BDNF Preserves the Dendritic Morphology of α and β Ganglion Cells in the Cat Retina after Optic Nerve Injury

Abstract

purpose. To examine whether brain-derived neurotrophic factor (BDNF), a potent neuroprotectant in the mammalian retina, also plays a role in preserving the dendritic integrity of the surviving ganglion cells after optic nerve injury.

methods. Single ganglion cells from cats that underwent unilateral optic nerve crush and received no treatment or nerve crush combined with intravitreous treatment of the affected eye with BDNF were labeled intracellularly, reconstructed using confocal microscopy, and compared quantitatively.

results. Optic nerve injury produced a significant decrease in the soma, dendritic field size, and dendritic complexity of α cells. β Cells also displayed a significant decrease in soma size, but their dendritic fields were not affected as severely as those of α cells. Intravitreous treatment of the eye with BDNF at the time of injury preserved the normal somal and dendritic morphologies of both α and β cells.

conclusions. BDNF, in addition to promoting ganglion cell survival, plays an important role in preserving the somal and dendritic morphologies of the surviving ganglion cells, a necessary precursor to maintaining normal visual function. Ganglion cells, however, are not created equal with respect to their responses to nerve injury or to treatment of the eye with BDNF. Thus, development of effective treatment strategies for preserving ganglion cell function in optic nerve–related diseases mandates a clearer understanding of the cellular response characteristics of the specific neurons involved.

Retinal ganglion cells play an important role in visual processing; they represent the final stage of signal integration within the eye, and their axons form the optic nerve through which visual information leaves the eye. Optic neuropathies such as glaucoma are characterized not only by degenerative changes to the optic nerve, but also the retrograde loss of ganglion cells within the retina itself. ¹ This loss is thought to reflect, in part, a reduction in the level of target-derived trophic material available to the affected neurons. ^{1 2 3 4 5} Numerous studies, including our own, have identified the neuroprotective effect that direct application of brain-derived neurotrophic factor (BDNF) has on ganglion cells in the mammalian retina after optic nerve injury. ^{6 7 8 9 10 11 12 13} What these studies have not included, however, is a detailed analysis of whether the surviving ganglion cells retain their normal structural integrity—a critical prelude to retaining normal function, since ganglion cells receive all their synaptic input via their dendritic arbors. ^{14 15}  

In the present study, we used an optic nerve crush model to examine the somata and dendritic field architecture of single ganglion cells in the feline retina. In particular, we compared the morphologies of α and β cells, the two major classes of ganglion cells in this species. α Cells comprise approximately 5% of the ganglion cells in the feline retina. They have the largest somata, and they contain large, radially oriented dendritic arbors that often arise from four to six thick primary dendrites. β Cells represent approximately 55% of the ganglion cells in the feline retina. These neurons have medium-sized somata and smaller, more compact dendritic trees that often arise from a single primary dendrite. ¹⁶ Functionally, α and β cells have been shown to be the anatomic correlates of the physiologically identified Y and X cells, which are considered to be involved with temporal and spatial processing, respectively. ¹⁷ Our results indicate that treatment of the eye with BDNF after optic nerve injury not only serves to prevent ganglion cell loss, but also helps to maintain dendritic integrity, an important intermediate step in our ongoing studies of the development of neuroprotection strategies to preserve ganglion cell function. 

Materials and Methods

A total of 263 ganglion cells were analyzed. The distribution of these neurons by cell class and eye treatment is provided in Table 1 . All experiments were approved by the Institutional Animal Care and Use Committee at Michigan State University, and all were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Nerve crush was achieved as described previously. ¹³ In brief, anesthesia was initiated in a Plexiglas chamber with 4% isoflurane (IsoFlo; Abbott Laboratories, Abbott, IL) in pure oxygen, delivered at 3 L/min. Each animal then was intubated, and anesthesia was maintained with a 2.5% to 3.0% isoflurane-oxygen mixture (1 L/min). Analgesia and sedation consisted of an intramuscular injection of glycopyrrolate (0.05 mg/kg; Fort Dodge Animal Health, Fort Dodge, IA), and SC injections of butorphanol tartrate (0.2 mg/kg, Torbugesic; Butler, Columbus, OH) and acepromazine (0.04 mg/kg; Butler). Heart and respiratory rates were monitored every 15 minutes. Body temperature was maintained at 37°C by using a thermostatically regulated heating pad. The head was stabilized with a vacuum-activated, beanbag-like, restraining device (Olympic Vac; Olympic Medical, Seattle, WA), and the eyes were treated with the topical anesthetic 0.5% proparacaine HCl (Alcaine; Alcon Laboratories, Ft. Worth, TX) and 2.5% hydroxypropyl methylcellulose (Gonak; Alcorn, Inc., Buffalo Grove, IL) to prevent corneal drying. 

