ABSTRACT
Following amputation of the eye of Strombus luhuanus, a morphologically equivalent one is regenerated. Early regenerates exhibited a relatively simple, single-peaked electroretinogram (ERG) that progressively matured to display two ‘on’ peaks, a ‘steady state’ during sustained illumination, and repetitive ‘off’ potentials. ERGs of regenerates were dependent on stimulus intensity, wavelength, light adaptation and temperature, as are those of mature eyes. Intracellularly recorded light-evoked depolarizations from retinas of early stages of regeneration tended to be simpler in waveform than those of more mature stages. In addition, certain cells exhibited light-evoked and electrically evoked action potentials, while others were hyperpolarized by light. Optic nerves of eyes in early stages of regeneration showed only ‘on’ activity, while those of later stages also exhibited ‘off’ responses.
INTRODUCTION
In the mature eye of Strombus luhuanus, a marine gastropod, retinal illumination can evoke an electroretinogram (ERG) exhibiting cornea-negativity and concomitant ‘on’ action potentials (APs) in afferent fibres in the optic nerve (Gillary, 1974, 1977). Both types of response are apparently linked to light-evoked depolarizations (LEDs) in retinal neurones, at least one type of which is a photoreceptor (Quandt & Gillary, 1979; Gillary & Gillary, 1979). The LEDs and ERG can exhibit two peaks which apparently arise from two separable light-evoked ionic conductance changes (Quandt & Gillary, 1980; Chinn & Gillary, 1980; Chinn, 1981). The cessation of illumination can evoke repetitive oscillations in the ERG, apparently due to synchronous potentials in a relatively small number of neurones hyperpolarized during retinal illumination, and rhythmic bursts of ‘off’ APs in their axons, which are considerably larger in diameter than the ‘on’ fibres (Gillary, 1974, 1977; Quandt & Gillary, 1979; Gillary & Gillary, 1979).
Following amputation of the eye of Strombus, a new one regenerates in its place, progressively assuming the characteristics of the original eye (Gillary, 1972; Gillary & Gillary, 1980). This paper surveys the maturation of the physiological responses of the regenerate, particularly the light-evoked activity of its retina. The retinal morphology of the regenerate, including its ultrastructure, will be considered more extensively elsewhere.
METHODS
In general, previously described methods used to study the electrophysiological activity of the mature eye of Strombus luhuanus (Gillary, 1974, 1977; Quandt & Gillary, 1979) were applied to several types of preparation derived from animals having eyes at different stages of regeneration. These were obtained by first amputating the distal end (including the eye) of one or both eyestalks from mature animals anaesthetized by immersion in 360mm-MgCl2; the animals were then returned to marine aquaria (approximately 23 °C or 28 °C) in which they recovered fully, behaved normally (except for certain visual deficiencies), and invariably regenerated new eyes. Physiological preparations were obtained by dissecting in sea water or a 1: 1 mixture of sea water and 360mm-MgCl2, the eyestalk or a distal portion of it, isolated from the re-anaesthetized animal.
In the simplest type of preparation, the eye remained intact, either covered with the original eyestalk epithelium, or dissected free of it (thus exposing the capsule). With these, the ERG could be recorded from the smallest stages of regenerating eye by applying suction electrodes of appropriate size (Figs 2–7). In other preparations, the retina was exposed in order to permit intracellular impalement of its neurones with microelectrodes. These included (1) preparations in which the retina and its attached capsule were peeled away from the vitreous body, permitting a microelectrode to be advanced to the vitreal surface of the retina (e.g. Figs 9, 10; for technical details, see Quandt & Gillary, 1979), and (2) preparations in which, after immersion (approximately 5 min) of the isolated eye in sea water containing pronase ( 1 mg ml−1 ), the capsule was peeled away from the retina, which remained attached to the vitreous body, held in position by suction applied to a fixed glass capillary (Fig. 8). Unfortunately neither technique was readily applicable to regenerates much smaller than 0·5 mm in diameter. Another type of preparation included the intact regenerated eye attached to the distal portion of the original optic nerve trunk, to which suction electrodes of appropriate size could be applied, either for electrophysiological recording or electrical stimulation (Fig. 11; see Gillary, 1974, 1977 for technical details).
