ABOUT 130,000 years ago, so the popular theory goes, the first humans carved
homes out of caves in Africa. In those early days, primitive people woke with
the Sun, chipped tools from rocks, and hunted and gathered. We’ve come a long
way since then, but in the eyes of evolution it was only yesterday. In the 7000
or so generations since modern humans emerged on Earth, we’ve had precious
little time to diverge genetically. We are all still 99.9 per cent
identical.
Compare that to chimpanzees. They tend to look pretty similar, but the same
gene in any two of them can show up to five-fold more variation than in any two
humans. We, despite our superficial differences, are much like one another, and
not that different from our rock-chipping ancestors. “We are all close cousins,”
says Eric Lander, director of the Whitehead Institute Center for Genome Research
in Cambridge, Massachusetts. All 6 billion of us.
Cruise the biotech hinterlands of California and Cambridgeshire, and you’ll
find many a company that is hoping to cash in on our interrelatedness. They are
hoping that our lack of diversity may just make it possible to decipher the
genetic variations that play a big role in deciding who gets cancer, diabetes,
heart disease or any other of the big killers.
Advertisement
The key to the next step in the genetics revolution, claim some geneticists,
lies with the burgeoning catalogues of tiny molecular beacons called single
nucleotide polymorphisms. These SNPs (pronounced “snips”) are variations that
occur when a single letter in our genetic code is altered. If the geneticists’
hunch is right, within years, SNPs will be able to pinpoint your risks of dying
from many common diseases, and hopefully do something to reduce them. “People
will look back on 20th-century pharmaceuticals developments and say, `Wow, was
it barbaric! It’s amazing anything got done’,” says Lander.
At the moment, when it comes to predicting your chances of, say, dying from a
heart attack, all a doctor usually has to go on is a brief medical history and a
quick chat about your lifestyle. So it’s not surprising that for a third of
people with coronary heart disease, the first warning that something is wrong is
the heart attack that kills them. But once the genetic combinations that
fine-tune disease susceptibility are known, doctors will be able to warn one
person to take more exercise and tell another to start taking one of a new
generation of drugs that will prevent—rather than treat—disease.
First, however, geneticists need to find some way of working out who carries
which genes.
Although it didn’t seem like an easy task at the time, tracking down the
single genes that cause the rarer diseases, like Huntington’s, or trigger a tiny
proportion of the more common diseases—BRCA1 and breast cancer,
for instance—was a far simpler nut to crack. Geneticists looked at
families with the disease and worked out which characteristic pieces of genetic
sequence—or “markers”—occurred in family members with the disease,
but not any of the healthy relatives. Using such linkage analysis, they could
home in on the interesting sections of the genome, and then sift through until
they found the gene that was causing the problem.
But that simple if time-consuming technique doesn’t work for the genes that
help trigger the big killers, such as heart disease. These so-called “complex
diseases” involve numerous genes, interacting in myriad ways to raise or lower a
person’s risk of disease. What’s more, each gene may come in several slightly
different forms. Add to that the scarcity of markers along the genome at the
geneticists’ disposal, and you can see why when it came to complex diseases it
was impossible to separate out the noise from the signal using traditional
linkage analysis.
Hot spots
Nothing can be done about the complexity of the big killers, but as it turns
out, something can be done about the number of markers along the genome. This is
where the SNPs come in.
SNPs are the hot spots on the chromosomes where variations commonly occur and
they account for most of the genetic differences between people. At these spots,
your genome may read “AAGGCTAA” while someone else’s may read “AT
GGCTAA”—a thymine nucleotide (T) has taken the place of adenine (A).
With the first rough draft of the human genome mapped out, geneticists are
now starting to add fine detail, including the position of some 3 to 10 million
SNPs. And although it’s not yet clear that SNPs help cause diseases, gene
hunters are banking on being able to use them as genetic signposts to conduct
mass screenings of human genomes for the gene combinations that do. They could
compare, for example, SNPs in people with and without heart disease, to see
which correlate with the disease. Doctors could then scan a patient’s genome to
see whether he or she carries the telltale patterns of SNPs associated with a
heightened disease risk.
