Lack of R-Ras Leads to Increased Vascular Permeability in Ischemic Retinopathy

Downloaded F om: h p://io s.a ojou nals.o g/pd access.ashx?u l=/da a/Jou nals/IOVS/935707/ on 11/30/2016 D Lack o R-Ras Leads o Inc eased Vascula Pe meabili y in Ischemic Re inopa hy Ma ia V¨ah¨a upa, 1,2 S ua P ince, 2 Su i Va aja, 1 Teija Me imo, 1 Ma ko Ka aja, 3 Ka i Kinnunen, 4 Va pu Ma jom¨aki, 5 Hannu Uusi alo, 1,3 Masanobu Koma su, 6 Te o A.H. J¨a inen, 2,7 and Hannele Uusi alo–J¨a inen 1,3 1 Depa men o Oph halmology, Uni e si y o Tampe e, Tampe e, Finland 2 Depa men o Ana omy, Uni e si y o Tampe e, Tampe e, Finland 3 Eye Cen e, Tampe e Uni e si y Hospi al, Tampe e, Finland 4 Depa men o Oph halmology, Kuopio Uni e si y Hospi al, Kuopio, Finland 5 Depa men o Biological and En i onmen al Science/Nanoscience Cen e , Uni e si y o Jy ¨askyl¨a, Jy ¨askyl¨a, Finland 6 San o d Bu nham P ebys Medical Disco e y Ins i u e a Lake Nona, O lando, Flo ida, Uni ed S a es 7 Depa men o Musculoskele al Diso de s, Tampe e Uni e si y Hospi al, Tampe e, Finland Co espondence: Hannele Uusi alo- J¨a inen, Depa men o Oph hal- mology, School o Medicine, 33014 Uni e si y o Tampe e, Finland; llhauus@u a. i Submi ed: Janua y 25, 2016 Accep ed: July 14, 2016 Ci a ion: Va¨h¨a upa M, P ince S, Va aja S, e al. Lack o R-Ras leads o inc eased ascula pe meabili y in ischemic e inopa hy. In es Oph hal- molVisSci . 2016;57:4898–4909. DOI:10.1167/io s.16-19212 P URPOSE . The ole o R-Ras in e inal angiogenesis and ascula pe meabili y was e alua ed in an oxygen-induced e inopa hy (OIR) model using R-Ras knockou (KO) mice and in human diabe ic neo ascula memb anes. M ETHODS . Mice de icien o R-Ras and hei wild- ype (WT) li e ma es we e subjec ed o 75% oxygen om pos na al day 7 (P7) o P12 and hen e u ned o oom ai . A P17 e inal ascula iza ion was examined om whole moun s, and e inal ascula pe meabili y was s udied using Miles assay. Real- ime RT-PCR, Wes e n blo ing, and immunohis ochemis y we e used o assess he exp ession o R-Ras in e ina du ing de elopmen o in he OIR model. The deg ee o pe icy e co e age and ascula endo helial (VE)-cadhe in exp ession on WT and R-Ras KO e inal blood essels was quan i ied using con ocal mic oscopy. The co ela ion o R-Ras wi h ascula endo helial g ow h ac o ecep o 2 (VEGFR2) and human se um albumin on human p oli e a i e diabe ic e inopa hy memb anes was assessed using immunohis ochemis y. R ESULTS . In e ina, R-Ras exp ession was mos ly es ic ed o he ascula u e. Re inal essels in he R-Ras KO mice we e signi ican ly mo e pe meable han WT con ols in he OIR model. A signi ican educ ion in he di ec physical con ac be ween pe icy es and blood essel endo helium as well as educed VE-cadhe in immunos aining was ound in R-Ras–de icien mice. In human p oli e a i e diabe ic e inopa hy neo ascula memb anes, R-Ras exp ession nega i ely co ela ed wi h inc eased ascula leakage and exp ession o VEGFR2, a ma ke o blood essel imma u i y. C ONCLUSIONS . Ou esul s sugges ha R-Ras has a ole in con olling e inal essel ma u a ion and s abiliza ion in ischemic e inopa hy and p o ides a po en ial a ge o pha macologic manipula ion o ea diabe ic e inopa hy. Keywo ds: e ina, diabe ic e inopa hy, neo ascula iza ion, e inal ischemia iabe ic macula edema (DME) and p oli e a i e diabe ic e inopa hy (PDR) a e he wo o ms o diabe ic e inop- a hy causing blindness. Vascula endo helial g ow h ac o (VEGF) plays a c ucial ole in hese diseases by causing ascula leakage and pa hologic neo ascula iza ion. 1–3 An i-VEGF d ugs a e ega ded as he mains ay o ea men o pa ien s a ec ed by DME. Howe e , VEGF inhibi ion ails o p o ide su icien e icacy in all DME cases; only hal o pa ien s a ain signi ican imp o emen in isual acui y. 4 An i-VEGF he apy is also accompanied by a e, bu se e e, ocula and sys emic side e ec s. In PDR, an i-VEGF may be associa ed wi h de elopmen o ac ional e inal de achmen 5,6 Fu he mo e, in a i eally adminis e ed an i-VEGF agen s may cause sus ained sys emic VEGF inhibi ion 7,8 This is why hey a e ela i ely con aindi- ca ed in p egnan women 9 and pa ien s wi h ca dio ascula insu iciency. 10,11 Thus, al e na i e he apies a e needed o mo e e ec i e and sa e ea men o diabe ic e inopa hy. 12,13 R-Ras is a small GTPase o he Ras amily o oncogenes. 14 Despi e he close s uc u al simila i y o o he membe s o he Ras amily, he unc ion o R-Ras is dis inc om o he Ras p o eins. 15 Whe eas all o he membe s o he Ras amily may cause malignan ans o ma ion, R-Ras has e y li le o no ans o ming ac i i y. 16 Fu he mo e, he opposing unc ions o R-Ras and H-Ras ex end o cell–ex acellula ma ix (ECM) adhesion, 17 cell di e en ia ion, 18 and signaling. 19 These oppos- ing unc ions ha e led o he sugges ion ha he balance be ween R-Ras and o he Ras membe s ac s as a swi ch ha con ols p oli e a ion and in asion e sus quiescence in cells. 15,20,21 Recen ly, ac i e Ras signaling has been implica ed as a key d i e o pa hologic angiogenesis in neo ascula diseases o he io s.a ojou nals.o g j ISSN: 1552-5783 4898 This wo k is licensed unde a C ea i e Commons A ibu ion-NonComme cial-NoDe i a i es 4.0 In e na ional License. Downloaded F om: h p://io s.a ojou nals.o g/pd access.ashx?u l=/da a/Jou nals/IOVS/935707/ on 11/30/2016 Lack o R-Ras and Vascula Pe meabili y in Re inopa hy IOVS jSep embe 2016 jVol. 57 jNo. 11 j4899 e ina. 22 Con e sely, exac ly he opposi e ole has ecen ly been shown o R-Ras in he egula ion o angiogenesis. 