Using sterile procedures, we removed the bone overlying the left frontal sinus and exposed the roof of the bony orbit. All openings to the nasal passages were then sealed with bone wax, to avoid disturbing the cat’s olfactory senses, which can result in a severe loss of appetite. A fine-tipped scalpel blade was used to make an opening in the dorsal surface of the orbit. Careful blunt dissection of the overlying tissues exposed the optic nerve without disturbing the nerve sheath or retinal artery. The optic nerve was stabilized with a small surgical hook, and a smooth-faced bulldog clamp that exerts about 1024 g of force was placed on the nerve for 15 seconds at a distance 2 to 3 mm behind the globe. The bone wax plugs then were checked, the frontal sinus packed with a sponge (Gelfoam; Upjohn, Kalamazoo, MI) moistened with sterile saline, the overlying skin sutured, and the eyes treated with sterile ophthalmic ointment. After removal of the intubation tube, each animal was monitored until it was able to move about freely and feed. Postoperative pain medication (0.2 mg/kg; Torbugesic, Fort Dodge Animal Health) was provided as needed. 

Single, intravitreous, injections of either 30 or 90 μL (1 μg/μL) of sterile recombinant BDNF (Regeneron Pharmaceuticals, Tarrytown, NY) were made into nine eyes (four at 30 μL, five at 90 μL) immediately after nerve crush. These two doses were selected based on our previous observation that, although the 30-μg dose provides slightly better overall cell survival, the 90-μg dose appears to be more effective in preserving large ganglion cells. An additional nine eyes underwent optic nerve crush and received no treatment. In all cases, the nonsurgical fellow eye served as the normal control, along with four animals that underwent no procedures at all. The injections were made through the opening in the frontal sinus at a point ∼5 mm posterior to the ora serrata. All injections were made with a syringe (Hamilton, Reno, NV) with a 30-gauge needle. Care was taken to ensure that the complete bevel of the needle was within the vitreous chamber, but that it did not hit the lens. ¹⁸ Intraocular injections were made over a 1-minute period, with the needle left in place for an additional 30 seconds to allow for diffusion of the drug away from the injection site. Backflow from the injection site was minimal due to the gelatinous vitreous. 

All animals received a 7-day survival period. This period was selected for two reasons. First, it matches that used in our previous BDNF-based neuroprotection studies, ¹³ and second, using our optic nerve injury model, it represents a point in the untreated retina at which a significant level of ganglion cell degeneration has occurred (50% loss), yet there remains sufficient dynamic range among the surviving neurons to evaluate both positive and negative aspects of the drug response. Although the focus of this study was to determine whether the surviving ganglion cells from our previous work had retained their normal dendritic morphologies, ongoing studies are focused on the ability of BDNF to sustain ganglion cell survival, integrity, and function over increasingly longer survival periods. 

After the survival period, the animals were anesthetized deeply with ketamine HCl (15 mg/kg IM), followed by an intravenous injection of pentobarbital sodium (35 mg/kg). The eyes then were removed quickly, the animals received an overdose of pentobarbital sodium, and it was perfused transcardially with 0.5 L of 0.9% saline, followed by a 2-L mixture of 1.5% paraformaldehyde and 2.0% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). The brains were removed and postfixed for future histologic examination of the visual thalamus. Immediately on enucleation, the anterior segment of each eye (from the ora serrata forward) was removed with a pair of fine scissors and the posterior eyecup placed in a solution of artificial cerebrospinal fluid ¹⁹ (aCSF, pH 7.4) saturated with a mixture of 95% O₂ and 5% CO₂ at room temperature. The vitreous body was removed and the posterior eyecup flattened and placed ganglion cell layer up in a Plexiglas chamber also perfused with oxygenated aCSF (6–8 mL/min). The tissue was held submersed in the chamber by a small nylon net, and the chamber was mounted on the stage of an upright microscope equipped with epifluorescence. A neutral-density filter (ND4) was used to reduce the intensity of the mercury vapor light reaching the tissue. Single ganglion cells were viewed using a 40× water immersion objective (NA 0.55; Nikon, Tokyo, Japan) with a working distance of 1.6 mm. Periodically, a few drops of a 1-mM solution of the vital dye acridine orange (cat. no. A-4921; Sigma-Aldrich, St. Louis, MO,) was added to the tissue chamber to aid neuron visualization. ^{20 21 22 23}  

