- Email: [email protected]
Age-related changes in the human retina
Age-related changes in the human retina Carlo Cavallotti, * MD, PhD; Marco Artico, * MD, PharmD; Nicola Pescosolido, t MD; Francesca Maria Tranquilli Leali, * BSc; Janos Feher,:j: MD ABSTRACT • RESUME Background: In a previous study, scanning electron microscopy (SEM) showed agerelated changes in the rat retina. We carried out a study to evaluate age-related changes in the human retina. Methods: Samples of fresh retinal tissue obtained from younger (age 22 years or less) and older (age 66 years or more) donors were studied by means of traditional histologic methods and by SEM. Eight retinas were obtained from four donors whose corneas had been used for transplantation, and four retinas were obtained from four subjects whose eyes had been enucleated owing to injury. All morphologic results were subjected to quantitative analysis of images. The concentration of cytoplasmic (free) and structural (tissue-associated) protein in retinal tissue homogenates was determined by means of biochemical methods. Results: There was a decrease in all features studied with the exception of structural protein concentration. The mean retinal thickness (and standard error of the mean) was 426 (34.2) IJm in the younger subjects and 261 ( 18.9) !Jm in the older subjects. The mean numbers of ganglion cells (and standard error of the mean) were 413.5/mm 2 (32.3/mm 2) and 256.2/mm 2 (26.8/mm 2) respectively, of capillaries 3.6/mm 2 ( 1.4/mm 2) and 1.8/mm 2 ( 1.2/mm 2) respectively, of synaptic bodies 122.4 (4.9) conventional units (CU)/area observed and 38.5 ( 1.6) CU/area observed respectively, of cellular processes 82.3 (3.1) CU/area observed and 13.1 ( 1.5) CU/ area observed respectively, and of intercellular connections 36.4 (2.5) CU/area observed and 14.3 ( 1.4) CU/area observed respectively. The mean concentration of total protein per milligram of fresh tissue (and standard error of the mean) was 92.1 ( 1.8) IJg in the younger subjects and 78.7 ( 1.3) IJg in the older subjects; the corresponding values for cytoplasmic protein were 27.6 ( 1.3) IJg and I 1.8 (0.8) !Jg, and for structural protein, 64.4 ( 1.6) IJg and 86.9 ( 1.4) IJg.AII differences between the younger and older subjects were significant (p < 0.00 I) with the exception of mean concentration of cytoplasmic and of structural protein. Interpretation: The human retina undergoes specific changes with aging. SEM provides new morphometric information regarding age-related changes in photoreceptor cells, bipolar cells and ganglion cells that
From the Departments of *Human Anatomy and tOphthalmology, University of Rome "La Sapienza," Rome, Italy, and :j:the Eye Clinic, University of Pees, Pees, Hungary
Cardiovascolari e Respiratorie, Sezione di Anatomia Umana, University of Rome "La Sapienza," Via A. Borelli 50,00161 Rome, Italy; fax 0039/06/4957669, cavallotti @uniroma1.it
Originally received Feb. 6, 2003 Accepted for publication July 12, 2003
This article has been peer-reviewed.
Correspondence to: Dr. Carlo Cavallotti, Dipartimento di Scienze
Can J Ophthalmol 2004;39:61-8
Age-related retinal changes-Cavallotti et al
61
Age-related retinal changes-Cavallotti et al
increases our understanding of this topic. Our results may be adopted as a model or as normal values when studying other changes that may occur in the human retina in pathological conditions.
Contexte : Dans une etude precedente, Ia microscopie a rayons X avait fait voir dans Ia retine des modifications associes au vieillissement chez les rats. Nous avons mene une etude pour evaluer les modifications de Ia retine lies au vieillissement chez les etres humains. Methodes : Des echantillons de tissus retiniens frais, obtenues chez des jeunes gens de 22 ans ou moins et des personnes agees de 66 ans ou plus, ont fait I' objet d'etudes selon les methodes histologiques traditionnelles et par microscopie a rayon X. Huit retines ont ete obtenues de quatre donateurs dont les cornees avaient servi pour une greffe et quatre autres, de personnes qui avaient subi !'enucleation a Ia suite d'une blessure. Les resultats ont ete soumis a une analyse quantitative d'images. La concentration de proteines cytoplasmiques (libres) et structurelles (associees au tissu) dans les homogenats de tissus retiniens a ete etablie au moyen de precedes biochimiques. Resultats : II y a eu une reduction de toutes les caracteristiques etudiees, sauf Ia concentration des proteines structurelles. L'epaisseur moyenne des retines (et erreur type sur Ia moyenne [ETM]) etait de 426 IJm (34,2 1-1m) chez les jeunes sujets et de 261 IJm (18,9 IJm) chez les autres. Le nombre moyen de cellules ganglionnaires (et ETM) etait de 413,5/mm 2 (32,3/mm2) et de 256,2/mm 2 (26,8/mm 2) respectivement; celui des capillaires, de 3,6/mm 2 ( 1,4/mm 2) et de I ,8/mm 2 (I ,2/mm 2) respectivement. Le nombre moyen (en unites conventionnelles par secteur observe) (et ETM) de corps synaptiques chez les deux groupes etait de 122,4 (4,9) et de 38,5 (I ,6) respectivement; celui des processus capillaires, de 82,3 (3, I) et de 13, I (I ,5) respectivement; et celui des liens intercellulaires, de 36,4 (2,5) et de 14,3 (I ,4) respectivement. La concentration moyenne des proteines entieres par milligramme de tissu frais (et ETM) etait de 92, I IJg (I ,8 IJg) chez les jeunes sujets et de 78,7 IJg (I ,3 IJg) chez les aines. Les valeurs correspondantes des proteines cytoplasmiques etaient de 27,6 IJg (I ,3 IJg) et de I I,8 IJg (0,8 IJg), et celles des proteines structurelles, de 64,4 IJg (I ,6 IJg) et de 86,9 IJg (I ,4 IJg). Les ecarts entre les jeunes sujets et les sujets ages etaient significatifs (p < 0,00 I), sauf pour Ia concentration moyenne des proteines cytoplasmiques et structurelles. Interpretation : La retine humaine subit des modifications particulieres en vieillissant. La microscopie a rayons X fournit de !'information morphometrique nouvelle en ce qui a trait aux modifications lies a l'age dans les cellules photoreceptrices, bipolaires et ganglionnaires, augmentant ainsi notre comprehension du sujet. Nos resultats peuvent etre adoptes comme modeles ou comme valeurs normales pour l'etude des autres modifications qui peuvent survenir dans Ia retine humaine dans des conditions pathologiques.
