Bevacizumab Reduces Neurocan Content and Gene Expression in Newborn Rat Retina In Vitro

Abstract

Purpose.: Extracellular matrix (ECM) and cellular membrane proteoglycans (PGs) play important roles in neural differentiation and cell adhesion. Vascular endothelial growth factor, an important signal protein in vascular and retinal neural cell development, is retained in the ECM due to its high affinity for PG. Bevacizumab, an anti-VEGF agent, has been extensively used for treating retinal diseases in adult and newborn patients, although its effect on the developing retina remains largely unknown. The purpose of this study was to investigate the effect of bevacizumab on neurocan, phosphacan, and syndecan-3 PG levels in newborn rat retina.

Methods.: Retinal explants of sixty 2-day-old Lister hooded rats were obtained after eye enucleation and maintained in culture media with or without bevacizumab for 48 hours. Immunohistochemical staining was assessed against neurocan, phosphacan, and syndecan-3. Proteoglycan content was quantified based on the intensity of immunohistochemical labeling. Gene expressions were quantified by real-time reverse-transcription polymerase chain reaction. The results from the treatment and control groups were compared.

Results.: No significant difference in the staining intensity and mRNA expression of phosphacan and syndecan-3 was observed between the groups. However, a significant decrease in neurocan content and mRNA expression was observed in bevacizumab-treated retinal explants compared with controls.

Conclusions.: Bevacizumab did not affect phosphacan and syndecan-3 levels but decreased neurocan content and gene expression. Therefore, it may interfere with early postnatal retinal cell differentiation. Although further studies are necessary to confirm our findings, we suggest anti-VEGF agents be used with caution in developing retinal tissue.

Introduction

Vascular endothelial growth factor (VEGF) is an important glycoprotein during vascular development influencing vessel remodeling, stabilization, and differentiation of endothelial cells ^1–3 and modulating the vascular component of pathological processes such as tumor growth and retinopathies. ^4–6 Vascular endothelial growth factor has also been shown to play a critical role in neurogenesis, with significant neurotrophic and neuroprotector activity in the nervous system, including the retina. ^7,8 The development of anti-VEGF agents such as bevacizumab, a humanized monoclonal antibody against all isoforms of VEGF, ^9,10 has allowed the use of intravitreous anti-VEGF antibody injection as an important therapy in the treatment of chronic vascular eye diseases in adults ^9,11–13 and, more recently, in the treatment of hyperoxia-induced vaso-obliteration followed by neovascular formation in developing eyes of premature infants with retinopathy of prematurity (ROP). ^14,15  

Since the adoption of anti-VEGF agents to treat retinal diseases, researchers have evaluated the possible consequences of exposure of retinal tissues to such compounds. Several animal studies using light microscopy and electrophysiological assessment of retinal function have failed to confirm retinal toxicity from bevacizumab administered intravitreally. ^16–20 On the other hand, other experimental in vivo and in vitro studies have shown ultrastructural abnormalities in the retina, including increased apoptotic activity ^21–24 and increased reactivity of glial cells ²⁵ in adult retinal tissues. Furthermore, Miguel et al. ²⁶ evaluated newborn rat retinal explants exposed to bevacizumab for 48 hours, and found increased vimentin labeling and a decrease of glial fibrillary acidic protein mRNA levels in treated tissue compared with controls, suggesting that the anti-VEGF activity of bevacizumab may interfere with glial cell maturation in the early stages of retinal development. 

During retinal development, cellular migration, dendrites, and axon outgrowth occur through the extracellular matrix (ECM). The ECM in the central nervous system is largely devoid of cells and is filled with a network of glycoproteins, proteoglycans (PGs), and hyaluronan. ²⁷ Proteoglycans are components of the ECM and the cell membrane that regulate numerous cellular processes such as adhesion and proliferation, differentiation, induction of neuritis, and neural network formation. ^28,29 Cell membrane- and ECM-bound PGs serve as coreceptors for growth factors such as VEGF, which adheres by way of its heparin-binding domain, providing a reservoir of biologically active VEGF. ³⁰ Neurocan is a nervous tissue-specific PG and one of the most important chondroitin sulfate PGs in the brain. ³¹ It is found throughout the retinal layers of rats during early development. ³² Other important PG molecules include phosphacan, a chondroitin sulphate PG, and syndecan-3 (SDC-3), a heparin sulphate PG. The latter promotes neurite outgrowth from retinal neuronal cells when the molecule binds to VEGF or basic fibroblastic growth factor and communicates with the cytoskeleton to enhance neurite outgrowth. ^33,34  