Intracellular injections of single retinal ganglion cells were made using glass microelectrodes and a hydraulic micromanipulator attached to the microscope stage. Glass micropipettes were pulled on a Brown-Flaming micropipette puller (model P-87; Sutter Instruments, Novato, CA) and filled with a solution of 3% Lucifer Yellow CH (cat. no. L-0259; Sigma-Aldrich) in 0.1 M LiCl (pH 7.6). Each cell was impaled by slowly advancing the electrode along its axis until the tip of the electrode was seen to penetrate the cell’s membrane. Complete filling of the soma, dendritic tree, and from 1 to 2 mm of the cell’s intraretinal axon segment was achieved by passing a negative current (1–5 nA) through the electrode. Typically, 1 to 2 minutes was sufficient to label completely the dendritic trees of both the α and β cells. The progress of each intracellular injection was monitored visually, with care taken to minimize the amount of time that individual cells were exposed to the mercury vapor light. After the last cell was injected, the retina was removed from the injection chamber and immersion fixed in 4.0% paraformaldehyde in 0.1 M sodium phosphate buffer (pH 7.4). Each retina then was rinsed with 0.1 M sodium phosphate buffer, wholemounted on a gelatinized glass slide, dehydrated, defatted, and coverslipped (DePex; BDH Laboratory Supplies, Poole, UK). 

We restricted our intracellular analysis to ganglion cells located in the superior midtemporal retina approximately 2 to 7 mm from the area centralis for several reasons (Fig. 1) . First, it was important to select a region away from the central retina, where the somata and dendritic field sizes of ganglion cells undergo significant changes, with small changes in retinal eccentricity. ^{16 19 20 21 24} Second, the intracellular technique does not allow one to inject a large number of cells at all retinal locations, and it was necessary to compare cells matched for retinal eccentricity. And third, this retinal region represents the sample area used in our previous BDNF-related work. ¹³  

The position of each labeled ganglion cell relative to the location of the optic disc and area centralis was recorded using a microscope-based digitizing system (AccuStage, Shoreview, MN). This retinal map then served as a guide for the reconstruction of each labeled ganglion cell using a confocal microscope (model LSM5; Carl Zeiss Meditec, Inc., Dublin, CA). Measurements of cell soma and dendritic field area, length, surface area, and complexity, as determined by Sholl analysis (resolution, 10 μm) were performed by importing the confocal images into an image-analysis system (Neurolucida; MicroBrightfield, Inc., Colchester, VT). 

Measurements for each experimental group were pooled, and all data comparisons are presented as the mean ± 1 SE. Mean values for each parameter across the different experimental and control groups were evaluated by using nonparametric ANOVA (Kruskal-Wallis with the Dunn multiple comparison post hoc test). All statistical analyses were performed with commercial software (InstaStat; GraphPad, San Diego, CA), with P = 0.05 as the level of significance. Since the quantitative measurements for the 30- and 90-μg–treated eyes did not differ significantly, these data were combined to represent a single, BDNF-treated, group of animals. 

Results

The photomicrographs in Figure 2compare, qualitatively, the soma morphologies of α and β cells from retinas under each of the different conditions studied. Normal ganglion cells of both classes displayed smooth, circular cell bodies with centrally located nuclei and a uniformly labeled cytoplasm. By contrast, the somata of ganglion cells from eyes receiving an optic nerve crush and no treatment typically contained numerous crenulations, the nuclei were displaced from the center of the cell, and the label within the cytoplasm was often nonuniform, because of the presence of numerous vacuoles, signs common to neurons undergoing degeneration. ²⁵ Treatment of the eyes with 30 μg of BDNF at the time of the nerve injury enhanced the qualitative appearance of the somata of both α and β cells; however, the β cells appeared to show greater improvement. Increasing the treatment dose to 90 μg resulted in α- and β-cell somata that were qualitatively indistinguishable from normal (Fig. 2) . These changes were confirmed quantitatively, as shown in Table 2and the histograms of Figures 3A and 3B . The somata of α cells from the retinas of animals receiving a nerve crush and no treatment had mean cross-sectional areas that were only 65% of normal. Treatment of the eye with BDNF resulted in a significant preservation of soma size (83% of normal). Much of this effect appeared to come from eyes receiving the higher dose of BDNF, which showed only a 9% decrease in α-cell soma size compared with a 28% decrease for those eyes treated with 30 μg of BDNF. β-Cell somata were not affected to the same extent, retaining approximately 77% of their size after the nerve crush alone. They also showed a stronger response, retaining 95% of their cross-sectional area after treatment with BDNF, with no difference in their response to either dose of the drug. 

