Effects of Brain-Derived Neurotrophic Factor and Neurotrophin-4 on Isolated Cultured Retinal Ganglion Cells: Evaluation by Flow Cytometry

Abstract

purpose. Effects of brain-derived neurotrophic factor (BDNF) and neurotrophin (NT)-4 on retinal ganglion cells (RGCs) isolated and cultured in a serum-free medium are evaluated objectively by using flow cytometry.

methods. RGCs from the retinas of 2-day-old rats were isolated in a two-step panning and cultured in a serum-free medium. BDNF (1, 10, and 100 pg/ml or 1, 10, and 100 ng/ml), NT-4 (0.1, 1, 10, and 100 ng/ml) or their vehicle, phosphate-buffered saline, were individually added to aliquots of the medium to be cultured for 48 hours. Then, after adding 5-chloromethylfluorescein diacetate, the survival of RGCs was evaluated using flow cytometry.

results. The method used allowed the authors to analyze 10,000 RGCs per sample in approximately 2 minutes, so that a much larger number of cells was evaluated in a shorter period than with previously reported methods. RGCs were classified into either large or small RGCs, and the survival of each of these groups was determined objectively by the amount of fluorescent emission. BDNF improved the survival rate of RGCs concentration-dependently. In particular, the survival rate of small RGCs was greatly improved. BDNF at 100 ng/ml increased the survival rate of small RGCs by 17.4% and that of large RGCs by 7.8% in comparison to the controls. NT-4 did not significantly improve the survival rates of either large or small RGCs.

conclusions. BDNF improved the survival rate of RGCs, particularly of small RGCs, concentration-dependently, but NT-4 had little influence on the survival rate. The current method was useful in evaluating the effects of neuroprotective factors or neurotoxic factors on cultured RGCs.

In recent years, retinal ganglion cells (RGCs) have been isolated with immunologic methods, ¹ and new findings have been reported. ^{2 3} However, in the previous reports, RGCs have been evaluated microscopically with relatively small numbers of cells and subjective judgment. Therefore, a more objective and accurate evaluation has been sought. 

Brain-derived neurotrophic factor (BDNF) discovered in 1975 is one of the most important neurotrophic factors and is involved in the classification and survival of neurons. ^{4 5 6 7 8 9} It has been reported that BDNF improved RGC survival in vitro and that BDNF suppressed RGC loss in axotomized animals and in the ocular hypertensive model. ^{6 7 10} Neurotrophin (NT)-4 is another major neurotrophic factor that has been reported to increase the RGC survival rate and is involved in retinal development, but these results are not necessarily consistent. ^{5 8 11 12 13 14}  

Rat RGCs have been classified into subgroups by their cell size, ^{2 7 15} and RGCs of different sizes had different susceptibilities to the loading conditions of these factors. ⁷ However, it has been difficult to evaluate precisely the effects of loading conditions on RGCs separated by their cell size. 

Flow cytometry has been recently developed to analyze cells very quickly and objectively. Cells can be rapidly separated according to their size, cell cycle, and other parameters with this method. 

In the present study, isolated RGCs cultured in serum-free medium were subject to flow cytometric analysis to investigate quantitatively and objectively the neurotrophic effects of BDNF and NT-4 on RGCs. 

Methods

All experiments were conducted and all laboratory animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and the Helsinki Treaty. 

Isolation of Retinal Ganglion Cells

RGCs were isolated according to the two-step panning method reported. ¹ Briefly, 2-day-old rats (Sprague–Dawley) were euthanatized to obtain approximately 60 eyes for each experiment. Retinas were separated from enucleated eyeballs, incubated for 20 minutes in a solution containing 5 mg/ml papain to dissociate the cells, and incubated for 5 minutes with rabbit antimacrophage antibody (Inter-Cell Technologies, Hopewell, NJ). Cell suspensions were treated for 45 minutes in 100-mm Petri dishes coated with goat anti-rabbit IgG (L + H chain) antibody (Southern Biotechnology Associates, Birmingham, AL). Suspensions containing cells that did not adhere to the Petri dish were treated for 1 hour in 100-mm Petri dishes coated with anti-Thy 1.1 antibody (from hybridoma T11D7e2; American Type Culture Collection, Rockville, MD). 

Cells that adhered to the second Petri dish were collected after treatment with 0.125% trypsin for 10 minutes for avulsion and incubated in 24-well plates. 

