Dendrites of rod bipolar cells sprout in normal aging retina

Abstract

The aging nervous system is known to manifest a variety of degenerative and regressive events. Here we report the unexpected growth of dendrites in the retinas of normal old mice. The dendrites of many rod bipolar cells in aging mice were observed to extend well beyond their normal strata within the outer plexiform layer to innervate the outer nuclear layer where they appeared to form contacts with the spherules of rod photoreceptors. Such dendritic sprouting increased with age and was evident at all retinal eccentricities. These results provide evidence of retinal plasticity associated with normal aging.

An extensive literature has documented a variety of age-related changes in the nervous system at the molecular (1, 2), cellular (3, 4), and behavioral (5–7) levels. Most scientific evidence points to a predominance of degenerative or regressive events with normal aging. For example, in the hippocampus, age-related changes include alterations in the structure of dendritic arbors and spines (8, 9), loss of synapses (10), changes in the functional properties of synapses (11), as well as a pronounced decrease in neurogenesis (12). In the autonomic nervous system, specific groups of neurons appear to be vulnerable to age-related apoptosis (13, 14) and loss of plasticity (15). In the visual system, advancing age brings on a spectrum of events from declining psychophysical responses in peripheral and scotopic vision (16–18) to atrophy and loss of neurons and glia in subcortical and cortical structures (19–22). Whereas some age-related functional changes could be construed as adaptive (11, 23, 24), to our knowledge there is no compelling evidence for neuronal growth in the aging visual system.

Results

Here we document a remarkable degree of dendritic sprouting in the normal aging retina. Dendrites of rod bipolar cells in aged animals extend processes beyond their normal confines of the outer plexiform layer (OPL) into the outer nuclear layer (ONL), normally a fiber- and synapse-free layer (Fig. 1 A and C–E). Such extended processes are not found in normal retinas of young adult animals (Fig. 1B). Shown in Fig. 1 is a montage of a complete retinal section taken through the optic nerve of an aged mouse (28 months) highlighting rod bipolar cells stained with PKC, a selective marker of rod bipolar cells (A), along with higher-magnification images of PKC-labeled tissue of young (B) and aged adult (C–D) animals. For clarity, the length of each process in the montage that extends beyond the OPL into the ONL has been colored white. As can be seen, aberrant processes extend into the ONL throughout the entire extent of the retinal section.

Fig. 1.

Demonstration of aberrant dendritic processes of rod bipolar cells projecting into the ONL of the aged retina. (A) PKC-labeled rod bipolar cells (red) in a photomontage of an entire retinal section taken through the optic nerve of a 28-month-old mouse. For clarity, the bipolar cell dendritic fibers that extend into the ONL were traced in white and overlaid onto the montage. (Scale bar: 200 μm.) (B) An image of PKC-labeled tissue taken at higher magnification from the retina of a young adult animal (3 months) illustrating the normal confinement of the rod bipolar cell dendritic arbors in the OPL. At the left is a narrow strip of the same retinal section labeled with DAPI to identify the retinal lamination (green). INL, inner nuclear layer. (C–E) Examples of high-magnification images from three aged animals showing the aberrant bipolar cell processes extending into the OPL. (Scale bar: 20 μm.)

The degree of dendritic sprouting in rod bipolar cell fibers was quantified by tracing the portion of each fiber that extended into the ONL within a captured field (Fig. 2 A and B). Examples of these fibers traced from animals representing a variety of ages ranging from 1 to 3 years are shown in Fig. 2C. Note that the aberrant processes tend to become longer and more numerous with age. The proportion of aberrant fibers to rod bipolar cells (number of fibers/number of rods) was compared in the young and aged groups. Young animals had very few fibers compared with rod bipolar cells (0.09 ± 0.05; n = 9) whereas the aged group had significantly more (0.63 ± 0.23, n = 15, t = 6.72, P < 0.01). A regression analysis of total normalized fiber length vs. age (Fig. 2D) demonstrated that there is a significant relationship between the age of the animal and the extent of dendritic growth in animals beyond 1 year of age (r = 0.81). No appreciable differences were seen in fiber length when comparing central, middle, and peripheral retinal regions in young or old animals (Fig. 2E).

