PF-2341066 and a Promise
I've been kinda busy, and tired. Haven't been 'round much of anywhere much.
But for today, a story inspired by events in June, and finally finished here. A story of progress and promise...
The first cancer gene took almost forty years to move from discovery to cancer cure.
One of the most recent, took less than forty *months*.
It was in 1960 that U. Penn scientists identified what would become known as the Philadelphia Chromosome, a mutant chromosome responsible for many cases of Chronic Myelogenous Leukemia. It contributes to nearly 20% of all adult leukemias, and thousands of deaths a year.
It was a pivotal discovery, especially given the primative scientific tools and knowledge of that time. After all, when the Philadelphia Chromosome was first discovered, not only did scientists have no idea why it was connected to cancer, scientists had no idea how chromosomes controlled anything at all. Scientists at that time hadn't even discovered *genes* yet. The first key step on that road wouldn't even be made until the following year, when scientists at the NIH in Bethesda figured out the triplet-codon system by which a gene encodes the structure of it's protein. Scientists at that time not only lacked the tools to figure out what the Philadelphia Chromosome did, they didn't even know what they should be looking *for*.
Decades of further study was necessary to learn what genes were and how they controlled the cell. To invent the tools needed to sequence DNA, identify genes, study the proteins that those genes made. It was like the situation faced by the first sea explorers, who had to invent everything from shipbuilding to navigation before they could start crossing the oceans...
It took another twenty years before scientists had enough knowledge to be able to zero in on what was actually happening. To handwave furiously, the Philadelphia Chromosome is formed when the bottom chunks of chromosome 9 and chromosome 22 break off and swap places. As it happens, the fracture line on chromosome 9 happens to cross through the middle of the Abl gene. The fracture line on chromosome 22 happens to cross through the middle of the BCR gene. When the broken bits swap places, the front half of the BCR gene gets stuck onto the back half of the Abl gene, which results in the body making a new, mutant half-BCR/half-ABL protein.
The problem is that both the normal BCR protein and the normal ABL protein are cellular control switches, part of the web of enzyme switches which control cell growth. If the cell is a giant clockwork robot, BCR and ABL are two of the gearshift boxes controlling it. Now imagine breaking each of those gearshift boxes in half and then jamming the broken halves together, so that the top half of the BCR gearbox suddenly finds itself driving the bottom half of the ABL gearbox. The levers and drive shafts which originally powered the BCR gearbox, now find themselves turning everything connected to the ABL gearbox instead. Needless to say, this leads to a lot of very screwed-up signals that drive the whole clockwork berzerk. In the case of the BCR/Abl mutant cellular control switch, it gets turned on in immune cells. It gets turned on and *stays* on, jamming those immune cells into uncontrolled growth -- and leads to the uncontrolled growth known as cancer.
BCR/Abl is the driving force behind many cases of Chronic Myelogenous Leukemia. It also, scientists realized, could be CML's Achilles heel. After all, the BCR/Abl fusion protein didn't normally exist, and the body usually operated completely happily without it. So if you could find a way to shut the mutant, berzerk BCR/Abl protein down, you wouldn't harm anything else at the same time. You'd have the perfect drug for CML -- a drug which could selectively shut down CML tumor cells, yet be ignored by the entire rest of the body.
Of course, before you could do that, you had to figure out *how* the BCR/Abl protein actually worked. Figure out how to design (or search for) a compound that would shut the protein down. Figure out how to *manufacture* those kinds of compounds in industrial amounts. All of that knowledge and technology in a dozen related fields had to be discovered or invented. Which is why it took almost fifteen years to get from the Philadelphia chromosome to the discovery of the BCR/Abl gene, and another fifteen years between the discovery of the BCR/Abl gene, and the first clinical trials of a drug targeted against it. The drug was known as STI 571. Later known as Gleevec. In the early 2000s, the first 500 patients with deadly CML were given Gleevec.
Almost a decade later, more than eight out of ten of them are still alive.
It wasn't just survival by which Gleevec succeeded -- although it was a significant improvement over the previous state-of-the-art. The promise of Gleevec, after all, was that, by specifically targeting the mutant BCR/ABL protein specific to CML, it would stop CML without stopping everything else -- and in that too, Gleevec succeeded.