During an experiment the preparation was immersed in sea water which remained constant within ±0·2 °C and, unless otherwise specified, was 20–24°C. Except for spectral sensitivity measurements, photic stimuli consisted of diffuse retinal illumination with white light (from a tungsten source), the maximum intensity of which was 400 J m−2s−1, although lower intensities were more usual. The general methods used for photic stimulation, electrophysiological recording and stimulation via suction electrodes and intracellular microelectrodes, and signal display were as described previously (Gillary, 1974, 1977; Quandt & Gillary, 1979). Microelectrodes were usually filled with 2 M potassium acetate (rather than 3 M-KCI), generally had resistances of 80–200 MΩ (although bevelled electrodes <50 MΩ could impale cells), and could be positioned with the aid of a hydraulic microdrive apparatus (David Kopf Insts., Tujunga, Calif.). In some experiments (e.g. Figs 9, 10) signals were recorded on magnetic tape (Model B Recorder, A. R. Vetter Co., Rebersburg, Pa.). The morphology of the regenerate was determined by examining dissected fresh specimens as well as sectioned, epon-embedded material under compound light and electron microscopes; these latter procedures are described elsewhere (Gillary, 1974; Gillary & Gillary, 1979). Other methodological details are included in the text.
RESULTS
Morphology of the regenerating eye
Light and electron micrographs of specimens obtained from animals reared at 28 °C (ambient temperature in the Marshall Islands, the natural habitat of Strombus luhuanus) have revealed that within a day after amputation, an eye cup (approximately 0·2 mm diameter) begins to form by invagination of the epithelium just external to the cut end of the optic nerve (Gillary & Gillary, 1980). This invagination closes off by 3 days to form a structure like an embryonic eye which progressively increases in diameter (Fig. 1). The early regenerates are approximately spherical; however, after attaining a mean diameter of approximately 0·5 mm, the eye’s dimensions along the pupillary axis tend to increase more slowly than the lateral growth, which can cause the more mature eyes to be somewhat oblate (e.g. Fig. 1D).
At 28 °C the rate of increase in mean diameter of the regenerate was approximately 50 μmday − 1; the approximate times required to obtain mean diameters of 0·2, 0·3. 0·4, 0·5, 0·6, 0·7 and 0·8 mm were, respectively, 2, 4, 6, 8, 10, 12 and 15 days. As the regenerating eye increases in size, its cells continue to increase in number and differentiate to form the various retinal layers, i.e. the rhabdome composed of photoreceptor distal segments, the pigment layer, the nuclear layer and the neuropile just internal to the eye’s capsule (Fig. 1; Gillary & Gillary, 1980; cf. Gillary & Gillary, 1979). When the regenerate is 0·8 mm in diameter, its general retinal ultrastructure appears quite mature. It continues to increase in size and can eventually, after many months, attain the diameter of fully grown mature eyes (approximately 2 mm). In fact such full grown regenerates are indistinguishable in appearance from normal mature eyes, except perhaps for somewhat less pigmentation of the newer distal eyestalk epithelium than that proximal to the original level of amputation.
Most of the regenerates used in the electrophysiological experiments were produced by animals reared at ambient temperatures of 23 ± 3 °C, at which the mean growth rate of the regenerates was significantly lower than at 28°C; to attain mean diameters of 0·3, 0·5, 0·7 and 0·8mm, respectively, approximately 10, 15, 22 and 30 days were required. Because of this dependence of growth rate on temperature, and because the extent of morphological and electrophysiological differentiation of the retina seemed to correlate more closely with the size of the regenerate than with its age, it was generally considered more reliable in comparing responses from different preparations to specify the regenerated eye in terms of its mean diameter.
The size, age, sex and nutritional state of the animal, or whether the left or right stalk was amputated had no noticeable influence on the results presented here. The level of amputation, usually within 3 mm of the distal tip of the eyestalk (approximately 1 cm long) also did not appear to influence the rate of eye regeneration. Amputation practically always resulted in complete regeneration of the portion amputated, which, for more proximal levels of amputation, could include a tentacle protruding medially. However, in two instances (out of well over a thousand amputations), two extensions of eyestalk were regenerated from the end of a single stump of original eyestalk, each of which bore its own distal regenerated eye.