Culture Conditions

Culture plates were coated with 0.1 mg/ml of polyornithine (Sigma Chemical Co., St. Louis, MO) for 5 hours or longer and then coated overnight with 5 μg/ml of EHS-laminin (Upstate Biotechnology, Lake Placid, NY). The medium developed by Politi et al.¹⁶ for monolayer culture of mixed mouse retinal neurons was modified for use in this experiment. The medium used was Dulbecco’s modified Eagle’s medium to which the following were added: insulin (1.6 × 10⁻⁶ M), progesterone (4.0 × 10⁻⁸ M), selenite (6.0 × 10⁻⁸ M), transferrin (12.5 × 10⁻⁸ M), putrescine (2 × 10⁻⁴ M), hydrocortisone (1.0 × 10⁻⁷ M), cytidine-5′-diphosphocholine (5.2 × 10⁻⁶ M) and cytidine-5′-diphosphoethanolamine (2.9 × 10⁻⁶ M). The seeding density was 1200 cells/mm². The cells were incubated at 37°C in humidified 10% CO₂ and 90% air. 

BDNF or NT-4, 10 μl, was added to each well containing RGCs, after preparation with phosphate-buffered saline (PBS) to make final concentrations of 1, 10, and 100 pg/ml and 1, 10, and 100 ng/ml for BDNF or 0.1, 1, 10, and 100 ng/ml for NT-4, respectively. As the control, 10 μl PBS without neurotrophic factors was added to each well. 

Treatment of Cells for Evaluation

After incubation for 48 hours, cells were treated for 10 minutes, keeping the same culture condition with 5-chloromethylfluorescein diacetate (Molecular Probes, Inc., Eugene, OR). After the reaction, the cells were detached from the culture dish by gently pipetting 20 to 30 times. To avoid cell damage as much as possible, pipetting was performed gently without making bubbles in the incubator. Then, RGCs in the supernatant were subjected quickly to flow cytometry. It took approximately 3 to 4 minutes from cell avulsion to complete the analysis for each culture well. The survival rate of RGCs was measured quantitatively by the amount of fluorescence. The condition of the RGCs was observed microscopically before and immediately after avulsion, and no attached cells were confirmed on the culture dish. 

Conditions of Flow Cytometry

Cells were evaluated using a flow cytometer (Facscaliber; Becton Co., San Jose, CA). The measurement conditions were determined from the preliminary experimental data as follows. Forward scatter (FSC) value reflecting cell size: voltage E00, amplifier gain 2.3; side scatter (SSC) value: reflecting cytoplasmic structure: voltage 400, amplifier gain 2.0; and fluorescent intensity: voltage 450, amplifier gain 1.5; flow rate: high. Ten thousand cells were evaluated in every sample automatically. Two sizes of standard particles, 6 and 10 μm, were measured to determine cell size before cell evaluation. 

Statistical Tests

The survival rates for the various conditions were tested with ANOVA and the post hoc method. A significant difference was defined as P < 0.05. 

Results

Morphologic Observation of Isolated Retinal Ganglion Cells

The percentage of RGCs recovered of all retinal neurons was 1.2 ± 0.54%. Isolated and cultured RGCs were classified morphologically into two classes: large cells of approximately 10 to 12μ m in diameter and small cells of approximately 6 to 8 μm in diameter (Fig. 1) . The large cells were somewhat spindle-shaped, and most of them had a comparatively long process. The small cells were round, and most of them had a somewhat thick and short process. There were more small cells than large cells. RGCs in the supernatant immediately after avulsion also showed a small number of large cells and a large number of small cells. However, these large cells in the supernatant showed slight changes of their appearance comparing with RGCs attached on the culture well. The single, long process seemed to be shortened, and cell body was comparatively round. RGCs in the supernatant were floated individually. 