Fig. 2.

Quantification of sprouting bipolar cell fibers in retinas of aged mice. (A) PKC-labeled images were used to trace sprouting dendritic processes that extended beyond the confines of the OPL into the ONL. (B) The Neurolucida tracing created from the image shown in A was used to quantify aberrant bipolar cell fibers. The border between the OPL and ONL was demarcated with a horizontal line drawn from photoreceptor cell somas in the DAPI-counterstained image. Fibers that extended beyond this line were traced for quantification. (Scale bar: 50 μm.) (C) Examples of fiber tracings from images of retinal sections averaging 325 μm in length from five aged animals ranging from 13 to 32 months old. (Scale bar: 50 μm.) (D) Correlation of normalized total fiber length and age. Length of aberrant fibers increases with age (r = 0.81). (E) Relationship between normalized total fiber length and retinal eccentricity. Sections from 6 young and 10 aged adults were compared at three retinal eccentricities: 100 μm from the optic nerve (central), 100 μm from the periphery (peripheral), and midway between those two positions (middle).

In the same sections analyzed for fiber growth, both the number of rod photoreceptors and the number of rod bipolar cells were quantified. In aged animals displaying dramatic fiber growth, there was no appreciable loss of rod photoreceptors (Fig. 3A). The lack of significant rod loss in these aged animals is consistent with findings in aging animals of this strain (25), suggesting that loss of afferents is not the impetus for dendritic sprouting. However, we cannot rule out the possibility that localized loss of rod photoreceptors, which would not have been revealed by our overall cell counts, could contribute to the observed sprouting of bipolar cell dendrites. Counts of PKC-labeled bipolar cells in young and aged retinas indicated that the incidence of these cells also did not decrease with age (Fig. 3B).

Fig. 3.

Rod photoreceptor and bipolar cell number in young and aged mice. (A) Aging C57BL/6 mice do not show significant rod loss as a function of age, even up to ≈3 years of age (P > 0.05). (B) Similarly, rod bipolar cell number remains constant as a function of age (P > 0.05).

To assess a possible linkage between bipolar cell processes and presynaptic inputs, photoreceptor spherules were labeled with an antibody directed against postsynaptic density protein (PSD)-95, which labels both cone peduncles and rod spherules (26). In young adult retinas, bipolar cell dendrites and labeled presynaptic sites were confined to the OPL (Fig. 4A). By contrast, in the aged retinas, both PKC and PSD-95 labeling can be seen to extend well into the ONL (Fig. 4B). Moreover, the labeled presynaptic sites were closely juxtaposed with sprouted dendritic processes, suggesting the possibility of functional synaptic sites (Fig. 4 C and D). This possibility was further supported by the localization of postsynaptic glutamate receptors within the ONL of aged adult retinas. Labeling of metabotropic glutamate receptor 6 (mGluR6) was confined to the OPL in the retinas of young adults (Fig. 4E) but could be seen along rod bipolar cell fibers extending beyond the limits of the OPL in aged animals (Fig. 4 F–I).

Fig. 4.

