- Email: [email protected]
Recovery of absolute threshold with UVA-induced retinal damage
Physiology &Behavior,Vol. 32, pp. 94%954. Copyright ©Pergamon Press Ltd., 1984. Printed in the U.S.A.
0031-9384/84 $3.00 + .00
Recovery of Absolute Threshold With UVA-Induced Retinal Damage WENDON
W. HENTON
AND STEPHEN
M. S Y K E S
U.S. Department o f Health and Human Services, Public Health Service, Food and Drug Administration National Center for Devices and Radiological Health, 12709 Twinbrook Parkway, Rockville, MD 20857 R e c e i v e d 20 S e p t e m b e r 1983 HENTON, W. W. AND S. M. SYKES. Recover3, of absolute thresholds with UVA-inducedretinal damage. PHYSIOL BEHAV 32(6)949--954, 1984.--A within-trial psychophysical procedure tracked the initial loss and subsequent recovery of visual thresholds in albino rats exposed to ultraviolet light at 350 nanometers and 0.4 milliwatts per square centimeter. Absolute thresholds increased up to 5 log units immediately following the 15 hour ultraviolet exposure, with a daily recovery of 1-2 log to asymptotic thresholds over a 7-day post-exposure period. The corresponding retinal damage on Day 1 included extensive vesiculation of the photoreceptor outer segments, vacuolation of the inner segments, and pyknosis of cell nuclei. The total number of photoreceptor nuclei and outer segments was unchanged relative to control eyes through post-exposure Day 3. Both nuclei and outer segment counts then consistently decreased 15-20 percent between Days 3-7. The two-stage loss of photoreceptors but daily recovery of absolute thresholds again suggests a significant dissociation of retinal structure and psychophysical function in light-induced ocular pathology. Absolute threshold
Histology
Pathology
Rats
(450× 10 6 W/cm 2 at 12 hr per day [27]). The first evidence of pathology is photoreceptor lesions in the outer segments with 10-16 weeks of chronic UVA, eventually followed by a complete loss of the entire photoreceptor cell layer by 70 weeks [27]. The available data then unanimously describe a disorganization of photoreceptor inner and outer segments, plus possible involvement of other retinal components from cell bodies to pigment epithelium, consequent to UVA radiation in aphakic and nocturnal subjects. Visual function and psychophysical ihresholds remain to be examined for this special circumstance of UVA-induced retinal pathology. In larger perspective, however, visible light-induced ocular pathology has been analyzed in several psychophysical studies. On one hand, behavioral psychophysical function can be virtually unchanged in spite of extensive structural damage to the retina and photoreceptors [1,2]. This apparent independence of behavioral stimulus control and rod photoreceptor pathology has been subsequently replicated by numerous investigators [4, 5, 13, 14]. Yet additional experiments find precisely the opposite result, with intensity functions and electrophysiological measures seemingly graded to the magnitude of photoreceptor damage [6, 11, 23]. Absolute thresholds for example progressively increase with high intensity light exposure, apparently associated with an increased vesiculation of outer segments and vacuolation of inner segments [11]. The data with visible wavelengths collectively suggest that light-induced retinal damage may have differential visual consequences, with threshold intensity functions but not suprathreshold pattern discrimination covarying with progressive photoreceptor pathology. The present study was then an initial comparison of the functional and histological consequences of UVA exposure
E X P E R I M E N T A L studies of the eye reveal a differential absorption of ultraviolet and visible radiation throughout the ocular media [7,21]. The differential absorption of high intensity radiation produces structural alterations along the optic axis as a function of the wavelength, intensity, and duration of surgical or accidental exposure. Ultraviolet radiation (UV), for example, primarily affects the cornea and crystaline lens at wavelengths between 280-400 nanometers (nm) [17,18]. In contrast, retinal lesions in the pigment epithelium and photoreceptors occur with the still longer and visible wavelengths of 400-700 nm [13, 15, 16, 22, 24]. Recent experimental effort, however, has identified a special case of U V A (315-400 nm) radiation damage to the retina rather than the cornea or lens in subjects with high ocular transmission of UV, such as aphakic (lensless) monkeys and nocturnal rodents [9, 19, 27]. The UVA-induced damage in the monkey retina includes ultrastructural changes in the photoreceptor inner and outer segments plus pyknotic cell nuceli within one hr of exposure to a 325 nm laser (3.8 J/cm 2 at 4.0 sec, [19]). Similar pathological changes were found at 24 hr following still lower laser exposures (0.15 J/cm 2 at 0.8 sec), with the most striking changes in the photoreceptor inner and outer segments. An action spectrum for retinal lesions throughout the UVA wavelengths has now been reported for aphakic monkeys [9]. The lesion thresholds increased as a direct function of wavelength, with thresholds approximately six times lower in the UVA at 325 and 350 nm relative to visible blue at 441 nm. Again, structural disorganization occurred within the outer segments, inner segments and cell bodies of the photoreceptors, with additional pathology including the pigment epithelium. Retinal damage also occurs in nocturnal rodents chronically exposed to low level UVA fluorescent lamps
949
950
H E N T O N AND SYKES
in nocturnal rats. More specifically, the initial loss and subsequent recovery of absolute thresholds were obtained for the left eye during daily sessions following a single 15 hr exposure to high intensity UVA. Enucleation of first the right eye and then the left eye provided a within-subject comparison of behavioral thresholds [10] and electron microscopic retinal lesions [22] at either 1, 3, or 7 days postexposure. METHOD
Animals Five male Sprague-Dawley albino rats, approximately 180 days of age, were trained in the psychophysical procedure. Four additional rats of the same age and sex were used as unexposed cage controls for comparative histology. Subjects were raised from birth under approximately 10 lux illumination (L:D 12:12). Each subject was maintained at 95 percent of the previous ad lib home cage weight, and dark adapted 60 min prior to each experimental session.
Apparatus The conditioning apparatus is described in detail in two previous publications [10,11]. Essentially, a neutral tint stimulus key, stainless steel drink tube, and food cup were mounted respectively at the left wall, apex, and right wall of an enclosed triangular area (8.8x8.8× 11.5 cm) behind the front panel of a standard rodent test chamber. A 3.8 cm diameter opening in the front panel limited manipulandum access to only the subject's head. For the present study, a 2800 Hz tone generator was placed in the ceiling, 7.5 cm directly above the subject's head, as a time out stimulus. A 10 cm speaker was placed immediately behind the stimulus key, as a "white" noise discriminative stimulus. The experiment was conducted in a ventillated, sound attenuating booth and controlled by a remote on-line computer. A 1.3 cm diameter white light was centered on the stimulus key by a light pipe (Edmonds Scientific BF 24) connected to an external wedge box. The wedge box included a 2.8 W tungsten bulb mounted behind a Metavac No. 1445 variable density wedge, plus a neutral density filter block to further attenuate light intensity. Stimulus intensity through the variable density wedge was controlled by a two rpm synchronous motor with a shaft mounted rotary switch. The irradiance at the surface of the key was measured by a calibration system with a total estimated error of -+25% [11]. The ultraviolet exposure apparatus for retinal damage consisted of a cylindrical hardware cloth cage (1.25 cm square mesh, 58x30 cm diameter) centered within a larger white enamel cylinder (82.5×57.5 cm diameter). Six blueblack fluorescent lamps (F20T12BLB) were mounted vertically at equal intervals around the circumference of the exterior cylinder. The spectral distribution was maximal at 350 nm (half-power bandwidth approximately 20 nm), with spiked lines at approximately 405 and 435 nm. The average irradiance in the horizontal plane at the center of the hardware cloth cage was 381×10 -6 W/cm 2, ranging from a minimum of 359 to a maximum of 403 × 10-6 W/cm z. Forced air through the chamber maintained the ambient temperature at 25_+2°C.
