Examining the structure of proteins

It's easy to understand what University of Toronto structural biologist Aled Edwards - and the SGC he heads - does. Yes, the Canadian-British-Swedish collaboration aims to determine the shape of the body's proteins, but stand with Edwards at a computer where data gathered from such a protein analysis is fit with a potential structural model and your brain quickly begins to feel like a giant "ouch."

"The data are represented by the blue things, and the red things reflect what people think the structure to contain this data should be," he explains while pointing to the rotating shape on the screen.

Easy enough to say, but to the untrained eye the tangled, twisting image looks as if one is trying to make sense of a structure that seems to have been created by a talented but quite mad Etch A Sketch artist.

"This is literally a puzzle where you know what proteins look like, you know it has to be this sequence, but then you have to overlay a sort of chicken-wire map and are continually saying, 'How the hell do I fit my protein into that tube?'" he continues.

Chicken-wire maps, protein models in tubes - "ouch," yet again. Hmm, you think, maybe what ordinary people need is a simple metaphor that gives the SGC effort a visual flavour, but as soon as Edwards begins searching for one, you suddenly find yourself feeling as if you have entered Alice-in-Metaphor-Land.

"People often talk about it like you are finding a key to fit into a lock," is how he first tries to explain it.

But no. You are not really opening anything but, rather, trying to understand how the shape of a protein might influence its response to drugs that try to block its action. "So maybe you could say it is like searching for different-shaped screwdriver blades to fit into different screw heads," he suggests, dressed casually as usual in shorts, bike shoes and a T-shirt.

But then no again. Because many of the biological screw heads interact and because fundamental differences between proteins can be seen only at the angstrom level - think the thickness of a slice of a slice of slice of a slice of a hair - maybe what's going on is more like trying to put together a biological jigsaw puzzle with microscopic dimensions.

Only the problem here, says Edwards, is that you don't even know what the border of the protein puzzle looks like or the shape of the 23,000 other puzzle pieces in the human genome.

"And if you don't know that, how are you going to know the piece you are holding is going to fit only into one other piece and nothing more?" says Edwards.

Now it's an "ouch." A scientific jigsaw puzzle where you don't even know the shapes of the pieces you are fitting together. "Ouch, ouch, ouch."

But if all this real and metaphorical imagery has served only to confuse you more, then you have already come a long way to appreciating why SGC exists. The basic ignorance about biological structure function is a seminal reason - maybe the seminal reason - why drug development in the 21st-century is threatening to become the poster child for what a failing and stagnant innovation process might look like.

"What is crippling the development of new medicines is the fact that as we discover potential new medicines and we start to test them in people, 90 per cent fail," is how Edwards tartly characterizes it.

"And 90 per cent fail not because scientists are dumb or make mistakes. They fail because we have a very poor understanding of human physiology and pharmacology. We just don't know. You could put 82 eggheads in a room, each with eight Nobel Prizes, and give them 10 medicines and say, 'Which one is going to work in a person?' No one would be able to predict. So until we conquer this fundamental problem, we're not going to have the new medicines for diseases we want to cure."

And this isn't just a cynical Aled Edwards talking. This is also the decidedly unhappy conclusion of a much-cited 2004 U.S. Food and Drug Administration report.

Despite a 57 per cent increase in U.S. biomedical research funding from 1994 to 2003, new drugs entering early-stage clinical trials have only an 8 per cent chance of reaching the market - down from 14 per cent 15 years before. Indeed, the year 2004 represented a 20-year low in the numbers of new molecular therapies, with 17 truly new drugs reaching market. Not to mention that failure rates during final-stage clinical trials were as high as 50 per cent - up from 20 per cent a decade before.

Why can't we tell what drugs will work and what won't at the beginning of the drug discovery process?

The simplest answer is that many drugs that block the bad effect of a disease-related protein - perhaps arthritis pain - are also likely to block the effect of many of the often hundreds of other proteins that, at anything but the most microscopic level, look precisely the same. It is this good/bad result of taking a drug - the trade-off between the effect you want and the sometimes lethal side effects you don't want - that has turned the drug development process into a fool's poker game where companies regularly ante up tens and hundreds of millions of dollars without being able to a look at all their cards.

With this in mind, the argument has become: If you know the exact structure of a disease-related protein, you've come a long way to being able to create a drug that you believe early on blocks that protein - and only that protein in the body. And this reasoning explains why three of the world's great universities, three of the world's biggest pharma companies, the U.K.'s Welcome Trust and charitable organizations (including the Canadian Institute of Health Research, Genome Canada and the Canada Foundation for Innovation) and government agencies in Canada, the U.K. and Sweden have come together in SGC to produce a very different model for research that enables drug development.