Double labeling of rod bipolar cell dendritic processes in the ONL and presynaptic sites of photoreceptors. (A) Section of retina from a young adult mouse labeled with PKC (red) and PSD-95 (green). Both dendritic fibers of rod bipolar cells and presynaptic sites of photoreceptors are confined to the OPL. (B) In aged adult retinas with bipolar cell dendritic processes extending into the ONL (red), presynaptic sites (green) can also be seen in the ONL. Bright green spots in the inner nuclear layer (INL) of A and B are blood vessels (arrowheads). (C and D) Two double-labeled images of retinal sections from aged mice showing rod bipolar cell fibers (red) and presynaptic sites (green) at high magnification. (E) Young adult mouse labeled with PKC (red) and mGluR6 (green). As in A, both dendritic fibers of rod bipolar cells and postsynaptic mGluR6 receptor sites are confined to the OPL. (F) In aged adult retinas with bipolar cell dendritic processes extending into the ONL (red), mGluR6 receptor sites (green) can also be seen in the ONL. (G–I) Three double-labeled images of retinal sections from aged mice showing rod bipolar cell fibers (red) and postsynaptic mGluR6 sites (green) at high magnification. (Scale bars: 20 μm in A, B, E, and F; 10 μm in C, D, and G–I.)

We also assessed the possibility that retinal cell types other than rod bipolar cells might demonstrate dendritic sprouting in the aged retinas. For this purpose, different antibodies were used to stain unique populations of amacrine, bipolar, horizontal, and ganglion cells in the retinas of young (Fig. 5 Left) and aged adult (Fig. 5 Right) mice. Horizontal cells were labeled with anti-calbindin, and different types of amacrine cells were labeled with anti-calretinin, anti-choline acetyltransferase, and anti-parvalbumin. Cone bipolar cells were labeled with anti-recoverin, whereas On cone bipolar cells and rods were stained with antibodies to the G protein G_oα. The labeled processes of these different types of retinal neurons did not appear to differ appreciably in the young and old retinas, although we did not attempt to make detailed reconstructions and quantitative measures of these cell types.

Fig. 5.

Other retinal cell populations do not develop aberrant fiber growth with age. Several antibodies were used to stain unique populations of amacrine, bipolar, horizontal, and ganglion cells in the retinas of young (Left; 3–6 months) and aged adult (Right; >1 year) mice. (A) Horizontal cells were stained with anti-calbindin. (B) Amacrine cells were labeled with anti-calretinin. (C) Cholinergic amacrine cells were labeled with anti-choline acetyltransferase. (D) Amacrine and ganglion cells were labeled with anti-parvalbumin. (E) Cone bipolar cells were labeled with anti-recoverin. (F) On cone bipolar cells and rods were labeled with anti-G_oα (red). Rods were labeled with anti-PKC (green). The absence of single-labeled red fibers demonstrates that aberrant fibers extending into the ONL emanated from rod bipolar cells, not On cone bipolar cells. (Scale bars: 50 μm in A–E; 25 μm in F.)

Discussion

This study has documented an unexpected phenomenon in the aging mouse retina. The dendrites of rod bipolar cells, normally confined to the OPL, were found to extend in old animals into the ONL. Such aberrant processes were noted in animals older than 1 year, and these tended to increase in length and incidence with the age of the animal. In the oldest animals studied, it was not uncommon to find rod bipolar cell dendrites that extended for 40 μm into the ONL. Such elongation of rod bipolar cell dendrites signifies a substantial degree of dendritic growth in the retinas of aged animals. Moreover, within the ONL, a retinal layer that is normally devoid of synapses, the sprouted dendritic processes appeared to make synaptic contacts with photoreceptor spherules. This is suggested by the fact that presynaptic sites of presumed rod spherules were closely juxtaposed with the postsynaptic sites of sprouted rod bipolar cell dendrites. To establish this point definitively, however, will require electron microscopy study of these synaptic sites in old retinas.

