Effects of triamcinolone acetonide on vessels of the posterior segment of the eye

Reagents

All culture reagents were obtained from Gibco (New York, NY). Corticosteroids were purchased from Sigma (Saint-Quentin-Fallavier, France) except for Kenacort Retard, which was purchased from Bristol-Myers Squibb (Paris, France).

Culture of bovine retinal endothelial cells and treatments

Bovine retinal endothelial cells (BRECs) were seeded in Dulbecco’s modified Minimal Essential Medium (DMEM)/glutamax supplemented with 10% calf serum and 20 ng of VEGF (Sigma, St. Louis, MO), at 37 °C in a humidified atmosphere containing 5% CO₂. After 24 h the cells were treated for either 24, 72 h, or 5 days with 0.1 or 1 mg/ml TA that had been dissolved in 1% ethanol (TA-E; 99.9% ethyl alcohol; VWR, Paris, France). Control cells were either not treated or treated with 1% ethanol in medium. Before treatment, cells were maintained in medium supplemented with 2% of fetal calf serum (FCS) for 24 h. For treatment, cells were cultured in medium supplemented with 2% FCS without VEGF. In some experiments, cells were treated after confluence was reached.

Analysis of TA effects on cultured cells

Evaluation of cell proliferation measuring DNA synthesis rate

For proliferation assay, TA was used at 0.1 mg/ml and 0.5 μg/ml. BRECs (10⁵ cells/well) were seeded in 24-well cell culture plates and allowed to attach for 24 h. Thereafter, the cells were incubated with TA-E for 24 h and 5-bromo-2-deoxyuridine (BrdU) was added into the medium, for a final concentration of 10 µM, and allowed to incubate for 18 h. The incorporation of BrdU into the genomic DNA was determined using the ELISA BrdU labeling and detection Kit III (Roche Molecular Biochemicals, Germany).

For other experiments, BRECs were seeded at 5x10⁴ cells/well in the Lab-Tek chambered Coverglass and grown at 37 °C in a humidified atmosphere containing 5% CO₂ and 95% air for 24 h. They were then treated with 0.1 mg/ml TA-E, which is a concentration that allows further analysis. Control cells were run using the culture medium alone or the culture medium containing 1% ethanol. All experiments were performed in triplicates and repeated twice. The Lab-Tek slides were mounted in Gel Mount (Biomeda, Burlingame, CA) before they were examined and photographed using an Olympus IX70 fluorescent microscope coupled to a digital camera. The following excitation/emission wavelengths were used: 365–420 nm; 470/510 nm; 510/560 nm.

Detection of apoptosis using terminal deoxynucleotidyl transferase dUTP nick end labeling assay

After treatment with 0.1 mg/ml TA-E for 24 h, 72 h, or five days, the cells were rinsed, air-dried, and fixed with 4% paraformaldehyde (Merck Eurolab, Fontenay sous-bois, France) one hour at room temperature. The cells were rinsed during 5 min in PBS, and permeabilized with 0.3% Triton X-100 in PBS for 20 min. Cells were incubated for 30 min with 50 µl of terminal deoxynucleotidyl transferase dUTP nick end labeling assay (TUNEL; Roche Diagnostics, Mannheim, Germany) reaction mixture at 37 °C. The cells were then rinsed once with PBS, stained with DAPI (Sigma), and rinsed with PBS five times for 5 min each time. Positive controls were obtained by inducing apoptosis with 1 µM staurosporine.

Annexin V binding

Annexin V binding was assayed with the Annexin V Apoptosis detection kit (Santa Cruz Biotechnology Inc., Santa Cruz, CA) according to the manufacturer instructions. Briefly, after treatment cells were rinsed with PBS, washed once for 5 min with 500 µl per well of assay buffer and incubated 15 min at room temperature with 2 µg/ml of Fluorescein isothiocyanate (FITC)–labeled Annexin V. Cells were then, rinsed with PBS, fixed with 4% paraformaldehyde, rinsed with PBS, stained with DAPI and rinsed five more times for 5 min with PBS.

