Antiangiogenic effects of β2-adrenergic receptor blockade in a mouse model of oxygen-induced retinopathy

Proliferative retinopathies including retinopathy of prematurity(ROP) are the leading causes of blindness in the pediatric age.ROP progresses in two phases. The first phase begins withdelayed retinal vascular growth after birth and partial regressionof existing vessels, followed by a second phase of hypoxia-induced hyper-vascularization (Chen and Smith 2007).

Pathogenic angiogenesis occurs on the surface of the retina(pre-retinal neovascular tuft formation) and causes functionalimpairments and severe vision loss, whereas intra-retinalangiogenesis is seen mostly in physiologic angiogenesisduring development (Zou et al. 2010). Vascular endothelialgrowth factor (VEGF), insulin-like growth factor-1 (IGF-1)and their receptors are involved in the pathogenic bloodvessel formation that characterizes ROP (Dal Monte et al.2007; Ristori et al. 2011).

The beta-adrenergic system interferes with several essen-tial steps of neovascularization opening novel therapeutic

Received May 19, 2011; revised manuscript received September 2,2011; accepted September 20, 2011.Address correspondence and reprint requests to Paola Bagnoli, Di-

partimento di Biologia, Universita di Pisa, Via San Zeno, 31 – 56127Pisa, Italy. E-mail: [email protected] used: b-AR, beta-adrenergic receptor; DMSO, dim-

ethylsulfoxide; DR, diabetic retinopathy; ERG, electroretinogram; GCL,ganglion cell layer; IGF-1, insulin-like growth factor-1; IGF-1R, insulin-like growth factor-1 receptor; IPL, inner plexiform layer; IR, immuno-reactivity; OD, optical density; OIR, oxygen-induced retinopathy; ONL,outer nuclear layer; OPs, oscillatory potentials; OPL, outer plexiformlayer; PD, postnatal day; PKA, protein kinase A; ROP, retinopathy ofprematurity; siRNA, small-interfering RNA; SOPs, sum OPs; VEGF,vascular endothelial growth factor; VEGFR, VEGF receptor.

*Department of Biology, University of Pisa, Via San Zeno, Pisa, Italy

�Neonatal Intensive Care Unit, Department of Critical Care Medicine, ‘‘A. Meyer’’ University

Children’s Hospital, Florence, Italy

Abstract

Oxygen-induced retinopathy (OIR) is a model for human ret-

inopathy of prematurity. In mice with OIR, beta-adrenergic

receptor (b-AR) blockade with propranolol has been shown to

ameliorate different aspects of retinal dysfunction in response

to hypoxia. In the present study, we used the OIR model to

investigate the role of distinct b-ARs on retinal proangiogenic

factors, pathogenic neovascularization and electroretino-

graphic responses. Our results demonstrate that b2-AR

blockade with ICI 118,551 decreases retinal levels of proan-

giogenic factors and reduces pathogenic neovascularization,

whereas b1- and b3-AR antagonists do not. Determination of

retinal protein kinase A activity is indicative of the fact that

b-AR blockers are indeed effective at the receptor level. In

addition, the specificity of ICI 118,551 on retinal angiogenesis

has been demonstrated by the finding that in mouse retinal

explants, b2-AR silencing prevents ICI 118,551 effects on

hypoxia-induced vascular endothelial growth factor accumu-

lation. In OIR mice, ICI 118,551 is effective in increasing

electroretinographic responses suggesting that activation of

b2-ARs constitutes an important part of the retinal response to

hypoxia. Lastly, immunohistochemical studies demonstrate

that b2-ARs are localized to several retinal cells, particularly to

Muller cells suggesting the possibility that b2-ARs play a role

in regulating vascular endothelial growth factor production by

these cells. The present results suggest that pathogenic

angiogenesis, a key change in many hypoxic/ischemic vision-

threatening retinal diseases, depends at least in part on b2-AR

activity and indicate that b2-AR blockade can be effective

against retinal angiogenesis.

Keywords: electroretinography, hypoxia, immunohistochem-

istry, pathogenic neovascularization, proangiogenic factors,

siRNA.

J. Neurochem. (2011) 119, 1317–1329.

JOURNAL OF NEUROCHEMISTRY | 2011 | 119 | 1317–1329 doi: 10.1111/j.1471-4159.2011.07530.x

� 2011 The AuthorsJournal of Neurochemistry � 2011 International Society for Neurochemistry, J. Neurochem. (2011) 119, 1317–1329 1317

opportunities for the use of beta adrenergic receptor (b-AR)blockers in the treatment of angiogenesis-dependent diseases.For instance, the non-selective b-AR blocker propranololinhibits the growth of capillary hemangiomas (Leaute-Labreze et al. 2008) as well as the proliferation, migration,differentiation and the VEGF-induced activation of theVEGF receptor-2 (VEGFR-2) in human endothelial cells(Lamy et al. 2010). In addition, propranolol down-regulatesproangiogenic factors, ameliorates proangiogenic effect ofhypoxia and repairs blood-retinal barrier breakdown in theretina of a mouse model of oxygen-induced retinopathy(OIR) (Ristori et al. 2011), which is a well-established modelto mimic the two phases of human ROP (Chen and Smith2007).

There are indications that distinct b-ARs are expressed inthe rodent retina (Smith et al. 2007). For instance, b1- andb2-ARs have been localized to cultured Muller cells of the ratretina (Walker and Steinle 2007). In the mouse retina, b3-ARs have been localized to the inner capillary network andtheir expression is up-regulated by hypoxia (Ristori et al.2011). In addition, b1- and b3-ARs are expressed in humanretinal endothelial cells (Steinle et al. 2003, 2005).

In line with previous findings of vascular protection afterb-AR blockade with propranolol in OIR mice (Ristori et al.2011), we hypothesized that treatment with distinct b-ARantagonists may result in ameliorated visual function. In thisrespect, the electroretinogram (ERG) has been extensivelyused to determine the function of photoreceptors and post-receptor cells in the retina. In fact, the a-wave directlyreflects photoreceptor responses to light, whereas the b-wave is produced by second-order bipolar cells. Theoscillatory potentials (OPs), which are believed to begenerated in the proximal retina, appear as a group ofwavelets superimposed on the leading edge of the b-wave.In OIR rats, ERG alterations have been observed inresponse to hypoxia (Liu et al. 2006a; b). ERG alterationshave been interpreted as the result of photoreceptor andpost-receptor abnormalities caused by abnormal retinalvasculature, although post-receptor neuron remodelingcaused by irreversible damage to the photoreceptors cannotbe excluded (Fulton et al. 2009). Profound ERG alterationshave been observed in OIR mice in which rod and coneresponses are drastically reduced by hypoxia (Vessey et al.2011) although little is known on b-AR regulation of ERGresponses in OIR.

In the present study, we investigated: (i) the effects ofselective b-AR antagonists on retinal proangiogenic factorsand pathogenic neovascularization, (ii) the effectiveness ofb-AR antagonists by determining retinal protein kinase A(PKA) activity, (iii) the selectivity of ICI 118,551 effects inthe retina by inhibiting b2-ARs with small interfering RNA(siRNA), (iv) the effects of ICI 118,551 on visual function byERG recording and (v) the localization of b2-ARs to themouse retina.

Materials and methods

AnimalsProcedures involving animals were carried out in compliance withthe Italian guidelines for animal care (DL 116/92) and the EuropeanCommunities Council Directive (86/609/EEC) and were approvedby the Ethics Committee in Animal Experiments of the Universityof Pisa. The ARRIVE guidelines were followed. All efforts weremade to reduce both animal suffering and the number of animalsused. Two months old male and female mice (C57BL/6J strain)were originally purchased from Charles River Laboratories Italy(Calco, Italy) and were mated in our breeding colony. Experimentswere performed on a total of 452 mouse pups of both sexes. Micewere killed at postnatal day (PD)17 (6 g body weight) in in vivoexperiments. Before killing, some mice were used in ERGexperiments (see below). In experiments with retinal explants, miceat PD12 (3.5 g body weight) were used as PD12 is the day at whichmice begin to experience hypoxia in the OIR model. Animals werekept in a regulated environment (23 ± 1�C, 50 ± 5% humidity) witha 12-h light/dark cycle (lights on at 8 AM) with food and water adlib. In all experiments, mice were anesthetized with halothane (4%),killed by cervical dislocation and the eyes were enucleated.

Model of oxygen-induced retinopathyIn a typical model of OIR (Smith et al. 1994), litters of mice pupswith their nursing mothers were exposed in an infant incubator tohigh oxygen concentration (75 ± 2%) between PD7 and PD12, priorto return to room air between PD12 and PD17. Oxygen was checkedtwice daily with an oxygen analyzer (Pro-Custom Elettronica,Milano, Italy). Individual litters were either oxygen or room airreared. The data were collected from both males and females and theresults combined as there was no apparent gender difference.

