G. Hide
Wellcome Unit of Molecular Parasitology, University of Glasgow, Andersen College 56 Dumbarton Road, Glasgow, UK.
References
Firstly, I would like to thank Dr Tibayrenc for an interesting paper and also the organisers for inviting me to contribute to this meeting.
Dr Tibayrenc discusses the problems of using molecular markers as a tool for the identification of strain types and the epidemiological tracking of strains. The problem is that to define a strain you need to take into account both the variation within that strain and the differences between that strain and another strain. This can be done relatively easily for a clonal strain - where genetic exchange and recombination do not reassort the genome (so within strain markers become stable) -but it is much more difficult where strains are not clonal because, of course, genes from one strain can become intermixed with genes of another and the defining edges of a strain group become blurred. To overcome these pitfalls, Dr Tibayrenc proposes a number of strategies. Firstly, it is necessary to define whether a species is clonal. In Trypanosoma cruzi it most certainly is. In Trypanosoma brucei there is some evidence for clonality and some which suggest the occurrence of recombination. I will come back to this issue. Secondly, Dr Tibayrenc proposes that we develop an approach which takes into account within and between strain/species variation and develop an understanding of the levels of that variation across all groups of micro-organisms such that we can define molecular levels of variation which distinguish taxonomic levels (eg species or subspecies) and can be used, epidemiologically, to define medically important strains. This, I think is a very desirable end-point but may be difficult to achieve because of different rates of evolution of the same markers in different organisms and different levels of clonality between and within the same species.
Coming back to T. brucei, we have evidence to suggest that two populations of T. brucei which were sampled in the same time and space (Tororo District, Uganda, 1990), have differing levels of clonality. Isolates taken from cattle during the 1990 epidemic are clearly not clonal while isolates taken from man (and human serum resistance isolates taken from cattle in Tororo) form a group which are probably almost clonal although may show remnants of genetic exchange (John Maynard Smith calls these an epidemic population because underlying levels of genetic exchange are masked by the spread of a single or small number of strains). Thus it appears that, at least in this region, two completely different transmission cycles are super-imposed. This has an important implication when considering evolutionary genetic approaches. This is that the sampling methods used may have important effects on the outcome of these types of analysis. In many analyses, a common approach is to pool data to ensure large samples. Three cases exist where this may lead to problems (such as, resulting in the detection of genetic linkage disequilibrium in a mating population): (1) When the population has been sampled from different places; (2) When the population has been sampled at different times and (3) when the population is made of subgroups which never have the opportunity of undergoing genetic exchange.
I think that the use of Dr Tibayrenc's approaches are important in developing approaches to molecular epidemiology - but we should not forget to employ the correct sampling strategies.
As a specific example of the uses of molecular epidemiology, I would like to consider some of our work on the epidemiology of human sleeping sickness in East Africa.
One of the fundamental aspects required for an understanding of the epidemiology of human sleeping sickness, is a suitable method for identifying and tracking strains. The complexity of the transmission cycles involved (man, tsetse, domestic animal reservoir and wild animal reservoir) and the comparative dynamics of transmission through parts of these cycles, means that sensitive genetic markers are needed to follow strains through these cycles to obtain meaningful epidemiological data. To achieve this, we set out to develop a set of epidemiologically useful markers which would provide a molecular fingerprint for individual trypanosome isolates (Hide et al., 1990). Our system was based on Restriction Fragment length Polymorphism (RFLP) analysis of repetitive DNA. Briefly, this involves extraction of DNA from a cloned trypanosome stock, digestion of the DNA with a range of (DNA sequence specific) restriction enzymes, separating digested DNA fragments on agarose gels, transfer of these separated fragments to a nylon (DNA binding) membrane and hybridisation to a series of repetitive DNA probes (see Hide, 1996, for methods). This sequence of steps results in a set of banding patterns (similar to those patterns seen in human DNA fingerprinting) which are unique to each trypanosome stock. However, many of the bands are shared between different stocks which allows us to calculate a similarity coefficient to describe the relatedness of one stock to another. Using these coefficients from a large number of stocks, cluster analysis techniques can be used to generate a dendrogram (or tree) of relationships where related stocks cluster as groups.
