Aiming for a more individualized form of medicine

When he was getting his PhD at the University of Toronto, Steve Scherer, now director of the Centre for Applied Genomics at the Hospital for Sick Children, was told something both fundamental and startling by his mentor and teacher, Lap-Chee Tsui. "He said to me, 'The only advice I'll ever give you in genetics is be the first one to get new equipment in your lab,'" recalls Scherer in an office in the MaRS complex where his current lab spills over two floors and has become a beta testing site for new equipment.

Tsui, who achieved world renown after becoming a co-discoverer of the gene that causes cystic fibrosis, is far from alone in his belief that progress in modern genetic research is not simply enhanced by technological advances, but indistinguishable from them in many ways.

Across the road from Scherer in Mount Sinai Hospital, Kathy Siminovitch, head of the Division of Genomic Medicine at the Toronto General Research Institute, verbally constructs what you might call The Relentless Advancement of Genetic Technology equation. "The sequencing machine that we just bought is at least a hundred to a thousand times faster and maybe a hundred to a thousand times less costly than two-year-old conventional sequencers," she says about her high-throughput genotyping laboratory in which hundreds of research projects besides her own take place. "But the technology moves so quickly I can see in two years' time there will be ones that are going to be 1,000 or 10,000 times faster and cheaper."

If you are looking for a more concrete expression of what Siminovitch's equation means, consider the following. The Human Genome Project took 13 years and $3 billion to produce the first human genome map. (A genome is the full complement of DNA sequence found in any living organism.) It cost Craig Venter and associates $70 million. And it took four years to produce the first individual's genome - Venter's own.

Venter estimated his team could duplicate the effort in someone else's genome for $300,000. In 2006, a California foundation offered a prize of $10 million to anyone who could map 100 individual genomes in 10 days at a cost of no more than $10,000 per genome. But this is just a step toward the point at which the technology becomes cheap enough so that every baby comes out of the hospital with a copy of his or her genetic makeup figuratively pinned to his or her diapers - the $1,000 genome.

The first effect of the onrush of faster, cheaper, better DNA sequencers, robotics and other genomic technologies has been a reconfiguration of how genetic science is done. "With better, faster computers and sequencers there was a shift from hypothesis-driven research to hypothesis-free-type research," Scherer says. "In the past, you would study a specific gene and try to figure out what it did in a cell or how it impacted on disease. The new approach was to map a chromosome or a genome and ask questions based on that data."

The way that this data-first, questions-second approach produces revolutionary results can be seen in discoveries made in Scherer's laboratory now known as copy number variation (CNV). Working in conjunction with Charles Lee of Harvard Medical School, Scherer discovered that when you looked at normal people's genomes there was a wide variety in the number of copies of genes they carried. Sometimes a gene or a chunk of gene-bearing DNA was entirely absent. Other times it was found in two or three or up to a dozen duplicates. And the more the researchers looked, the more different humans seemed to be when it came to copy variation. It is now estimated that 5 per cent of any single person's DNA is present in two or more places in his or her genome.

The consequences of duplications and deletions are still uncertain. Some double copies have been found to protect against diseases such as HIV and malaria, while others have been associated with genetically rooted diseases such as Alzheimers, Parkinson's and, recently by Scherer's group, a type of autism.

In other cases, a person dies if there are not two copies of a given gene. Following up, Scherer and his colleagues are now exploring whether lethal CNVs are the explanation for why upwards of a third of pregnancies self-abort.

But beyond the CNV specifics, the data-first, hypothesissecond approach has also produced the beginnings of the road toward a medical nirvana known as personalized medicine. The raison d'être for personalized medicine grows out of the deep contextual ignorance that doctors are now faced with when treating most illnesses.

Siminovitch, whose laboratory studies the genetics of immune-related diseases and chronic inflammatory diseases such as asthma and arthritis, describes the confusion that arises after someone receives a generalized diagnosis of rheumatoid arthritis. "Ten people walk in my door after being told they have rheumatoid arthritis and they're all very different in terms of their disease. Somebody's going to get a disease that's really not so bad. It's going to require almost no treatment. Someone else is going to get another version of the disease that is so fulminating and so terrible that every joint will be involved and that person is going to be in severe, terrible pain. And not only that, they're probably not going to respond to any of my first-line medications."

"But today I have no way of telling who is who," she says. "Basically, everybody gets an individual disease almost unique to that person, and we don't yet have a way to predict outcome." The hope is that as the cost of individual genome mapping falls and the understanding of the associations between gene insertion, deletions and mutations rises, people in the near future will be given treatments that fit them and not the "average" patient.

Indeed, even as scientists struggle to make sense of the masses of data that faster and more accurate DNA sequencing is throwing up in their faces, the first outlines of genome-specific, personalized medicine is appearing.

Craig Venter has changed his lifestyle to fit a disease profile his DNA suggested might lie in wait for him.

"His genome shows that he carries genetic variations associated with high cholesterol," says Scherer. "I recently had lunch with him and can report he eats salads and he's on cholesterol pills. Now we don't yet know if the genes will express themselves and cause a problem, but based on his genes, Craig's taken decisions that should better his chances for good health."

But the rush toward personalizing medicine has also seen individualized gene testing begin to appear on the market. The most famous is California-based 23andMe, in which Google has a stake. Using spit samples it can determine whether a person carries genes that have been shown to have a linkage to 89 diseases or conditions.

And as another sign of things to come, in September of 2008 the company announced that it was dropping the price of its tests from $1,000 to $400 (U.S.) because a faster, cheaper-tooperate genetic analysis technology had arrived on the scene.

What does all this technology-driven change mean to science and medicine in Ontario?

You can hope for a more predictable medical treatment, for one thing.

"What I am hoping in, say, 10 years is that we will be able to predict which child has a high risk of something like rheumatoid arthritis and we will treat them before they have the disease so that they don't even get sick," says Siminovitch.

Scherer forecasts a better understanding of who will respond best to what medications. "I think there are somewhere in the neighbourhood of 100,000 deaths a year in North America due to negative drug reactions," he points out. "If we could figure out who we shouldn't give certain drugs to, that would be a great advance."

But beyond all this lies a larger lesson for both future research and future medicine. It is a lesson embodied in Tsui's 20th-century Confucian words of wisdom to his graduate students.

If Canada and Ontario are to prosper in the personalized medicine era, it's clear to all scientists that standing still technologically is a kind of research death sentence. For example, Scherer estimates that by the end of three years almost all his current machinery will have become technologically obsolete and need to be replaced.

Piggybacking on others elsewhere who have newer technology isn't an alternative - or at least not a good alternative. "When you are the owner of the equipment, you are always the one with the power. Although the word 'collaboration' sounds nice in theory, the truth is you could be really low on the priority list of whoever it is you are collaborating with," says Siminovitch.

And, when you get new technologies, what also arrives is the need to have sufficient money to operate them. "The new DNA sequencer that we have costs a half a million bucks to run, and the minimum cost per run is $8,000. If you don't have a tech who knows how to do those runs in their sleep, you'll fail. It's like a Formula One Ferrari in a race. The guy in the street, if he tries to drive it, he crashes, so you need a team. You need someone who knows how to drive, someone who knows how to change the tires, someone who know how to maintain the car," says Scherer.

Then he thinks for a minute.

"And because it is a race, the next year you had better make certain you have a better Ferrari and the next year after that even better, because if you don't, even with the best team you won't win anything."