ERG
Attempts were made to record ERGs via suction electrodes applied to more than 70 regenerates that ranged from the earliest stages to those that were fully mature. Light-evoked potentials were identified as ERGs on the basis of their waveforms and their dependence on light intensity, light adaptation, wavelength and ambient temperature (see below). Most recordings were made via an electrode applied to the corneal epithelium of the eye (Figs 2, 3, 4A,B, 6), particularly for the smaller stages, which were most resistant to isolation by dissection. Recording from the larger stages of eye, which were more amenable to isolation, also included the application of electrodes to the lens at the pupil (Figs 4C, 5A, 7, 8A), to the capsule at the back of the eye (Fig. 4D), or to the retina from which the capsule had been removed (Fig. 8B). Results obtained with all of the types of electrode placement yielded a consistent picture of ERG ontogeny as described below.
The earliest stages from which a definitive ERG could be recorded were between 0·2 and 0·3 mm in diameter, and had grown for approximately 6 days at about 23 °C. In response to brief flashes of light, these early responses exhibited, after a brief latency, a comea-negative potential, with only a single ‘on’ peak, followed by a return to the dark value. For slightly later stages (<0·3 mm diameter, approximately 9 days old at 23 °C), in addition to a single ‘on’ peak evoked by brief stimuli (Figs 2, 3A – D), prolonged illumination evoked a smaller, sustained, ‘steady state’ negativity which persisted until illumination ceased (Fig. 3E – G). The above responses were considerably simpler than the ERGs of mature eyes in exhibiting no sign of a second ‘on’ component or any ‘off’ activity. In ERGs of somewhat later stages (0·4 – 0·5mm diameter), the ‘steady state’ negativity was more pronounced (Fig. 4). In addition, some indication of an active ‘off’ response could occasionally be detected (Fig. 4D) as well as deviations in waveform suggesting the contribution of an emerging second, slower phase of cornea-negativity. This slower phase could be considerably more pronounced in the ERGs of 0·6 mm diameter eyes (Fig. 5B). Signs of a rhythmic ERG ‘off’ component were not usually encountered until stages 0·6 – 0·7 mm in diameter. The ERGs of eyes 0·8 mm in diameter were qualitatively very similar to that of the mature eye in that they could exhibit two well-defined cornea-negative ‘on’ components (Fig. 6) and repetitive ‘off’ potentials (Fig. 7), although these components were generally less pronounced than for mature ERGs (cf. Gillary, 1974). ERGs from regenerates more than 1 mm in diameter were indistinguishable from those of mature, larger eyes.
The ERG amplitudes from all stages of regenerated eye varied with light intensity in a manner similar to ERGs from mature eyes; as intensity was increased above a threshold, the amplitude increased approximately linearly over several log units of intensity before attaining a maximum value (Figs 2, 6A – D, G – J). Background illumination or a decrease in interstimulus interval decreased the ERG amplitude, indicating light adaptation, as seen for mature responses (Figs 3C,D, 4A,B, 5, 6E,F,K,L, 8A). As in the mature eye, the ERGs of regenerates were temperature dependent, exhibiting longer latencies and durations at lower temperatures (Figs 2, 3, 6). The spectral sensitivities of regenerates, including 0·3 mm stages, were not significantly different from that of mature ERGs, which are maximal at 485 nm and approximated by the Dartnall nomogram for a single visual pigment (Gillary, 1974).
The maximum ERG amplitudes recordable from the earliest stages of regenerate were smaller than for later stages, which could reach values attained by mature eyes (more than 10 mV). However, such variations in the amplitude of this extracellular potential need not have direct implications regarding the amplitudes of the underlying cellular events (see Discussion).
Experimental animals were usually exposed to normal cycles of daylight. However, after eye amputation, several animals were reared in total darkness, to assess the effects of light deprivation on regeneration. These preliminary experiments indicated that regeneration can occur in the dark and there were no dramatic differences, with regard to waveform and spectral sensitivity, between the ERGs of such regenerates and those exposed to daylight.