Proced, re Behavioral procedure. Initially, a continuous reinforcement (CRF) schedule reinforced nose poking responses on
the stimulus key in the presence of either a red or white visual stimulus (approximately 5×10 -s W/cmZ). After 3-5 sessions, onset of a 7.0 sec stimulus trial was contingent upon a licking response on the drink tube. Trials with red and white light were irregularly alternated on a quasirandom schedule. The lick contingency was gradually extended to a variable ratio 15 (VR, range 2-35 responses) over approximately 30 sessions. The final conditioning schedule was thus an observing response schedule of chain VR 15 sig FR 1, producing approximately 250-350 signalled reinforcement trials per 40 min session. A response on the stimulus key during the intertrial interval produced a 20 sec timeout (paired with the overhead 2800 Hz tone). The timeout was followed by a fixed requirement of 35 responses to the next signalled reinforcement trial. The conditioning schedule then reverted back to the standard chain VR 15 sig FR 1. After initial training, signalled reinforcement was altered to a within-trial threshold procedure with intertrial titration. First, "white" noise was substituted for the red light as the standard stimulus for signalled reinforcement, and trials with the white light were only scheduled once every 90 sec (range 35-180 sec). This modification preserved the dark adapted state of the subject by minimizing light exposure during the threshold conditioning sessions. For absolute threshold trials, the intensity of the white light was no longer constant but slowly increased approximately 0.14 log per sec throughout the 7.0 sec trial. Intensity was set at one of nine nominal wedge positions (rotary switch positions) for stimulus onset, and progressively increased until emission of a nose poking response on the stimulus key. A criterion stimulus response was defined by the emission of a two-response sequence--licking followed within 1.5 sec by a response on the lighted key. A trial was repeated if the stimulus response was not preceded by licking, or if the changeover latency was greater than 1.5 sec. The titration of light intensity from Trial N to Trial N + I was determined by the intensity controlling the criterion response sequence. The initial intensity at trial onset was increased to the next higher setting, remained the same, or decreased to the next lower setting for Trial N + I given a lick-stimulus response sequence at 0.0-1.5, 1.5-4.0, or 4.07.0 sec, respectively, during Trial N. The titrated change of initial intensity was approximately 0.2-0.4 log, depending on the specific position of the neutral density wedge. Two exceptions were a downward titration from the lowest wedge setting No. 1, or an upward titration from the highest wedge setting No. 9. The previous intensity was repeated in both cases, with appropriate adjustment of supplemental neutral density filters for subsequent sessions. In sum, the programmed intensity for trial onset would titrate to progressively lower intensities across the first 5-10 trials, and then either titrate down, remain the same, or titrate up for succeeding trials of each session. Given titration of intensity for trial onset, light intensity would then slowly increase approximately 1.0 log within each 7.0 sec trial. The result is an overlapping range of intensities scanned for threshold across successive trials. Absolute threshold was defined within each trial as the minimum intensity controlling the criterion response sequence, excluding the first ten trials as preliminary titration down to stable threshold. Subjects received a total of 25 conditioning sessions using the within-trial threshold procedure prior to the acute UVA exposure. Exposure procedure. The exposure design provided unexposed surgical controls plus right and left eyes enucleated
951
VISUAL DEFICIT AND RETINAL DAMAGE
:Z
TABLE 1 ENCULEATION SCHEDULE OF RIGHT AND LEb'T EYES FOR FIVE RATS ~o
Experimental Day Subject N umber
23
24
25
26
27
28
29
39
31
Baseline
Day 2
Day 1
Day 3
-7
32
| 1
2 3 4 5
R
R R R R
-10
L
L
-11
L L L
-12
..........
i 21
...........
i 25
,,,i,
. . . . . .
t
24
.............
i 28
Trials
on Day 1 (immediately following exposure), Day 3, and Day 7 post-exposure. Subjects 1 and 5 served as controls to determine the effect of enucleating the right eye upon subsequent absolute thresholds for the left eye. The right eyes were enucleated immediately after Session 23 during 30-60 sec surgery under ether anesthesia. Subjects were then returned to the home cage, with baseline thresholds resumed during the regular morning conditioning sessions for the next two days. The right eyes of Subjects 1 and 5 also served as unexposed, normal controls for histological analysis. Subjects 1-5 were U V A exposed overnight for 15 hr prior to Session 26. The enucleation schedule (Table 1) provided comparison of absolute thresholds and retinal damage for two control eyes, three eyes at Day 1, two eyes at Day 3, and three eyes at Day 7 post-exposure. The enucleation schedule also provided five within-subject comparisons of left eye threshold versus right and left eye retinal damage over the 7 day post-exposure period. The exposure routine then included (a) 15 hr U V A exposure from 6 p.m. to 9 a.m., (b) one hr dark adaptation, (c) 40 rain threshold session, (d) immediate enucleation and initiation of histological processing, if scheduled, (e) return to home cage with supplemental feeding, (f) resumption of normal daily schedule with dark adaptation at 9 a.m. Four additional, untrained rats provided unexposed control eyes for corroborative retinal histology. Histological procedure. Eyes were enucleated under ether anesthesia and immediately fixed in cold 2.5% glutaraidehyde and 6% sucrose buffered to pH 7.2 with 50 mM cacodylate. After fixation, the eyes were hemisected vertically through the optic nerve. The nasal hemisphere was embedded in glycol methacrylate, sectioned at 3 microns, and stained with toluidine blue. Samples from the temporal hemisphere were post-fixed in 1% osmium tetroxide, dehydrated by a series of acetone dilutions, and embedded in epoxy resin. Sections for electron microscopy were double stained with uranyl acetate and lead citrate, then photographed with a J E O L 100C electron microscope. In all cases, the retinas were sectioned transversely with the photoreceptots cut as nearly as possible along their long axis. The number of photoreceptor nuclei were determined from light micrographs of four regions of the temporal retina, each 1-2 mm from the optic disk (total magnification of 1000×). Photoreceptqr outer segment density was determined from electron micrographs of the same regions (magnification of 4500x). Outer segment density was determined by counting the number of outer segments intersecting a line 15 /zm from and parallel to the apical edge of the pigment epithelium.
FIG. 1. Comparison of absolute thresholds before and after UVA exposure for Subject 1. The filled data points in each panel represent a single threshold determination during the final baseline session, and three consecutive daily sessions following UVA exposure. A trial with no obtained threshold is indicated by an open square.