20 R-Ras unc ion has been shown o be essen ial o he es ablishmen o ma u e and unc ional blood essels in umo s by s abiliza ion o imma u e, leaky essels, enhancing pe usion and ul ima ely educing plasma leakage. 15,21 Based on hese unc ions, R-Ras has been classi ied as an an iangiogenic molecule. Howe e , R-Ras is unc ionally di e en om classic an iangiogenic agen s as i does no induce endo helial cell apop osis as do o he an iangiogenic molecules, bu ac ually p omo es endo helial cell su i al. 20 Thus, R-Ras could p o ide a po en ial agonis ic he apeu ic a ge o ea men o a a ie y o e inal diseases associa ed wi h ischemia, neo ascu- la iza ion, and ascula leakage. In he p esen s udy, he ole o R-Ras in he e ina was s udied by analyzing exp ession o R-Ras in he e ina and by analyzing he e ec o R-Ras de iciency on he s uc u al and unc ional in eg i y o e inal ascula u e du ing de elopmen and ischemic e inopa hy using an oxygen-induced e inopa hy model (OIR). To add ess he ole o R-Ras in PDR, associa ion o i s exp ession wi h ascula imma u i y and leakiness was examined in human e inal neo ascula memb anes. METHODS Mice Fo he expe imen s, wild- ype (WT) C57BL/6 mice om Ha lan Sp ague Dawley (Indianapolis, IN, USA) and homozy- gous R-Ras knockou (KO) mice we e used. R-Ras KO mice wi h an inse ion dele ing R-Ras exp ession be ween exons 4 and 5 o R as on ch omosome 7 (R-Ras KO mice) ha e been desc ibed p e iously. 20 Be o e any expe imen s, R-Ras he e o- zygous mice we e backc ossed eigh imes wi h he C57BL/6 s ain o ob ain homozygous KO and WT mice in he same gene ic backg ound. The mice we e b ed, and he geno ype was de e mined by PCR. Mice we e ed wi h s anda d labo a o y pelle s and wa e ad libi um. All animal expe imen s we e pe o med acco ding o he ARVO S a emen o he Use o Animals in Oph halmic and Vision Resea ch in acco dance wi h p o ocols app o ed by he Na ional Animal E hics Commi ee o Finland. Oxygen-Induced Re inopa hy Model The expe imen s on he OIR model we e ca ied ou as desc ibed in de ail p e iously. 23,24 B ie ly, neona al mice a pos na al day 7 (P7) we e exposed o 75% oxygen o 5 days. A P12, hey we e e u ned o no mal oom ai . 23 Animals we e eu hanized a P12 o assess he deg ee o ascula eg ession and a P17 o de e mine he a e o e inal e ascula iza ion and p e e inal neo ascula iza ion. As pos na al weigh gain has been shown o a ec ou come in he OIR model, 25 only he pups weighing be ween 6.3 and 7.5 g a P17 we e included in he s udy. Immunohis ochemis y (IHC) and Isolec in GS-IB 4 S aining Fo he analysis o e inal ascula u e, eyes we e enuclea ed, ixed wi h 4% pa a o maldehyde (PFA), and e inas dissec ed. Fla -moun e inas we e blocked in 20% no mal goa and 20% e al bo ine se ums o 2 hou s, incuba ed o e nigh Isolec in (Isolec in GS-IB 4 , 1:200; In i ogen, Ca lsbad, CA, USA) and wi h an i-NG2 chond oi in sul a e p o eoglycan an ibody (1:80; Millipo e, Bille ica, MA), ollowed by Alexa Fluo -conjuga ed seconda y an ibody. Re inas we e imaged ia con ocal mic o- scope (LSM 700; Ca l Zeiss, Obe kochen, Ge many) and he a e o angiogenesis was de e mined du ing de elopmen (P0– P6) and in he OIR model as p e iously desc ibed. 24 B ie ly, e inas we e imaged using con ocal mic oscopy (Ca l Zeiss LSM 700) wi h 5 3 objec i e. By ocusing jus abo e he inne limi ing memb ane o he e ina, he p e e inal neo ascula u s we e eadily dis inguished om he unde lying supe icial ascula plexus. A eas o ascula obli e a ion and pa hologic neo ascula iza ion (meaning neo ascula u s) we e quan i- ied om hese images using image edi ing so wa e (Adobe Pho oshop CS3; Adobe Sys ems, Inc., San Jose, CA, USA). The a e o de elopmen al angiogenesis a P0 o P6 was de e mined by measu ing he leng h o ascula u e om he op ic ne e o he ips o he blood essels. Fou measu emen s pe e ina we e aken and an a e age was calcula ed. Fo immunohis ochemis y, he eyes we e ixed wi h 4% PFA and embedded in pa a in o immedia ely ozen in o ni hine ca bamoyl ans e ase embedding compound in iso- pen ane cooled wi h liquid ni ogen and la e ixed wi h ace one. The IHC was ca ied ou on 4- o 6-lm- hick issue sec ions using he ollowing p ima y an ibodies: abbi an i–R- Ras (1:50) and abbi an i-VEGFR2 (1:100; bo h Cell Signaling Technology, Dan e s, MA, USA), abbi an i–R-Ras (1:750) and mouse an i-VEGFR2 (1:200, clone sc.6251; bo h San a C uz Bio echnology, Dallas, TX, USA), a an i-mouse CD31 (1:50; BD Pha mingen, San Diego, CA, USA), guinea pig an i-NG2 an ibody (1:100, gi om William S allcup), 26 abbi an i- Syndecan (1:500) and abbi an i-human ascula endo helial (VE)-cadhe in (1:500; bo h om Abcam, Camb idge, UK), and abbi an i-human se um albumin (1:100; Li espan Biosciences, Sea le, WA, USA), ollowed by ho se adish pe oxidase (HRP) o luo escein-conjuga ed seconda y an ibodies. Hema oxylin s aining was used as a coun e s ain. Samples we e moun ed wi h Vec ashield moun ing medium wi h 4 0 ,6-diamidino-2- phenylindole ( Vec o Labo a o ies) and analyzed ia con ocal mic oscope. Each s aining expe imen included sec ions s ained wi hou p ima y an ibody as nega i e con ols. Pe cen age o R-Ras posi i e cells in p e e inal neo essels we e quan i ied manually by using he coun ool in Adobe Pho oshop (Adobe Sys ems, Inc.) ( n ¼ 6 mice, 4–6 sec ions/ e ina). Quan i ica ion o Pe icy e Co e age and VE- Cadhe in Colocaliza ion Wi h CD31 Pe icy e co e age o he blood essels was quan i ied om WT and R-Ras KO P17 OIR model whole-moun e inas. Samples we e imaged using con ocal mic oscopy (Ca l Zeiss LSM 700) wi h a 63 3 objec i e lens, and 3D pic u es we e made om Z- s acks. Pic u es selec ed o he analysis we e andomly aken om Isolec in IB 4 and NG2-s ained la -moun s om he mos supe icial ascula plexus a he ips o he blood essels nex o a ascula a ea. This is a egion o e ina whe e he blood essels a e g owing owa d he op ic ne e. Di ec con ac be ween pe icy es (NG2) and endo helial cells (Isolec in B 4 ) we e quan i ied by colocaliza ion analysis wi h BioImageXD (p o ided in he public domain a h p://www.bioimagexd. o g). 