The important interaction of VEGF with PG in the developing retina raises the possibility of adverse effects of anti-VEGF agents on the ECM and therefore on early neurites and other retinal cell components. No previous study, however, has evaluated the effect of anti-VEGF agents on PG in the early developing retina. The purpose of this study was therefore to evaluate in vitro the influence of bevacizumab on neurocan, phosphacan, and syndecan-3 levels and gene expression in newborn rat retinas. 

Materials and Methods

All procedures were performed in accordance with the recommendations of the Guide for the Care and Use of Laboratory Animals, as described in the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research, and by the National Institutes of Health, and approved by the ethics committees of animal research of our institutions. 

Materials

Eagle's basal medium (BME), HEPES, penicillin G, streptomycin sulfate, and L-glutamine were obtained from Sigma-Aldrich Corp. (St. Louis, MO, USA). Bevacizumab was from Genentech, Inc. (Avastin; San Francisco, CA, USA). The RNA extraction kit (RNeasy Mini Kit) and the PCR kit (Rotor Gene SYBR Green PCR Kit) were obtained from Qiagen (São Paulo, Brazil). The reverse transcriptase kit (SuperScript II Reverse Transcriptase Kit) was obtained from Invitrogen (Carlsbad, CA, USA). Primary mouse antibodies for neurocan, phosphacan, and syndecan-3 were from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA); Fluorescent secondary antibodies, Alexa 488 rabbit anti-goat IgG (for neurocan and phosphacan immunostaining) were from Invitrogen (Carlsbad, CA, USA) and donkey anti-goat IgG (for syndecan-3), were purchased from Abcam (Cambridge, UK). Blue staining DAPI for cell nuclei labeling were from Invitrogen. All other reagents were of analytical grade. 

Animals

Sixty newborn 10-g Lister hooded rats were used. The matrices were kept under a 12-hour light/dark cycle with free access to water and food while nursing. 

Retinal Tissue Culture

The preparation of rat retinal explants has been described elsewhere. ³⁵ Two-day-old rats were killed by decapitation, the eyeballs were removed, and the retinas dissected free from scleral tissue and RPE using a calcium- and magnesium-free balanced salt solution (CMF). The retinas were cut into fragments of approximately 1 mm² in complete BME supplemented with 2 mM L-glutamine, 100 μg/mL streptomycin sulfate, 100 U/mL penicillin G, 20 mM HEPES, and 5% fetal calf serum. The retinal explants were maintained in 25-mL tight-lidded Erlenmeyer flasks and incubated at 37°C, in an atmosphere of 5% CO₂ and 95% air in an orbital shaker at 80 to 90 rpm. After dissection, some explants were assigned as controls while the remainder received culture medium, with or without the addition of 0.5 mg/mL bevacizumab. After 48 hours, retinal explants were either fixed in paraformaldehyde for immunohistochemistry or immersed in RNA stabilization agent (RNAlater; Qiagen) for quantitative real-time PCR analysis (qRT-PCR). 

Histological Procedures

The retinal explants were fixed by immersion in 4% paraformaldehyde in phosphate buffer, 0.1 M, pH 7.4, for 1 hour, followed by immersion in a 30% sucrose solution in phosphate buffer. Tissue orientation was verified under a stereoscopic microscope. Cross-sections that were 10-μm thick were prepared with a cryostat (Leica CM3000; Leica Microsystems, Wetzlar, Germany), mounted on poly-L-lysine-coated slides, and maintained at −20°C until processing. The orientation procedure ensured that sectioning through the tissue was almost exactly orthogonal to the surface of the retina. 