Intracellular potentials
Intracellular potentials were recorded from two types of preparation made from regenerated eyes: (1) those in which the microelectrode approached the retina from its vitreal side (three preparations; minimum diameter approximately 0·65 mm), and (2) decapsulated preparations in which the electrode approached from the capsular side (six preparations; minimum diameter approximately 0·55 mm). The general experimental procedure was (i) to advance a microelectrode through the retina and impale a cell, as indicated by appropriate increases in resting potential (RP), input resistance and membrane charging time (see Quandt & Gillary, 1979), (ii) to estimate the depth of impalement within the retina (with the microdrive apparatus) and (iii) to examine light-evoked potentials, including their dependence on light intensity and light adaptation. Such data were obtained from about 40 cells yielding stable potentials, in a variety of regions of the retina from a range of sizes of regenerating eye. Impalements were estimated to have been in the nuclear layer, usually nearer the neuropile than rhabdome, although this was not ascertained morphologically. In addition, for the decapsulated preparations, the ERG was usually recorded simultaneously with the microelectrode potentials, via a suction electrode applied to the vitreous body; this electrode also served to hold the preparation in a fixed position.
The earliest stages of regenerated eye (five preparations <0·7 mm diameter) yielded results similar in many ways to those from later stages and mature retinas (cf. Quandt & Gillary, 1979). These included (1) stable RPs as large as – 77mV, (2) LEDs as large as 40 mV, with waveforms dependent upon stimulus intensity and light adaptation, (3) cells that hyperpolarized in response to light (N = 2, in eyes 0·55 mm diameter), (4) cells that had action potentials (APs) accompanying the LEDs (TV = 3), and (5) repetitive APs of variable amplitude, usually from cells of decreasing RP, apparently injured by the electrode. In addition, membrane resistances and charging times for cells in regenerates were not remarkably different from those of mature eyes. Furthermore, no physiological effects could be obviously attributed to differences in the two types of retinal preparation used (i.e. ‘vitreal’ and ‘decapsulated’); both yielded similar data (e.g. RPs and LEDs) from comparable stages of regenerating eye. In addition, such data from well impaled cells (more than 14) in four control preparations of decapsulated mature eyes were typical of those obtained from hundreds of vitreal preparations used in other studies (Quandt & Gillary, 1979; Chinn & Gillary, 1980).
The most notable feature of the LED waveforms of the earlier stages was their failure to exhibit a second slow phase of depolarization seen in later stages and mature retinas. In conformity with the ERG data presented previously (Figs 2 – 4), the LED waveforms of stages less than 0·6 mm in diameter were relatively simple and monophasic (Fig. 8). Stages greater than 0·6 mm in diameter could exhibit a deviation from this simple waveform. For example, the differences between the effects on LED waveform of varying stimulus intensity (Fig. 9A–D) and prestimulus dark interval (Fig. 9E – H) suggest an emerging second phase of depolarization that is relatively more susceptible to light adaptation. In regenerates greater than 1 mm in diameter, the LEDs could exhibit a pronounced second phase and were indistinguishable from those of mature eyes (for records, see Quandt & Gillary, 1979, 1980).
Cells that exhibited LEDs accompanied by light-evoked APs (Fig. 10A,B) were encountered relatively infrequently in earlier regenerates (three cells, in stages 0·55, 0·65 and 0·84 mm diameter, respectively, out of approximately 35 cells in stages less than 0·9mm diameter). Nevertheless, the frequency of encounter was noticeably greater than in mature retinas, in these studies (none encountered in mature and regenerated retinas greater than 0·9 mm in diameter) as well as in other studies on the mature retina (three out of more than 300 impaled cells, by Quandt & Gillary, 1979, and none out of hundreds more, by Chinn & Gillary, 1980). Furthermore, in such cells in the present studies, depolarizing pulses of current passed via the recording electrode could evoke regenerative APs in the dark (Fig. 10C – H); no such regenerative activity in the mature retina of Strombus has been previously reported.
Other unquantified impressions were conveyed by the above experiments. For example, APs apparently evoked by injury inflicted by the advancing microelectrode seemed more prevalent and more diffusely recordable throughout the retinas of earlier regenerates than in more mature retinas, where such APs tended to be localized in the layer of neuropile. In addition, it seemed more difficult to separate non-spiking cells depolarized by light into two relatively distinct categories on the basis of RP, LED waveform and other properties than in mature eyes (Quandt & Gillary, 1979). Most cells in the regenerates tended to be more similar to cell ‘type II’ (e.g. Figs 8, 9) than to cell ‘type I’, encountered far less frequently in mature eyes (Quandt & Gillary, 1979). Finally, the maximum recordable RPs and average maximum LED amplitudes seemed slightly smaller in the earlier regenerates than in more mature retinas. However, these observations require further corroboration.
