Living and dying with the tubercle bacillus
This essay was written by Douglas Young and Sebastien Gagneux and was first published in the 2009 Mill Hill Essays. An updated version was published in the Mill Hill Essays anthology.
The public profile of Tuberculosis (TB) has increased over the last decade. While this is largely a result of its inclusion as one of the “big three” global killer diseases, there is a lingering sense amongst TB afficionados that we suffer the neglect of the middle child syndrome. This is reflected for example in the most recent disbursement of support from the Global Fund to Fight AIDS, TB and Malaria in the ratio of 61:14:25 between the three diseases. Increased funding for infectious diseases has been driven by an economic argument, ably championed by economist Jeffrey Sachs, that investment in disease control will generate a multi-fold financial dividend in global productivity. It will be interesting to see how the balance of this argument holds up as the world moves to an era more cautious of the purely capitalist principle. There are many other neglected infectious diseases queuing for a bigger slice of what may prove a diminishing pie.
Renewed public interest and funding has fuelled renewed enthusiasm for TB control. The immediate goal set out in the report Global Plan to Stop TB is to cut the prevalence and mortality of TB in half by the year 2015, in line with the UN Millennium Development Goals. Whilst admirable and challenging in itself, achieving this goal will still leave a million people dying every year from a treatable disease, and the longerterm goal is to eliminate TB by the middle of the century; elimination being defined as less than one case per million of the global population.
Surely this is achievable? We do have a track record in controlling TB. In the mid-19th century, the annual mortality from TB in the UK was around 300 per 100,000 of the population, about ten times today’s global rate. With some ups and downs in times of war, this declined slowly over the following century and, with the advent of effective drugs, TB incidence was dropping by ten percent per year in post-war Europe. However, this success has not been reproduced at a global level. At best, there is a slow annual decline of a few percentage points in some countries; at worst, in sub-Saharan Africa, the last two decades have witnessed a dramatic increase in TB. There are certainly some obvious immediate factors that we can identify as contributing to this failure. Coinfection with HIV, the AIDS virus, significantly impairs the body’s ability to deal with TB; a growing number of strains of Mycobacterium tuberculosis, the causative agent of human TB, have developed resistance to the available drugs; and there are major problems with diagnosis and access to health care in impoverished communities. In addition to these relatively recent issues, it is also useful to think about the epidemiology of TB in a longer-term perspective.
In his book Guns, Germs and Steel, Jared Diamond takes the long view on the evolution of human society, drawing on a very broad historical and geographic canvas. He advances a compelling argument that societies develop differently due not to inherent differences between human populations but to the way that the same instincts and behaviours have different consequences dependent on variations in environmental circumstance. Western societies evolved as a result of fortunate (or perhaps unfortunate, from the standpoint of many nonwestern societies) conjunctions of the availability of cultivable crops and domesticable mammals rather than some particular genius of western man. It’s interesting to try and think about TB in a similar framework.
The ability to determine the genome sequences of M. tuberculosis isolates from different parts of the world allows us to begin to reconstruct an evolutionary history for the microbe and its relationship with humans. All strains of M. tuberculosis are very similar to each other, sharing around 99.95% genetic identity. This represents a level of diversity similar to that across different human populations, and is quite sufficient to generate a range of distinctive appearances and personalities. Based on the available data, we know that M. tuberculosis consists of six main lineages, each associated with different regions of the world (Figure 1). Interestingly, some of these lineages are as genetically distinct from each other as they are from the animal-adapted forms of tubercle bacilli, which are known as Mycobacterium bovis. This can be seen from the length of the branches, which indicate genetic distances in the phylogenetic tree. The association of different M. tuberculosis lineages with different geographic regions and human populations suggests a history of co-evolution between the pathogen and its human host. Our current model envisages a prolonged cohabitation of early humans with some ancestral form of “prototuberculosis”. A dramatic change occurred with the emergence of a particularly successful bacterial variant which came to dominate human infection and which gave rise to almost all of the strains currently in circulation. While our grasp of the timescale of mycobacterial evolution is shaky at best, it is attractive to place the emergence of this new variant as coincident with the emergence of modern man in Africa some 200,000 years ago. The reason for its success may be its aptitude for aerosol transmission, a highly advantageous characteristic for a microbe with a human host increasingly drawn to socially-interactive stable settlements which provide ample opportunity for intimate exchange of respiratory microbes.
M. tuberculosis diversified alongside its human host in Africa, giving rise to two lineages currently found only in West Africa and a third that is found along the Indian Ocean coastal route of human migration out of Africa, which occurred 50-70,000 years ago (Figure 2A). In addition to these three evolutionary “ancient” lineages, a set of “modern” lineages emerged along a separate evolutionary trajectory, expanding in Europe, China and North India; the three major centres of human population explosion in the last millennium. Survival of the ancient strains in low population-density hunter-gatherer societies encouraged them to preserve their limited supply of susceptible hosts by a strategy of prolonged asymptomatic infection and occasional bursts of reactivation disease. This phenomenon is known as latency and is an important characteristic of human tuberculosis. A long latency period may have evolved in order to allow the pathogen to access new susceptible hosts by effectively jumping human generations, and allowing for migration of infected hosts into previously tuberculosis-free areas. By contrast, because of the crowded conditions in the growing cities of Europe, China and India, human life was cheap for the modern TB strains and they tended towards a more aggressive form of disease with a much shorter latency period. The modern strains were able to gain an advantage by triggering rapid progression to disease (and death) whilst retaining a good chance that a new susceptible host would be within coughing distance. Adopting the conquistador characteristics of their hosts, the modern strains thrived in the colonial era (Figure 2B). TB in the Americas is dominated by strains from the European lineage, for example, with no trace of the ancient West African lineages which must surely have been brought over on the slave ships. Most likely, these ancient African strains were out-competed by the modern, more aggressive European strains. Similarly, waves of travel, trade, and conquest out of India and China led to dissemination of the other two modern lineages.
