Another promising obesity drug bites the dust…but at least we know it. Ruminations on pharmaceutical R&D knowledge transfer.
Nature News story, a well-written example of similar stories to hit the wires this week, highlights the clinical failure of Merck’s NPY antagonist MK-0557, a once-promising treatment for obesity. The clinical study referenced in the story was published this week in Cell Metabolism (subscription req).
The usual spin on this MK-0557 story is that this failure marks another in a long line of clinical failures for anti-obesity medications. Or that for weight-loss drugs to have a good chance of success they’ll have to be tested and used in practice in combinations from the outset. Yeah, we know. So, what else is new?
Maybe you’ll read the investment angle–how the clinical failure of MK-0557 is another late-phase failure that Merck can ill-afford, as it tries to re-establish itself as a major to count on for innovation. Or how about the tried and true–that pharma needs better preclinical models of obesity (not to mention nearly every other chronic disease) if it ever hopes to make major gains over current therapies.
Yawn.
There’s gotta be a more interesting angle to this drug’s failure. Maybe something along the lines of my recent rant about publishing negative studies and the recently launched Journal of Negative Results in Biomedicine.
Well, for starters there was the fact that this paper is likely going to be missed by many clinicians who don’t do much basic research, and many that do. Cell Metabolism isn’t exactly a top-20 clinical research journal. And this was, after all, an important, 52-week clinical efficacy and safety study that enrolled over 1600 overweight or obese subjects. But I’m nitpicking. Those who are interested in obesity drugs will find this article. And the truth is I don’t know whether Merck wanted the article to pass into publication oblivion or whether Merck actively supported its prompt publication in a journal that the outside authors’ sanctioned. What’s important is that it was published somewhere, and it was published in a reasonable time frame relative to the actual study conduct.
Yawn.
There’s gotta be something. I turned to the literature to check on the clinical-study publications for other investigational drugs in this class. Not recalling any off the top of my head. Maybe there’s something important to be said about Pharma’s recent record of publishing investigational drug studies in the peer-reviewed literature? Now that online clinical-trial data registries are all the rage, are more or less pre-market clinical studies being published in peer-reviewed journals? That’s an interesting angle, and NPY should make a great test case. The target (targets actually; there are at least five NPY receptor subtypes and some are found outside the CNS) has been around for a long time. Many drugs have been synthesized around it (them). I was just having trouble remembering some of these clinical-study publications, so I turned to PubMed to help my failing memory.
Neuropeptide Y or its abbreviated version “NPY” turns up nearly 10,000 hits on PubMed (Medline is the primary database for PubMed). An impressive 1,128 are review articles. The earliest to refer to the neuropeptide was published in Nature in April, 1982: “Neuropeptide Y–a novel brain peptide with structural similarities to peptide YY and pancreatic polypeptide.” The complete amino acid sequence was published in PNAS later that year (Sept, 1982). Great, you might think, nearly 25 years since the endogenous peptide was discovered. Should be tons of clinical studies here.
So, let’s get right to those clinical publications. Let’s see…limiting the search by publication types ‘clinical trial’ or ‘randomized controlled trial’ yields 116 pubs. A decent number, but I’ve been fooled by these PubMed delimiters before. Sometimes they miss stuff. Using my own, custom, homemade, proprietary, one-of-a-kind, special-order delimiters for clinical studies yields 341 publications [apologies to George Carlin for stealing his bit].
Much better. If these 341 studies are all obesity treatments targeting NPY that should certainly be enough to examine recent publication trends. So, I limit further, to articles including the words obesity or overweight. Uh oh. I’m down to just 27 articles, and the first two are the Cell Metabolism article and the accompanying editorial, leaving 25.
Here’s the list (NPY Clinical Studies in Obesity).
I’ll save you the trouble of reading it. Not a single clinical study of an NPY receptor antagonist, other than the one of Merck’s MK-0557, appears in the peer-reviewed literature. Okay, I lied a little before. It’s not that my memory was failing me. I pretty much knew before I started the PubMed search that it would come out this way. Not a single clinical study published.