Optic nerve activity
Previous studies examined the electrical activity in the optic nerve of mature eyes in response to stimulation of the eye with light as well as to electrical stimulation of the optic nerve (Gillary, 1974, 1977). Similar studies were conducted on preparations (approximately 40) that had different stages of regenerating eye (ranging from the earliest, barely visible, to mature regenerates, approximately 2 mm in diameter) forming at the end of the original optic nerve trunk. The above methods were not applied to new portions of optic nerve in the distal regenerate, which were difficult to isolate by dissection.
Preparations with the earliest stages of regenerating eye (less than 0·4 mm diameter) failed to yield light-evoked activity in the optic nerve. Later stages (greater than 0·4 mm diameter) exhibited only ‘on’ activity (Fig. 11 A) during the corneanegative phase of the ERG recorded simultaneously. As in preparations with mature eyes, the light-evoked ‘on’ impulses recorded from the optic nerve were usually very small in amplitude and largely obscured by baseline noise in visual displays; however, they were clearly detectable by using an audio monitor.
The earliest stage to exhibit unequivocal ‘off’ activity was approximately 0·7 mm in diameter; this activity was usually an abrupt increase in the frequency of impulses of relatively low amplitude (Fig. 11B), most reliably detectable with an audio monitor.
A burst of impulses comparable in amplitude to those of the ‘off’ response also occurred at the onset of stimulation, as can occur for preparations with mature eyes (Gillary, 1974). Slightly larger stages of regenerate (approximately 0·8 mm diameter) could exhibit discrete bursts of ‘off’ impulses and concomitant ‘off’ oscillations in the ERG, although the impulse amplitude and number of impulses per burst were still generally lower than for preparations with mature eyes. The light-evoked activity recorded from the optic nerve of preparations with mature regenerates (>l mm diameter) was indistinguishable from preparations with normal mature eyes.
In addition to the above light-evoked activity, optic nerves in preparations with regenerating eyes could occasionally exhibit impulses not clearly correlated with a photic stimulus. Such activity, never seen in preparations with mature eyes, appeared to be mediated by non-visual fibres such as mechanoreceptors (see below).
In preparations with all stages of regenerated eye, it was possible to evoke compound action potentials (CAPs) by electrical stimulation of the optic nerve trunk (Fig. 11C) that were similar to those evoked in optic nerves with mature eyes in exhibiting a major slow component (reflecting impulses in a large population of small diameter fibres) and a smaller, faster component (reflecting impulses in a markedly smaller number of fibres of larger diameter). This suggests that the original optic nerve fibres remained viable during regeneration of a new eye. Contributions to the CAPs from possibly newly regenerated fibres were not detected; however they could well have been obscured by the responses evoked in the original fibres.
Invasion of the mature eye by the slow component of the CAP can evoke, after a delay of about 10 ms, ‘answering bursts’ of impulses in ‘off’ fibres that propagate away from the eye and can be inhibited by retinal illumination (Gillary, 1977). Preparations with ‘mature’ regenerated eyes (>1 mm diameter) showed identical activity. So did those with somewhat earlier regenerates (0·7 mm diameter; Fig. 11C,D), except that, as described for light-evoked ‘off’ bursts, the bursts were usually less distinct than those from preparations with mature eyes. CAPs evoked in preparations with earlier stages of regenerate were often followed by bursts of impulses after a similar latency; however, unlike those of preparations with more mature eyes, these impulses generally required a higher intensity of electrical stimulation, were not readily inhibited by light, and often accompanied visible movement within the eyestalk.
