were treated (i.p.) with M, melarsoprol; D, diminazene; S, suramin I, isometamidium and observed for any relapse for 60 days. Isolates marked with this symbol in thee first column were from relapsed sleeping sickness patients, the rest were new cases. à Dosage in mg/kg body weight (bwt). Apparent sensitivity status is either resistant (R) or not. For those, which are apparent resistant the first letter of drug concerned is or are given in brackets. If more than two infected mice remained parasitaemic at any dose after 60 days post-treatment, the isolate was considered resistant (R) and selected for further testing. Conversely if two or less infected mice remained parasitaemic at a given dose, the isolate was considered sensitive. Isolates which were sensitive to all drugs by the initial in vivo screening test and selected as sensitive reference control. All isolates were cured by diminazene at 56 mg/kg. Isolate TMRS 11(2) could not be cured after treatment at the dose of 14 mg/kg and TMRS 12(4) could not be cured at 14 mg/kg and yet relapsed on day 48 post-treatment at a high dose of 28 mg/kg. The positive control isolate (KETRI 1989) could not be cured at 14 mg/kg and also relapsed ª 2006 Blackwell Publishing Ltd 149

7 Table 2 In vitro drug resistance screening test for 13 isolates that appeared resistant in the in vivo initial screening test Results (ng/ml) Melarsoprol Diminazene Suramin Isometamidium Trypanosome isolates IC 50 MICà IC 50 MICà IC 50 MICà IC 50 MICà TMRS 10(7) ND TMRS 11(2) ND TMRS 12(4) ND TMRS 13(2) ND TMRS 3(11) ND TMRS 4(1) ND TMRS 1(13) ND TMRS 2(11) ND TMRS 2(12) ND TMRS 12(2) ND TMRS 15(10) ND TMRS 10(6) ND TMRS 8(14) ND TMRS 15(6) S ND TMRS 3(7) S ND TMRS 10(3) S ND 2, KETRI 1989 R $ ND KETRI 2653 R $ ND Inhibitory concentration, IC 50 values were determined by the short-term viability assay (24 h). à Minimum inhibitory concentration, MIC values were determined by the long-term viability assay. Isolates, which were susceptible (s) to all drugs upon testing as outlined in Table 1 and used as reference in the in vitro screening test. $ KETRI 1989: known to be resistant (R) to diminazene and suramin. $ KETRI 2653: known to be resistant (R) to melarsoprol. ND, not done. on day 41 post-treatment at a dose of 28 mg/kg. The two isolates, which had apparently high IC 50 values in vitro were also resistant at high doses in the in vivo screening test. Therefore, they were confirmed to be resistant to diminazene: TMRS 11(2) at 14 mg/kg and TMRS 12(4) at both 14 and 28 mg/kg doses. All isolates were sensitive to suramin at the 10 mg/kg dosage except the control isolate KETRI 1989, which is known to be resistant to suramin. However, TMRS 4(1) and TMRS 2(11) relapsed at dose of 5 mg/kg, although none had shown resistance in vitro. Both of these isolates were considered to be resistant to suramin at a dose of 5 mg/kg. Isometamidium administered i.p. at dosage 1.0 mg/kg failed to cure the critical number (3) of mice infected with isolates TMRS 1(13), TMRS 2(11) and TMRS 10(6); moreover these isolates tested resistant in vitro. Hence, TMRS 1(13), TMRS 10(6) and TMRS 2(11) were confirmed to be resistant to isometamidium at 1.0 mg/kg dosage. TMRS 11(2) and TMRS 12(4) appeared to be resistant to both melarsoprol (5 and 10 mg/kg) and diminazene (14 and 28 mg/kg). TMRS 2(11) appeared to be resistant to suramin (5 mg/kg) and isometamidium (1 mg/kg). There was a clear relationship between the time of relapse and the dose of the drugs used: All untreated (controls) infected mice died between days 7 and 12. Where resistance was apparent at more than one dose, relapse occurred sooner at lower doses than at higher doses (Table 4) Discussion Trypanosoma b. rhodesiense isolates (n ¼ 35) collected from patients in Tanzania between 2000 and 2002 were