by Reto Weiler and Mark Pottek

Adaptation mechanisms allow the visual system to work perfectly under very different lighting conditions. There are a number of neuronal adaptation mechanisms in the retina that are dependent on information about the prevailing light conditions. We have recently discovered that retinoic acid may provide this information and could therefore be the long sought-after light switch in the eye. Retinoic acid is formed in the retina as a light-dependent by-product of the phototransduction cycle. Retinoic acid induces the formation of spinules, modulates the light responses and the electrical coupling of horizontal cells as well as a light adaptation of the retina.

Retinoic Acid: The Light Switch in the Eye?

Several mechanisms of adaptation allow the visual system its enormous working range. In the retina, there are a number of so-called neuronal adaptational mechanisms which all depend on the appropiate information about the ambient light situation. We recently discovered that retinoic acid might be the source of this information and therefore represent a light switch in the eye. In the retina, retinoic acid is produced as a light- dependent by-product of the phototransduction cycle. Retinoic acid induces the formation of spinules, modulates the light responses of horizontal cells and affects the electrical coupling of these cells. In all cases its effects match the effects of light adaptation.

Perhaps you are reading the lines of this article at your holiday destination somewhere in the sunny south on the beach in direct sunlight, or perhaps you have decided to make it your bedtime reading and are skimming the lines in the dim light of a bedside lamp. In both situations, you will have no trouble recognising the characters, even though the intensity of the lighting can vary a billion-fold. Obviously, our visual system can work perfectly within an extraordinarily wide range of intensities.

This everyday experience, which we take for granted, is a scientific challenge for neurobiologists. If we analyse the individual neuronal elements of the visual system, i.e. the photoreceptors and the neurons involved in processing, in the retina and in the visual centres of the brain in more detail, we find that they generally do not have such a large intensity range. Typically, they can only process an intensity range of around a thousand times that. In order for the visual system to be able to process a much larger range, special mechanisms must therefore be present, which are generally referred to as adaptation. These adaptation mechanisms are found in the eye and particularly in the retina. The control of pupil width is almost irrelevant for adaptation, as it does not even achieve a factor of 10. The fact that we have two types of photoreceptors, rods and cones, which differ in their sensitivity, is of greater importance, but is still not completely sufficient to explain the large intensity range. In addition, there are adaptation mechanisms of the retina, which we will refer to below as neuronal adaptation and which are based on the interactions of neurons.

A kind of "neuronal light switch"

The task of neuronal adaptation mechanisms is therefore to always adjust the relatively narrow working range of neuronal processing so that it corresponds as closely as possible to the prevailing light conditions. However, this means that the neuronal network of the retina must be informed about the prevailing light conditions. There must therefore be a "light signal" that can influence the interactions between the neurones, i.e. a kind of neuronal light switch. Signals that do this are known as "neuromodulators". Such a signal should be clearly dependent on the prevailing light conditions and be able to trigger certain processes in the retinal neurones that are known to correlate with light and dark adaptation.

In the search for such a light signal, our research group was the first to analyse retinoic acid. Why retinoic acid? In biology, retinoic acid is known as one of the most important factors controlling development. As a derivative of vitamin A, which is ingested with food, retinoic acid controls the development of the embryo by switching the transcription of certain genes on and off. Without retinoic acid there is therefore no normal development. Retinoic acid has also recently been used in the treatment of certain carcinomas and is also found in many cosmetics to prevent wrinkles. It is therefore a highly active biological molecule that could also be considered to have a neuromodulatory effect.

Retinoic acid in the eye

Retinoic acid is actually also produced in the eye. The visual pigment in the photoreceptors, which absorbs the light quanta, is made up of a protein, opsin, and a chromophore, retinaldehyde. This visual pigment (rhodopsin) breaks down into its two components when it has absorbed a quantum of light, triggering a biochemical cascade that ultimately results in a change in the release of the neurotransmitter glutamate by the photoreceptors. In order for the photoreceptors to continue to function, the decayed rhodopsin must be regenerated. Part of this regeneration process (dark regeneration) takes place in the pigment epithelial cells that line the outer limbs of the photoreceptors. These cells also contain the enzyme aldehyde dehydrogenase, which can oxidise retinaldehyde to retinoic acid. As the oxidation is irreversible, a certain amount of retinoic acid is produced depending on the concentration of retinaldehyde. Since the concentration of retinaldehyde depends directly on the decay of rhodopsin and thus on the light quanta that trigger this process, the concentration of retinoic acid must also depend directly on the prevailing light conditions. We and other research groups in the USA have recently succeeded in demonstrating this dependence directly.

Light-dependent synaptic plasticity

Retinoic acid therefore fulfils two important requirements for a light signal, namely to be produced in a light-correlated manner and to be biologically active. We therefore developed a research programme to analyse the physiological role of retinoic acid in the retina in more detail and to discover a putative neuromodulatory effect. Firstly, we investigated the effect of retinoic acid on the formation of spinules on the dendrites of retinal cells. The formation of these spinules is dependent on the adaptation state of the eye: In a light-adapted retina there are very many spinules, in a dark-adapted retina they have almost completely disappeared. In a cross-section through the synaptic terminal of the photoreceptor, the profiles of the horizontal cells that invaginate into the terminal of the photoreceptors can be clearly recognised under the electron microscope. The spinules emanating from these profiles are just as clearly recognisable in the light-adapted retina. In the dark-adapted retina, the profiles are rounded and spinules are no longer recognisable. Accordingly, the number of spinules is a measure correlated with the adaptation state of the retina and can therefore be used to analyse the potential signalling effect of a neuromodulator.