Chemotherapy had always been based on the idea of targeting cancer cells as tightly as possible. The problem had always been that, before the molecular biology revolution, physicians didn't really have a lot of ways of doing that. Before the molecular biology revolution, the best physicians could do was to use drugs that targeted fast-growing cells. This by definition included cancer cells, which killed by growing out of control. But this *also* would target every other fast growing cell in the body. Like the cells that lined one's mouth and gut, leading to nausea and vomiting and painful sores. The cells that grow hair, causing the prototypical baldness seen in cancer patients. And the same corrosively powerful chemotherapy would damage many other cells as well.
Many cancer therapies would literally make you sterile, damage your heart, fry your kidneys and liver, and even *cause* other forms of cancer ten years down the line. Of course, without treatment, you wouldn't *have* ten years down the line -- or even ten weeks -- so physicians did the best they could. That best was certainly far better than nothing. And even before molecularly targeted therapies, oncologists had already identified many other ways cancer cells differed from normal cells, and developed therapies to take advantage of that.
The ultimate goal, however, had always been to *specifically* target cancer cells. To develop drugs that could truly kill a tumor cell and *just* a tumor cell. BCR/Abl was one of the best opportunities to do just that. At the end of forty years of effort, Gleevec was proof that it could be done. But Gleevec was always meant as only the beginning. After all, the whole idea was that the techniques, tactics, and knowledge assembled on the long road from the Philadelphia Chromosome to Gleevec, could be applied to other cancers -- and other diseases -- as well.
The human body is an enormously complex puzzle made of a vast number of pieces -- but a *finite* number of pieces. Often, the same pieces, the same controls, are employed by many different organs and systems. Which means that learning about the biology behind one disease, suddenly teaches us important things about a seemingly totally different disease that happens to use the same mechanism for a different purpose. And that a drug that was originally designed to target one diseases' control system, suddenly becomes useful in a totally different disease using the same control system.
Technology has advanced substantially since the Philadelphia Chromosome and BCR/Abl were first studied. The completion of the human genome project gave scientists everywhere a full map of the normal human genome, giving us a normal template against which we can compare the mutant genomes of cancer cells. Today's gene sequencing engines, built by private corporations based on publically funded basic science, can do in days what once took entire armies of scientists months and years. Similarly, the tools to do everything else -- from studying protein function to drug design and screening -- have improved by orders of magnitude. And most importantly, we *know* vastly more than we did back then. Have learned and mapped and probed the function of hundreds of genes and systems throughout the human body. Pubmed, the federally supported database, now contains over twenty million scientific articles. Two of which, published just months apart, kicked off a revolution in lung cancer...
In contrast to the progress made against leukemia, other cancers remain far more intractable. Lung cancer is one of them. In part, it is because lung cancer is a far more complex problem. In contrast to leukemia, where the tumor arises from one single cell gone bad, most lung cancer is born in lungs warped after a lifetime's exposure to mutagenic cigarette smoke into a giant, pre-cancerous mass. There isn't just *one* mutation in there -- there's potentially dozens, all just waiting to be tipped over the line. Even if you target one mutation, there's plenty of others waiting, from all the genetic damage caused by constant exposure to tobacco smoke. Not to mention the other damage caused by cigarette smoking beyond just mutagenesis -- reduced lung function, reduced heart function, damaged blood vessels, even *other* cancers in mouth and throat. All of which limit the room within which oncologists can work, and limit how much chemotherapy a lung cancer patient can even tolerate. All of which contribute to the far higher fatality rate in lung cancer in general.
There are, however, a subset of lung cancers which *aren't* associated with smoking. Lung cancers which arise in people who have never smoked a ciagrette in their lives. *These* lung cancers are more similar to CML, for example, in that there's likely a single series of responsible genetic changes, rather than an entire lung mutated by tobacco smoke in dozens of different ways. Lung cancer in smokers might have dozens of separate genetic causes. Lung cancer in *non*-smokers usually only has one. It's the difference between finding out what's wrong with a car that has been shot once with a laser, vs. what's wrong with a car that has gone through a Mad-Max demolition derby and then tumbled off the top of Mt. Everest. Lung cancer in smokers and non-smokers are both important medical questions; but the latter is the easier scientific problem to solve. And in August 2007, a Japanese group discovered EML4-Alk...
