N-(4-hydroxyphenyl) Retinamide Augments Laser-Induced Choroidal Neovascularization in Mice

Abstract

purpose. To evaluate the effect of N-4-hydroxyphenyl retinamide (4-HPR) on experimental laser-induced choroidal neovascularization (CNV) and on the expression and secretion of relevant growth factors by cultured human retinal pigment epithelial (RPE) cells.

methods. CNV was induced by laser photocoagulation in C57BL/6 mice. 4-HPR (0.2 or 1 mg) or vehicle, was injected intraperitoneally twice daily for 14 days. Plasma and tissue levels of 4-HPR were measured by HPLC. CNV was evaluated by fluorescein angiography, histology, and quantitative confocal analysis of isolectin B4 histochemistry on days 7 and 14. Induction of apoptosis and expression and secretion of growth factors was studied in 4-HPR-treated RPE cultures.

results. Mice treated with 4-HPR exhibited time- and dose-dependent increases in plasma and tissue 4-HPR levels. CNV lesions showed increased volume with increased vascular leakage and contained fewer lesion-associated RPE in treated versus untreated mice. Treatment of nonpolarized RPE cultures with 4-HPR in the presence of serum resulted in RPE apoptosis; however, apoptosis was minimal in similarly treated highly polarized RPE. Treatment of RPE cells with 4-HPR resulted in the upregulation of VEGF-A and -C (P < 0.05) and Ang-1 (P < 0.01) mRNA and increased secretion of VEGF-A and -C (P < 0.05), whereas pigment epithelium–derived growth factor (PEDF) and thrombospondin (TSP)-1 mRNA expression and secretion were downregulated (P < 0.05).

conclusions. 4-HPR increases lesion size and leakage in laser-induced CNV and is associated with the upregulation of key proangiogenic factors and the downregulation of antiangiogenic factors. Consistent with the preferential loss of RPE in CNV lesions in vivo, 4-HPR induces apoptosis of nonpolarized RPE in the presence of serum.

Age-related macular degeneration (AMD), the most common cause of blindness and vision impairment in Americans aged 60 and older, manifests in two forms. ¹ The nonexudative (dry) form of AMD is characterized by degenerative changes in the retinal pigment epithelium (RPE) and deposition of drusen beneath the RPE. The exudative (wet) form of AMD is characterized by the growth of neovascular choroidal vessels under the retina. These vessels leak fluid and blood and may ultimately result in a blinding disciform scar. ² Several theories of the pathogenesis of AMD have been proposed, involving genetic and environmental factors related to primary RPE senescence, ocular perfusion abnormalities, increased inflammation, excessive lipofuscin accumulation, and oxidative stress. ^{1 2}  

Choroidal neovascularization (CNV) is a complex, multistep process that is associated with alterations in the balance between pro- and antiangiogenic factors. ^{3 4 5} Animal models of CNV are important tools for studying the potential effects of drugs on angiogenesis. The most widely accepted and most frequently used mouse model for studying CNV is laser-induced Bruch’s membrane photocoagulation. ⁶  

There is increasing evidence suggesting that an imbalance between angiogenesis-stimulating factors and angiogenesis inhibitors results in a proangiogenic environment in ocular disorders, with hypoxia, ischemia, inflammation, and tumor ^{5 7} ; hence, the assessment of the pro- and antiangiogenic profile is an important component of the evaluation of drug efficacy. Vascular endothelial growth factor (VEGF) is a critical mediator of angiogenesis in the eye. ^{8 9} Its effects are mediated through binding to the tyrosine kinase receptors VEGF-R1(flt-1) and VEGF-R2 (flk), which are predominantly expressed in the vascular endothelium. ¹⁰ VEGF is secreted in a biologically active form and its receptors are found at sites of angiogenesis. ¹¹ A marked increase has been reported in VEGF receptors in endothelial cells participating in CNV. ¹² Normal RPE cells secrete low levels of VEGF, but under pathophysiological conditions such as CNV, VEGF triggers angiogenesis and increased vascular permeability. Studies in our laboratory and those from others have shown that induction of oxidative stress stimulates VEGF-A ^{13 14 15 16} and -C ¹⁵ in RPE cells. Angiopoietins (Ang-1 and -2) constitute a second group of proangiogenic factors that are localized in CNV membranes ¹⁷ and are produced by RPE cells. ^{17 18 19} Both Ang-1 and -2 facilitate VEGF-induced angiogenesis, with Ang-1 promoting vascular network maturation and Ang-2 initiating angiogenesis. ²⁰ Among the antiangiogenic molecules, pigment epithelium–derived growth factor (PEDF) is the most potent inhibitor of angiogenesis that is expressed in the RPE and is involved in the pathogenesis of angiogenic eye diseases. ^{21 22} Thrombospondin (TSP) is an extracellular matrix glycoprotein that inhibits angiogenesis both in vitro and in vivo. ^{23 24 25} TSP-1 has been described as an inhibitor of angiogenesis, as it blocks the formation of new blood vessels in the cornea in vivo in response to basic FGF ²⁶ and blocks endothelial cell tube formation and cell migration in vitro. ²⁷  

N-(4-hydroxyphenyl) retinamide (4-HPR, fenretinide) is a synthetic analogue of all-trans-retinoic acid (ATRA) and is widely used as an anticancer agent. ^{28 29 30 31} It was reported recently that prolonged administration of 4-HPR to ABCA4 ^{− / −} mice decreased serum retinol levels and reduced toxic lipofuscin deposition, suggesting that 4-HPR is a potential therapy for nonexudative AMD. ³² This drug is currently in phase II clinical trials for treatment of geographic atrophy in patients with AMD (clinicaltrials.gov/ identifier: NCT00429936). 