Results of Analysis Using Flow Cytometry

From each of culture well, a total of 10,000 RGCs were subject to analysis and each analysis was completed within 2 minutes. Two sizes of RGCs were observed in flow cytometry. According to the positions of standard beads, smaller cells and larger cells were observed at locations that would correspond to the sizes observed microscopically. The ratio of the small RGCs to the large RGCs was approximately 4.2 to 1 in the control, which was similar to that observed microscopically. The current fluorometric assay showed very stable intensity of fluorescence among the all experiments, and the RGCs were divided by their intensity of fluorescence and cell size. Figure 2 shows a representative result of a density blotting image. Some cells of both sizes emitted strong fluorescence (intensity ranges: approximately 0.5 × 10 to 5 × 10² for the small cells and 0.3 × 10² to 3 × 10³ for the large cells), which were considered survival cells, whereas others emitted weak fluorescence (intensity ranges: approximately 0.4 × 10 to 6 × 10 for the small cells, 2 × 10 to 0.3 × 10² for the large cells), which were considered nonsurvived cells. We estimated the survival rate by calculating the cell numbers in each intensity range shown as peaks by histogram images (Figs. 3a 3b ). 

Effects of BDNF

Although the survival rate of the small cells in the control group was 78.6%, BDNF improved the survival rate concentration-dependently. The survival rates of small RGCs subjected to BDNF concentrations of at 100 pg/ml, 1 ng/ml, 10 ng/ml, and 100 ng/ml increased significantly from that of the control by 5.9%, 8.2%, 17.2%, and 17.4%, respectively (Fig. 4) . 

The survival rate of the large cells in the control group was 83.1%. As for the small cells, BDNF improved the survival rate concentration-dependently. The survival rates of large cells subjected to BDNF concentrations of 10 and 100 ng/ml increased significantly from that of the control by 7.3% and 7.8%, respectively (Fig. 4) . 

Effects of NT-4

NT-4 slightly improved the survival rate of small RGCs in a somewhat dose-dependent manner. The survival rate of small RGCs with 100 ng/ml of NT-4 increased by 4.1% from that of the vehicle control. However, this improvement was not significant. The survival rate of large RGCs did not change with any of the NT-4 concentrations used. 

Discussion

We confirmed the purification of RGC by recovery rate, morphologic appearance, and retrograde labeling of isolated RGCs by DiI injection from the superior colliculus. ^{1 2 3} The sizes and distribution of both small and large RGCs classified in flow cytometry were the same as those of RGCs on the culture dish or in previous reports. 

The evaluation of RGC survival using fluorescence intensity or morphologic observation in the previous reports had a limit of evaluated number of cells and took some time, during which the cell conditions may have changed. Some reports evaluated RGC survival according to the degree of process elongation. ³ However, many cells with short processes were seen emitting strong fluorescence of 5-chloromethylfluorescein diacetate, indicating that RGC survival could not necessarily be evaluated accurately by measuring the length of the process. It is not easy to draw the line accurately between survived cell and nonsurvived cells by previous methods. The present method can classify a total of 10,000 RGCs in approximately 2 minutes per sample automatically by their cell size and fluorescence intensity. Therefore, we believe that the present method evaluates cell conditions much more objectively and accurately than previous methods. However, the present method requires approximately twice the cell density needed in previous methods. ^{1 2}  

In observations of RGCs under a fluorescent microscope before and after avulsion, cells immediately after avulsion were found to be round with a short process. However, the intensity of fluorescence did not change, suggesting that there should be no problem in evaluating of cell survival. In preliminary studies, we tried to use some enzymes to dissociate and recover cells from the culture dish. In these cases, however, cell viability was decreased and not suitable for evaluation. 

The cell size of surviving cells was always slightly larger than nonsurviving cells, either in the small-cell group or large-cell group in the present study. Because it is well known that cells shrink their cell body by apoptosis, nonsurviving cells in the present study were thought to expire by apoptosis. Indeed, electron-microscopic observations of currently isolated RGCs showed the condensation or defragmentation of nuclei, which represent typical apoptotic changes (data not shown). 

In the present study, BDNF improved the survival rate concentration-dependently, having nearly the same range as that reported in previous studies. The maximum improvements of the survival rates of small and large cells were 17.4% and 7.8%, respectively, which seem to be lower than the values of previously reported in in vivo studies. ¹⁷ Several explanations of this difference are possible. Because previous studies were conducted by injection into the hyaloid body or by coculture, the effects of unknown factors were undeniable, and accurate quantitative analysis was difficult. Several factors have neurotrophic effects on RGCs, and multiple factors should be involved in neurotrophic action. ^{7 8 10 18 19 20 21 22} Indeed, the number of RGCs in the BDNF knock-out mouse was the same as that of the wild type. ^{19 23} In the present study, isolated RGCs were cultured in a serum-free medium, eliminating the participation of unknown factors. Second, the degree of improvement should differ depending on cell type. Moreover, the maturity of RGCs may greatly affect the results. Johnson et al. ⁶ reported that BDNF rescued retinal neurons by 7% in E17 rat retina. However, this improvement decreased in retinas of rats older than E17. ²⁴ The number of trk-B receptors, a major BDNF receptor, was variable, depending on retinal maturity. ^{25 26 27 28}  

