The factors responsible for the development of resistance to antitrypanosomal compounds are not well known. The exposure of parasites to subtherapeutic drug concentrations (owing to underdosing) has been considered as the most important factor for the development of resistance (Whiteside, 1960; 1962; Boyt, 1986). Boyt suggested that the evolution of drug resistance in trypanosomes is fundamentally different from resistance in insects, helminths or micro-organisms. For the latter, it is generally accepted that resistance genes are present in a very small proportion of the population and that these pre-existing resistant individuals are selected by drug pressure. Boyt suggested, however, that in a manner similar to antigenic variation, chemoresistance is another example of the remarkable ability of the trypanosome to adapt defensively in the face of unfavourable changes appearing in its environment. However, no concrete evidence has been brought forward to show that resistance is indeed an adaptation rather than a selection process. Nevertheless it is possible to take drug-sensitive clones and induce drug resistance rapidly by repeated underdosing and passage. Furthermore, Osman, Jennings and Holmes, (1992) showed that this occurs much more rapidly in immunosuppressed hosts (mice). Similar observations were also reported by Hess et al. (1997), who showed that malnourished children (malnutrition adversely affects the immune status) were at a higher risk from treatment failure resulting from drug-resistant Plasmodium falciparum than well-nourished children. These findings serve to illustrate the importance of immune responses in drug clearance of parasites. The type and degree of host immunity - either generalized (regulation of the overall level of infection) or specific immunity (each clone within the infection is regulated independently) - might also influence the rate of development of drug resistance, as was shown for Plasmodium (Hastings, 1997). This aspect needs to be examined thoroughly for trypanosomes.
How do trypanosomes develop resistance to trypanocidal drugs? Selection by drugs takes place during asexual multiplication in the animal or human host. During the passage through the tsetse fly genetic exchange (sexual recombination) may occur, at least in T. brucei (Jenni et al., 1986; Tait and Turner, 1990), and is strongly suspected in T. congolense given the high degree of genetic diversity observed (G. Hide, personal communication). The genetic structure of a parasite population (clonal or panmictic) is an important parameter that is influenced by the transmission intensity, which in turn might influence the rate of development of drug resistance as is suggested by recent research on drug resistance in Plasmodium (Paul et al., 1995). Similar to Plasmodium (Thaitong, 1983), populations of trypanosomes in infected animals are polyclonal with different sensitivities to antitrypanosomal drugs (Peregrine et al., 1991; Mutugi, Boid and Luckins, 1995). Therefore drug resistance in trypanosomes is likely to occur under the same circumstances as for Plasmodium and many other parasites, that is: i) under large-scale drug use; ii) by using inadequate dosing; and iii) by using correct dosing with drugs that are slowly eliminated from the body (White, 1992). Furthermore, some trypanocidal drugs are well-known mutagenic compounds and might induce mutations, the most resistant of which are certainly selected under drug pressure (Hayes and Wolf, 1990). Taking into account the high basic mutation rate in trypanosomes, which is estimated at 10-9 per base pair per cell generation in T. brucei (Valdes et al., 1996), the effects of this phenomenon should not be underestimated. Figure 2 summarizes some important factors influencing the development of drug resistance in trypanosomes.
Up to now the most important guidelines on the avoidance or delay of the development of drug resistance were considered to be: i) use of the "sanative pair" of drugs (ISMM or ethidium and diminazene); and ii) avoidance of the exposure of trypanosomes to subtherapeutic drug concentrations (Whiteside, 1960; Boyt, 1986). It is clear, however, that the application of these guidelines may not be sufficient to maintain the efficacy of the existing drugs, especially since they lack recommendations concerning a reduction of the treatment frequency.
Based on current knowledge in the field of trypanocide resistance and on experience in the control of resistance to insecticides, anthelmintics, antibiotics and other drugs (Boray, Martin and Routh, 1990; Bergogne-Bérézin, 1997; Geerts, Coles and Gryseels, 1997; Routh, 1993) the following re-commendations are proposed in order to delay the development of resistance.
It is widely agreed that the most efficient way to delay the development of drug resistance remains the reduction of selection pressure by the drugs, i.e. decreasing the number of treatments. This is of particular importance in areas of high tsetse challenge, which are commonly associated with reduced periods of chemoprophylaxis (Whiteside, 1960; 1962). In such situations the treatment frequency is commonly increased and drug resistance often emerges as a constraint to further drug usage. Very intensive drug treatment schedules, as described by Stevenson et al. (1995), who administered ISMM six to seven times a year and ethidium up to11 or 12 times a year, might be able to control the resistance problem temporarily, but are no solution in the long term. Furthermore, frequently repeated trypanocidal drug treatments have been associated with toxicity problems (Stevenson, Munga and Dolan, 1993; Eisler et al., 1997b). This kind of approach inevitably increases the selection pressure - in the absence of any measures such as the use of the sanative pair to counteract the development of resistance - and must lead to increased levels of drug resistance. It has been shown in other areas that there is a strong correlation between the treatment frequency and the rate of development of resistance (Conder and Campbell, 1995; Routh, 1993). It is therefore strongly recommended that in high tsetse challenge areas control of trypanosomiasis should not rely solely on drugs but that an integrated approach should be adopted using vector control, to reduce the tsetse challenge, along with reduced frequency of drug dosing. Where such measures have been adopted the results have been impressive (Fox et al., 1993; Peregrine et al., 1994). In situations in West and Central Africa the use of trypanotolerant livestock and drugs may be appropriate in areas of high tsetse challenge (Diall et al., 1992).