Movement
Illumination of regenerated eyes, ranging from the largest to those below 0·5 mm in diameter, could cause visible movement in the retina. It was most clearly visible in isolated, enucleated eyes and exhibited a time course approximated by the corneanegativity of the ERG, similar to that reported for mature eyes (Gillary, 1977). It is apparently mediated by muscle fibres just external to the capsule (see Gillary & Gillary, 1980). This movement could be eliminated by dissecting away those fibres, or more easily, by immersing the eye (for about 2 min) in sea water containing pronase (1 mg ml − 1); these procedures were needed to facilitate recording from cells impaled with microelectrodes. In addition to light-evoked movement, mature eyestalks isolated in sea water but not dissected further could exhibit intermittent ‘spontaneous’ writhing movements, although usually the eyestalk remained motionless in a contracted state. These movements also occurred in newly regenerated eyestalks (with eyes as small as 0·3mm diameter), but more frequently and with greater rapidity and vigour. As mentioned previously, CAPs evoked in optic nerves of preparations with regenerated eyes by relatively intense electrical stimuli were occasionally accompanied by slight visible twitches in the eyestalk near the eye, possibly mediated by motor fibres, antidromic impulses in afferent neurones, or the spread of stimulating current directly to the muscle fibres. Such movement did not accompany CAPs in preparations with mature eyes.
DISCUSSION
ERG
The results of experiments in which ERGs were recorded from eyes at different stages of regeneration indicate that as the regenerate grows and matures morphologically, its ERG waveform progressively matures from the earliest recordable response, a relatively simple cornea-negative ‘on’ potential, to the more complex response characteristic of normal adult eyes, which can exhibit two cornea-negative ‘on’ peaks, a maintained ‘steady-state’ cornea-negativity during sustained retinal illumination, and repetitive ‘off’ potentials when illumination ceases. The ERG can be used in Strombus to make inferences regarding the cellular bases of the underlying activity; previously published data indicate that the cornea-negative components of the ERG reflect LEDs that can exhibit two depolarizing ‘on’ phases, and the repetitive ‘off’ potentials correlate with neurones that hyperpolarize during retinal illumination. The ERG is useful because it reflects simultaneous neuronal activity throughout the retina and was recordable from a wide range of stages of regenerating eye, including the smallest, in which neurones were not impaled.
Inferences based on ERG data regarding events at the cellular level are subject to limitations because the retina of Strombus can act as an electrically inhomogeneous volume conductor. Extracellular microelectrodes advanced through it appeared to encounter regions of relatively high electrical resistance at the eye’s external capsule as well as in the pigmented apical region between the layer of rhabdome and the nuclear layer (H. Gillary, unpublished observations), where glia wrap around neurones and form junctional complexes with them (Gillary & Gillary, 1979). The amplitude of an ERG component depends on the electrical proximity of the underlying cellular current generators to the recording electrode, which might be expected to vary with the stage of regenerate, since the early stages are not simply scaled down versions of more mature stages (Fig. 1). Such a factor could have reduced the maximum ERG amplitudes recordable from early regenerates more than in later stages, and could have postponed the earliest stages at which certain ERG components (e.g. the ‘steady state’ cornea-negativity or repetitive ‘off’ potentials) were first detected. An attempt was made to minimize misinterpretation by comparing ERGs recorded with electrodes of different sizes, and placed at various positions with respect to the retina.
Intracellular retinal potentials
Intracellular recording can provide more precise information than ERGs on the absolute amplitude and waveform of potentials from different retinal cell types, although care must be taken to avoid sampling bias introduced by such factors as electrode properties, impalement techniques, dissection procedures and retinal morphology (including cell size), the latter two of which could have varied systematically with the stage of regenerate. In conformity with the ERG data, intracellular recordings indicate that the LEDs of early regenerates exhibit only a single phase and that a second does not emerge until later stages. Since the two LED components appear to arise in mature retinas from separable light-evoked conductance increases mediated by two populations of membrane sodium channels (Quandt & Gillary, 1980; Chinn & Gillary, 1980), the ontogeny of the LED during regeneration could reflect shifts in the populations of such ion conducting channels, such as occur in other developing neurones (Spitzer, 1979). Cells in the regenerate that exhibited LEDs and no APs, the type impaled most frequently, were more similar to ‘type II’ cells in the mature retina (also most frequently encountered) than ‘type I’ cells (Quandt & Gillary, 1979). Failure to distinguish two such categories of response in earlier regenerates suggests that the cells may have been functionally less differentiated.
