Young RW. Solar radiation and age-related macular ... spectral photic damage to primate cones. Vis Res. ... noy G. Dark adaptation in age-related maculopathy.
Aging of the Human Photoreceptor Mosaic: Evidence for Selective Vulnerability of Rods in Central Retina Christine A. Curcio,*^ C. Leigh Millican,* Kimberly A. Allen,X and Robert E. Kalina%
Purpose. Because previous studies suggested degeneration and loss of photoreceptors in aged human retina, the spatial density of cones and rods subserving the central 43° of vision as a function of age was determined. Methods. Cones and rods were counted in 27 whole mounted retinas from donors aged 27 to 90 years with macroscopically normal fundi. Photoreceptor topography was analyzed with new graphic and statistical techniques. Results. Changes in cone density throughout this age span showed no consistent relationship to age or retinal location, and the total number of foveal cones was remarkably stable. In contrast, rod density decreased by 30%, beginning inferior to the fovea in midlife and culminating in an annulus of deepest loss at 0.5 to 3 mm eccentricity by the ninth decade. Space vacated by dying rods was filled in by larger rod inner segments, resulting in a similar rod coverage at all ages. At the temporal equator, cone density declined by 23%, but rods were stable throughout adulthood. Conclusions. The stability of both rod coverage and rhodopsin content despite decreasing cell number suggests plasticity of the adult rod system and that age-related declines in scotopic sensitivity may be due to postreceptoral factors. There is no evidence for the massive loss of foveal cones required to explain even modest decrements in acuity, consistent with evidence that visual deficits at high photopic levels may be largely due to optical factors. Why the rods of central retina, which share a common support system and light exposure with the neighboring cones, are preferentially vulnerable to aging remains to be determined. Invest Ophthalmol Vis Sci. 1993;34:3278-3296.
JMLost older persons have deficits in both photopic1 and scotopic2 vision that are unrelated to detectable disease but may still impair daily tasks. Normal visual function depends on the integrity of a number of factors, including clarity of optical media, sampling and sensitivity characteristics of the photoreceptor mosaic, intraretinal pathways, and connections among brain structures subserving vision, all of which may show age-related changes. Regarding the optical media, pupillary constriction and lens opacification together result in greatly reduced retinal illumination in the elderly.3 One of the major questions in the study of the From the Departments of *Ophthalmology and ~fCell Biology and Anatomy, University of Alabama at Birmingham, and the Department of %Ophlhalmology, University of Washington, Seattle. Supported by National Eye Institute grant EY06109 and a Research Career Development Award from Research to Prevent Blindness, New York, Nero York (CAC). Submitted for publication December 30, 1992; accepted April 14, 1993. Proprietary interest category: N. Reprint requests: Christine A. Curcio, PhD, Department of Ophthalmology, University of Alabama at Birmingham, Birmingham, AL 35294-0009.