RESULTS For the enucleation control procedure, removal of the right eye did not affect the subsequent thresholds for the left eye. Median thresholds for the two sessions before and after right eye enucleation were 1.1 and 1.4 vs. 0.9 and 0.8× 10-12 W/cm ~ for Subject 1, compared to 2.4 and 2.3 vs. 2.2 and 2.0×10 - u W/cm 2, respectively, for Subject 5. In consequence, any loss of absolute threshold following U V A exposure would not be a simple artifact of the enucleation procedure. Figure 1 illustrates the daily change in thresholds following U V A exposure for Subject 1. Each data point is a single threshold determination during the final baseline session and post-exposure sessions on Days l, 2, and 3. Excluding the first 10 trials as preliminary titration, baseline thresholds ranged from 5.6 (Trial 15) to 10.0 (Trials 11 and 17), with a median of 7.7x 10-12 W/cm 2, for the final preexposure session (left panel Fig. 1). After 15 hr U V A and 1 hr dark adaptation (Day 1), programmed intensities on Trial 1 and Trial 2 failed to control the changeover to stimulus responses. Threshold was first recorded on Trial 3 only after intensity titration up to 1 . 6 x l 0 -6 W/cm 2. Subsequent thresholds ranged between 0.4 and 3.5 x 10-6 W/cm ~ over the succeeding trials of Day 1. Absolute thresholds thus stabilized approximately 5 log units above previous baselines with acute UVA exposure. Thresholds did not progressively deteriorate but systematically recovered over subsequent post-exposure sessions. Daily thresholds decreased towards baseline and recovered more than 2 log by Day 2, and 3 log by Day 3 for Subject 1. Figure 2 plots this recovery of function following acute UVA exposure for all subjects. Thresholds were tracked either 3 days (Subjects 1 and 2) or 7 days (Subjects 3-5) prior to enucleation of the remaining left eye and experiment termination. For all subjects, the largest threshold shift was on Day l immediately following 15 hr U V A exposure. Threshold increase ranged from more than 3 log (Subjects 3 and 4) to more than 5 log (Subjects 1, 2, and 5) relative to previous baselines. F o r Subject 5, threshold change exceeded the maximum system capability, with no behavioral thresholds found at even the highest available intensity of 4×10 -6 W/cm 2. Threshold gradually recovered by 1-2 log per postexposure day, with apparent asymptote at post-exposure
952
H E N T O N AND SYKES -5
$1
$3
T
30 7i
300
-6
E
E 2k
-7
O
-8
E 0
~a . 0
-9
_=
C u)
-5
E
-6
4-' U~
g, --I
e®
E ffl
-10
-12
20
0 0 0 0 e¢,1
-11 (/)
200
I B
L 1
I 2
I 3
I B
L 1
I 2
I 3
I 1
$4
I 2
I 3
I 4
I 5
I 6
t~ >4 x 10-6 \ \
I 7
S5
I c
I 1
I
i 3
I
I
_
I
J 7
0
Days After Exposure FIG. 3. Counts of photoreceptor nuclei (left axis) and outer segments (right axis) for control retinas (C) and retinas at 1,3 and 7 days following 15 hr exposure to UVA.
-7
-8 -9 -10
•
-11 -12
t B
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
1
2
3
4
5
6
7
B
1
2
3
4
5
6
7
Days Post Exposure FIG. 2. Absolute thresholds before and after UVA exposure for each of five rats. Each data point is the median threshold for the last baseline session (B) and successive post-exposure sessions (I-7). Subjects 1 and 2 and Subjects 3-5 were trained for three and seven post-exposure days, respectively, until enucleation of the remaining left eye. For Subject 5, the threshold loss on post-exposure Day 1 exceeded the maximum intensity of the optical system at 4× 10-6 W/cm2
Day 3 (Subject 3) or Day 5 (Subjects 4 and 5). Absolute thresholds virtually recovered by the final post-exposure sessions, stabilizing approximately 0.5 log above preexposure baselines for Subjects 3-5. Figure 3 presents separate counts of photoreceptor nuclei and outer segments for both control and exposed eyes. The density of photoreceptors was unchanged relative to control eyes at postexposure Days 1-3. However, cell nuclei then decreased approximately 15% by Day 7. Photoreceptor outer segments also remained relatively unchanged throughout postexposure Day 3, then decreased approximately 20% by Day 7. Further, continued pyknosis of cell nuclei and vesicuolation of outer segments suggests that the photoreceptor degeneration had not yet stabilized by post-exposure Day 7. For all subjects, photoreceptor pathology was variable across full cross sections of the retina, with the superior retina routinely more damaged than the inferior retina at each post-exposure time interval. In contrast, the pigment epithelium and inner retinal layers were relatively unchanged across the 7 post-exposure days for all subjects. Figure 4 compares the ultrastructural characteristics of control and exposed eyes at 1, 3, and 7 days post-exposure. The upper and lower rows of electron micrographs represent the minimal and maximal pathology observed at each time interval. For the control eyes, all retinal layers appeared relatively normal (Fig. 4A). However, the distal ends of o c -
casional outer segments were bulbous and somewhat separated from the pigment epithelium (Fig. 4B). Scattered pyknotic nuclei and vacuoles in the pigment epithelium were also occasionally observed in the control eyes. The rod photoreceptor morphology was clearly disrupted immediately following UV exposure and testing on Day 1 (Fig. 4C and 4D). Photoreceptors were routinely bent and misshapen, with irregular cell profiles and vesiculation of the distal portion of all outer segments. Inconsistent and nonspecific vacuole formations appeared in the inner segments. In contrast, prominent organelles such as the mitochondria appeared normal in the inner segments. Changes in the pigment epithelium and other retinal layers were not observed at Day 1. This range of photoreceptor damage was associated with an absolute threshold change of 3-5 log units. The reinal damage on post-exposure Day 3 was an intermixed pattern of partial recovery of outer segment vesiculation in some photoreceptors (Fig. 4E) combined with a continuing vesiculation and nuclei degeneration of neighboring photoreceptors (Fig. 4F). In particular, the outer segment vesiculation was at least equal to or perhaps increased in most photoreceptors relative to Day 1 (Fig. 4F compared to Fig. 4C and 4D). The inner segments continued to contain nonspecific vacuole formations. However, the distorted orientation of the individual photoreceptors would prevent any accurate association between the inner segment morphology and outer segment recovery. This differential pattern of photoreceptor degeneration/repair resulted in uneven damage across the retinal sections, with severely vesiculated cells directly adjacent to relative normal cells. The composite photoreceptor morphology was associated with a partial recovery of absolute thresholds to approximately 1.0-2.5 log above previous control baselines. On post-exposure Day 7, distinct recovery approximating normal morphology was apparent in various retinal sections (Fig. 4G). The outer segments recovered a relatively normal shape and internal architecture, although the total number of outer segments was somewhat reduced even in the relatively normal retinal sections. In contrast, outer segment vesiculation, prominent inner segment vacuolation, and cell pyknosis were readily apparent with no evidence of recovery in the more damaged superior retina (Fig. 4H). In sum, the final appearance of the retina after seven days of recovery from UVA exposure included a 15--20% loss of photoreceptor nuclei and outer segments, with a composite pattern of continued degeneration intermixed with progres-
VISUAL DEFICIT AND RETINAL DAMAGE
953
L E A S T
M 0 S T
FIG. 4. Electron micrographs of cOntrol (A, B) and UVA exposed retinas at post-exposure Days 1 (C, D), 3 (E, F), and 7 (G, H). The upper row represents the least amount of damage, and the lower row represents the most damage, observed at each time period (3600×). (See text for description).
