The complexity of the afterpotential of rabbit A-type horizontal cells.

Afterpotentials of A-type horizontal cells (HCs) are believed to be rod-induced. They are, however, complex potentials and evidently of multiple causation. That part of the HC potential immediately after light-off is not entirely rod-determined because it has the same spectral sensitivity as the response to light-on, which is cone-induced with only some rod influence. It persists during a strong blue adapting light, which suppresses rod activity. The afterpotential may also be influenced by feedback from HCs to photoreceptors. The later part of the afterpotentials of A-type HCs is, however, rod dominated, as are the afterpotentials of axon terminals of B-type HCs.

The effects of GABA and vigabatrin on horizontal cell responses to light and the effect of vigabatrin on the electroretinogram.

When used as an antiepileptic drug in humans vigabatrin, which is a GABA analogue and an inhibitor of GABA-aminotransferase, often causes peripheral visual field loss. This could result from increases in endogenous GABA levels. Accordingly we have investigated the effects of GABA on horizontal cells (HCs) of the rabbit retina, and of vigabatrin, when applied for only a few minutes, on HCs and on the electroretinogram (ERG). The intracellular HC and ERG records were first obtained from isolated rabbit retinas during perfusion with a physiological solution. The perfusate was then changed to one containing GABA (2 mM) or vigabatrin (25, 40 or 150 microM) for at least 5 min, and then returned to the control solution. 2 mM GABA significantly but reversibly reduced the light responses of HCs elicited by diffuse light (at -4 log intensity) to 52 +/- 17% (SD, n = 7). Vigabatrin had no significant effect on the light responses of HCs (n = 7), and no effect on the b-wave (n = 4), but the PIII-component of the ERG was slightly but significantly reduced to 84 +/- 5% (SD, n = 5). The high dosage of GABA needed to affect the light responses of HCs could be due to strong GABA uptake systems in the intact rabbit retina. It is, however, possible that in humans receiving long-term treatment with vigabatrin, high levels of GABA occur because of the inhibition of GABA- aminotransferase. It seems, from these observations, that neurons like on-bipolar cells, which are contributors to the b-wave, and HCs are uninfluenced by vigabatrin in short-term experiments. The slightly reduced slow PIII-component, however, indicates an influence on the glial Müller cells which are the main contributors to the slow PIII-component.

The influence of HEPES on light responses of rabbit horizontal cells.

HEPES-buffered solutions, mostly used in studies of isolated cells, and bicarbonate-buffered solutions, mostly used in studies of isolated retinal tissues, have both been used to superfuse an isolated rabbit retina preparation. The responses of horizontal cells (HCs) to light, detected by intracellular microelectrodes filled with Lucifer Yellow, were recorded. Buffering of the superfusate with 100% HEPES completely, but reversibly, abolished the responses of A-type HCs, and is not, therefore, suitable for studies on isolated rabbit retinas. The responses remained when buffering was partially with HEPES and partially with bicarbonate, but were changed: in A-type HCs the overshoot was reduced and the afterpotential was increased. The overshoot may be caused by feedback of HCs on the cones and might be dependent on pHi at the synaptic structure between HCs and photoreceptors.

On-bipolar cells and depolarising third-order neurons as the origin of the ERG-b-wave in the RCS rat.

In the retinas of Royal College of Surgeons (RCS) rats light induces an increase in distal extracellular potassium irrespective of the age, between days 19-24 and days 29-35 postpartum, but by days 29-35 the ERG b-wave has become reduced. The synaptic blocker 2-amino-4-phosphonobutyric acid (APB) causes the abolition of both the b-wave and the potassium increase at any age. MgCl2 greatly reduces the b-wave at all ages and abolishes the potassium increase in older rats, but in younger rats the potassium increase is enlarged. Since this increase occurs in the absence of the b-wave it is unlikely that the on-bipolar cells are the only sources of the b-wave. Because the NMDA receptor blocker ketamine reduces the b-wave, third order neurons, which possess NMDA receptors, could contribute to the b-wave.

