Protective Effects of β1/2 Adrenergic Receptor Deletion in a Model of Oxygen-Induced Retinopathy

Abstract

Purpose.: β-adrenergic receptors (β-ARs) regulate angiogenesis in proliferative retinopathies. We studied the effects of β1/2-AR deletion in a model of oxygen-induced retinopathy (OIR) to confirm the role of β1- and/or β2-ARs in regulating angiogenesis and to get insights into the role of β3-ARs.

Methods.: Mice with β1/2-AR deletion (KO) were used. Levels of norepinephrine (NE), β3-ARs, transcription, and proangiogenic factors were evaluated. Retinas were analyzed for avascular area and neovascular tufts in the superficial plexus. Deep plexus and blood–retinal barrier (BRB) were also analyzed. Neovascularization, proangiogenic factors, protein kinase A (PKA) activity, and nitrite production were assessed after BRL 37344, a β3-AR agonist.

Results.: Oxygen-induced retinopathy was characterized by NE upregulation with higher levels in wild type (WT) than in KO. Wild type and KO displayed comparable levels of β3-ARs, transcription, and proangiogenic factors, but differed in VEGF receptor (VEGFR) expression with VEGFR-1 in WT lower than in KO and VEGFR-2 in WT higher than in KO. Blood–retinal barrier dysfunction did not differ between WT and KO. Vascular abnormalities in the superficial plexus were abolished by β1/2-AR deletion, which also helped the development of the deep plexus. In both WT and KO, β3-AR agonism, acting through the nitric oxide pathway, caused enhanced neovascular responses with increased levels of VEGF.

Conclusions.: We confirm that β1- and β2-ARs play a pivotal role in retinal angiogenesis. In their presence, β3-ARs potentiate angiogenic responses, whereas, in their absence, β3-ARs sustain the angiogenic drive. These results suggest β-ARs as promising targets for therapies aimed to counteract proliferative retinopathies.

Introduction

In premature infants, the relative hyperoxia of the extrauterine environment as well as supplemental oxygen may induce an impairment of retinal blood vessel development. This may lead to retinopathy of prematurity (ROP), an ischemia-induced proliferative retinopathy, which is the main cause of blindness in children.¹ The rodent models of oxygen-induced retinopathy (OIR) are used as surrogate for investigating the pathogenesis of ROP. In mice, OIR is characterized by capillary disappearance from the central retina followed by abnormal formation of new blood vessels in the peripheral retina, which results from a drastic upregulation of proangiogenic factors including VEGF.² In addition, OIR is characterized by blood–retinal barrier (BRB) dysfunction presumably caused by VEGF upregulation that would act as a permeability factor.^3–5 

Although hypoxia is considered a trigger for neovascularization, unambiguous demonstrations that in OIR the retina is hypoxic, are still lacking. In particular, there is evidence that, in addition to hypoxia, inflammatory responses to ischemia or malnutrition may play a pivotal role in the vascular pathogenesis characteristic of OIR and contribute to proangiogenic factor upregulation.^6,7 In this respect, treatments that counteract inflammation have been shown to reduce retinal neovascularization,⁸ whereas insufficient nutritional supply or poor postnatal weight gain may profoundly alter the course of retinopathy.⁹ In addition, not only hypoxia, but also inflammatory cytokines are known to cause extensive BRB breakdown in OIR.¹⁰ 

The main aspects of ROP are replicated in OIR. However, certain caveats may be taken into account when using the mouse model of OIR. For instance, in mice, the central retinal vessels, but not the peripheral vessels, are obliterated during exposure to hyperoxia.¹¹ In addition, the contribution of BRB breakdown to ROP remains unclear as the markers of BRB are not always affected in ROP.¹² Despite these key differences, OIR very closely recapitulates the pathological events that characterize proliferative retinopathies.¹³ In addition, a great advantage of OIR in mice is the ease of genetic manipulation, which may facilitate the study of the roles of different genes in pathological neovascularization.¹¹ 

In the last few years, among the possible targets of antiangiogenic drugs, an intense research activity, recently reviewed by Casini et al.,¹⁴ has focused around the β-adrenergic system. There are several indications that β-adrenergic receptors (β-ARs) may regulate retinal pathological angiogenesis. In particular, β-ARs are expressed in the retina where β1- and β2-ARs are mainly localized to Müller cells,^15,16 which are the primary source of VEGF,¹⁷ whereas β3-ARs are localized to retinal capillaries.⁵ The fact that the retina in OIR is characterized by high levels of norepinephrine (NE) suggests the possibility that enhanced NE may overstimulate β-ARs and potentially activate mechanisms leading to proangiogenic factor upregulation.¹⁸ In this respect, drugs that can restore sympathetic homeostasis have been shown to ameliorate vascular dysfunctions. For instance, β1/2-AR blockade with propranolol or β2-AR blockade with selective antagonists reduces the formation of pathological vessels and ameliorates BRB dysfunction in concomitance with reducing retinal levels of proangiogenic factors.^5,15,19 The recent finding that oral propranolol counteracts retinal angiogenesis in preterm infants suffering from ROP has given valuable feedback to the results obtained by basic research.²⁰ 

Less is known about a possible role of β3-ARs in retinal angiogenesis except for the finding that OIR is characterized by upregulation of β3-AR expression^5,21 and that β3-ARs regulate VEGF production suggesting a role for these receptors in the neovascular response.²² 

Mechanisms mediating β-AR regulation of proangiogenic factors are less known, although their elucidation would be a great step forward in helping the development of novel drugs targeting retinal vascular pathologies. In OIR, propranolol reduces VEGF upregulation through the destabilization of the α isoform of hypoxia-inducible factor 1 (HIF-1α) and the inhibition of signal transducer and activator of transcription 3 (STAT3) phosphorylation^5,19 indicating that these transcription factors may participate in the mechanisms coupling β-ARs to VEGF regulation. Also in human pancreatic tumor cells, a β2-AR/HIF-1/VEGF axis mediates VEGF upregulation induced by β2-AR activation.²³ In the retina, β3-AR regulation of VEGF is likely to be mediated by nitrite production indicating that nitric oxide (NO) would act on VEGF through either those endothelial cells that express β3-ARs or the surrounding Müller cells.²² 