Whether 4-HPR possesses pro- or antiangiogenic properties in noncancerous tissues and cell types is controversial. Although several studies have revealed that retinoic acid and its derivatives modulate angiogenesis, results have been conflicting, and both stimulatory ^{33 34 35} and inhibitory ^{36 37 38} activities of these compounds have been found. A recent study reports that ATRA induces angiogenesis via the retinoic acid receptor (RAR) by stimulation of endothelial cell proliferation and increased signaling of growth factors such as VEGF, hepatocyte growth factor (HGF), and Ang-2. ³⁵ In ARPE-19 cells, Chen et al. ³⁹ found that relatively low concentrations of 4-HPR induce neuronal differentiation, whereas high 4-HPR doses cause cell shrinkage and death. ⁴⁰ These studies also showed that 4-HPR-induced cell death in ARPE-19 cells is RAR mediated. ⁴⁰ The purpose of the present study was to evaluate the effect of 4-HPR treatment in a laser-induced murine CNV model, to assess its angiogenic properties, and to delineate the role of 4-HPR-induced cell death in cultured human RPE cells. 

Materials and Methods

Animals

C57BL/6 male mice were purchased from the National Cancer Institute (Frederick, MD). Mice between 6 and 8 weeks old were fed the standard laboratory chow in an air-conditioned room equipped with a 12-hour light/12-hour dark cycle. All procedures were performed in compliance with the Keck School of Medicine Institutional Animal Care and Use Committee approved protocols and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. For all surgical procedures, the mice were anesthetized and the pupils were dilated with topical 1% tropicamide (Alcon, Fort Worth, TX). 

Treatment Protocol for 4-HPR

The 4-HPR used in this study was obtained from the National Cancer Institute Repository (Bethesda, MD) and was of GMP grade. 4-HPR was delivered to adult mice by intraperitoneal (IP) administration of either 0.2 mg per injection or 1 mg per injection in 0.4 mL of sterile carrier solution (12.5% ethanol, 12.5% hydrogenated castor oil [Cremophor EL; BASF, Parsippany, NJ], and 75% normal saline). Injections were made twice daily for 14 days after laser treatment. The two 4-HPR doses are referred to as low (0.2 mg) and high (1 mg). Control mice were injected with 0.4 mL carrier solution alone. The levels of 4-HPR in the blood and retina after IP treatment were measured by HPLC. ⁴¹  

Laser Photocoagulation

Diode laser photocoagulation (75-μm spot size, 0.1-second duration, 140 mW) was performed on both eyes of each mouse on day 0. ⁴² For the purposes of CNV fluorescein angiography (FA), CNV volume, and histologic analysis, three laser photocoagulation burns were delivered to the retina lateral to the optic disc, through a slit lamp, with a coverslip used as a contact lens. Only lesions in which a subretinal bubble or focal serous detachment of the retina developed were used for experiments. ⁴³  

Tissue 4-HPR and 4-MPR Levels by HPLC

On days 7 and 14 after the laser procedure, mice from the 4-HPR-treated groups and the untreated control group were euthanatized by CO₂ in the morning, 12 hours after receiving the last dose. Blood from killed animals was collected by cardiac puncture into a preheparinized 1-mL syringe, transferred to a 2-mL brown pre-heparinized tube (Eppendorf, Fremont, CA) and centrifuged for 10 minutes at 9000g. The supernatant (plasma) from two animals of the day-7 cohort and two animals of the day-14 cohort was pooled to obtain an adequate sample size for analysis and stored at −80°C until analysis. Three pooled samples of blood (i.e., from six animals/sample) were collected for each group. For analysis of 4-HPR in the retina, the posterior poles of the eyes (including the retina, RPE, and sclera) were isolated and stored at −80°C until analysis. Three mice served as the control for the day-7 animals and two as the control for the day-14 animals. A previously described HPLC method was used for the analysis of 4-HPR and its major metabolite 4-MPR (N-(4-methoxyphenyl)retinamide) from tissues, plasma, and cultured human RPE. ⁴¹  

Fluorescein Angiography

The effect of 4-HPR treatment on the development of CNV was evaluated by semiquantitative assessment of late-phase fluorescein angiograms, captured 3 minutes after IP injection of 0.1 mL of 2.5% fluorescein sodium (Akorn, Decatur, IL), as previously described. ²⁷ Leakage was defined as the presence of a hyperfluorescent lesion that increased in size with time in the late-phase angiogram. Angiography was graded in a masked fashion by two examiners using reference angiograms. Angiograms were graded as follows: 0, no leakage; 1, slight leakage; 2, moderate leakage; and 3, prominent leakage. ⁴³  

Histology

For histopathologic analysis, the eyes were enucleated and snap frozen. Sections (8 μm) were stained with hematoxylin and eosin (H&E), to assess the histology of retina with 4-HPR treatment, laser lesions, and subsequent CNV development. 

Choroidal Flatmount Preparation and CNV Volume Quantification

The eyes were enucleated at various times after laser photocoagulation and fixed with 2% paraformaldehyde for 1 hour at 4°C. Eyecups were obtained by removing the anterior poles and neurosensory retina and were washed three times in PBS. The remaining eyecups containing the RPE–choroid–sclera complex were incubated with blocking buffer (PBS containing 1% BSA and 0.5% Triton X-100) for 1 hour at room temperature. Fluorescence volume measurements were made by creating image stacks of optical slices within lesions. Flatmounts stained with 10 μg/mL FITC-isolectin B4 were visualized using the 20× objective of a scanning confocal microscope (model LSM510; Carl Zeiss Meditec, Inc., Thornwood, NY). The image stacks were generated in the Z-plane, with the confocal microscope set to excite at 488 nm and to detect at 505 to 530 nm. Images were further processed using the microscope’s system software (LSM; Carl Zeiss Meditec, Inc.), by closely circumscribing and digitally extracting the fluorescent lesion areas throughout the entire image stack. The extracted lesion was processed through the topography software to generate a digital topographic image representation of the lesion and an image volume. The topographic analysis program determines and displays the objects’ surface contours by detecting fluorescent signal from the top of the image stack and then measures everything under the surface to yield a final volume (square micrometers ± SD), which reflects the CNV fluorescence volume. 