And once again, a piece of evolutionary good fortune may have made things
easy for geneticists. They wager that most SNPs have no effect on our ability to
procreate, so that they go unchecked by evolutionary pressures—and that
includes the SNPs linked to the big killer diseases that strike after the
childbearing years. With no outside pressures to dictate when and where they
appear, scientists say, SNPs have become randomly scattered across the genome,
occurring roughly once in every 1300 nucleotide bases, providing the perfect
signposts. “SNPs are a wonderful advance,” says David Altshuler, an
endocrinologist at the Whitehead.
Public research teams and companies worldwide seem to agree. They are now
drafting genome maps stacked with SNP information gleaned from the blood samples
collected from dozens of volunteers. Just this September, Celera Genomics of
Rockville, Maryland, the company that turbocharged the race to sequence the
human genome, announced that for a fee, subscribers could view its database of
2.8 million SNPs.
Then, a few weeks later, researchers at the Sanger Centre in Cambridge
unfurled a map of 2730 SNPs on chromosome 22 (Nature, vol 407, p 516).
The Sanger Centre, along with various other academic centres, drugs companies
and the Wellcome Trust, is part of a SNP consortium that has pledged to create a
free SNP map by next year. The consortium has located at least 800,000 so
far.
Meanwhile, Allen Roses, director of genetics at Glaxo Wellcome in Research
Triangle Park, North Carolina, and his colleagues, have been testing a draft SNP
map to see how useful it is in practice. Alzheimer’s is one of the few complex
diseases in which a major susceptibility gene has been analysed, so it’s an
ideal test of the power of SNPs. One gene, ApoE, comes in three
varieties: ApoE2, ApoE3 and ApoE4. People with the
ApoE4version of the gene are most likely to develop Alzheimer’s, while
those with ApoE2 are least likely. People who have ApoE3 seem
to have intermediate disease risk.
Roses and his colleagues mapped an area of the genome surrounding the
ApoE gene in three groups of people: unrelated individuals with
Alzheimer’s, members of families who had the disease, and healthy individuals.
Sixty SNPs sat in the analysed region. Sure enough, an ApoE4
SNP—as well as two other SNPs sitting close by—occurred far more
often in people with the disease. “żěè¶ĚĘÓƵs used to spend years, with little
success, hunting for these complex disease genes,” remarks Roses. “Now it takes
just months, and it will get much faster than that.”
But the study also highlights some of the limitations of current SNP
technology. The Glaxo Wellcome team scanned a sliver of the human
genome—1.5 million out of a total 3 billion DNA nucleotide bases—and
turned up just a small number of relevant SNPs. But even this mini-version of a
whole genome scan took considerable time and money, says Roses. “It could cost
$100,000 to $200,000 to scan the genome of a single person,” Roses
says.
Many scientists expect SNP costs to plummet, as they have with other gene
sequencing technologies. But for now, costly SNP analyses are primarily being
used to streamline routine linkage analysis, or to work out how a single gene
that is already implicated in a disease changes in health and sickness (see “What’s up doc?”).
Early-warning system
Ultimately, researchers hope to harness SNPs to track down the subtle genetic
patterns that signal susceptibility to different diseases, and allow doctors to
screen for those diseases. Alongside the spatulas and stethoscopes, clinic
drawers could one day house “SNP chips”—tiny microarrays studded with the
DNA sequences that bind to different SNPs
(żěè¶ĚĘÓƵ, 14 November 1998, p 32).
A technician would wash a patient’s DNA over the chip, and
fragments that matched the sequences would bind to the chip and light up. With a
little computer analysis, doctors would know which gene variations each person
carries—and with that head start, they could intervene long before the
telltale memory loss or high blood pressure began.
Affymetrix, a biotech company in Santa Clara, California, is leading the SNP
chip brigade. “We can [already] put millions of DNA probes on a piece of glass
the size of your thumbnail,” says Janet Warrington, director of health
management at Affymetrix. “And we’re only [going to] pack more information on.”