You’re probably thinking that I’m now going to rant about how pharma should be publishing their clinical studies of investigational drugs in Phases 1 and 2. Nope. In fact, I’m happy to argue why pharma should not feel compelled to do so. My argument would depend heavily on the premise that the benefits of peer-reviewed publication of investigational drug studies in early phases accrue far more to publishers and academics that must publish or perish than they do to the industry and to the patients we all serve. As a drug is advanced towards the marketplace, obligations for industry to publish become increasingly weighty, as the benefits of publishing become more evenly distributed, and the risks of not publishing have an increasingly real chance of causing some indirect harm to patients.
Despite industry’s lack of obligation to publish early clinical-phase work, pharmaceutical firms do have a strong interest in advancing the state of the art as quickly as feasible, without risking their intellectual property or their market position. The patent system provides a mechanism to transfer knowledge throughout the scientific community, but there is a significant time lag in this transfer, and many experimental details, including findings from clinical studies, are not usually needed to support patents and therefore never find their way into this literature.
Short of the peer-reviewed and patent literatures, there are other mechanisms being used to transfer knowledge among industry scientists and from industry to the public. In order of decreasing efficiency of transfer these are trial registries, scientific meetings, investor-oriented consultations, voluntary and “mandatory” research disclosures from industry to investors, and informal discussions among scientists, within and outside of industry (including transfers of knowledge that accompany transfers of scientists when they switch employers).
Registries. So far, industry has strongly resisted disclosing key early-phase study details in clinical-trial registries, so while they’re potentially efficient sources of knowledge transfer, they’re not effective sources today. Scientific meetings. Besides the formal poster and oral presentations occurring at such meetings, there’s a significant amount of knowledge transfer occurring during the chatter taking place among scientists, including chatter not sanctioned by industry. Investor-organized consultations. Professional investors are “hooked up” with clinical investigators via intermediaries, such Gerson-Lehrman Group and Leerink-Swann’s MedaCorp. Although the practice of such informal meetings have come under increased scrutiny by regulators, no outside parties police these interactions, least of all industry representatives. As a result, a good bit of proprietary investigational-drug information circulates in loops between scientists and investment analysts. Company disclosures. Small public companies have to disclose more, because more of their early-phase results are considered material information. But, as you can see, few of the companies granted drug patents were small, and only one (Neurogen) was small and publicly funded. Big companies disclose early-phase information only when they believe it will benefit them or their investors. While such voluntary information disclosures serve as a source of “competitive intelligence,” most astute industry scientists recognize them for the propaganda they are. As a result, rather than contributing to advancement of the state of the science, industry-sponsored, voluntary information disclosures to investors at research days and the like will tend to slow down scientific advancement by gumming up knowledge flows with information detritus.
I’d love to be able to trace the network of knowledge transfer flows that propelled forward clinical development efforts in the class of NPY antagonists. I can’t, because I’ve not been privy to the many behind-the-scenes discussions that undoubtedly played some role. What I can try to do, though, is deduce some sense of the importance of these hidden flows of knowledge by examining the plainly visible sources of knowledge transfer.
Elements of the pharmaceutical industry have been interested in NPY ever since its powerful orexigenic (appetite-increasing) effects were first demonstrated (circa 1984). But the real excitement began after 1996 with the discovery of a subtype of NPY receptor, Y5, responsible for mediating NPY effects on eating behavior (see the article here). Any NPY antagonists discovered before 1996 were found using screens that were non-selective for the orexigenic effects of the endogenous peptide.