Data in favor of positive effects of β-AR blockade against retinal pathological neovascularization seem to be contradicted by numerous studies demonstrating a proangiogenic action of β-AR blockade and a beneficial effect of β-AR activation. Indeed, the lack of sympathetic activity, as in sympathectomized rats and in mice with genetic deletion of either dopamine-β-hydroxylase or distinct β-AR subtypes, causes microvascular and neural changes characteristics of diabetic retinopathy (DR).^24–28 In addition, β-AR blockade increases VEGF levels and the phosphorylation of insulin-like growth factor 1 (IGF-1) receptor (IGF-1R) in rat retina.²⁹ In addition, in human retinal endothelial cells, β1-AR activation reduces IGF-1R signaling, thus reinforcing the concept that β-AR blockade may induce angiogenic processes which, on the contrary, may be counteracted by β-AR activation.³⁰ Accordingly, in DR, isoproterenol and Compound 49b, two nonselective β-AR agonists, prevent diabetic-like changes in the retina.^31,32 In this line, a beneficial effect of β-AR blockade on pathological vascularization has been excluded since propranolol fails to suppress retinal neovascularization or VEGF upregulation in OIR.²¹ 

The discrepancy between the results obtained after propranolol administration has been debated^33,34 although more work is needed to clarify the reason of this discrepancy. In this respect, since β-AR pharmacology may pose some concerns in vivo,³⁵ the analysis of animals carrying genetic deletion of one or more β-ARs may help to better define the role of the β-adrenergic system in the neovascular response that characterizes OIR. 

As a first aim of this study, we used β1/2-AR knockout mice (from now on referred to as KO) to investigate the role of β1- and β2-ARs in the angiogenic response characteristic of OIR by determining whether retinal vascularization, the expression of factors known to be involved in angiogenic processes and BRB dysfunction may differ from wild type mice (from now on referred as WT). As a second aim of this study, KO were also used to get insights into the functional role of β3-ARs, which are the only β-AR remaining after β1/2-AR deletion. To this aim, mice were treated systemically with BRL 37344, a selective β3-AR agonist,³⁶ in order to unravel β3-AR involvement in retinal neovascular responses including β3-AR role in regulating VEGF and its receptors. Signal transduction pathways known to be coupled to β3-ARs were also studied. Effects of β1/2-AR deletion on angiogenic responses characteristic of OIR have been preliminarly reported in a recent review.¹⁴ 

Materials and Methods

Chemicals

The ELISA kit for the detection of NE was purchased from IBL International (Hamburg, Germany). Primary antibodies directed to β3-ARs, HIF-1α, STAT3, the phosphorylated form of STAT3 (pSTAT3), VEGF, VEGF receptor 1 (VEGFR-1), VEGFR-2, the phosphorylated form of VEGFR-2 (pVEGFR-2), IGF-1, the β subunit of IGF-1R and albumin, as well as the rabbit antigoat peroxidase-labeled secondary antibody, and the mouse anti-rabbit horseradish peroxidase-labeled secondary antibody, were from Santa Cruz Biotechnologies (Santa Cruz, CA, USA). The primary antibody directed to occludin was purchased from Zymed Laboratory (South San Francisco, CA, USA). Polyvinylidene difluoride (PVDF) membranes and the iQ Sybr Green Supermix were obtained from Bio-Rad Laboratories (Hercules, CA, USA). The enhanced chemiluminescence reagent was from Millipore (Billerica, MA, USA). The primary antibody directed to CD31 was obtained from BD Pharmingen (San Diego, CA, USA). The secondary antibody Alexa Fluor 488 was from Molecular Probes (Eugene, OR, USA). The RNeasy Mini Kit and the QuantiTect Reverse Transcription Kit were purchased from Qiagen (Valencia, CA, USA). The ELISA kit for the detection of VEGF was from R&D Systems (Minneapolis, MN, USA). The PepTag nonradioactive PKA assay kit was from Promega (Madison, MI, USA). The colorimetric assay kit for the detection of nitrite production was from Enzo Life Sciences (Plymouth Meeting, PA, USA). All other chemicals, including BRL 37344 and 4′,6-diamidino-2-phenylindole (DAPI), were obtained from Sigma-Aldrich Corp. (St Louis, MO, USA). 

Animals

Procedures involving animals were carried out in agreement with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and in compliance with the Italian guidelines for animal care (DL 116/92) and the European Communities Council Directive (86/609/EEC). Procedures were approved by the Ethical Committee in Animal Experiments of the University of Pisa (Pisa, Italy). All efforts were made to reduce both animal suffering and the number of animals used. 

Knockout mice (Adrb1/Adrb2 null mice; Jackson Laboratories, Bar Harbor, ME, USA) are derived from a mixed FVB/C57/129/DBA genetic background (for references see Ref. 37). They were mated with C57BL/6 mice to produce F1 mice heterozygous for both knockout genes, according to previous studies.^38,39 Thereafter, F1 heterozygous mice were mated to obtain F2 WT and KO derived from the same mixed genetic background. Genotyping was performed on tail DNA using standard protocols. Knockout mice show no gross physical or behavioral abnormalities or deficits in basal metabolic rate, although they display attenuated effects of exercise and lack the tachycardia and hypotensive responses to the β-AR agonist isoproterenol.⁴⁰ In addition, increased expression of the transcription factor Egr1 and re-expression of fetal genes were observed in the heart,^41,42 indicating subtle differences in the phenotype of KO in respect to that of WT. 

Experiments were performed on a total of 252 mouse pups of both sexes, which were euthanized at postnatal day (PD) 17. In some experiments, mice at PD12 (46 animals), PD15 (46 animals) and PD21 (24 animals) were also used. Animals were kept in a regulated environment (23 ± 18°C, 50% ± 5% humidity) with a 12-hour light/dark cycle (lights on at 8 AM) with food and water ad libitum. In all experiments, mice were anesthetized with halothane (4%), killed by cervical dislocation, and the eyes were enucleated. 

Model of Oxygen-Induced Retinopathy

To obtain OIR, litters of mice pups with their nursing mothers were exposed in an infant incubator to high oxygen concentration (75% ± 2%) between PD7 and PD12, before return to room air between PD12 and PD17. Oxygen was checked twice daily with an oxygen analyzer (Pro-Custom Elettronica, Milano, Italy). Individual litters were reared in either oxygen or room air (controls). All experiments were performed at the same time of day to exclude possible circadian influences. The data were collected from both males and females and the results combined, as there was no apparent sex difference. 

Pharmacological Treatment

BRL 37344 is a classical selective β-AR agonist.³⁶ It was considered to act as a full agonist at human or mouse β3-ARs and binding studies with human β3-ARs reported an affinity in the high nanomolar range with similar values for mouse β3-ARs.^43–45 In addition, a comparative study of all human β-ARs found a rank order of potency of β3- greater than β2- greater than β1-ARs with an approximately 20- to 100-fold selectivity for β3- than for both β2- and β1-ARs.⁴³ Results in hypoxic retinal explants support the selectivity of BRL 37344 at mouse β3-ARs.²² In the present study, BRL 37344, dissolved in dimethylsulfoxide and diluted at the final concentration in sterile saline, was given three times a day subcutaneously from PD12 to PD16 at 2 mg kg⁻¹ dose⁻¹, a dose previously used to activate β3-ARs in mouse adipose tissue.^46,47 The same dose was also demonstrated to act at β3-ARs in mouse mesenchymal stem cells.⁴⁸ Sham injections were performed with sterile saline. In all experiments no differences were observed between untreated and vehicle-treated mice. 