27 The pe cen age o EC a ea colocalized wi h pe icy e a ea was quan i ied in each image, and he esul exp essed as a mean o wo di e en images o each e ina. The in ensi y o VE-cadhe in s aining and colocaliza ion o VE-cadhe in and CD31 was quan i ied om ozen sec ions using BioImage XD. Real-Time Quan i a i e PCR (qPCR) Eyes we e collec ed a P0, P4, P7, P12, P17, and P22 and placed immedia ely in cold RNA s abiliza ion eagen (RNA- la e ; Sigma-Ald ich Co p., S . Louis, MO, USA) o he Downloaded F om: h p://io s.a ojou nals.o g/pd access.ashx?u l=/da a/Jou nals/IOVS/935707/ on 11/30/2016 Lack o R-Ras and Vascula Pe meabili y in Re inopa hy IOVS jSep embe 2016 jVol. 57 jNo. 11 j4900 du a ion o dissec ion o he e inas. RNA was ex ac ed using RNA isola ion ki (RNeasy Mini Ki ; Qiagen, Hilden, Ge many) ollowing he manu ac u e ’s ins uc ions. The in eg i y o ex ac ed RNA was e i ied on o maldehyde aga ose gel. Re e se ansc ip ion o RNA was ca ied ou using cDNA syn hesis ki (Maxima Fi s S and cDNA Syn hesis Ki ; The mo Fishe Scien i ic, Bos on, MA, USA), acco ding o he manu ac u e ’s ins uc ions. Real- ime qPCR was done using he PCR dye (SYBR G een; The mo Fishe Scien i ic) me hod acco ding o he manu ac u e ’s ecommenda ions and pe o med in whi e 96-well pla es (Mul iply PCR pla es; Sa s ed , Nu¨mb ech , Ge many). The luo escence signal was de ec ed using a sequence de ec ion sys em (ABI P ism 7000 Sequence De ec ion Sys em 1.2; Applied Biosys ems, Ca ls- bad, CA, USA). The ollowing in on-spanning p ime s speci ic o exon 3 and 4 o R as we e used: 5 0 -ACAGGCA GAGTTTCAATGAG-30( o wa d), 50-GTTCTCCAGATCTGCC TTG-3 0 ( e e se). Ppia was used as an endogenous con ol gene (50-CACCGTGTTCTTCGACATC-30and 50-ATTCTGTGA AAGGAGGAACC-3 0 ), allowing he compa ison o samples. 20 Two o h ee eplica es we e analyzed o each sample, and he esul s we e exp essed as a mean o each sample. Nega i e, no empla e (NTC) and no e e se ansc ip ase (no RT), con ols we e included in e e y qPCR analysis. A mel ing cu e analysis was pe o med o check o unspeci ic PCR p oduc s (which didn’ occu ). Reac ion e iciencies o bo h p ime pai s we e es ed wi h s anda d cu e analysis in di e en ime poin s. Due o simila e iciencies o bo h p ime pai s, he di e ences in gene exp ession le els we e de e mined by he compa a i e C me hod (DDC me hod). 28 Vascula Pe meabili y Assay Vascula pe meabili y was quan i ied by Miles assay in e inas as desc ibed p e iously. 29 B ie ly, he mice we e injec ed wi h 2% E ans Blue ([EB], 150 lL/20g; Sigma-Ald ich Co p.) in ape i oneally, and he ci cula ion ime was 24 h o P16 o 1 hou o P17 injec ed mice. Phospha e-bu e ed saline– injec ed mice we e used as a nega i e con ol. Animals we e eu hanized a P17, and blood was collec ed om he le en icle. Re inas we e dissec ed and weighed. Blood was cen i uged a 3,550 g o 15 minu es, and plasma was dilu ed 1:100 and 1:1000 in N, N-dime hyl o mamide (Sigma-Ald ich Co p.). All samples we e incuba ed in a shake (100 pm) in 200 lL N, N-dime hyl o mamide o e nigh a 78 8 C. Re inas we e cen i uged a 17,000 g o 45 minu es a 4 8 C, and he supe na an s we e collec ed. The EB abso bance o each sample was measu ed a 620 nm by spec opho ome e . EB concen a ion in he e ina was p opo ioned and no malized by he weigh o he e ina and he amoun o EB concen a ion in he plasma. 30 Wes e n Blo ing Re inas we e lysed in 10 lL o cold RIPA bu e pe 1 mg o issue wi h added p o ease inhibi o (c O mple e; Roche, Basel, Swi ze land) and phospha ase inhibi o (Hal ; The mo Fishe Scien i ic). The issues we e homogenized using CK14 beads and a issue homogenize (P ecellys; Be in Technologies, Mon igny le- B e onneux, F ance). F om each sample, 40 lg o p o ein was loaded pe well o a 4-12% g adien gel (NuPAGE; In i ogen, Ca lsbad, CA, USA) and elec oblo ed on poly inylidene luo ide memb anes (Immun-Blo; Bio-Rad Labo a o ies, He cu- les, Cali o nia). Fo de ec ion o speci ic p o eins by immuno- blo ing, he ollowing p ima y an ibodies we e used: abbi an i–R-Ras (1:500; Cell Signaling Technology), goa an i-GAPDH (1:500: AbCam) and ho se an i-mouse IgG (1:2000; Cell Signaling Technology). P ima y an ibodies we e de ec ed by HRP-coupled an ibodies. Wes e n blo images we e cap u ed ia so wa e (ImageQuan ; GE Heal hca e, Chal on S . Giles, UK) and quan i ied by densi ome y using Adobe Pho oshop CS3 so wa e, whe e GAPDH was used o no malize o p o ein loading. A calib a o con ol sample was included in e e y memb ane o enable sample compa ison be ween di e en memb anes. Re inal and Neo ascula Samples F om Human Pa ien s P e e inal neo ascula memb anes we e ob ained om eigh ype I diabe ic pa ien s who we e unde going pa s plana i ec omy o he ea men o PDR. All pa ien s we e Caucasians, and al oge he he e we e i e emales and h ee males. A he ime o pa s plana, i ec omy pa ien s’ mean age was 33 yea s ( ange, 27–56 yea s) and mean du a ion o diabe es was 24 yea s ( ange, 16–32 