3278
aging visual system is to determine the relative contributions of optical and neural factors to diminished photopic and scotopic functions. In this study, we have asked how the spatial density and size of the entrance aperture of cones and rods in the posterior pole of the human retina change with age, because previous studies on the aging neurosensory retina have suggested degeneration and outright loss of photoreceptors. Both rod4 and cone5 outer segments become disorganized, and refractile bodies, presumably lipofuscin, accumulate in cone inner segments with age.6-7 Displacement and patchy loss of photoreceptor nuclei may occur without significant defects in underlying pigment epithelium (PE), Bruch's membrane, or choriocapillaris.8 The number of photoreceptor nuclei are inversely correlated with age in black donors and with lipofuscin levels in the adjacent PE in white donors.9 A recent study10 suggested that a loss of rods occurs early in adult life and that cone density steadily diminishes over adulthood in
Investigative Ophthalmology & Visual Science, November 1993, Vol. 34, No. 12 Copyright © Association for Research in Vision and Ophthalmology
Aging Human Photoreceptor Mosaic
far peripheral retina. In contrast, there is no evidence for a decrease in the peak density of foveal cones, although the variability among individuals is very high,11 and there may be loss of foveal photoreceptors in persons older than 90 years.12 Because many previous studies specified retinal location imprecisely, it is difficult to compare changes in retinal anatomy to age-related changes in visual function. Thus, we used a retinal whole-mount preparation and computer methods for quantifying retinal topography,1113 confining our analysis to donor eyes without evidence of ocular disease to separate age-related from pathological changes. We find that within the central 43° diameter of vision, rods but not cones are lost during adulthood. We consider the functional implications and cellular mechanisms that might underlie selective cell loss in the retina. Some results were reported in abstract form.14
METHODS Tissue Collection, Preparation, and Selection Criteria Human tissues used in this study were obtained according to the tenets of the Declaration of Helsinki, following informed consent and approval by the institutional review boards at the University of Washington and University of Alabama at Birmingham. Eyes obtained from eye bank donors within 3 hours of death were fixed by immersion in 4% paraformaldehyde and 0.5% glutaraldehyde in 0.1 M phosphate buffer after the cornea and lens had been removed. After fixation for at least several weeks, eyes were examined under a dissecting microscope and those with grossly visible macular drusen, pigmentary disturbances, signs of past ocular surgery or disease, or postmortem folds and detachment in the posterior retina were rejected. Eyes with common age-related degenerations in the far peripheral retina, such as microcystoid or paving stone degeneration15 were used. One eye was obtained at surgery from a 32-year-old woman with recurrent muco-epidermoid carcinoma of the ethmoid sinus. The final sample consisted of 16 men and 11 women aged 27 to 90 years (Table 1), a male bias that is representative of the donor pool of the two eye banks providing our tissue. Photoreceptor distributions of the seven youngest persons (Table 1) were described previously.11-16 A 12-mm-wide belt of retina containing the fovea, optic disk, and horizontal meridian was prepared as a whole mount.11 The retina was dissected free from the PE, flattened on a slide, rinsed in water, and cleared under a coverslip overnight in dimethyl sulfoxide. Either 100% glycerol or a combination of glycerol and polyvinyl alcohol17 was applied to the tissue, and a cov-
3279
erslip was mounted and sealed with nail polish. Retinas exhibited a 2 to 13% expansion in area during processing without any evidence of age-related differences (linear regression, log volume changes vs age, r = 0.264, NS). Cell density estimates were not corrected for this small expansion. Whole mounts were viewed with a combination of Nomarski differential interference contrast and video to exclude specimens with microscopic artifacts in the foveal center1' such as optically indistinct photoreceptors, breaks in the external limiting membrane, and steep slopes that presented photoreceptors in longitudinal rather than cross-sectional view. Twenty foveas that met these criteria were studied in detail (Table 1). Seven retinas with minor foveal defects were used for less detailed maps that allowed reasonable estimates of total cone number but not peak cone density.