"In 1999 the companies got together and said, 'You know, we want this information about protein shape but we don't have the resources to generate it ourselves and we don't mind if our competitors also have it. Why not pool the money, get one person to generate it, make the information free and, in so doing, we all would benefit,'" is how Edwards characterized the thinking.

After much searching for a chief executive, they seized upon Welsh-born, Montreal-raised, McGill and Stanford-educated Edwards. Why him? One reason was the belief research wasn't just about knowing. Not only was Edwards a good scientist, he was rapidly turning into a serial entrepreneur, with two start-up companies to his credit at that time.

The initial SGC mandate was to determine the structure of 386 proteins with medical relevance in three years. This was anything but easy, even when aided by that master of 3-D visualization, the computer.

"The data tell you quite unequivocally what the structure looks like, and programs create the models sans humans," Edwards muses, "But what the structure means..." - there is a deep pause - "...that requires great thought."

Nonetheless, in terms of its protein structure determination goals, the SGC has become the world's exemplar for turning blue experimental protein data into red structural protein models. By the end of its first three years of operation in 2007, the SGC had determined the structure of not 386, but 455 human proteins. This Herculean feat was done at costs that are somewhere between one-third to one-eighth the cost in other laboratories.

Encouraged by its success and renewed funding in 2007, SGC now aims at determining the structure of 650 new human disease-related proteins by 2011.

But these numbers alone don't fully capture what has been so surprising about the SGC. The consortium was a collaboration specifically set up to freely and openly dispense all of its data. But almost by accident, it has also turned into something of a profit-making enterprise.

"We developed a process that allowed for simpler and more controlled cell culture growth and immediately thought, 'Oh, man, people have gotta use this. It really makes life easier for a scientist,'" says Edwards. "And people who saw the machines wanted them too."

But how to do that? After unsuccessfully trying to propagate the use of the technology - by placing the engineering drawings online - serial entrepreneur Edwards and associates set up Harbinger Biotechnology and Engineering to market the concept. "Now Harbinger is selling the technology and it's employing people," Edwards says. "We didn't set out to do that. But if you create good science, stuff like this happens."

Other unforeseen shortcuts and benefits - in what is normally a long, arduous and failure-ridden 15-year drug discovery process - have started to emerge. As part of its next research efforts, the Toronto branch of the SGC consortium has decided to specialize in so-called neglected diseases, the orphan illnesses that governments, biotechnology companies and traditional pharma companies often ignore because the dollar payback for coming up with a new drug is small.

With this need in mind, Ray Hui in Edwards' lab has begun looking at the structure of the proteins produced by the parasite that causes cryptosporidiosis - a water-borne disease most often found in Africa. In healthy people the disease usually runs its course but in the sick and immune-compromised - think AIDS patients - it can be deadly.

To date there is no standard treatment for cryptosporidiosis. After Hui and others determined the structures of cryptosporidium proteins, they compared them to known structures in human proteins and noted an amazing similarity to proteins associated with bone-wasting conditions such as osteoporosis. This led to a eureka moment. Could drugs designed to stop osteoporosis also stop the cryptospirosis parasite?

"We've found that an existing family of drugs already clinically approved for osteoporosis actually work very well to kill the parasite," says Hui.

Suddenly a medical "ouch" was turning into a drug "wow." And finally, the SGC model for how to make drug discovery more efficient - Edwards has estimated that knowing a protein structure can knock 18 months off the drug development time frame - has not been lost on other scientific enterprises. Chas Bountra, the head of the Oxford branch of the SGC, has been in active discussion with people at the U.S. National Institutes of Health to free up another sticking place in the drug development pipeline - reliable tests to determine, as early as possible, whether a drug works or doesn't or has unwanted side effects. The NIH mandate is to develop and make chemical screening tests publicly available in a way similar to what SGC has done for many years with respect to protein structures.

Ergo, a partnership between the two efforts is now forming. A vision, hard work and the best technology in the world matter. But something else has also made SGC a success - the effusive, positive, pulsating personality of Aled Edwards.

"His energy and passion are infectious," says Bountra. "And that is what we need in this business because drug discovery is hard. We need to put egos aside. We need to put all the brains together - and, in that regard, everyone knows Aled arrives with no personal agenda. All he cares about is good science."

No "ouches" there.