yea s). No mal human e inas we e ob ained om pa ien s whose eyes we e enuclea ed due o cho oidal melanoma. The p o ocol o collec ing human issue samples was app o ed by he ins i u ional e iew boa ds o he Pi kanmaa Hospi al Dis ic and he Kuopio Uni e si y Hospi al. The s udy was conduc ed in acco dance wi h he Decla a ion o Helsinki. All pa ien s ga e w i en in o med consen . Du ing i ec o- my, he ib o ascula memb anes we e isola ed, g asped wi h i eous o ceps, and pulled ou h ough a scle o omy. The sample was immedia ely ixed wi h 4% o maldehyde o 3 h, ans e ed o 70% e hanol, embedded in pa a in, and p ocessed o immunohis ochemis y. The numbe o R-Ras– posi i e neo essels was calcula ed om CD31 and R-Ras– double-s ained issue sec ions, and co ela ion be ween R- Ras and VEGFR2 was quan i ied om R-Ras and VEGFR2 double-s ained sec ions. Immunohis ochemis y agains hu- man se um albumin (HSA) was done o quan i y he amoun o ascula leakage ou side he neo essels. Quan i ica ion was done using IHC P o ile plugin in ImageJ so wa e (h p://imagej.nih.go /ij/; p o ided in he public domain by Na ional Ins i u es o Heal h, Be hesda, MD, USA). The a ea o posi i e s aining o HSA was compa ed o he o al analyzed a ea, and co ela ion analysis be ween HSA and pe cen age o R-Ras–posi i e neo essels was pe o med o each sample. S a is ical Analysis S uden ’s - es was conduc ed o no mally dis ibu ed da a and nonpa ame ic Mann-Whi ney U es using s a is ical so wa e (G aphPad P ism 6.01; G aphPad So wa e, San Diego, CA, USA, and IBM SPSS S a is ics; IBM, Amonk, NY, USA) o nonno mally dis ibu ed da a o es he s a is ical signi icance o he esul s. Spea man’s ank co ela ion es was conduc ed o es co ela ion be ween wo a iables. P alues less han 0.05 we e conside ed s a is ically signi ican . RESULTS R-Ras Exp ession Peaks Du ing Re inal De elopmen al Angiogenesis Bu Does No In luence he Ra e o De elopmen al Angiogenesis R-R as KO mice a e e ile and show no ob ious mo phologic abno mali ies, 20,31 and hei eyes appea no mal upon his o- logic examina ion. To add ess he ole o R-Ras on e inal angiogenesis in i o, he exp ession p o ile o R-Ras in he Downloaded F om: h p://io s.a ojou nals.o g/pd access.ashx?u l=/da a/Jou nals/IOVS/935707/ on 11/30/2016 Lack o R-Ras and Vascula Pe meabili y in Re inopa hy IOVS jSep embe 2016 jVol. 57 jNo. 11 j4901 F IGURE 1. Hypoxia-d i en R-Ras exp ession in he OIR model. Hypoxia-induced angiogenesis in e ina was s udied wi h he OIR model. Re inas we e ha es ed immedia ely a e he exposu e o hype oxia (P12) and a e hypoxia-d i en pa hologic angiogenesis has eached i s maximum a P17. Re inas om no mal mice we e ha es ed a co esponding ime poin s. Re inas we e subjec ed o ei he quan i a i e mRNA ( eal- ime qPCR) analysis using SYBR G een me hod o p o ein (Wes e n blo ing) analysis. Fo immunoblo ing, e inal p o ein supe na an s we e elec opho esed on g adien gels, s anda d Wes e n blo ing was ca ied ou wi h R-Ras–speci ic an ibody, and GAPDH de ec ion was used as a loading con ol. ( A ) The g aph ep esen s old changes (2 - DDC me hod) in he R as mRNA exp ession le el ela i e o he R as mRNA exp ession le el o no mal mice a P12. The le el o R as mRNA exp ession be ween no mal and OIR model e inas is equal a P12, whe eas a P17 hypoxia leads o 2.5- old up egula ion o R as mRNA ( P ¼ 0.016, * nonpa ame ic Mann-Whi ney U es ). E o ba s ep esen he minimum and maximum o he old change. (P12, P17 OIR: n ¼ 4; P17, P12 OIR: n ¼ 5.) ( B , C ) The le el o R-Ras p o ein exp ession was quan i ied by densi ome ic analysis o immunoblo ed p o ein. R-Ras p o ein le el a P17 shows a 7.5- old inc ease ( P ¼ 0.016) du ing hypoxia-induced angiogenesis. E o ba s ep esen 6 95% con idence in e als. (P12: n ¼ 6; P12 OIR: n ¼ 7; P17: n ¼ 5; P17: n ¼ 5.) The esul s a e analyzed wi h nonpa ame ic Mann-Whi ney U es . The samples p esen ed abo e we e un on he same gel. Rep esen a i e samples we e c opped and p esen ed side-by-side. de eloping mouse e ina as well as he de elopmen al a e o angiogenesis in WT and R-Ras KO mice we e cha ac e ized. To assess he empo al exp ession pa e n o R-Ras du ing de elopmen al e inal angiogenesis, qPCR and Wes e n blo analysis we e used. Quan i a i e PCR demons a ed ha he R as mRNA exp ession inc eased om P0 onwa d, peaked a P17, and s a ed o decline om P17 un il P22. Wes e n blo analysis demons a ed a R-Ras p o ein exp ession p o ile simila o he mRNA p o ile: an inc ease in exp ession om P0 un il P17 and a decline om P17 o P22 (Supplemen a y Fig. S1). The inc ease in R as mRNA and p o ein le els om P0 o P17 co ela es wi h he de elopmen o e inal ascula plexuses. 32 The R-Ras le els peaked a P17, when all h ee ascula plexuses a e o med and ascula densi y peaks in he e ina 33 (Supplemen a y Fig. S1). The indings a e consis en wi h he no ion ha R-Ras appea s in he blood essels du ing hei di e en ia ion and is mos s ongly exp essed in ully di e en ia ed, quiescen blood essels. 20 The educ ion in R as mRNA and p o ein exp ession le els om P17 o P22 can be explained by he emodeling o e inal ascula u e (including ascula p un- ing), he ou come o which is dec eased endo helial cell densi y a P28. 