Immunohistochemistry

For immunodetection, sections were incubated with 0.1% Triton X-100 in PBS, washed once with PBS/Triton, and treated for 60 minutes at room temperature with a blocking solution (10% bovine serum albumin [BSA]) diluted in PBS. After blocking, sections were incubated overnight at 4°C with primary polyclonal antibodies at 1:100 dilution (anti-neurocan: C-12 clone, anti-phosphacan: C-19 clone PTPζ, and anti-syndecan-3: sc-1110 clone; all purchased from Santa Cruz Biotechnology, Inc.) in 5% BSA diluted in PBS. Further processing was done by incubation for 2 hours at room temperature with specific fluorescent secondary antibody at 1:400 dilution (donkey anti-goat IgG-FITC: sc-2024, Santa Cruz Biotechnology, Inc.) diluted in 1% PBS-BSA. After the reactions, preparations using fluorescent-conjugated secondary antibodies were mounted in mounting medium (Fluoromount Aqueous Mounting Medium; Sigma-Aldrich Corp.) and conserved at 4°C in a darkened refrigerator. Negative control sections were processed in the absence of primary antibodies (data not shown). 

The sections were observed, first under a conventional fluorescence microscope (Zeiss Axioskop 2 Plus; Carl Zeiss Optical, Inc., Baltimore, MD, USA), then under a confocal microscope (Leica TCS-SP5; Leica Microsystems) at 543-nm wavelength of excitation for the fluorochrome used. All images were acquired at a resolution of 512 × 512 pixels, using 75% laser transmittance, 0.7-μm optical slices, and 1-minute scanning-time in the same value of the detector gain. Negative and positive controls for the reactions were prepared (data not shown). Sections containing the central or peripheral retinal areas from treated and control eyes were evaluated both under the light microscope (×63 objective) and the confocal microscope. Three fields of each section of the retina (nine fields of treated and control retinas) were chosen at random. 

Real-Time Polymerase Chain Reaction (RT-PCR)

For quantitative RT-PCR, total RNA was isolated from rat retinal explants using an RNA extraction kit (Qiagen) according to the manufacturer's protocol. Subsequently, cDNAs were generated from 2 μg total RNA using a reverse transcriptase kit (Invitrogen). The Table shows the primer sequences used for RT-PCR amplification. Following the manufacturer's PCR kit (Qiagen) protocol, the resulting cDNA was submitted to a 40-cycle PCR amplification in order to quantify neurocan, phosphacan, and syndecan-3 transcripts. For each gene analysis, three replicates of the acidic ribosomal phosphoprotein (ARBP) gene were run. We used ARBP as an endogenous reference gene due to the absence of significant expression abnormalities from all groups (data not shown). 

Table.

 

 Primer Sequences, Expected Product Length and GenBank Accession Number Used in RT-PCR Analysis

Gene	Primer Sequences		Length	Accession Number
Neurocan	Sense:	5′-ACC TGG TAA CCC TGG AAG TGA-3′	77 bp	NM_031653.1
	Antisense:	5′-AGC GAA GGT CAA CGC ATA GC-3′
Phosphacan	Sense:	5′-TGG GAC TGA AAG TGT TTA GTC GTT TT-3′	78 bp	NM_001170685.1
	Antisense:	5′-GGA GTT TGG CAG GAG GTT CTG-3′
SDC-3	Sense:	5′-CTT GGC CTC CAC GAC AAT-3′	84 bp	NM_053893.3
	Antisense:	5′-GCA CCT CCT TCC GCT CTA AGT-3′
ARBP	Sense:	5′-ACATTTGTCTACATTGGTACTCT-3′	89 bp	NM_098898.2
	Antisense:	5′-GGCGCCCTATATACGATGAC-3′

 bp, base pairs.

Adequate standard curve slopes for each gene indicated efficient amplification of neurocan, phosphacan, and syndecan-3. The relative quantification method (ΔΔCt) was used, with the ratio of target mRNA, normalized with respect to ARBP mRNA and relative to a calibrator sample using high-resolution melting analysis software (Rotor-Gene ScreenClust HRM; Qiagen). 

Statistical Analysis

The results from the two groups (control and treatment) were expressed as median and interquartile range (IQR) and compared using the Mann-Whitney U test. The level of statistical significance was set at 5% (P < 0.05). 