Enzyme-Linked Immunosorbent Assays

To measure NE levels, five samples from 10 different mice, each containing four retinas from four different mice, were used for each experimental condition. Retinas were sonicated in 330 μL of a solution containing 0.1 M HCl and 1 mM EDTA and were centrifuged at 22,000 g for 15 minutes at 4°C. Vascular endothelial growth factor levels were measured in cytosolic protein-containing supernatants used in Western blot analysis (see below). Norepinephrine and VEGF levels were measured using commercially available kits in line with previous works.^5,15,18,19 The ELISA plates were evaluated spectrophotometrically (Microplate Reader 680 XR; Bio-Rad Laboratories). Data were expressed as picomole NE/mg retina or picogram VEGF/mg protein. All experiments were run in duplicate. After statistical analysis, data from the different experiments were plotted and averaged in the same graph. 

Western Blot Analysis

To perform Western blot experiments, five samples from five different mice, each containing two retinas from two different mice, were used for each experimental condition. Retinal samples were sonicated in 150 μL of 10 mM Tris-HCl, pH 7.6, containing 5 mM EDTA, 3 mM EGTA, 250 mM sucrose, protease, and phosphatase inhibitor cocktails, and centrifuged at 22,000 g for 30 minutes at 4°C. The supernatants, containing cytosolic proteins, were used to detect HIF-1α, STAT3, pSTAT3, VEGF, IGF-1, and albumin. The supernatants were also used to measure VEGF levels (see above) and nitrite production (see below). Pellets were resuspended in 20 mM HEPES, pH 7.4, containing 150 mM NaCl, 5 mM EDTA, 3 mM EGTA, 4 mg/mL n-dodecyl-β-maltoside, protease, and phosphatase inhibitor cocktails, and centrifuged at 22,000 g for 30 minutes at 4°C. The supernatants, containing membrane proteins, were used to detect β3-ARs, VEGFR-1, VEGFR-2, pVEGFR-2, IGF-1R, and occludin. Protein concentration was determined using a fluorometer (Qubit; Invitrogen, Carlsbad, CA, USA). Aliquots of each sample containing equal amounts of protein (30 μg) were subjected to SDS-PAGE. β-actin was used as the loading control. The gels were transblotted onto PVDF membrane using a transfer system (TransBlot Turbo transfer system; Bio-Rad Laboratories) and the blots were blocked in 3% skim milk for 1 hour at room temperature. Blots were then incubated overnight at 4°C with goat polyclonal antibody directed to β3-ARs (1:200), pVEGFR-2 (1:100), and albumin (1:100), mouse monoclonal antibodies directed to HIF-1α (1:100) and pSTAT3 (1:100), or rabbit polyclonal antibodies directed to STAT3 (1:100), VEGF (1:200), VEGFR-1 (1:100), VEGFR-2 (1:100), IGF-1 (1:100), IGF-1R (1:100), and occludin (1:250). The same membrane was reblotted with a mouse monoclonal antibody directed to β-actin (1:10,000) as loading control. Finally, blots were incubated for 1 hour at room temperature with a rabbit anti-goat peroxidase-labeled secondary antibody (1:5,000), a rabbit anti-mouse horseradish peroxidase-labeled secondary antibody (1:25,000), or a mouse anti-rabbit horseradish peroxidase-labeled secondary antibody (1:5,000), and developed with the enhanced chemiluminescence reagent. Images were acquired (Chemidoc XRS+; Bio-Rad Laboratories) and the optical density of the bands was evaluated (Image Lab 3.0 software; Bio-Rad Laboratories). The data were normalized to the level of β-actin, STAT3 or VEGFR-2, as specified. All experiments were run in duplicate. After statistical analysis, data from the different experiments were plotted and averaged on the same graph. 

Immunohistochemistry and Quantitative Analysis

A rat CD31 monoclonal antibody was used to visualize blood vessels on retinal whole-mounts or 10-μm cryostat sections. Dissected retinas or enucleated eyes were immersion-fixed for 1.5 hours in 4% paraformaldehyde in 0.1 M phosphate buffer (PB) at 4°C, transferred to 25% sucrose in 0.1 M PB and stored at 4°C. Retinal whole-mounts were rinsed in 0.1 M PB and incubated for 72 hours at 4°C in the CD31 rat monoclonal antibody (1:50) diluted in 0.5% Triton X-100-containing 0.1 M PB. After incubation, the whole mounts were rinsed in 0.1 M PB and incubated for 48 hours at 4°C in Alexa Fluor 488 (1:200) in 0.1 M PB. Finally, they were rinsed in 0.1 M PB, mounted on gelatin-coated glass slides, and cover-slipped with a 0.1 M PB-glycerine mixture. Fixed eyes were embedded in Killik medium (Bio Optica Milano, Milan, Italy), frozen at −20°C, serially sectioned at 10 μm on a cryostat, mounted on gelatin-coated slides, and stored at −20°C. Sections were rinsed in 0.1 M PB and incubated overnight in the CD31 rat monoclonal antibody (1:50) in the presence of 0.5% Triton X-100 at 4°C. Sections were rinsed in 0.1 M PB and incubated in secondary antibodies conjugated with Alexa Fluor 488 (1:200) for 2 hours at room temperature. Sections were then washed in 0.1 M PB and the slides were cover-slipped with 0.1 M PB-glycerin mixture containing 0.5 μg/mL DAPI. Immunofluorescence images were viewed with a microscope (Eclipse E800; Nikon, Badhoevedorp, The Netherlands) and acquired using a digital camera (DS-Fi1c; Nikon). Electronic images were processed using an image-editing software (Adobe Photoshop CS3; Adobe Systems, Inc., Mountain View, CA, USA). In preliminary experiments, we used retinal whole-mounts to assess whether β1/2-AR deletion might alter the normal development of the superficial plexus and we found that the time course of retinal vascularization did not differ between WT and KO in terms of retinal coverage or vessel morphology (Supplementary Fig. S1). Quantification of the avascular area and neovascular tufts was performed in whole retina montages created on the basis of specific landmarks, such as the optic disc and major vessels. In each whole mount, the extent of the avascular area and the total area of preretinal neovascular tufts were measured (in pixels) using the freehand selection tool of an image-editing software (ImageJ; http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA) and were expressed as the percentage of the respective average calculated in controls. For each experimental condition, quantitative data originated from six retinas from six different mice. After statistical analysis, averaged data were plotted on the same graph. 