34 Consis en wi h his inding, he esul s show a simila dec ease in he in ensi y o Isolec in IB 4 s aining om P12 onwa d (Supplemen a y Fig. S2). No R-Ras p o ein o mRNA exp ession was de ec ed in he e inas collec ed om R-Ras KO mice. To explo e whe he he inc ease in R-Ras exp ession le el seen du ing e inal de elopmen in luences he a e o supe icial ascula plexus o ma ion (de elopmen al angio- genesis), e inas om P0, P2, P4, and P6 WT and R-Ras KO mice we e analyzed. In P0, P2, P4, and P6 neona al mice, he diame e o e inal supe icial ascula plexuses we e simila bo h in R-Ras KO and WT mice, indica ing ha R-Ras exp ession is no necessa y o in i o angiogenesis du ing neona al de elopmen (Supplemen a y Fig. S3). The lack o Downloaded F om: h p://io s.a ojou nals.o g/pd access.ashx?u l=/da a/Jou nals/IOVS/935707/ on 11/30/2016 Lack o R-Ras and Vascula Pe meabili y in Re inopa hy IOVS jSep embe 2016 jVol. 57 jNo. 11 j4902 F IGURE 2. R-Ras is s ongly, bu selec i ely exp essed in he endo helial cells and pe icy es o blood essels in e ina. Oxygen-induced e inopa hy was induced by exposing WT pups o 75% oxygen a P7 o 5 days and e u ning hem o no mal oom ai a P12. A e 5 days in no moxia (a P17), he R-Ras exp ession was de e mined om he e ascula ized e inas by IHC and immuno luo escence (IF) using R-Ras–speci ic p ima y an ibody. (A) Rep esen a i e images o R-Ras exp ession in he OIR model. Mo e han 30% o he p e e inal blood essels a e nega i e o R-Ras ( a ow ), whe eas blood essels in e ina a e R-Ras posi i e ( a owhead ). Righ , nega i e con ol (no p ima y an ibody). The e is also ain R-Ras exp ession om o he e inal cells, p esumably om neu al cells in he e ina. GCL, ganglion cell laye ; IPL, inne plexi o m laye ; INL, inne nuclea laye , OPL, ou e plexi o m laye ; ONL, ou e nuclea laye ; IS/OS, pho o ecep o inne /ou e segmen s; RPE, e inal pigmen epi helium. ( B–E ) Rep esen a i e con ocal images showing he colocaliza ion o R-Ras ( g een ) and CD31 ( ed ) in he e ina o ozen sec ions a e immuno luo escence s aining. ( C ) R-Ras is exp essed in he endo helial cells in he e ina and ( D ) also in he p e e inal blood essels. ( E ) Some o he p e e inal blood essels a e nega i e o R-Ras. ( F ) R-Ras is exp essed in he pe icy es (NG2, ed ) in he e ina. Some pe icy es o p e e inal blood essels a e nega i e o R-Ras in OIR ( a ow ), bu ( G ) R-Ras is also exp essed in he pe icy es o p e e inal neo essels. Scaleba s : 100 lm ( A , B ), 20 lm ( C – G ). in luence on de elopmen o e inal ascula u e by R-Ras is in line wi h he abo e- epo ed low exp ession o R-Ras a he ea ly s ages o e inal ascula de elopmen (Supplemen a y Fig. S1). Induc ion and Selec i e Disappea ance o R-Ras F om he Pa hologic Neo ascula u e in OIR Nex , he R-Ras exp ession p o ile in he OIR model was analyzed using WT mice. Using qPCR, a 2.5- old inc ease o R as mRNA om no mal P17 o P17 OIR model e inas was de ec ed. Using Wes e n blo analysis, a 7.5- old inc ease in R- Ras p o ein le els was seen in OIR P17 e inas compa ed o heal hy P17 e inas. A P12, he exp ession o R-Ras was educed in OIR. The amoun o R-Ras exp ession is in line wi h eg ession o he ascula u e a P12 and essel eg ow h a P17 in he OIR model (Fig. 1). Using R-Ras immunohis ochemis y, R-Ras was ound o be localized mainly o he blood essels in OIR e inas (Fig. 2, Supplemen a y Fig. S4). A ain R-Ras exp ession was also de ec ed ou side o he e inal blood essels in mouse e ina (Fig. 2A). This immunohis ochemical signal mos p obably ep esen s low exp essions om e inal neu onal cells, which a e known o exp ess R-Ras. 35 Using R- Ras and endo helial cell (CD31), as well as R-Ras and pe icy e (NG2), double-s aining, R-Ras exp ession was shown o be con ined o endo helial cells and pe icy es in e inal essels (Fig. 2). I appea ed ha all o he e inal blood essels exp essed R-Ras, whe eas only 68% o he p e e inal neo essels had R-Ras exp ession a P17 in he OIR model (Fig. 2A, Supplemen a y Fig. S4). This inding was con i med by double immuno luo escen imaging o he e ina showing some p e e inal neo essels de oid o R-Ras exp ession in OIR (Figs. 2E, 2F). By cos aining he e inal issue sec ions o R-Ras and endo helial cells, as well as o R-Ras and pe icy es, he R-Ras exp ession was con i med o be con ined o bo h endo helial c ell s an d pe i cy e s (Fig s. 2B–G ). Downloaded F om: h p://io s.a ojou nals.o g/pd access.ashx?u l=/da a/Jou nals/IOVS/935707/ on 11/30/2016 Lack o R-Ras and Vascula Pe meabili y in Re inopa hy IOVS jSep embe 2016 jVol. 57 jNo. 11 j4903 F IGURE 3. Nei he e ascula iza ion no pa hologic neo ascula iza ion is a ec ed by R-Ras de iciency in he OIR model. Wild- ype and R-Ras KO mice pups we e exposed o hype oxia as p e iously desc ibed, and e inas we e ha es ed a P17. A eas o ascula obli e a ion and pa hologic neo ascula iza ion (meaning neo ascula u s) we e quan i ied om Isolec in IB 4 –s ained whole moun s using Adobe Pho oshop CS3. ( A ) Rep esen a i e e inas o WT and R-Ras KO mice a P17. The e ascula iza ion a e was de e mined by quan i ying he a ascula a eas ( yellow ) in e inal la moun s. The amoun o pa hologic neo ascula iza ion ( u s, ed ) was also de e mined. ( B ) Summa y o quan i a i e analysis o ascula obli e a ion ( yellow ) and neo ascula iza ion ( ed ). The a ascula and pa hologic neo ascula iza ion a eas we e measu ed, and esul s a e ep esen ed as an a ea ela i e o WT a ea. E o ba s ep esen 6 95% con idence in e als. (WT: n ¼ 63, KO: n ¼ 48 e inas.) R-Ras Does No Change he Hypoxic Re ascula iza ion Ra e o he Re ina As all o he an iangiogenic molecules used o ea neo ascula e inal diseases ha e been shown o inhibi sp ou ing angiogenesis induced in he OIR model, 36–38 he ole o R-Ras in hypoxia-d i en neo ascula iza ion in he e ina was ex- plo ed nex . Fi s , we de e mined whe he eg ession o essels unde hype oxic condi ions in he OIR model is compa able be ween R-Ras KO and WT mice. A e exposu e o 75% oxygen be ween P7 and P12, e inas we e e alua ed a P12. Typical o his model, 39 la ge a eas o he cen al ascula ne wo k we e obli e a ed, wi h only a ew majo essels emaining cen ally in bo h WT and R-Ras KO mice (Supplemen a y Fig. S5). Quan i a i e analysis o obli e a ed a eas in e ina con i med ha he e inal ascula u e in R-Ras KO and WT mice a e a ec ed in a simila ashion by hype oxic exposu e (Supplemen a y Fig. S5). Taken oge he , hese da a sugges ha dele ion o R-Ras has no obse able e ec on ascula eg ession and emodeling in he e ina in esponse o hype oxic condi ions. On e u n o no moxia, he a ascula and hypoxic cen al e ina s imula es apid eg ow h o essels. 24 The a e o e inal e ascula iza ion in WT and R-Ras KO e inas was de e mined by quan i ying a ascula e inal a ea 5 days a e he mice we e e u ned o no moxia. The e was no signi ican di e ence in he a ascula e inal a ea be ween he WT and R-Ras KO, indica ing ha he a e o e inal e ascula iza ion is simila be ween he WT and R-Ras KO mice a P17 (Fig. 3). In addi ion o e ascula iza ion o he e ina, he s ong hypoxic s imulus om he cen e o he e ina also d i es abno mal misdi ec ed sp ou ing o blood essels in o he i eous a he in e ace be ween he cen ally obli e a ed and pe iphe ally pe used e ina. 24 P e e inal (i.e., pa hologic) neo ascula iza ion eaches i s maximum in WT mice 5 days a e e u ning o no moxia (a P17). 24 No signi ican di e ences in he numbe o p e e inal neo ascula u s and clus e s we e ound be ween R-Ras KO mice and WT mice (Fig. 3). The esul s indica e ha nei he he hypoxia-induced e ascula iza ion a e o he obli e a ed a eas in e ina no he a e o p e e inal neo ascula iza ion is a ec ed by R-Ras dele ion. R-Ras De icien Re inal Blood Vessels Ha e Inc eased Pe meabili y in OIR I has been ecen ly shown ha loss o R-Ras wo sens blood essel s uc u e and pe usion and he pa hologic plasma leakage in umo angiogenesis. 15,21 In o de o assess he ole o R-Ras in con olling he ascula pe meabili y o he e inal ascula u e, he OIR model was used. The ascula leakage was quan i ied by measu ing EB dye ex a asa ion om e inal essels. The R-Ras KO e inas had app oxima ely 100% inc ease in he e inal EB leakage compa ed wi h he WT e inas in OIR ( P ¼ 0.0223 a 1 hou and P ¼ 0.0006 a 24 hou s; Figs. 4A–C). No di e ence was de ec ed in he amoun o EB leakage when compa ing no mal WT and R-Ras KO e inas (Supplemen a y Fig. S6). To con i m he enhanced e inal leakage in R-Ras KO animals in OIR, we also measu ed he amoun o IgG hea y chain molecule accumula ion in he e ina, as his an ibody subclass is no mally compa men alized wi hin he ci cula- ion. 40 We could no de ec a di e ence in IgG leakage in he no mal e inas be ween WT and R-Ras KO animals. A 4- old inc ease in he amoun o IgG was de ec ed in he R-Ras KO e inas o e he WT e inas in OIR as a sign o inc eased ascula pe meabili y ( P ¼ 0.0007; Fig. 4D). To ule ou he possibili y ha he inc eased le els o IgG in he R-Ras KO OIR e inas we e due o plasma cells, immunohis ologic s aining o CD138/Syndecan-1 was pe o med. The e was no inc ease in he numbe o plasma cells in KO e inas compa ed o WT in he OIR samples (Supplemen a y Fig. S7). Downloaded F om: h p://io s.a ojou nals.o g/pd access.ashx?u l=/da a/Jou nals/IOVS/935707/ on 11/30/2016 Lack o R-Ras and Vascula Pe meabili y in Re inopa hy IOVS jSep embe 2016 jVol. 57 jNo. 11 j4904 F IGURE 4. R-Ras de iciency inc eases essel pe meabili y in mouse OIR model. Wild- ype and R-Ras KO mice pups we e exposed o hype oxia as desc ibed p e iously. The EB dye was injec ed IP, and e inas and blood samples we e ha es ed a e 1 hou and 24 hou s a P17. E ans Blue concen a ions we e measu ed spec opho ome ically a 620 nm, and EB concen a ion o each sample was calcula ed om a s anda d cu e. Resul s a e exp essed as ela i e o EB concen a ion in he plasma. ( A ) Rep esen a i e R-Ras KO and WT OIR e inas a e sys emic injec ion o EB dye. In many o he R-Ras KO e inas, EB dye was isible, whe eas mos o he WT e inas we e colo less. ( B ) S a is ical analysis o a ep esen a i e expe imen shows a signi ican inc ease in ascula leakage in R-Ras KO OIR e inas compa ed o WT OIR e inas al eady a a 1-hou ime poin ( P ¼ 0.0223, *, nonpa ame ic Mann-Whi ney U  es , WT n ¼ 6, R-Ras KO n ¼ 8). ( C ) The e is a highly signi ican di e ence in ascula leakage be ween R-Ras KO and WT mice a he 24-hou ime poin ( P ¼ 0.0006, ***, nonpa ame ic Mann-Whi ney U  es ; WT n ¼ 20, R-Ras KO n ¼ 18). E o ba s ep esen SEM. To assess he IgG p o ein accumula ion, OIR e inas we e ha es ed a P17, p o eins we e ex ac ed and samples we e elec opho esed on g adien gels, and Wes e n blo ing was done using an i-mouse IgG an ibody. ( D ) Rep esen a i e images o immunoblo ing om whe e mouse IgG hea y chain densi ies we e quan i ied and no malized agains GAPDH. Rela i e compa ison o IgG hea y chain p o ein om WT OIR and R- Ras KO OIR samples gi es a 3.5- old di e ence ( P ¼ 0,0007, ***, nonpa ame ic Mann-Whi ney U es ). The e is no di e ence in he IgG p o ein le el be ween heal hy WT and KO mice a P17 ( P ¼ 0.7879). E o ba s ep esen 6 95% con idence in e als. (WT: n ¼ 6, KO: n ¼ 6, WT OIR: n ¼ 14, KO OIR: n ¼ 16.) R-Ras De iciency Se e ely Impai s Pe icy e Co e age o Angiogenic Blood Vessels and Reduces VE-Cadhe in Exp ession in OIR Model Close in e ac ion be ween sp ou ing endo helium and pe i- cy es is c ucial o essel ma u a ion and s abili y. 41,42 Pa hologic, angiogenic ascula u e in umo s is cha ac e ized by insu icien pe icy e associa ion a ound he blood essels. 