Qualitatively, the dendritic arbors of the α and β cells from the experimental eyes were similar to those from normal eyes; α-cell arbors were large and branched openly, whereas the β-cell arbors were smaller, more compact, and displayed a more complex branching pattern (Fig. 4) . After optic nerve injury and no treatment, the dendrites of both the α and β cells displayed greater variability with respect to diameter along their lengths, especially near branch points, and less uniformity in the density of label within individual processes. In particular, untreated cells often showed very dense labeling of the soma and proximal dendrites, whereas the distal portions of the dendritic trees were very pale. These features were much less common in cells from eyes treated with BDNF, the exception being that β cells from eyes treated with 30 μg BDNF were, on average, qualitatively more normal in appearance than those from eyes receiving the 90 μg BDNF treatment. As with soma size, these qualitative observations were supported by the quantitative analyses. For α cells, nontreatment of the injured eye resulted in a significant decrease in mean dendritic field size to only 64% of normal. 

Similar reductions also were seen with respect to the total amount of dendritic material associated with these neurons, as depicted by significant decreases in dendritic length and complexity (Figs. 5A 5C 5E ; Table 2 ). The combined result of these changes was a highly significant decrease in dendritic surface area (Fig. 5G ; Table 2 ). Treatment of the eye with BDNF resulted not only in a significant preservation of α-cell dendritic arbor cross-sectional area, but also the total dendritic length, complexity, and surface areas of these neurons (96%–98% of normal). Optic nerve injury also had an adverse affect on the dendritic field size and complexity of β cells, but to a much lesser degree than that measured for α cells; on average, the different components of the β-cell dendritic trees were reduced to 84% of normal, compared with 64% of normal for the α cells. As with the α cells, treatment of the eye with BDNF at the time of the injury resulted in a level of β-cell dendritic morphology that was comparable to normal (Figs. 5B 5D 5F 5H ; Table 2 ). 

Discussion

Previously, we demonstrated that glaucoma-related optic nerve injury results in a pattern of retinal ganglion cell degeneration that begins with the dendritic tree. ²³ More recently, we reported that these changes in dendritic morphology contribute to abnormal visual responses within the glaucomatous eye. ²² Having shown BDNF to be a potent neuroprotectant in primate-sized eyes, ¹³ the focus of the current study was to determine whether the surviving neurons retain their normal dendritic morphology, a critical step toward maintaining ganglion cell function after optic nerve injury. 

To examine the effect that BDNF has on ganglion cell structure after optic nerve injury, we applied intracellular staining and reconstruction techniques to ganglion cells from feline eyes subjected to an optic nerve crush and no treatment as well as those treated with BDNF at the time of the nerve injury. The results indicate that, in addition to promoting ganglion cell survival, BDNF also is effective in maintaining the normal dendritic integrity of surviving ganglion cells, and ongoing studies suggest that this protection results in enhanced preservation of ganglion cell function (Weber AJ, et al. IOVS 2006;47:ARVO E-Abstract 1569). 

When considering these results, it is important to keep in mind several factors that could influence the data. These include ganglion cell location, identification, and classification, as well as the experimental paradigm applied. Because ganglion cell soma and dendritic field sizes vary with retinal eccentricity, it is possible that the morphologic differences we describe simply reflect sampling biases related to the retinal locations of the neurons studied. This bias seems unlikely, for two reasons. First, we restricted our intracellular injections to a specific region of the midtemporal retina, where the somata and dendritic field sizes of α and β cells are relatively constant with respect to changes in retinal eccentricity. ^{16 26} Second, as shown in Figure 1 , comparison of the spatial distributions of the ganglion cells studied did not indicate any systematic sampling bias with respect to retinal location across the different animals studied. 