investigated for sensitivity to melarsoprol, diminazene, suramin and isometamidium. The aim was to find out if these trypanosomes would exhibit reduced drug sensitivity or if other reasons could have caused the relapses in the patients. Our results probably indicate the presence of drug resistant T. b. rhodesiense in Tanzania. We used in vivo tests involving treating mice with doses higher than the equivalent used to treat trypanosomiasis in the field, the end point indicator of resistance. On this basis, three isolates were resistant to melarsoprol, two to diminazene, two to suramin and three to isometamidium. Although the 150 ª 2006 Blackwell Publishing Ltd

8 Table 3 In vivo drug resistance screening test using high doses (a) Number of mice tested/relapsed upon treatment with the following doses Melarsoprol (mg/kg) Diminazene (mg/kg) Suramin (mg/kg) Isometamidium (mg/kg) Trypanosome isolates 5à 10à 20à 14à 28à 56à 5à 10à 20à 1à 2à 4à TMRS 10 (7) 5/0 5/0 5/0 5/2 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 TMRS 11 (2) 5/3 5/1 5/0 5/4 5/1 5/0 5/0 5/0 5/0 5/1 5/0 5/0 TMRS 12 (4) 5/4 5/3 5/0 5/0 5/3 5/0 5/0 5/0 5/0 5/0 5/1 5/0 TMRS 13 (2) 5/0 5/0 5/0 5/2 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 TMRS 3 (11) 5/4 5/2 5/0 5/0 5/0 5/0 4/0 5/0 4/0 5/0 5/0 5/0 TMRS 4 (1) C 5/0 5/0 5/0 5/0 4/0 5/4 5/0 5/0 5/1 5/0 4/0 TMRS 1 (13) 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/3 5/0 5/0 TMRS 2 (11) 5/2 5/0 4/0 5/0 5/0 4/0 5/4 5/0 5/0 5/5 5/0 4/0 TMRS 2 (12) 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 TMRS 12 (2) 5/0 5/0 5/0 5/1 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 TMRS 15(10) 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 TMRS 10 (6) 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/5 5/0 5/0 TMRS 8 (14) 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 4/0 5/0 5/0 5/0 TMRS 15(6) S 5/0 5/0 4/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 TMRS 3(7) S 5/0 5/0 5/0 5/0 5/0 5/0 5/0 4/0 5/0 5/1 5/0 5/0 TMRS 10(3) S 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 5/0 KETRI 1989 R $ 5/1 5/0 5/0 5/5 5/4 5/0 5/4 5/3 5/0 5/0 5/0 5/0 KETRI 2653 R $ 5/4 5/4 5/0 5/0 5/1 5/0 5/0 5/0 5/0 5/0 5/0 5/0 Response of the 13 isolates in the mouse model upon treatment (i.p.) with melarsoprol, Diminazene, suramin and isometamidium; the data indicate the number treated vs. relapsed after 60 days. Isolates TMRS 15(2) S, TMRS 3(7) S and TMRS 10(3) S were used as drug sensitive reference isolates. KETRI (1989), which is known to be resistant to suramin and diminazene and KETRI 2653 known to be resistant to melarsoprol were used as positive control. (b) Mean relapse interval (days) post-treatment for mice which survived beyond 60 days. The number mice tested were five and the value after/indicates the number of mice which relapsed. à Dosage in mg/kg body weight. See Table 2. $ See Table 2. Indicates the isolate was resistant at the given dosage of the given drug. If more than 2 of 5 mice tested relapsed at the given dose the isolate was considered resistant. Table 4 Mean relapse interval (days) post-treatment for mice which survived beyond 60 days Melarsoprol Diminazene Suramin Samorin Isolates 5 mg/kg 10 mg/kg 14 mg/kg 28 mg/kg 5 mg/kg 10 mg/kg 1 mg/kg 2 mg/kg TMRS 11 (2) 24 ± ± 3.25 TMRS 12 (4) 16 ± ± ± ± 5.75 TMRS 3 (11) 43 ± 4.67 TMRS 4 (1) 21 ± 2.45 TMRS 2 (11) 15 ± ± 1.67 TMRS 1 (13) 34 ± 2.05 TMRS 10 (6) 19 ± 3.45 KETRI ± ± ± ± 5.10 KETRI ± ± 1.45 Values given are mean ± SE. in vitro culture system provides a convenient way of determining the drug resistance, whereby large number of isolates can be screened under the same conditions, it presents a few obstacles: The first is to isolate trypanosomes constituting a true representative of the population of parasites in humans or animals with the infection. The ª 2006 Blackwell Publishing Ltd 151