Retinal explant culturesRetinal explants were cultured according to Ogilvie et al. (1999), asspecified in Appendix S1. Retinal explants were cultured in eithernormoxia (20% O2, 5% CO2, 24 h, 37�C) or hypoxia (1% O2, 5%CO2, 24 h, 37�C). Hypoxia was achieved using a MiniGalaxy Aincubator (RS Biotech, Irvine, Scotland, UK).

b-AR antagonistsAtenolol is a b1-AR blocker with limited effects at b2-ARs (Vrydagand Michel 2007). In the OIR model, atenolol was used atconcentrations ranging from 10 to 30 mg/kg/dose in line withprevious studies in diabetic rat retina (Phipps et al. 2007; Wilkin-son-Berka et al. 2007). In retinal explants, atenolol was used at10 lM in line with previous studies in in vitro models (Deng et al.2004; Jun et al. 2004; Al Zubair et al. 2008; Sartoretto et al. 2009).ICI 118,551 is a selective b2-AR antagonist. It has limited effects atb1-ARs (Vrydag and Michel 2007), but may also act as inverseagonist at b3-ARs (Hoffmann et al. 2004). ICI 118,551 has noknown clinical applications, but it is used in experiments because ofits strong receptor specificity. Common systemic doses used inrodent research are about 1–2 mg/kg (Yalcin et al. 2010; Yu et al.2010) although in some experimental models, ICI 118,551 efficacyhas been demonstrated at doses up to 5–30 mg/kg (Tan et al. 2003;Kitamura et al. 2007; Mantsch et al. 2010). In OIR mice, ICI118,551 was used at concentrations ranging from 1 to 30 mg/kg/

Journal of Neurochemistry � 2011 International Society for Neurochemistry, J. Neurochem. (2011) 119, 1317–1329� 2011 The Authors

1318 | D. Martini et al.

dose, whereas in retinal explants, ICI 118,551 was used atconcentrations ranging from 0.1 to 10 lM according to previousstudies in human cancer cell lines (Abraham et al. 2004) and mousemesenteric arteries (Al Zubair et al. 2008). SR59230A has beenwidely used as b3-AR antagonist in mice (Lenard et al. 2003; AlZubair et al. 2008), but it may also act as partial agonist at b3-ARs(Vrydag and Michel 2007). In the present study, SR59230A wasused at concentrations ranging from 5 to 20 mg/kg/dose in line withprevious studies in mice (Lenard et al. 2003). In retinal explants,SR59230A was used at 10 lM according to previous studies inmouse cultured cells (Russell and Tisdale 2005).

Pharmacological treatmentsIn OIR mice, b-AR antagonists were given three times a daysubcutaneously from PD12 to PD16. Atenolol was dissolved incitrate buffer, whereas ICI 118,551 and SR59230A were dissolvedin dimethylsulfoxide (DMSO) and diluted with citrate buffer (finalconcentration of DMSO 1%). Sham injections were performed withthose vehicles. In experiments using retinal explants, atenolol wasdissolved in citrate buffer, whereas ICI 118,551 and SR59230Awere dissolved in DMSO, at 10 mM. DMSO at 0.1% was used incontrol experiments. Atenolol, ICI 118,551 or SR59230A wereapplied for 24 h in hypoxia. To test the effect of ICI 118,551 onretinal explants in which b2-ARs were silenced, 10 lM ICI 118,551was added to the medium after 24 h incubation in normoxiaimmediately before the transfer of retinal explants to hypoxia.

siRNA transfectionPre-designed Flexitube siRNAs directed to b2-ARs, non-silencingand mitogen-activated protein kinase 1 positive control siRNAswere used (Qiagen, Valencia, CA, USA). Retinal explants weretransfected according to the manufacturer’s instructions. Afterpreliminary experiments performed to identify the siRNA that gavethe maximum reduction in the expression of b2-ARs at the lowestconcentration (Appendix S1 and Figure S1), we used b2-AR-siRNA3 at 50 nM. After transfection with b2-AR-siRNA3 or non-silencing siRNA, retinal explants were cultured for 24 h innormoxia and for further 24 h in hypoxia.

RNA isolation and cDNA preparationFrom each retina, total RNA was extracted (RNeasy Mini Kit;Qiagen), purified, resuspended in Rnase-free water and quantifiedspectrophotometrically (SmartSpec 3000; Bio-Rad, Hercules, CA,USA). First-strand cDNA was generated from 1 lg of total RNA(QuantiTect Reverse Transcription Kit; Qiagen). Aliquots of cDNAwere stored at )20�C.

Quantitative real time RT-PCRQuantitative real-time RT-PCR was performed using publishedprimers for b2-AR, VEGF, VEGFR-1, VEGFR-2, IGF-1 and IGF-1R, as previously described in mouse retinas (Ristori et al. 2011).For each experimental condition, 12 samples from 12 different micewere used in studies using OIR mice, whereas eight samples fromeight different mice were used in studies using retinal explants. Eachsample refers to the mRNA extracted from one retina. Primersequences are listed in Table S1. Amplification efficiency was closeto 100% for each primer pairs (Opticon Monitor 3 software; Bio-Rad). Each target gene was run concurrently with the ribosomal

protein L13A gene (Rpl13a) which encodes a ribosomal protein thatis a component of the 60S subunit (Mazumder et al. 2003). Asreported in Appendix S1, preliminary experiments were performedto validate Rpl13a as a stable housekeeping gene. After performingquantitative PCR, the results were analyzed using the comparativecycle threshold method (Livak and Schmittgen 2001). Changes inmRNA expression were relative to the respective normoxic orhypoxic retinas after normalization to Rpl13a. All reactions wererun as triplicate. After statistical analysis, the data from the differentexperiments were plotted and averaged in the same graph.

ELISAVEGF (content and release) and IGF-1 (content) were measuredusing commercially available kits (R&D Systems, Minneapolis,MN, USA) as detailed in Appendix S1. Either VEGF or IGF-1content was measured in the retina of OIR mice. VEGF content wasalso measured in retinal explants. For each experimental conditioneight samples from eight different mice, each containing two retinasfrom two different mice, were used. VEGF release was measured inthe culture medium of retinal explants. Eight samples for eachexperimental condition were used. In each experiment, all samplesand standards were measured in duplicate. Data were collected as pgVEGF or IGF-1 per mg of total protein (content) or pg VEGF permL culture medium (release) and, after statistics, averaged in thesame graph.

Western blot analysisWestern blot analysis using antibodies against mitogen-activatedprotein kinase 1, b2-ARs, VEGFR-1, VEGFR-2 and IGF-1R wasperformed on proteins extracted as previously described (Ristoriet al. 2011), and specified in Appendix S1. For each experimentalcondition eight samples from eight different mice, each containingtwo retinas from two different mice, were used. All experimentswere run as duplicate. After statistics, data were averaged in thesame graph.

PKA activityPKA activity was measured in the retina of OIR mice and in retinalexplants. For each experimental condition, eight samples from eightdifferent mice, each containing two retinas from two different mice,were used. PKA activity was measured with the PepTag non-radioactive PKA assay kit (Promega, Madison, MI, USA), asspecified in Appendix S1. Data were expressed as incorporatedpmol of phosphate per min per mg of protein and values representedthe mean of duplicate determination. After statistics, data wereaveraged in the same graph.

ImmunohistochemistryImmunohistochemistry on retinal whole mounts or 14 lm cryostatsections was performed in line with previous works (Dal Monteet al. 2007; Ristori et al. 2011), as detailed in Appendix S1. A ratCD31 monoclonal antibody was used to visualize blood vessels inretinal whole mounts in agreement with previous studies in themouse retina (Shen et al. 2007; Cao et al. 2010). To localize b2-ARin retinal sections, a b2-AR rabbit polyclonal antibody (Walker andSteinle 2007) was used either alone or in combination with aglutamine synthetase mouse monoclonal antibody, which is knownto label Muller cells in the mouse retina (Dal Monte et al. 2007).

� 2011 The AuthorsJournal of Neurochemistry � 2011 International Society for Neurochemistry, J. Neurochem. (2011) 119, 1317–1329

b2-AR blockade ameliorates ROP | 1319

The b2-AR rabbit polyclonal antibody specificity was evaluated bythe use of pre-immune serum instead of the primary antibody, bypre-adsorption with the corresponding blocking peptide and by theuse of rabbit IgG negative control instead of the primary antibody.As secondary antibodies, Alexa Fluor 488 and/or Alexa Fluor 546were used. Immunofluorescent materials were observed withconfocal microscopy (Laser Scanning Microscope Radiance Plus;Bio-Rad) using ·10 or ·40 objective lens. In CD31-immunolabeledwhole mounts, pre-retinal neovascular tufts could be readilydistinguished focusing just above the inner limiting membrane ofthe whole mount. Serial optical sections (1 lm apart) were scannedthrough the thickness of each whole mount or section at the samepre-determined z-axis position.