To visualize the deep plexus, immunofluorescent materials were observed with confocal microscopy (Laser Scanning Microscope Radiance Plus; Bio-Rad Laboratories). Overlapping stacks of confocal optical sections were acquired with ×10 objective and a detector resolution of 1024 × 1024 pixels. Each individual image was converted to 2 × 2 inches with 600 pixel/inch resolution using an image software (Adobe Photoshop CS3). 

Measurement of Retinal Vascular Leakage

The blood–retinal vascular leakage was qualitatively evaluated using Evans' blue dye. After the mice were deeply anesthetized with Avertin, Evans' blue dye, dissolved in normal saline (30 mg/mL), was injected through the femoral vein under microscopic inspection. Immediately after Evans' blue infusion, the mice turned visibly blue, confirming uptake and distribution of the dye. The mice were kept on a warm pad to ensure the complete circulation of the dye and were killed 2 hours after Evans' blue infusion. The eyes were removed and immediately immersed in 2% paraformaldehyde. After 2 hours, the retinas were dissected, and flat mounts were obtained and mounted on glass slides. Retinal flat-mounts were analyzed by fluorescence microscopy (Eclipse E800; Nikon) and images were acquired (DS-Fi1c; Nikon). For each experimental condition, six retinas from six different mice were viewed. 

Isolation of RNA and cDNA Preparation

After death, the tissues were rapidly dissected, immediately frozen in liquid nitrogen, and stored at −80°C until analysis. Total RNA was extracted (RNeasy Mini Kit; Qiagen), purified, resuspended in RNase-free water and quantified using a fluorometer (Qubit; Invitrogen). First-strand cDNA was generated from 1 μg of total RNA (QuantiTect Reverse Transcription Kit; Qiagen). 

Real-Time Quantitative RT-PCR

To perform Quantitative real-time RT-PCR (QPCR) experiments, five samples from five different mice, each containing two retinas from two different mice, were used for each experimental condition. To evaluate gene expression, QPCR experiments were performed using a kit (SYBR Green PCR Kit; Qiagen). Quantitative real-time RT-PCR primer sets were obtained from Primer Bank (VEGF)⁴⁹ or obtained from RTPrimerDB (Rpl13a).⁵⁰ The primer set for VEGF was designed in order to match the VEGF isoforms expressed in the retina. Forward and reverse primers were chosen to hybridize to unique regions of the appropriate gene sequence. Their sequences were as follows: VEGF forward 5′-GCACATAGGAGAGATGAGCTTCC-3′; VEGF reverse 5′-CTCCGCTCTGAACAAGGCT-3′; Rpl13a forward 5′-CACTCTGGAGGAGAAACGGAAGG-3′; Rpl13a reverse 5′-GCAGGCATGAGGCAAACAGTC-3.′ Amplification efficiency was close to 100% for each primer pair (Opticon Monitor 3 software; Bio-Rad Laboratories). Vascular endothelial growth factor gene was run concurrently with Rpl13a, a constitutively expressed gene encoding a ribosomal protein that is a component of the 60S subunit.⁵¹ Rpl13a is a stable housekeeping gene in OIR.¹⁵ Samples were compared using the relative threshold cycle (Ct Method).⁵² The increase or decrease (x-fold) was determined relative to a control after normalizing to Rpl13a. All reactions were run in triplicate. After statistical analysis, the data from the different experiments were plotted and averaged in the same graph. 

PKA Activity

To measure PKA activity, six samples from six different mice, each containing two retinas from two different mice, were used for each experimental condition. Protein kinase A activity was measured with the PepTag nonradioactive PKA assay kit. Retinas were sonicated in 25 mM Tris-HCl, pH 7.4, containing 0.5 mM EDTA, 0.5 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulphonyl fluoride, 1 μg/mL leupeptin, and 1 μg/mL aprotinin. Retinal cytosolic fractions were isolated by centrifugation at 14,000 g for 10 minutes at 4°C. Protein concentration was determined using a fluorometer (Qubit; Invitrogen). Protein kinase A activity was measured at 30°C for 45 minutes using 15 μg protein per sample in a final volume of 25 μL containing 5 μL PepTag PKA reaction buffer, 2 μg PepTag A1 Peptide (L-R-R-A-S-L-G kemptide), and 1 μM cAMP. The reaction was stopped by placing the samples on a 95°C heating block for 10 minutes. Agarose gel electrophoresis (0.8%) was used to separate the phosphorylated (net charge −1) from nonphosphorylated (net charge +1) kemptide. The phosphorylated kemptide was excised from the gel and PKA activity was quantified spectrophotometrically (SmartSpec 3000; Bio-Rad Laboratories). Data were expressed as incorporated picomole phosphate/mg protein/min. All experiments were run in duplicate. After statistical analysis, data from the different experiments were plotted and averaged on the same graph. 

Nitrite Production

Nitric oxide levels were indirectly detected by measuring the production of nitrite, a stable product of NO metabolism. Nitrite production was measured in cytosolic protein-containing supernatants used in Western blot analysis and was assessed by the Griess reaction using a colorimetric assay kit (IBL International), according to the manufacturer's instructions. The absorbance at 540 nm was quantified spectrophotometrically (Microplate Reader 680 XR; Bio-Rad Laboratories). Each experiment was performed as triplicate. After statistics, data were plotted and averaged in the same graph. 

Statistics

All data were analyzed by the Shapiro-Wilk test to verify their normal distribution. Statistical significance was evaluated using ANOVA followed by Newman-Keuls Multiple Comparison post-test. The results are expressed as mean ± SEM of the indicated n values (Prism 4; GraphPad Software, San Diego, CA, USA). Differences with P less than 0.05 were considered significant. 

Results

Retinal Levels of NE and β3-ARs

ELISA and Western blot were performed in order to establish whether β1/2-AR deletion might influence retinal levels of NE and β3-ARs. In controls, NE levels were approximately 43% higher in WT than in KO (P < 0.05; Fig. 1A). Oxygen-induced retinopathy caused an increase in retinal NE (~90%, P < 0.001) in line with previous results.¹⁸ In KO, NE levels associated with OIR were decreased by approximately 20% (P < 0.01) indicating a reduced stimulation of the β-adrenergic system (Fig. 1A). Retinal expression of β3-ARs did not differ between WT and KO with comparable β3-AR upregulation that characterizes OIR⁵ (~120%, P < 0.001; Fig. 1B). 