9 second obstacle is that trypanosomes which establish in culture may not be representative of the parasites in a sample. There is therefore, inevitably, selection for some trypanosomes which adapt more readily to the culture system, with the faster-growing trypanosomes being more abundant. The in vitro technique will be much improved when bloodstream trypanosomes can be adapted rapidly to culture, and exposed to trypanocides (and possibly their metabolites) for a minimum time to give an easily measurable end-point. The result obtained may then be correlated with results of field treatment (Geerts & Holmes 1998). The amastigote-macrophage assay of Leishmania sp. is currently the only model able to correlate clinical response to the in vitro drug sensitivity of the isolate, as demonstrated in relation to pentavalent antimonials (Croft 2004). The results of the in vitro tests indicate that, compared with laboratory reference stocks, the test trypanosomes are more, or less, sensitive to a particular concentration of a trypanocide. To interpret these results in the light of clinical findings is not straightforward. It is important to be able to state that a trypanocidal effect exerted in vitro at a concentration of, say, 100 ng/ml melarsoprol is equivalent to a curative dose of, say, 3.6 mg/kg bodyweight. Too few data are currently available for this to be done. In addition, the test does not reflect the role of an unimpaired immune system of the natural host. This is very important because some drugs are cytostatic and their treatment success depends on the immune system (Dee Gee et al. 1983). But even in the absence of immune factor, the in vitro drug sensitivity test can then give a clue of differences in drug sensitivities between different isolates. For better results, in vitro drug sensitivity test has to be confirmed by in vivo (Enyaru et al. 1998). Therefore, our isolates were probably resistant. Nevertheless, sensitivity from the mouse model does not always reflect that in human patients (Sones et al. 1988). Due to their higher metabolic rate, mice can alter or remove the drug faster than the natural host or vice versa. Hence, in mice the drug may not exert the same trypanocidal effect at the same dose as in humans (Zweygarth & Röttcher 1989). However, the fact that these isolates were obtained from patients who failed to be cured by the drugs tested in the mouse model leaves us with the conclusion that the isolates we found resistant by the mouse model are indeed resistant. Despite some discrepancies the in vitro test could sometimes corroborate the in vivo test in determining the drug resistance. For example the in vitro and in vivo results support the possibility of resistance of TMRS 12(4) to melarsoprol, since the high dose of 10 mg/kg could not cure it. The maximum dose for humans is 3.6 mg/kg, and it had an IC 50 value of 15.8 ng/ml. The relapse of TMRS 11(2) after treatment with 5 mg/kg melarsoprol may also indicate reduced susceptibility considering that the other isolates were completely cured. This is corroborated by its in vitro IC 50 and MIC values, which are higher than those of the melarsoprol-resistant control. TMRS 3(11) had low IC 50 and MIC values, but unexpectedly relapsed when treated with 5 mg/kg melarsoprol. Thus, the resistance of this isolate requires further investigation. TMRS 12(4), which turned out resistant by all tests used, was isolated from a female patient who was reporting for the third time to Kasulu District Hospital with late-stage sleeping sickness. Isolate TMRS 11(2) was isolated from a late stage patient reported as a first-time relapse case. The trypanosomes from TMRS 12(4) and TMRS 11(2) were confirmed to be melarsoprol-resistant, thus indicating relapse caused