sive recovery across the remaining rod photoreceptors. This composite range of photoreceptor degeneration and repair was associated with an asymptotic recovery of absolute thresholds to approximately 0.5 log of previous baselines. DISCUSSION Behavioral absolute thresholds are immediately altered by U V A exposure in rats. The initial UVA-induced threshold shift required stimulus intensities 3-5 log higher than preexposure baselines. The magnitude of threshold change is particularly large given the previous context of only 1-2 log loss with 12-36 hr white light exposure using similar subjects and procedures [1 l]. In direct comparison, the immediate functional loss was up to 10,000 times greater with UVA relative to our previous white light exposure. Nevertheless, the massive threshold loss was not permanent, with 1-2 log recovery uver each succeeding postexposure day. The apparent asymptotic recovery approximately 0.5 log above preexposure thresholds does suggest a minimal long-term functional loss from acute UVA exposure. Consistent with previous data [9,19], the acute UVA exposure produced an immediate vesiculation of the distal
one-third to one-half of the rod outer segments coupled with inner segment vacuolation and cell pyknosis. The disorganization of lamellar outer segment structure is also the initial site of chronic retinal damage in mice exposed to low-level UVA [27]. Using the electron microscope, Sykes and coworkers similarly report that outer segment vesiculation is a predominant histological consequence of low-level white light damage in monkeys [22] and rats [l l]. The qualitatively similar electron microscopic detail in turn suggests that the rhodopsin-bearing outer segment is a common initial focus of radiation damage from both low-level UVA and visible light in the nocturnal and diurnal eye [13, 16, 24]. The present study systematically replicates previous comparisons of psychophysical function and photoreceptor structure in light-induced retinal pathologies. The 3-5 log recovery of absolute thresholds markedly contrasts with the corresponding 15-20 percent loss of cell nuclei and outer segments. In consequence, histologically verified photoreceptor loss may be associated with a broad functional recovery within the same experimental subject. The differential rates of structural and functional recovery then suggest that UVA-induced psychophysical loss is not simply reducible to the electron microscopic structure of the rod photo-
954
HENTON AND SYKES
receptor. The experimental separation of photoreceptor structure and psychophysical function is directly consistent with the previous findings of visual function with extensive photoreceptor damage in rats exposed to visible white light [l, 2, 4, 5, 14, 24]. The experimental evidence thus continues to suggest a significant degree of dissociation between photoreceptor structure and psychophysical function in light-induced retinal damage. Functional recovery in spite of photoreceptor degeneration would not support a physiological reductionism [3] attributing absolute thresholds to the number of surviving photoreceptors or the total amount of available rhodopsin. The operational and epistemological problems of reducing overt behavioral functions to covert structure are discussed through the history of light-induced visual pathology in particular [6, 23, 24], as well as operant and interbehavioral psychology in general [8, 12, 20]. Noell [16], Williams and Howell [25], and Young [26] similarly review visual pathology as an array of causal mechanisms, involving interactive absorption cycles and complex photochemical processes throughout the entire retina. Williams and Howell conclude
that meaningful analyses of light-induced retinal pathology may require neurohistological methods that integrate rather than isolate photoreceptor function within the complex retinal environment. Similar integrational analyses are also commonly applied to behavioral function. To paraphrase Delprato [8], psychophysical thresholds are an aggregate or conjoint function of myriad interdependent variables, each perhaps individually contributing but certainly not individually sufficient to define psychophysical function. In this latter view, photoreceptor pathology and psychophysical function are neither parallel nor redundant measures reducible to a common elemental cause or structure. Rather, photoreceptor damage and behavioral thresholds are different and separable processes, mutually coparticipating in a manydimensioned matrix of aggregate causality. To juxtapose Williams and Deiprato, the integration of electron microscopy and behavioral psychophysics may reveal important differences as well as significant relationships which more accurately characterize the composite structural and functional processes of light-induced visual pathology.
REFERENCES I. Anderson, K. V. and W. K. O'Steen. Black-white discrimination in rats without photoreceptors. Exp Neurol 34: 446-454, 1972. 2. Anderson, K. V. and W. K. O'Steen. Altered response latencies on visual discrimination tasks in rats with damaged retinas. Physiol Behav 12: 633-637, 1974. 3. Ayala, F. J. and T. Dobzhansky (Eds.). Studies in the Philosophy o f Biology. Berkeley: University of California Press, 1975. 4. Bennet, M. H., R. F. Dyer and J. S. Dunn. Light induced retinal degeneration: Effect upon light-dark discrimination. Exp Neurol 34:434 a.45, 1972. 5. Bennet, M. H., R. F. Dyer and J. D. Dunn. Visual-deficit following long-term light exposure. Exp Neurol 38: 80-89, 1973. 6. Birch, D. G. and G. H. Jacobs. Effects of constant illumination on vision in the albino rat. Physiol Behav 19: 255-259, 1977. 7. Boettner, E. A. and J. R. Wolter. Transmission of the ocular media. Invest Ophthalmol l: 776-783, 1962. 8. Delprato, D. J. The interbehavioral alternative to brain-dogma. Psyehol Ree 29: 40%418, 1979. 9. Ham, W. T., H. A. Mueller, J. J. Ruffolo, D. Guerry and R. K. Guerry. Action spectrum for retinal injury from nearunltraviolet radiation in the aphakic monkey. Am J Ophthalmol 93: 299-306, 1982. 10. Henton, W. W. and S. M. Sykes. Within-trial psychophysics with intertrial titration. Behav Anal Lett 3: 43-50, 1983. I 1. Henton, W. W. and S. M. Sykes. Changes in absolute threshold with light-induced retinal damage. Physiol Behav 31: 17%185, 1983.
12. Kantor, J. R. lnterbehavioral Psychology. Granville, OH: Principia Press, 1959. 13. Lanum, J. The damaging effects of light on the retina: Empirical findings, theoretical and practical implications. Surv Ophthaltool 22: 221-249. 1978. 14. Lemmon, V. and K. V. Anderson. Behavioral correlates of constant light-induced retinal degeneration. Exp Neurol 63: 3~49, 1979.
15. Noell, W. K, V. S. Walker, B. S. Kang and S. Berman. Retinal damage by light in rats. Invest Ophthalmol Vis Sei 5: 450-473, 1966. 16. Noell, W. K. Possible mechanisms of photoreceptor damage by light in mammalian eyes. Vis Res 20: II63-1171, 1980. 17. Pitts, D. The human ultraviolet action spectrum. Am J Optom Physiol Opt 51: 946-960, 1974. 18. Pitts, D. C., A. P. Cullen and P. D. Hacker. Ocular effects of ultraviolet radiation from 295-365. Im,est Ophthalmol Vis Sci 16: 932-939, 1977. 19. Schmidt, R. E. and J. A. Zuclich. Retinal lesions due to ultraviolet laser exposure. Invest Ophthalmol Vis Sei 19:11661175, 1980. 20. Skinner, B. F. Are theories of learning necessary? Psyehol Rev 57: 193-216, 1950. 21. Sliney, D. and M, Wolbarsht. Safi, ty with Lasers and Other Optical Sources. New York: Plenum Press, 1980, 22. Sykes, S. M., W. G. Robison, Jr., M. Waller and T. Kuwabara. Damage to the monkey retina by broad-spectrum fluorescent light. Invest Ophthalmol Vis Sei 20: 425-434, 1981. 23. Weizenbaum, F. and F. Colavita. Behaviroal and electrophysiological changes in visual sensitivity following prolonged exposure to constant light. Exp Neurol 48: 440-446, 1975. 24. Williams, T. P. and B. N. Baker tEds.). Ihe EJ.]~,cts qf('onstant Light on Visual Processes. New York: Plenum Press, 1979. 25. Williams, T. P. and W. L. Howell. Action spectrum of retinal light-damage in albino rats. Invest Ophthalmol Vis Sci 24: 285287, 1983. 26. Young, R. A theory of central retinal disease. In: New Directions in Ophthalmic Research, edited by M. L. Sears. New Haven, CT: Yale University Press, 1981. 27. Zigman, S. and T. Vaughn. Near-ultraviolet effects on the lenses and retinas of mice. Invest Ophthalrnol 13: 462-465, 1974.