Figure 1

 

Levels of endogenous NE and β3-ARs in WT and KO mice. Mice were exposed to room air or to 75% ± 2% oxygen from PD7 to PD12 and returned to room air from PD12 to PD17. Retinas were explanted at PD17. (A) Norepinephrine levels were evaluated by ELISA. Norepinephrine levels were increased in OIR with an increase in KO lower than in WT (*P < 0.05 and **P < 0.001 versus control WT, ^#P < 0.05 versus control KO, ^$P < 0.01 versus OIR WT; ANOVA). (B) The expression of β3-ARs was evaluated by Western blot and densitometric analysis. Protein expression was relative to the loading control β-actin. Retinal expression of β3-ARs did not differ between WT and KO with comparable β3-AR upregulation (**P < 0.001 versus respective control; ANOVA). Each column represents the mean ± SEM of data from five independent samples, each containing four retinas from four different mice (NE levels) or two retinas from two different mice (β3-AR expression).

Retinal Neovascularization

Effects of β1/2-AR deletion on vascular alterations characteristic of OIR were investigated by immunohistochemistry. In OIR, the superficial vascular plexus was characterized by a pronounced avascular area without morphologic signs of neovascularization at PD12 followed by neovascular tuft formation starting to emerge at PD15 between the central avascular area and the peripheral vascularized retina. In line with previous results,² intravitreal neovascularization reached its maximal density at PD17 to regress at PD21 together with the progressive revascularization of the central retina (Figs. 2A–D). In contrast to what found in WT, the superficial plexus of KO did not exhibit an evident avascular area and only rare neovascular tufts could be seen at PD17 indicating ameliorative effects of β1/2-AR deletion on intravitreal neovascularization (Figs. 2E–H). Quantitative analysis confirmed these observations. As shown in Figure 2I, the avascular area was significantly larger in WT as compared with KO at PD12 (~60%, P < 0.001), PD15 (~81%, P < 0.001), PD17 (~92%, P < 0.001), and PD21 (~100%, P < 0.001). As shown in Figure 2J, the neovascular tuft area was significantly larger in WT than in KO at PD15 (~100%, P < 0.001) and PD17 (~90%, P < 0.001). 

Figure 2

 

Flat-mounted retinas from WT (A–D) and KO (E–H) immunolabeled with a rat monoclonal antibody directed to CD31. The superficial plexus is shown. Mice were exposed to 75% ± 2% oxygen from PD7 to PD12 and returned to room air from PD12 to PD21. Retinas were explanted at PD12 (A, E), PD15 (B, F), PD17 (C, G), or PD21 (D, H). The avascular (red boundaries) and the tuft (yellow boundaries) areas are shown. In contrast to what was found in WT, the superficial plexus of KO did not exhibit an evident avascular area and only rare neovascular tufts could be seen at PD17. Scale bar: 1 mm. The extent of the avascular area (I) and the tuft area (J) were quantitatively evaluated in WT (black bars) and KO (grey bars; *P < 0.001 versus WT; ANOVA). Each column represents the mean ± SEM of data from six retinas.

In WT, the development of the superficial plexus (Figs. 3A–C) was accompanied by a delayed development of the deep plexus. In particular, the deep plexus could not be seen at PD12 (Fig. 3D), started to be observed at PD15 (Fig. 3E) and continued to growth to cover the same area as the superficial plexus until PD17 (Fig. 3F). In KO, on the contrary, β1/2-AR deletion not only reduced vascular abnormalities in the superficial plexus (Figs. G–I) but also helped the development of the deep plexus, which was already present at PD12 (Fig. 3J), reached the same area of the superficial plexus at PD15 (Fig. 3K) and continued to follow the development of the superficial plexus at PD17 (Fig. 3L). 

Figure 3

 

Flat-mounted retinas from WT (A–F) and KO (G–L) immunolabeled with a rat monoclonal antibody directed to CD31. The superficial plexus (A–C, G–I) and the deep plexus (D–F, J–L) are shown. Mice were exposed to 75% ± 2% oxygen from PD7 to PD12 and returned to room air from PD12 to PD17. Retinas were explanted at PD12 (A, D, G, J), PD15 (B, E, H, K), or PD17 (C, F, I, L). Delayed development of the deep plexus that characterizes OIR could be observed in WT but not in KO. Scale bar: 1 mm.

The effects of β1/2-AR deletion on vascular abnormalities that characterize OIR were also determined in vertical sections of the midperipheral retina of WT (Figs. 4A–C) and KO (Figs. 4D–F) taken at PD12 (Figs. 4A, D), PD15 (Figs. 4B, E), and PD17 (Figs. 4C, F). In WT, blood vessel profiles in the superficial plexus could be observed at PD12 and started to form engorged neovascular tufts at PD15 when vertical vessels diving down from the superficial plexus began to form the deep plexus. At PD17, vessel profiles were seen in the deep plexus with communicating vessels connecting them to the superficial plexus in which neovascular tufts were clearly visible. Knock out mice were characterized by a decreased formation of neovascular tufts in the superficial plexus and by an anticipated development of the deep plexus, which started to be observed already at PD12. 

Figure 4

 

Vertical sections of midperipheral retinas of WT (A–C) and KO (D–F) immunolabeled with a rat monoclonal antibody directed to CD31 (green). Mice were exposed to 75% ± 2% oxygen from PD7 to PD12 and returned to room air from PD12 to PD17. Sections were taken from retinas at PD12 (A, D), PD15 (B, E), or PD17 (C, F). Nuclei were counterstained with DAPI (blue). In WT, blood vessel profiles were observed only in the superficial plexus at PD12. At PD15, neovascular tufts were observed in the superficial plexus along with development of the deep plexus. At PD17, vessel profiles were seen in the deep as well as in the superficial plexus in which neovascular tufts were clearly visible. In KO, the deep plexus could be observed already at PD12. The superficial plexus was characterized by a reduced neovascular tuft formation. Deep, deep plexus; Int, intermediate plexus; Sup, superficial plexus; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer. Scale bar: 20 μm.

Transcription and Proangiogenic Factors

Effects of β1/2-AR deletion on the levels of transcription and proangiogenic factors known to play a role in OIR were investigated.^5,19 As shown in Figure 5A, levels of HIF-1α, pSTAT3, STAT3, VEGF, VEGFR-1, pVEGFR-2, VEGFR-2, IGF-1, and IGF-1R increased from PD12 to PD17 in both WT and KO. As evaluated at PD17 (Fig. 5B), levels of VEGFR-1, pVEGFR-2, and VEGFR-2 differed between WT and KO. In particular, VEGFR-1 levels were approximately 64% lower in WT than in KO (P < 0.01), while pVEGFR-2 and VEGFR-2 levels were approximately 49% higher in WT than in KO (P < 0.01), suggesting a reduced activation of VEGF signaling in KO. This difference between WT and KO was not found at PD12 or PD15 (not shown). 