by trypanosomes selected against the therapeutic drug level during previous treatment rather than re-infection. In the mouse model experiment the possibility that relapse might be caused by re-invasion of the blood stream by trypanosomes from the CNS is also ruled out because relapse of this nature cannot occur in mice treated earlier than 14 days post-infection (Enyaru et al. 1998). In this experiment mice were treated 36 h post-infection, so trypanosomes had not yet crossed the blood brain barrier. The relapses after treating infected mice with 14 and 28 mg/kg diminazene also indicate development of resistance to this drug, since the recommended doses for cattle are mg/kg (Mbwambo et al. 1988; Kaminsky et al. 1989; Enyaru et al. 1998). Diminazene resistance was also indicated in vitro, where the IC 50 values were higher than that of the diminazene-resistant control. Therefore, TMRS 11(2) and TMRS 12(4) were resistant to diminazene. The isolates which relapsed after treatment with 5 mg/kg suramin may not be resistant in humans, as the maximum recommended dose in man is 20 mg/kg. Also, suramin treatment failure has never been observed in the areas where these isolates were obtained (District Medical Officers, personal communication). The relapsed isolates (TMRS 4(1) and TMRS 2(11)) also had low IC 50 and MIC values compared to the suramin-resistant control isolate (KETRI 1989). It cannot be said with certainty whether suramin resistance is emerging in the area. The mice infected with isolates TMRS 2(11), TMRS 1(13) and TMRS 10(6) relapsed after treatment with isometamidium at 1 mg/kg, meaning that it is a partially effective dose, which gave temporary clearance of the parasites. Considering that the recommended dose for cattle is mg/kg, the data indicate the reduced susceptibility of these isolates to isometamidium. Although Nyeko et al. (1988) did not consider trypanosomes 152 ª 2006 Blackwell Publishing Ltd

10 resistant when relapsed at 1 mg/kg isometamidium, the high IC 50 values observed in our study do suggest resistance. Our data curiously indicate cross-resistance between apparently unrelated drugs. For example isolates TMRS 11(2) and TMRS 12(4), which relapsed after 5 and 10 mg/kg melarsoprol, also relapsed after at 14 and 28 mg/kg diminazene. Isolate TMRS 2(11) was resistant to both suramin (5 mg/kg) and isometamidium (1 mg/kg). Although the drugs used to treat trypanosomiasis in humans may be chemically different from those for animals, cross-resistance may develop. Cross-resistance between melamine-based arsenicals and diamidines has been repeatedly observed in trypanosomes with laboratoryinduced drug resistance (Rollo & Williamson 1951). Indeed cross-resistance between diminazene (for animal trypanosomiasis) and melarsoprol (for HAT) can be selected with relative ease (Barrett & Fairlamb 1999) because both drugs, and their active metabolites, can enter trypanosomes via the P2 aminopurine transporter, and alteration of this transporter can impair the uptake of both drugs by parasites. It is highly probable that melarsoprol resistance comes by virtue of selection of diminazene resistance of the human trypanosomes through treatment of the animal reservoir (Barrett & Fairlamb 1999). One shortcoming of this study is that the results of the short-term viability assay were not necessarily reflected in the long-term viability assay, especially for diminazene and isometamidium (Table 2). Iten et al. (1997) made a similar observation whereby differences in susceptibility to melarsoprol in their short-term viability assay (hypoxanthine incorporation) could not be confirmed by the longterm feeder-layer viability assay. A possible explanation of this phenomenon could be the short test duration of the 24-hour assay, which may bias results by