Faster and more open collaboration among scientists could yield a wealth of discoveries
Floor 3A of the venerable Banting Institute on College Street looks like any other biochemical lab in the world. Refrigerators hum, technicians titrate and researchers cogitate at workbenches separated by ceiling-high storage cabinets. Lab gear fills the shelves along the walls and there’s a continuous buzz of intellectual activity.
This is the Toronto home of the Structural Genomics Consortium, an international non-profit research group. As ordinary as it may seem, the consortium’s lab is actually the centre of what director and CEO Aled Edwards hopes will be a new way of doing the science of drug discovery.
Edwards, a 46-year-old professor of structural biology, sprawls on his office sofa, props sandal-shod feet on a coffee table, and notes that “there are dark days coming for the development of new medicines.” Only one drug candidate in 10 makes it through the hideously expensive clinical trials process and last year only 17 new medications received regulatory approval to be sold in the North American market. That “discouraging number” is the lowest in recent history, Edwards says.
A key barrier, he says, is the misuse of intellectual property claims. Edwards argues that too many scientists and corporations turn discoveries that should be available to everyone into their own secret possessions. This is where the consortium comes in.
With financing from industry, governments and trusts, the group aims to identify the three-dimensional structures of 2,400 key human proteins (as well as a few from human parasites) and then make them publicly available. Anyone could then use the protein structures for any purpose – even to create profit-making drugs.
The idea runs counter to the prevailing industrial ethos of patenting every discovery and innovation, no matter how minor, as soon as possible, if only to keep them away from the competition.
Comparing the two approaches, Edwards says opening up knowledge of human proteins to the world is “certainly the right way to go.”
For most of us, protein is what we need for a balanced diet, along with fruits and vegetables. But to a biomedical researcher, proteins are simply us. Most of the organic compounds in the human body are proteins (DNA is one vitally important exception) made up of combinations of simpler chemicals
called amino acids.
The estimated 22,000 human proteins carry messages between and within cells. They recognize and communicate with each other in complicated ways that are largely determined by their shape. The three-dimensional structure of a protein governs how it interacts with the rest of its biochemical world.
From the point of view of the pharmaceutical industry, proteins are targets for drug development. Choose the right protein, block it or enhance it properly, and you have a medication. But it’s not really that simple – witness the 90 per cent failure rate of potential drugs.
The current intellectual property model doesn’t make it any easier. For example, about 20 per cent of potential drugs fail because they turn out to be toxic in humans. A test that could predict toxicity would therefore be useful to a drug company and would probably increase its success rate.
But that “Toxotest” – as Edwards calls the hypothetical discovery – would be only a first version, an initial draft. The drug company that developed it would almost certainly keep it under wraps, preventing the cross-pollination of ideas that might revise it and improve it.
Under a model of open innovation, the test would be made freely available. Researchers all over the world would be encouraged to use it and, more importantly, to make it better – not for monetary reward but for public recognition and scientific kudos. The result, in the long run, would be that all companies developing medications would have a better chance of successfully navigating them through clinical trials.
Exactly why the drug development process is broken is complicated. Regulatory hurdles are high, the public’s tolerance for risk in medications is low and big pharmaceutical companies tend to be poorly organized. According to Edwards, however, the biggest problems are gaps in our knowledge of human physiology.
“At the end of the day, if you don’t know what’s going to happen when you put a molecule into a human body, it’s a lottery,” he says. The consortium aims to help scientists make better guesses about which drug candidates will work and which won’t.
In some respects, the idea of a more collaborative scientific method is similar to the open-source software movement. Such open innovation systems “demonstrate that greater value can be realized when firms selectively share their intellectual property,” says Don Tapscott, an adjunct professor at U of T’s Rotman School of Management.
With open software, a program’s code is made publicly available, allowing computer engineers anywhere to work on it, leading in theory to better software programs. The practice seems to bear out the theory – the popular Firefox web browser is an example of an open-source program that competes on roughly equal terms with Microsoft’s proprietary Internet Explorer.
Tapscott, the co-author with Anthony D. Williams of Wikinomics: How Mass Collaboration Changes Everything, says that an open approach to science “speeds up the metabolism of scientific inquiry. Valid discoveries will be more quickly confirmed, and bogus claims will be exposed sooner.”