By comparing the genomes of non-smokers with one type of lung cancer to the master human genome, they discovered that a subset of their lung cancer patients had a mutant fusion protein called EML4-Alk. Like BCR/Abl, these patients had a chromosome rearrangement which fused part of the ELM4 gene to part of the ALK gene. So in theory, one could use the same tactics used for BCR/ABL to try to help patients with EML4/ALK. Especially important, as mean survival time for these particular lung cancer patients was measured in months to weeks, even with the best chemotherapy, surgery, and radiation therapy known.
Thanks to the massive advances in technology since BCR/ABL was discovered, the Japanese project identifying BCR/ABL took months rather than years, and was the work of a few scientists rather than an army. Further, they didn't have to figure out what ALK did -- that had already been discovered by other groups. ALK was already known as control gene in lymphomas. Had already been studied extensively and -- most critically -- Pfizer had already made a drug that could block ALK activity.
Actually, Pfizer had originally designed PF-2341066 to target the control protein c-Met. But PF-2341066 was *also* discovered to block the control protein ALK. The paper announcing PF-2341066's ability to block ALK kinase was published in May of 2007, and oncologists were already preparing to test PF-2341066 in ALK-containing lymphomas. Suddenly, three months later, with the finding that there were *lung* cancers that *also* carried ALK mutants, oncologists immediately connected the dots. An international clinical trial was underway just weeks later, using the new EMK4-ALK gene tests to look for lung cancer patients with the EMK4-ALK mutation, and then taking the new experimental ALK lymphoma drug and giving it to dying EMK4-ALK lung cancer patients who had no other hope, to see if the new drug could work one last miracle.
Just a year and a half later, just a few weeks ago, the lead oncologists announced the result of their trials. It worked. Spectacuarly.
In the initial cohort of fifty patients who had already failed all other therapy, suddenly nine out of ten patients saw their tumors completely halted with *just* crizotinib, as PF-2341066 is now known. More than half saw their tumors actually *shrink* by more than 30%. Side effects were mostly limted to nausea, vomiting and diarrhea -- a hell of a lot better than other chemotherapy agents for lung cancer, which didn't even work anyway. And to top it all off, crizotinib comes in a pill. You don't even need an IV to take it.
Just two years ago, every one of these patients would have died within weeks to months. Now, for most of them, one single pill has halted the cancer right in its tracks. Put the cancer on hold and allowed them to move on with their lives. How long will the tumor stay on hold? Only time will tell. But every week and month is a week and month and year won from the Black Racer. And for all we know, the tumor might *stay* on hold. After all, most of the original Gleevec patients are still going, ten years later.
This is just one of many such stories. Like Herceptin and breast cancer. Or Avastin and colon cancer. Or like UCN-01 or HA-22 or all the other experimental drugs I work with every day on the oncology services here at NIH or at Hopkins. And we're only just beginning.
Physicians and scientists all over the world are working on turning the same trick on dozens of other kinds of cancer. And on countless other medical problems as well. Identify the genes, figure out how they work, find a drug that will fix the problem, on everything from depression to multiple sclerosis to arthritis and so on. More and more often, we discover that the gene we find is one we already knew about from somewhere else. Or already has a drug we can borrow from somewhere else. The entire point of a physician-scientist is someone who can turn things from bedside to bench and back again. These days, we're able to do that faster than ever before. And on the horizon are even more profound achivements, the future hinted at in Stem Cell Saltarello and A Tale of Chrétien.
For generations, we as physicians and scientists have held a trust from all of you. Roosevelt created the federally-funded NIH. Successor presidents expanded it. The Clinton Administration built it into the largest scientific enterprise in human history. Funded the completion of the public human genome project, placing the sequence of mankind's DNA in public, not corporate, hands. Doubled the amount of money available to fund the basic science for-profit companies had little interest in, and that even the largest private philanthropies couldn't afford on their own. Vastly expanded training programs for scientists and physicians alike. The massive momentum Clinton touched off carried on even while George W Bush's administration largely failed to keep the NIH budget up with cost of living increases, and federal scientists butted heads with restrictions imposed by his religious conservative appointees.
The promise behind those generations of investment in us, from Roosevelt to Obama, was that the more we learned, the faster cures would come.
The promise we made to the parents of the children we fail, is that what we learned might help others.
With your help, those were the promises we made.
With your help, those are the promises we keep.