Figure 5

 

Retinal levels of transcription and proangiogenic factors of WT and KO. Mice were exposed to 75% ± 2% oxygen from PD7 to PD12 and returned to room air from PD12 to PD17. Retinas were explanted at PD12, PD15, or PD17. (A) Expression of HIF-1α, pSTAT3, STAT3, VEGF, VEGFR-1, pVEGFR-2, VEGFR-2, IGF-1, and IGF-1R as evaluated by Western blot. Levels of transcription and proangiogenic factors were increased from PD12 to PD17 in both WT and KO. (B) Densitometric analysis performed at PD17 in WT (black bars) and KO (grey bars). Protein expression was relative to the loading control β-actin (HIF-1α, STAT3, VEGF, VEGFR-1, VEGFR-2, IGF-1, and IGF-1R), to STAT3 (pSTAT3), or to VEGFR-2 (pVEGFR-2). Vascular endothelial growth factor receptor–1 levels were lower in WT than in KO, while pVEGFR-2 and VEGFR-2 levels were higher in WT than in KO (*P < 0.001 versus WT; ANOVA). Each column represents the mean ± SEM of data from five independent samples, each containing two retinas from two different mice.

Blood–Retinal Barrier

The finding that in OIR, propranolol partially restores BRB integrity⁵ prompted us to investigate the effects of β1/2-AR deletion on BRB. To this aim, we measured retinal levels of occludin, a tight junction protein, and albumin, a marker of vascular leakage. In agreement with previous results,⁵ OIR was characterized by occludin levels lower than in controls (~46%; P < 0.001), but albumin levels higher than in controls (~130 %; P < 0.001). No significant differences were found between WT and KO (Figs. 6A, B). The use of Evan's blue dye in the evaluation of BRB leakage may pose several problems as discussed by Berkowitz et al.,⁵³ but these problems are particularly relevant for quantitative evaluation of BRB leakage.⁵⁴ In the present study, qualitative assessment of vascular permeability with Evans' blue demonstrated dye retention within the vessel lumen in controls (Figs. 6C, E), whereas OIR was characterized by dye leakage into the retinal parenchyma without any difference between WT and KO (Figs. 6D, F). 

Figure 6

 

Evaluation of BRB dysfunction in WT and KO. Mice were exposed to room air or to 75% ± 2% oxygen from PD7 to PD12 and returned to room air from PD12 to PD17. Retinas were explanted at PD17. Levels of occludin (A) and albumin (B) were evaluated by Western blot and densitometric analysis. Protein expression was relative to the loading control β-actin. In OIR, occludin levels were decreased, while albumin levels were increased with no differences between WT and KO (*P < 0.001 versus relative control; ANOVA). Each column represents the mean ± SEM of data from five independent samples, each containing two retinas from two different mice. (C–F) Blood–retinal vascular leakage as qualitatively evaluated with Evans' blue in control WT (C), OIR WT (D), control KO (E) and OIR KO (F). Arrows: vascular leakage. Six mice were used for each experimental condition. Scale bar: 200 μm.

β3-AR Agonism

The nearly abolished angiogenic response observed in OIR KO, together with the reduced activation of VEGF signaling, seems to point to a major role of β1- and/or β2-ARs in the regulation of angiogenic processes in the retina, thus excluding any residual role of β3-ARs. On the other hand, the possibility exists that retinal levels of NE must be ineffective in activating β3-ARs, whose possible function in regulating angiogenesis might be revealed only after their exogenous activation. 

The effect of BRL 37344 on pathological neovascolarization was evaluated by CD31 immunohistochemistry at PD17. In WT, BRL 37344 did not affect the avascular area, whereas it caused an evident increase in blood vessel tufts with respect to untreated OIR (Figs. 7A, B). In KO (Figs. 7C, D), the extension of the avascular area was not affected by BRL 37344, which, in contrast, caused the appearance of numerous blood vessel tufts. Quantitative analysis confirmed these observations. As shown in Figure 7E, the avascular area did not change after BRL 37344 in both WT and KO. In contrast, BRL37344 significantly increased the neovascular tuft area in both WT and KO (~41%, P < 0.05 and ~1200%, P < 0.001, respectively; Fig. 7F). 

Figure 7

 

Flat-mounted retinas from WT (A, B) and KO (C, D) treated with BRL 37344, immunolabeled with a rat monoclonal antibody directed to CD31. Mice were exposed to 75% ± 2% oxygen from PD7 to PD12 and returned to room air from PD12 to PD17. Mice were untreated (A, C) or treated with subcutaneous injection of either vehicle (not shown) or 2 mg kg⁻¹ BRL 37344 three times a day (B, D) from PD12 to PD16. Retinas were explanted at PD17. The avascular area (red boundaries) and the tuft area (yellow boundaries) are shown. BRL 37344 did not affect the avascular area, while caused an increase in neovascular tuft formation in both WT and KO. Scale bar: 1 mm. The extent of the avascular area (E) and the tuft area (F) were quantitatively evaluated (*P < 0.05 and **P < 0.001 versus respective untreated OIR; ANOVA). Each column represents the mean ± SEM of data from six retinas.

In OIR, KO was characterized by similar levels of VEGF (Figs. 8A–C), increased expression of VEGFR-1 (~66%, P < 0.001; Fig. 8D), and decreased expression of both pVEGFR-2 and VEGFR-2 (~34% and ~33%, P < 0.001, respectively; Figs. 8E, F) in respect to WT. Vascular effects of BRL 37344 were accompanied by increased levels of VEGF with no difference between the two strains. In both WT and KO, BRL 37344 increased VEGF mRNA (~28% and ~26%, respectively, P < 0.001; Fig. 8A) and protein, as evaluated by Western blot (~35% and ~27%, respectively, P < 0.001; Fig. 8B) or ELISA (~20% and ~26%, respectively, P < 0.001; Fig. 8C). In contrast, BRL 37344 did not affect levels of VEGFR-1 (Fig. 8D), VEGFR-2 (Fig. 8E), and pVEGFR-2 (Fig. 8F). 

Figure 8

 

Levels of VEGF and its receptors in WT and KO after BRL 37344. Mice were exposed room air or to 75% ± 2% oxygen from PD7 to PD12 and returned to room air from PD12 to PD17. Mice exposed to oxygen were untreated or treated with subcutaneous injection of either vehicle or 2 mg kg⁻¹ BRL 37344 three times a day from PD12 to PD16. Retinas were explanted at PD17. (A–C) Vascular endothelial growth factor levels as evaluated by QPCR (A), Western blot and densitometric analysis (B) and ELISA (C). (D–F) Vascular endothelial growth factor receptor–1 (D), pVEGFR-2 (E), and VEGFR-2 (F) levels as evaluated by Western blot and densitometric analysis. In qPCR experiments data were analyzed by the formula 2^−ΔΔCT using Rpl13a as internal standard. In Western blot experiments protein expression was relative to the loading control β-actin (VEGF, VEGFR-1, VEGFR-2) or to VEGFR-2 (pVEGFR-2). BRL 37344 caused an increase in VEGF levels in both WT and KO (*P < 0.01 and **P < 0.001 versus respective control; ^§P < 0.05 versus respective untreated OIR, ^$P < 0.001 versus untreated WT; ANOVA). Each column represents the mean ± SEM of data from five independent samples, each containing two retinas from two different mice.