0031-9384/84 $3.00 + .00
Recovery of Absolute Threshold With UVA-Induced Retinal Damage WENDON
W. HENTON
AND STEPHEN
M. S Y K E S
U.S. Department o f Health and Human Services, Public Health Service, Food and Drug Administration National Center for Devices and Radiological Health, 12709 Twinbrook Parkway, Rockville, MD 20857 R e c e i v e d 20 S e p t e m b e r 1983 HENTON, W. W. AND S. M. SYKES. Recover3, of absolute thresholds with UVA-inducedretinal damage. PHYSIOL BEHAV 32(6)949--954, 1984.--A within-trial psychophysical procedure tracked the initial loss and subsequent recovery of visual thresholds in albino rats exposed to ultraviolet light at 350 nanometers and 0.4 milliwatts per square centimeter. Absolute thresholds increased up to 5 log units immediately following the 15 hour ultraviolet exposure, with a daily recovery of 1-2 log to asymptotic thresholds over a 7-day post-exposure period. The corresponding retinal damage on Day 1 included extensive vesiculation of the photoreceptor outer segments, vacuolation of the inner segments, and pyknosis of cell nuclei. The total number of photoreceptor nuclei and outer segments was unchanged relative to control eyes through post-exposure Day 3. Both nuclei and outer segment counts then consistently decreased 15-20 percent between Days 3-7. The two-stage loss of photoreceptors but daily recovery of absolute thresholds again suggests a significant dissociation of retinal structure and psychophysical function in light-induced ocular pathology. Absolute threshold
Histology
Pathology
Rats
(450× 10 6 W/cm 2 at 12 hr per day [27]). The first evidence of pathology is photoreceptor lesions in the outer segments with 10-16 weeks of chronic UVA, eventually followed by a complete loss of the entire photoreceptor cell layer by 70 weeks [27]. The available data then unanimously describe a disorganization of photoreceptor inner and outer segments, plus possible involvement of other retinal components from cell bodies to pigment epithelium, consequent to UVA radiation in aphakic and nocturnal subjects. Visual function and psychophysical ihresholds remain to be examined for this special circumstance of UVA-induced retinal pathology. In larger perspective, however, visible light-induced ocular pathology has been analyzed in several psychophysical studies. On one hand, behavioral psychophysical function can be virtually unchanged in spite of extensive structural damage to the retina and photoreceptors [1,2]. This apparent independence of behavioral stimulus control and rod photoreceptor pathology has been subsequently replicated by numerous investigators [4, 5, 13, 14]. Yet additional experiments find precisely the opposite result, with intensity functions and electrophysiological measures seemingly graded to the magnitude of photoreceptor damage [6, 11, 23]. Absolute thresholds for example progressively increase with high intensity light exposure, apparently associated with an increased vesiculation of outer segments and vacuolation of inner segments [11]. The data with visible wavelengths collectively suggest that light-induced retinal damage may have differential visual consequences, with threshold intensity functions but not suprathreshold pattern discrimination covarying with progressive photoreceptor pathology. The present study was then an initial comparison of the functional and histological consequences of UVA exposure
E X P E R I M E N T A L studies of the eye reveal a differential absorption of ultraviolet and visible radiation throughout the ocular media [7,21]. The differential absorption of high intensity radiation produces structural alterations along the optic axis as a function of the wavelength, intensity, and duration of surgical or accidental exposure. Ultraviolet radiation (UV), for example, primarily affects the cornea and crystaline lens at wavelengths between 280-400 nanometers (nm) [17,18]. In contrast, retinal lesions in the pigment epithelium and photoreceptors occur with the still longer and visible wavelengths of 400-700 nm [13, 15, 16, 22, 24]. Recent experimental effort, however, has identified a special case of U V A (315-400 nm) radiation damage to the retina rather than the cornea or lens in subjects with high ocular transmission of UV, such as aphakic (lensless) monkeys and nocturnal rodents [9, 19, 27]. The UVA-induced damage in the monkey retina includes ultrastructural changes in the photoreceptor inner and outer segments plus pyknotic cell nuceli within one hr of exposure to a 325 nm laser (3.8 J/cm 2 at 4.0 sec, [19]). Similar pathological changes were found at 24 hr following still lower laser exposures (0.15 J/cm 2 at 0.8 sec), with the most striking changes in the photoreceptor inner and outer segments. An action spectrum for retinal lesions throughout the UVA wavelengths has now been reported for aphakic monkeys [9]. The lesion thresholds increased as a direct function of wavelength, with thresholds approximately six times lower in the UVA at 325 and 350 nm relative to visible blue at 441 nm. Again, structural disorganization occurred within the outer segments, inner segments and cell bodies of the photoreceptors, with additional pathology including the pigment epithelium. Retinal damage also occurs in nocturnal rodents chronically exposed to low level UVA fluorescent lamps
949
950
H E N T O N AND SYKES
in nocturnal rats. More specifically, the initial loss and subsequent recovery of absolute thresholds were obtained for the left eye during daily sessions following a single 15 hr exposure to high intensity UVA. Enucleation of first the right eye and then the left eye provided a within-subject comparison of behavioral thresholds [10] and electron microscopic retinal lesions [22] at either 1, 3, or 7 days postexposure. METHOD
Animals Five male Sprague-Dawley albino rats, approximately 180 days of age, were trained in the psychophysical procedure. Four additional rats of the same age and sex were used as unexposed cage controls for comparative histology. Subjects were raised from birth under approximately 10 lux illumination (L:D 12:12). Each subject was maintained at 95 percent of the previous ad lib home cage weight, and dark adapted 60 min prior to each experimental session.
Apparatus The conditioning apparatus is described in detail in two previous publications [10,11]. Essentially, a neutral tint stimulus key, stainless steel drink tube, and food cup were mounted respectively at the left wall, apex, and right wall of an enclosed triangular area (8.8x8.8× 11.5 cm) behind the front panel of a standard rodent test chamber. A 3.8 cm diameter opening in the front panel limited manipulandum access to only the subject's head. For the present study, a 2800 Hz tone generator was placed in the ceiling, 7.5 cm directly above the subject's head, as a time out stimulus. A 10 cm speaker was placed immediately behind the stimulus key, as a "white" noise discriminative stimulus. The experiment was conducted in a ventillated, sound attenuating booth and controlled by a remote on-line computer. A 1.3 cm diameter white light was centered on the stimulus key by a light pipe (Edmonds Scientific BF 24) connected to an external wedge box. The wedge box included a 2.8 W tungsten bulb mounted behind a Metavac No. 1445 variable density wedge, plus a neutral density filter block to further attenuate light intensity. Stimulus intensity through the variable density wedge was controlled by a two rpm synchronous motor with a shaft mounted rotary switch. The irradiance at the surface of the key was measured by a calibration system with a total estimated error of -+25% [11]. The ultraviolet exposure apparatus for retinal damage consisted of a cylindrical hardware cloth cage (1.25 cm square mesh, 58x30 cm diameter) centered within a larger white enamel cylinder (82.5×57.5 cm diameter). Six blueblack fluorescent lamps (F20T12BLB) were mounted vertically at equal intervals around the circumference of the exterior cylinder. The spectral distribution was maximal at 350 nm (half-power bandwidth approximately 20 nm), with spiked lines at approximately 405 and 435 nm. The average irradiance in the horizontal plane at the center of the hardware cloth cage was 381×10 -6 W/cm 2, ranging from a minimum of 359 to a maximum of 403 × 10-6 W/cm z. Forced air through the chamber maintained the ambient temperature at 25_+2°C.