Low b-wave amplitudes in a strain of rabbits with a pigment epithelium defect.

When preparing isolated rabbit retinas we found in some animals fundi which were not uniformly dark but had abnormal areas of red coloration. The in situ electroretinograms (ERG) of 82 rabbits recorded after 1 h of dark adaptation were checked for abnormalities indicative of a degenerative disorder. The ERGs of eight rabbits with small dark adapted b-waves (< or = 250 microV) were re-recorded and their b-waves found to decline with time. The greatest reduction, in three rabbits, was > or = 150 microV over 2.5 years. After 1 year, however, the light adapted b-waves were similar to those of rabbits with normal dark adapted b-waves. The majority of the progeny of these rabbits also had small b-waves, which became still smaller in 2 years. Ultrastructural studies of two rabbit retinas of the first generation showed pathological changes of the pigment epithelium (Wrigstad, Hanitzsch & Nilsson, Ultrastructural and electrophysiological studies of the retina and the retinal pigment epithelium in rabbits with low b-wave amplitudes, in preparation). Evidently there is an inheritable defect in the pigment epithelium which first impairs the rod pathway.

Impaired function of bipolar cells in the Royal College of Surgeons rat.

The retina of the Royal College of Surgeons (RCS) strain of rat, which is being used as an animal model for human retinal degenerations, has been employed in the study of the function of second order neurons. By about the 33rd postnatal day the dendritic branching of isolated bipolar cells is more sparse than in bipolar cells of the normal rat retina, but their GABA channels are as in the normal rat retina. The normally occurring light-induced distal potassium increase has been used as the indicator of the functional competence of second order neurons in the isolated RCS rat retina. These are dependent upon the integrity of ionotropic and metabotropic synapses. At about the 22nd postnatal day MgCl2 enlarges the light-induced distal potassium increase in the young RCS rat retina as in the normal rat retina. It seems that MgCl2 does not block the metabotropic synapses of on-bipolar cells. At about postnatal day 33, at which time the photoreceptors of the RCS rat retina had become severely damaged, the size of this light-induced distal potassium increase was not changed, but it was abolished by MgCl2. This indicates that bipolar cells are still active but that the synaptic function of on-bipolar cells has become vulnerable to MgCl2. The conclusion is that at a time when photoreceptor degeneration is already severe bipolar cells are still active, but that on-bipolars, mainly rod bipolar cells, have some functional deficit.

Light potentiation of horizontal cells in the isolated rabbit retina.

When the retinas of some fishes and amphibians are dark-adapted the hyperpolarising response of horizontal cells (HCs) to a light stimulus is suppressed (= dark suppression) but is restored to that of light adaptation when a light stimulus is rapidly repeated (= light potentiation (LP). This phenomenon, which had not been previously demonstrated in a mammal, has been recorded from isolated rabbit retinas. The steady amount of LP of HCs after several light stimuli was expressed as the percentage of induced hyperpolarisation when that of the dark-adapted retina is taken as 100%. The LP was small (113+/-13%, n=14), when the HCs had stable and large responses (>15 mV), but varied greatly between HCs. Those HCs with slight LP (106+/-3% n=8) designated as x-HCs, mostly exhibited larger overshoots than did those HCs with stronger LP (123+/-14, n=6) and designated y-HCs. These mostly had a smaller overshoot. Other HCs (z-HCs) were unstable and were only slightly hyperpolarised by the first light stimulus after dark adaptation but showed the strongest LP (193+/-48%, n=7). These results could indicate a variable degree of light potentiation in normally-functioning HCs, which can be classified accordingly. It is possible, however, that the degree of LP is small in all normally-functioning HCs, but as the HCs in isolated mammalian retinal preparations deteriorate, the phenomenon of LP is progressively exaggerated. LP is not peculiar to HCs and can also occur in cells which depolarise in response to a light stimulus, and can be evident in the PIII component of the ERG.