Quantitative analysis of CD31 immunohistochemistryQuantitative analysis of pre-retinal neovascular tuft formation andavascular areas were performed in accordance with previous worksin the mouse model of OIR (Banin et al. 2006; Zou et al. 2010), asdetailed in Appendix S1. For each experimental condition, quan-titative data originated from eight retinas from eight different mice.After statistical analysis, averaged data were plotted in the samegraph.

ElectroretinographyScotopic full-field ERG was recorded from anesthetized mice darkadapted overnight according to Vessey et al. (2011), as specified inAppendix S1. Mixed (rod and cone) ERGs were evoked by flashesof light intensities ranging from )3.4 to 1 log cd-s/m2 generatedthrough a Ganzfeld stimulator (Biomedica Mangoni, Pisa, Italy).The a- and b-waves, as well as OPs were measured. Intensity-response functions for the b-wave were fit to a modified Naka-Rushton function (Naka and Rushton 1966): V(I ) = V0 + (Vmax Æ In)/(In + kn). V is the amplitude of the b-wave (in lV), I is the stimulusintensity (in log cd-s/m2), V0 is the non-zero baseline effect, Vmax isthe saturated amplitude of the b-wave (in lV) and refers to the post-receptor response amplitude, k is the stimulus intensity that evokes ab-wave of half-maximum amplitude (in log cd-s/m2) and refers tothe retina sensitivity, n, which was constrained to unity, is adimensionless constant controlling the slope of the function. Of thefive OPs that can be isolated from the mouse ERG, only OP2, OP3and OP4 were analyzed as OP1 and OP5 are not measurable atPD17 in OIR mice (Vessey et al. 2011). For each OP, the trough-to-peak amplitude was measured and the wavelet amplitudes wereadded to create the Sum OPs or SOPs (OP2 + OP3 + OP4)(Bresnick and Palta 1987). Mean amplitudes of ERG responseswere plotted as a function of increasing light intensity. For eachexperimental condition, ERG analysis was performed on nine mice.

Statistical analysisAll data were analyzed by the Shapiro-Wilk test to verify theirnormal distribution with the exception of those from experimentsperformed to validate Rpl13a as a stable housekeeping gene whichwere analyzed by the Kolmogorov–Smirnov test. In all experimentsusing OIR mice with the exception of ERG experiments, one-wayANOVA followed by Newman–Keuls Multiple Comparison post-testwas used. In experiments using retinal explants, unpaired Student’st-test or one-way ANOVA followed by Newman–Keuls MultipleComparison post-test were used as appropriate. Two-way ANOVA for

repeated measures was performed on ERG responses. If anystatistically significant difference was found the data were furtheranalyzed using post hoc comparison Bonferroni test. Naka-Rushtonparameters were analyzed using unpaired t-test. The results wereexpressed as means ± SE of the indicated n values. Prism 4(GraphPad Software, San Diego, CA, USA) was used to analyzedata. Differences with p < 0.05 were considered statisticallysignificant.

ChemicalsRetinal culture reagents were from Lonza (Basel, Switzerland),except Fungizone (Sigma-Aldrich, St Louis, MO, USA). The PCRmastermix (iQ Sybr Green Supermix) was from Bio-Rad. All thematerials for siRNA transfection were from Qiagen. Primers wereobtained from MWG Biotech (Ebersberg, Germany). Antibodieswere from Santa Cruz Biotechnologies (Santa Cruz, CA, USA),except antibodies against CD31 (BD Pharmingen, San Diego, CA,USA) and glutamine synthetase (Chemicon, Temecula, CA, USA).Alexa Fluor 488 and Alexa Fluor 546 were from Molecular Probes(Eugene, OR, USA). The enhanced chemiluminescence reagent wasfrom Millipore (Billerica, MA, USA). All other chemicals, includingatenolol, ICI 118,551 and SR59230A, were from Sigma-Aldrich.

Results

Role of b-AR antagonists on proangiogenic factorsIn OIR mice, we evaluated whether selective b-AR antag-onists affected retinal levels of VEGF, IGF-1 and theirreceptors. In all experiments, no effects were observed aftervehicle treatment (not shown). In agreement with previousresults in OIR mice (Ristori et al. 2011), hypoxic retinasdisplayed �109% (p < 0.01 vs. normoxic) and �365%(p < 0.001 vs. normoxic) increase in VEGF mRNA (Fig. 1a)and protein (Fig. 1b), respectively. In addition, hypoxiasignificantly increased VEGFR-2 mRNA by �92%(p < 0.01 vs. normoxic, Fig. 1c) and VEGFR-2 protein by�65% (p < 0.001 vs. normoxic, Fig. 1d). Hypoxic levels ofVEGF and VEGFR-2 mRNA and protein were unaffected byatenolol or SR59230A. ICI 118,551 at 1 or 5 mg/kg/dose didnot influence VEGF and VEGFR-2 mRNA and protein,whereas at 30 mg/kg/dose decreased hypoxic levels ofVEGF mRNA by �39% (p < 0.01 vs. hypoxic, Fig. 1a),VEGF protein by �54% (p < 0.001 vs. hypoxic, Fig. 1b),VEGFR-2 mRNA by �37% (p < 0.01 vs. hypoxic, Fig. 1c)and VEGFR-2 protein by �33% (p < 0.001 vs. hypoxic,Fig. 1d). Hypoxia increased VEGFR-1 mRNA by �47%(p < 0.01 vs. normoxic, Figure S2a) and VEGFR-1 proteinby �68% (p < 0.001 vs. normoxic, Figure S2b). VEGFR-1mRNA and protein were unaffected by atenolol, ICI 118,551or SR59230A (Figure S2a and b). According to previousresults in OIR mice (Ristori et al. 2011), hypoxia increasedIGF-1 mRNA and protein by �62% (p < 0.01 vs. normoxic,Fig. 1e) and �45% (p < 0.001 vs. normoxic, Fig. 1f),respectively. Hypoxic levels of IGF-1 mRNA and proteinwere unaffected by atenolol or SR59230A. ICI 118,551 at 1



or 5 mg/kg/dose did not influence IGF-1 mRNA and protein,whereas at 30 mg/kg/dose decreased IGF-1 mRNA by�30% (p < 0.01 vs. hypoxic, Fig. 1e) and IGF-1 proteinby �32% (p < 0.001 vs. hypoxic, Fig. 1f). Hypoxiaincreased IGF-1R mRNA and protein by �56% (p < 0.01vs. normoxic, Figure S3a) and �46% (p < 0.001 vs. norm-oxic, Figure S3b), respectively. IGF-1R mRNA and proteinwere unaffected by atenolol, ICI 118,551 or SR59230A(Figure S3a and b).

Pre-retinal tuft formation and intra-retinal vascularizationTo investigate the effects of b-AR blockade on angiogenesis,we performed CD31 immunohistochemistry in the retina of

OIR mice. Figure 2 shows the vascular pattern of flat mountsin normoxic (Fig. 2a) and OIR mice treated with eithervehicle (Fig. 2b) or b-AR antagonists (Fig 2c–e). In OIRmice, a large avascular area could be observed in the centerof the retina. The mid-peripheral region was instead charac-terized by excessive regrowth of abnormal superficial vesselsleading to pre-retinal neovascular tufts with most of themforming at the border between the vascularized peripheraland the obliterated central regions. No differences in theextent of CD31-positive neovascular tufts and avascular areawere observed between mice treated with vehicle and thosetreated with atenolol at 30 mg/kg/dose or SR59230A at20 mg/kg/dose (Fig. 2c and e, respectively). In contrast, ICI

Fig. 1 Levels of VEGF, VEGFR-2 and IGF-1 mRNA and protein after

b-AR antagonists. VEGF, VEGFR-2 and IGF-1 mRNA (a, c, e) were

evaluated by quantitative real-time RT-PCR. VEGF and IGF-1 protein

(b, f) were evaluated by ELISA, whereas VEGFR-2 protein (d) was

evaluated by western blot. VEGF, VEGFR-2 and IGF-1 mRNA and

protein were increased by hypoxia (*p < 0.01 and **p < 0.001 vs.

normoxic; one-way ANOVA followed by Newman–Keuls multiple com-

parison test). Hypoxic mice were treated with atenolol at 10 (+10), 20

(+20) or 30 (+30) mg/kg/dose; ICI 118,551 at 1 (+1), 5 (+5) or 30

(+30) mg/kg/dose; SR59230A at 5 (+5), 10 (+10) or 20 (+20) mg/kg/

dose. Either atenolol and SR59230A or ICI 118,551 at 1 and 5 mg/

kg/dose did not affect hypoxic levels of VEGF, VEGFR-2 and IGF-1

mRNA and protein. ICI 118,551 at 30 mg/kg/dose reduced hypoxic

levels of VEGF, VEGFR-2 and IGF-1 mRNA and protein (§p < 0.01

and §§p < 0.001 vs. hypoxic; one-way ANOVA followed by Newman–

Keuls multiple comparison test). In quantitative real-time RT-PCR

experiments, data were analyzed by the cycle threshold (CT) method

using Rpl13a as internal standard. Each column represents the

mean ± SE of data from 12 samples. In ELISA and western blot

experiments, each column represents the mean ± SE of data from

eight samples. In western blot experiments, protein levels were

evaluated using b-actin as the loading control. OD, optical density.