differences in drug uptake or metabolism of the drugs by different trypanosome stocks (Matovu et al. 1997). Reduced susceptibility of T. b. rhodesiense from Tanzania to melarsoprol, diminazene and isometamidium is a clear sign of the emergence of drug resistance in the sleeping sickness endemic areas of Tanzania. The factors responsible for the emergence of resistance cannot be established by this study, although an important one is subtherapeutic drug concentration. This was not uncommon some years back in Tanzania before the government s decision to provide free treatment for sleeping sickness. At that time most patients could not afford to meet the cost of a complete course of treatment with suramin or melarsoprol due to shortages and the relatively high cost of these drugs. Under dosing may also occur in treating animal trypanosomiasis, for similar reasons. Block treatment of cattle in endemic areas is another factor, which may impose a high selection pressure; this is exacerbated by its frequency. This may be the reason why so many more resistant trypanosomes are isolated from animals than from humans, who are normally treated under the hospital conditions. Sometimes generic products are used, which are less efficacious (Geerts & Holmes 1998). We conclude that there are conditions present in Tanzania, which can lead to selection of trypanosomes resistant to trypanocides. Since it is very unlikely that new trypanocidal drugs will be released on the market in the near future, it is essential to try to maintain the efficacy of the currently available drugs. Authorities in Africa need to adopt an integrated disease management strategy to slow down the development of this resistance. Acknowledgement This study was financially supported by UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases (Project ID A00304 linked to ID ). All laboratory work was carried out at the Faculty of Veterinary Medicine, Makerere University, Kampala, Uganda. We are grateful to the Kenya Trypanosomiasis Research Institute, Kikuyu, Kenya for providing drug resistant T. b. rhodesiense isolates which were used as controls. References Afewerk Y, Clausen PH, Abebe G, Tilahun G & Mehlitz D (2000) Multiple-drug resistant Trypanosoma congolense populations in village cattle of Metekel district, north-west Ethiopia. Acta Tropica 76, Anene BM, Onah DN & Nawa Y (2001) Drug resistance in pathogenic African trypanosomes: what hopes for the future? Veterinary Parasitology 96, Atouguia J & Costa J (1999) Therapy of human African trypanosomiasis: current situation. Memorial Institute Oswaldo Cruz Rio de Jeneiro 94, Baltz T, Baltz D, Groud C & Crockett J (1985) Cultivation in a semi-defined medium of animal infective forms of Trypanosoma brucei, T. equiperdum, T. evansi, T. rhodesiense and T. gambiense. The EMBO Journal 4, Barrett MP & Fairlamb AH (1999) The biochemical basis of arsenical- diamidine cross-resistance in African trypanosomes. Parasitology Today 15, Brun R, Jenni L, Schonenberger M & Schell KF (1981) In vitro cultivation of blood stream forms of Trypanosoma brucei, T. rhodesiense and T. gambiense. The Journal of Protozoology 38, Brun R, Schumacher R, Schmid C, Kunz C & Burri C (2001) The phenomenon of treatment failures in Human African ª 2006 Blackwell Publishing Ltd 153