Immunostaining for Cytokeratin

The eyecups were serially sectioned, and the middle of the laser lesion was determined. Frozen sections (8 μm) of the center of the lesions were air dried for 15 minutes before being fixed in acetone for 5 minutes. Slides were washed for 5 minutes in PBS and then placed in 0.3% hydrogen peroxide in PBS for 20 minutes. Sections were blocked for 15 minutes in 1% bovine serum albumin (BSA; Sigma-Aldrich, St. Louis, MO) in PBS. The samples were incubated with anti-pan cytokeratin antibody (Dako, Carpinteria, CA) at 1:300 dilution for 1 hour at room temperature. The slides were washed three times and incubated with biotinylated anti-rabbit secondary antibody at 1:200 dilution for 30 minutes. After additional washes, the slides were incubated with avidin-biotin-peroxidase complex (ABC; Vector Laboratories, Burlingame, CA) for 30 minutes at room temperature and washed in PBS, followed by another incubation in the dark for 6 minutes in the peroxidase substrate 3-amino,9-ethyl-carbazole (Vector Laboratories). The slides were counterstained with modified Mayer’s hematoxylin (Master Tech, Lodi, CA), washed with distilled water, and mounted (Gel/Mount; Biomeda, Foster City, CA). 

Cell Culture and Treatment

RPE cells were isolated from human fetal eyes obtained from Advanced Bioscience Resources, Inc. (Alameda, CA) and cultured in Dulbecco’s minimum Eagle’s medium (Fisher Scientific, Pittsburgh, PA) with 2 mM l-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin (Sigma-Aldrich), and 10% heat-inactivated fetal bovine serum (FBS; Irvine Scientific, Santa Ana, CA), as previously described. ⁴⁴ Third- to fourth-passage cells at >90% confluence were used in all experiments. RPE culture medium was switched to 0% FBS overnight and then replaced with medium containing 4-HPR (3, 5, or 10 μM) with 0%, 2%, and 5% FBS for 24 hours. A 5-mM 4-HPR stock in DMSO was prepared for serial dilution, whereas the control cells received vehicle alone. After the incubation period, the extracellular medium was collected for protein-secretion analysis, and the cells were used for mRNA and protein quantitation studies. 

Highly differentiated fetal human RPE cells were grown on polyester permeable supports (12 mm diameter; 0.4 μm pore size; Transwell; Corning Costar, Corning, NY) as described earlier ⁴⁵ with slight modifications. We cultured RPE cells on the filters in 10% FBS containing medium for 1 day and changed to 1% FBS for the remaining culture period. The integrity and resistance of the RPE monolayer were evaluated by measurement of transepithelial resistance (TER) ¹⁵ and staining for tight-junction proteins. ⁴⁵  

TUNEL Staining

Apoptosis (DNA fragmentation) was detected by the terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin nick end-labeling (TUNEL), method according to the manufacturer’s protocol (ApopTag peroxidase in situ apoptosis detection kit; Chemicon, Temecula, CA). In short, cells (differentiated RPE cells grown on polyester permeable supports [Transwell; Corning Costar] and RPE cells in chamber slides) were fixed in 1% paraformaldehyde solution and rinsed with PBS. After treatment with 3% H₂O₂ at room temperature for 5 minutes, the cells were incubated with TdT enzyme for 1 hour at 37°C in a humidified chamber. The DIG-labeled nucleotides incorporated into DNA breaks were detected by applying anti-digoxigenin conjugate and peroxidase substrate. 

Immunofluorescence for Activated Caspase-3

RPE cells grown in chamber slides and transdifferentiated RPE cells grown on filters after treatment with 4-HPR for 24 hours in 0% and 5% FBS-containing medium were fixed in 2% paraformaldehyde and blocked in 10% normal goat serum in PBS and 0.3% Triton-X for 60 minutes at room temperature. The samples were incubated overnight with anti-caspase-3 rabbit polyclonal antibody (Cell Signaling, Danvers, MA) at 1:500 dilution in PBS/Triton-X, followed by incubation with anti-rabbit FITC-labeled secondary antibody (Chemicon) for 30 minutes at room temperature. The samples were mounted with fluorescence mounting medium with DAPI (4′,6′-diamino-2-phenylindole) and viewed under a laser scanning confocal microscope (LSM; Carl Zeiss Meditec, Inc.). 