O
cular tissues in both humans and animals have been widely studied by scanning electron microscopy (SEM). 1- 3 The retinal neurons,Z the inner and outer segments of the photoreceptors, 4 and the retinal pigment epithelium (RPE) and its relations with the photoreceptors 3·5 have also been well described. The entire retina in both animals
62
CAN J OPHTHALMOL-VOL. 39, NO. I, 2004
and humans has been analysed by SEM.6-8 SEM has been used primarily for studying surface morphology, enabling great advances in the investigation of various biologic materials. It has also been used in biologic research, for examining the structural organization of tissues and cells. Subsequently, other biologic applications of SEM have been introduced. 8
Age-related retinal changes-Cavallotti et al
Recently, our group studied the SEM features of the rat retina by means of quantitative analysis of images. 9 We carried out a study to determine whether the age-related changes that we observed in the rat retina9 also occur in the human retina. METHODS
We studied 12 human retinas: 8 obtained from four donors whose corneas had been used for transplantation, and 4 obtained from four subjects whose eyes had been enucleated owing to injury. Four of the subjects were aged 22 years or less and were classified as young, and the other four were aged 66 years or more and were classified as old. Subjects with metabolic, tumoral or vascular retinal damage were excluded from the study. In a previous paper we described age-related changes in the human optic nerve obtained from cadaveric donors. 10 In the current study we used only samples of retinas from living humans because postmortem phenomena may bring about early modifications in retinal morphology. 11 Our study was approved by the ethics committees of the hospitals involved, and patients or their relatives gave informed written consent. The investigations were performed according to the guidelines of the Declaration of Helsinki. Retinal thickness is known to vary between the central and peripheral regions and between the temporal and nasal regions. 1 SEM of the rat retina showed that the number of neurons and their morphology varied within each retinallayer. 2 Moreover, wide individual variations have been described in the neuronal layers of the human retina. 4 Consequently, all retinal quantifications must be performed precisely, and the topographic location must be the same for each retina investigated. After removal, the globe was dissected with a razor blade, and samples of intact retinal tissue (located precisely in the same site in each eye, equatorially in the nasal region) were harvested. The entire thickness of the retina was measured with a micrometric gauge. Small fragments of the retina were then placed in ice-cold fixative for morphologic staining or submerged in ice-cold buffer for the production of retinal homogenates.
Production of retinal homogenates Samples of whole human retina were weighed and homogenized (with the modified Potter-Elvehjem tissue homogenizer) 1/10 w/v into an ice-cold
homogenization buffer (veronal acetate, pH 7.4). Retinal homogenate was centrifuged in a 0.656 M sucrose gradient containing 1 mM ethylenediarninetetraacetic acid. The tubes were centrifuged for 30 minutes at 47 000 x g (21 500 rpm) in the Spinco SW 25.1 rotor (Spinco Metal Products, New York). Six fractions were obtained: 1) supernatant A, 2) membranes, 3) supernatant B, 4) mitochondria, 5) cytosol and 6) pellet. Layers 1, 3 and 5 constituted the liquid phase, and layers 2, 4 and 6, the solid phase.
Protein concentration We determined tissue protein concentrations using bovine serum albumin as standard and Falin phenol as reagent. 12
Light microscopy Retinal samples were immediately prefixed in 2% osmium tetraoxide at pH 7.4 in verona} acetate buffer for 5 minutes at 4 oc. After fixation, the specimens were washed with verona} acetate buffer, dehydrated in a graded ethanol series and embedded in paraffin. Thin sections (about 4 J.llll) were made for morphologic staining with toluidine blue (0.05% for 1 minute) for light microscopic analysis.
Scanning electron microscopy After prefixation, retinal samples were oriented, and the exposed surface was coated with gold-carbon vapour and examined with a JEM-lOOB electron microscope with the EM Asid high-resolution scanning device (University of Pees, Pees, Hungary). ORWO NR 5R NP 20 film (ORWO, Budapest) was used for photographs. Photomicrographs were printed in black and white.
Quantitative analysis of images For detailed evaluation of the effects of aging on retinal morphology, we performed quantitative analysis of images on slides and on photomicrographs. This technique makes it possible to summarize a great number of results obtained from a great number of photomicrographs. We used a Quantimet image analyser (Leica Microsystems Imaging Solutions Ltd., Cambridge, UK) equipped with specific software that permits determination of retinal thickness, the number
CAN JOPHTHALMOL-VOL. 39, NO. I, 2004
63
Age-related retinal changes-Cavallotti et al of ganglion cells/mm2 of retina, the number of capillaries/mm2 of retina, the number of synaptic bodies/area observed, the number of cellular processes/area observed and the number of intercellular connections in the matrix between the photoreceptors. The results obtained by quantitative analysis of images may be incorrect because the main choices (the instructions for the software) are decided by each operator according to personal preference. For this reason, the data are unclear rather than clear, and it is necessary to follow very careful rules. The counts must be repeated at least three times using the doublemasked technique. All the counts should be performed by different observers using different analysers and with samples identified only by a number or a letter. Final results must be evaluated by another worker, who examines experimental protocols to identify each sample and attribute specific values. Final values must be submitted to statistical analysis of data. The values that we report represent the "values" of staining for the two age groups and are expressed in conventional units (CU) (and standard error of the mean). CU are arbitrary units furnished and printed directly by the Quantimet system.