With respect to ganglion cell identification and classification, these are relatively straightforward processes when comparing intracellularly labeled neurons, since the presence of an axon identifies the injected cell as a ganglion cell, versus displaced amacrine, and visualization of the soma and dendritic arbor permits unambiguous classification. Although the feline retina contains several morphologic classes of ganglion cells, ^{16 19 24 27 28 29 30 31 32 33 34} it is not difficult, particularly in the temporal retina where our injections were made, to target α and β cells with a high degree of specificity. In the acridine-stained retina the large cells are readily identified as α cells, medium-sized cells are predominantly β cells, and the remaining cells with small somata are mainly displaced amacrine or non-α/β cells. ^{27 28 30 34} In this study, we chose to focus our analysis on α and β cells because they provide the primary input to the visual thalamus, ^{26 34 35} and their morphologies and functional roles are well documented. ^{16 36 37} We were able to identify and thus include only α and β cells because, as in other studies, the basic morphologic features needed to characterize the different types of ganglion cells remained evident, even for degenerating neurons. This ease of identification is primarily because the initial structural abnormalities associated with ganglion cell degeneration after optic nerve injury involve the distal, rather than proximal, regions of the dendritic arbor. ^{23 25 38}  

While ganglion cell death after optic nerve injury is thought to reflect primarily a decrease in target-derived trophic support, ^{4 39 40} other contributing factors, such as the changes in ganglion cell morphology described by us and others, ^{23 25 41} most likely also play a role; it is reasonable to assume that the dendritic atrophy that characterizes the initial stage of ganglion cell degeneration results in conformational changes to synaptic contact sites, ion channels, and trophic factor receptors which, in turn, affect ganglion cell activity and various cellular functions. These include intracellular levels of trophic factors and cAMP, translocation of receptors from the cytoplasm to the plasma membrane, maintenance of normal ratios of various receptor isoforms, balanced activation of specific cell survival/death signaling pathways, and maintenance of the cytoskeleton. ^{25 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56} That disruption of the microtubule array that comprises the bulk of the dendritic cytoskeleton contributes to the degenerative process is suggested by our observation that many of the ganglion cells from the nontreated eyes showed strong levels of label in the soma and proximal dendrites, compared with the distal processes. Although it could be argued that the reductions in dendritic field size simply reflect a general inability to label the distal processes of degenerating neurons, the reality is that without proper cytoskeletal support, these distal regions are dysfunctional and are not sustainable. They are synonymous with the arbor of a dying tree, where although the distal branches remain, they contain no foliage and thus are of no functional use to the tree. 

The ability of BDNF to preserve the dendritic integrity of both α and β cells after optic nerve injury is not surprising, given its role in shaping the dendritic morphologies of central nervous system (CNS) neurons, including retinal ganglion cells, during development. ^{57 58} In addition, it is well-known that ganglion cells in the normal adult retina express BDNF and its primary receptor, TrkB, ^{59 60} and that both are upregulated as a result of injury to the optic nerve. ^{42 61} Finally, numerous studies have demonstrated that intravitreous delivery of BDNF or viral-mediated enhancement of ganglion cell TrkB receptor levels can delay ganglion cell death after optic nerve injury. ^{7 8 9 11 13 48 49 51 62} The mechanism by which BDNF acts to counter the degenerative events most likely involves not only its ability to enhance neuronal activity, ⁴⁶ but also its ability to activate the phosphatidylinositol-3-kinase-protein kinase B (PI3K-Akt) and mitogen-activated protein kinase (MAPK) survival pathways via its interaction with full-length TrkB receptors. ^{48 49 51} Although retinal ganglion cells also possess truncated TrkB receptor isoforms, their role is less clear, since they lack the intracellular signaling domain characteristic of the full-length receptors. Typically, these receptors are not considered to be involved in signal transduction, but rather act to modulate the function of the full-length receptors. ⁶³ Nevertheless, more recent work has shown that truncated TrkB receptors are capable of mediating ligand-induced changes in cellular physiology, although the mechanism and purpose remain undefined. ⁶⁴ While the present study indicates that activation of these intracellular pathways at the time of the nerve injury prevents the degenerative events from developing, what remains to be determined is the duration of this protective influence, as well as the ability to reverse previously established degenerative changes. 