Western blot analysis

Total protein extracts were obtained by lysis of BRECs in Laemmli sample buffer. Extracted proteins were separated by SDS–PAGE, immobilized on nitrocellulose membrane (Millipore, Billerica, MA) and blotted with affinity purified rabbit anti-active Caspase-3 polyclonal antibody in a 1:500 dilution (Calbiochem, San Diego, CA). The secondary goat anti-rabbit IgG (Vector Laboratories, Burlingame, CA) was used in a 1:5,000 dilution. The amount of total protein extracts analyzed was 30 μg/lane. Positive controls for caspase-3 activation were run on HL-60 cells, cultured in RPMI 1640 supplemented with 10%. Fetal calf serum, treated for 24 h with 50 μM etoposide [19].

Immunocytochemistry

After the different treatments BREC cells were stained with the following antibodies: anti-cytochrome C [20] (Sigma, Steinheim, Germany) diluted 1:250 in PBS containing 1% BSA (BSA 98%; Sigma, Steinheim, Germany); anti-apoptosis-inducing factor (AIF) [21] (Sigma, Saint-Quentin Fallavier, France) diluted 1:100 in PBS containing 1% BSA; anti-microtubules-associated protein and light chain 3 (MAP-LC3) [22] (Santa Cruz Biotechnology) diluted 1:50 in PBS containing 1% fat-free milk; and polyclonal anti- L -DNase II antibody [23,24] diluted 1:100 in PBS containing 1% skimmed milk. After incubation with the specific primary antibodies, the cells were rinsed three times for 5 min with PBS, and incubated for 1 h with goat Alexa Fluor anti-rabbit IgG (Molecular Probe Invitrogen, NewYork, NY), or rabbit anti goat Texas red dye-conjugated (Jackson Immuno Research Laboratories Inc., Suffolk, UK) diluted 1:250 in PBS containing 1% BSA. PBS containing 1% BSA or 1% skimmed milk was used instead of the primary antibody as negative controls.

For autophagic positive controls, BRECs were cultured for 24 h in amino acid-depleted medium, e.g Hanks Buffered Salt Saline (HBSS, Gibco, New York, NY) complemented with 20 mM glucose [25]. Slides were mounted in Gel Mount and examined with a fluorescent microscope Olympus IX70 coupled to a digital camera. The following excitation/emission wavelengths were used: 365–420 nm; 470/510 nm; 510/560 nm.

Semithin and ultrathin structure analysis

Lewis rats used in this study were obtained from Roger Janvier, Le Genest Saint Isle, France. The use of animals adhered to the ARVO statement for Ophthalmic and Vision Research and protocols were approved by the ethical committee of René Descartes University of Paris. Lewis rats (6-week-old, weighing 150 g) were intravitreally injected with either 10 μl of PBS (n=3) or with 10 μl of PBS containing 40 μg of TA (n=3). Thirty days after injection, the rats were euthanized with a CO₂ overdose, and their eyes processed for semithin and transmission electron microscopy analysis. The eyes were fixed in 2.5% glutaraldehyde cacodylate buffer (Na 0.1 M, pH 7.4) for 1 h. They were then dissected, and the posterior segments were fixed for 3 h. Specimens were fixed in 1% osmium tetroxyde in cacodylate buffer (Na 0.1 M, pH 7.4) and progressively dehydrated in graduated ethanol solution (50, 70, 95, and 100%). Each posterior segment was separated in two samples, included in epoxy resin and oriented. Sections were cut to 1 µm thickness with an ultra microtome Reichert Ultracut (Leica, Biel, Switzerland) and stained with toluidine blue. Sections cut to 80 nm thick were stained by uranyl acetate and lead citrate, and analyzed with a Philips CM10 electron microscope.

