Persistent, Non-Dividing: TB with Carl Nathan

tl;dr Nathan says the way forward in TB drug discovery is to target non-dividing, persistent mycobacterial cell populations. Do this by identifying pathways that are required for the “phenotypcially-resistant” state - targets like the mycobacterial proteosome and other components of proteolyses-regulaiton machinery. Then use classic screening and medchem to find inhibitors of these targets. These types of drugs should be complementary to the current set of drugs and can get rid of the bugbear of clinical TB: the lurking antibiotic non-responsive bugs that will start dividing down the road.

When Carl Nathan comes to speak, arrive early. I made the mistake of being a few minutes late to his recent lecture and ended up crawling in the front door of the packed lecture hall to claim the only free space in the hall (probably a higher-capacity venue should have been used). I have never heard Carl Nathan deliver a lecture but have had several people independently remark on the elegant style of his reviews - and man of that level of lucid thought must be worth seeing; and he did not disappoint.

Carl’s career seems to have been upper-eastside-centric. A quick perusal of his papers shows that in the late 70’s he was at Rockefeller and publishing papers with Zanvil Cohn, the man who probably would have shared the Nobel Prize with Ralph Steinmann for the discovery of dendritic cells had he lived long enough. One of the areas of research on Cohn/Dubos laboratory was how tuberculosis interacts with monocytes/macrophages and although i don’t know if Nathan worked on these problems himself, his knowledge of the literature on the mycobacterial lifecycle seems to be comprehensive enough that it wouldn’t surprise me if he had.

Nathans talk revolved around the idea of bacterial “persisters.” He noted that persisters can generally fall into three classes - resisters with heritable resistance, acquired resistance and phenotypic resistance. In the first two cases we can think of as classic bacterial resistance to antibiotics due to a genetic resistance mechanism while the last class can be broadly induced. He pointed out a paper from 1944 in which it was first realized that there are several conditions in which reversible resistance to penicillin can be induced. This “phenotypic” resistance is now knowns, he stated, to be induced through a number of mechanisms like nutrient starvation (C/N/P?), acidification, sublethal concentrations of antibiotics. Bacteria that are rendered resistant to penicillin via these mechanisms can be replated and become susceptible. It was hypothesized that the cells that were initially resistance were not dividing at the time of antibiotic administration, a hypothesis that has ben shown to be true, for example, here. So Nathan led us to this idea of non-replicating persistence and I was sure he was going to talk about stringent response and ppGpp and how you could interfere/block the process. Instead, however, he presented the following thought: dividing cells can be killed by antibiotics but non-dividing cells cannot. Could it be that we have only found bactericidal compounds for growing cells because our screens are all designed to pick those up? It is an intriguing thought and led us into his research to identify and attach such targets.

So how does one go about finding these new targets? Using a tandem approach: use whole cell screens (transposon strains) to identify mutants that are compromised in the their ability to grow in stressed conditions. Simultaneously, you must pursue classic target-based screening against high-value cellular targets obtained by this, or other, means. From Nathan’s presentation, it looks as if he has been rather successful: he showed at least two classes of compounds of small molecules targeting the mycobacterial proteosome. Apparently a proteosome-like protein lit up in one of their phenotypic screens leading to the idea that targeting protein homeostasis could be a good way to try to kill these persisters. Of the two drugs he showed, both were relatively simple and one would have thought theres not much to get specificity with. It turns out that there are small differences in the mycobacterial proteosome binding pocket that preclude water from the binding pocket. The reaction intermediate can be atacked by water or by an amide on the protein. While water is more favored in the human proteosome, the lack of water allows this relatively simple chemical to have a ~1000x specificity for TB. Will it work in the clinic? I dunno.

Nathan showed us another TB-active proteosome inhibitor with a similar specificity story and seemed keen to show us that there is vertically integrated story for identifying pathways on required for bacterial persistence, and then screening and med-cheming the targets until a potent and selective compound is found. In his final minutes he showed the outline of a screen using sub-lethal concentrations of Kanamycin to induce protein disrepair. Apparently irreparably oxidized proteins get sequestered in the cell and can be detected by excess carbonyl (how do they detect: dye? Raman spectroscopy?). Apparently there are a number of pathways governing what to do with this stuff: supposedly it is asymmetrically parsed out to daughter cells (“cellular altruism”) allowing one of the daughter cells to divide when replication conditions arise once more. But like human cells, when sufficient damage is obtained, there is a self-imposed death (or at least death). So Nathan’s group is looking for proteins regulating the repair pathways and looking for potential targets of the homestatic system. A great idea and i wish him luck.

In his FAQ, Jeffrey Ravitch asked him about the role of the innate immune system in fighting TB. This was the sort of soft lob that allows him to speculate widely about TB’s evolution/co-evolution with the human immune system and you can see here the early influence of the Cohn/Dubos school. What are the cells that come around and detect foreign stuff? That scoop up these intruders and send the alarm: how does a pathogen evolve to handle them? Nathan’s answer was that mycobacterium has a strategy that absolutely depends on antagonizing innate immunity in order to thwart it. Supposedly the most invariant parts of all sequenced mycobacterial strains are the regions that are in the most immunogenic portions. Mycobacterial strains evolve ALL genes faster than those bits that macrophages recognize. So they recruit macrophages to hide out inside them, but they also need the macrophages to destroy the lung tissue and make the lesions in which they can camp out and live comfortably and slowly over long periods of time, until an inflamed-lung-induced cough send it flying into the next victim’s lung. The long view of an infectious foe that has tuned its entier existence to gettig around human immunity is a somewhat chilling thought and perhaps that scary thought is what keeps the already successful Nathan working at full tilt.