If retinoic acid were the appropriate light signal for the formation of spinules, then it should be possible to induce the formation of spinules in a dark-adapted eye by the injection of retinoic acid alone. We carried out these experiments, and it was indeed possible to induce spinule formation by retinoic acid that was no different from that induced by light. Of course, we carried out a whole series of control experiments, all of which confirmed the effect of retinoic acid. A very important finding was also the proof that by inhibiting the enzyme aldehyde dehydrogenase, which catalyses the synthesis of endogenous retinoic acid in vivo (see above), the direct light-dependent formation of spinules during normal light adaptation was inhibited.

Light responses of horizontal cells

It is possible to use an electrode to record intracellularly the membrane potential changes of horizontal cells as a result of light stimulation. These light responses show characteristic differences depending on the adaptation state of the retina. In a dark-adapted retina, stimulation with a small light spot only leads to a relatively small change in the membrane potential, whereas in the light-adapted retina the same stimulus leads to a much larger change. Conversely, the cell in the dark-adapted retina reacts more sensitively to stimulation with a larger, ring-shaped light stimulus than in the light-adapted retina. The ratio between the light responses to the two different stimuli can now be formed (A/S) and plotted as a function of the light intensities used. As can be seen, the two curves for light- and dark-adapted retinas differ very clearly. If a dark-adapted eye is treated with retinoic acid, the curve obtained then shows the same course as that for a light-adapted eye, even though it was kept dark-adapted the whole time. We can therefore conclude that retinoic acid changes the typical light response characteristics of horizontal cells in a dark-adapted retina as if this retina had been light-adapted.

Electrical synapses

Based on previous work from our laboratory and many others, we know that one of the reasons for the altered light response behaviour of horizontal cells in the light-adapted retina compared to the dark-adapted retina is the electrical coupling of these neurons. Horizontal cells are connected to each other via electrical synapses that enable a direct current flow from one cell to the other. These electrical synapses can be regulated, i.e. the current flow can be facilitated, impeded or stopped completely. In a light-adapted retina, the majority of these electrical synapses are closed and in a dark-adapted retina, the majority are open.

The state of the electrical synapses can be visualised directly: If a certain dye is injected into a cell using an electrode, this dye can spread into the neighbouring cells via the open electrical synapses; if they are closed, the dye remains in the injected cell. If the electrical synapses in the dark-adapted retina are open, an entire network of coupled, identical cells will be found; if the electrical synapses in the light-adapted retina are completely closed, only a single cell will be found.

It was now natural to investigate the extent to which the effect of retinoic acid on the light responses of the horizontal cells was also based on a change in the permeability of the electrical synapses. For this purpose, the dye was injected into individual horizontal cells both in dark-adapted retinas and in dark-adapted retinas treated with retinoic acid. If our assumption is correct that retinoic acid influences the light responses of the horizontal cells by changing the permeability of the electrical synapses between these cells, then one would expect that in the retinoic acid-treated retina a whole network of cells would no longer be labelled, but only very few or even just a single cell.

A fluorescence microscope image of sections of the corresponding retinas clearly shows that several cells are labelled in the dark-adapted retina, but only a single cell in the dark-adapted retinoic acid-treated retina. These findings show that retinoic acid can indeed directly regulate the permeability of electrical synapses. Also in this case, retinoic acid treatment simulates the effect of light, which, as mentioned above, reduces the permeability of electrical synapses.

The data presented so far have all been obtained in the fish retina, which has some experimental advantages. As we are interested in the generalisability of our findings, we have started to investigate the function of retinoic acid in the mammalian retina as well. We can already state that retinoic acid can modulate the electrical synapses between the horizontal cells in the rabbit retina as well as in the mouse retina. We were thus not only able to show that retinoic acid can modulate the light responses of horizontal cells in the same way as the ambient brightness, but were also able to uncover the first possible mechanism underlying this modulation, namely the regulation of the permeability of electrical synapses.

Conclusions for beach and bed

Let us summarise: We have shown that retinoic acid induces the formation of spinules, modulates the light responses of horizontal cells and thereby regulates the electrical coupling between cells. In all three cases, the effect of retinoic acid corresponds to that of light. Together with the finding that retinoic acid is formed in a light-correlated manner in the eye, these findings support the hypothesis that retinoic acid is the long-sought light signal for neuronal adaptation in the retina.

This reveals a new principle that once again demonstrates the elegance of biological problem solving: an inevitable, stimulus-correlated by-product of an essential biochemical reaction cascade is used as a modulator of the stimulus-processing network and thus guarantees the highest sensitivity of the visual system over a wide intensity range, or in other words, reading on the beach or in bed.

The authors

Prof. Dr Reto Weiler, Professor of Neurobiology and Ethology at the Department 7 of Biology and Environmental Sciences, studied biology at the University of Zurich and then became an assistant professor at the University of Munich, where he obtained his doctorate in 1977 and habilitated in 1982. His research focus is on analysing neuronal interaction in the retina at the cellular and molecular level. Weiler has received several awards for his pioneering work, including the Max Planck Research Award in 1990 and the International Research Award of the Australian Research Council in 1997. In Oldenburg, Weiler was instrumental in setting up the special research centre "Neurocognition". Mark Pottek studied biology in Oldenburg and completed his studies in 1994 with a Diplom thesis in the neurobiology research group. He was then a doctoral candidate in a project funded by the German Research Foundation in the same research group. He is currently writing his dissertation.