Interestingly, sprouting of rod bipolar cell dendrites has been noted in animals with the neural retina detached from the pigmented epithelium (27), in humans with retinitis pigmentosa (28), and in Bassoon (29) and Cacna1f (30) mutant mice, strains with defective or absent ribbon synapses between photoreceptors and second-order neurons that exhibit deficits in synaptic transmission. A study of the RCS rat, whose retina undergoes progressive photoreceptor degeneration, also demonstrated sprouting of rod bipolar cell dendrites (31). These studies on young adult animals with defective retinas also described sprouting of horizontal cell dendrites into the ONL, a phenomenon that we did not observe in old mice. No dendritic sprouting was noted in the old retinas labeled with a variety of cell-specific retinal markers other than PKC (anti-calretinin, anti-choline acetyltransferase, anti-parvalbumin, anti-recoverin, and anti-G_oα). Thus, the age-related sprouting of rod bipolar cell dendrites we observed appears to be a more specific effect than the retinal restructuring observed in mutant animals or those with detached retinas. Nevertheless, we cannot conclude that the dendrites of rod bipolar cells are the only processes that sprout in an age-related manner because many other cell types remain to be examined.

What underlies the growth of rod bipolar cell dendrites in the aging retina is unknown. One possibility is that rod spherules retract their axons with age, in which case the growth of rod bipolar cell dendrites could reflect mechanical tension generated by the retracting axon (32). Another possibility is that this phenomenon reflects impaired synaptic transmission in the old retinas between rods and rod bipolar cells because reduced synaptic efficacy has been observed to induce new neuronal growth and the formation of ectopic synapses in other systems (29). Although the mechanism for the phenomenon reported here remains to be established, the results of the present study provide evidence that the aging nervous system is capable of manifesting a remarkable degree of plasticity. Although plasticity often implies an adaptive or compensatory reaction, this is not always the case, as has been aptly demonstrated by the well established consequences of monocular deprivation in the developing visual cortex (33). In the case of the dendritic growth described here in the aged retinas, it remains to be established whether this phenomenon reflects a compensatory response to the aging process. It would be of interest to address this issue by assessing the functional properties of the bipolar cells in old retinas. Irrespective of the outcome of such studies, the fact that dendrites in old retinas are capable of extensive growth may prove useful in designing future strategies for overcoming the myriad deficits that are known to impact the aging nervous system.

Previous studies dealing with age-related changes in the mammalian visual system have concluded that these are primarily cortical in nature (5, 34–36). Moreover, these studies emphasized the debilitating effects of the age-related changes, in particular the loss of orientation selectivity by cortical neurons, which appears to reflect a decrease in GABA-dependent neural communication (37). By contrast, our study demonstrates that substantial age-related change occurs in the retina, just one synapse removed from photoreceptors cells. A related study now in progress in our laboratory has revealed that the sprouting of rod bipolar cell dendrites documented here in the aged mouse retina also occurs in the retina of old people (K.E., L.C.L., and L.M.C., unpublished work). The unexpected finding that old neurons are capable of significant new growth offers a more optimistic view of the aging process than what has been reported in the literature heretofore.

Materials and Methods

Animal and Tissue Preparation.

Retinas were harvested from young adult (3–6 months) and aged adult (older than 1 year) C57BL/6 mice after they were given a lethal dose of Eutha-6 (0.2 ml of pentobarbital sodium; Western Medical Supply, Arcadia, CA). After enucleation, eyes were hemisected and postfixed in 0.4% paraformaldehyde (Sigma, St. Louis, MO) in PBS (EM Science, Gibbstown, NJ) for 30–120 min. The eye cup was subsequently washed in PBS in preparation for further processing. All experiments were performed in accordance with National Institutes of Health and institutional guidelines regarding animal use. Ophthalmological exams revealed no abnormalities in a representative sample of the aged animals.

Immunohistochemistry.

