Our laboratory is interested in the mammalian visual system and, in particular, the retina, the light-absorbing sheet of cells that lines the back of the eye. Our general approach is to use the tools of molecular genetics to identify and study genes involved in development, function, and disease. A second interest is in elucidating the mechanisms of pattern formation during animal development and, in particular, the role of the Frizzled family of cell-surface receptors. These two interests have converged with our discovery of a Frizzled-based system for controlling the development and integrity of retinal blood vessels.
Mammalian genomes encode 10 distinct Frizzled receptors. Fifteen years ago, we showed, in collaboration with the laboratory of Roel Nusse (HHMI, Stanford University), that Frizzleds are the receptors for the Wnt family of ligands. Mammalian genomes encode 19 Wnts, and the ligand-receptor relationships among Wnts and Frizzleds are complex. We have constructed knockout and/or conditional knockout mice for each of the 10 Frizzleds to define their roles, singly and in combination, in mammalian development. This summary focuses on experiments that reveal the roles of Frizzled3 and Frizzled6 in the development of the brain, spinal cord, inner ear, and skin; Frizzled1, Frizzled2, and Frizzled7 in the development of the heart and palate; and Frizzled4 in the development of the vasculature.
Mice lacking Frizzled6 exhibit unusual hair patterns. Frizzled6 is expressed in hair follicles and in the skin. In its absence, these structures look normal under a microscope, but instead of the normal pattern of parallel hairs, the hairs over much of the body surface are organized into large-scale patterns, such as whorls. In analyzing the developmental origin of these macroscopic patterns, we observe that the nearly parallel alignment of hairs on the body of a wild-type mouse is produced by a reorientation of follicles that are subtly misaligned at birth, and that the diverse Frizzled6 mutant hair patterns are produced by a much larger reorientation of follicles that are severely misaligned at birth.
These observations show that, despite their large size, developing hair follicles display substantial plasticity, reorienting within the dermis in what appears to be a self-organized process. The essential features of this process can be studied with simple cellular automata models in which a local consensus "rule" acts iteratively to bias each hair's orientation in favor of the average orientation of its neighbors. These experiments define two systems for hair orientation: a global system that acts early in development and is dependent on Frizzled6 and a local self-organizing system that acts later and is independent of Frizzled6. Several hair follicle–associated structures, including the muscle that mediates hair erection ("goose bumps") and axons that asymmetrically innervate the hair follicle, reorient along with their target hair follicle.
Complex hair patterns are prevalent in a variety of mammals, including humans. In many species, including our own, individual variations in hair patterning are well documented. At present, the molecular and cellular mechanisms responsible for orienting hair follicles are largely unknown. We suspect that interactions between the individual elements in a variety of complex biological structures—such as cells within an epithelium or axons within a fiber tract—use some of the same "rules" that we have observed in the hair follicle system, even if the detailed molecular mechanisms differ. It is interesting that local consensus interactions, conceptually similar to those that align hair follicles, influence group behavior in a variety of complex social systems, including locust migration and fish schooling. These interactions presumably represent an economical way to propagate global signals across a population and to enhance precision in the context of imprecise individual responses.
Frizzled3 controls the directional growth of axons in the embryonic brain and spinal cord. Although hair patterning and axonal navigation may seem like unrelated processes, both require local structures (hairs or axons) to sense the larger environment around them. Frizzled3 and Frizzled6 are highly homologous, and mice that are missing both genes exhibit additional defects not seen with either single mutant. In particular, their neural tubes fail to close, a common congenital defect in humans. Double-mutant mice also exhibit a defect in the orientation of sensory cells in the inner ear. We suspect that each of these developmental processes—neural tube closure, axonal path finding, hair patterning, and orienting inner ear sensory cells—utilizes the same basic molecular machinery to read out the spatial coordinates of the tissues in which they operate. This system appears to be the body's equivalent of the global positioning system—a system that provides positional information to cells that must coordinate their movements or orientations with the overall body plan.
Frizzled1, Frizzled2, and Frizzled7 also act redundantly and are important in a different set of developmental processes: closure of the palate, the structure that forms the roof of the mouth, and closure of the ventricular septum, the wall that separates the left and right ventricles of the heart. Defects in these two closure processes are among the most common birth defects in humans. The observation that Frizzled signaling plays a central role in a variety of tissue closure processes suggests that these processes share a fundamental mechanistic similarity despite their diverse anatomic contexts.
The function of Frizzled4 is quite different from the functions of the Frizzleds described previously. Loss of Frizzled4, in either mice or humans, results in a failure to fully develop a vascular system in the retina. We discovered that Frizzled4 is the receptor for an unusual ligand, Norrin, that is structurally unrelated to the Wnt proteins, which, as already noted, are the principal ligands for the Frizzled receptor family. In humans and mice, mutations in Norrin also cause hypovascularization of the retina, as do mutations in two Frizzled-associated membrane proteins, Lrp5 and Tspan12, that are part of a signaling complex. Norrin is produced by glial cells in the retina and brain, and Frizzled4 is expressed on vascular endothelial cells. We have recently found that Norrin-Frizzled4 signaling also plays a central role in maintaining the blood-brain and blood-retina barriers. This work defines a signaling system that controls central nervous system vascular development and homeostasis, and it unifies a set of previously unrelated disorders of human retinal vascular development.
Grants from the National Institutes of Health, the Foundation Fighting Blindness, and the Ellison Medical Foundation provided partial support for these projects.
As of April 29, 2016