Vascular corrosion casts

Lewis rats (12, 8-week-old, weighing 150 g) were intravitreally injected with either 10 µl of PBS (n=6) or with 10 µl PBS containing 40 µg of TA (n=6). After 30 days the animals were sacrificed by CO₂ inhalation followed by a thoracotomy. A catheter was introduced into the aorta through the left heart ventricle, and the right auricle was cut to allow evacuation of injected products. A perfusion was performed with a mixture of red Mercox resin and catalyst (Ladd Research, Williston, VT). Eyes were extracted and lenses were removed. Tissues were conserved overnight at 37 °C in PBS to allow complete polymerization, and then digested by 5% KOH for two weeks at 37 °C until only the vascular corrosion casts remained. Distilled water was used to remove salt, and the mold was dried. Only corrosion casts with completely filled iris vessels were used. This was to exclude corrosion casts from incomplete perfusion. Retinal vasculature was removed with forceps. The specimens were mounted on Stretched Specimen Mount (SSM) stubs, coated with gold palladium, and scanned at an accelerating voltage of 117 kV. To measure the thickness of the choriocapillary lumen, we cut corrosion casts paracentrally (1 mm from the aperture of the optic nerve) and positioned for perpendicular views of the choriocapillaries. To analyze the intercapillary space (avascular area), we positioned the casts for frontal views of the choriocapillaries. Electron micrographs were scanned and analyzed using Image J Software. The avascular areas were measured on frontal views and expressed as the percentage of intercapillary surface (space between the plastic capillary casts) of the whole area (n=6 rats; five pictures/cast).

Effect of TA on VEGF and COX-2 expression in primary retinal pigment epithelium (RPE) cells

Outer segments isolation and phagocytosis assay

Outer segments (OS) were isolated following established protocols [27]. Briefly, 20 pig retinas were dissected and homogenized in 20% sucrose buffer, 20 mM Tris, 2 mM MgCl₂, and 0.13 mM NaCl (pH 7.2). Retina samples were centrifuged on a sucrose gradient (50%, 27%) 1h at 100,000× g, and OS were harvested at the ring interface and diluted in DMEM. Centrifugation for 10 min at 10,000× g was performed, and the pellet was resuspended in DMEM to obtain a stock solution of 10⁸ OS/ml.

Before treatment, RPE cells were maintained in medium supplemented with 2% of FCS for 24 h. Confluent rat primary RPE cells were incubated with 0.1 mg/ml TA-E. Control cells were run using the culture medium alone. After 24 h of treatment, the cells were rinsed once with PBS and incubated with 150 μl of 10⁸ outer segments/ml. After 1 h, 850 µl of complete medium was added. After 6 h, cells were washed and 350 μl of guanidine thiocyanate lysis buffer (RNeasy mini kit; Qiagen, Courtaboeuf, France) was added for RNA extraction.

Reverse transcription and real-time polymerase chain reaction

Total RNA was isolated with RNeasy plus Mini Kit (Qiagen, Courtaboeuf, France). Single-stranded cDNA was synthesized from total RNA using random primer and superscript reverse transcriptase (Invitrogen). Subsequent real-time polymerase chain reaction (RT–PCR) was performed using cDNA, qPCR SuperMix-UDG Platinum SYBR Green (Invitrogen), and the and 0.5 pmol/µl of the primers described on Table 1.

PCR reactions were performed in 40 cycles of 15 s at 95 °C, 45 s at 60 °C. No product was obtained in control reactions where reverse transcriptase was omitted.

Statistical analysis

Data are represented as mean±standard error of the mean (SEM) and compared using the nonparametric Mann–Whitney U test. P values were calculated for a confidence interval of 95% and p values of less than 0.05 were considered significant.

Effect of TA on BREC viability

The percentage of living cells was reduced to 75±1.1%, p<0.05 after exposure to 0.1 mg/ml of TA for 24 h. This survival was reduce to 52±2.9% p<0.05 as concentration of TA increased to 1 mg/ml. The rate of survival was further reduced as exposure time increased, being of 58±2.3% and 50±3.7% for 0.1 mg/ml, and 1mg/ml of TA, respectively, after 72 h and of 55±2% and 39±0.9% p<0.05, respectively, for the same concentrations after 5 days of treatment. TA therefore decreased the number of living cells as compared to control cells, with an increased effect upon exposure duration. Photomicrographs of DAPI-stained BREC nuclei in culture showed that after 72 h (not shown) or five days (Figure 1D-F) of treatment with 0.1 mg/ml TA-E, the cells were not only reduced in number (Figure 1D), but also organized in a ramified manner (Figure 1E). After treatment with 1 mg/ml for 72 h or five days (Figure 1F), the number of cells was reduced and many cellular debris were present in the culture medium. To evaluate whether the reduction in the number of living cells resulted from a reduction in proliferation or in TA-induced cell toxicity, we performed the same experiments on confluent cells. As shown on Figure 1B, TA induced a similar reduction in the number of living cells, this effect increased upon the duration of exposure and the concentration, suggesting that the reduction in the number of living cells did not result solely from an inhibition of cell proliferation.