To evaluate possible effects of β1/2-AR deletion on β3-AR coupling to signal transduction pathways, PKA activity and nitrite production were evaluated after BRL 37344 as PKA and NO are β3-AR downstream effectors in the retina.^15,18,22 Oxygen-induced retinopathy was characterized by lower PKA activity (~53% in WT and ~46% in KO, P < 0.001; Fig. 9A) and higher nitrite production (~235% in WT and ~210% in KO, P < 0.001; Fig. 9B) in respect to controls. BRL 37344 did not affect PKA activity, while further enhanced nitrite production (~25%, P < 0.05) without any difference between WT and KO. 

Figure 9

 

Protein kinase A activity and nitrite production in WT and KO after BRL 37344. Mice were exposed to room air or to 75% ± 2% oxygen from PD7 to PD12 and returned to room air from PD12 to PD17. Mice exposed to oxygen were untreated or treated with subcutaneous injection of either vehicle or 2 mg kg⁻¹ BRL 37344 three times a day from PD12 to PD16. Retinas were explanted at PD17. (A) Measurements of PKA activity. Representative agarose gel containing phosphorylated and nonphosphorylated kemptide bands is shown. The positive control (+) was provided by the manufacturer. In the negative control (–), PKA activity was measured in the absence of the cytosol fraction. Protein kinase A activity was quantified spectrophotometrically from bands excised from the gel. Neither in WT nor in KO, BRL 37344 affected PKA activity (*P < 0.001 versus respective control; ANOVA). (B) Nitrite production was evaluated by the Griess method. In both WT and KO, BRL 37344 caused an increase in nitrite production (*P < 0.001 versus respective control, ^§P < 0.05 versus respective untreated OIR; ANOVA). Each column represents the mean ± SEM of data from six independent samples, each containing two retinas from two different mice (PKA activity) or from five independent samples, each containing two retinas from two different mice (nitrite production).

Discussion

In the present study, we confirm and expand the notion that increased activity of the sympathetic system may play a key role in regulating retinal angiogenesis and that drugs that restore sympathetic homeostasis can effectively ameliorate vascular dysfunctions. 

NE Levels and β3-AR Expression

Although little is known on NE sources in the retina, there are indications that most of the retinal NE is associated with sympathetic nerves directed to choroidal vessels. However, in ischemic conditions retinal neurons and/or endothelial cells may represent additional sources of NE (for references see Ref. 14). As shown by the present results, in the absence of β1- and β2-ARs NE upregulation that characterizes OIR is lower than in their presence suggesting that β1- and/or β2-ARs may contribute to regulate NE levels in the retina. 

As also shown by the present results, the marked upregulation of β3-ARs in OIR^5,18 is not influenced by β1/2-AR deletion. This finding seems to exclude that the lack of β1- and/or β2-ARs may affect the expression of β3-ARs and is in line with previous findings demonstrating that β3-ARs do not compensate for the lack of β1- and β2-ARs in mouse myocytes.⁵⁵ 

Effects of β1/2-AR Deletion on Retinal Vasculature

The present results show that β1/2-AR deletion has little or no impact on physiological vascularization unless ischemic conditions are experienced by the retina. Previous finding, in contrast, have demonstrated that in physiologic conditions, the lack of β1-AR causes capillary degeneration supporting the idea that maintenance of β-AR signaling is beneficial for retinal homeostasis.²⁵ 

As also shown by the present results, hyperoxia-induced vaso-obliteration and subsequent neovascular tuft proliferation are almost prevented by β1/2-AR deletion suggesting that in OIR β-AR overstimulation may potentially act as a proangiogenic switch. The finding that β1/2-AR deletion prevents the formation of the avascular area implies that β-AR overstimulation may cause vaso-obliteration in the central retina. In this respect, there are several evidence that β1- and β2-ARs are coupled to main regulators of retinal vasculature, which, lacking autonomic innervation, is mostly influenced by local factors released by endothelial cells and surrounding retinal tissue.⁵⁶ 

Effects of β1/2-AR deletion in OIR are in line with previous data demonstrating that β-AR blockade with propranolol or β-AR desensitization after isoproterenolol reduces retinal neovascularization.^5,18,19 Effects of β-AR blockade in OIR have been confirmed by clinical data. In fact, in preterm infants, propranolol counteracts ROP²⁰ and, in patients with AMD, reduces the need for repeated intravitreal injections of bevacizumab.⁵⁷ 

At variance with β1/2-AR deletion, β-AR blockade does not prevent the formation of the central avascular area presumably because the administration of β-AR blockers begins at PD12 when the avascular area is already formed.^5,15,19 In addition, genetic deletion of a receptor and antagonist treatment at a discrete time do not necessarily result in similar phenotypes. In this respect, mice lacking β-ARs do not show any gross abnormalities,⁴⁰ while β-AR blockade is a way to reduce hypertension. 

The additional finding that the delay in deep plexus development that characterizes OIR² is not observed in KO implies that the lack of β1 and/or β2-ARs not only blocks pathological neovascularization in the superficial plexus, but also aids the development of the inner retinal vessels. This would have implication for treatments involving β1- and β2-AR antagonists as ameliorative effects on inner retinal vessels would affect visual function by rescuing, for instance, the function of amacrine cells that contribute to electroretinographic responses⁵⁸ and are highly susceptible to ischemia.⁵⁹ In general, treatments reducing retinal vascular pathology may be expected to ameliorate neuronal defects that characterize OIR. For instance, β2-AR blockade has been shown to reduce neovascularization in the superficial plexus and to restore electroretinographic responses.¹⁵ However, there is also evidence indicating that treatments preventing pathological angiogenesis may not necessarily improve retinal function.⁶⁰ 

The notion that the lack of β1/2-ARs represents a protective mechanism to avoid pathological neovascularization in OIR is in apparent contrast with previous results demonstrating that β-AR agonism ameliorates retinal vascularization in DR.^31,32 This can be explained by the fact that OIR and DR, although sharing some similarities, are based on different experimental models and reproduce distinct stages of vascular pathology (early proliferative in OIR versus late nonproliferative in DR).¹⁴ 