Statistical analysis The statistical methods that we used must be interpreted as an accurate description of the data rather than a statistical inference of the data. We performed the preliminary studies of each value using basic sample statistics. Mean values, maximum and minimum limits, variations, standard deviation, standard error of the mean and correlation coefficients were determined. 13 We performed correlative analysis of the morphologic and biochemical data by comparing the significant differences for each age group with the corresponding values for the other homogeneous groups. Correlation coefficients denote a significance level of p < 0.001 and are not significant when p > 0.05. We also calculated the correlation coefficient (probability). 14 RESULTS
Table 1 shows the age and sex of the subjects and the reason for eye removal. The morphometric and biochemical results are given in Table 2. All differences between the younger and older subjects were significant (p < 0.001) with the exception of mean retinal weight and mean
64
CAN J OPHTHALMOL-VOL. 39, NO. I, 2004
concentration of cytoplasmic (free) and of structural (tissue-associated) protein. All the data regarding retinal thickness, number of capillaries, number of synaptic bodies, number of cellular processes and number of intercellular connections were confirmed by quantitative analysis of images. The mean concentration of cytoplasmic protein was lower in the older subjects than in the younger subjects, whereas the reverse was true for mean concentration of structural protein (Table 2). The SEM microanatomic details of the retina of an 18-year-old subject are shown in Fig. 1. The ganglion cell layer, outer nuclear layer or bipolar cell layer, and photoreceptor cell layer can be seen. Fig. 2 shows the outer and inner segments of the photoreceptors in the retina of a 68-year-old subject. The external limiting membrane, the bipolar cell layer and numerous capillaries can be distinguished. Comparison of the two figures shows a decrease in the number of capillaries and in retinal thickness with increased age. Fig. 3 shows the outer segments of the photoreceptors at higher magnification in the retina of a 21-yearold subject. The cellular matrix is crossed in places by intercellular connections. These structures are likely due to a collapsed interphotoreceptor matrix, which normally fills the subretinal spaces. 15 The outer segments of photoreceptors are composed of a series of discs. Occasionally, intercellular connections may be found between the neighbouring outer segments. The inner and outer segments of the cones are much wider than the inner and outer segments of the rods. Fig. 4 shows the intercellular spaces of the outer layers of the retina from a 19-year-old subject. Numerous cellular processes and synaptic bodies can be seen. A presumed interplexiform cell can also be identified.
Table 1-Age and sex of the subjects and reason for eye removal Subject no. I
2 3
4 5 6 7 8
Age,yr/ sex
Reason for eye removal
22/M 18/M 19/M 21/F 68/F 70/M 66/M 68/M
Eye donor Eye donor Injury Eye donor Injury Injury Eye donor Injury
Age-related retinal changes-Cavallotti et al
Table 2-Morphometric and biochemical features of retinas from the younger and older subjects* Group; mean (and standard error of mean) Feature
Younger subjects (n = 7)
Older subjects (n = 5)
p value
426 (34.2) 413.5 (32.3) 3.6 (1.4)
261 (18.9) 256.2 (26.8) 1.8 ( 1.2)
< 0.001 < 0.001 < 0.001
122.4 (4.9)
38.5 (1.6)
< 0.001
82.3 (3.1)
13.1 ( 1.5)
< 0.001
36.4 (2.5) 3.1 (0.9)
14.3 ( 1.4) 2.5 ( 1.1)
< 0.001 NS
92.1 ( 1.8)
78.7 ( 1.3)
< 0.001
27.6 (1.3)
11.8 (0.8)
NS
64.4 (1.6)
86.9 ( 1.4)
NS
Retinal thickness,t IJm No. of ganglion cells* per mm 2 No. of capillaries per mm 2 No. of synaptic bodies/area observed, CU No. of cellular processes/area observed, CU No. of intercellular connections/ area observed, CU Retinal weight, mg Protein concentration, IJg/mg fresh tissue Cytoplasmic protein concentration, IJg/mg fresh tissue Structural protein concentration, IJg/mg fresh tissue
*Values obtained by means of quantitative analysis of images with the exception of protein concentration. CU =conventional unit, NS =not significant. tWithout the pigment epithelium that is detectable in all remaining layers. :j:Corresponds to mean number of optic nerve fibres.
Fig. 5 depicts a sample from a 19-year-old subject. The bulges of the outer nuclear layer appear tightly attached to one another, with only very narrow bridges of cellular matrix between them. The outer plexiform layers are wider and more complex than the inner
Fig. !-Scanning electron photomicrograph of retina of 18-year-old subject, showing ganglion cell layer (G), bipolar cell layer (B) and photoreceptor cell layer (P) (magnification x 1600). Scale bar = I IJm.
plexiform layers. The ganglion cells appear as large, spherical bodies arranged in a fibrous network. Most of the cell bodies are smooth, but some appear to have
Fig. 2-Scanning electron photomicrograph of outer layers of retina of 68-year-old subject. Outer segments (os) and inner segments (is) of photoreceptor cell, external limiting membrane (elm) and bipolar cell layer can be seen.Arrows = intercellular spaces; N = pigmented epithelium (magnification x8000). Scale bar= I 1Jm.
CAN J OPHTHALMOL-VOL.39, NO. I, 2004
65
Age-related retinal changes-Cavallotti et al
Fig. 3-Scanning electron photomicrograph of retina of 21-year-old subject, showing outer segments of photoreceptor cell. Note cellular matrix between outer segments, intercellular connections (ic) and apex (a) of head of photoreceptors (magnification x 16 000). Scale bar = I IJm.
Fig. 4--Scanning electron photomicrograph of outer nuclear layer of retina of 19-year-old subject. Numerous synaptic bodies (S), a presumed interplexiform cell (I PC) and numerous cellular processes (CP) can be seen. Many capillary vessels (C) are also evident (magnification x9600). Scale bar = I !Jm.
a villose surface. Numerous synaptic bodies, cellular processes and intercellular spaces can be observed. Comparison with images from an older subject (Fig. 2) shows a smaller number of capillaries, markedly fewer synaptic bodies and cellular processes, and a larger number of intercellular spaces in the latter. INTERPRETATION
Our results show that the human retina undergoes specific changes with aging. 16•17 These changes are responsible for all the age-related functional optical
66
CAN J OPHTHALMOL-VOL. 39, NO. I, 2004
Fig. 5-Scanning electron photomicrograph of outer nuclear layer of retina of 19-year-old subject. Numerous synaptic bodies, cellular processes and capillary vessels are evident (magnification x9600). Scale bar = I !Jm.