Underdosing is one of the major causes of resistance development. Subtherapeutic drug concentrations exert a strong selective pressure for the emergence of resistant clones that pre-exist in the trypanosome population. Unfortunately, underdosing occurs very frequently. Farmers have the tendency to underestimate the weight of their animals when they have to treat them (Besier and Hopkins, 1988). Sometimes generic products are used, which have a reduced efficacy, as has been shown in the field of anthelmintics (Van Wyk et al., 1997), while anecdotal evidence suggests that it also occurs with trypanocides. Given the fact that in many countries unskilled persons are allowed to administer drugs, errors easily occur in calculating the correct doses for the treatment of the animals. Packaging of isometamidium as a one-dose treatment - similar to diminazene - would undoubtedly help to reduce this problem. In addition, as the drugs are relatively expensive there is a temptation to overdilute the drug and hence underdose.
Data on the pharmacokinetics of ISMM in goats (Kinabo and McKellar, 1990) suggest that following intramuscular administration the bioavailability of ISMM in that species is very low, approximately half that in cattle under similar circumstances (Eisler, 1996). If this is confirmed, higher doses might be used in goats than in cattle, although caution is required because of the low therapeutic index of ISMM. Similar observations were made for benzimidazoles and levamisole, the dosage of which should be 1.5 to 2 times higher in goats than in sheep (Hennessy, 1994).
The use of improved formulations of existing drugs is another possible way to avoid subtherapeutic concentrations. Controlled release devices which provide more stable drug concentrations and a sharper cut off at the end of the release period might have particular advantages in this respect. The polymer devices containing ISMM or ethidium as described by Geerts et al. (1997) are a step in this direction.
Unlike human sleeping sickness, animal trypanosomiasis is commonly controlled with mass treatments which can be highly successful over many years in ranch cattle for example (Trail et al., 1985). However, this form of treatment exerts a strong selection pressure on the trypanosome population. The higher the proportion of the trypanosome population exposed to the drug and the lower the proportion in refugia (i.e. the proportion of trypanosomes present in the fly population or in other hosts), the higher the selection pressure. The percentage of the total parasite population that is exposed to the drug at the time of treatment might, thus, have an impact on resistance development. It has been shown by computer models that leaving 20 percent of the herd untreated significantly decreases the rate of development of anthelmintic resistance (Barnes, Dobson and Barger, 1995). Although no experimental data are available for trypanocide resistance, experiences with other drugs and pesticides indicate that systematic mass treatments hasten the development of resistance. Therefore, in well monitored situations there may be a case for limiting treatment to individual clinical cases. In such situations, drug-resistance problems can be minimized and acquired immunity encouraged (Scott and Pegram, 1974). A similar approach is currently being used in South Africa to control anthelmintic-resistant Haemonchus contortus in sheep (Van Wyk, Malan and Bath, 1997). Instead of carrying out systematic mass treatments, only those sheep that are strongly anaemic are treated. The identification of animals not able to cope with H. contortus was done by clinical appraisal of the colour of the ocular mucous membranes using a colour chart (FAMACHAR). This allowed a significant reduction in the number of treatments without an increase in the mortality rate. The identification of trypanosome-infected animals by farmers and veterinarians would be greatly assisted by the development of reliable and cheap pen-side diagnostic tests. The need for such tests is widely recognized and they remain an important goal for the research community
Quinapyramine was widely used in cattle in Africa during the period 1950 to 1970. In 1976, it was withdrawn from sale for cattle use because of problems with toxicity and resistance development. It is still available for use in camels, however, and it is likely that it is still mistakenly used in cattle in some situations in Africa. The use of quinapyramine was the suggested cause of the multiple drug-resistance problem in the Ghibe valley of Ethiopia referred to earlier. Ndoutamia et al. (1993) showed that, after artificial induction of resistance to quinapyramine in T. congolense, multiple resistance to ISMM, homidium and diminazene was expressed at the level of the individual trypanosome and could be transmitted by tsetse flies. This confirms the results obtained in earlier field studies by Whiteside (1962). The use of quinapyramine as a trypanocide in cattle is, therefore, completely contraindicated.