Do horizontal cells contribute to the rabbit ERG?

An increase in the amplitude of horizontal cell (HC) potentials, a decrease in the extracellular potassium around the photoreceptors and an increase in the PIII component of the ERG that occur with increasing light stimulus intensity are shown to be concomitant occurrences, which could, therefore, be causally related. With high intensity light stimuli the PIII component is prominent and the HC potentials exhibit an afterpotential. Within the range of intensities tested the peak times of the HC afterpotentials coincide with the peak times of PIII. This indicates the likelihood that the HC potential contributes to the PIII, although these results do not allow a quantitative assessment of the contribution of HC potentials to the PIII-component.

Two neuronal retinal components of the electroretinogram c-wave.

Although the rising phase of the b-wave seems to be generated mainly in the rod bipolar cells and the cone on-bipolar cells, the slow component of the electroretinogram, the c-wave, evidently originates in the Müller cells and the pigment epithelium. The c-wave has three components. One cornea-positive component derives from the pigment epithelium, while a distal cornea-negative component (slow PIII) and a proximal slow component originate in the Müller cells. This third proximal component of the c-wave differs between mammalian species: it is negative in the rat retina, positive in the rabbit and human retina and may be lacking in the cat retina.

The influence of MgCl2 and APB on the light-induced potassium changes and the ERG b-wave of the isolated superfused rat retina.

The ERG and the extracellular potassium concentration, [K+]o, of the isolated superfused rat retina were measured in a physiological solution and in solutions containing 10 mM MgCl2 or 100 mu M APB. MgCl2 nearly abolished the b-wave, but the light-induced distal [K+]o increase was enlarged from 0.13 +/- 0.05 to 0.28 +/- 0.08 mM. There was also an increase in the light-induced [K+]o in the proximal retina. APB abolished the b-wave completely, and the distal light-induced [K+]o increase was then replaced by a [K+]o decrease. Upon return to the control solution, there was a larger transitory [K+]o increase than under control conditions, and this occurred before the b-wave had returned. Under these experimental conditions, the distal [K+]o increase could not be correlated with the b-wave, and so the Muller cells are unlikely to be the main source of the rising phase of the b-wave. More probable sources of the b-wave are the on-bipolar cells with their metabotropic and ionotropic receptors, with only the latter apparently being blocked by MgCl2. The extracellular [K+]o changes, however, had an influence upon the slow potentials of the ERG.

The c-wave of the electroretinogram possesses a third component from the proximal retina.

ERG and light-induced extracellular potassium ([K+]o) changes have been measured in isolated retinas of both Rana esculenta and Rana temporaria. The conditions of the preparations have been varied. Isolated frog retinas kept receptor side-upward in a moist chamber without perfusion showed the well-known slow PIII in the ERG. Retinas superfused from the receptor side, with O2 enrichment at their vitreal surface, however, exhibit a slow cornea-positive potential in the ERG. The slow ERG-potentials relate to different light-induced potassium changes in the proximal retina. There was a long lasting and larger proximal potassium increase in adequately maintained retinas but a smaller and shorter one in preparations lacking superfusion and oxygen. There was no significant difference between the size of potassium decrease around receptors of retinas superfused from their vitreal side and those superfused from receptor side. A reduction of slow PIII should therefore not be responsible for the slow cornea-positive potential. The long lasting and larger (by 59%) potassium increase in the proximal retina may counteract the potential in the Müller cells caused by the potassium decrease around receptors and thereby cancel slow PIII and generate a third component of the electroretinogram c-wave.

Comparison between the slow cornea-negative PIII component of the ERG and potassium changes in the isolated rabbit retina.