Effects of β1/2-AR Deletion on Factors Involved in Retinal Angiogenesis

The present results show that retinal levels of several transcription and proangiogenic factors increase between PD12 and PD17 without any difference between WT and KO indicating that β1/2-AR deletion is protective against neovascular responses without influencing proangiogenic cascade. However, despite similar VEGF levels in WT and KO, a reduced angiogenic drive can be observed in KO. This can be interpreted as the result of an increased expression of VEGFR-1 and a decreased activation of VEGFR-2, which characterize KO. Vascular endothelial growth factor preferentially binds to VEGFR-1, which, according to its function as a decoy receptor,⁶¹ would reduce VEGFR-2 signaling thus interfering with the main role of VEGFR-2 in angiogenic processes.⁶² 

Effects of β1/2-AR Deletion on BRB

As shown by the present results, BRB damage persists in the absence of β1- and β2-ARs in contrast with the fact that pharmacological β-AR blockade ameliorates BRB dysfunction.⁵ This can be explained by assuming that in KO, the persisting high level of VEGF may act as permeable factor to induce alterations in tight junction proteins resulting in BRB leakage.⁶³ This hypothesis implies that reduced VEGFR-2 signaling that characterizes KO is not sufficient to prevent BRB dysfunction. One possibility is that VEGFR-2 signaling alone is not responsible for BRB leakage that instead requires cross-communication between VEGFR-1 and VEGFR-2.⁶⁴ Another possibility to explain the persisting damage of BRB in KO is that the newly formed vessels colonizing the avascular area remain leaky until PD17, thus permitting the extravasation of plasma proteins. In this respect, newly formed blood vessels are known to be functionally defective with their stabilization depending on their maturation.⁶⁵ 

Role of β3-ARs

There are several findings suggesting a role of β3-ARs in retinal angiogenesis. Indeed, β3-ARs are localized to blood vessels in rodent retinas,^5,66 they are upregulated in OIR,^5,21 their activation increases the expression of proangiogenic factors in human retinal and choroidal endothelial cells,^67,68 and their blockade or silencing reduces the expression of angiogenic factors in retinal explants cultured in reduced oxygen tension.²² The fact that β1/2-AR deletion protects the retina from pathological angiogenesis suggests that β3-ARs are normally under stimulated. They would become fully active solely when adequately stimulated, suggesting that β3-ARs have no tonic role in regulating retinal angiogenesis, whereas such a role can be supposed for β1- and/or β2-ARs. 

In the present study, β3-AR role in the angiogenic response has been disclosed for the first time using KO and β3-AR agonism with BRL 37344. The binding affinity of BRL 37344 at human β3-ARs is 10-fold higher than that of NE⁴³ and our results indicate that sufficient BRL 37344 penetrates the BRB although it is difficult to conjecture about the actual concentration of BRL 37344 reaching the retina. Retinal penetration of BRL 37344 may be enhanced by BRB leakage characteristic of OIR.^69,70 

As shown by the present results, BRL 37344 stimulates the tuft formation in line with the possibility that β3-ARs, localized to neovascular tufts,⁵ regulate endothelial cell proliferation. In this respect, the proliferative potential of β3-AR activation has been demonstrated in human retinal endothelial cells.⁶⁷ The finding that in KO, BRL 37344 causes neovessel growth suggests that β3-AR activation may replace β2-AR function. In this respect, one can assume that β2-ARs may normally mask β3-AR activity suggesting a redundancy of the β-AR system. In addition, reduced NE upregulation in KO may not provide sufficient ligand concentration to β3-ARs as the increase in tuft formation after BRL 37344 is greater in KO than in WT. The additional finding that BRL 37344 does not influence the avascular area in the central retina, while increases tuft formation in the midperipheral retina is in line with the possibility that vessel growth in the avascular region and neovascular tuft formation are two processes inversely correlated: when vessels grow in the central retina, neovascular tufts are reduced in the periphery or vice versa.⁷ 

The finding that β3-AR activation sustains the angiogenic response in the absence of β1- and β2-ARs may help to explain the discrepancy between our studies reporting antiangiogenic effects of propranolol^5,19 and those excluding them, although in a mouse strain in which OIR is characterized by an enormous increase of β3-AR expression,²¹ which may be responsible for the reported insensitivity to propranolol in this strain. 

β3-AR Signaling

In line with previous findings,^15,18,22 OIR is characterized by reduced PKA activity and increased nitrite production indicating that PKA downregulation and NO upregulation are both involved in mechanisms by which the retina adapts to decreased oxygen tension. Although more work is required to determine the molecular pathways underlying β-AR regulation of angiogenesis, the current study indicates that β1/2-AR deletion suppresses pathological angiogenesis without influencing PKA activity. The fact that, in contrast, β-AR antagonism prevents PKA reduction^15,18 suggests that β1/2-AR deletion is not sufficient to restore PKA further demonstrating that pharmacological blockade and genetic deletion often result in different effects. 

The present results show that PKA activity is not modulated by BRL 37344, indicating that PKA is not (or only minimally) involved in mediating the angiogenic effects of selective activation of β3-ARs. This is in line with the observation that BRL 37344 causes the proliferation of human retinal endothelial cell through a PKA-independent pathway.⁶⁷ Rather, the present results suggest that NO pathway is a downstream signaling of β3-ARs as BRL 37344 increases nitrite production. This finding is in line with previous evidence demonstrating β3-AR coupling to NO²² and suggests that β3-AR–induced production of NO is a nodal point in the control of angiogenesis-related diseases in the retina. 

Conclusions

Overall, our results confirm the pivotal role that β1- and/or β2-ARs play in retinal angiogenesis, thus expanding the importance of the use of β-AR blockers to counteract angiogenesis-driven retinal pathologies. As to β3-ARs, the fact that their agonism increases the neovascular response in WT and causes neovascularization in KO indicates that, in the presence of β1- and/or β2-ARs, β3-ARs may potentiate retinal angiogenic responses, whereas, in their absence, may sustain the angiogenic drive. Although extrapolation of these data to the human situation is difficult, these results may help to further explore the role of β-ARs in angiogenesis-driven diseases and suggest β-ARs as promising, novel targets for the development of therapies aimed at counteracting proliferative retinal pathologies. 

Acknowledgments

The authors thank Giovanni Casini for his valuable comments and the extensive revision of the manuscript. The authors also thank Irene Fornaciari for her technical assistance in immunohistochemistry. 

Supported by grants from Regione Toscana (PB; Florence, Italy), Meyer Foundation (MDM, PB; Florence, Italy), and Ente Cassa di Risparmio di Firenze (LF; Florence, Italy). 

Disclosure: M. Dal Monte, None; M. Cammalleri, None; E. Mattei, None; L. Filippi, None; P. Bagnoli, None 

Footnotes

 MDM and MC contributed equally to the work presented here and should therefore be regarded as equivalent authors.