What's New:

Light-sensing cells in retina develop before vision

By Jim Dryden

Dec. 21, 2005 — Investigators at Washington University School of Medicine in St. Louis have found that cells making up a non-visual system in the eye are in place and functioning long before the rods and cones that process light into vision. The discovery should help scientists learn more about the eye's non-visual functions such as the synchronization of the body's internal, circadian clock, the pupil's responses to light and light-regulated release of hormones.

The researchers report in the Dec. 22 issue of Neuron that in the mouse retina, intrinsically photosensitive retinal ganglion cells (ipRGCs) are active and functioning at birth. That was surprising because the mouse retina doesn't develop fully until a mouse is almost three weeks old, and the first rod cells don't appear until about 10 days after birth.

"We were stunned to find these photoreceptors were firing action potentials on the day of birth," says Russell N. Van Gelder, M.D., Ph.D., associate professor of ophthalmology and visual sciences and of molecular biology and pharmacology. "Mice are very immature when they're born. It takes about three weeks after birth for the retina to fully develop. No one previously had detected light-dependent cell firing in a mouse before 10 days."

Van Gelder says the ganglion cells react to light in two ways, sending messages to parts of the brain that control circadian rhythms, and (on the first day or two of life) also setting off a wave of activity that spreads through the retina, possibly helping visual cells develop.

Van Gelder and colleagues have spent the last few years learning how blind animals (and people) can sense light and use it to set their circadian clocks. The ipRGCs were first identified in 2002 — by David M. Berson, Ph.D., and colleagues at Brown University — as the cells that could sense light even in visually blind eyes. But it was very difficult and time consuming to isolate and study the cells, requiring precise injection of a tracing dye into the brains of animals to label and identify the ipRGCs.

That has changed as the result of a technical advance developed by Daniel C. Tu and Donald Zhang, both Medical Scientist Training Program students in Van Gelder's lab, and co-first authors of this study. Tu and Zhang used a multi-electrode array technique in which tiny, individual electrodes are placed about 200 microns apart. Each electrode is a mere 30 microns in size — there are 25,400 microns per inch —and 60 electrodes are contained on a grid.

"This spacing turns out to be perfect for a retina," Van Gelder says. "You can remove the retina and place it, ganglion cell-side down, on this array. Then the electrodes pick up the impulses of the ganglion cells when those cells react to light."

Whereas the original brain injection technique allowed researchers to study only one or two ipRGCs per day, the multi-electrode array allows Van Gelder's team to study 30 times that many. Those studies have revealed a cell population that reacts quickly and consistently to light.

"If you give the cells a series of identical pulses of light and look at how fast they fire, the reaction is identical every time," Van Gelder says. "The ganglion cells detect brightness, and they're extremely good at it. You could make a good light meter for a camera out of these cells because they are consistent in their response to brightness over the equivalent of almost 10 f-stops on a camera. That's completely different from the rods and cones in the retina. Those visual cells can't detect brightness very well. They detect contrast, sensitivity and motion."

Studying these populations of ipRGCs, Van Gelder also found the cells require a protein called melanopsin to sense and react to pulses of light. When the group examined retinas of mice that were genetically engineered to lack melanopsin, they found that the ganglion cells lost all sensitivity to light.

The ability to study many of these cells at once allowed Van Gelder's team to learn that there are three distinct populations of ipRGCs, and each cell type reacts to light differently. Some fire quickly when a light turns on but take longer to stop firing when it goes out. Other cells take a while to ramp up their response but then quickly stop firing when the area gets dark. A third cell type is slow to turn on when exposed to light and takes its time shutting down in darkness.

In addition, the cells tend to react to light in groups. Electrically, some of the cells work almost like a chorus, sending several synchronized "harmonies" to the brain as part of one big "song" that responds to light impulses.

"We were able to detect about 20 percent of the ganglion cells were coupled to other ganglion cells," he says. "That's probably a low estimate because if we had a finer grid and could record the activities of more individual cells, we might well find more interactions."

Van Gelder believes the early activity and the interactions of the ipRGCs may somehow enhance survival by helping animals detect light and set their circadian clocks prior to the development of vision. And he says because retinas tend to be very similar in most mammals, human ganglion cells also may develop and begin to function earlier than rods and cones.

Although ipRGCs sense light in mice and humans, they don't connect to the brain's visual cortex. Instead, they send signals to deeper, more ancient parts of the brain, such as the hypothalamus, from which they project to the brain regions that control the circadian clock as well as the response of the pupil to light.

"The multi-electrode array technique that Dan Tu and Don Zhang have brought into this field should help us learn a lot more about how these retinal ganglion cells influence all kinds of non-visual functions and reinforce the fact that the eye is responsible for more than just vision," Van Gelder says.