Light-induced extracellular potassium changes were measured in the isolated rabbit retina superfused by a plasma saline mixture and compared with the electroretinogram. When the transmission to second-order neurons was blocked by aspartate and glutamate or by Mg2+, the electroretinogram consisted of the receptor potential and the cornea-negative slow PIII. Since the onset of PIII could then be seen to precede the decrease in extracellular potassium concentration ([K+]0) around photoreceptors, the [K+]0 decrease could not be the cause of the onset of slow PIII. A possible source for the initial phase of slow PIII could be the electrogenic Na+/bicarbonate transporter mechanism of glial cells. Slow PIII depended highly on the extracellular sodium concentration, and it was larger in solutions buffered with bicarbonate than with HEPES, while the [K+]0 decrease around receptors was unchanged.

The neural retina of the frog contributes a slow cornea-positive potential to the electroretinogram.

A c-wave-like cornea-positive potential in the isolated rabbit retina has been described. In this study, frog retinas were investigated to see if the neural retina contributes a slow cornea-positive component to the c-wave of the electroretinogram. The eye cups of both Rana esculenta and Rana temporaria exhibited a normal electroretinogram with c-wave, a larger proximal light-induced extracellular potassium increase, a small distal extracellular potassium increase and an extracellular potassium decrease around photoreceptors. Isolated frog retinas kept receptor side-upward in a moist chamber without perfusion showed the well-known slow PIII generated by the potassium decrease around receptors. If the isolated retinas were well perfused, the slow PIII was not seen, but a cornea-positive d.c. potential sometimes appeared after the b-wave. The different slow potentials seemed to relate to different light-induced potassium changes in the proximal retina. There was a long-lasting proximal potassium increase in the superfused retinas but a quick return of the proximal potassium increase to the baseline in the retinas lacking oxygen at the vitreal side. The lasting proximal potassium increase in adequately maintained retinas may counteract the potassium decrease around receptors and cancel slow PIII.

Measurements of the extracellular potassium concentrations in the isolated rabbit retina with different kinds of potassium-sensitive microelectrodes.

When the dark-adapted isolated rabbit retina has been superfused from one side with a plasma-saline mixture containing 3.5 mM potassium, the extracellular potassium concentration within the dark-adapted retina was significantly higher (4.5-6.1 mM) than the potassium of the superfusate, when measured with an electrode filled with Corning 477317. The substantial difference between the [K+]0 of the perfusate and the [K+]0 within the retina is difficult to explain and could be an instrumental artefact. To probe this possibility measurements have now been repeated using 2 different K+ electrodes, one filled with Corning 477317 and the other filled with a valinomycin-based ion exchanger. With the latter electrodes there was practically no gradient between the [K+]0 in the retina (3.2-4.0 mM) and the potassium concentration of the superfusate. It is thus evident that the earlier enigmatic values for [K+]0 relate to the use of Corning 477317-filled K+ electrodes. When, however, changes in [K+]0 were induced by a light stimulus, the measured magnitudes of change in [K+]0 were the same with the 2 types of electrode.

Changes in [K+]0 at the vitreal surface compared with those around receptors in the isolated rabbit retina.

Isolated rabbit retinas were superfused from the receptor side with a plasma-saline mixture kept at 35 degrees C. The vitreal side was exposed to an atmosphere of humidified warm oxygen. In one study the second-order neuronal activity was suppressed with aspartate and glutamate; in another study transmission was not blocked. When all neurons were active, [K+]0 around receptors was 4.5 +/- 0.4 mM in the dark. During a long (60s) exposure to light stimulus, [K+]0 dropped to 73% of the dark value and reaccumulated to 80%. At the vitreal surface, [K+]0 in the dark was 4.7 +/- 0.8 mM. During the 60s light stimulus, [K+]0 increased transiently, dropped to 83% of the dark value, then increased again to 91%. A continuous decrease of [K+]0 at the vitreal surface during long light stimuli concurrent with the increase of [K+]0 around receptors would indicate that the spatial buffering capability of the Müller cells contributes to the reaccumulation of potassium. Such a decrease, however, was not detected. After the blockage of transmission, [K+]0 values did not vary significantly from those after light stimulus in unblocked preparations. In the dark, [K+]0 was 5.2 +/- 0.9 mM at the vitreal surface and 4.6 +/- 0.4 mM around the receptors.