43 Consis en wi h his no ion, a signi ican educ ion in he di ec physical con ac be ween pe icy es and he endo helium o blood essels was ound in R-Ras KO e inas in he OIR model a P17 (Figs. 5A–B). When he pe icy e co e age a ound he ips o he sp ou ing angiogenic blood essels was quan i ied, 40% educ ion in he pe icy e co e age was eco ded in he R-Ras KO e inas o e he WT e inas ( P ¼ 0.033; Fig. 5C). Vascula endo helial–cadhe in is a membe o cadhe in supe amily exp essed exclusi ely on endo helial cells. 21 I is a c ucial ac o o adhe ens junc ions be ween he endo helial cells. 15 R-Ras is known o s abilize adhe ens junc ions by inhibi ing VEGF-induced VE-cadhe in in e naliza ion. 15 We s udied he exp ession o VE-cadhe in in OIR model, and we obse ed ha bo h he in ensi y o he VE-cadhe in immuno- s aining and colocaliza ion o VE-cadhe in wi h CD31 we e signi ican ly educed in KO mice compa ed o WT mice (Fig. 5D). Colocaliza ion analysis showed signi ican educ ion in VE-cadhe in and CD31 colocaliza ion in R-Ras KO mice compa ed o WT mice (o e lap coe icien acco ding o Mande s:  ¼ 0.51 o WT, and R ¼ 0.43 o KO ( P ¼ 0.04)). Ou inding on VE-cadhe in exp ession is in line wha has been epo ed in umo ascula u e o R-Ras KO and WT mice. 21 Reduced R–Ras Exp ession in he Neo essels o Human Diabe ic Re inopa hy Co ela es Wi h Inc eased Vascula Leakage and he Imma u e S a e o he Neo essels R-Ras exp ession was es ic ed exclusi ely o blood essels in no mal human e ina, and i was exp essed in all blood essels in no mal human e inas (Supplemen a y Fig. S8). To add ess he ele ance o ou indings on he ole o R-Ras in he OIR model o human ischemic e inopa hies, he pa hologic e inal neo ascula memb anes ha de elop in human diabe ic e inopa hy pa ien s we e s udied. These neo ascula mem- b anes we e collec ed om pa ien s su e ing om ype I diabe es, who had al eady de eloped ac ional e inal de ach- men due o ib osis o neo ascula memb anes. Thus, he samples ep esen he end s age o he disease, whe e subs an ial amoun o ib osis is associa ed wi h neo essels, bu hey s ill con ain egions wi h ac i e pa hologic angiogen- esis. In he diabe ic neo ascula memb anes, R-Ras exp ession was exclusi ely es ic ed o blood essels (Fig. 6). Like pa hologic p e e inal neo essels in he OIR model, human diabe ic neo ascula memb anes had a ying exp ession o R- Ras (Fig. 6). Only app oxima ely 80% o he blood essels in he human neo ascula memb anes exp essed any R-Ras p o ein a all (Fig. 6C). To explo e whe he R-Ras exp ession has any ela ion o he blood essel ma u i y, he human issue samples we e double-s ained o R-Ras and VEGFR2, a ma ke o blood essel imma u i y. 44,45 A s ong in e se co ela ion be ween exp ession o he wo p o eins was iden i ied (Spea man’s q ¼ - 0.821, P ¼ 0.023, R 2 ¼ 0.563); he highe he pe cen age o imma u e blood essels (VEGFR2 þ ), he less R-Ras exp ession he e was in he blood essels (Fig. 6). To add ess whe he lack o R-Ras in imma u e blood essels in luences ascula pe meabili y in human diabe ic neo ascu- la memb anes, samples we e s ained o HSA. Samples wi h a low pe cen age o R-Ras–exp essing blood essels showed a s ong accumula ion o HSA ou side he essels, whe eas samples wi h a high pe cen age o R-Ras–exp essing blood essels had HSA s aining es ic ed mainly inside he essels (Figs. 6A–B). The e was a s ong nega i e co ela ion be ween a ea o ex a ascula HSA and he numbe o R-Ras–posi i e essels (Spea man’s q:  ¼ - 0.886, P ¼ 0.019, R 2 ¼ 0.835); he mo e ascula leakage ou side o he blood essels, he less R- Ras exp ession was in he blood essels. Fu he mo e, we could demons a e ha leaked, ex a ascula HSA was de ec ed mainly a ound indi idual blood essels ha lack R-Ras (Fig. 7). This was especially e iden in samples ha had R-Ras in he majo i y o blood essels as ex a ascula HSA mainly accumula ed a ound hose ew blood essels de oid o R-Ras exp ession (Fig. 7). DISCUSSION The p esen s udy demons a es ha he small GTPase R-Ras egula es he pa hologic pe meabili y o blood essels in hypoxia-d i en angiogenesis wi hou al e ing he a e o Downloaded F om: h p://io s.a ojou nals.o g/pd access.ashx?u l=/da a/Jou nals/IOVS/935707/ on 11/30/2016 Lack o R-Ras and Vascula Pe meabili y in Re inopa hy IOVS jSep embe 2016 jVol. 57 jNo. 11 j4905 F IGURE 5. Pe icy e co e age and VE-cadhe in exp ession is educed in R-Ras KO in he OIR model in angiogenic e inal blood essels. Oxygen- induced e inopa hy was induced in WT and R-Ras KO mice as desc ibed p e iously. Re inas we e ha es ed a P17, and whole-moun e inas we e double-s ained wi h Alexa Fluo –conjuga ed Isolec in IB 4 and wi h an an ibody agains he pe icy e ma ke NG2 p o eoglycan. Th ee-dimensional images we e aken om he mos supe icial ascula plexus a he ips o he blood essels om he egion whe e essels g ow owa d he op ic ne e. ( A ) Rep esen a i e images o WT and R-Ras KO blood essel endo helial cells ( ed ) su ounded by pe icy es ( g een ). ( B ) The majo i y o images om WT mice e ina had a lo o pe icy es a ound he blood essels, whe eas he majo i y o pic u es aken om he KO e inal blood essels we e lacking o had e y ew pe icy es. ( C ) Di ec con ac be ween endo helial cells and pe icy es was quan i ied by colocaliza ion analysis. The esul is shown as a pe cen age o