modifications observed in humans. 18 •19 However, the retina may be altered by early postmortem damage or by common vascular diseases, such as hypertension, arteriosclerosis and diabetic angiopathy. In some circumstances these diseases are associated with capillary occlusion, neural cell damage and modification of glial cells. We avoided both these occurrences by obtaining retinas from living subjects rather than from cadaveric donors and by harvesting samples from precisely the same site (equatorially in the nasal region) in undamaged retinal tissue. We found that retinal thickness decreases significantly with aging. The number of ganglion cells also decreases with increased age. These cells seem to be more vulnerable to age-related loss than other retinal cells. The number of retinal capillaries, intercellular connections, cellular processes and synaptic bodies also decreases markedly with aging. A study by Ono and colleagues20 gives inconclusive data about the thickness of the retinal nerve fibre layer and does not provide good evidence of a significant correlation between modifications to this structure and increased age. However, Funaki and associates 21 reported that retinal nerve fibre layer thickness decreases significantly with aging. Our results are not in keeping with those of Radnot and Follman,6•7 who found a marked decrease with increased age in the number of rods and RPE cells, which also became irregular in size and shape and accumulated massive amounts of lipofuscin. A more recent study showed changes in ganglion cell topography with aging. 22
Age-related retinal changes-Cavallotti et al
Gao and Hollyfield23 reported substantial progressive loss of neurons (photoreceptors and ganglion cells) with aging. They also found a rate of loss of equatorial RPE cells of about 14 cells/mm2 per year from the second to the ninth decade. These findings are not in keeping with those of Panda-Jonas and coworkers,24 who reported controversial RPE cell count, distribution and correlations in normal human eyes. The discrepancy is presumably due to the relatively small number of eyes analysed. Gao and Hollyfield23 also reported substantial loss of equatorial rods and cones in older subjects, whereas the density of cones in the fovea appeared to be more stable. They reported a similar pattern in the loss of equatorial rods and ganglion cells. The findings of Curcio and Allen22 suggest relative variability in the total number of ganglion cells and a larger number of ganglion cells in the superior retina, probably related to the higher density of rods in this quadrant. These authors also described an increase in cone density in the far nasal periphery; the significance of this finding remains controversial. They affirmed that comparison of ganglion cell topography with the visual field representation in area Vl suggests that cortical magnification is proportional to ganglion cell density throughout the visual field. However, these results were obtained from patients with a limited age range (27-37 years). Linberg and Fisher5 reported the existence of two morphologic types of interplexiform cells within the retina, one between the rods or cones and the bipolar cells in the outer plexiform layer, and the other between the bipolar cells and the amacrine cells in the inner nuclear layer. Our results are in keeping with these findings. In fact, our observations confirm that the interplexiform cells undergo age-related changes. Impairment of visual function in older subjects has long been considered the consequence of opacity of the ocular media; little attention has been paid to agerelated changes in the retina. Our findings suggest that the age-related decrease in visual acuity is influenced, at least in part, by age-related changes in retinal tissues. The changes in the patterns of cytoplasmic and structural protein that we observed may be interpreted as confirmation of the morphologic findings; this strengthens our hypothesis concerning the correlation between age-related changes in visual function and modification of retinal function and metabolism. In fact, our finding of a decrease in cytoplasmic protein concentration in older subjects may be correlated to a general rearrangement of the metabolic functions in
the tissues of these subjects. It is possible that this decrease in cytoplasmic proteins with aging may explain the impairment of some functional retinal mechanisms (e.g., visual function, local metabolism of proteins), especially in retinal disorders whose onset is significantly related to increased age. We are greatly indebted to Drs. Alessandro Kovacs and Massimo Bucci for their useful criticisms and suggestions. We gratefully acknowledge the informatics consulting services of Dr. Mauro Cameroni, the secretarial work of Mrs. Silvana Casamento, the photographic assistance of Mr. Giuseppe Leoncini and the assistance of Mrs. Sharon Hobby in revising the English. This study was supported in part by grant 97.05.016 from the University of Rome "La Sapienza," Ricerche di A teneo. REFERENCES
1. Lewis ER, Zeevi YY, Werblin FS. Scanning electron microscopy of vertebrate visual receptors. Brain Res 1969;15:559-62. 2. Hansson HA. Scanning electron microscopic studies on the synaptic bodies in the rat retina. Z Zellforsch Mikrosk Anat 1970;107:45-53. 3. Steinberg RH. Scanning electron microscopy of the bullfrog's retina and pigment epithelium. Z Zellforsch MikroskAnat 1974;143:451-63. 4. Antal M. Scanning electron microscopy of photoreceptors. Ophthalmologica 1977;174:28~. 5. Leuenberger P. Stereo-ultrastructure de la retine. Arch Ophtalmol (Paris) 1971;31:813-22. 6. Radnot M. Scanning electron microscopic study of human retina. Ophthalmologica 1978;176:308-12. 7. Follman P, Radnot M. Some scanning electron microscopic observations on human retina. Klin Oczna 1979; 81:513-4. 8. Tanaka K. Frozen resin cracking method and its role in cytology. In: Hayat MA, editor. Principles and techniques of scanning electron microscopy: biological applications. Vol I. New York: van Nostrand Reinhold Company; 1974. p. 125-34. 9. Cavallotti C, Artico M, Pescosolido N, Feher J. Agerelated changes in rat retina. Jpn J Ophthalmol 2001; 45:68-75. 10. Cavallotti C, Pacella E, Pescosolido N, Tranquilli-Leali FM, Feher J. Age-related changes in the human optic nerve. Can J Ophthalmol2002;37:389-94. 11. Huang JC, Voaden MJ, Zarbin MA, Marshall J. Morphologic preservation and variability of human donor retina. Curr Eye Res 2000;20:231-41. 12. Lowry OH, Rosebrough NJ, Farr AL, Randall J. Protein measurement with the Polin phenol reagent. J Biol Chern 1951;193:265-75.
CAN J OPHTHALMOL-YOL. 39, NO. I, 2004
67
Age-related retinal changes-Cavallotti et al
13. Serio A. Appunti dalle lezioni di statistica sanitaria. Rome: Kappa; 1986. 14. Castino M, Roletto E. Statistica applicata. Padua (Italy): Piccin; 1992. 15. Hageman GS, Johnson LV. Structure, composition and function of the retinal interphotoreceptor matrix. Prog Retin Res 1991;10:207-49. 16. Morrison JD, McGrath R. Assessment of the optical contributions to the age-related deterioration in vision. Q J Exp Physio/1985;70:249-69. 17. Weale RA. Aging and vision. Vision Res 1986;26: 1507-12. 18. Elliott DB. Contrast sensitivity decline with ageing: A neuronal or optical phenomenon? Ophthalmic Physiol Opt 1987;7:415-9. 19. De Laey JJ. Ophtalmologie geriatrique. In: Oosterhosch W, editor. Traite de geriatrie. Brussels: Societe Scientifique de Medecine Generale; 1988. p. 263-488. 20. Ono J, Kimura T, Ishii R, Fujiki K, Kanai A. Relationship between the retinal nerve fiber layer thickness and the
68
CAN J OPHTHALMOL-VOL. 39, NO. I, 2004
21.