In our initial study, we used populations of trypanosome isolates (or stocks) from West Africa (from humans and from cattle), from Kenya and Uganda (from humans, cattle and tsetse) and from Zambia (from humans). The first thing that became apparent was that the classical definitions did not describe the groups we saw. Trypanosoma brucei gambiense (classically defined as West African human infective stocks) grouped into two quite distinct groups. Furthermore, the T. b. rhodesiense stocks (classically defined as human infective strains from East Africa) collected in Kenya/Uganda and Zambia were also of two different types. This suggested to us that T. b. rhodesiense wasn't a monophyletic group, that is, a single strain which has spread throughout East Africa but that human infectivity has arisen more than once. To investigate this aspect further, we characterised a larger collection of trypanosome strains from both the Busoga focus (Kenya/Uganda) and the Zambian focus (Hide et al., 1991). It was clear from this study that the, so-called, T. b. rhodesiense strains from the two foci were as different from each other as each was from T. b. gambiense. This raises some important questions. Firstly, if human infectivity has arisen more than once, how frequently does it arise in field populations? Secondly, is the biological mechanism for resistance to human serum different in populations of human infective trypanosomes from different areas? Thirdly, do different human trypanosome populations differ in their susceptibility to drug treatment? (For example T. b. gambiense is effectively treated with DFMO but T. b. rhodesiense is not). Finally, is genetic exchange important in spreading the human infective genotype to other human serum sensitive strains?
To address some of these questions, we carried out an analysis of 88 trypanosome stocks collected at the height of a human sleeping sickness epidemic in the Tororo District of Uganda in 1990 (Hide et al., 1994). In addition to the RFLP analysis, we also analysed the stocks for sensitivity to human serum and isoenzyme profiles. We wanted to address a number of specific questions: (1) Could human infective strains (T. b. rhodesiense) be distinguished from non-human infective strains (T. b. brucei)? Prior to this study, it had been thought that T. b. rhodesiense was a host range variant of T. b. brucei (Tait et al., 1985) and that they could not easily be distinguished (Godfrey et al., 1990). (2) As humans and cattle lived in close proximity in the Tororo District, how significant were the cattle as a reservoir host during the epidemic? (3) Does genetic exchange occur in this natural population of trypanosomes? Considerable controversy existed, in the literature, between those who found no evidence of genetic exchange in natural populations of trypanosomes (the clonal theory -Tibayrenc et al., 1990; Tibayrenc, 1995, 1997; Stevens and Welburn, 1993) and those who had demonstrated the occurrence of genetic exchange in laboratory experiments (Jenni et al., 1986; Gibson, 1989; Tait and Turner, 1990; Tait et al., 1996). The Tororo collection was an ideal collection for carrying out population genetic studies. In this analysis we found that the isolates grouped into two overall groups: (1) a highly homogeneous group, in terms of molecular fingerprints, which were all human serum resistant or isolated from humans. These were T. b. rhodesiense, 23% of the isolates from cattle were of this type. Furthermore, the banding patterns from these isolates were very closely related to stocks isolated from humans in the 1960 epidemic in the Kenya and the 1980 epidemic in Uganda. As noted previously these stocks were completed unrelated to the T. b. rhodesiense stocks from Zambia. (2) A group of stocks which were highly heterogeneous (by molecular fingerprint banding patterns) and which were all human serum sensitive and were never found in any of the human isolates. These were T. b. brucei. With regard to the genetic exchange question, we did not observe any evidence of genetic exchange between the cattle strains and the human strains. This was also supported by the apparent stability of the genotype of the human strains from 1960-1990 in this focus. Using the isoenzyme patterns, we investigated the applicability of clonality to our populations. We showed that neither the populations of cattle stocks nor the human stocks conformed to the Hardy-Weinberg predictions for random mating suggesting that genetic exchange was not occurring. However, the Hardy-Weinberg approach really only addresses the question as to whether random mating is occurring. Another approach, proposed by John Maynard Smith et al. (1993), addresses the question from the opposite viewpoint - can mating be detected? Using this approach, the answer was yes in both cases suggesting that these populations are not clonal. However, the population structure of the human stocks was distorted by the rapid spread of a small number of strains - an epidemic population (Maynard Smith et al., 1993). More recent analyses, using this method, suggest that other T. brucei populations are clonal (Stevens et al., 1996). It seems, therefore, that much of the debate may be due to the fact that some T. brucei populations are clonal while others are not.
The importance of the animal reservoir became clear from the Tororo study. As 23% of cattle harboured human infective trypanosomes, we calculated that the cattle-fly-man transmission cycle was five times more probable than the man-fly-man cycle (Hide et al., 1996; Hide, 1997). If this is the case then the priorities for the control of an epidemic may lay in treating the cattle not the people!