It has been reported that rat RGCs can be classified into subgroups by cell size, and those reactions against neurotoxic events or neurotrophic factors were different depending on the cell size. ^{7 15} Large RGCs were very resistant to axotomy in adult rats. ⁷ The existence of trk-B receptors in retina is well known. ^{25 26 28 29} Although many trk-B receptors exist as a homodimers in the cell membrane, heterodimers composed of truncated trk-B and full-length trk-B receptors have inhibitory effects on BDNF signaling, ^{30 31} and those truncated trk-B receptors were detected to a much greater extent in large RGCs than in small RGCs in the adult rat. ³¹ The trk-B receptor is also the major receptor for NT-4. This evidence may explain why the improvements in the survival of large RGCs with BDNF or NT-4 was lower than that for small RGCs in the present study. 

The neuroprotective effects of NT-4 on RGCs are not necessarily consistent. ^{5 8 11 12} The expression of NT-4 in the retina is weaker than that of BDNF. This may be the reason why significant neuroprotective effects of NT-4 on RGCs were not observed in the present study. Because it was expected that the neurotrophic effects should be clear when neurotoxic factors were challenged to the RGCs, 100 μM glutamate was administrated to RGCs with or without 100 ng/ml of NT-4. Glutamate decreased the survival rate in both small and large RGCs. However, the improvement of survival rate by NT-4 was not significant (data not shown). Overall, it is implied that the neuroprotective effect of NT-4 on RGCs is weaker than that of BDNF, at least in the postnatal rat in vitro model. 

In addition to BDNF and NT-4, several other neuroprotective factors have been reported such as nerve growth factor (NGF), NT-3, and NT-6, ciliary neurotrophic factor (CNTF), basic-fibroblast growth factor (b-FGF). On the other hand, many neurotoxic factors like glutamate have been discovered. Using the current experimental system, the effects of various neuroprotective and neurotoxic factors on RGCs can be evaluated objectively and accurately. 

 Supported by Grant 11307036 from the Japanese Minister of Health and Welfare (KK).

 Submitted for publication November 3, 1999; revised February 9, 2000; accepted March 8, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Phase-contrast image of retinal ganglion cells. The phase-contrast image shows retinal ganglion cells that were separated using a two-step panning method and cultured for 2 days. There are many relatively small round cells and a few large and spindle-shaped cells. Although small cells have a comparatively short process, large cells have a long process. Bar, 10 μm.

Figure 2.

 

Representative result of flow cytometry. Density blot image: RGCs are classified into a large number of small cells (diameter: approximately 6 to 8 μm, indicated by a quadrangular area S) and a small number of large cells (diameter: approximately 10–12 μm, indicated by a quadrangular area L). Surviving cells, which emit strong fluorescence after reaction to 5-chloromethylfluorescein diacetate, are seen in the upper portion of the figure. Nonsurviving cells emitting weak fluorescence are seen in the lower portion of the figure. The diameter of nonsurviving cells is somewhat smaller than that of surviving cells. In total, approximately 10,000 cells were analyzed in approximately 2 minutes. FSC, forward scatter, corresponding to cell size; FL, intensity of fluorescein.

Figure 3.

 

Histogram of fluorescein distribution of RGCs. Histograms of fluorescein distribution of quadrangular areas S and L in Figure 2 . In both histograms for small RGCs (A) or for large RGCs (B), the right peaks representing cells emitting greater amount of fluorescein represent surviving cells (A), and the left peaks representing cells emitting weak fluorescein represent nonsurviving cells (B). The number of cells in each peak was counted and estimates of survival rates were determined from them.

Figure 4.

 

Effects of BDNF on RGCs. The survival rates of small and large RGCs increased concentration-dependently compared with those of the vehicle controls. The survival rate of the small and large RGCs increased significantly with concentrations of BDNF higher than 100 pg/ml and 10 ng/ml, respectively. Bar, SD. BDNF, brain-derived neurotrophic factor.* P < 0.05 vs. control, §P < 0.01 vs. control.