Proced, re Behavioral procedure. Initially, a continuous reinforcement (CRF) schedule reinforced nose poking responses on
the stimulus key in the presence of either a red or white visual stimulus (approximately 5×10 -s W/cmZ). After 3-5 sessions, onset of a 7.0 sec stimulus trial was contingent upon a licking response on the drink tube. Trials with red and white light were irregularly alternated on a quasirandom schedule. The lick contingency was gradually extended to a variable ratio 15 (VR, range 2-35 responses) over approximately 30 sessions. The final conditioning schedule was thus an observing response schedule of chain VR 15 sig FR 1, producing approximately 250-350 signalled reinforcement trials per 40 min session. A response on the stimulus key during the intertrial interval produced a 20 sec timeout (paired with the overhead 2800 Hz tone). The timeout was followed by a fixed requirement of 35 responses to the next signalled reinforcement trial. The conditioning schedule then reverted back to the standard chain VR 15 sig FR 1. After initial training, signalled reinforcement was altered to a within-trial threshold procedure with intertrial titration. First, "white" noise was substituted for the red light as the standard stimulus for signalled reinforcement, and trials with the white light were only scheduled once every 90 sec (range 35-180 sec). This modification preserved the dark adapted state of the subject by minimizing light exposure during the threshold conditioning sessions. For absolute threshold trials, the intensity of the white light was no longer constant but slowly increased approximately 0.14 log per sec throughout the 7.0 sec trial. Intensity was set at one of nine nominal wedge positions (rotary switch positions) for stimulus onset, and progressively increased until emission of a nose poking response on the stimulus key. A criterion stimulus response was defined by the emission of a two-response sequence--licking followed within 1.5 sec by a response on the lighted key. A trial was repeated if the stimulus response was not preceded by licking, or if the changeover latency was greater than 1.5 sec. The titration of light intensity from Trial N to Trial N + I was determined by the intensity controlling the criterion response sequence. The initial intensity at trial onset was increased to the next higher setting, remained the same, or decreased to the next lower setting for Trial N + I given a lick-stimulus response sequence at 0.0-1.5, 1.5-4.0, or 4.07.0 sec, respectively, during Trial N. The titrated change of initial intensity was approximately 0.2-0.4 log, depending on the specific position of the neutral density wedge. Two exceptions were a downward titration from the lowest wedge setting No. 1, or an upward titration from the highest wedge setting No. 9. The previous intensity was repeated in both cases, with appropriate adjustment of supplemental neutral density filters for subsequent sessions. In sum, the programmed intensity for trial onset would titrate to progressively lower intensities across the first 5-10 trials, and then either titrate down, remain the same, or titrate up for succeeding trials of each session. Given titration of intensity for trial onset, light intensity would then slowly increase approximately 1.0 log within each 7.0 sec trial. The result is an overlapping range of intensities scanned for threshold across successive trials. Absolute threshold was defined within each trial as the minimum intensity controlling the criterion response sequence, excluding the first ten trials as preliminary titration down to stable threshold. Subjects received a total of 25 conditioning sessions using the within-trial threshold procedure prior to the acute UVA exposure. Exposure procedure. The exposure design provided unexposed surgical controls plus right and left eyes enucleated
951
VISUAL DEFICIT AND RETINAL DAMAGE
:Z
TABLE 1 ENCULEATION SCHEDULE OF RIGHT AND LEb'T EYES FOR FIVE RATS ~o
Experimental Day Subject N umber
23
24
25
26
27
28
29
39
31
Baseline
Day 2
Day 1
Day 3
-7
32
| 1
2 3 4 5
R
R R R R
-10
L
L
-11
L L L
-12
..........
i 21
...........
i 25
,,,i,
. . . . . .
t
24
.............
i 28
Trials
on Day 1 (immediately following exposure), Day 3, and Day 7 post-exposure. Subjects 1 and 5 served as controls to determine the effect of enucleating the right eye upon subsequent absolute thresholds for the left eye. The right eyes were enucleated immediately after Session 23 during 30-60 sec surgery under ether anesthesia. Subjects were then returned to the home cage, with baseline thresholds resumed during the regular morning conditioning sessions for the next two days. The right eyes of Subjects 1 and 5 also served as unexposed, normal controls for histological analysis. Subjects 1-5 were U V A exposed overnight for 15 hr prior to Session 26. The enucleation schedule (Table 1) provided comparison of absolute thresholds and retinal damage for two control eyes, three eyes at Day 1, two eyes at Day 3, and three eyes at Day 7 post-exposure. The enucleation schedule also provided five within-subject comparisons of left eye threshold versus right and left eye retinal damage over the 7 day post-exposure period. The exposure routine then included (a) 15 hr U V A exposure from 6 p.m. to 9 a.m., (b) one hr dark adaptation, (c) 40 rain threshold session, (d) immediate enucleation and initiation of histological processing, if scheduled, (e) return to home cage with supplemental feeding, (f) resumption of normal daily schedule with dark adaptation at 9 a.m. Four additional, untrained rats provided unexposed control eyes for corroborative retinal histology. Histological procedure. Eyes were enucleated under ether anesthesia and immediately fixed in cold 2.5% glutaraidehyde and 6% sucrose buffered to pH 7.2 with 50 mM cacodylate. After fixation, the eyes were hemisected vertically through the optic nerve. The nasal hemisphere was embedded in glycol methacrylate, sectioned at 3 microns, and stained with toluidine blue. Samples from the temporal hemisphere were post-fixed in 1% osmium tetroxide, dehydrated by a series of acetone dilutions, and embedded in epoxy resin. Sections for electron microscopy were double stained with uranyl acetate and lead citrate, then photographed with a J E O L 100C electron microscope. In all cases, the retinas were sectioned transversely with the photoreceptots cut as nearly as possible along their long axis. The number of photoreceptor nuclei were determined from light micrographs of four regions of the temporal retina, each 1-2 mm from the optic disk (total magnification of 1000×). Photoreceptqr outer segment density was determined from electron micrographs of the same regions (magnification of 4500x). Outer segment density was determined by counting the number of outer segments intersecting a line 15 /zm from and parallel to the apical edge of the pigment epithelium.
FIG. 1. Comparison of absolute thresholds before and after UVA exposure for Subject 1. The filled data points in each panel represent a single threshold determination during the final baseline session, and three consecutive daily sessions following UVA exposure. A trial with no obtained threshold is indicated by an open square.
