In one tank at the zebrafish facility at Harvard's Cardiovascular Research Center, fish just under three weeks old dart around like small slits in the water, each barely the length of a newborn human's fingernail. Hundreds of other tanks, most containing a transgenic or mutant line, fill the room. Next door in the main lab, Randall Peterson, whose group shares the facility with four others, pulls a Petri dish containing a 48-hour-old fish out of an incubator. Under a microscope, it is completely translucent, its eyes and heart well formed, blood pumping vigorously and tail extending straight back like an arrow. A fluorescent transgene selectively labels the organism's myocytes, outlining the two chambers of its heart with red and enabling Peterson to monitor the effects of different chemicals on the heart.
Peterson became interested in using zebrafish to screen for novel drug candidates in 1999. Increasingly, drug discovery researchers are following his lead, reasoning that the model affords a quick, easy, and cheap way to see the effects of chemicals on a living organism early in the discovery process.
Four years ago, a colleague convinced Ulrich Rodeck, a zebrafish neophyte at Philadelphia's Thomas Jefferson University Hospital, to try the model as a "whole-body readout" of epidermal growth factor receptor blockade. A couple weeks later, the student working on the project waltzed into the lab and said, "Doc, want to see some fish?" Looking into the microscope, Rodeck says, "I could see the heartbeat, the blood, everything that you wanted." Rodeck, part of a researcher-physician team with Adam Dicker, was sold. The duo is using the model to look for compounds protecting cells from radiation and chemotherapy toxicity (Cell Cycle, 7:1-7, 2008).
The zebrafish's translucence during development, prolific breeding, and short development time give quick results. Zebrafish are orders of magnitude cheaper than mice, and they are easier to raise. They're also vertebrates, with a relatively close homology to humans. "It's the only model system that has organ systems that are comparable to ours, that also fit into 96-well plates," says Michael Pack, a physician and researcher at the University of Pennsylvania. Plunk a single embryo or young fish into each well, add your compounds, and "you can screen for drugs against the disease, as opposed to against some theoretical target," says Roger Cone, CEO of Znomics, a biotech using a transgenic fat zebrafish as an obesity model (FASEB J, 21:2042-9, 2007). "You can really just imagine zebrafish as a really complicated cell line."
Cone started Znomics in 2002 with his former PhD student, Wenbiao Chen, who had developed an efficient technique for making transgenic zebrafish. The company sells its library of more than 11,000 lines of zebrafish, 354 of which contain genes known to be involved in human diseases. In November 2007, it received $4.88 million in venture funding to pursue drug discovery. Another small biotech, UK-based Summit, is also using zebrafish for drug discovery.
No drugs discovered in zebrafish have been commercialized yet, or are even in clinical trials, so most Big Pharma companies have so far stayed away from setting up their own fish facilities for drug discovery. The main exception is Novartis; Mark Fishman, one of the first researchers to use zebrafish to model human disease and screen for novel drugs (and postdoc advisor to Peterson and Pack), established a discovery group at Novartis in 2002 to investigate axon demyelination and other disease models.
But promising drug candidates are beginning to crop up in academic labs. Peterson recently used zebrafish to identify a novel group of compounds with the potential for treating anemia (Nature Chem Biol, 4:33-41, 2008). Now, he says he's excited about looking for drugs for psychiatric disorders, which haven't yet been modeled in fish. Leonard Zon, also at Harvard, has been working on zebrafish disease models since the mid-1990s. Recently, he identified a prostaglandin that boosts stem cell production in cord blood, for which he will soon begin clinical testing (Nature, 447:1007-11, 2007). "Basically it's taken us three to four years to go from basic discovery" to the clinic, says Zon, when most drugs take a decade or more. "That time frame is very attractive."