Expression of Pro- and Antiangiogenic Genes

Total RNA was isolated (TRIzol extraction protocol; Invitrogen, Carlsbad, CA), and treated with DNase (Ambion, Austin, TX), to remove contaminating genomic cDNA. Reverse transcription was performed with 1 μg total RNA, oligo(dT)₁₅ primer, and AMV reverse transcriptase according to the manufacturer’s protocol (Promega, Madison, WI). The PCR experiments were then performed on a thermocycler (model LC 480 light cycler; Roche Diagnostics, Indianapolis, IN), with SYBR Green (Roche Diagnostics) as the interaction agent. Each 20-μL PCR contained 5 μL cDNA template, 10 μL SYBR Green PCR master mix, and 0.5 μM of each gene-specific primer. Quantification analysis of mRNA was normalized with β-actin or GAPDH used as the housekeeping gene. The specificity of the PCR amplification products was checked by performing dissociation melting-curve analysis or by 1% agarose gel electrophoresis. Reaction conditions were as follows: 5 minutes at 95°C followed by 45 cycles of 5 seconds at 95°C, 5 seconds at 55°C, and 10 seconds at 72°C. The sequences of primers used for human VEGF-A were forward: 5′-CTA CCT CCA CCA TGC CAA GTG-3′, reverse: 5′-TGC GCT GAT AGA CAT CCA TGA-3′; VEGF-C forward: 5′-CTG CCG ATG CAT GTC TAA ACT G-3′, reverse- 5′-TCT TGT TCG CTG CCT GAC ACT-3′, PEDF forward: 5′-CGA CCA ACG TGC TCC TGT CT-3′, reverse: 5′-GAT GTC TGG GCT GCT GAT CA-3′; TSP-1 forward: 5′-CTG ATC TGG GTT GTG GTT GTA-3′, reverse: 5′-CCT GTG ATG ATG ACG ATG A-3′; Ang-1 forward: 5′-GCA AAT GTG CCC TCA TGT TA-3′, reverse: 5′-TAG ATT GGA GGG GCC ACA-3′; and Ang-2, forward: 5′-TGC AAA TGT TCA CAA ATG CTA A-3′, reverse: AAG TTG GAA GGA CCA CAT GC-3′. Relative multiples of change in mRNA expression was determined by calculation of 2^{−ΔΔC T}. Results are reported as the mean difference in relative multiples of change in mRNA expression ± SD. 

ELISA Analysis

At the end of 4-HPR treatment, the extracellular medium from control and treated nonpolarized RPE groups and the medium from the apical and basal compartments of the highly polarized RPE groups were collected and stored at −80°C until further analysis. VEGF-A (Quantikine; R&D Systems, Minneapolis, MN), VEGF-C (IBL Co., Tokyo, Japan), and PEDF (BioProducts, Middletown, MD) secretion from the RPE cells was measured according to the manufacturers’ protocols. In separate experiments, we measured the VEGF-A, VEGF-C, and PEDF cellular levels, as described previously. ¹⁵ Data derived from standard curves were expressed as picograms per milliliter for growth factors secreted into medium and as relative difference (x-fold) in growth factor protein relative to the untreated control in cellular homogenates. 

Western Blot Analysis of TSP-1

After 4-HPR treatment as just described, the supernatant was collected and concentrated to equal amounts through a filter device (Millipore, Billerica, MA). Subsequently, medium was subjected to Western blot analysis under reducing conditions on a 7.5% Tris-HCl gel (Ready Gel; Bio-Rad, Hercules, CA) and then transferred to PVDF membrane (Millipore). The membrane was blocked with 5% BSA in TBST for 1 hour at room temperature and incubated with antiTSP-1 mouse monoclonal antibody (Abcam, Cambridge, MA) at 1:250 dilution in 1% BSA overnight at 4°C. After incubation with horseradish peroxidase–conjugated goat anti-mouse secondary antibody in 5% BSA (Vector Laboratories), protein bands were detected by chemiluminescence (Pierce, Rockford, IL). To verify equal loading, the gels were stained with Coomassie stain (Bio-Safe Coomassie Stain; Bio-Rad) and the intensity of the bands was measured (Scion Image Beta software; Scion Corp., Frederick, MD). 

Statistics

All statistical analyses were performed with Student’s t-test. Data reported in the figures are expressed as the mean ± SD. P < 0.05 was considered statistically significant. 

Results

Plasma and Retina 4-HPR and 4-MPR Levels

Plasma 4-HPR levels of laser-treated or control mice were determined by HPLC on days 7 and 14 after laser injury. In low-dose 4-HPR-treated mice, the mean 4-HPR plasma level was 1.2 μM on day 7 and 1.7 μM on day 14 after laser treatment. In high-dose 4-HPR-treated mice, the mean plasma drug level increased to 13.4 μM on day 7 and 53 μM on day 14 (Table 1) . The level of 4-MPR, the chief metabolite of 4-HPR in plasma, also showed a dose-dependent increase, but the absolute levels were much lower than those of the parent drug on days 7 and 14. The levels of accumulation of 4-HPR in the posterior poles of mouse eyes after IP administration are also shown in Table 1 . In untreated mice used as the negative control, no extraneous interfering peak was detected in HPLC analysis verifying the specificity of the assay. 

Histologic Analysis of Mice Treated with 4-HPR Alone

To ensure that exposure to 4-HPR in the absence of laser irradiation did not cause spontaneous CNV, we performed histopathology in mice that did not receive laser treatment but were treated with low- and high-dose 4-HPR for 7 and 14 days. A previous study had demonstrated that treatment with 4-HPR results in no alterations in RPE or retinal histology at the light microscopic level. ³² Consistent with the previous report, 4-HPR treatment alone caused no adverse changes in retinal morphology at the light microscopic level and no evidence of spontaneous CNV (Fig. 1) . ³²  

Fluorescein Angiogram of Laser- and 4-HPR-Treated Mice

To determine the effect of 4-HPR on the development of laser-induced CNV, choroidal angiogenesis was evaluated by fluorescein angiogram (FA). Late phase FAs of both eyes in each mouse from the control and low- and high-dose groups were evaluated according to the grading system described in the Materials and Methods section. Representative examples of the angiographic pattern in low- and high-dose 4-HPR, as well as in control mice on days 7 and 14, are shown in Figure 2A . The 4-HPR-treated mice exhibited increased CNV and leakage compared with vehicle-treated mice. Mice treated with 4-HPR exhibited a significantly higher FA score than did the control animals on days 7 and 14 after laser photocoagulation (Fig. 2B) . There was no significant difference between the scores of the low- and high-dose 4-HPR groups. These data suggest that 4-HPR treatment was associated with enhanced choroidal angiogenesis. No spontaneous CNV lesions were noted in any of the mice. 