Hemisected eyes were cryoprotected in 25% sucrose overnight before being embedded in optimal cutting temperature medium (Ted Pella, Inc., Torrence, CA) and cut at a thickness of 10 μm on a cryostat (Leica, Deerfield, IL). Sections chosen for immunohistochemistry and analysis were taken from the center of the eye passing through the optic nerve. All sections were incubated in a blocking solution containing 10% normal donkey serum, 0.2% BSA, and 0.3% Triton X-100 in pH 7.35 phosphate buffer, for 1 h at room temperature. Primary antibodies were diluted in fresh blocking solution, and the sections were then incubated overnight at 4°C. After washing in PBS, the sections were incubated for 1 h at room temperature in secondary antibodies diluted in PBS. Finally, sections were counterstained with DAPI (1:500; Kirkegaard & Perry Laboratories, Gaithersburg, MD) and mounted with VECTASHIELD mounting medium (Vector Laboratories, Burlingame, CA). The following primary antibodies were used: a rabbit polyclonal and a mouse monoclonal anti-PKC (1:1,000; Upstate Biotechnology, Charlottesville, VA); a mouse monoclonal anti-PSD-95 (1:200; Affinity Bioreagents, Golden, CO); a mouse monoclonal anti-mGluR6 (1:1,000; kind gift of Catherine Morgans, Oregon Health & Science University, Portland, OR); a rabbit anti-calretinin (1:1,000; Swant, Bellinzona, Switzerland); a rabbit anti-parvalbumin (1:1,000; Swant); a rabbit anti-recoverin (1:1,000; kind gift of Alexander Dizhoor, Pennsylvania College of Optometry, Elkins Park, PA); a rabbit anti-calbindin (1:500; Swant); a goat anti-choline acetyltransferase (1:50; Chemicon International, Temecula, CA); and a rabbit anti-G_oα (1:500; Chemicon International). The appropriate fluorescent secondary antibodies (all diluted to 1:500) were conjugated to Cy3, FITC (both from Jackson ImmunoResearch), Alexa Fluor 488, or Alexa Fluor 594 (both from Molecular Probes, Eugene, OR).

Fluorescent Imaging and Analysis.

Fluorescent images were acquired on an FV500 series confocal microscope (Olympus, Tokyo, Japan) equipped with a 405 diode, multiline argon and krypton lasers. Image stacks were collected through the z-axis with either a ×20 or ×40 oil-immersion objective and a pixel resolution of 1024 × 1024. z-steps were made at 0.62 and 0.311 μm, respectively. For enhanced clarity, DAPI was pseudo-colored green with Olympus Fluoview software. Laser intensity and photomultiplier tube levels were adjusted by using Fluoview during capture to optimize elongating dendrites (which sometimes caused the more heavily labeled somas to be overexposed). The contrast and brightness of whole images were adjusted by using Photoshop (Adobe Systems, Mountain View, CA), and the montage image was constructed in Illustrator (Adobe Systems). The two high-magnification images of PKC/PSD-95 labeling (Fig. 4 C and D) were captured on an Olympus BX61 upright microscope equipped with a CCD camera (Hamamatsu Photonics, Bridgewater, NJ) and a ×60 oil-immersion objective by using SlideBook software (Intelligent Imaging Innovations, Denver, CO). For analysis, the fluorescent images were imported into Neurolucida (Microbrightfield, Colchester, VT) where the PKC fibers were traced. Only the portions of the fibers that extended beyond the OPL were traced, and the border between the OPL and ONL was determined by drawing a line along the inner edges of the photoreceptor somas in the DAPI-stained images. From the tracings, aberrant fiber length and density were calculated in Neurolucida Explorer (Microbrightfield). Normalized total fiber length was used as a composite measure of fiber length and density. This measure reflects the added lengths of all fibers in an image normalized by the length of the field imaged. Rod photoreceptor counts were obtained by labeling all photoreceptors in a section with DAPI, labeling all cones in a section with the lectin peanut agglutinin conjugated to an Alexa Fluor 594 chromafluor, and subtracting the cone count from the total photoreceptor count.

Statistical Analysis.

Values are presented as mean ± SEM. Student’s t tests were performed on data in figures to compare mean values. Significant difference between groups was set at P < 0.05.

Abbreviations

mGluR6

metabotropic glutamate receptor 6

ONL

outer nuclear layer

OPL

outer plexiform layer

PSD

postsynaptic density protein.