/ 30 March 2010

Fast machines, genes and the future of medicine

Francis Collins, who helped map the human genome, did not get around to having his own genes analysed until last summer. And he was surprised by what he learned.

Collins has a predisposition for type-2 diabetes, something he had never suspected. The lanky, former director of the National Human Genome Research Institute (NHGRI) discovered this through tests offered by Navigenics, 23andMe and DecodeMe — companies that charge customers a few hundred dollars for a peek at their genetic makeup.

“I signed up for all three because I wanted to see if they gave the same answer,” he said. “They all agreed my diabetes risk is higher.”

Armed with that information, he eventually lost 11kg. But as a rule, he doesn’t consider such tests especially useful — at least not yet.

“Admittedly, right now your family history may be your best bet and it doesn’t cost anything,” he said.

And so it goes in the fledgling genome field.

Some experts say the world is on the cusp of a “golden age” of genomics, when a look at the DNA code will reveal your risk of cancer, diabetes or heart disease, and predict which drugs will work for you. Yet the $3-billion international Human Genome Project, whose first phase was completed a decade ago, has not led to a single blockbuster diagnosis or product.

To be sure, there have been some tantalising glimpses:

  • A personalised blood test can tell whether a patient’s cancer has spread or come back. Dr Bert Vogelstein of Johns Hopkins University in Baltimore and colleagues found stretches of DNA in colon and breast tumors with extra DNA copies, or fused-together chromosomes.
  • A gene-based test called Oncotype DX made by Genomic Health helps identify breast cancer patients who are not likely to benefit at all from chemotherapy.
  • Dr James Lupski of the Baylor College of Medicine in Houston studied his own entire DNA map and sequenced the genomes of family members — including his deceased grandfather — to diagnose the mutation causing his rare genetic nerve disease, called Charcot-Marie-Tooth syndrome.
  • Genetic tests are now able to pick out poor responders to Plavix, or clopidogrel, a common life-saving anticlotting drug made by Sanofi-Aventis SA and Bristol-Myers Squibb.

Still, Collins describes this as low-hanging fruit. He says the hard work is only just beginning.

In a sense, the field is a victim of its own success. Companies are beating down the price of genetic sequencing, competing to make the machine that every biotech lab will have as standard equipment to sequence a person’s entire genome on the spot. But all this genome sequencing is creating what current NHGRI director Dr. Eric Green calls a “tsunami of information” that is overloading the brains of scientists and the capacity of computers.

Paradoxically, this reflects the fact that people have relatively few actual genes, the stretches of DNA that instruct a cell to make a protein, or what Green refers to as “bricks and mortar.” Humans have just 20 500 of them, compared with up to 30 000 for mice and 50 000 in rice. That was one of the big surprises from the Human Genome Project.

As a result, much of the most important information lies in what used to be called “junk DNA,” which makes up two-thirds of the human genetic code.

“There is this dark matter of the genome that is lurking out there, waiting to be uncovered,” says Collins.

PUT IT IN THE BIOBANK

One part of the hunt starts in a nondescript building in west London, where volunteers line up to reveal their innermost secrets. While many have given blood before, this time they are donating their DNA and medical records, both past and future, to a vast experiment that will track them to the grave.

It sounds Orwellian. Yet volunteers checking into UK Biobank — backed by the government and the Wellcome Trust — are keen to participate in something that might help their children or grandchildren.

This age group, 40 to 69 years, has been chosen because the volunteers won’t keep researchers waiting too long before developing interesting conditions such as cancer, arthritis, diabetes, heart disease and dementia.

So far some 450,000 Britons have signed up, consenting to have their DNA sequenced and their health tracked, anonymously, through the National Health Service.

The target of 500,000 should be reached around July, by which time the project’s giant freezer facility in northern England will have the equivalent of two road tankers worth of frozen blood samples.

Principal investigator Dr. Rory Collins says it is only by doing such large-scale sampling that scientists can uncover how lifestyle factors interact with a long list of rare genetic variants to cause common diseases.

“If you are looking for the effect of lots and lots of different genetic variants that are producing modest effects and they’re interacting with a lot of non-genetic factors, then you need to be able to do studies that are very, very big,” he said. “It’s only just now that the technology allows those experiments to be done.”