The time course of the light-induced extracellular potassium change around receptors and at the vitreal surface compared with the time course of slow PIII wave in the isolated rabbit retina.

The experimental preparation was a piece of isolated rabbit retina, the lower side of which was superfused in a chamber with a plasma-saline mixture kept at 35 degrees C to which, in some experiments, either sodium aspartate and glutamate, or Mg2+, were added to suppress second order neuronal activity. The upper side was exposed to an atmosphere of humidified oxygen. With longlasting light stimuli (30-180 s) the [K+]0 around receptors first falls from its dark-adaptation value (4.6 +/- 0.4 mM) to 3.3 +/- 0.6 mM and then reaccumulates nearly to the pre-stimulus value. The time course of reaccumulation varies greatly (30-180 s) between preparations. The peak time (90% value) of the [K+]0 decrease is several seconds. It seems that the more rapidly the retina is prepared the shorter the peak time of the [K+]0 reaccumulation and the more rapid the [K+]0 reaccumulation. This suggests that the [K+]0 reaccumulation is effected by active, oxygen-dependent mechanisms. PIII always precedes the [K+]0 change, its peak time is several 100 ms. During longlasting light stimuli (30-180 s) slow PIII reflects the difference in the [K+]0 changes around the photoreceptors and at the vitreal surface. The rapidly and carefully isolated preparations exhibit a slow cornea-positive potential similar to a c-wave, when second order neuronal activity is not suppressed.

Coding of light intensity and stimulus duration in the receptor potential of the isolated rabbit retina.

The receptor potential of the isolated rabbit retina was recorded between a microelectrode approximately 80 microns deep from the inner limiting membrane and a reference electrode at the receptor side. Amplitude and slope of the receptor potential and the area under it were measured as a function of increasing stimulus intensity and the amplitude and duration of the receptor potential as a function of stimulus duration. The amplitude depends on the intensity and duration of the stimulus. The slope is the single parameter which changes only with light intensity. The slope follows a power function with n = 0.36. The duration of rabbit receptor potential is linearly related to the duration of long stimuli (greater than 1 s for intensities tested).

A chamber preserving cellular function of the isolated rabbit retina suited for extracellular and intracellular recordings.

The isolated retina of rabbit lies on nylon tissue fixed in a ring, which is placed in a chamber. The lower side of the retina, by choice the receptor or ganglion cell side, is superfused by a plasma saline mixture kept at 35 degrees C. At the upper side of the retina, warm humidified oxygen is added. The two sides are isolated from each other by a rubber joint. ERGs with b waves of about 1 mV indicate a good function of retinal cells. The receptor potential is recorded extracellularly between an intraretinal microelectrode and a reference electrode at the receptor side (amplitude typically about 300 microV). The few intracellular records were hyperpolarizations to white and red light with an initial overshoot and a plateau.

Dependence of the b-wave on the potassium concentration in the isolated superfused rabbit retina.

The amplitude of the b-wave of the isolated superfused rabbit retina is drastically reduced with increasing potassium concentration (10 and 20 mM respectively) in the perfusate like in frog retina. These results are in agreement with the idea of the glial origin of the b-wave, but an influence of potassium on synaptic transmission remains a possibility. For these results the conditions for tissue survival are imperative. When the retina was superfused with a plasma saline mixture kept at 35 degrees C, b-wave amplitudes for different preparations varied between 300 microV and 900 microV and loss of sensitivity was tolerated till 15% in one preparation. The temperature quotient for the amplitude of b-wave was 4-6 between 35 degrees and 25 degrees C, for the peak time about two.