Douglas Young

Department of Medical Microbiology, Imperial College School of Medicine at St Mary's Hospital, Norfolk Place, London

From a symposium on the theme of'Genetics and Tuberculosis', it is attractive to consider tuberculosis as a meeting of the mycobacterial genome with the human genome, and suggest that identification of the genes involved in this interaction will hold the key to understanding the disease. There are two caveats to such a conclusion, however. The first came up in the presentations from Richard Bellamy and Igor Kramnik. Variations in a large number of genes — probably at least 100 — appear to influence the host response to infection, and there may be many additional background genes that, although not having a direct role in infection, can influence those genes that are involved. Igor showed that, in a mouse model, certain genes could have a beneficial effect in some backgrounds and a detrimental effect in others, for example. Susceptibility or resistance is a complex phenotype, in which understanding of interactions between multiple genes may well prove more important than identification of individual genes.

Similarly in bacteria, genes encoding factors that are directly involved in host cell interaction are most probably influenced by other background physiology genes. The second caveat is that, whatever the genotype, factors related to nutrition, stress and endocrine status all have a profound influence on disease susceptibility. To understand tuberculosis, therefore, we need not only to identify the key host and microbial genes, but also to understand how their function is regulated by interactions with other genes and by physiological and environmental variables. Genetics is not going to provide a simple one-word answer that will solve all our problems in tuberculosis; we have to regard genetics as a powerful approach that we can add on to our existing ways of trying to dissect these problems. In summarizing our discussions over the last few days, I will divide my comments into ways that genetics are being used to answer questions firstly about the microbe, then about mechanisms of protective immunity and finally about the epidemiology of tuberculosis.

In terms of the bacteria, drug resistance is an area in which genetics has already answered important questions and generated improved tools for disease control. However, as Julian Davies and Lynn Miesel pointed out, there is more to be learned about drug resistance than simply identification of genotypic markers.

Julian encouraged us to think of what we might learn about the biology of Mycobacterium tuberculosis by looking more closely at the development of resistance. He referred to the horizontal transfer of genes. Although, so far, this has not played a role in resistance in M. tuberculosis, phage-like elements have been identified by genome sequencing, and the possibility that horizontal gene transfer has contributed to virulence, and might make a future contribution to resistance, merits careful evaluation.

Stewart Cole's presentation of the near-complete sequence of the genome of M. tuberculosis provided an exciting vision of future research directions. He illustrated the way in which the process of accumulation of sequence data is being transformed into functional genomics, with novel and unexpected implications that can be fed back to the 'wet' lab for experimental analysis. With rapid progress in bacterial genome sequencing, comparative genomics will revolutionize the study of bacterial physiology and pathogenesis over the next decade. The development of expertise in bioinformatics will be an important aspect in ensuring that tuberculosis research can optimally exploit progress in the field of bacterial genomics.

Paul van Helden discussed ways of studying mycobacterial virulence by looking at transmission of different strains in an epidemiological setting. He discussed questions associated with the evolution of M. tuberculosis, and the 'virulence' or 'fitness' of individual strains. A limitation of current studies in this area is that, although we now have useful genotypic markers for strain variation (such as IS6110), we have yet to identify variable genetic elements that have an influence on biological phenotype. Genomic analysis will provide a rich source of candidates for testing as potential phenotypic markers. Considerable benefit might be anticipated in this area from interactions with experts in the study of other bacterial diseases, in which the science of population genetics is more advanced.

The interaction of mycobacteria with macrophages is a central event in the infectious process and was discussed briefly by David Russell. He described the development of elegant approaches to study the biology and biochemistry of the mycobacterial phagosome, in combination with genetic tools to investigate the way in which the pathogen subverts the normal host response. By establishing the precise sequence of events leading to mycobacterial killing, or survival, within the macrophage, we can set a framework for understanding the significance of the other immunological events associated with infection.

We spent a lot of time discussing the immune response to infection. We agreed that macrophage activation by the y-interferon (IFN-y) pathway plays a central role in protection, although additional cell—cell interactions may also influence the killing of intracellular mycobacteria. There was much discussion of why, if IFN-y is protective, we can have IFN-y production and still have disease. Where does the immune response go wrong? What is the mechanism of disease-associated immunopathology? The effect of IFN-y might depend on the presence or absence of other immune mediators. A combination of IFN-y with excessive levels of inflammatory cytokines, such as tumour necrosis factor (TNF), may be detrimental, for example. Another suggestion was that TNF in the presence of T helper 2 cytokines might contribute to pathology. Ian Orme suggested that immune activation driven predominantly by cytokines might provide protection, whereas a chemokine-driven response might be associated with delayed-type hypersensitivity. Graham Rook discussed mechanisms by which the endocrine system can influence the immune response; potentially antagonizing or synergizing with the IFN-y pathway. Finally, Peter Donald cautioned that humoral responses should not be ignored, because they may influence the initial stages of infection or the transfer of bacteria from one cell to another.

There is no simple way to disentangle this complex network of interacting immune mediators. Animal models represent one useful approach, with knockout mice providing a powerful tool to determine whether or not individual cytokines play an essential role in protection. It is more difficult to analyse the hypothesis that protection versus pathology is determined by quantitative relationships between different mediators, however. In the context of human tuberculosis, Gilla Kaplan and Graham Rook discussed the way in which immunotherapy trials might be used to evaluate the contribution of particular cytokines or T cell subsets to the outcome of the immune response. Paul Fine pointed out that the differential effect of Bacillus Calmette—Guerin vaccination, protecting some populations but not others, provides an opportunity to test the role of particular types of immune response in protection.

Richard Bellamy discussed two genetic approaches that could be used to study susceptibility in human populations. The candidate gene approach could have an important role in evaluating the contribution ofparticular cytokines to protection. Alternatively, screening based on a whole-genome approach has the potential to identify genes, and consequently mechanisms, that had not been seen by conventional biology. NRAMP was initially identified by whole-genome screen in mice, for example, with the product now being actively studied for its biological function. Recent candidate gene approaches suggest that NRAMP may also have some role in tuberculosis susceptibility in humans. The problem of the whole-genome approach is that its 'power' is relatively low in a situation in which multiple genes each make a relatively small individual contribution to overall susceptibility. It may be possible to increase the power of the genetic screen by focusing on susceptibility to particular types of disease — reactivation versus reinfection disease, for example. A combination of human genetics with bacterial genetics (to identify particular disease situations) might be advantageous. It is clear that we are only at the beginning of exploitation of human genetic analysis in tuberculosis research.

I have focused my comments particularly on the mechanisms underlying the biology of tuberculosis from the perspective of fundamental scientific understanding. With the urgent need for improved disease control, it is clear that we cannot wait for a complete scientific understanding before proceeding with attempts to develop new drugs and vaccines for tuberculosis. Progress in combating tuberculosis will ultimately depend on a synergy between the 'push' of basic science and the 'pull' of the public health priority. Our discussions over the last few days have highlighted the exciting potential that novel genetic approaches provide in facing the formidable problems of this most complex and persistent of human infections. I'm sure that if Rene Dubos had been here, he would wish once more to be a young researcher setting out on the fight against tuberculosis, armed with today's powerful new genetic tools.

Novartis 217: Genetics and Tuberculosis.

Copyright © 1998 John Wiley & Sons Ltd Print ISBN 0-471-98261-X elSBN 0-470-84652-6

Was this article helpful?

0 0

Post a comment