China, Sweden and other countries have also set up biobanks but the British one is the most comprehensive in terms of the number of factors studied. Organizers hope it will go beyond what earlier biobanks produced — like one in Iceland that helped create gene-hunting firm Decode Genetics.

Working out of a glass-and-steel building on the outskirts of Reykjavik, Decode’s scientists have peppered the scientific literature with reports on common DNA variants linked to schizophrenia, cancer and other diseases by trawling the country’s genetic heritage, which has changed little since the Vikings arrived more than 1,000 years ago.

Understanding a few of the pieces of the gene puzzle, however, was not enough to shore up Decode’s ailing business and the former Nasdaq-listed company filed for bankruptcy protection last November. It re-emerged as a private business in January.

Decode was one of a number of biotech start-ups that rode the first wave of genomics, offering the technological tools needed to understand the links between genes and diseases. Many fell by the wayside after just a couple of years — but not all.

Human Genome Sciences Inc is one that finally looks set for prime time. Its shares have skyrocketed since last year, when it reported unexpectedly strong data from a trial of its experimental lupus drug Benlysta.

Last March the company was trading as low as 45 cents; now its shares hover around $30. If approved, the drug, which is being developed in partnership with GlaxoSmithKline Plc, would be the first new treatment for lupus, a serious immune system disease, in more than 50 years.

But such winners are rare and investors remain wary of biotech drug developers over-selling the promise of genomics, given the fact that new medicines face a risky, 10 to 15-year path to market.

In fact, the past decade has turned out to be the worst in the history of the drugs industry, with a dearth of new medicines and an unprecedented cliff of patent expiries.

“There is no question that people have felt that they got their fingers burnt and the enthusiasm has decreased a great deal,” said Glaxo’s head of genetics Lon Cardon.

The problem for drug developers and investors is that greater knowledge has brought with it greater complexity, frustrating early hopes for relatively simple fixes to complex diseases.

Yet Cardon, too, now sees a turning point, driven by cheaper, faster sequencing and clear advances in one key disease area — cancer.

Genes and cancer
For many cancer patients, a major fear is that their surgeon missed something and their cancer will grow back. The only way to tell now is to wait until tumors are big enough to be spotted by imaging machines.

That could soon change. A gene-based test that can search a patient’s blood for tiny bits of DNA shed from tumours may soon give doctors an early warning that they may have missed something.

“That’s only become possible through the advent of so-called next-generation sequencing technology,” said Dr Bert Vogelstein of Johns Hopkins University in Baltimore and the Howard Hughes Medical Institute, who is developing the blood test.

The test takes advantage of rapid advances in the technology to sequence whole genomes. The latest machines from companies like Illumina and Life Technologies can map out a patient’s whole DNA code in just a few weeks for as little as $5 000, a far cry from 13 years and $3-billion it took Collins and his international collaborators to get the first human genome.

Vogelstein said the rapidly falling cost of genome sequencing means the blood test could be affordable enough to be on the market within two years. Before long, all cancer patients could have their tumours sequenced routinely to find the genetic defects that cause them to grow.

“Cancer is maybe the best disease to cut our teeth on,” said Yale Medical School geneticist Richard Lifton. “The reason for that is we know that cancer is largely a disease in changes of DNA sequence.”

Matthew Meyerson of the Dana-Farber Cancer Institute and the Broad Institute of Harvard and the Massachusetts Institute of Technology said he is impressed by the pace of change. “The first cancer gene sequence was reported in 2008. There were probably 100 done last year. Maybe there will be many hundreds or even 1 000 this year,” he said.

Lifton predicts that within the next two years, scientists will have the genetic sequence of every major human cancer. “Many of these will identify new genes that we had not previously known about with a role in cancer,” he said. “Some of these will turn out to be incredibly important new drug targets.”

Or new tests, as Collins suggests. “If you’ve just discovered the molecular basis of a rare disease, you can turn it into a [test],” he said.

A new industry is just emerging to help them get there — and its smart kit is bowling over scientists.

Shrinking technology
Dr Eric Green had never seen anything like it. At a meeting of what he calls “sequencing geeks” at Marco Island, Florida in February a small start-up called Ion Torrent was demonstrating its new DNA sequencer.

“It’s the size of a computer printer,” he said. “They were sequencing in a hotel room.”

