Throughout history, tuberculosis (TB) has taken a remarkable toll on human health. Even today, more than 50 years after the advent of effective antibiotic therapy, this bacterial infection kills millions every year. The resilience of this disease in the face of modern treatment is largely a result of the unusual cell biology of the causative agent, Mycobacterium tuberculosis. Like many other intracellular pathogens, M. tuberculosis resides largely within immune cells of the host. However, this is where the similarity with these other well-studied pathogens ends. M. tuberculosis is covered with a uniquely impermeable cell wall, grows remarkably slowly, and has a cellular architecture that is completely distinct from other bacterial pathogens.
With its unique biology, it is perhaps not surprising that the genome sequence of M. tuberculosis provides few recognizable clues to the molecular mechanisms used to interact with the host and cause disease. We can predict the functions of only a relatively small minority of this organism's genes, leaving thousands uncharacterized and likely responsible for novel interactions with the host.
Probing the Genomic Dark Matter
These large tracts of genome with no known function are not unique to M. tuberculosis. While the ease with which microbial genomes can be sequenced has increased exponentially over the past decade, the methods used to determine the functions of the identified genes have changed very little. Even in the best-studied model organisms, such as Escherichia coli or yeast, precise functional predictions can only be made for perhaps half of the genes in the genome. This fraction is even smaller in medically important, but phylogenetically divergent, organisms such as mycobacteria. To understand the molecular pathogenesis of TB and to develop fundamentally new treatments, we have devised new methods to probe the functions of these genetic orphans rapidly.
A critical first step to defining a gene's function is to identify the specific conditions under which it is essential for the organism to survive. To do this systematically and on a genome-wide scale, we devised a method called transposon site hybridization (TraSH). This approach uses either microarray hybridization or newer deep-sequencing methods to characterize the composition of large libraries of bacterial mutants in order to identify the complete set of genes that is required for survival under any particular condition. Thus far, we have identified distinct sets of hundreds of genes that are required in laboratory medium, for virulence in animals, and in models of microbial latency. We have derived phenotypic information for approximately one-third of the genes in the genome, including hundreds for which no functional information was previously available.
The next step toward understanding how genes function is to define how they work together as distinct functional units or pathways. To do this, we have adapted TraSH to identify "genetic interactions." This allows the rapid identification of genes that rely on each other for their function and therefore work in a concerted fashion. This approach provides a wealth of functional information, since orphan genes can often be placed in pathways with predictable functions. We have used this on a small scale to characterize individual virulence genes, and we are expanding this effort to define virulence pathways more globally and to understand their interdependencies.
As described below, we have also used these high-throughput genetic methods to unravel the bacterium's eating and sleeping habits, and we are investigating whether this information can be used to devise more-effective therapies.
The "Big Mac" Diet
Once M. tuberculosis is inhaled into the lung, it is quickly engulfed by resident immune cells called macrophages. Although these cells are designed to kill invading microbes, some pathogens, including M. tuberculosis, have the ability to survive and grow within a vacuole-like compartment of this cell. One of the most fundamental, but unexplored, challenges facing any pathogen living in this niche is acquiring nutrients while apparently sequestered inside a host-derived membrane. We have begun to explain this ability by demonstrating that M. tuberculosis can extract cholesterol from the host cell and has the unusual ability to use this compound as a source of carbon and energy. Thus, M. tuberculosis survives inside of the macrophage vacuole by eating a ubiquitous component of the mammalian membrane. Cholesterol is clearly not the only carbon source used by the bacterium during infection, and we are characterizing additional systems involved in the uptake of nutrients and determining if new TB treatments can be designed that limit the bacterium's access to these compounds.
Understanding Latent TB
The most unique feature of TB is the long period of clinical latency that often precedes disease. During this phase, the bacteria are thought to replicate slowly, if at all. This behavior is largely responsible for both the staggering rate of asymptomatic carriage (approximately 1 billion individuals) and the relative ineffectiveness of antibiotics. We have applied our genetic tools to understand the physiology of these dormant bacteria, and we have identified two classes of bacterial mutants in which this response is altered. One class is unable to enter the quiescent state and continues to replicate under dormancy-inducing conditions. A second class is unable to survive long periods of stasis. By characterizing the corresponding genes, we are beginning to understand how M. tuberculosis regulates its growth rate in response to stress and manages to maintain its cellular integrity for months, or even years, of stasis.
Interrupting Cellular Dormancy to Increase Antibiotic Efficacy
Arguably the most significant limitation of current TB control efforts is the remarkably long antibiotic regimen that is necessary to cure disease and prevent relapse. This situation is not uncommon, as many chronic infections involve populations of bacteria that replicate slowly or not at all and are difficult to eradicate with antibiotics. In the case of TB, shortening the treatment regimen, even modestly, could have a profound impact, both by increasing cure rates and decreasing the selection of drug-resistant strains. Since virtually all bacteria are more sensitive to antibiotics when actively growing and metabolizing, we hypothesized that inhibiting or interrupting dormancy could increase the effectiveness of antibiotics and speed cure. Indeed, we have identified multiple genetic mutations that increase drug efficacy by altering the metabolic state of dormant bacteria. Using this as proof of principle, we are pursuing high-throughput chemical screening approaches to identify small molecules that produce similar effects and could synergize with existing drugs to accelerate TB cure.