A lot more. Last month, Affymetrix announced the birth of a spin-off company,
Perlegen Sciences, which plans to copy the entire genomes of 50 people onto
13-centimetre glass wafers within a year. Once Perlegen has located sufficient
SNPs, the company wants to attract drugs companies into partnerships that would
use the wafers to link different SNPs to diseases or unfavourable reactions to
drugs.
“Everybody and their brother would like a way of making drugs safer,” Roses
says. At Glaxo Wellcome, researchers are already using abbreviated SNP maps,
dubbed “medicine response profiles”, to search for genes that cause bad
reactions to drugs.
But even with your genotypes in hand, your future health will be far from
certain. Complex disease genes usually add up to moderate risk, not a certain
curse, with the deciding role often played by your environment and lifestyle.
You might learn, for instance, that your risk of hypertension is just a third
higher than normal. Still, health experts are betting that such extra
information will be just what you need to put that salt shaker back in the
cabinet, or remember to take your morning medicine—even before symptoms
strike. “At the moment, we’re reluctant to start treatment before symptoms,
because of the poor return and potential side effects,” says Mark McCarthy, an
endocrinologist at Imperial College, London. A patient’s SNPs, he says, could
“tilt the balance” in favour of preventative drug therapies.
And that could have major benefits for public health. At the moment, genetic
testing only identifies those people who are at very high risk of getting a
relatively rare disease. But the vast majority of people have a low to medium
risk of getting a common disease. Find a way to reduce their risks, even by a
small amount, and you’ll have a much bigger impact on public health. “If we
discover that you have a cholesterol of 240, can we predict that you’re going to
die of a heart attack? Absolutely not,” says Altshuler. “But can we say you have
a moderate increase in risk? Yes. And that’s worth something.” In the same way,
he says, SNP profiles will uncover moderate risks that you may not have been
aware of.
SNP chips could make the future brighter for a few lucky people. After all,
with some genetic makeups you might just get away with a life of pigging out,
boozing and vegging in front of the telly.
SCATTERED across the genome are tiny genetic variations that may help
researchers figure out which bits of a person’s genome code for green eyes, a
knack for mathematics or an increased risk of diabetes, cancer or heart disease.
These variations, called SNPs, can also reveal the workings of genes already
implicated in disease.
Stephen O’Brien’s lab at the National Cancer Institute near Washington DC has
amassed gene data on some 10,000 people at risk of HIV infection. Of those who
have contracted HIV, some have developed AIDS more rapidly than others.
To find out why, O’Brien’s team went SNP-hunting inside eight genes that code
for proteins affected by HIV, such as CCR2, a cell receptor that the virus
hijacks to break into immune cells.
And the targeting worked. This autumn, the NCI team described eight SNPs
within the genes. Some, including a CCR2 SNP, are linked to a delay in
the onset of AIDS; others to more rapid disease progression (Annual Review
of Genetics, vol 34, p 563). Those findings raise the possibility that SNPs
may be used in the future to more closely tailor a patient’s drug therapy to
their genetic makeup.
In a separate study (Nature Genetics, vol 26, p 76), David Altshuler
of the Whitehead Institute Center for Genome Research in Cambridge,
Massachusetts, and his colleagues, found that people carrying an SNP in a gene
called PPAR-gamma face a 25 per cent higher risk of getting diabetes.
Other researchers are exploring variations in genes that affect blood clotting,
heart disease and autoimmune disorders.
“Finally, we can begin to ask, `What roles, if any, do all these common gene
variants play in disease?'” says Altshuler.
What’s up, doc?
-
Further reading:
An SNP map of human chromosome 22
by J. C. Mullikin and others, Nature, vol 407, p 516 (2000) -
Pharmacogenetics and the practice of medicine
by Allen D. Roses, Nature, vol 405, p 857 (2000) -
For information on SNPs, see the SNP consortium website:
http://snp.cshl.org