CNV Volume in Laser- and 4-HPR-Treated Mice

A choroidal flatmount analysis with fluorescein-conjugated isolectin B4 was used to assess the volume of the CNV lesions on day 7 after laser photocoagulation. Mice treated with 4-HPR demonstrated an enhanced lesion size compared with that in the control mice (Fig. 3A) . Quantitative measurement of the volume of CNV showed that low-dose 4-HPR treatment resulted in an approximate 48% increase in choroidal vascular volume compared with that in control mice 1 week after laser photocoagulation (Fig. 3B) . High-dose 4-HPR treatment resulted in an approximate 55% increase of CNV volume over that in the control animals (Fig. 3B) . Consistent with the FA score, no statistical significance was found between the high- and low-dose groups. 

Histologic Analysis of Laser Lesions from Mice Treated with 4-HPR

Histopathologic studies in mice that received both laser irradiation and 4-HPR treatment suggested an increase in lesion diameter and thickness on days 7 and 14 after laser photocoagulation (Fig. 4) . The laser lesions in 4-HPR-treated mice showed increased subretinal space, which was more apparent in the high-dose drug-treated mice than in the low-dose group (5/5 mice on day 7 and 6/7 on day 14, respectively). Although it is possible that the increased subretinal space is an artifact related to tissue sectioning, we suggest that the subretinal space contained increased fluid consistent with the associated accumulation of mononuclear cells within the space (Fig. 4) . In the control mice not treated with 4-HPR, only 3 of 12 showed a minimal increase in subretinal fluid, which was much less in magnitude when compared to the mice treated with 4-HPR (Fig. 4) . The eyes with prominent enlargement of the subretinal space also showed prominent disorganization of the retinal layers. 

Histologic Localization of RPE Surrounding the Laser Lesion

To determine whether 4-HPR treatment causes loss of RPE, we immunostained retinal sections from control mice and from low- and high-dose 4-HPR-treated mice with cytokeratin after laser photocoagulation (Fig. 5) . As expected, staining was abundant in the control tissue in the RPE layer. In the untreated laser-induced CNV lesion cytokeratin-positive cells lined its surface and were present within the lesion itself. With low-dose 4-HPR treatment for 14 days, there was discontinuity of the RPE overlying the lesion (Fig. 5B) . With high-dose 4-HPR treatment, the loss of RPE was more severe, and pigment-containing macrophages were found in the subretinal fluid, suggesting prior RPE cell death (Fig. 5C) . 

Serum-Dependent RPE Cell Death Induced by 4-HPR Treatment

A cytotoxicity assay of human RPE cells incubated with 4-HPR in the different serum conditions showed that cell death induced by 4-HPR was serum dependent (data not shown). Consistent with this finding, TUNEL staining revealed more TUNEL-positive nuclei with increasing 4-HPR dose in 5% serum-containing conditions than in serum-free conditions (Figs. 6A 6B) . We further verified these results by staining RPE cells with an antibody that specifically detects active caspase-3 (Fig. 6C) . According to these analyses, the amount of active caspase-3 increased with 4-HPR treatment and was significantly (P < 0.01) higher in 5% FBS-containing medium than with serum-free medium (Fig. 6Ca 6Cb 6Cc 6Cd 6Ce 6Cf 6Cg 6Ch 6D) . These findings were also confirmed by Western blot analysis for cleaved caspase-3 expression (Fig. 6E) . To determine the relationship between intracellular drug levels and extent of cell death, we measured the intracellular 4-HPR levels in serum-containing and -free conditions (Fig. 7) . The drug levels were higher in serum-containing than in serum-free conditions with all 4-HPR doses, and the increase was statistically significant (P < 0.05) at 10 μM. 

To determine whether 4-HPR induces apoptosis in highly differentiated RPE, we performed parallel experiments on RPE cells grown on polyester permeable supports (Transwell; Corning Costar). Polarization of the filter-cultured RPE monolayer was confirmed by measurement of TER; mean resistance measured 359 ± 17.9 Ω/cm² and varied <5% after a 24-hour 4-HPR treatment. Only very rare TUNEL⁺ or active caspase-3⁺ cells were seen when the polarized RPE was treated with 5 μM 4-HPR in 0% or 5% serum for 24 hours (Figs. 8A 8B) . Similarly, Western blot of 4-HPR-treated polarized cultures showed undetectable active caspase-3 (Fig. 8C) . 

Regulation of Pro- and Antiangiogenic Factors with 4-HPR Treatment

To examine whether 4-HPR treatment alters gene expression of pro- and antiangiogenic factors in nonpolarized RPE cultures, the relative abundance of mRNA of VEGF-A, VEGF-C, Ang-1, Ang-2, TSP-1, and PEDF was determined by real-time RT-PCR. The gene expression of proangiogenic VEGF-A increased as a function of 4-HPR dose (Fig. 9A) . The increase in VEGF-A mRNA with 5 μM 4-HPR was 2.4-fold compared with that in untreated control cells (P < 0.05). The trend was similar for VEGF-C mRNA, where treatment with 3 and 5 μM 4-HPR significantly (P < 0.05) upregulated its gene expression versus that in the untreated control cells. Similar to VEGF-A, the maximum increase in gene expression was observed for VEGF-C at 5 μM 4-HPR (Fig. 9A) . Among the angiopoietins, only Ang-1 showed a significant (P < 0.01) dose-dependent increase in gene expression, whereas no significant change in the transcript level of Ang-2 was observed (Fig. 9A) . The antiangiogenic factors PEDF and TSP-1 showed a significant downregulation (P < 0.05) in gene expression in 4-HPR-treated versus untreated control cells (Fig. 9A) . 