RESULTS For the enucleation control procedure, removal of the right eye did not affect the subsequent thresholds for the left eye. Median thresholds for the two sessions before and after right eye enucleation were 1.1 and 1.4 vs. 0.9 and 0.8× 10-12 W/cm ~ for Subject 1, compared to 2.4 and 2.3 vs. 2.2 and 2.0×10 - u W/cm 2, respectively, for Subject 5. In consequence, any loss of absolute threshold following U V A exposure would not be a simple artifact of the enucleation procedure. Figure 1 illustrates the daily change in thresholds following U V A exposure for Subject 1. Each data point is a single threshold determination during the final baseline session and post-exposure sessions on Days l, 2, and 3. Excluding the first 10 trials as preliminary titration, baseline thresholds ranged from 5.6 (Trial 15) to 10.0 (Trials 11 and 17), with a median of 7.7x 10-12 W/cm 2, for the final preexposure session (left panel Fig. 1). After 15 hr U V A and 1 hr dark adaptation (Day 1), programmed intensities on Trial 1 and Trial 2 failed to control the changeover to stimulus responses. Threshold was first recorded on Trial 3 only after intensity titration up to 1 . 6 x l 0 -6 W/cm 2. Subsequent thresholds ranged between 0.4 and 3.5 x 10-6 W/cm ~ over the succeeding trials of Day 1. Absolute thresholds thus stabilized approximately 5 log units above previous baselines with acute UVA exposure. Thresholds did not progressively deteriorate but systematically recovered over subsequent post-exposure sessions. Daily thresholds decreased towards baseline and recovered more than 2 log by Day 2, and 3 log by Day 3 for Subject 1. Figure 2 plots this recovery of function following acute UVA exposure for all subjects. Thresholds were tracked either 3 days (Subjects 1 and 2) or 7 days (Subjects 3-5) prior to enucleation of the remaining left eye and experiment termination. For all subjects, the largest threshold shift was on Day l immediately following 15 hr U V A exposure. Threshold increase ranged from more than 3 log (Subjects 3 and 4) to more than 5 log (Subjects 1, 2, and 5) relative to previous baselines. F o r Subject 5, threshold change exceeded the maximum system capability, with no behavioral thresholds found at even the highest available intensity of 4×10 -6 W/cm 2. Threshold gradually recovered by 1-2 log per postexposure day, with apparent asymptote at post-exposure
952
H E N T O N AND SYKES -5
$1
$3
T
30 7i
300
-6
E
E 2k
-7
O
-8
E 0
~a . 0
-9
_=
C u)
-5
E
-6
4-' U~
g, --I
e®
E ffl
-10
-12
20
0 0 0 0 e¢,1
-11 (/)
200
I B
L 1
I 2
I 3
I B
L 1
I 2
I 3
I 1
$4
I 2
I 3
I 4
I 5
I 6
t~ >4 x 10-6 \ \
I 7
S5
I c
I 1
I
i 3
I
I
_
I
J 7
0
Days After Exposure FIG. 3. Counts of photoreceptor nuclei (left axis) and outer segments (right axis) for control retinas (C) and retinas at 1,3 and 7 days following 15 hr exposure to UVA.
-7
-8 -9 -10
•
-11 -12
t B
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
1
2
3
4
5
6
7
B
1
2
3
4
5
6
7
Days Post Exposure FIG. 2. Absolute thresholds before and after UVA exposure for each of five rats. Each data point is the median threshold for the last baseline session (B) and successive post-exposure sessions (I-7). Subjects 1 and 2 and Subjects 3-5 were trained for three and seven post-exposure days, respectively, until enucleation of the remaining left eye. For Subject 5, the threshold loss on post-exposure Day 1 exceeded the maximum intensity of the optical system at 4× 10-6 W/cm2
Day 3 (Subject 3) or Day 5 (Subjects 4 and 5). Absolute thresholds virtually recovered by the final post-exposure sessions, stabilizing approximately 0.5 log above preexposure baselines for Subjects 3-5. Figure 3 presents separate counts of photoreceptor nuclei and outer segments for both control and exposed eyes. The density of photoreceptors was unchanged relative to control eyes at postexposure Days 1-3. However, cell nuclei then decreased approximately 15% by Day 7. Photoreceptor outer segments also remained relatively unchanged throughout postexposure Day 3, then decreased approximately 20% by Day 7. Further, continued pyknosis of cell nuclei and vesicuolation of outer segments suggests that the photoreceptor degeneration had not yet stabilized by post-exposure Day 7. For all subjects, photoreceptor pathology was variable across full cross sections of the retina, with the superior retina routinely more damaged than the inferior retina at each post-exposure time interval. In contrast, the pigment epithelium and inner retinal layers were relatively unchanged across the 7 post-exposure days for all subjects. Figure 4 compares the ultrastructural characteristics of control and exposed eyes at 1, 3, and 7 days post-exposure. The upper and lower rows of electron micrographs represent the minimal and maximal pathology observed at each time interval. For the control eyes, all retinal layers appeared relatively normal (Fig. 4A). However, the distal ends of o c -
casional outer segments were bulbous and somewhat separated from the pigment epithelium (Fig. 4B). Scattered pyknotic nuclei and vacuoles in the pigment epithelium were also occasionally observed in the control eyes. The rod photoreceptor morphology was clearly disrupted immediately following UV exposure and testing on Day 1 (Fig. 4C and 4D). Photoreceptors were routinely bent and misshapen, with irregular cell profiles and vesiculation of the distal portion of all outer segments. Inconsistent and nonspecific vacuole formations appeared in the inner segments. In contrast, prominent organelles such as the mitochondria appeared normal in the inner segments. Changes in the pigment epithelium and other retinal layers were not observed at Day 1. This range of photoreceptor damage was associated with an absolute threshold change of 3-5 log units. The reinal damage on post-exposure Day 3 was an intermixed pattern of partial recovery of outer segment vesiculation in some photoreceptors (Fig. 4E) combined with a continuing vesiculation and nuclei degeneration of neighboring photoreceptors (Fig. 4F). In particular, the outer segment vesiculation was at least equal to or perhaps increased in most photoreceptors relative to Day 1 (Fig. 4F compared to Fig. 4C and 4D). The inner segments continued to contain nonspecific vacuole formations. However, the distorted orientation of the individual photoreceptors would prevent any accurate association between the inner segment morphology and outer segment recovery. This differential pattern of photoreceptor degeneration/repair resulted in uneven damage across the retinal sections, with severely vesiculated cells directly adjacent to relative normal cells. The composite photoreceptor morphology was associated with a partial recovery of absolute thresholds to approximately 1.0-2.5 log above previous control baselines. On post-exposure Day 7, distinct recovery approximating normal morphology was apparent in various retinal sections (Fig. 4G). The outer segments recovered a relatively normal shape and internal architecture, although the total number of outer segments was somewhat reduced even in the relatively normal retinal sections. In contrast, outer segment vesiculation, prominent inner segment vacuolation, and cell pyknosis were readily apparent with no evidence of recovery in the more damaged superior retina (Fig. 4H). In sum, the final appearance of the retina after seven days of recovery from UVA exposure included a 15--20% loss of photoreceptor nuclei and outer segments, with a composite pattern of continued degeneration intermixed with progres-
VISUAL DEFICIT AND RETINAL DAMAGE
953
L E A S T
M 0 S T
FIG. 4. Electron micrographs of cOntrol (A, B) and UVA exposed retinas at post-exposure Days 1 (C, D), 3 (E, F), and 7 (G, H). The upper row represents the least amount of damage, and the lower row represents the most damage, observed at each time period (3600×). (See text for description).