Inhibition of BREC proliferation

To further characterize the effect of TA on endothelial cell proliferation, we evaluated the DNA synthesis rate using the BrdU test. As shown in Figure 1C, 0.1 mg/ml TA-E significantly reduced the proliferation rate of BRECs from OD=2.06±0.09 in control cells to OD=1.62±0.02, n=4, *p<0.05 in TA-E treated cells. This effect was not observed with a lower concentration of TA-E: 0.5 µg/ml (OD=1.8±0.03).

Evaluation of retinal endothelial cell death after exposure to TA

Caspase-dependent apoptosis

Caspase-dependent apoptosis was evaluated by TUNEL assay, annexin-V binding, caspase 3 activation, and cytochrome C release. As shown in Figure 2, no TUNEL-positive cells were detected after treatment with 1% ethanol (Figure 2A-C). In addition, no TUNEL-positive cells could be detected in cells treated with 0.1 mg/ml TA-E after 72 h (Figure 2D-F) or even after five days of exposure (Figure 2G-I). However, when cells were exposed to 1 µM staurosporine, a known inducer of caspase-dependent cell death, TUNEL-positive cells with condensed nuclei were observed (Figure 2J-L).

During the early phases of any type of apoptosis, phosphatidyl serine translocates to the external surface membrane of the cell [28] and binds annexin-V [29]. As shown in Figure 3, while no annexin V was found to bind on the surface of control cells (Figure 3A-C), positive, stained cells were detected after 72 h and five days of exposure with 0.1 mg/ml TA-E (Figure 3D-F,G-I). After 72 h of exposure to TA, annexin-V positive cells displayed condensed or fragmented nuclei (Figure 3D-F insets), while after five days some cells displaying normal (uncondensed) nuclei also started annexin-V binding (Figure 3G-I insets).

The results discussed in the previous section suggested the activation of apoptosis with no activation of caspase. To assess this point, we performed western blot analysis of caspase-3 activation. The results obtained revealed the presence of pro-caspase-3 and the absence of activated caspase-3 in BRECs treated during 72 h or five days with either nothing, 1% ethanol medium, or in 0.1 mg/ml TA (Figure 4). In positive control HL-60 cells treated with etoposide, pro-caspase-3 as well as activated caspase-3 were detected (Figure 4). These experiments showed that a noncaspase dependent form of apoptosis could intervene.

After a toxic stress caspase 3 is, in most of the cases, activated though the intrinsic pathway, in which, the release of cytochorme C from the mitochondria is a key event. We investigated then if cytochrome C was released. In BRECs treated during 72 h or five days with 0.1 mg/ml of TA-E (Figure 5D-F,G-I), no release of cytochrome C from mitochondria into the cytoplasm could be observed. The fluorescent labeling appeared indeed located in mitochondria, and no diffuse or nuclear fluorescence was present, indicating that mitochondria have not released cytochrome C. In contrast, in BRECs treated with staurosporine, cytochrome C was located in condensed nuclei of apoptotic cells (Figure 5 J-L, arrows).

Caspase-independent apoptosis in triamcinolone acetonide-induced toxicity

As the cell death induced by TA seemed to be caspase independent, we analyzed two markers of caspase-independent apoptosis in TA-treated BRECs: leukocyte elastase inhibitor/ L-DNase II (LEI/L-DNase II) and apoptosis inducing factor (AIF).