Retinal Topography: Data Collection Counts were made from Nomarski differential interference contrast video images of the whole mounted retina, using the stylus of a digitizing tablet to mark the inner segments of counted cells.1113 Photoreceptor types were distinguished by the diameter of their inner segments, because cones are larger than rods by a factor of about three throughout the retina. Foveal cones and rods were viewed with a 100X objective, for a 37-Aim square counting window at a final video magnification of 5000X. Nonfoveal cones were viewed with either a 40X or 60X objective (window dimensions/ final magnification of 94 /urn/ 2000X and 62 ixm/ 3000X, respectively). At each location, counts from two to four adjacent windows were pooled, and for both rods and cones, the half-width of the 95% confidence interval for mean density was 11 to 12%. Locations were more closely spaced near the fovea than in the periphery to capture rapidly changing gradients of cell density. In the rod-free zone of the 20 well-preserved foveas (Table 1), counts were made in as many adjacent 100X windows as possible within a 7 X 7 grid to detect peak cone density. In six other foveas, counts were made in a grid of 25 100X windows that were 0.076 mm apart. Outside the rod-free zone up to 6 mm eccentricity in all retinas, data were collected from as many locations as possible in a foveocentric spiral of 72 points that evenly covered the retinal surface.18 Thus, the maximum number of sample points for any one retina was 49 + 72 = 121. Cell densities from each retina were indexed by coordinates on the retinal sphere. The origin of the coordinate system was either the point of highest cone density if measurable or the bottom of the external foveal pit if not. For convenience, we consider a line through the foveal center and optic disk as the horizontal meridian, although the optic disk is actually slightly above the horizontal meridian. The directions nasal
Investigative Ophthalmology & Visual Science, November 1993, Vol. 34, No. 12
3280 TABLE
l. Donor Eyes PMD (hr), Death
¥1
A*
Age Eye (years) Sex Enucleation Fix
1
L
2 3 4 5
R L L, Rt L L R R L L L L R R R R R R L R R L L R L L
6 7 8 9 10 11 12 13 14 15 16
17 18 19 20
21 22 23 24 25 26 27 Totals
Retina Mapped*
A
Patient No.
R
27 32 34 35 35 36 37 44 48 52 52 56 56 58 58 61 66
68 72 73 74 75 82 82 86 89 90
M F M F F M F F M M
M M F M F M M M F M M M M M F F F
0:15 —
1:35 0:27 1:30 1:51 2:00 0:22 0:40 1:25 NA :44 :30 :43
:25 :41 ]1:15 2:05 2:14 1:35 1:15 2:18 2:33 1:50 1:08 1:05 2:00
Cause of Death or Reason for Enucleation
NA Multiple trauma
0:15 2:00 0:55 NA 2:26 2:30 1:07 2:15 NA
3:25 3:00 1:58 2:13 1:55 2:11 1:45 NA
2:29 NA
1:30 3:30 NA
1:50 1:58 1:05 2:30
Mucoepidermoid carcinoma Head injury and respiratory arrest Head injury Brain tumor Pulmonary embolism Head injury Subarachnoid hemorrhage Metastatic carcinoma, non-Hodgkin lymphoma Cerebrovascular accident Renal failure, adenocarcinoma of lung Lymphoma, cardiac arrest Septicemia Cardiac arrest Cardiac arrest Cardiac arrest Multiple myeloma Cardiac arrest Aortic stenosis Multiple trauma Cardiac arrest Cerebrovascular accident Cardiomvopalhy ASHD Cardiac arrest Mel astatic breast cancer Cardiac arrest
Age Cones FovClr Rods Group X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X X
X
X
X
X
X
X
X X
X X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X X
X
X X
X
X
26
20
24
2 2 2 9 9
2 2 2 3 3 3 3 3 3 3 4 4 4 4 4
NA = not applicable. * Cones: outside the rod-free zone and s6 mm eccentricity. FovCtr = rod-free zone; all rods 45%) around the edge of the rod free zone, high (30
3283
Aging Human Photoreceptor Mosaic
to 45%) within 1 mm of the fovea and also around the optic disk, and low (< 15%) elsewhere. Both the maps (Fig. 6) and the similarity statistic reveal significant between-group differences in rod topography (P = 0.007), and in contrast to cones, these age-related differences are monotonic with age and consistent across retinal locations. With increasing age (left column of Fig. 6), rod density is reduced in both extrafoveal and foveal retina, as demonstrated by a shift from orange to yellow isodensity contours and an expansion of the low density blue-green contours, respectively. The difference maps (right column of Fig. 6) indicate that loss of rods relative to Group 1 begins in the inferior retina of Group 2 (green in Fig. 6C) and is widespread across the central retina of Group 3 (Fig. 6E), in which densities are on average 81% of those in Group 1. Rod densities in the oldest group (Group 4) are only 72.4% of those in Group 1, and the loss is particularly great in an annulus from 0.5 to 3 mm eccentricity (purple in Fig. 6G), where rod densities are