In which I disagree with a Nobel Prize winner...

This last week my MRU had the privilege in hearing a talk by Dr. J. Michael Bishop entitled "The Cancer Genome of Therapeutics." It was an interesting talk and Dr. Bishop told several good stories, which isn't that surprising since I expect you don't get a Nobel if you can't give a good talk. However, just because you give an interesting, engaging, and thought provoking talk, doesn't mean I have to agree with you.


The thrust of the talk was based on the idea that by comparing cancer cells with "normal" cells, we can identify things that are different. Once differences are identified, these are now targets for therapeutic interventions. Now I want to be clear for those not knee-deep into the biological sciences, when I say "targets for therapeutic development," I am not suggesting that therapies exist. This is the 4000 lb. gorilla in the room no one likes to talk about in many scientific areas. Just because you identify a target, does not mean you have any way of hitting it. Two thousand years ago if a hungry hunter saw a fat elk on the other side of a 500 foot gorge (the target to solve the problem of hunger), there was nothing the hunter could do about it. Even if said hunter had a bow and arrow or more likely an atlatl, the elk is still a useless target, since the dead elk would be eaten by wolves, lions, and other animals long before the hunter could climb down the gorge and back up the other side. So having a target does not necessarily mean much.


One way in which these target identification approaches is done is to identify genes that are expressed in one cell type but not another, such as expressed in cancer cells but not in normal cells. This is usually the direction these approaches work too, we look for things that are expressed in the undesired cell, not things that are absent. (Its easier, but not impossible, to target something that's there not something that's missing.) Also, we generally go for genes and not proteins because it is currently much easier to determine what genes are or are not expressed over essentially the entire genome than it is to determine all the proteins expressed in a given cell population. This means our measure for expressed targets is somewhat indirect.


So lets say we find 12 genes are expressed in a specific kind of cancer cell, like breast cancer cells, but no in normal breast tissue. (The reality is many more than 12, but let's keep the numbers small.) Do these 12 genes specify 12 new targets? Well, the short answer is no. See these 12 genes are found in the genome because human beings probably need this gene for something other than causing uncontrolled cell growth in the breast. This is a huge limitation to this kind of approach. Just because you identify an expressed gene specific in a cancer compared to the otherwise normal tissue, does not mean and almost certainly doesn't mean the gene is not normally expressed somewhere. So your newly discovered cancer target may also be a pancreas development target or bacterial combating lymphocyte target too.


This is where I was left wanting. The nobel winning scientist begins their talk by establishing the overarching theme: by comparing cancer and normal cells/tissues, we will identify new targets based on these differences, and begin curing cancers at a previously unknown rate. Sadly, the first thing that came to my mind was all those microbes that have been killing us for generations and are still pretty damn good at it. Bacteria are about as different from us as you can get and you know what, we really have no new ways of combatting them. They kill more of us than cancer, but pharmaceutical companies have been shutting down their anti-microbial divisions to the point few actually exist anymore. (Don't blame the pharmaceutical companies, which can make a ton more money making guys hard, women skinny, and kids easier to oversee in factories schools.) Bacteria, fungi, protozoan parasites. All are extremely different from us at least when compared to cancer cells which are essentially clones of all your other cells.


So I am skeptical that knowing all the differences in expression between cancer cells and normal cells will pay dividends in any rapid way at least not as sold. However, I am not against this approach scientifically (I am against how it is sold to the public though). This will definitely tell us much about cancer biology, it will reveal commonalities and distinctions between different cancers, it may reveal genetic risk factors in patient populations that could impact screening and lifestyle choices (think about the current mammogram controversy), it may also lead to new treatments just not in the one gene = one target paradigm.


There are two ways I can envision genomics leading to targets that are susceptible to therapeutic intervention.


1. Often cancer progression is associated with chromosomal rearrangements. One chromosome recombines with another making a fusion chromosome not found in normal cells. A recombinant chromosome is not necessarily a bad thing and normal cells frequently contain them. However, the recombination can lead to the generation a protein that would never normally be generated. If the recombination occurs in the middle of two distinct genes a fusion gene can be created. One famous case of this is Bcr-Abl, which is associated with certain leukemias. Abl and Bcr are both kinases, although the specific function of Bcr is still not clear. The Bcr-Abl fusion removes an inhibitory domain of Abl, which leads to hyperactive Abl and that is oncogenic (cancer promoting). These types of rearrangements can be detected using new deep sequencing genomic approaches. 


2. Cancer is complicated and not due to a single cellular defect. Cancer requires numerous genetic changes. What we often see is that for a given type of cancer a similar set of cellular pathways act differently, although in the same way in the cancer. Much like specific targets, like a protein, a pathway can be the target for a therapeutic intervention. Actually, pathways are much larger targets since a pathway can be targeted by disrupting any of the proteins that make up the pathway. As before these pathways exist in normal cells as well. However, we could target two or three different pathways that are hyperactive in cancer cells with different therapeutics, which could kill the cancer cells. A normal cell may require one or two of these pathways, but not all three and thus would be 'immune' to the treatment. Admittedly there are a lot of ifs associated with this approach, but it is a viable approach. In fact, this is the approach that has made HAART so successful in treating many HIV infections.


I don't mean to be all gloom and doom, but the scientific community has gotten a fair bit of well-deserved blow back from overstating the impact of our studies to the lay public. Also, the fact that a given approach may not lead to a life-altering new product does not mean it is not worthwhile. If we look back through the history of science, many of the biggest advances were not front page news at the time. One ready example is the initial identification of penicillin by Fleming occurred in 1928 (earlier reports existed but this was the one that stuck), however it took 11 years and another group (Chain and Florey) to purify it for use as an antibiotic. This is a rapid turn around time and probably a poor example, because penicillin changed our lives. None of us were alive when death from bacterial infection was common and normal. People did not generally die of cancer or heart disease. The splinter in your finger you got chopping wood, that could kill you, if you got the wrong bacterium in the wound. Of the ~350,000 Union soldiers that died in the US Civil War, ~220,000 died of disease. Or pre-penicillin, in WWI 16.5/1000 soldiers died of disease/year whereas post-penicillin, in WWII 0.6/1000 soldiers died of disease/year¹. The take away point is that new treatments take time and are not often clear from the initial findings. The second point is that we'll probably never see another medical intervention that has the same societal effects as the antibiotic generation.