> Yanshu Wang and Jeremy Nathans at work in the lab.

In the late 1980s, many molecular biologists took up fishing, and Jeremy Nathans was one of them. “It's common now, but back then finding genes by high throughput sequencing was just beginning, depending as it did on development of automated genome sequencing machines,” he says. “So we went looking for genes important in retinal development. There was no other way to find them aside from sequencing.” Several of the genes that early search uncovered looked interesting, but only one—Frizzled—turned out to be the equivalent of Melville's great white whale, an elusive muse that Nathans and his students have pursued intensely through the ensuing years, with powerful results.

All Frizzled proteins appear to serve as cell surface receptors, and one, Frizzled-4, and its ligand, Norrin, play a central role in vascular development in the eye and ear, Nathans and colleagues discovered. Moreover, defects in the Frizzled-4-Norrin signaling pathway create serious retinal blood vessel disorders, some leading to blindness. Learning that hasn't come easily, says Nathans, “though it's all obvious, of course, after the fact.” It's taken his lab several years to conclude that the congenital Norrie disease—which produces complete blindness and progressive deafness—and familial exudative vitreoretinopathy (FEVR) are related. Though FEVR may impair vision only slightly, and by that mildness seem far removed from Norrie, the Nathans lab found both are tied to Frizzled-4-Norrin signaling gone awry.

“The common path was a wild guess,” says Yanshu Wang, who has worked with Nathans for 17 years, first as a graduate student, then as a postdoc and more recently as a Howard Hughes Institute research specialist. “But it turned out to be right.”

As a postdoctoral fellow, Wang showed that the Frizzled family of genes was conserved across species from nematodes to humans, suggesting a major role in development. “At that point we had no idea what the function of the gene was,” says Wang. “I'd chosen to work on Frizzled-4 mostly because it was expressed in the retina, and this is a retina lab.” The Frizzled-4 knockout mice had problems—esophageal enlargement and dysfunction, progressive hearing loss, cerebellar degeneration, ataxia. And though their retinas looked nearly normal, Wang discovered, after examining the tissue, that Frizzled-4 knockout mice completely lacked intraretinal capillaries. To compensate for that loss, arteries and veins had proliferated madly and were swollen, distorted and bleeding.

Soon after, a group of researchers in Nova Scotia linked mutations in Frizzled-4 to a hereditary eye disease characterized by retinal bleeding—familial exudative vitreoretinopathy. Though FEVR exists in both mild and severe forms, both involve some degree of capillary failure, hypervascularization and bleeding, just like Wang's knockout mice.

Intrigued by the parallels between the human disease and symptoms that Wang noted in the mice, Nathans combed the clinical literature, looking at other eye diseases related to vascular development, such as diabetic retinopathy and macular degeneration. But when he came to literature on Norrie disease, he made an intuitive leap. Like FEVR, Norrie disease is characterized by pathologic growth of new blood vessels. Unlike FEVR, its damage is far more dramatic. People with Norrie disease are often born blind. And though Norrie disease is linked to a mutated gene—Norrin—its gene product, is a secreted protein, not a receptor. Could Norrin and Frizzled-4 be related? Was a common pathway involved?

The Nathans lab swung into action. Nathans asked postdoctoral fellow Qiang Xu to look for biochemical connections, and the results astonished lab members. Not only did Frizzled-4 bind very tightly to Norrin, but when Xu did a gene activation assay and added an Lrp co-receptor to Frizzled and Norrin, the assay lit up like a Christmas tree. “We got a thousandfold response compared to the usual tenfold or hundredfold response that ensues when a Frizzled is combined with signaling ligands like Wnt,” says Wang. The results were so dramatic that Nathans repeated some of Xu's experiments himself just to be sure. “Jeremy wanted to look for anything that might be an artifact,” Wang says. “He said, ‘It's too good to be true.'”

Keeping an eye on the clinical literature helped reveal connections that might otherwise have eluded him, Nathans adds. And collaborations with clinicians have also been important. “A lot of the work we do would not have happened without their insights.”