118,551 at 30 mg/kg/dose reduced drastically the vessel tuftarea, but did not influence the extent of the avascular area(Fig. 2d). Quantitative analysis confirmed our qualitativeobservations and revealed that retinas of pups treated withICI 118,551 had a significantly reduced neovascular tuft area(�62%, p < 0.001 vs. vehicle-treated hypoxic, Fig. 2f). Noeffects of b-AR antagonists were observed on the avasculararea (Figure S4).

PKA activity after b-AR blockadeTo investigate whether b-AR blockade was effective in theretina, we measured PKA activity used as a biomarker thatb-AR blockers were reaching the retina and eliciting anormal cellular response, according to a previous study in theretina of diabetic rats (Jiang et al. 2010). In all experiments,no effects were observed after vehicle treatment (not shown).As shown in Fig. 3a, in the retina of OIR mice, hypoxiadecreased PKA activity by �47% (p < 0.001 vs. normoxic).After atenolol at 30 mg/kg/dose, PKA activity was �23%higher than in hypoxia (p < 0.001 vs. hypoxic), but �35%lower than in normoxia (p < 0.001 vs. normoxic). Either ICI118,551 at 30 mg/kg/dose or SR59230A at 20 mg/kg/doseabolished the effect of hypoxia on PKA activity, which didnot differ significantly from that measured in normoxia. Asshown in Figure 3b, in retinal explants, hypoxia decreasedPKA activity by �48% (p < 0.001 vs. normoxic). Afteratenolol at 10 lM, ICI 118,551 at 10 lM or SR59230A at10 lM, PKA activity was �49%, �47% and �45% higherthan in hypoxia (p < 0.001 vs. hypoxic), respectively, but�23%, �24% and �25% lower than in normoxia (p < 0.001vs. normoxic), respectively.

Selectivity of ICI 118,551 effectsTo investigate whether the effect of ICI 118,551 on OIR wasrelated to b2-AR antagonism, we used retinal explants culturedin hypoxia in which b2-ARs were knocked-down using asiRNA approach and we determined the effect of ICI 118,551on VEGF content (mRNA and protein) and release. ICI118,551 was administered at concentrations ranging from 0.1to 10 lM. No effects were observed after vehicle treatment(not shown). As shown in Figure 4, hypoxia increased VEGFmRNA (�179%, p < 0.001 vs. normoxic, Fig. 4a) andprotein (�271%, p < 0.001 vs. normoxic, Fig. 4b), as wellas VEGF release (�365%, p < 0.001 vs. normoxic, Fig. 4c).ICI 118,551 at 0.1 lM did not affect hypoxic levels of VEGFcontent and release. ICI 118,551 at 1 lM decreased hypoxiclevels of VEGF mRNA (�19%, p < 0.001 vs. hypoxic,Fig. 4a) and protein (�46%, p < 0.001 vs. hypoxic, Fig. 4b),as well as VEGF release (�39%, p < 0.001 vs. hypoxic,Fig. 4c). ICI 118,551 at 10 lMabolished the effect of hypoxiaon VEGF content and release, which did not differ signifi-cantly from those measured in normoxia. b2-AR silencingreduced the hypoxia-induced increase in VEGF content andrelease. In fact, after b2-AR-siRNA3, VEGF mRNA andprotein were lower than in hypoxia (�28% and �27%,p < 0.001 vs. hypoxic, respectively, Fig. 4a and b), but higherthan in normoxia (�102% and �169%, p < 0.001 vs.normoxic, respectively, Fig. 4a and b). In addition, VEGFrelease was lower than in hypoxia (�31%, p < 0.001 vs.hypoxic, Fig. 4c), but higher than in normoxia (�223%,p < 0.001 vs. normoxic, Fig. 4c). ICI 118,551 at 10 lM didnot affect VEGF content and release in hypoxic explantstreated with b2-AR-siRNA3.

(a) (b)

(c) (d)

(e) (f)

Fig. 2 Flat-mounted retinas immunolabeled with a rat monoclonal

antibody directed to CD31. Mice exposed to room air (a) or to

75% ± 2% oxygen from PD7 to PD12 (b–e), vehicle-treated (b) and

treated with atenolol at 30 mg/kg/dose (c), ICI 118,551 at 30 mg/kg/

dose (d) or SR59230A at 20 mg/kg/dose (e) from PD12 to PD16.

Hyperoxia followed by normoxia for 5 days produced the central loss

of blood vessels and the formation of tufts. No apparent differences in

the extent of CD31-positive neovascular tufts and avascular area were

observed between mice treated with vehicle and those treated with

atenolol or SR59230A. In contrast, ICI 118,551 reduced drastically the

vessel tuft area, but did not influence the extent of the avascular area.

The extent of neovascular tuft area was quantitatively evaluated (f).

*p < 0.001 vs. vehicle-treated hypoxic; one-way ANOVA followed by

Newman–Keuls multiple comparison test. Each column represents the

mean ± SE of data from eight retinas. Scale bar, 1 mm.



ERG responsesIn OIR mice, we determined the effects of ICI 118,551 onERG responses to light stimulation. Representative mixed a-,b-waves, and OPs recorded from normoxic, OIR and ICI118,551 at 30 mg/kg/dose-treated mice are shown in Fig. 5a.In Fig. 5b and c, a- and b-wave amplitudes averaged as afunction of increasing light intensities are reported innormoxic and OIR mice. An increase in a- and b-waveamplitudes with increasing stimulus intensity was observed.A clear a-wave developed at a light intensity of approx-imately )1.6 log cd-s/m2. As shown in Fig. 5d, in normoxicmice SOPs increased as the light intensity was increased.Consistent with previous findings (Liu et al. 2006a; Vesseyet al. 2011), OIR mice displayed significantly reduced a- andb-wave amplitudes compared with normoxic mice (a-wave:p < 0.01 for )1.6 log cd-s/m2 and p < 0.001 from )1 to

1 log cd-s/m2 vs. normoxic, Fig. 5b; b-wave: p < 0.001from )2.1 to 1 log cd-s/m2 vs. normoxic, Fig. 5c). Asdemonstrated by a Naka-Rushton function, Vmax was260.3 ± 7.93 lV, whereas k was )1.46 ± 0.14 log cd-s/m2.In OIR mice, Vmax and k were 190.4 ± 6.66 lV and)1.06 ± 0.12 log cd-s/m2, respectively. As shown by thesample fits in Fig. 5c, hypoxia significantly reduced the Vmax

and k to approximately 73.1% and 72.6% of the normoxicvalues (p < 0.0001 and p < 0.05 vs. normoxic, respec-tively). As previously described (Liu et al. 2006b; Vesseyet al. 2011), hypoxia significantly reduced SOPs across thelight intensities analyzed (p < 0.01 for )0.3 log cd-s/m2,p < 0.001 from 0 to 1 log cd-s/m2 vs. normoxic, Fig. 5d).Neither vehicle treatment influenced retinal responses innormoxic and OIR mice, nor ICI 118,551 affected theamplitude of ERG components in normoxic mice (notshown). As shown in Fig. 5b and c, ICI 118,551 significantlyincreased a- and b-wave amplitudes as compared withuntreated hypoxic mice (a-wave: p < 0.01 for )1.6 log cd-s/m2,p < 0.001 from )1 to 1 log cd-s/m2 vs. untreated hypoxic;b-wave: p < 0.05 for )2.1 log cd-s/m2, p < 0.001 from )1.6to 1 log cd-s/m2 vs. untreated hypoxic). In ICI 118,551-treated mice, Vmax and k were 262 ± 5.40 lV and)1.21 ± 0.1 log cd-s/m2, respectively, and did not signifi-cantly differ from values measured in normoxic mice. Asshown in Fig. 5d, ICI 118,551 significantly increased SOPsas compared with untreated hypoxic mice (p < 0.05 for0 log cd-s/m2, p < 0.01 for 0.7 and 1 log cd-s/m2 vs.untreated hypoxic). After ICI 118,551, SOPs were similarto those measured in normoxic mice.