Some cells depolarized by light, exhibited light-evoked or electrically evoked intracellular APs. They were encountered more frequently in stages less than 1 mm in diameter. This, and the more widespread occurrence in regenerates of APs evoked by injury, apparently reflect maturational changes during regeneration. These could include changes in the distribution of voltage-sensitive ionic channels (Spitzer, 1979), or morphological changes that might affect the spread of APs to the soma or the ease of impalement of other neuronal regions from which APs are recordable. The tendency for processes in the neuropile of earlier stages of regenerate to be larger in diameter than those of more mature retinas (Gillary & Gillary, 1980) is compatible with such hypotheses. Jacklet & Rolerson (1982) found that LEDs in Aplysia photoreceptors, most often soma-impaled, usually exhibited no APs; however, infrequently, an axonal site of impalement could yield LEDs accompanied by APs.
Intracellular recording also revealed cells (in stages <0·6 mm diameter) that hyperpolarized in response to light. Such cells encountered in the mature retina apparently underlie ‘off’ responses exhibited by the ERG and optic nerve activity (cf. Gillary, 1974, 1977; Quandt & Gillary, 1979).
Correlation of responses with retinal morphology
The general picture provided by the electrophysiological results is compatible with the morphological development of the regenerate (Fig. 1, and Gillary & Gillary, 1980). The eye cups of the earliest stages ( <0·3 mm diameter) consist of relatively undifferentiated cells bearing short apical microvilli (<10 μm long); distinct layers of rhabdome, neuropile and opaque pigmented granules are not yet present. As might be expected, such stages did not yield light-evoked electrical activity. In somewhat later stages (0·3 – 0·4mm diameter), the retina exhibits greater morphological differentiation, including distinct layers of neuropile and opaque pigment granules (approximately 0·5 μm in diameter), glia and cells that exhibit densely packed, clear cytoplasmic vesicles (approximately 50 nm in diameter) and numerous microvilli (approximately 20 μm long) that extend towards the cornea directly from their apical ends to form a layer of rhabdome. These cells are apparently developing photoreceptors; such vesicles and apical microvilli are characteristic of photoreceptors in other gastropods (for references see Gillary & Gillary, 1979; Messenger, 1981). Stages such as the above yielded only simple single-peaked ERGs, as do the mature eyes of stylommatophoran gastropods (Gillary & Wolbarsht, 1967; Gillary, 1970), with similar rhabdome morphology (cf. Eakin & Brandenburger, 1967). Furthermore, in developing insect eyes, the stage at which the ERG is first recordable correlates with the formation of photoreceptor microvilli; as in the regenerating eye of Strombus, the ERG initially appears to be a relatively simple cornea-negative receptor potential and progressively increases in complexity as the eye matures (White, Brown, Hurley & Bennett, 1983). Since the rhabdome presumably contains visual pigment that mediates light-evoked photoreceptor responses, one might expect its morphology (including its thickness and ultrastructure), to be correlated with the maturation of the ERG, as well as with other light-evoked activity.
In contrast to the retinas of early regenerates, most retinal microvilli in stages 0·4 – 0·5 mm in diameter emanate from and swirl around apical extensions of photoreceptors, forming distal segments that extend towards the cornea across the entire layer of rhabdome, approximately 40 μm thick. This is the general configuration in the mature retina (Gillary & Gillary, 1979), although the rhabdome is thicker (approximately 100 μ m). The ERGs of such stages were somewhat more mature than those of earlier stages (e.g. they could exhibit a larger ‘steady state’ component and a primitive ‘off’ response).
In later stages (approximately 0·6mm in diameter), the rhabdome thickness (approximately 60 μm) and retinal ultrastructure are more like that of the mature eye, although not yet as differentiated with regard to the complexity of the neuropile and the appearance of cytologically distinct cell types. ERG and LED waveforms were still relatively simple in these stages, and cells yielding LEDs seemed less readily separable into different classes than in the mature eye, observations compatible with the apparently lower degree of morphological differentiation. It could not be ascertained that LED amplitudes were lower than in mature eyes, although one might also have expected this in view of the smaller thickness of rhabdome in these stages.