sive recovery across the remaining rod photoreceptors. This composite range of photoreceptor degeneration and repair was associated with an asymptotic recovery of absolute thresholds to approximately 0.5 log of previous baselines. DISCUSSION Behavioral absolute thresholds are immediately altered by U V A exposure in rats. The initial UVA-induced threshold shift required stimulus intensities 3-5 log higher than preexposure baselines. The magnitude of threshold change is particularly large given the previous context of only 1-2 log loss with 12-36 hr white light exposure using similar subjects and procedures [1 l]. In direct comparison, the immediate functional loss was up to 10,000 times greater with UVA relative to our previous white light exposure. Nevertheless, the massive threshold loss was not permanent, with 1-2 log recovery uver each succeeding postexposure day. The apparent asymptotic recovery approximately 0.5 log above preexposure thresholds does suggest a minimal long-term functional loss from acute UVA exposure. Consistent with previous data [9,19], the acute UVA exposure produced an immediate vesiculation of the distal
one-third to one-half of the rod outer segments coupled with inner segment vacuolation and cell pyknosis. The disorganization of lamellar outer segment structure is also the initial site of chronic retinal damage in mice exposed to low-level UVA [27]. Using the electron microscope, Sykes and coworkers similarly report that outer segment vesiculation is a predominant histological consequence of low-level white light damage in monkeys [22] and rats [l l]. The qualitatively similar electron microscopic detail in turn suggests that the rhodopsin-bearing outer segment is a common initial focus of radiation damage from both low-level UVA and visible light in the nocturnal and diurnal eye [13, 16, 24]. The present study systematically replicates previous comparisons of psychophysical function and photoreceptor structure in light-induced retinal pathologies. The 3-5 log recovery of absolute thresholds markedly contrasts with the corresponding 15-20 percent loss of cell nuclei and outer segments. In consequence, histologically verified photoreceptor loss may be associated with a broad functional recovery within the same experimental subject. The differential rates of structural and functional recovery then suggest that UVA-induced psychophysical loss is not simply reducible to the electron microscopic structure of the rod photo-
954
HENTON AND SYKES
receptor. The experimental separation of photoreceptor structure and psychophysical function is directly consistent with the previous findings of visual function with extensive photoreceptor damage in rats exposed to visible white light [l, 2, 4, 5, 14, 24]. The experimental evidence thus continues to suggest a significant degree of dissociation between photoreceptor structure and psychophysical function in light-induced retinal damage. Functional recovery in spite of photoreceptor degeneration would not support a physiological reductionism [3] attributing absolute thresholds to the number of surviving photoreceptors or the total amount of available rhodopsin. The operational and epistemological problems of reducing overt behavioral functions to covert structure are discussed through the history of light-induced visual pathology in particular [6, 23, 24], as well as operant and interbehavioral psychology in general [8, 12, 20]. Noell [16], Williams and Howell [25], and Young [26] similarly review visual pathology as an array of causal mechanisms, involving interactive absorption cycles and complex photochemical processes throughout the entire retina. Williams and Howell conclude
that meaningful analyses of light-induced retinal pathology may require neurohistological methods that integrate rather than isolate photoreceptor function within the complex retinal environment. Similar integrational analyses are also commonly applied to behavioral function. To paraphrase Delprato [8], psychophysical thresholds are an aggregate or conjoint function of myriad interdependent variables, each perhaps individually contributing but certainly not individually sufficient to define psychophysical function. In this latter view, photoreceptor pathology and psychophysical function are neither parallel nor redundant measures reducible to a common elemental cause or structure. Rather, photoreceptor damage and behavioral thresholds are different and separable processes, mutually coparticipating in a manydimensioned matrix of aggregate causality. To juxtapose Williams and Deiprato, the integration of electron microscopy and behavioral psychophysics may reveal important differences as well as significant relationships which more accurately characterize the composite structural and functional processes of light-induced visual pathology.
REFERENCES I. Anderson, K. V. and W. K. O'Steen. Black-white discrimination in rats without photoreceptors. Exp Neurol 34: 446-454, 1972. 2. Anderson, K. V. and W. K. O'Steen. Altered response latencies on visual discrimination tasks in rats with damaged retinas. Physiol Behav 12: 633-637, 1974. 3. Ayala, F. J. and T. Dobzhansky (Eds.). Studies in the Philosophy o f Biology. Berkeley: University of California Press, 1975. 4. Bennet, M. H., R. F. Dyer and J. S. Dunn. Light induced retinal degeneration: Effect upon light-dark discrimination. Exp Neurol 34:434 a.45, 1972. 5. Bennet, M. H., R. F. Dyer and J. D. Dunn. Visual-deficit following long-term light exposure. Exp Neurol 38: 80-89, 1973. 6. Birch, D. G. and G. H. Jacobs. Effects of constant illumination on vision in the albino rat. Physiol Behav 19: 255-259, 1977. 7. Boettner, E. A. and J. R. Wolter. Transmission of the ocular media. Invest Ophthalmol l: 776-783, 1962. 8. Delprato, D. J. The interbehavioral alternative to brain-dogma. Psyehol Ree 29: 40%418, 1979. 9. Ham, W. T., H. A. Mueller, J. J. Ruffolo, D. Guerry and R. K. Guerry. Action spectrum for retinal injury from nearunltraviolet radiation in the aphakic monkey. Am J Ophthalmol 93: 299-306, 1982. 10. Henton, W. W. and S. M. Sykes. Within-trial psychophysics with intertrial titration. Behav Anal Lett 3: 43-50, 1983. I 1. Henton, W. W. and S. M. Sykes. Changes in absolute threshold with light-induced retinal damage. Physiol Behav 31: 17%185, 1983.
12. Kantor, J. R. lnterbehavioral Psychology. Granville, OH: Principia Press, 1959. 13. Lanum, J. The damaging effects of light on the retina: Empirical findings, theoretical and practical implications. Surv Ophthaltool 22: 221-249. 1978. 14. Lemmon, V. and K. V. Anderson. Behavioral correlates of constant light-induced retinal degeneration. Exp Neurol 63: 3~49, 1979.
15. Noell, W. K, V. S. Walker, B. S. Kang and S. Berman. Retinal damage by light in rats. Invest Ophthalmol Vis Sei 5: 450-473, 1966. 16. Noell, W. K. Possible mechanisms of photoreceptor damage by light in mammalian eyes. Vis Res 20: II63-1171, 1980. 17. Pitts, D. The human ultraviolet action spectrum. Am J Optom Physiol Opt 51: 946-960, 1974. 18. Pitts, D. C., A. P. Cullen and P. D. Hacker. Ocular effects of ultraviolet radiation from 295-365. Im,est Ophthalmol Vis Sci 16: 932-939, 1977. 19. Schmidt, R. E. and J. A. Zuclich. Retinal lesions due to ultraviolet laser exposure. Invest Ophthalmol Vis Sei 19:11661175, 1980. 20. Skinner, B. F. Are theories of learning necessary? Psyehol Rev 57: 193-216, 1950. 21. Sliney, D. and M, Wolbarsht. Safi, ty with Lasers and Other Optical Sources. New York: Plenum Press, 1980, 22. Sykes, S. M., W. G. Robison, Jr., M. Waller and T. Kuwabara. Damage to the monkey retina by broad-spectrum fluorescent light. Invest Ophthalmol Vis Sei 20: 425-434, 1981. 23. Weizenbaum, F. and F. Colavita. Behaviroal and electrophysiological changes in visual sensitivity following prolonged exposure to constant light. Exp Neurol 48: 440-446, 1975. 24. Williams, T. P. and B. N. Baker tEds.). Ihe EJ.]~,cts qf('onstant Light on Visual Processes. New York: Plenum Press, 1979. 25. Williams, T. P. and W. L. Howell. Action spectrum of retinal light-damage in albino rats. Invest Ophthalmol Vis Sci 24: 285287, 1983. 26. Young, R. A theory of central retinal disease. In: New Directions in Ophthalmic Research, edited by M. L. Sears. New Haven, CT: Yale University Press, 1981. 27. Zigman, S. and T. Vaughn. Near-ultraviolet effects on the lenses and retinas of mice. Invest Ophthalrnol 13: 462-465, 1974.