Most axons move out from the retina in a straight path to make their brain connections, crossing the brain at midline to connect with the opposite lobe of the brain. But other axons seem as if they are about to cross, only to have their torpedo-shaped tips splay out against the midline as if being thrown against an invisible wall. After a couple of tries, the axons eventually give up and turn back. These maneuvers of the tips of growing axons as they wend through the optic chiasm were documented at Columbia by means of video time-lapse by the lab of Dr. Carol Mason, professor of pathology and anatomy & cell biology in the Center for Neurobiology and Behavior.
What is it about crossing the midline that repels certain axons but not others? Dr. Mason and her team describe the reason in two papers published in Cell and Neuron. In addition, with the help of clinical geneticist Dr. Steve Brown, her lab has also discovered that a protein known to be involved in spinal cord and cerebellar development controls retinal axonal fate and may help explain the process of other similar midline crossings in the central nervous system.
For binocular vision to function, the brain must receive information from both eyes. Nerve fibers from each retina cross at the optic chiasm, at the midline of the brain. The right eye is wired to the left brain, and the left eye is wired to the right brain. A small number of axons, however, are repelled at midline and project to the same side of the brain. The "uncrossed" axonal connections are required for an animal's binocular vision: the ability to calculate how far objects, particularly prey, lie in the distance. The efficiency of an animal's binocular vision is a result of the proportion of uncrossed axons in the brain. In humans, who have fairly good binocular vision, about 40 percent of axons from the retina are uncrossed; in mice, which have poor binocular vision, only 3 percent are uncrossed. In most birds and fish, which lack binocular vision, no retinal axons are uncrossed.
What happens at the optic chiasm that repels a specific group of axons? As reported in the Sept. 11 issue of Neuron, the lab found that retinal axons destined to turn away from the brain midline detect cues displayed by special glial cells there, and the cues abruptly stop the axons from crossing over the midline.
Finding the cue, called ephrinB2, a ligand in a large family of receptor tyrosine kinases, wasn't a great surprise, since previous research in frogs had already determined that this family of molecules plays a role in directing the uncrossed retinal pathway. The lab, however, had to painstakingly sort through seven possible receptors to this ligand on retinal axons to learn which one they use to intercept the message.
"If my graduate student, Scott Williams, hadn't carefully sectioned the eye and looked for each and every candidate receptor at several different stages of development, we would have missed it," Dr. Mason says. Mr. Williams, a student in the Center for Neurobiology and Behavior, didn't find the right receptor, EphB1, until the very end of his search. The receptor is expressed only in the uncrossed axons during a brief (three day) developmental period in the mouse embryo.
Experiments with an EphB1 knockout mouse then proved the receptor was necessary for the re-direction of the axons away from the midline. In these animals, few axons wired the same side of the brain and instead, apparently crossed to the other side at the midline.
The researchers then wondered why only some neurons from the retina carry the EphB1 receptor. Hints came from work by the lab of Dr. Tom Jessell, professor of biochemistry & molecular biophysics, who has found motor neurons in the spinal cord express different combinations of transcription factors. These "codes" are specific to different subpopulations depending on which muscles in the limb they innervate. One such group of transcription factors was already known to control Eph receptors, but even though they are expressed in the eye, the Mason lab found that none of these appeared to designate the uncrossed group of retinal axons.
"We had abandoned our search for identifying genes," Dr. Mason says. "And then one day, Dr. Steve Brown, [assistant professor of obstetrics and gynecology] walked in and said he worked on midline developmental events." Dr. Brown had found that the Zic2 transcription factor helps demarcate the midline in the developing forebrain and head. Mutations in the gene in humans cause holoprosencephaly a rare disorder in which the embryonic forebrain fails to divide into two halves.
If the researchers hadn't already searched for more promising candidates, they never would have looked for Zic2, but postdoctoral researcher Dr. Elo92sa Herrera found that in the embryonic period when the retina sends axons to the brain, Zic2 comes on and appears to instruct axons to turn around at the midline. As reported in the Sept. 4 issue of Cell, when the amount of Zic2 is reduced in these cells, their axons cross instead of turning around in the optic chiasm. Further evidence that this gene directs the uncrossed retinal projection comes from comparative anatomical studies: In ferrets, mice, frogs and chicks, the numbers of retinal cells expressing Zic2 correlates exactly with the size and timing of the uncrossed retinal projection to the brain. Though not all of the evidence is in, the lab thinks Zic2 controls axon destiny by controlling the expression of the EphB1 receptor on the retinal cells.
"Zic2 is the first gene recognized to control axon avoidance of the midline. It will be interesting to see if Zic proteins are also involved in the midline decisions made by other types of neurons," Dr. Mason says.
The pioneering finding from Dr. Jessell's lab that different groups of neurons express specific transcriptional regulators, which in turn, regulate the direction and target of axon growth, "opens a new door on understanding how the circuitry of the nervous system is laid down," Dr. Mason says. As is true for many genes during development, the Zic gene seems to have a role in events that happen in the midline at radically different developmental timepoints if perturbed early, holoprosencephaly ensues; if perturbed in the eye later, binocular vision is impaired.