I’m confused. Read this and see if you end up the same way. TechCrunch has the story of Isaac Yonemoto, who’s crowdsourcing a project around a potential oncology compound. It’s a derivative of sibiromycin, a compound I hadn’t come across, but it seems that it was first studied in Russia, and then at Maryland. Yonemoto’s own work on the compound is in this paper from 2012, which looks reasonable. (Here’s more). And the crowdfunding pitch is also reasonable, in lay-audience terms:
The drug candidate 9DS was developed at the University of Maryland. The last work done on the drug showed that it had activity against cancer competitive with leading cancer drugs such as taxol. Moreover, 9DS is also likely to have lower side effects than most chemotherapies, since a related compound, SJG-136, seems to have low side effects in early clinical trials.
Project Marilyn involves: production of more 9DS, and submitting 9DS to a xenograft study (‘curing cancer in mice’). This is the next step in drug development and an important one on the way to doing clinical (human) studies. The process we’re seeking to fund should take approximately 6 months. If we recieve more funding, we will add stretch goals, such as further preclinical experiments on 9DS, development 9DS analogs, or other exciting anti-cancer ideas.
But here’s where things begin to swerve off into different territory. Yonemoto isn’t just talking about some preclinical spadework on yet another oncology compound (which is what the project actually is, as far as I can tell). He’s pitching it in broader terms:
. . .Some drugs can cost upwards of $100,000 a year, bankrupting patients. This level of expense is simply unacceptable, especially since 1/3 of people will get cancer in their lifetime.
One solution to this problem is to develop unpatented drugs – pharmaceutical companies will have to sell them at a reasonable price. To those who believe that drugs cannot be made without patents we remind them:
When Salk and Sabin cured polio, they didn’t patent the vaccine. It’s time to develop a patent-free anticancer drug for the 21st century.
The software industry and the open-source movement have shown that patenting is not necessary for innovation. Releasing without a patent means the drugs will be cheaper and it will be easier to build on the work to make improved drugs or drug combinations. Releasing without a patent means expanded access to drugs in countries that can’t afford extensive licensing and export agreements.
OK, let’s take this one apart, piece by piece, in good old classic blogging style. Yes, some oncology drugs are indeed very expensive. This is more of a problem for insurance companies and governments, since they’re paying nearly all of these costs, but the topic of drug prices in oncology has come up around here many times, and will do so again. It’s especially worrisome for me that companies are already up close to the what-the-market-will-possibly-bear price with things that are not exactly transformative therapies (what pricing structure will those have?)
But are unpatented drugs the solution? It seems to me that pharmaceutical companies will not “have to sell them at a reasonable price”. Rather, unpatented compounds will simply not become drugs. Yonemoto, like so many others who have not actually done drug development, is skipping over the longest, most difficult, and most expensive parts of the process. Readers of the crowdsourcing proposal might be forgiven if they don’t pick up on this, but getting a compound to work in some mouse xenograft models does not turn it into a drug. Preparing a compound to go into human trials takes a lot more than that: a reliable scale-up route to the compound itself, toxicology studies, more detailed pharmacokinetic studies, formulation studies. This can’t be done by a handful of people: a handful of people don’t have the resources and expertise. And that’s just setting the stage for the real thing: clinical trials in humans. That crowdsourcing proposal skates over it, big-time, but the truth is that the great majority of money in drug development is spent in the clinic. The amount of money Yonemoto is raising, which is appropriate for the studies he’s planning, is a roundoff error in the calculations for a decent clinical campaign.
So who’s going to do all that? A drug company. Are they going to take that on with an unpatented compound that they do not own? They are not. Another thing that a lay reader won’t get from reading Yonemoto’s proposal is that the failure rate for new oncology compounds in the clinic is at least 90%, and probably more like 95. If you are going to spend all that money developing compounds that don’t make it, you will need to make some money when one of them finally does. If a compound has no chance of ever doing that, no one’s even going to go down that road to start with.
Now we get to the Salk/Sabin patent example. There are plenty of persistent myths about the polio vaccine story (this book review at Technology Review is a good intro to the subject). Jonas Salk created one of the most enduring myths when he famously told Edward R. Murrow in an interview that “There is no patent. Would you patent the sun?”. But the idea of patenting his injected, killed-virus vaccine had already been looked into, and lawyers had determined that any application would be invalidated by prior art. (Salk himself, in his late work on a possible HIV vaccine, did indeed file patent applications).
Sabin’s oral attenuated-virus vaccine, on the other hand, was indeed deliberately never patented. But this does not shed much light on the patenting of drugs for cancer. The Sabin polio vaccine protected all comers after a single dose. The public health implications of a polio vaccine were obvious and immediate: polio was everywhere, and anyone could get it. But Yonemoto’s 9SDS is not in that category: cancer is not a single disease like polio, and is not open to a single cure. Even if a sibiromycin derivative makes it to market (and they’ve been the subject of research for quite a while now), it will do what almost every other cancer drug does: help some people, to a degree, for a while. The exceptions are rare: patients who have a tumor type that is completely dependent on a particular mechanism, and that doesn’t mutate away from that phenotype quickly enough. Most cancer patients aren’t that fortunate.
So here’s the rough part of cancer drug discovery: cancer, broadly speaking, is indeed a big public health issue. But we’re not going to wipe it out the way the polio and smallpox vaccines wiped out their homogeneous diseases. Cancer isn’t caused by a human-specific infectious agent that we can eliminate from the world. It crops up over and over again as our cells divide, in thousands of forms, and fighting it is going to take tremendous diagnostic skill and an array of hundreds of different therapies, most of which we haven’t discovered yet. And money. Lots of money.
So when Yonemoto says that “The software industry and the open-source movement have shown that patenting is not necessary for innovation”, he’s comparing apples and iguanas. Drug discovery is not like coding, unfortunately: you’re not going to have one person from San Jose pop up and add a chlorine atom to the molecule while another guy pulls an all-nighter in St. Louis and figures out the i.v. formulation for the rat tox experiments. The pitch on Indysci.org, which is really about doing some preliminary experiments, makes it sound like the opening trumpet of a drug discovery revolution and that it’s going to lead to “releasing” a drug. That’s disingenuous, to say the least. I wish Yonemoto luck, actually, but I think he’s going to be running into some very high-density reality pretty soon.
Update: Yonemoto has added this to the comments section, and I appreciate him coming by:
“Thanks Derek! You’ve basically crystallized all of my insecurities about the future of open-source drugs. But that’s okay. I think there are business models wherein you can get this to work, even under the relatively onerous contemporary FDA burden. To answer a few questions. I think sibiromycin is not a bad candidate for several reasons: 1. (I’m not sure I buy this one but) it’s a NP derived and NP derived tends to do well. 2. A molecule with a similar mechanism has made it into phase III and phase I/II show only mild hepatotoxicity and water retention, which are prophylactically treatable with common drugs. 3. There is reportedly no bone marrow suppression in these compounds, and importantly it appears to be immune-neutral, which would make PBDs excellent therapies to run alongside immune-recruitment drugs.”