Results

Our retinal explants displayed adequate histological organization with nuclear layers alternating with plexiform layers. Within the period of culture (48 hours), no noticeable changes occurred in either general structure or lamination of the retinal tissue as compared to the retina in situ (data not shown), similarly to what was observed in previous studies. ³⁶ The maintenance of an adequate histological organization is a strong indication of the validity of the study model. 

The immunohistochemical findings for the two groups (bevacizumab-treated and untreated) with regard to neurocan, phosphacan, and SDC-3 levels are displayed in Figure 1. Figure 2 shows quantitative immunohistochemical findings, with box plot graphs for each group. No significant difference in labeling was observed between the groups with regard to phosphacan (median = 0.26 and IQR = 0.32 in the study group versus median = 0.48 and IQR = 1.59 in the control group; P = 0.361) or syndecan-3 (median = 0.36 and IQR = 0.36 in the study group versus median = 0.24 and IQR = 0.25 in the control group; P = 0.075). However, neurocan antibody levels were significantly lower in bevacizumab-treated retinas (median = 0.16, IQR = 0.46) than in controls (median = 1.41, IQR = 1.74; P = 0.028). 

Figure 1

 

Confocal laser scanning photomicrographs of bevacizumab-treated and control retinal explants stained (in green) immunohistochemically against: (A) neurocan, (B) phosphacan, and (C) syndecan-3. A significant reduction in immunohistochemical staining (in green) for neurocan in the bevacizumab-treated compared with the control group was observed while no significant difference occurred in both groups regarding phosphacan and syndecan-3 staining. GCL, ganglion cell layer.

Figure 2

 

Bar graphs for bevacizumab-treated and control retinal explants showing the intensity of immunofluorescence labeling antibodies against: (A) neurocan, (B) phosphacan, and (C) syndecan-3.

Figure 3 shows box plot graphs of the mRNA expression for neurocan, phosphacan, and syndecan-3. Bevacizumab treatment did not affect the mRNA expression for phosphacan (median = 0.53 and IQR = 0.11 in the study group versus median = 0.99 and IQR = 0.88 in the control group; P = 0.221) or syndecan-3 (median = 0.89 and IQR = 3.45 in the study group versus median = 0.07 and IQR = 0.05 in the control group; P = 0.05), with similar mRNA levels in the two groups. However, the neurocan mRNA content was significantly decreased in bevacizumab-treated retinas (median = 0.54, IQR = 0.20) compared with controls (median = 1.17, IQR = 0.28, P = 0.014). 

Figure 3

 

Bar graphs showing the results of mRNA transcripts of: (A) neurocan, (B) phosphacan, (C) syndecan-3 in bevacizumab-treated and control retinal explants.

Discussion

For many years bevacizumab has been used as a therapeutic choice for a wide array of retinal diseases, although its possible toxicity to the retina has been an issue of much concern, particularly when used in children. ^37,38 Despite that concern, studies in general have failed to show significant toxicity to the retina; on histologic or electrophysiological analysis of adult ^16–19,21 or juvenile ²⁰ animals or in cultured adult human retinal pigmented cells, ³⁹ human vascular endothelial cells, ³⁹ or adult rat retinal neurosensory cells. ³⁹  

However, other experimental animal studies have concluded that bevacizumab is not completely innocuous and can induce adverse structural and functional changes in retinal tissue. Inan et al. ²¹ found mitochondrial disruption in the inner segments of photoreceptors on electron microscopy and intense apoptotic expression in the retina after intravitreal bevacizumab injection in adult rabbits. Avci et al. ²² found that bevacizumab caused a significant increase in apoptotic activity in rabbit photoreceptor cells. Using TUNEL assays and caspase-3 immunostaining, Jee et al. ²⁴ observed significantly more apoptotic activity in bevacizumab-injected eyes of adult rabbits than in controls. Likewise, juvenile rabbits injected intravitreally with bevacizumab by Fusco et al. ⁴⁰ displayed more apoptotic activity and gliosis in the retina than untreated controls. 