endo helial cell a ea colocalized wi h pe icy es in WT and R-Ras KO e inal blood essel ips. The pe icy e co e age is signi ican ly educed in R-Ras KO animals compa ed o WT by 40% ( P ¼ 0.033, *). E o ba s ep esen 6 95% con idence in e als. (WT: n ¼ 15; R-Ras KO: n ¼ 15. Two pic u es we e aken om each e ina, and an a e age was calcula ed). The exp ession o VE-cadhe in was s udied in he OIR model by s aining ozen c oss-sec ions o e ina wi h an ibodies agains VE-cadhe in and endo helial cells (CD31). The in ensi y as well as he colocaliza ion o VE-cadhe in s aining wi h endo helial cells was quan i ied using BioImageXD. ( D ) Rep esen a i e images o VE- cadhe in ( g een ) and CD31 ( ed ) in P17 OIR model in WT and R-Ras KO a e p esen ed. Colocaliza ion analysis showed signi ican educ ion in VE- cadhe in and CD31 colocaliza ion in R-Ras KO mice compa ed o WT mice (o e lap coe icien acco ding o Mande s:  ¼ 0.51 o WT and R ¼ 0.43 o KO [ P ¼ 0.04]), n ¼ 5 mice o WT and n ¼ 5 mice o KO. e ascula iza ion o he hypoxic e ina. Fu he mo e, by analyzing a se o p e e inal ascula memb anes ob ained om diabe ic e inopa hy pa ien s, we we e able o demon- s a e ha neo ascula memb anes wi h educed R-Ras exp ession in imma u e essels display inc eased pa hologic ascula leakage. Gi en ha inc eased ascula leakage is he pa hognomonic ea u e in human diabe ic e inopa hy, ou s udy implica es ha loss o R-Ras may ha e a pa hologic ole in his disease. R-Ras educed he pa hologic ascula pe meabili y in ischemia-induced e inopa hy wi hou inhibi ing e inal angio- genesis, whe he i was de elopmen al o induced by hypoxia in OIR. This is a clea di e ence o all cu en a ailable an iangiogenic he apies ha ha e been shown o unc ion in OIR by inhibi ing sp ou ing angiogenesis. 36 Fu he mo e, he mechanism o esis ance o he cu en ly a ailable an iangio- genic he apies bo h in umo s and e inopa hy is ac ually ela ed o he e adica ion o he neo essels, which, in u n, wo sens he unde lying ischemia and d i es he o ma ion o new, leaky blood essels by al e na i e molecula mecha- nism. 13 Thus, he p oposed molecula mechanism o u u e an iangiogenic he apies is one in which he angiogenic blood essels a e ‘‘no malized’’ o s able ones o alle ia e he hypoxia and s op he de imen al abe an ascula leakage. 46,47 Taken wi h ea lie demons a ions o a causal ela ionship be ween R- Ras exp ession and essel ma u a ion, 15,20,48,49 ou esul s sugges po en ial u ili y o R-Ras in ea ing e inopa hies in ol ing angiogenesis. Modula ing R-Ras exp ession ep e- Downloaded F om: h p://io s.a ojou nals.o g/pd access.ashx?u l=/da a/Jou nals/IOVS/935707/ on 11/30/2016 Lack o R-Ras and Vascula Pe meabili y in Re inopa hy IOVS jSep embe 2016 jVol. 57 jNo. 11 j4906 F IGURE 6. Reduced R-Ras exp ession co ela es wi h leakage o human se um albumin in human diabe ic e inopa hy ascula u e, and i is nega i ely co ela ed wi h VEGFR2 exp ession. P e e inal neo ascula memb anes ob ained om i ec omies om diabe ic e inopa hy pa ien s we e analyzed by immunohis ochemis y o hei p o ein exp ession. Immunohis ochemical s aining om adjacen sec ions was done wi h an i- CD31 þ an i–R-Ras, an i-HSA, and an i–R-Ras þ an i-VEGFR2 an ibodies. The numbe o R-Ras–posi i e essels was calcula ed om R-Ras and CD31 double-s ained sec ions ( n ¼ 7; 5–9 di e en sec ions analyzed om each sample). Co ela ion analysis be ween R-Ras and VEGFR2 was done om R- Ras þ VEGFR2 double-s ained sec ions and compa ed o CD31 s aining om adjacen sec ions ( n ¼ 7; 5–9 di e en sec ions analyzed om each sample). Vascula memb anes wi h s ong R-Ras exp ession show limi ed ex a ascula s aining o HSA ( A ), whe eas samples wi h weak R-Ras exp ession show s ong s aining o HSA ou side he blood essels ( B ). Scaleba s : 100 lm. The e is a s ong in e se co ela ion be ween HSA- posi i e ex a ascula a ea and he pe cen age o R-Ras–posi i e neo essels (Spea man’s q:  ¼- 0.886, P ¼ 0.019, *, R 2 ¼ 0.835, n ¼ 6 pa ien s) ( D ). ( C ) App oxima ely 20% o he neo essels do no exp ess any R-Ras a all (R-Ras–posi i e essels 81 6 10%, mean þ SD, n ¼ 6 pa ien s). Da a a e shown as a box plo wi h median and 95% con idence in e al. ( E , F ) Double-s aining o R-Ras and VEGFR2 shows an in e se co ela ion (Spea man’s q:  ¼- 0.821, P ¼ 0.023, R 2 ¼ 0.563) be ween he numbe o R-Ras–exp essing and VEGFR2-posi i e blood essels. When a majo i y o blood essels has R-Ras exp ession, a majo i y o blood essels a e nega i e o VEGFR2 exp ession, whe eas he VEGFR2 exp ession is opposi e when e y ew blood essels exp ess R-Ras. Scaleba s : 200 lm. sen s a no el he apeu ic app oach ha may be capable o add essing pa hologic ascula pe meabili y wi hou simul a- neously wo sening he unde lying pa hobiology o he disease— hypoxia. The di e ences be ween he mechanisms o R-Ras and o he an iangiogenic he apies a e e idenced by a ecen epo implica ing ac i a ed Ras signaling as a key media o o pa hologic neo ascula iza ion induced by many cy okines and g ow h ac o s, including VEGF, in blinding neo ascula eye diseases. 22 Wes enskow e al. 22 epo ed ha abe an Ras signaling can be inhibi ed by a molecule called p120RasGAP. The inhibi o y unc ion o p120RasGAP on Ras-d i en e inal neo ascula iza ion was shown o ake place h ough he inhibi ion o sp ou ing angiogenesis. 22 In con as , R-Ras unc ions in he e ina p ima ily o s abilize ascula u e and p e en leakage and does no inhibi sp ou ing angiogenesis in OIR. The mechanis ic di e ences be ween R-Ras and p120Ras-