22. 23. 24.
25.
effect of aging in normal eyes. Nippon Ganka Gakkai Zasshi 1998;102:752-8. Funaki S, Shirakashi M, Abe H. Relationship between age and thickness of the retinal nerve fiber layer in normal subjects. Nippon Ganka Gakkai Zasshi 1998;102:383-8. Curcio CA, Allen KA. Topography of ganglion cells in human retina. J Camp Neuro/1990;300:5-25. Gao H, Hollyfield JG. Aging of the human retina. Invest Ophthalmol Vis Sci 1992;33:1-17. Panda-Jonas S, Jonas JB, Jakobczyk-Zanija M. Retinal pigment epithelial cell count, distribution, and correlations in normal human eyes. Am J Ophthalmoll996;121: 181-9. Linberg KA, Fisher SK. An ultrastructural study of interplexiform cell synapses in the human retina. J Camp Neuro/1986;243:561-76.
Key words: age, retina, photoreceptors, bipolar cells, ganglion cells, quantitative analysis of images
From the Departments of *Human Anatomy and tOphthalmology, University of Rome "La Sapienza," Rome, Italy, and :j:the Eye Clinic, University of Pees, Pees, Hungary
Cardiovascolari e Respiratorie, Sezione di Anatomia Umana, University of Rome "La Sapienza," Via A. Borelli 50,00161 Rome, Italy; fax 0039/06/4957669, cavallotti @uniroma1.it
Originally received Feb. 6, 2003 Accepted for publication July 12, 2003
This article has been peer-reviewed.
Correspondence to: Dr. Carlo Cavallotti, Dipartimento di Scienze
Can J Ophthalmol 2004;39:61-8
Age-related retinal changes-Cavallotti et al
61
Age-related retinal changes-Cavallotti et al
increases our understanding of this topic. Our results may be adopted as a model or as normal values when studying other changes that may occur in the human retina in pathological conditions.
Contexte : Dans une etude precedente, Ia microscopie a rayons X avait fait voir dans Ia retine des modifications associes au vieillissement chez les rats. Nous avons mene une etude pour evaluer les modifications de Ia retine lies au vieillissement chez les etres humains. Methodes : Des echantillons de tissus retiniens frais, obtenues chez des jeunes gens de 22 ans ou moins et des personnes agees de 66 ans ou plus, ont fait I' objet d'etudes selon les methodes histologiques traditionnelles et par microscopie a rayon X. Huit retines ont ete obtenues de quatre donateurs dont les cornees avaient servi pour une greffe et quatre autres, de personnes qui avaient subi !'enucleation a Ia suite d'une blessure. Les resultats ont ete soumis a une analyse quantitative d'images. La concentration de proteines cytoplasmiques (libres) et structurelles (associees au tissu) dans les homogenats de tissus retiniens a ete etablie au moyen de precedes biochimiques. Resultats : II y a eu une reduction de toutes les caracteristiques etudiees, sauf Ia concentration des proteines structurelles. L'epaisseur moyenne des retines (et erreur type sur Ia moyenne [ETM]) etait de 426 IJm (34,2 1-1m) chez les jeunes sujets et de 261 IJm (18,9 IJm) chez les autres. Le nombre moyen de cellules ganglionnaires (et ETM) etait de 413,5/mm 2 (32,3/mm2) et de 256,2/mm 2 (26,8/mm 2) respectivement; celui des capillaires, de 3,6/mm 2 ( 1,4/mm 2) et de I ,8/mm 2 (I ,2/mm 2) respectivement. Le nombre moyen (en unites conventionnelles par secteur observe) (et ETM) de corps synaptiques chez les deux groupes etait de 122,4 (4,9) et de 38,5 (I ,6) respectivement; celui des processus capillaires, de 82,3 (3, I) et de 13, I (I ,5) respectivement; et celui des liens intercellulaires, de 36,4 (2,5) et de 14,3 (I ,4) respectivement. La concentration moyenne des proteines entieres par milligramme de tissu frais (et ETM) etait de 92, I IJg (I ,8 IJg) chez les jeunes sujets et de 78,7 IJg (I ,3 IJg) chez les aines. Les valeurs correspondantes des proteines cytoplasmiques etaient de 27,6 IJg (I ,3 IJg) et de I I,8 IJg (0,8 IJg), et celles des proteines structurelles, de 64,4 IJg (I ,6 IJg) et de 86,9 IJg (I ,4 IJg). Les ecarts entre les jeunes sujets et les sujets ages etaient significatifs (p < 0,00 I), sauf pour Ia concentration moyenne des proteines cytoplasmiques et structurelles. Interpretation : La retine humaine subit des modifications particulieres en vieillissant. La microscopie a rayons X fournit de !'information morphometrique nouvelle en ce qui a trait aux modifications lies a l'age dans les cellules photoreceptrices, bipolaires et ganglionnaires, augmentant ainsi notre comprehension du sujet. Nos resultats peuvent etre adoptes comme modeles ou comme valeurs normales pour l'etude des autres modifications qui peuvent survenir dans Ia retine humaine dans des conditions pathologiques.