The effect of 4-HPR on pro- and antiangiogenic cellular protein expression and secretion from RPE was then assessed. The expression of VEGF-A protein showed a dose-dependent increase with 4-HPR treatment, as did its secretion into supernatants of RPE cultures (Figs. 9B 9C) . Analysis of VEGF-C revealed that the secretion into RPE supernatants increased significantly; however, the cellular protein levels remained unaltered compared with control levels (Figs. 9B 9C) . On the other hand, cellular protein expression and secretion of the antiangiogenic factor PEDF decreased significantly as a function of 4-HPR dose (Figs. 9B 9C) . The decrease in gene expression of TSP-1 with 4-HPR treatment was also reflected in the significant (P< 0.05) decrease in TSP-1 protein secretion (Fig. 9D) . The VEGF/PEDF mRNA and protein ratio increased significantly (P< 0.05) with 4-HPR treatment (Fig. 10) . 

To determine whether 4-HPR alters secretion of VEGF and PEDF in highly differentiated RPE, we performed parallel experiments on polarized RPE cells grown on polyester permeable supports (Transwell; Corning Costar). Measurement of apical and basal secretion of VEGF and PEDF revealed predominantly basolateral secretion of VEGF and apical secretion of PEDF, and these values were not significantly altered after a 24-hour incubation with 5 μM 4-HPR (results not shown). 

Discussion

4-HPR (fenretinide) is a synthetic analogue of ATRA that is currently in phase II clinical trials for treatment of geographic atrophy in patients with AMD. ³² In other systems, there is controversy as to whether 4-HPR possesses pro- or antiangiogenic properties. ^{33 34 35 36 37 38} It is therefore important to determine the effect of 4-HPR on choroidal angiogenesis, to determine the potential for induction or exacerbation of CNV in patients with AMD. In the present study, prolonged 4-HPR treatment enhanced choroidal angiogenesis in a laser-induced murine model of CNV. CNV lesions from the 4-HPR-treated animals were larger, which suggested the possibility that 4-HPR promotes a proangiogenic growth factor environment. Indeed, analysis of 4-HPR-treated RPE cultures demonstrated a dysregulation of the growth factor environment, with increased expression and/or secretion of proangiogenic growth factors (VEGF-A, VEGF-C, and Ang-1) and decreased expression and/or secretion of antiangiogenic growth factors (PEDF and TSP-1). CNV lesions from 4-HPR-treated animals also showed increased fluorescein leakage, particularly at day 14 after the laser procedure, indicating increased breakdown of the blood–retina barrier. Although this may be due in part to the increased expression of vasogenic proteins such as VEGF, histology of the lesions also revealed decreased a number of cytokeratin⁺ RPE cells overlying the CNV lesions, suggesting that attempts of stromal transdifferentiated RPE to re-establish the outer blood–retinal barrier had been hindered. This possibility was explored further by evaluation of 4-HPR-induced RPE apoptosis in vitro. 4-HPR induced apoptosis in nonpolarized RPE cultures. Of importance, the induction of apoptosis occurred only in the presence of serum, consistent with a loss of RPE in CNV lesions with breakdown of the blood–retina barrier, but not in the normal RPE monolayer. 

Support for the proangiogenic role of 4-HPR is provided by a recent study showing that retinoic acid increases bovine aortic endothelial cell proangiogenic behavior in vitro and in vivo via enhanced RARα-dependent FGF-2 production. ³⁴ In human microvascular endothelial cells, ATRA and 9-cis retinoic acid stimulate capillary tube formation. ³³ Another study performed in HUVEC (human umbilical vascular endothelial cells) and NHDF (normal human dermal fibroblast) cells found that ATRA induced in vitro angiogenesis by enhancement of VEGF signaling and by upregulation of RAR. Further, tube formation induced by ATRA was completely blocked by VEGF neutralizing antibody. With dosages similar to those that we used, these reaearchers found increased VEGF secretion. Of note, the retinoic acid effect was due to VEGF-R2 but not VEGF-R1 in VEGF signaling. ³⁵  

The safety of 4-HPR with respect to normal tissues has been demonstrated in multiple studies. The anticancer effects of 4-HPR are well known—hence, its extensive use in cancer therapy. ^{46 47} Clinical studies in patients with cancer have shown that 4-HPR is well tolerated and is devoid of any serious adverse side effects, even during chronic exposure. ^{48 49} However, one 5-year chemoprevention trial in patients with breast cancer reported decreased night vision caused by plasma retinol depletion as the major clinical toxicity symptom. ⁵⁰ Recently, Radu et al. ³² found no deleterious effects of 4-HPR on the retina in their electrophysiologic and morphologic analysis. Consistent with this report, we found no significant histologic abnormalities in the retinas of mice treated with 4-HPR for 7 or 14 days. It should be noted that the dosage used in Radu et al. ³² in ABCA4 ^{− / −} mice (10 mg/kg, 28 days) corresponds approximately to the lower dose used in our present study in C57BL/6 mice (0.2 mg per injection) and is about five times lower than the high dose that we used (1 mg per injection). The dosage regimen we used in our mouse model studies is similar to that used in cancer treatment protocols in humans, where 4-HPR was given orally for 1 week every 3 weeks, resulting in a median plasma level of 6 to 13 micromolar, with a range of 3 to 20 micromolar. ⁵¹ To our knowledge, this is the first report on quantification of 4-HPR in retinal tissue. The levels of 4-HPR and 4-MPR were much higher in the retinal tissue than in plasma, suggesting facile penetration of the blood–retina barrier by 4-HPR in our experimental conditions. 