Finally, it is of interest to note that the data presented herein, in agreement with results of previous studies, ^{38 41 62 65} suggest that α and β cells differ with respect to their responsiveness to optic nerve injury, as well as treatment with BDNF (Table 2 ; Figs. 2 4 5 ). In response to the nerve injury, α cells displayed highly significant changes with respect to all the parameters measured, whereas β cells showed only a significant reduction in mean soma and dendritic field sizes (Table 2 , Fig. 5 ). Although treatment with BDNF enhanced preservation of the morphologic features of both α and β cells, the latter appeared to respond best to a lower dose of drug, whereas the former responded best to the higher dose (Figs. 2 4) . A more detailed analysis is needed, but it is reasonable to assume that such differential responses by α and β cells reflect differences in the type, number, ratios, and turnover rates, of the different trophic factor receptors each possesses, ⁶⁶ as well as potential variations in the signal transduction pathways activated. 

In summary, the data demonstrate the capacity of BDNF not only to prevent ganglion cell death after optic nerve injury, but also its ability to preserve the fine dendritic architecture of the surviving neurons. Preliminary investigations using noninvasive electrophysiological techniques (data not shown) suggest an increased retention of ganglion cell function as well and support the previous receptive field studies of Miyoshi et al. ⁶⁷ after optic nerve section and peripheral nerve transplantation. Ongoing studies will investigate the molecular basis for the different response characteristics of α and β cells to nerve injury and BDNF treatment. Only by understanding the response characteristic of the individual neurons involved will one be able to design effective, long-term neuroprotective treatment strategies for preventing blindness. 

 Supported by National Eye Institute Grant EY11159 (AJW).

 Submitted for publication October 15, 2007; revised January 17, 2008; accepted March 31, 2008.

 Disclosure: A.J. Weber, None; C.D. Harman, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Summary of Cells Studied

Condition	Total Cells	α Cells	β Cells
Normal	69	22	47
Crush-NT	88	60	28
Crush+30 μg BDNF	57	23	34
Crush+90 μg BDNF	49	26	23

 NT, no treatment.

Figure 1.

 

Spatial distribution of the α and β cells labeled intracellularly under each experimental condition, showing significant overlap among all groups of cells sampled.

Figure 2.

 

Comparison of the somal morphologies of α and β cells after intracellular injection, under each experimental condition studied.

Table 2.

 

Summary of Morphological Features

	Soma Area (μm2)	DF Area (μm2)	DF Length (μm)	Sholl Intersections	DF Surface Area (μm2)
α Cells
Normal
Mean	896.7	225,600.7	8,907.7	736.8	52,726.5
SE	55.3	20,815.1	539.9	47.4	5,961.1
N	22.0	22.0	22.0	22.0	22.0
Crush Only
Mean	588.4	145,306.6	6,033.1	483.9	29,908.6
SE	26.3	7,711.2	327.7	26.7	2,075.3
n	60.0	60.0	60.0	60.0	60.0
% of normal	65.6	64.4	67.7	65.7	56.7
BDNF
Mean	741.3	216,864.5	8,744.2	723.3	51,998.9
SE	33.9	11,882.3	319.9	27.0	2,822.2
n	49.0	49.0	49.0	49.0	49.0
% of normal	82.7	96.1	98.2	98.2	98.6
β-Cells
Normal
Mean	383.4	21,229.1	2,552.4	195.3	7,222.8
SE	14.4	1,176.5	94.1	7.5	388.9
n	47.0	47.0	47.0	47.0	47.0
Crush Only
Mean	294.4	18,017.9	2,154.4	164.6	6,214.7
SE	17.7	1,009.7	116.7	8.7	401.9
n	28.0	28.0	28.0	28.0	28.0
% of normal	76.8	84.9	84.4	84.3	86.0
BDNF
Mean	365.2	19,096.6	2,361.7	182.3	7,155.7
SE	17.5	1,184.4	107.4	8.8	429.7
n	57.0	57.0	57.0	57.0	57.0
% of normal	95.3	90.0	92.5	93.3	99.1

Figure 3.

 

Histograms comparing the soma sizes for α and β cells from normal feline eyes, eyes receiving an optic nerve injury and no treatment, and those receiving an optic nerve injury and treatment with BDNF at the time of the injury. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 4.

 

Comparison of the dendritic morphologies of α and β cells labeled intracellularly under each experimental condition.

Figure 5.

 

Histograms comparing the dendritic field properties of α and β cells under each condition studied. NT, no treatment. **P < 0.01; ***P < 0.001.