In physics, scientists have been using an open approach for years. They post early drafts of papers online in “pre-print” archives. Eventually, after much back-and-forth, some of the research goes on to formal peer-reviewed publication. The online process, as Tapscott observes, ferrets out flaws and confirms validity much faster than in the old days when all research first saw the light of day in paper journals.
Tapscott believes the potential for a more collaborative approach to science extends far beyond drug discovery. “A new economics of intellectual property is starting to take shape,” he says. “Increasingly (and to a degree paradoxically) firms in electronics, biotechnology and other fields find that maintaining and defending a proprietary system of intellectual property often cripples their ability to create value.”
Tapscott argues that the existing proprietary system is starting to topple under its own weight of secrecy. In its place will be more open sharing, using Web 2.0 tools such as blogs, wikis, shared bookmarks and tagging.
Harvard Business School professor Karim Lakhani calls this kind of collaborative work on a problem “broadcasting.” In a 2007 study, he and colleagues found that when 166 previous unsolved scientific problems selected from companies’ research laboratories were opened to outside scrutiny, a solution was found for a third of them.
“Opening up the scientific problem-solving process,” they concluded, “can yield innovative technical solutions, increase the probability of success in science programs and ultimately boost research productivity.”
Surprising as it may sound, the notion of scientific collaboration is relatively recent. The followers of Pythagoras, the Greek whose name is associated with the famous theorem about the square of the hypotenuse of a right-angle triangle, collaborated among themselves, but were forbidden to reveal their secrets to outsiders. Scipio Ferro, an Italian who first figured out how to solve equations of the form x3+mx = n, kept this knowledge secret for 30 years.
But in modern times, scientists have been rewarded – by acclaim, tenure and sometimes cash – for publishing their work. And publication is traditionally the first step in what is a kind of slow collaboration with other scientists all over the world.
Edwards says that what his consortium is doing in its three labs in Toronto, England and Sweden falls within this slowly collaborative tradition. The difference now is that they want to speed things up. So, when he and his colleagues have identified the shape of a protein – as they have done for more than 700 proteins so far – they immediately deposit the information information in a protein data bank.
Publications follow later, Edwards says, but they’re still an essential part of the process. The time-honoured practice of peer review ensures that any findings are scientifically valid. And for the author, publication in a peer-reviewed journal confers respect from peers. Since most scientists don’t choose their line of work for the money, there need to be other rewards. “The buzz for scientists is when everybody knows that you discovered something,” Edwards says. “That’s what drives us.”
Some Corporations have begun to learn that by aggressively protecting their intellectual property, they may, in fact, be stifling scientific progress. Some are warming to the idea of open development.
The members of the Structural Genomics Consortium, which include several large pharmaceutical companies, have pledged $30 million a year through 2011 to support some 200 scientists and technical staff at the University of Toronto, Oxford University in England and the Karolinska Institute in Stockholm, Sweden.
For their contribution, the pharmaceutical giants hope eventually to gain information that will allow them to make better decisions about what drugs to test in the highly expensive game of clinical trials.
Running a clinical trial is like paying between $30 million and $300 million for a lottery ticket, with poor odds of winning. “It’s virtually a crapshoot,” says Edwards. The consortium’s work won’t change the price of the ticket, but the odds of winning may improve. In the long run, this will be good – not only for the pharmaceutical industry but for the rest of society – because more drugs will make it to market faster.
The current model of drug development still works, but it’s yielding fewer results. “We’re spending more and more and more and we’re at best treading water,” Edwards says. In his model, industry pays up front for an encyclopedia of information that individual companies can use as they see fit, with the hope of improving the odds of developing a winner.
Not everyone is convinced the open approach will work, however. Some universities are reluctant to give up the chance of getting in on the ground floor of a blockbuster discovery, even though the odds are extremely low. If they continue to focus on that small probability, says Edwards, the science will grind to a halt. Most corporations (with some laudable exceptions) remain wedded to the patent-everything approach, he says, even though just about everybody agrees the current drug development model is broken.
Edwards says these hurdles can be overcome if he and his colleagues can present a convincing alternative. “We’re going to need to create the case for open domain – we are the vanguard,” he says. “The best way forward is just to do it. We’ll walk the walk and people will say, ‘Look, it works.’”
Michael Smith is a science writer in Toronto.