11 Trypanosomiasis. Tropical Medicine & International Health 6, Croft SL (2004) Drug resistance in leishmaniasis WHO/TDR Scientific Working Group (SWG) Report on Leishmaniasis. Dee Gee ALW, McCann PP & Mansfield JM (1983) Role of antibody in elimination of trypansomes after DFMO chemotherapy. The Journal of Parasitology 69, Enyaru J, Matovu E, Lubega G, & Kaminsky R (1998) Response of a T. b. rhodesiense stock with reduced susceptibility in vitro to treatment in mice and cattle. Acta Tropica 69, Fairbairn H (1948) Sleeping sickness in Tanganyika territory Tropical Diseases Bulletin 45, Geerts S & Holmes PH (1998) Drug management and parasite resistance in Bovine Trypanosomiasis in Africa. PAAT Technical Sciences Series, No.1, FAO, p. 31. Huber W & Koella JC (1993) A comparison of three methods of estimating EC50 in studies of drug resistance of malaria parasites. Acta Tropica 55, Iten M, Mett H, Evans A, Enyaru JCK, Brun R & Kaminsky R (1997) Alterations in ornithine decarboxylase characteristics account for tolerance of Trypanosoma brucei rhodesiense to DL-a-difluoromethylornithine. Antimicrobial Agents and Chemotherapy 41, Kaminsky R, Chuma F & Zweygarth E (1989) Trypanosoma b. brucei; expression of drug resistance in vivo. Parasitology 69, Kaminsky R & Zweygarth E (1989) Feeder layer-free in vitro assay for screening antitrypanosomal compounds against T. b. brucei and T. b. evansi. Antimicrobial Agents and Chemotherapy 33, Legros D, Evans S, Maiso F, Enyaru JCK & Mbulamberi D (1999) Risk factors for treatment failure after melarsoprol for Trypanosoma brucei gambiense in Uganda trypanosomiasis. Transactions of the Royal Society of Tropical Medicine and Hygiene 93, Matovu E, Seebeck T, Enyaru J & Kamisnky R (2001) Drug resistance in T. brucei spp., the causative agents of sleeping sickness in man and nagana in cattle. Microbes and Infection 3, Matovu E, Iten M & Enyaru JCK et al. (1997) Susceptibility of Ugandan Trypanosoma brucei rhodesiense isolated from man and animal reservoirs to diminazene, isometamidium and melarsoprol. Tropical Medicine & International Health 2, Mbwambo HA, Mella PM & Lekaki KA (1988) Berenil (Diminazene aceturate) resistant Trypanosoma congolense in cattle under natural tsetse challenge at Kibaha, Tanzania. Acta Tropica 45, Nyeko JHP, Golder TK & Otieno A (1988) Selection of drug resistance in T. congolense during cyclic transmissions through Glossina m.morsitans and drug treated rabbits. Acta Tropica 45, Ormerod WE (1961) The epidemic spread of Rhodesian sleeping sickness Transactions of the Royal Society of Tropical Medicine and Hygiene 55, Pepin J, Milord F, Khonde A, Niyonsenga T, Loko L & Mpia B (1994) Gambiense trypanosomiasis: frequency of, and risk factors for, failure of melarsoprol therapy. Transactions of the Royal Society of Tropical Medicine and Hygiene 88, Perez Martin O, Lastre Gonzalez M, Urban M & Shwalbach J (1991) In vivo resistance to strains of Trypanosoma rhodesiense, Mozambique Revista Cubana de Medicina Tropical 43, Rollo IM & Williamson J (1951) Acquired resistance to melarsen, ryparsamide and amidines in pathogenic trypanosomes after treatment with melarsen alone. Nature 167, Sones KR, Anjou AR & Holmes PA (1988) Assessment of sensitivity of T. Congolense to Isometamidium chloride: a comparison of tests using cattle and mice. Acta Tropica 45, Robertson DH & Hawking F (1960) Increasing Mel B resistance in consecutively isolated strains of T. rhodesiense from a case of sleeping sickness refractory to treatment. EATRO Report 1960, Ruppol JF & Burke J (1977) Follow-up des traitements contre la trypanosomiase experimentes a Kimpangu (Republic du Zaire). Annales de la Societe Belge de Medecine Tropicale 57, Van Nieuwenhove S (1999) Present strategies in the treatment of human African trypanosomiasis (Chapter 15). In: Progress in Human African Trypanosomiasis Sleeping Sickness (eds M Dumas, B Bouteille & A Buguet) Springer-Verlag, Paris, France, pp Zweygarth E & Röttcher D (1989) Efficacy of trypanocidal compounds against multiple resistant Trypanosoma brucei brucei stock in mice. Parasitology Research 75, Corresponding Author Prof G. W. Lubega, Faculty of Veterinary Medicine, Makerere University, P. O. Box 7062, Kampala, Uganda. Tel.: ; Fax: ; glubega@vetmed.mak.ac.ug 154 ª 2006 Blackwell Publishing Ltd