Having examined the trypanosome strains circulating in an epidemic, it was then important to investigate the strains circulating during an endemic period. A collection of stocks were made from the Busia District of Kenya, some 20km from Tororo, during an endemic period. This study showed that the same genotypes of cattle stocks and human stocks were circulating in cattle in this region (Hide et al, in preparation). Furthermore, we also sampled individual animals over a long time period (3 months). We found, surprisingly, that although different animals had different genotypes of circulating trypanosomes, a single animal always had the same trypanosome strain present. This suggested that T. brucei trypanosomes are clonal within an individual animal. However, in a number of cases, these cattle were also infected with T. congolense and T. vivax. This raises the interesting questions as to what allows the cattle to procure mixed infections with different trypanosome species but to be clonal with respect to the genotype of the T. brucei strain. Perhaps the answer lies in some trypanosome communication or social system which maintains the predominance of one genotype in an animal or perhaps it is a result of the population dynamics of transmission leading to a population bottle-neck. However, the implications of this for the control of disease may be important. If cattle are an important reservoir for human sleeping sickness, if mixed infections do not exist, if only a proportion of cattle carry human infective trypanosomes then specific identification of those cattle carrying human trypanosomes followed by specific drug treatment of those cattle may be all that is required to damp down or prevent an epidemic. If 23% is a realistic figure for the proportion of cattle carrying human infective trypanosomes, then specific treatment of those cattle will cost a quarter of the cost of treating all cattle. More studies need to be carried out in this area and a priority is the development of a sensitive, quick assay for human infective trypanosomes in cattle.
Gibson, W.C., 1989. Analysis of a genetic cross between Trypanosoma brucei rhodesiense and T. b. brucei. Parasitology, 99: 391-402.
Godfrey, D.G., Baker, R.D., Rickman, L.R. and Mehlitz, D., 1990. The distribution, relationships and identification of enzymic variants within the subgenus Trypanozoon. Advances in Parasitology, 29: 1-74.
Hide, G., 1996. The molecular identification of trypanosomes. Methods in Molecular Biology, 50:243 - 262.
Hide, G., 1997. The molecular epidemiology of trypanosomatids. In: Trypanosomiasis and Leishmaniasis: biology and control. CAB International, Wallingford, Oxon, UK., Eds: Hide, G., Mottram, J.C., Coombs, G.H. and Holmes, P.H., pp. 289-304.
Hide, G., Angus, S., Holmes, P.H., Maudlin, I., and Welburn, S.C., 1997. Comparison of Trypanosoma brucei strains circulating in an endemic and an epidemic area of a sleeping sickness focus. In preparation.
Hide, G., Buchanan, N., Welburn, S.C., Maudlin, I., Barry, J.D., and Tait, A., 1991. Trypanosoma brucei rhodesiense: Characterisation of stocks from Zambia, Kenya and Uganda using repetitive DNA probes. Experimental Parasitology, 72: 430-439.
Hide, G., Cattand, P., LeRay, D., Barry, J.D., and Tait, A., 1990. The identification of T. brucei subspecies using repetitive DNA sequences. Molecular and Biochemical Parasitology, 39: 213-226.
Hide, G., Tait, A., Maudlin, I., and Welburn, S.C., 1996. The origins, dynamics and generation of Trypanosoma brucei rhodesiense epidemics in East Africa. Parasitology Today, 12: 50 - 55.
Hide, G., Welburn, S.C., Tait, A., and Maudlin, I., 1994. Epidemiological relationships of Trypanosoma brucei stocks from South East Uganda: evidence for different population structures in human infective and non-human infective isolates. Parasitology, 109: 95-111.
Stevens, J.R. and Tibayrenc, M., 1996. Trypanosoma brucei S.I.: evolution, linkage and the clonality debate. Parasitology, 112:481-488.
Stevens, J.R. and Welburn, S.C., 1993. Genetic processes within an epidemic of sleeping sickness. Parasitology Research, 79: 421-427.
Tait, A., Barry, J.D., Wink, R., Sanderson, A., and Crowe, J.S., 1985. Enzyme variation in Trypanosoma brucei ssp II. Evidence for T. b. rhodesiense being a subset of variants of T. b. brucei. Parasitology, 90: 89-100.
Tait, A., Buchanan, N., Hide, G., and Turner, C.M.R., 1996. Evidence for self-fertilisation in T. brucei. Molecular and Biochemical Parasitology, 76: 31 - 42.
Tibayrenc, M., 1995. Population Genetics of Parasitic Protozoa and other Microorganisms. Advances in Parasitology, 36: 47 115.
Tibayrenc, M., 1997. Evolutionary genetics of Trypanosoma, Leishmania and other microorganisms: epidemiological, taxonomical and medical implications. In: Trypanosomiasis and Leishmaniasis: biology and control. CAB International, Wallingford, Oxon, UK., Eds: Hide, G., Mottram, J.C., Coombs, G.H. and Holmes, P.H., pp. 305 -315.
Tibayrenc, M., Kjellberg, F. and Ayala, F.J., 1990. A clonal theory of parasitic protozoa: the population structures of Entamoeba, Giardia, Leishmania, Naegleria, Plasmodium, Trichomonas and Trypanosoma and their medical and taxonomic consequences. Proceedings of the National Academy of Sciences of the USA, 87: 2414-2418.