O
cular tissues in both humans and animals have been widely studied by scanning electron microscopy (SEM). 1- 3 The retinal neurons,Z the inner and outer segments of the photoreceptors, 4 and the retinal pigment epithelium (RPE) and its relations with the photoreceptors 3·5 have also been well described. The entire retina in both animals
62
CAN J OPHTHALMOL-VOL. 39, NO. I, 2004
and humans has been analysed by SEM.6-8 SEM has been used primarily for studying surface morphology, enabling great advances in the investigation of various biologic materials. It has also been used in biologic research, for examining the structural organization of tissues and cells. Subsequently, other biologic applications of SEM have been introduced. 8
Age-related retinal changes-Cavallotti et al
Recently, our group studied the SEM features of the rat retina by means of quantitative analysis of images. 9 We carried out a study to determine whether the age-related changes that we observed in the rat retina9 also occur in the human retina. METHODS
We studied 12 human retinas: 8 obtained from four donors whose corneas had been used for transplantation, and 4 obtained from four subjects whose eyes had been enucleated owing to injury. Four of the subjects were aged 22 years or less and were classified as young, and the other four were aged 66 years or more and were classified as old. Subjects with metabolic, tumoral or vascular retinal damage were excluded from the study. In a previous paper we described age-related changes in the human optic nerve obtained from cadaveric donors. 10 In the current study we used only samples of retinas from living humans because postmortem phenomena may bring about early modifications in retinal morphology. 11 Our study was approved by the ethics committees of the hospitals involved, and patients or their relatives gave informed written consent. The investigations were performed according to the guidelines of the Declaration of Helsinki. Retinal thickness is known to vary between the central and peripheral regions and between the temporal and nasal regions. 1 SEM of the rat retina showed that the number of neurons and their morphology varied within each retinallayer. 2 Moreover, wide individual variations have been described in the neuronal layers of the human retina. 4 Consequently, all retinal quantifications must be performed precisely, and the topographic location must be the same for each retina investigated. After removal, the globe was dissected with a razor blade, and samples of intact retinal tissue (located precisely in the same site in each eye, equatorially in the nasal region) were harvested. The entire thickness of the retina was measured with a micrometric gauge. Small fragments of the retina were then placed in ice-cold fixative for morphologic staining or submerged in ice-cold buffer for the production of retinal homogenates.
Production of retinal homogenates Samples of whole human retina were weighed and homogenized (with the modified Potter-Elvehjem tissue homogenizer) 1/10 w/v into an ice-cold
homogenization buffer (veronal acetate, pH 7.4). Retinal homogenate was centrifuged in a 0.656 M sucrose gradient containing 1 mM ethylenediarninetetraacetic acid. The tubes were centrifuged for 30 minutes at 47 000 x g (21 500 rpm) in the Spinco SW 25.1 rotor (Spinco Metal Products, New York). Six fractions were obtained: 1) supernatant A, 2) membranes, 3) supernatant B, 4) mitochondria, 5) cytosol and 6) pellet. Layers 1, 3 and 5 constituted the liquid phase, and layers 2, 4 and 6, the solid phase.
Protein concentration We determined tissue protein concentrations using bovine serum albumin as standard and Falin phenol as reagent. 12
Light microscopy Retinal samples were immediately prefixed in 2% osmium tetraoxide at pH 7.4 in verona} acetate buffer for 5 minutes at 4 oc. After fixation, the specimens were washed with verona} acetate buffer, dehydrated in a graded ethanol series and embedded in paraffin. Thin sections (about 4 J.llll) were made for morphologic staining with toluidine blue (0.05% for 1 minute) for light microscopic analysis.
Scanning electron microscopy After prefixation, retinal samples were oriented, and the exposed surface was coated with gold-carbon vapour and examined with a JEM-lOOB electron microscope with the EM Asid high-resolution scanning device (University of Pees, Pees, Hungary). ORWO NR 5R NP 20 film (ORWO, Budapest) was used for photographs. Photomicrographs were printed in black and white.
Quantitative analysis of images For detailed evaluation of the effects of aging on retinal morphology, we performed quantitative analysis of images on slides and on photomicrographs. This technique makes it possible to summarize a great number of results obtained from a great number of photomicrographs. We used a Quantimet image analyser (Leica Microsystems Imaging Solutions Ltd., Cambridge, UK) equipped with specific software that permits determination of retinal thickness, the number
CAN JOPHTHALMOL-VOL. 39, NO. I, 2004
63
Age-related retinal changes-Cavallotti et al of ganglion cells/mm2 of retina, the number of capillaries/mm2 of retina, the number of synaptic bodies/area observed, the number of cellular processes/area observed and the number of intercellular connections in the matrix between the photoreceptors. The results obtained by quantitative analysis of images may be incorrect because the main choices (the instructions for the software) are decided by each operator according to personal preference. For this reason, the data are unclear rather than clear, and it is necessary to follow very careful rules. The counts must be repeated at least three times using the doublemasked technique. All the counts should be performed by different observers using different analysers and with samples identified only by a number or a letter. Final results must be evaluated by another worker, who examines experimental protocols to identify each sample and attribute specific values. Final values must be submitted to statistical analysis of data. The values that we report represent the "values" of staining for the two age groups and are expressed in conventional units (CU) (and standard error of the mean). CU are arbitrary units furnished and printed directly by the Quantimet system.
Statistical analysis The statistical methods that we used must be interpreted as an accurate description of the data rather than a statistical inference of the data. We performed the preliminary studies of each value using basic sample statistics. Mean values, maximum and minimum limits, variations, standard deviation, standard error of the mean and correlation coefficients were determined. 13 We performed correlative analysis of the morphologic and biochemical data by comparing the significant differences for each age group with the corresponding values for the other homogeneous groups. Correlation coefficients denote a significance level of p < 0.001 and are not significant when p > 0.05. We also calculated the correlation coefficient (probability). 14 RESULTS
Table 1 shows the age and sex of the subjects and the reason for eye removal. The morphometric and biochemical results are given in Table 2. All differences between the younger and older subjects were significant (p < 0.001) with the exception of mean retinal weight and mean
64
CAN J OPHTHALMOL-VOL. 39, NO. I, 2004
concentration of cytoplasmic (free) and of structural (tissue-associated) protein. All the data regarding retinal thickness, number of capillaries, number of synaptic bodies, number of cellular processes and number of intercellular connections were confirmed by quantitative analysis of images. The mean concentration of cytoplasmic protein was lower in the older subjects than in the younger subjects, whereas the reverse was true for mean concentration of structural protein (Table 2). The SEM microanatomic details of the retina of an 18-year-old subject are shown in Fig. 1. The ganglion cell layer, outer nuclear layer or bipolar cell layer, and photoreceptor cell layer can be seen. Fig. 2 shows the outer and inner segments of the photoreceptors in the retina of a 68-year-old subject. The external limiting membrane, the bipolar cell layer and numerous capillaries can be distinguished. Comparison of the two figures shows a decrease in the number of capillaries and in retinal thickness with increased age. Fig. 3 shows the outer segments of the photoreceptors at higher magnification in the retina of a 21-yearold subject. The cellular matrix is crossed in places by intercellular connections. These structures are likely due to a collapsed interphotoreceptor matrix, which normally fills the subretinal spaces. 15 The outer segments of photoreceptors are composed of a series of discs. Occasionally, intercellular connections may be found between the neighbouring outer segments. The inner and outer segments of the cones are much wider than the inner and outer segments of the rods. Fig. 4 shows the intercellular spaces of the outer layers of the retina from a 19-year-old subject. Numerous cellular processes and synaptic bodies can be seen. A presumed interplexiform cell can also be identified.