Our study demonstrates that while 4-HPR apparently does not affect the histology of normal retina, there is selective loss of stromal RPE in CNV lesions. 4-HPR has been shown to possess apoptotic activity in a variety of transformed cell lines ^{52 53} ; however, there have been only limited studies in normal cells. ^{40 54} Our studies revealed that, in early passage, nonpolarized human RPE cultures the apoptotic effect of 4-HPR was associated with caspase-3 activation, consistent with a recent report in ARPE-19 cells. ⁴⁰ Of interest, we found that 4-HPR induced apoptosis in nonpolarized human RPE cells in the presence of serum and that the apoptosis was minimal under serum-free conditions. This result suggests that the growth-promoting effects of serum may be critical for the induction of apoptosis by 4-HPR and may explain why the RPE layer away from the CNV lesion was resistant to cell death. To investigate the resistance of the normal RPE monolayer to 4-HPR-induced apoptosis further, we evaluated the differentiation state of the RPE. In the normal monolayer, RPE cells are highly differentiated and polarized, whereas in the CNV lesion, stromal RPE are transdifferentiated and nonpolarized. When we tested highly polarized RPE monolayers grown on polyester permeable supports (Transwell; Corning Costar), we found that they were highly resistant to apoptosis induced by 4-HPR. Thus, both the presence of serum and a nonpolarized differentiation state of the RPE play a role in promoting 4-HPR-induced RPE apoptosis within CNV lesions compared with the resistance of RPE to apoptosis in the intact monolayer. 

Neovascularization is a complex multistep biological process, and each of these steps is regulated by the delicate balance of a variety of agonistic and antagonistic effector molecules. ⁵ Our study found that 4-HPR treatment of nonpolarized RPE cultures significantly increased expression and/or secretion of proangiogenic growth factors (VEGF-A, VEGF-C, and Ang-1) and decreased expression and/or secretion of antiangiogenic growth factors (PEDF, TSP-1). The upregulation of VEGF mRNA, cellular protein, and secretion is consistent with that found in another study in which retinoic acid treatment of RPE and Y79 cells significantly induced VEGF mRNA and protein secretion. ⁵⁵ 4-HPR treatment of cultured RPE resulted in significant upregulation of Ang-1 but not of Ang-2. Of note, in subfoveal membranes surgically removed from patients with AMD, Ang-1 mRNA was found to be significantly higher than Ang-2 mRNA. ¹⁷ Ang-1 stimulates the tyrosine kinase activity of the Tie-2 receptor and enhances blood vessel maturation. ²⁰ PEDF is the most potent inhibitor of angiogenesis that is expressed in the RPE and its mRNA expression and cellular protein expression and secretion were significantly downregulated in RPE treated with 4-HPR. The ratio of VEGF to PEDF has been suggested to be an indicator of a dysregulated growth factor balance that can be found in angiogenic ocular disorders. ²³ In our study, the ratio of VEGF to PEDF was significantly increased at both the mRNA and cellular protein levels. TSP is an extracellular matrix glycoprotein that inhibits angiogenesis both in vitro and in vivo, with both direct and indirect effects on endothelial cells. ^{23 24 25 56 57 58 59 60 61} Of the five known subtypes of TSPs, TSP-1 is the most common, and several cell types, including RPE cells, ⁵⁷ produce TSP-1. In the context of eyes with AMD, TSP-1 immunoreactivity has been reported to be significantly decreased, especially in Bruch’s membrane and the choriocapillaris in the submacular region, ⁶² and thus impaired expression of TSP-1 in AMD may promote CNV formation in AMD. 

We did not find any evidence of spontaneous CNV in mice treated with 4-HPR, suggesting that alterations in growth factor expression may be localized to the site of the laser lesions. Indeed, when we examined highly differentiated, polarized RPE cells for dysregulated growth factor expression after treatment with 4-HPR, we found that there was no significant effect on the secretion of VEGF or PEDF into basal or apical compartments. Future studies evaluating the effects of serum, RPE differentiation status, and laser injury on RAR signaling in RPE will be of interest. 

We conclude that the angiogenic effects of 4-HPR are highly context-dependent both in vitro and in vivo. Nonpolarized RPE are preferentially susceptible to 4-HPR induction of RPE apoptosis and of a proangiogenic growth factor phenotype. Similarly, in vivo studies revealed that 4-HPR augmented laser-induced CNV lesions but did not result in spontaneous CNV lesions or histologic damage to normal retina. These results warn against the use of 4-HPR in patients with AMD with CNV and suggest that although the polarized RPE monolayer appears to be resistant, patients with geographic atrophy being treated with 4-HPR should be evaluated for development of CNV. 

²  Contributed equally to the work and therefore should be considered equivalent authors.

 Supported by The Arnold and Mabel Beckman Foundation, National Institutes of Health Grants (EY01545, EY03040), Research to Prevent Blindness.

 Submitted for publication June 4, 2007; revised November 26, 2007; accepted February 11, 2008.

 Disclosure: P.G. Sreekumar, None; J. Zhou, None; J. Sohn, None; C. Spee, None; S.J. Ryan, None; B.J. Maurer, None; R. Kannan, None; D.R. Hinton, 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.

 

Plasma and Retinal Tissue Levels of 4-HPR and 4-MPR in Mice after 4-HPR Treatment

	4-HPR	4-MPR
Plasma (μM)
Control	ND	ND
Low dose
Day 7	1.2 ± 0.14	0.2 ± 0.1
Day 14	1.71 ± 0.42	0.3 ± 0.14
High dose
Day 7	13.4 ± 1.8	4.03 ± 0.95
Day 14	53.0 ± 22.3	13.3 ± 4.5
	Day 7	Day 14
4-HPR retinal tissue (μg/g)
Control	ND	ND
Low dose	11.7 ± 0.7	11.6 ± 0.85
High dose	29 ± 7.8	91.2 ± 47.4

 Data are expressed as the mean ± SD (n = 3). ND, not detectable.