The rhythmic ‘off’ potentials of the ERG, seen for stages >0·6 mm in diameter, apparently arise from synchronous activity in cells hyperpolarized during retinal illumination (Gillary, 1974), and since ‘off’ activity appears to be synaptically mediated (Gillary, 1974, 1977), one might expect some correlation with the maturation of synaptic structures in the neuropile. The absence of these rhythmic potentials in earlier stages (cf. Figs 4D, 7) may also be due to less synchrony between the ‘off’ cells. Regenerates >l mm in diameter were indistinguishable, ultrastructurally or electrophysiologically, from the mature retina.
Optic nerve
The ontogeny of light-evoked optic nerve activity in preparations with regenerating eyes parallels the maturational sequence of the retina in which the earliest appearance of ERG cornea-negativity, presumably a reflection of LEDs, precedes that of the ‘off’ responses. The earliest stages for which ‘on’ and ‘off’ responses, respectively, were detectable in the optic nerve were somewhat later than those for which such retinal activity was first observed. This may have been due to the additional time needed for the growth of optic nerve fibres, which must extend from the retina, as well as to factors related to the recording conditions (e.g. signal strength and signal/noise), which could have postponed the detection of the optic nerve responses.
In evaluating the implications of the optic nerve responses with respect to fibre regeneration and reinnervation of the regenerating retina, it is useful to bear in mind certain morphological considerations. The optic nerve of the mature eye, which runs to the optic lobe of the brain more than 1 cm away, exhibits two populations of fibre; one, several tens of thousands of relatively small diameter (approximately 0·4 μm), includes the ‘on’ fibres, and another, several hundred in number and larger in diameter (>l μm), includes ‘off fibres (Gillary & Gillary, 1979). Lucifer yellow (Stewart, 1978), introduced into the cut end of the optic nerve (Mulloney, 1973), can pass along the large fibres and enter retinal somas (H. Gillary, unpublished observations), apparently those hyperpolarized during retinal illumination (Quandt & Gillary, 1979). Lucifer yellow similarly introduced can also enter certain somas in the brain (H. Gillary, unpublished observations). These may be efferent neurones, as are found in the optic nerves of other molluscs (Lange & Hartline, 1974; Luborsky-Moore & Jacklet, 1976), although they have not been demonstrated physiologically.
In the present studies, optic nerve responses were recorded from the original optic nerve trunk, rather than from the distal region that presumably contained newly forming fibres. Furthermore, new fibres running along the surface of the original optic nerve may have often been stripped away during dissection. The persistence of the CAP indicates that the original optic nerve fibres can remain viable for a considerable time during eye regeneration as can severed axons in other invertebrates (Muller & Carbonetto, 1979; Anderson, Edwards & Palka, 1980).
The relation between these original fibres and those that innervate the regenerate and exhibit light-evoked activity is uncertain. It seems probable that new axons sprout from the somas within the regenerating retina and grow towards the central nervous system, as in other regenerating and developing invertebrate sense organs (Edwards & Palka, 1976; Bate, 1978). These may invade the original optic nerve trunk and eventually replace the severed axons, as in other systems (Muller & Carbonetto, 1979). The present studies indicate that during regeneration, fibres other than those types normally serving the mature eye may associate with the original optic nerve, including mechanoreceptors or motor fibres that innervate nonvisual portions of the regenerating eyestalk. Examining the ultrastructure and electrophysiological activity at sequential points along the optic nerve of preparations with different stages of regenerating eye should help clarify the fate of the original fibres following eye amputation, and the extent and direction of new fibre growth.
Behaviour
Under appropriate conditions, Strombus will gradually extend and orientate its eyestalks toward a visual stimulus such as a light source, or rapidly retract them in response to a sudden decrease in light intensity (e.g. a shadow) or, less readily, to an abrupt intensity increase (H. & E. Gillary, unpublished observations). These responses, which may aid the animal, an herbivore, in avoiding predators, appear to be mediated, respectively, by light-evoked ‘on’ activity and by ‘off’ fibre activity (Gillary, 1974). Amputation of the eyes abolishes these responses, which return to normal as new eyes are regenerated. Many details regarding the morphology, physiology and behavioural role of the regenerating eye await further investigation.
ACKNOWLEDGEMENTS
The author is grateful to E. W. Gillary for preparing the micrographs, to F. N. Quandt and K. S. Chinn for technical assistance, and to I. M. Cooke for comments on the manuscript. This study was supported by research grants from the U.S. National Science Foundation (GB32091) and the National Institutes of Health (EY01531).