Several consequences flow from this history. Firstly, M. tuberculosis (or something similar) has been a constant factor throughout human evolution, and we might well expect to see some consequence of this hard-wired in our immune system. Reciprocally, when we compare mycobacterial genome sequences, we can look for changes in the bacteria that have been selected during adaptation to different social conditions. These genetic variations identify molecules that may play a particularly important role modulating the precise balance of host-pathogen interactions. Finally, it encourages us to think of TB in different parts of the world in terms of a series of non-synchronous epidemic curves each extending over centuries. If post-war Europe represented the tail-end of an epidemic curve, it might not be surprising if an intervention that was successful in this context has less impact if we are trying to apply it during the upslope of a modern-lineage epidemic in Africa.
While Jared Diamond takes note of the colonial spread of TB in Guns, Germs and Steel, he invokes an evolutionary model that was commonly assumed prior to the genome sequence information, envisaging that TB was introduced to humans as a zoonotic infection during the domestication of cattle by Neolithic farmers. We now know that the strains which cause TB in cattle and other mammals (“M. bovis”) are part of the overall M. Tuberculosis complex, forming a fourth branch in the group of ancient lineages (Figure 1). While the original “prototuberculosis” may have been shared between humans and other animals, the structure of the evolutionary relationships shown in the phylogenetic tree and the greater overall diversity seen amongst human strains hints that we may have transmitted highly-infectious M. tuberculosis to animals rather than vice versa.
In this context, it’s interesting to take a closer look at the strains that cause bovine TB in Africa. We have been studying TB in Ethiopian cattle over several years. Despite the absence of any form of control programme, the prevalence of bovine TB in the Ethiopian countryside is low, with only a few percent of cows showing a positive response to the tuberculin skin test. When we look in detail at the bacteria responsible for bovine TB in Ethiopia, rather than some novel ancient lineages what we find are strains of M. bovis that look suspiciously similar to those circulating in the cows of colonial Europe. We also find significant numbers of human M. tuberculosis strains, again suggesting that we should keep in mind the potential for human-to-animal transmission alongside conventional animal-to-human concerns. The picture is very different as you move from the countryside to the dairy farms around Addis Ababa. Here, a concerted government programme is replacing the indigenous low milk-yield zebu cows (derived from original domestication in India, with the characteristic “Brahman” hump) by high-yield Holstein Friesians, farmed in western-style highdensity intensive herds. The Holsteins are significantly more susceptible and rates of TB in these herds can rise as high as 80 or 90% tuberculinpositivity. It is M. bovis rather than M. tuberculosis that is responsible for this high-prevalence pattern of disease. It may be that variations amongst members of the M. tuberculosis complex, causing the preferential association of particular strains with different animal hosts, represent not so much the ability to cause some pathology in an individual infected animal, but rather the ability to fine-tune this process into an effective transmission cycle. This is the essential art that distinguishes the successful tubercle bacillus, and it is likely to depend on subtle complementary interactions between particular host and pathogen strains. Economic development certainly brings benefits in terms of milk yield, but it looks as though changing the social structure of Ethiopian cows has unfortunate knock-on effects on the population structure of the local tubercle bacilli.
Back to human TB, and the Jared Diamond adage that the same human activity can have different consequences in a different environment is very evident at the level of primary health clinics. During visits over the last year, it was interesting to compare a couple of small-town primary health care centres in Mexico and in South Africa. In outward appearance, they are interchangeable, almost identical. It is a testament to the architects of global TB control that the pattern of service delivery is identical across continents, relying on a highly motivated community nurse who records appointments and drug treatments in a meticulous and tightly-prescribed form of double-entry book keeping. While the systems are identical, the epidemics are worlds apart. In Orizaba in Mexico, a diminishing handful of TB cases are marked by a few coloured pins in a weathered and fading map; in Worcester in South Africa the growing list of TB patients is peppered with deaths, HIV clinic appointments, and notes on drug resistance. Neither centre approaches the trauma of a tertiary healthcare facility in the southern tip of Korea, where patients with advanced multidrug-resistant TB undergo increasingly desperate interventions to prolong their lives. The war photographer, James Nachtwey, has mounted an exhibition with a harrowing series of photographs that document the contemporary horror of the wasting deaths from drug-resistant TB. The high mortality experienced by extensively drug-resistant patients in tertiary hospitals provides an important driving force for the accelerated testing and introduction of new drugs. To have a global impact on TB control, however, new drugs will have to contribute to improved therapy at the level of primary care centres; in particular, to allow the shortening of current 6-month treatment protocols.
The key to success in TB control is to combine prompt diagnosis and treatment that interrupts transmission with interventions that reduce the risk that someone who is infected will progress to active disease. Circumstances were favourable in post-war Europe, perhaps simply reflecting an impact of economic development on general health status. While global TB statistics stand to benefit from anticipated economic uplift in India and China, it is worth keeping in mind the negative impact of development on the cows in Addis Ababa, and watching out for other changes in society that could affect TB trends. A caveat to the positive picture of TB control in the Mexican population discussed above is growing evidence of an association between TB and diabetes for instance. Similarly, epidemiological data from India indicate that while around 5% of TB is linked to HIV-coinfection, 15% of TB is associated with the growing epidemic of diabetes. This suggests a disturbing scenario in which the non-communicable diseases that follow in the wake of rapid economic development might synergise with residual poverty-related infections to generate new patterns of disease. On the research front, we need to find ways to shortcut the economic equation with other interventions that reduce progression to disease. New vaccines and efficient ways of treating latent infection are high on the agenda.