|
| http://www.wired.com/wired/archive/...start.html?pg=6
It's a harsh reality of the drug business: The only way to make money is to keep
the pipeline full of new products. Yet a drug takes something like $800 million
and 10 years to get from the lab bench to patients, and at any time in the
process, side effects or adverse outcomes can scuttle its approval. So biotech
companies are always looking for ways to catch pipeline-clogging troubles early.
Their latest guinea pigs aren't animals or people, but machines.
Hardware that simulates human physiology lets drugmakers experiment in bulk, on
the cheap. Most of the new high-speed tools use robot-driven, microscopic arrays
- like miniature test tube racks - to try thousands of chemicals at once. In the
worst case, a potential medication might be toxic. Because the human organ that
copes with toxicity is the liver (it metabolizes compounds into either benign or
harmful byproducts), Solidus Biosciences of Troy, New York, sandwiches drug
candidates between liver enzymes on one microscope slide and cells from various
human organs on another. The enzymes break down the drugs, exposing organ cells
to the metabolites and revealing any toxicity.
In another, somewhat broader approach, a San Diego company called Kalypsys uses
a robotic arm to fill 1,536 tiny wells with a mixture that includes human cells
and up to 1.5 million different chemicals per day. The resulting data show
toxicity, purity, and metabolic activity.
"It gives you a glimpse of what's at the end of the tunnel," says Kalypsys CEO
John McKearn.
Right now, the drive toward early screening is motivated by profit (and an
increasingly anxious population of pill-poppers - thank you, fen-phen and
Vioxx). But drugmakers also see a broader future in medicines tailored to
individual genetic variations. Already Affymetrix makes an index-card-sized
plate that exposes DNA snippets from thousands of different genes to RNA from a
drug-treated cell to check which genes get turned on or off. "You really need to
look at the big picture of what's going on in the body to know if a drug works,"
says Affymetrix's John Blume. "In the late '80s, we could look at eight or 12
genes at a time. Now we can look at more than 30,000."
What that means is researchers can collect data on how a drug might affect
people based on their genes. The Food and Drug Administration has started asking
companies to voluntarily share pharmacogenomic information. "The hope is to
avoid problems down the line in the human population," says Felix Frueh,
associate director for genomics at the FDA. And hey, if they save a few guinea
pigs along the way, so much the better.
- Eilene Zimmerman
Alan
http://unitedeuropeanworkersunion.blogspot.com/
http://theoriginalfirebird.blogspot.com/
http://lordcerneabbastoo.blogspot.com/
http://lordcerneabbas.blogspot.com/
http://veloceraptor.blogspot.com/
http://www.stopwar.org.uk/
|
|