Table 1-Age and sex of the subjects and reason for eye removal Subject no. I
2 3
4 5 6 7 8
Age,yr/ sex
Reason for eye removal
22/M 18/M 19/M 21/F 68/F 70/M 66/M 68/M
Eye donor Eye donor Injury Eye donor Injury Injury Eye donor Injury
Age-related retinal changes-Cavallotti et al
Table 2-Morphometric and biochemical features of retinas from the younger and older subjects* Group; mean (and standard error of mean) Feature
Younger subjects (n = 7)
Older subjects (n = 5)
p value
426 (34.2) 413.5 (32.3) 3.6 (1.4)
261 (18.9) 256.2 (26.8) 1.8 ( 1.2)
< 0.001 < 0.001 < 0.001
122.4 (4.9)
38.5 (1.6)
< 0.001
82.3 (3.1)
13.1 ( 1.5)
< 0.001
36.4 (2.5) 3.1 (0.9)
14.3 ( 1.4) 2.5 ( 1.1)
< 0.001 NS
92.1 ( 1.8)
78.7 ( 1.3)
< 0.001
27.6 (1.3)
11.8 (0.8)
NS
64.4 (1.6)
86.9 ( 1.4)
NS
Retinal thickness,t IJm No. of ganglion cells* per mm 2 No. of capillaries per mm 2 No. of synaptic bodies/area observed, CU No. of cellular processes/area observed, CU No. of intercellular connections/ area observed, CU Retinal weight, mg Protein concentration, IJg/mg fresh tissue Cytoplasmic protein concentration, IJg/mg fresh tissue Structural protein concentration, IJg/mg fresh tissue
*Values obtained by means of quantitative analysis of images with the exception of protein concentration. CU =conventional unit, NS =not significant. tWithout the pigment epithelium that is detectable in all remaining layers. :j:Corresponds to mean number of optic nerve fibres.
Fig. 5 depicts a sample from a 19-year-old subject. The bulges of the outer nuclear layer appear tightly attached to one another, with only very narrow bridges of cellular matrix between them. The outer plexiform layers are wider and more complex than the inner
Fig. !-Scanning electron photomicrograph of retina of 18-year-old subject, showing ganglion cell layer (G), bipolar cell layer (B) and photoreceptor cell layer (P) (magnification x 1600). Scale bar = I IJm.
plexiform layers. The ganglion cells appear as large, spherical bodies arranged in a fibrous network. Most of the cell bodies are smooth, but some appear to have
Fig. 2-Scanning electron photomicrograph of outer layers of retina of 68-year-old subject. Outer segments (os) and inner segments (is) of photoreceptor cell, external limiting membrane (elm) and bipolar cell layer can be seen.Arrows = intercellular spaces; N = pigmented epithelium (magnification x8000). Scale bar= I 1Jm.
CAN J OPHTHALMOL-VOL.39, NO. I, 2004
65
Age-related retinal changes-Cavallotti et al
Fig. 3-Scanning electron photomicrograph of retina of 21-year-old subject, showing outer segments of photoreceptor cell. Note cellular matrix between outer segments, intercellular connections (ic) and apex (a) of head of photoreceptors (magnification x 16 000). Scale bar = I IJm.
Fig. 4--Scanning electron photomicrograph of outer nuclear layer of retina of 19-year-old subject. Numerous synaptic bodies (S), a presumed interplexiform cell (I PC) and numerous cellular processes (CP) can be seen. Many capillary vessels (C) are also evident (magnification x9600). Scale bar = I !Jm.
a villose surface. Numerous synaptic bodies, cellular processes and intercellular spaces can be observed. Comparison with images from an older subject (Fig. 2) shows a smaller number of capillaries, markedly fewer synaptic bodies and cellular processes, and a larger number of intercellular spaces in the latter. INTERPRETATION
Our results show that the human retina undergoes specific changes with aging. 16•17 These changes are responsible for all the age-related functional optical
66
CAN J OPHTHALMOL-VOL. 39, NO. I, 2004
Fig. 5-Scanning electron photomicrograph of outer nuclear layer of retina of 19-year-old subject. Numerous synaptic bodies, cellular processes and capillary vessels are evident (magnification x9600). Scale bar = I !Jm.