Fig. 3 Measurements of PKA activity after b-AR antagonists. PKA

activity was measured in the retina of OIR mice (a) and in hypoxic

retinal explants (b). Representative agarose gels containing phos-

phorylated and non-phosphorylated kemptide bands are shown. The

positive control (+) was provided by the manufacturer. In the negative

control ()), PKA activity was measured in the absence of the retinal

cytosol fraction. PKA activity was quantified spectrophotometrically

from bands excised from the gel. PKA activity was decreased by hy-

poxia (*p < 0.001 vs. normoxic; one-way ANOVA followed by Newman–

Keuls multiple comparison test). After b-AR blockers (atenolol at

30 mg/kg/dose, ICI 118,551 at 30 mg/kg/dose or SR59230A at 20 mg/

kg/dose in OIR mice; atenolol at 10 lM, ICI 118,551 at 10 lM or

SR59230A at 10 lM in retinal explants), PKA activity was higher than

in hypoxia (§p < 0.001 vs. hypoxic; one-way ANOVA followed by New-

man–Keuls multiple comparison test). In the retina of OIR mice, PKA

activity recovered normoxic values after ICI 118,551 or SR59230A.

After atenolol, PKA activity was lower than in normoxia (*p < 0.001 vs.


parison test). In retinal explants, PKA activity was lower than in

normoxia after atenolol, ICI 118,551 or SR59230A (*p < 0.001 vs.


parison test). Each column represents the mean ± SE of data from

eight samples.



b2-AR localizationThe ameliorative effects of ICI 118,551 indicated a pivotalrole of b2-AR on retinal angiogenesis and visual impairmentsleading to investigate the localization of b2-ARs in the retinaThe representative images of Fig. 6 are taken from retinalsections of normoxic mice. The controls for immunohisto-chemistry resulted in the absence of immunoreactivity (IR)(Fig. 6a and b). As shown in Fig. 6c and d, dense b2-AR-IRwas present in the outer plexiform layer as well as invertically oriented fibers through the outer nuclear layer

(ONL) directed to the outer limiting membrane. b2-AR-IRwas also localized to sparse fibers directed to the innerplexiform layer, several cell profiles in the inner nuclear layer(INL) and numerous processes in the ganglion cell layer(GCL). Colocalization studies (Fig. 6d–f) indicate that a fewb2-AR immunolabeled cell profiles in the inner nuclear layerexpressed glutamine synthetase-IR, a marker of Muller cells.In contrast, the great majority of b2-AR immunoreactivevertically oriented fibers in the ONL and some of theimmunolabeled processes observed in the GCL wereobserved to express glutamine synthetase-IR, indicatingb2-AR expression by Muller cell fibers in the ONL andtheir endfeets in the GCL. Comparable b2-AR distributionpattern was found in retinal sections of OIR mice (notshown).

Discussion

Effects of b-AR blockade on proangiogenic factors andpathogenic neovascularizationPrevious findings indicate that the noradrenergic systemplays an important role in retinal angiogenesis. For instance,in human choroidal endothelial cells, b-AR activation withisoproterenol increases growth factors that are active duringvascular remodeling (Steinle et al. 2008), whereas in diabeticrats, isoproterenol inhibits vascular remodeling that occurs indiabetic retinopathy (DR) suggesting beneficial effects ofb-AR activation (Jiang et al. 2010). In rats, b-AR blockadewith propranolol produces retinal alterations common to DR

Fig. 4 Effect of ICI 118,551 on VEGF content (mRNA and protein)

and release in hypoxic retinal explants after b2-AR silencing. VEGF

mRNA (a) was evaluated by quantitative real-time RT-PCR, whereas

VEGF protein (b) and VEGF release in the culture medium (c) were

evaluated by ELISA. VEGF content and release were increased by

hypoxia (*p < 0.001 vs. normoxic; one-way ANOVA followed by New-

man–Keuls multiple comparison test). Hypoxic levels of VEGF content

and release were not affected by ICI 118,551 at 0.1 lM, but were

decreased by ICI 118,551 at 1 or 10 lM (§p < 0.001 vs. hypoxic; one-

way ANOVA followed by Newman–Keuls multiple comparison test).

VEGF content and release after ICI 118,551 at 1 lM were higher than

in normoxia (*p < 0.001 vs. normoxic; one-way ANOVA followed by

Newman–Keuls multiple comparison test), whereas they did not differ

from normoxic values after ICI 118,551 at 10 lM. b2-AR-siRNA3 re-

duced the hypoxia-induced increase in VEGF content and release

(§p < 0.001 vs. hypoxic; one-way ANOVA followed by Newman–Keuls

multiple comparison test). VEGF content and release after b2-AR-

siRNA3 were higher than in normoxia (*p < 0.001 vs. normoxic; one-

way ANOVA followed by Newman–Keuls multiple comparison test). ICI

118,551 at 10 lM did not affect VEGF content and release in hypoxic

explants treated with b2-AR-siRNA3. In quantitative real-time RT-PCR

experiments, data were analyzed by the cycle threshold (CT) method

using Rpl13a as internal standard. Each column represents the

mean ± SE of data from eight samples. In ELISA experiments, each

column represents the mean ± SE of data from eight samples.



(Jiang and Steinle 2010) and loss of sympathetic innervationor lack of norepinephrine causes retinal dysfunctions similarto those characteristic of DR (Wiley et al. 2005; Steinle et al.2009). However, propranolol has been recently shown toameliorate retinal angiogenesis in response to hypoxia inOIR mice (Ristori et al. 2011). Results of the present studymay help to elucidate the role that distinct b-ARs play onretinal angiogenesis.

As shown by the present results, atenolol does not affectretinal levels of proangiogenic factors and pathogenicneovascularization in OIR mice. This finding seems toexclude that b1-ARs play a role in regulating retinalangiogenesis in agreement with previous results in rats withDR (Phipps et al. 2007; Wilkinson-Berka et al. 2007). In thisline, the b1-AR agonists dobutamine and xamoterol do notaffect angiogenic phenotype in human choroidal and retinalendothelial cells (Steinle et al. 2003, 2005).

As also shown by the present results, ICI 118,551 reducesretinal levels of proangiogenic factors similarly to whatfound after propranolol (Ristori et al. 2011). Our additionalfinding that in the OIR model, ICI 118,551 reducesneovascular tuft formation without affecting intra-retinalvascularization demonstrates an angiostatic effect of b2-ARblockade on pathogenic angiogenesis and suggests a key roleof b2-ARs in the regulation of angiogenic processes in theproliferative phase of ROP. A decrease in neovascular tuftformation without effects on intra-retinal vascularization hasbeen observed in the mouse model of OIR after treatmentwith antiangiogenic compounds (Banin et al. 2006).

Our finding that in the OIR model, ICI 118,551 is noteffective until a dose of 30 mg/kg, raises the concern that theeffect of ICI 118,551 on OIR may be unrelated to b2-ARantagonism. Although no information is available in theretina, there are previous studies demonstrating that ICI118,551 is effective in mice at doses of 1–2 mg/kg (Yalcinet al. 2010; Yu et al. 2010) although there are also studiesdemonstrating ICI 118,551 effects at higher doses in rodents(Tan et al. 2003; Kitamura et al. 2007; Mantsch et al. 2010).In OIR mice, we have previously demonstrated that thedecrease in retinal VEGF after propranolol at 20 mg/kg issignificantly more pronounced than at 2 mg/kg (Ristori et al.2011). Propranolol is a highly lipophilic drug such as ICI118,551, achieving high concentrations in the brain. Inpreliminary studies, we have found that in mice withsubcutaneous administration of 20 mg/kg propranolol giventhree times a day from PD12 to PD16, the concentration ofpropranolol in the retina is 20.02 ± 3.21 lg/g (unpublishedresults). Although it is difficult to conjecture about the ICI118,551 concentration that actually reaches the retina, ourresults seem to indicate that sufficient ICI 118,551 penetratesthe blood–retinal barrier to achieve retinal levels adequate toelicit biological effects. However, pathogenic angiogenesis ischaracterized by chaotic and leaky blood vessels with high