As far as I can tell, the first US patent for a drug antagonizing NPY was granted in 1996 to Eli Lilly (5,567,714). Lilly was granted patents for other novel NPY antagonists in 1997, 1998, 1999, 2001 and 2003. Other recipients of U.S. patents for inhibitors of NPY receptors include: Pfizer (with or without contribution from Neurogen) in 1999, 2000 (2 different chemical series), 2001, 2002 (5 different chemical series), 2003 (3 different chemical series), 2004 and 2006; Bayer in 1999 (2 different chemical series), 2000 and 2006; Banyu and/or Merck in 2000 (3 different chemical series), 2001 (5 different chemical series) and 2003 (2 different chemical series); Ortho-McNeil in 2001 (2 different chemical series), 2002, 2003 (2 different chemical series), 2005 (2 different chemical series) and 2006; Synaptic (later partnered with Novartis) in 2001, 2002, 2003 (2 different chemical series), 2004 and 2006; Aventis in 2002; Wyeth in 2002; Servier in 2003; Schering-Plough in 2003, 2005 and 2006; AstraZeneca in 2005.
The above granted-patent list isn’t exhaustive, and does not include patent applications pending or withdrawn. So, for example, some of the known NPY Y5 antagonists, such as those from GSK aren’t included. But you get the idea. Industry has been both interested in and successful in finding drugs that inhibit NPY signaling beginning prior to the Y5 discovery and cloning. It’s also been successful more recently in finding drugs that selectively antagonize NPY at the Y5 receptor.
How many of these patented drugs made it to the clinic? This is a tough question to answer. I searched all of the sources mentioned above as well as I could without relying on second-hand sources (e.g. a reported rumor, or a scientific meeting report without a company-written abstract attached to it). This July, Shionogi disclosed results of its Phase 2a study of its NPY Y5 antagonist S-2367. And that’s it. I could find no other evidence of clinical trials of NPY antagonists disclosed other than the Shionogi study and the Merck study.
Perhaps these were the only two clinical studies conducted and reported out since the first patents were granted for NPY antagonists in 1996. It’s possible. However, I think it’s likely that quite a few clinical trials of NPY antagonists in overweight or obese patients have been conducted. If I’m correct, then a substantial amount of knowledge gained from prior clinical experience with NPY antagonists occurred in the world invisible to the researcher armed with nothing more than the internet’s primary literature.
Is this bad? One thing I’ve learned about innovation from other R&D-intensive industries is that the rate of knowledge transfer among R&D groups is directly linked to the rate of product innovation, all else being equal. If that lesson applies to the pharmaceutical industry, then industry practices that impede scientific knowledge transfer are anti-innovative. This says nothing about the value of such practices to the individual firm, but I would argue intuitively that market-protective practices such as knowledge-hoarding are diminishingly valuable to the firm the more innovative (i.e. challenging or expensive) the R&D efforts.
Generally speaking, policies protecting early-phase clinical studies as if they were museum pieces strike me as decidedly anachronistic. In today’s pharma R&D reality, where the most valuable fruit is hidden among the higher leaves and harvested by rapacious hordes, R&D managers charged with the responsibility of producing innovative and important therapies must do what they can to maximize their odds of success (by taking many calculated risks). Note the important difference between a calculated risk-taking strategy and a calculated risk-mitigation strategy. Locking knowledge in a vault is a risk-mitigation strategy. Sharing knowledge is a risk-taking strategy.
I’m getting very long-winded, but I want to wrap up with a few words about mechanisms for disseminating investigational drug knowledge. As you might have gathered, I believe that publishing early-phase trial findings in the peer-reviewed literature as it is currently administered is not an ideal solution to improving the efficiency of pharma R&D knowledge transfer. It’s simply too slow to be very helpful. Ditto scientific meetings; plus meetings have the added problem that their proceeding aren’t uniformly published. Simultaneous internet deposition and online review, as has been happening for years in the physical sciences, seems like a solution worthy of experimentation by pharmaceutical researchers. The first medical-oriented online publication of this type, PLoS One, is expected to go live soon. Another plausible solution is for industry to relax its objections to publishing early-phase (what PhRMA calls “exploratory”) clinical results in trial registries.
The mechanisms are in place, the question is which company will be first to experiment with an open-air research publications strategy. Whichever one does will understand that to win the innovation race you have to be willing to take risks and upend the status quo regularly.