Figure 1.

 

Representative retinal histology in mice treated with low (C, D) or high (E, F) doses of 4-HPR compared with control mice treated with carrier solution (A, B). Tissue sections were stained with H&E. No significant morphologic changes were observed on day 7 or 14 of 4-HPR treatment.

Figure 2.

 

FA analysis of CNV. (A) Representative late-phase FAs of 4-HPR-treated mice and control mice on days 7 and 14 after laser photocoagulation. (B) Comparison of semiquantitative CNV FA score between low- and high-dose 4-HPR-treated (n = 12/group) and control mice treated with carrier solution alone (n = 24). *P < 0.05 versus the untreated control.

Figure 3.

 

4-HPR treatment resulted in increased volume of CNV. (A) The topographic representation of an FITC-isolectin B4–stained CNV lesion is shown from 4-HPR-treated and control mice at day 7 after laser photocoagulation. (B) Comparison of CNV volume between 4-HPR-treated and control mice. n = 5 per group. *P < 0.05 versus the untreated control.

Figure 4.

 

Histology of representative CNV lesions of low- and high-dose 4-HPR-treated and control mice at days 7 and 14 after laser photocoagulation. Tissue sections were stained with H&E (n = 6 per group). Arrows: subretinal space.

Figure 5.

 

Immunoperoxidase staining for pancytokeratin in 4-HPR-treated and control mice 14 days after laser treatment. In the control sections, the cytokeratin⁺ RPE (arrows) covered the surface of the CNV lesion. In 4-HPR-treated lesions, the arrows mark zones where the lesions were not covered by cytokeratin-positive RPE. (*) The appearance of pigment-containing cells in the subretinal space.

Figure 6.

 

Effect of 4-HPR on RPE apoptosis in vitro. Human RPE cells were treated with 4-HPR in different serum conditions, and apoptosis was assessed by TUNEL and cleaved caspase-3 staining. (A) A much higher proportion of 5 μM 4-HPR-treated cells were apoptotic by TUNEL staining in 5% serum-containing medium but not in serum-free conditions. (B) Quantification of the dose response for 4-HPR treatment is shown (5 μM 4-HPR in serum versus control, P < 0.01). Activation of caspase-3 by immunofluorescence confocal microscopy is shown in Ca–Ch). Green: cleaved caspase (a, c, e, g); blue: nuclear staining (b, d, f, h). (a, b) RPE cells treated with vehicle alone in 5% FBS; (c, d) cells treated with 4-HPR in serum-free FBS containing medium; (e, f) RPE cells treated with vehicle alone in 5% serum-containing medium; (g, h) cells treated with 4-HPR in 5% serum-containing medium. Quantification by immunofluorescence of 4-HPR-induced cell death through activation of caspase-3 showed significantly higher apoptosis with 5% FBS (D). All values are compared to corresponding control data and are expressed as the percentage of positive versus control cells. Western blot analysis of active caspase-3 (E) also confirmed higher expression of cleaved caspase with 5 μM 4-HPR in 5% serum. **P < 0.01 versus the control.

Figure 7.

 

Intracellular 4-HPR levels in nonpolarized RPE cells treated with 4-HPR. RPE cells were treated with various doses of 4-HPR for 24 hours in different serum conditions, and intracellular 4-HPR levels were measured. Intracellular HPR levels increased significantly with 5% serum and 10 μM 4-HPR. Data are representative of the results in three independent samples. *P < 0.05 versus the serum-free medium.

Figure 8.

 

Effect of 4-HPR on apoptosis of RPE in highly polarized monolayers. RPE cells were grown on polyester permeable supports inserts for 4 weeks as highly polarized monolayers. The cells were then starved in serum-free medium overnight and treated with 5 μM 4-HPR for 24 hours. Analysis of apoptosis and activation of caspase-3 was performed as described in Figure 6 . No significant cell death was observed in highly polarized RPE cells in response to 4-HPR (A). Caspase-3 activation in the same experimental conditions was also negligible in serum-free (a, b) and 5% serum containing (c, d) conditions (B). Western blot analysis for cleaved caspase-3 in 4-HPR-treated RPE on filter inserts did not show any visible bands at 17 and 19 kDa (C).

Figure 9.

 

Gene expression, cellular protein expression and secretion of key angiogenic and antiangiogenic factors in RPE cells exposed to 4-HPR for 24 hours. mRNA expression (A) and protein levels (B–E) were determined by real-time PCR, ELISA, and Western blot analysis. Quantification analysis of mRNA in (A) was normalized to β-actin or GAPDH as internal controls and untreated control data were taken as 1. (B) VEGF-A, VEGF-C, and PEDF cellular protein levels were measured by ELISA. Protein results were normalized versus the untreated control cells and the values are expressed as multiples of change. (C) VEGF-A, VEGF-C, and PEDF secretion measured by ELISA. *P < 0.05 versus the control. (D) TSP-1 protein expression in medium was quantified by Western blot analysis. A 43% decrease (P < 0.05) with 5 μM 4-HPR in TSP-1 protein secretion was observed by densitometry analysis. Data are representative of the results in three independent experiments. *P < 0.05 versus the control.

Figure 10.

 

Increase in the VEGF-A-to-PEDF mRNA (A) and protein (B) ratios in RPE cells after 4-HPR treatment (n = 3). *P < 0.05 versus the control.

The authors thank Patricia Becerra of the National Eye Institute for advice on PEDF assays, Fernando Gallardo for help in the laser treatment studies, and Ernesto Barron for technical assistance.