Fig. 5 ICI 118,551 effects on ERG responses. (a) Representative ERG

waveforms in normoxic, hypoxic and ICI 118,551-treated mice. (b)

a-wave amplitudes (means ± SE) in normoxic (open squares; n = 9),

hypoxic (filled squares; n = 9) and ICI 118,551-treated (filled circles;

n = 9) mice plotted as a function of the stimulus intensity. (c) Fits of a

Naka-Rushton function to b-waves in normoxic (solid line), hypoxic

(thick dashed line) and ICI 118,551-treated (thin dashed line) mice. ICI

118,551 at 30 mg/kg/dose significantly increased the amplitude of a-

and b-waves as compared with untreated hypoxic mice (for the a-wave:

p < 0.01 for )1.6 log cd-s/m2, p < 0.001 from )1 to 1 log cd-s/m2 vs.

untreated hypoxic; for the b-wave: p < 0.05 for )2.1 log cd-s/m2,

p < 0.001 from )1.6 to 1 log cd-s/m2 vs. untreated hypoxic; two-way

ANOVA followed by Bonferroni post-test). In ICI 118,551-treated mice,

the amplitude of a- and b-waves was similar to that measured in

normoxic mice. (d) SOPs (means ± SE) in normoxic (open squares;

n = 9), hypoxic (filled squares; n = 9) and ICI 118,551-treated (filled

circles; n = 9) mice plotted as a function of the stimulus intensity. In ICI

118,551-treated mice, SOPs were larger than in untreated hypoxic mice

(p < 0.05 for 0 log cd-s/m2, p < 0.01 from 0.7 to 1 log cd-s/m2 vs. un-

treated hypoxic; two-way ANOVA followed by Bonferroni post-test) and

were not different from those measured in normoxic mice.



interstitial fluid pressure and inefficient blood flow, whichcan impede drug delivery to the tissue (Yang et al. 2005). Inaddition, hypoxia and acidosis consequent to poor oxygendelivery may also contribute to reduced drug uptake by cells.This may explain the fact that in the retina of OIR mice, ICI118,551 is not effective until a higher dose of 30 mg/kg.Finally, the fact that ICI 118,551 is not effective in hypoxicretinal explants in which b2-ARs have been knocked downwith siRNA seems to exclude that ICI 118,551 can exert anoff-target effect and suggests that ICI 118,551 acts specif-ically at b2-ARs. In this respect, the result that b2-ARsilencing reduces VEGF accumulation in response tohypoxia adds further evidence to the hypothesis thatb2-ARs play a key role in regulating the levels of proangi-ogenic factors in the retina.

Our finding that b3-AR blockade with SR59230A iswithout effects on both proangiogenic factors and pathogenicvascularization seems to exclude that b3-ARs play a role inregulating angiogenesis in the retina although previousresults suggest that b3-ARs can be related to pathogenicangiogenesis (Steinle et al. 2003; Ristori et al. 2011).

However, the use of SR59230A may lead in some cases tomisleading conclusions and many responses attributed tob3-AR blockade may need reevaluation in light of thedevelopment of more selective blockers. For instance,SR59230A may also act as partial agonist at b3-ARs (Vrydagand Michel 2007) or it may not suppress the receptorconstitutive activity, which is typical of b3-ARs (Perrone andScilimati 2010).

Effectiveness of b-AR blockadeIt is well known that activated b-ARs couple to Gas toactivate adenylyl cyclase, which induces intracellular cAMPconcentration and subsequently stimulates PKA. To ourknowledge, there is no information on the effects of hypoxiaon PKA activity in the retina. However, several studies havedemonstrated that PKA activity may be either increased orreduced by hypoxia in different experimental models(Beitner-Johnson et al. 1998; Millen et al. 2006; Wang andYang 2009; Zhang et al. 2010). Our results demonstrate forthe first time that hypoxia reduces PKA activity in the retinaof OIR mice as well as in hypoxic retinal explants,

(a) (b) (c)

(d) (e) (f)

Fig. 6 Immunohistochemical studies of b2-AR distribution pattern in

the mouse retina. Confocal images from retinal sections that are

representative of the results obtained in three retinas. Specificity of the

rabbit polyclonal antibody directed to the b2-AR evaluated with the

blocking peptide solution (a) and in the presence of rabbit IgG negative

control instead of the primary antibody (b). (c, d) b2-AR immunola-

beling was mainly observed in vertically directed processes in the

ONL, in the OPL, in cell profiles in the INL and in processes in the

GCL. (e) The same field as in (d) showing glutamine synthetase im-

munolabeling. (f) Merged image showing the overlapping (yellow) of

b2-AR (green) and glutamine synthetase (red) immunoreactivities.

ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nu-

clear layer; IPL, inner plexiform layer; GCL, ganglion cell layer. Scale

bar, 20 lm.



suggesting that down-regulation of PKA activity is involvedin mechanisms by which retina adapts to hypoxia. Inaddition, the present results provide the first demonstrationthat hypoxia effects on PKA activity are influenced by b-ARblockade indicating the effectiveness of b-AR blockers. Ouradditional result that in the retina of OIR mice, b-ARblockers modulate PKA activity similarly to what found inretinal explants adds further support to the hypothesis that b-AR antagonists administered in vivo reach the retina atconcentrations that selectively affect the b-AR-mediatedsignaling. The fact that in the retina of OIR mice, PKAactivity completely recovers its normoxic value after ICI118,551 or SR59230A, whereas recovers it partially afteratenolol is in line with the finding that atenolol displaysmuch lower tissue partition coefficients as compared withlipophilic b-AR antagonists (Kadam and Kompella 2010).

Effects of b2-AR blockade on ERG amplitudesAs shown by the present results, ICI 118,551 is effective inincreasing ERG responses indicating that b2-AR blockadecan result in improved retinal function. Our hypothesis is thatin OIR mice, vascular protection by b-AR blockade mayresult in ameliorated retinal function by an associatedbeneficial effect on both photoreceptors and post-receptorcells in the neural retina. In this respect, the retinalvasculature and the neural retina are in close physicalproximity, are immature at the same ages and develop underthe cooperative control of growth factors that direct bothangiogenesis and neurogenesis (Akula et al. 2008). Ourresult that b-AR blockade increases ERG responses in OIR isin apparent contrast with the finding that b-AR activationameliorates retinal function in DR (Jiang et al. 2010) andsuggests that b-AR control on angiogenesis can be regulatedby diverse mechanisms in OIR and DR. This possibility issupported by the fact that propranolol reduces drasticallyVEGF accumulation in the retina of OIR mice (Ristori et al.2011), whereas does not affect or even increases VEGFproduction in the retina of rats with DR (Zheng et al. 2007;Jiang and Steinle 2010). The additional fact that ICI 118,551improves ERG amplitudes in OIR mice when Jiang andSteinle (2010) have found that propranolol producesdecreased ERG amplitudes in normal rats can be explainedby assuming that b2-AR blockade acts differently onproliferating retinal vessels than on normal vessels. In thisrespect, functional, morphologic and molecular differencesbetween proliferating blood vessels and normal vasculaturehave been described (Dua et al. 2005; Yang et al. 2005;Huang et al. 2006). In line with this hypothesis, ICI 118,551does not affect ERG in control mice. In addition, differentdosage and administration route in mice and rats (ICI118,551 at 30 mg/kg three times a day for five days in micewhile propranolol at 1 mg/kg each day delivered by osmoticpumps at a rate of 2.5 lL/h for 21 days in rats) may alsoexplain the fact that ERG amplitudes are unaffected by ICI

118,551 in control mice, whereas propranolol producesdecreased ERG amplitudes in rats.

b2-AR localizationAs shown by our immunohistochemical studies, several cellsand processes in the mouse retina express b2-ARs. Inparticular, most b2-ARs-immunoreactive fibers passingthrough the ONL and some b2-ARs-immunoreactive pro-cesses in the GCL belong to Muller cells as demonstrated byour colocalization studies. This is in line with previousfindings demonstrating that cultured Muller cells of the ratretina express b2-AR-IR (Walker and Steinle 2007). Inter-estingly, the distribution pattern of b2-AR-IR found in thepresent study almost coincides with that of immunoreactivityfor VEGF, IGF-1 and their receptors as observed in themouse retina (Dal Monte et al. 2007). Thus, our hypothesisis that b2-AR blockade inhibits pathogenic vascularizationthrough a control of angiogenic factor production by retinalcells. In this respect, the fact that b2-ARs are localized toMuller cells suggests the possibility that b2-ARs play a rolein regulating VEGF production by these cells. Indeed, Mullercells are known to express VEGF and hypoxia has beenshown to increase their VEGF expression and release (Yafaiet al. 2004; Bai et al. 2009; Yanni et al. 2010). In addition,Muller cell-derived VEGF significantly contributes to pro-liferative neovascularization in the retina of OIR mice (Baiet al. 2009) and Muller cells seem to be involved inproliferative DR in rats (Guidry 2005).