LEI and L-DNase II are two forms of one protein that are recognized by the same antibody. However, while LEI is a cytoplasmic anti-protease, its nuclear translocation parallels its change into L-DNase II activity [30]. While LEI/L-DNase II fluorescent labeling was present in the cytoplasm of the control cells (Figure 6A-C), surrounding non-fluorescent nuclear shapes, a clear nuclear staining was observed in cells treated with 0.1 mg/ml TA for 72 h or for five days (Figure 6D-F,G-I, arrows). Nuclei positive for L-DNase II were condensed or fragmented, indicating a caspase-independent apoptotic pathway, taking place in BRECs cells treated with TA.

In control cells (Figure 7A-C), anti-AIF showed a cytoplasmic labeling; corresponding to the normal localization of AIF at the mitochondria. After 72 h of TA-E treatment no change in AIF localization was seen (Figure 7D-F). However, after exposure to 0.1 mg/ml TA-E for 5 days, some cells with condensed or fragmented nuclei showed positive nuclear staining. This suggested that caspase-independent apoptotic pathways were activated in TA-treated cells (Figure 7G-I, arrows).

Autophagy

After activation of autophagy, MAP-LC3 concentrated in autophagic vesicles, giving a dotted pattern to the cells. After 72 h (not shown) or five days of treatment with 0.1 mg/ml TA-E (Figure 8D-F), no MAP-LC3-positive vesicles could be observed—meaning that autophagy was not taking place under these conditions. In cells cultured in amino acid-deprived medium, a condition known to induce authophagy, cells showed large intracytoplasmic vesicles that stained positive with anti-MAP-LC3 antibody (Figure 8G-I).

Structural analysis of rat retinal and choroidal vasculature, one month after TA exposure

As demonstrated, the results obtained indicated that TA could affect endothelial cell viability. We investigated this hypothesis in vivo. Semithin analysis of the retina did not show evidence of alteration of the retinal vessels or capillaries, but did show that a reduction of the diameter of the superficial capillaries was evident on sections of TA-treated retinas (Figure 9B) versus control ones (Figure 9A). More striking was the significant thinning of the choriocapillaries and choroid in the TA-treated rats (Figure 9D) as compared to PBS-injected rats (Figure 9 C and E). Moreover, as previously described [26] in TA-treated rats, there were alterations of the outer segments and vacuolization of the RPE cells (Figure 9D stars).

Analysis of rat choroidal vascular corrosion casts one month after TA exposure

To better analyze and quantify the modifications observed in the choroidal vascular network, we performed resin casts. In rats exposed to TA for a month, the casts showed that the spaces between choriocapillaries were increased and that capillary diameter was reduced as compared to PBS-injected rats. Figure 10A,B presents frontal views of choriocapillary casts of a PBS-treated rat and a TA-treated rat. Figure 10C,D shows cross-sectional cuts of pericentral choroidal corrosion casts of a PBS-treated rat and a TA-treated rat. In the TA-treated rat eyes, the avascular areas between choriocapillaries were significantly increased (n=6, *p<0.05 versus PBS) and the choriocapillary diameter significantly reduced (n=6, *p<0.05 versus PBS; Figure 10E,F).

Modulation of VEGF and COX-2 expression in RPE cells exposed to OS phagocytosis

To explore a potential indirect mechanism responsible for TA-induced choriocapillaries alterations, we evaluated whether the expression of cyclo-oxygenase-2 (COX-2) and VEGF in RPE cells submitted to OS phagocytosis was modulated by TA exposure. We have previously shown that COX-2 knockout mice develop significant choriocapillary involution [26]. So we hypothesed that beside a direct toxic effect of TA, COX-2 downregulation could also be induced by TA on the choriocapillary in vivo.

In primary culture rat RPE cells, OS phagocytosis significantly enhanced COX-2 expression (Figure 11A; ***p<0,001, OS versus control). TA decreased significantly the expression of COX-2 under basal condition (Figure 11A; *p<0.05 versus control) and also significantly reduced the OS phagocytosis-induced COX-2 upregulation (*p<0.05 OS-TA versus OS). Interestingly, while TA also significantly downregulated basal VEGF expression (Figure 11B; *p<0.05 TA versus control), it did not influence OS phagocytosis-induced VEGF upregulation (Figure 11B; ***p<0.001 OS versus control).