It was a shock to researchers who had used rooms the size of a football field full of sequencers for the original human genome.

Green said his institute has directed some funding to Guilford, Connecticut, and San Francisco based Ion Torrent for its $50 000 sequencer.

“This may completely crash and burn,” he acknowledged. But he and others were intrigued at the company’s compact system to detect individual molecules of hydrogen as a way to sequence the A, C, T, G code of DNA.

In another room was Pacific Biosystems’ $750 000 sequencer the size of a conference table.

“It is the Wild West,” Green said. “It is emblematic of what is going on in the field now, with not one, not two, not three, not four but multiple technologies.”

They are doing what the companies and the researchers want, and prices are plummeting. “There is a mix of science and business here,” said Green. “It’s breathtaking, what is happening.”

Taste of the future of medicine
If human genome sequencing is to transform medicine, it will have to be quick and easy to do. “One could imagine that acquiring a complete genome sequence of an individual might become the standard of care one day,” Green said.

“What these companies are doing is giving us a taste of the future of medicine.”

Dr Richard Gibbs, who directs the Human Genome Sequencing Centre at Baylor College of Medicine in Houston, said there are about 20 to 30 different sequencing companies out there that are trying new things.

He said scientists had expected to have to wait for these machines to produce real breakthroughs, but the ones by Illumina and Life Technologies are cranking out so much data and the price is falling so quickly that they are likely to be the ones that transform medicine.

“The current things are performing so damn well they are blowing out of the water the pie-in-the-sky numbers that were pushed around by the next next-generation companies,” Gibbs said.

Illumina’s chief executive Jay Flatley said storing and analysing the trillions of bits of data generated by their machines will likely be the biggest future stumbling block for sequencing companies.

Collins is confident they can do it. “We are not going to hit any limits any time soon,” he said. “No laws of physics need to be broken … We’ll have the $1 000 genome in a few years. It will [eventually] be possible for people to make a $100 genome.”

New markets for drugs
The high expense is not stopping China, which is making a big push into genomics. The Beijing Genomics Institute just bought 128 Illumina machines and is employing 1 000 researchers to focus on illnesses that are specific to Asian populations. They may also find new markets for drugs.

AstraZeneca Plc’s lung cancer pill Iressa, or gefitinib, was found to work far better than chemotherapy in people from East Asia whose tumors had specific mutations in genes for the epidermal growth factor receptor or EGFR. The discovery saved the drug, which only works in about 10 to 15% of lung cancer patients in Europe and which was headed for the trash bin.

“Most genomics research has been done on Caucasians based in Europe or the United States and we are only just starting to understand about how applicable these findings are to worldwide or Asian populations,” said Martin Hibberd of the Genome Institute of Singapore.

Analyse this
None of this means people should rush out to have their DNA tested by companies such as 23andMe and DecodeMe.

“I’d hesitate to call them a scam,” said Frances Flinter, a member of Britain’s Human Genetics Commission which is coordinating a global working group on the new guidelines. “Some of the companies offering these tests are trying very hard to develop something that has some scientific validity. But at the moment the scientific knowledge doesn’t exist to demonstrate they are useful.”

The Human Genetics Commission’s working group — which also includes members from Iceland’s Decode and California’s Navigenics — say firms should ensure buyers understand what they are doing, and what the outcome of the actions might be.

The draft guidelines, due to be revised by the end of the year, also suggest firms should only offer tests for incurable conditions like Huntington’s or Parkinson’s disease with “before and after” counseling, and on other tests, “make clear the limitations … and the relative roles of genetic makeup and lifestyle.”

“For the vast majority of people, decisions on lifestyle will probably have far more impact on health,” said Flinter, a consultant in clinical genetics.

Collins agrees. After all, the leading causes of death in the developed world — cancer, heart disease, stroke, diabetes — all can be prevented to a large degree with exercise, by avoiding tobacco and by eating less fat and sugar and more fruits and vegetables.

Then again, he had no family history of diabetes.

“Everybody in my family is lean and athletic so perhaps they were fighting off their genetic risks by their individual behavior choices,” he said.

“I wasn’t looking that lean and athletic when I got these test results last summer, and it caused me to pay attention to something I should have paid attention to all along.” – Reuters