modifications observed in humans. 18 •19 However, the retina may be altered by early postmortem damage or by common vascular diseases, such as hypertension, arteriosclerosis and diabetic angiopathy. In some circumstances these diseases are associated with capillary occlusion, neural cell damage and modification of glial cells. We avoided both these occurrences by obtaining retinas from living subjects rather than from cadaveric donors and by harvesting samples from precisely the same site (equatorially in the nasal region) in undamaged retinal tissue. We found that retinal thickness decreases significantly with aging. The number of ganglion cells also decreases with increased age. These cells seem to be more vulnerable to age-related loss than other retinal cells. The number of retinal capillaries, intercellular connections, cellular processes and synaptic bodies also decreases markedly with aging. A study by Ono and colleagues20 gives inconclusive data about the thickness of the retinal nerve fibre layer and does not provide good evidence of a significant correlation between modifications to this structure and increased age. However, Funaki and associates 21 reported that retinal nerve fibre layer thickness decreases significantly with aging. Our results are not in keeping with those of Radnot and Follman,6•7 who found a marked decrease with increased age in the number of rods and RPE cells, which also became irregular in size and shape and accumulated massive amounts of lipofuscin. A more recent study showed changes in ganglion cell topography with aging. 22
Age-related retinal changes-Cavallotti et al
Gao and Hollyfield23 reported substantial progressive loss of neurons (photoreceptors and ganglion cells) with aging. They also found a rate of loss of equatorial RPE cells of about 14 cells/mm2 per year from the second to the ninth decade. These findings are not in keeping with those of Panda-Jonas and coworkers,24 who reported controversial RPE cell count, distribution and correlations in normal human eyes. The discrepancy is presumably due to the relatively small number of eyes analysed. Gao and Hollyfield23 also reported substantial loss of equatorial rods and cones in older subjects, whereas the density of cones in the fovea appeared to be more stable. They reported a similar pattern in the loss of equatorial rods and ganglion cells. The findings of Curcio and Allen22 suggest relative variability in the total number of ganglion cells and a larger number of ganglion cells in the superior retina, probably related to the higher density of rods in this quadrant. These authors also described an increase in cone density in the far nasal periphery; the significance of this finding remains controversial. They affirmed that comparison of ganglion cell topography with the visual field representation in area Vl suggests that cortical magnification is proportional to ganglion cell density throughout the visual field. However, these results were obtained from patients with a limited age range (27-37 years). Linberg and Fisher5 reported the existence of two morphologic types of interplexiform cells within the retina, one between the rods or cones and the bipolar cells in the outer plexiform layer, and the other between the bipolar cells and the amacrine cells in the inner nuclear layer. Our results are in keeping with these findings. In fact, our observations confirm that the interplexiform cells undergo age-related changes. Impairment of visual function in older subjects has long been considered the consequence of opacity of the ocular media; little attention has been paid to agerelated changes in the retina. Our findings suggest that the age-related decrease in visual acuity is influenced, at least in part, by age-related changes in retinal tissues. The changes in the patterns of cytoplasmic and structural protein that we observed may be interpreted as confirmation of the morphologic findings; this strengthens our hypothesis concerning the correlation between age-related changes in visual function and modification of retinal function and metabolism. In fact, our finding of a decrease in cytoplasmic protein concentration in older subjects may be correlated to a general rearrangement of the metabolic functions in
the tissues of these subjects. It is possible that this decrease in cytoplasmic proteins with aging may explain the impairment of some functional retinal mechanisms (e.g., visual function, local metabolism of proteins), especially in retinal disorders whose onset is significantly related to increased age. We are greatly indebted to Drs. Alessandro Kovacs and Massimo Bucci for their useful criticisms and suggestions. We gratefully acknowledge the informatics consulting services of Dr. Mauro Cameroni, the secretarial work of Mrs. Silvana Casamento, the photographic assistance of Mr. Giuseppe Leoncini and the assistance of Mrs. Sharon Hobby in revising the English. This study was supported in part by grant 97.05.016 from the University of Rome "La Sapienza," Ricerche di A teneo. REFERENCES
1. Lewis ER, Zeevi YY, Werblin FS. Scanning electron microscopy of vertebrate visual receptors. Brain Res 1969;15:559-62. 2. Hansson HA. Scanning electron microscopic studies on the synaptic bodies in the rat retina. Z Zellforsch Mikrosk Anat 1970;107:45-53. 3. Steinberg RH. Scanning electron microscopy of the bullfrog's retina and pigment epithelium. Z Zellforsch MikroskAnat 1974;143:451-63. 4. Antal M. Scanning electron microscopy of photoreceptors. Ophthalmologica 1977;174:28~. 5. Leuenberger P. Stereo-ultrastructure de la retine. Arch Ophtalmol (Paris) 1971;31:813-22. 6. Radnot M. Scanning electron microscopic study of human retina. Ophthalmologica 1978;176:308-12. 7. Follman P, Radnot M. Some scanning electron microscopic observations on human retina. Klin Oczna 1979; 81:513-4. 8. Tanaka K. Frozen resin cracking method and its role in cytology. In: Hayat MA, editor. Principles and techniques of scanning electron microscopy: biological applications. Vol I. New York: van Nostrand Reinhold Company; 1974. p. 125-34. 9. Cavallotti C, Artico M, Pescosolido N, Feher J. Agerelated changes in rat retina. Jpn J Ophthalmol 2001; 45:68-75. 10. Cavallotti C, Pacella E, Pescosolido N, Tranquilli-Leali FM, Feher J. Age-related changes in the human optic nerve. Can J Ophthalmol2002;37:389-94. 11. Huang JC, Voaden MJ, Zarbin MA, Marshall J. Morphologic preservation and variability of human donor retina. Curr Eye Res 2000;20:231-41. 12. Lowry OH, Rosebrough NJ, Farr AL, Randall J. Protein measurement with the Polin phenol reagent. J Biol Chern 1951;193:265-75.
CAN J OPHTHALMOL-YOL. 39, NO. I, 2004
67
Age-related retinal changes-Cavallotti et al
13. Serio A. Appunti dalle lezioni di statistica sanitaria. Rome: Kappa; 1986. 14. Castino M, Roletto E. Statistica applicata. Padua (Italy): Piccin; 1992. 15. Hageman GS, Johnson LV. Structure, composition and function of the retinal interphotoreceptor matrix. Prog Retin Res 1991;10:207-49. 16. Morrison JD, McGrath R. Assessment of the optical contributions to the age-related deterioration in vision. Q J Exp Physio/1985;70:249-69. 17. Weale RA. Aging and vision. Vision Res 1986;26: 1507-12. 18. Elliott DB. Contrast sensitivity decline with ageing: A neuronal or optical phenomenon? Ophthalmic Physiol Opt 1987;7:415-9. 19. De Laey JJ. Ophtalmologie geriatrique. In: Oosterhosch W, editor. Traite de geriatrie. Brussels: Societe Scientifique de Medecine Generale; 1988. p. 263-488. 20. Ono J, Kimura T, Ishii R, Fujiki K, Kanai A. Relationship between the retinal nerve fiber layer thickness and the
68
CAN J OPHTHALMOL-VOL. 39, NO. I, 2004
21.
22. 23. 24.
25.
effect of aging in normal eyes. Nippon Ganka Gakkai Zasshi 1998;102:752-8. Funaki S, Shirakashi M, Abe H. Relationship between age and thickness of the retinal nerve fiber layer in normal subjects. Nippon Ganka Gakkai Zasshi 1998;102:383-8. Curcio CA, Allen KA. Topography of ganglion cells in human retina. J Camp Neuro/1990;300:5-25. Gao H, Hollyfield JG. Aging of the human retina. Invest Ophthalmol Vis Sci 1992;33:1-17. Panda-Jonas S, Jonas JB, Jakobczyk-Zanija M. Retinal pigment epithelial cell count, distribution, and correlations in normal human eyes. Am J Ophthalmoll996;121: 181-9. Linberg KA, Fisher SK. An ultrastructural study of interplexiform cell synapses in the human retina. J Camp Neuro/1986;243:561-76.
Key words: age, retina, photoreceptors, bipolar cells, ganglion cells, quantitative analysis of images