Conclusions and clinical implications

Most of recent therapeutic interventions against ROP havefocused on the mechanisms and the factors leading to newvessel growth. Our results suggest that pathological angio-genesis, a key change in many hypoxic/ischemic vision-threatening retinal diseases, depends at least in part on b2-ARactivity and indicate that b2-AR blockade can effectivelymodulate angiogenic processes in the retina. Althoughextrapolation of these experimental findings to the humansituation of ROP is difficult, these results may help to explorethe possible therapeutic use of beta blockers in ROP. In thisrespect, encouraging experiences have already been under-taken in a pilot clinical study performed to evaluate safetyand efficacy of propranolol administration in pre-termnewborns with ROP (Filippi et al. 2010).

Acknowledgements

Supported by the Meyer Foundation – ‘A. Meyer’ UniversityChildren’s Hospital and by the Foundation of the Cassa di Risparmiof Livorno to PB. The authors declare that they have no conflict ofinterest. The authors thank Prof. Maria Francesca Romano for herhelp in statistical analysis. The authors also thank Dr AngeloGazzano and Gino Bertolini for assistance with the mouse colonies.



Supporting information

Additional supporting information may be found in the onlineversion of this article:

Appendix S1. Supplementary methods.Table S1 Primers used for PCR analysis.Figure S1. Evaluation of siRNA transfection in hypoxic explants.Figure S2. Levels of VEGFR-1 mRNA and protein after b-AR

antagonists.Figure S3. Levels of IGF-1R mRNA and protein after b-AR

antagonists.Figure S4. Quantitative analysis of CD31 immunoreactivity.As a service to our authors and readers, this journal provides

supporting information supplied by the authors. Such materials arepeer-reviewed and may be re-organized for online delivery, but arenot copy-edited or typeset. Technical support issues arising fromsupporting information (other than missing files) should beaddressed to the authors.

References

Abraham G., Kneuer C., Ehrhardt C., Honscha W. and Ungemach F. R.(2004) Expression of functional beta2-adrenergic receptors in thelung epithelial cell lines 16HBE14o(-), Calu-3 and A549. Biochim.Biophys. Acta 1691, 169–179.

Akula J. D., Mocko J. A., Benador I. Y., Hansen R. M., Favazza T. L.,Vyhovsky T. C. and Fulton A. B. (2008) The neurovascular rela-tion in oxygen-induced retinopathy. Mol. Vis. 14, 2499–2508.

Al Zubair K., Bexis S. and Docherty J. R. (2008) Relaxations to beta-adrenoceptor subtype selective agonists in wild-type and NOS-3-KO mouse mesenteric arteries. Eur. J. Pharmacol. 587, 216–223.

Bai Y., Ma J. X., Guo J., Wang J., Zhu M., Chen Y. and Le Y. Z. (2009)Muller cell-derived VEGF is a significant contributor to retinalneovascularization. J. Pathol. 219, 446–454.

Banin E., Dorrell M. I., Aguilar E., Ritter M. R., Aderman C. M., SmithA. C., Friedlander J. and Friedlander M. (2006) T2-TrpRS inhibitspreretinal neovascularization and enhances physiological vascularregrowth in OIR as assessed by a new method of quantification.Invest. Ophthalmol. Vis. Sci. 47, 2125–2134.

Beitner-Johnson D., Leibold J. and Millhorn D. E. (1998) Hypoxiaregulates the cAMP- and Ca2+/calmodulin signaling systems inPC12 cells. Biochem. Biophys. Res. Commun. 242, 61–66.

Bresnick G. H. and Palta M. (1987) Oscillatory potential amplitudes.Relation to severity of diabetic retinopathy. Arch. Ophthalmol.105, 929–933.

Cao R., Xue Y., Hedlund E. M., Zhong Z., Tritsaris K., Tondelli B.,Lucchini F., Zhu Z., Dissing S. and Cao Y. (2010) VEGFR1-mediated pericyte ablation links VEGF and PlGF to cancer-asso-ciated retinopathy. Proc. Natl. Acad. Sci. USA 107, 856–861.

Chen J. and Smith L. E. (2007) Retinopathy of prematurity. Angiogen-esis 10, 133–140.

Dal Monte M., Cammalleri M., Martini D., Casini G. and Bagnoli P.(2007) Antiangiogenic role of somatostatin receptor 2 in a model ofhypoxia-induced neovascularization in the retina: results fromtransgenic mice. Invest. Ophthalmol. Vis. Sci. 48, 3480–3489.

Deng J., Muthu K., Gamelli R., Shankar R. and Jones S. B. (2004)Adrenergic modulation of splenic macrophage cytokine release inpolymicrobial sepsis. Am. J. Physiol. Cell Physiol. 287, C730–C736.

Dua R. S., Gui G. P. and Isacke C. M. (2005) Endothelial adhesionmolecules in breast cancer invasion into the vascular and lymphaticsystems. Eur. J. Surg. Oncol. 31, 824–832.

Filippi L., Cavallaro G., Fiorini P. et al. (2010) Study protocol: safetyand efficacy of propranolol in newborns with Retinopathy ofPrematurity (PROP-ROP): ISRCTN18523491. BMC Pediatr. 10,83.

Fulton A. B., Hansen R. M., Moskowitz A. and Akula J. D. (2009) Theneurovascular retina in retinopathy of prematurity. Prog. Retin.Eye. Res. 28, 452–482.

Guidry C. (2005) The role of Muller cells in fibrocontractive retinaldisorders. Prog. Retin. Eye Res. 24, 75–86.

Hoffmann C., Leitz M. R., Oberdorf-Maass S., Lohse M. J. and Klotz K.N. (2004) Comparative pharmacology of human beta-adrenergicreceptor subtypes-characterization of stably transfected receptors inCHO cells. Naunyn Schmiedebergs Arch. Pharmacol. 369, 151–159.

Huang Y., Shi H., Zhou H., Song X., Yuan S. and Luo Y. (2006) Theangiogenic function of nucleolin is mediated by vascular endo-thelial growth factor and nonmuscle myosin. Blood 107, 3564–3571.

Jiang Y. and Steinle J. J. (2010) Systemic propranolol reduces b-waveamplitude in the ERG and increases IGF-1 receptor phosphoryla-tion in rat retina. Invest. Ophthalmol. Vis. Sci. 51, 2730–2735.

Jiang Y., Walker R. J., Kern T. S. and Steinle J. J. (2010) Application ofisoproterenol inhibits diabetic-like changes in the rat retina. Exp.Eye Res. 91, 171–179.

Jun J. Y., Choi S., Yeum C. H., Chang I. Y., Park C. K., Kim M. Y.,Kong I. D., So I., Kim K. W. and You H. J. (2004) Noradrenalineinhibits pacemaker currents through stimulation of beta 1-adreno-ceptors in cultured interstitial cells of Cajal from murine smallintestine. Br. J. Pharmacol. 141, 670–677.

Kadam R. S. and Kompella U. B. (2010) Influence of lipophilicity ondrug partitioning into sclera, choroid-retinal pigment epithelium,retina, trabecular meshwork, and optic nerve. J. Pharmacol. Exp.Ther. 332, 1107–1120.

Kitamura H., Shiva D., Woods J. A. and Yano H. (2007) Beta-adrenergicreceptor blockade attenuates the exercise-induced suppression ofTNF-alpha in response to lipopolysaccharide in rats. Neuroimmu-nomodulation 14, 91–96.

Lamy S., Lachambre M. P., Lord-Dufour S. and Beliveau R. (2010)Propranolol suppresses angiogenesis in vitro: inhibition of prolif-eration, migration, and differentiation of endothelial cells. Vascul.Pharmacol. 53, 200–208.

Leaute-Labreze C., Dumas de la Roque E., Hubiche T., Boralevi F.,Thambo J. B. and Taıeb A. (2008) Propranolol for severehemangiomas of infancy. N. Engl. J. Med. 358, 2649–2651.

Lenard N. R., Gettys T. W. and Dunn A. J. (2003) Activation of beta2-and beta3-adrenergic receptors increases brain tryptophan.J. Pharmacol. Exp. Ther. 305, 653–659.

Liu K., Akula J. D., Falk C., Hansen R. M. and Fulton A. B. (2006a) Theretinal vasculature and function of the neural retina in a rat modelof retinopathy of prematurity. Invest. Ophthalmol. Vis. Sci. 47,2639–2647.