Studies in vitro have also detected adverse effects of bevacizumab. Chen et al. ⁴¹ showed that treatment with bevacizumab induced an epithelial-to-mesenchymal transition in the lineage of human retinal pigment epithelium cells (ARPE 19) by upregulation of the connective tissue growth factor, a profibrotic factor associated with the process of fibrosis. ⁴² Using the same cell line, Schnichels et al. ⁴³ also observed a slightly enhanced effect on cell death after bevacizumab treatment. Cell death and reduced viability were also detected for rat ganglion RGC-5 cells and mouse photoreceptor 661W cell lines. 

Only one previous study has evaluated the effect of bevacizumab on early developing retinas. The subject is important since VEGF, which is expressed by astrocytes in the retinal tissue during its development, ^44,45 apparently acts as a neuroprotective and neurotrophic factor for retinal cells, influencing neuronal growth, differentiation, and survival. ^7,23 Thus, Miguel et al. ²⁶ observed changes in glial maturation with upregulation of vimentin and downregulation of GFAP mRNA in retinal explants of newborn rats treated with bevacizumab for 2 days in vitro compared with controls. 

In this study, we employed the same rat retinal explant model to investigate cell development based on the behavior of three ECM and cell membrane PGs. While phosphacan and syndecan-3 levels remained unchanged, neurocan content was significantly decreased on imunohistochemistry and mRNA content was reduced on RT-PCR in the bevacizumab-treated group compared to controls. Neurocan is a nervous tissue-specific molecule, a major constituent of PGs in the brain and also abundantly expressed in the developing rat retina. ³³ Inatani et al. ³² found that neurocan is expressed at early postnatal stages (postnatal day [P]0–P3) and increases progressively in the neural retina, reaching a peak on P7 and decreasing thereafter to a faint expression in the adult retina. The molecule is believed to have an inhibitory effect on neurite outgrowth from cultured neuronal cells ^46,47 and in postnatal rat retinal ganglion cells. ⁴⁸ It is therefore likely to play a regulatory role in the formation of the neural network. ⁴⁸ Neurocan is synthesized normally by neurons ⁴⁹ but also by reactive glial cells. ^50,51 Since Miguel et al. ²⁶ found changes in retinal glial maturation produced by bevacizumab, it is possible that such abnormality contributed to the reduced neurocan content observed in our study. 

The finding of reduced neurocan content and gene expression in rat retina exposed to anti-VEGF bevacizumab is also interesting since previous studies have demonstrated that transient retinal ischemia upregulates neurocan expression ⁵² and that retinal ischemia is associated with increased VEGF content in the retina. ^53,54 Because neurocan is a nervous tissue–specific PG, changes in neurocan expression may influence retinal damage and repair in eyes with transient ischemia. Since neurocan has an inhibitory effect on neurite outgrowth, ⁴⁸ it conceivably reduces neuronal plasticity thereby possibly preventing the development of abnormal neuronal networks. The significant reduction in neurocan content and gene expression observed in our study therefore suggests that bevacizumab may interfere in neurite proliferation by causing a reduction in the inhibitory effect of neurocan. However, further studies are necessary to confirm our findings and better understand how anti-VEGF agents modify neurocan content in the developing retina. 

In conclusion, our study suggests that bevacizumab influences the early development of retinal cells in 2-day-old Lister hooded rat explants by downregulating neurocan gene expression and reducing neurocan content. Bevacizumab should therefore be used with caution when treating developing eyes, especially in newborn premature infants. 

Acknowledgments

Supported by grants from Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP, Grant No. 2011/12271-3), São Paulo, Brazil; Rio de Janeiro State Foundation for the Advancement of Science (FAPERJ, Grant No. E-26/111.777/2012), Rio de Janeiro, Brazil; and Conselho Nacional de Pesquisa (CNPq, Grants Nos. 476392/2011-0 and 306487/2011-0). The authors alone are responsible for the content and writing of the paper. 

Disclosure: P.G. Krempel, None; M. Matsuda, None; M.V. Marquezini, None; T.G. Seixas, None; G.M. Ventura, None; A. Sholl-Franco, None; N.C. Miguel, None; M.L. Monteiro, None