Liu K., Akula J. D., Hansen R. M., Moskowitz A., Kleinman M. S. andFulton A. B. (2006b) Development of the electroretinographicoscillatory potentials in normal and ROP rats. Invest. Ophthalmol.Vis. Sci. 47, 5447–5452.

Livak K. J. and Schmittgen T. D. (2001) Analysis of relative geneexpression data using real-time quantitative PCR and the 2(-DeltaDelta C(T)) Method. Methods 25, 402–408.

Mantsch J. R., Weyer A., Vranjkovic O., Beyer C. E., Baker D. A. andCaretta H. (2010) Involvement of noradrenergic neurotransmis-sion in the stress- but not cocaine-induced reinstatement ofextinguished cocaine-induced conditioned place preference inmice: role for b-2 adrenergic receptors. Neuropsychopharmacology35, 2165–2178.



Mazumder B., Sampath P., Seshadri V., Maitra R. K., DiCorleto P. E.and Fox P. L. (2003) Regulated release of L13a from the 60Sribosomal subunit as a mechanism of transcript-specific transla-tional control. Cell 115, 187–198.

Millen J., MacLean M. R. and Houslay M. D. (2006) Hypoxia-inducedremodelling of PDE4 isoform expression and cAMP handling inhuman pulmonary artery smooth muscle cells. Eur. J. Cell Biol. 85,679–691.

Naka K. I. and Rushton W. A. (1966) S-potentials from luminosity unitsin the retina of fish (Cyprinidae). J. Physiol. 185, 587–599.

Ogilvie J. M., Speck J. D., Lett J. M. and Fleming T. T. (1999) A reliablemethod for organ culture of neonatal mouse retina with long-termsurvival. J. Neurosci. Methods 87, 57–65.

Perrone M. G. and Scilimati A. (2010) b(3)-Adrenoceptor agonists and(antagonists as) inverse agonists history, perspective, constitutiveactivity, and stereospecific binding.Methods Enzymol. 484, 197–230.

Phipps J. A., Wilkinson-Berka J. L. and Fletcher E. L. (2007) Retinaldysfunction in diabetic ren-2 rats is ameliorated by treatment withvalsartan but not atenolol. Invest. Ophthalmol. Vis. Sci. 48, 927–934.

Ristori C., Filippi L., Dal Monte M., Martini D., Cammalleri M., For-tunato P., la Marca G., Fiorini P. and Bagnoli P. (2011) Role of theadrenergic system in a mouse model of oxygen-induced retinopa-thy: antiangiogenic effects of beta-adrenoreceptor blockade. Invest.Ophthalmol. Vis. Sci. 52, 155–170.

Russell S. T. and Tisdale M. J. (2005) The role of glucocorticoids in theinduction of zinc-alpha2-glycoprotein expression in adipose tissuein cancer cachexia. Br. J. Cancer 92, 876–881.

Sartoretto J. L., Jin B. Y., Bauer M., Gertler F. B., Liao R. and Michel T.(2009) Regulation of VASP phosphorylation in cardiac myocytes:differential regulation by cyclic nucleotides and modulation ofprotein expression in diabetic and hypertrophic heart. Am. J.Physiol. Heart Circ. Physiol. 297, H1697–H1710.

Shen J., Xie B., Dong A., Swaim M., Hackett S. F. and Campochiaro P.A. (2007) In vivo immunostaining demonstrates macrophagesassociate with growing and regressing vessels. Invest. Ophthalmol.Vis. Sci. 48, 4335–4341.

Smith L. E., Wesolowski E., McLellan A., Kostyk S. K., D’Amato R.,Sullivan R. and D’Amore P. A. (1994) Oxygen-induced retinop-athy in the mouse. Invest. Ophthalmol. Vis. Sci. 35, 101–111.

Smith C. P., Sharma S. and Steinle J. J. (2007) Age-related changes insympathetic neurotransmission in rat retina and choroid. Exp. EyeRes. 84, 75–81.

Steinle J. J., Booz G. W., Meininger C. J., Day J. N. and Granger H. J.(2003) Beta 3-adrenergic receptors regulate retinal endothelial cellmigration and proliferation. J. Biol. Chem. 278, 20681–20686.

Steinle J. J., Zamora D. O., Rosenbaum J. T. and Granger H. J. (2005)Beta 3-adrenergic receptors mediate choroidal endothelial cellinvasion, proliferation, and cell elongation.Exp. Eye Res. 80, 83–91.

Steinle J. J., Cappocia F. C., Jr and Jiang Y. (2008) Beta-adrenergicreceptor regulation of growth factor protein levels in human cho-roidal endothelial cells. Growth Factors 26, 325–330.

Steinle J. J., Kern T. S., Thomas S. A., McFadyen-Ketchum L. S. andSmith C. P. (2009) Increased basement membrane thickness,

pericyte ghosts, and loss of retinal thickness and cells in dopaminebeta hydroxylase knockout mice. Exp. Eye Res. 88, 1014–1019.

Tan L. B., Burniston J. G., Clark W. A., Ng Y. and Goldspink D. F.(2003) Characterization of adrenoceptor involvement in skeletaland cardiac myotoxicity Induced by sympathomimetic agents: to-ward a new bioassay for beta-blockers. J. Cardiovasc. Pharmacol.41, 518–525.

Vessey K. A., Wilkinson-Berka J. L. and Fletcher E. L. (2011) Char-acterization of retinal function and glial cell response in a mousemodel of oxygen-induced retinopathy. J. Comp. Neurol. 519, 506–527.

Vrydag W. and Michel M. C. (2007) Tools to study beta3-adrenoceptors.Naunyn Schmiedebergs Arch. Pharmacol. 374, 385–398.

Walker R. J. and Steinle J. J. (2007) Role of beta-adrenergic receptors ininflammatory marker expression in Muller cells. Invest. Ophthal-mol. Vis. Sci. 48, 5276–5281.

Wang S. M. and Yang W. L. (2009) Circulating hormone adrenomed-ullin and its binding protein protect neural cells from hypoxia-induced apoptosis. Biochim. Biophys. Acta 1790, 361–367.

Wiley L. A., Rupp G. R. and Steinle J. J. (2005) Sympathetic innervationregulates basement membrane thickening and pericyte number inrat retina. Invest. Ophthalmol. Vis. Sci. 46, 744–748.

Wilkinson-Berka J. L., Tan G., Jaworski K. and Ninkovic S. (2007)Valsartan but not atenolol improves vascular pathology in diabeticRen-2 rat retina. Am. J. Hypertens. 20, 423–430.

Yafai Y., Iandiev I., Wiedemann P., Reichenbach A. and Eichler W.(2004) Retinal endothelial angiogenic activity: effects of hypoxiaand glial (Muller) cells. Microcirculation 11, 577–586.

Yalcin I., Tessier L. H., Petit-Demouliere N., Waltisperger E., Hein L.,Freund-Mercier M. J. and Barrot M. (2010) Chronic treatment withagonists of beta(2)-adrenergic receptors in neuropathic pain. Exp.Neurol. 221, 115–121.

Yang A. D., Bauer T. W., Camp E. R., Somcio R., Liu W., Fan F. andEllis L. M. (2005) Improving delivery of antineoplastic agents withanti-vascular endothelial growth factor therapy. Cancer 103, 1561–1570.

Yanni S. E., McCollum G. W. and Penn J. S. (2010) Genetic deletion ofCOX-2 diminishes VEGF production in mouse retinal Muller cells.Exp. Eye Res. 91, 34–41.

Yu N. N., Wang X. X., Yu J. T., Wang N. D., Lu R. C., Miao D., TianY. and Tan L. (2010) Blocking beta-adrenergic receptor attenuatesacute stress-induced amyloid beta peptides production. Brain Res.1317, 305–310.

Zhang Y. L., Tavakoli H. and Chachisvilis M. (2010) Apparent PKAactivity responds to intermittent hypoxia in bone cells: a redoxpathway? Am. J. Physiol. Heart Circ. Physiol. 299, H225–H235.

Zheng Z., Chen H., Xu X., Li C. and Gu Q. (2007) Effects of angio-tensin-converting enzyme inhibitors and beta-adrenergic blockerson retinal vascular endothelial growth factor expression in ratdiabetic retinopathy. Exp. Eye Res. 84, 745–752.

Zou H., Otani A., Oishi A., Yodoi Y., Kameda T., Kojima H. andYoshimura N. (2010) Bone marrow-derived cells are differentiallyinvolved in pathological and physiological retinal angiogenesis inmice. Biochem. Biophys. Res. Commun. 391, 1268–1273.



Antiangiogenic effects of β2-adrenergic receptor blockade in a mouse model of oxygen-induced retinopathy

Documents