Contacting co-culture of human retinal microvascular endothelial cells alters barrier function of human embryonic stem cell derived retinal pigment epithelial cells

Con en s lis s a ailable a ScienceDi ec Expe imen al Cell Resea ch jou nal homepage: www.else ie .com/loca e/yexc Con ac ing co-cul u e o human e inal mic o ascula endo helial cells al e s ba ie unc ion o human emb yonic s em cell de i ed e inal pigmen epi helial cells H. Sko man a , J. Mu anen a,1 , H. Lähdeko pi a,1 , E. Pajula a , K. Mäkelä b , L. Koi usalo a , A. Kois inen c , H. Uusi alo d,e , K. Kaa ni an a c, , K. Juu i-Uusi alo a,⁎ a Facul y o Medicine and Li e Sciences, BioMediTech, Uni e si y o Tampe e, Tampe e, Finland b Depa men o Obs e ics and Gynecology, Tampe e Uni e si y Hospi al, Finland c Depa men o Oph halmology, Ins i u e o Clinical Medicine, Uni e si y o Eas e n Finland, Kuopio, Finland d SILK, Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Finland e Tays Eye Cen e , Tampe e, Finland Depa men o Oph halmology, Kuopio Uni e si y Hospi al, Kuopio, Finland ARTICLE INFO Keywo ds: HESC-RPE HREC Human e inal mic o ascula endo helial cell Co-cul u e Ou e blood- e inal ba ie model In i o model ABSTRACT He e we e alua ed he effec s o human e inal mic o ascula endo helial cells (hREC) on ma u e human emb yonic s em cell (hESC) de i ed e inal pigmen epi helial (RPE) cells. The hESC-RPE cells (Regea08/017, Regea08/023 o Regea11/013) and hREC (ACBRI 181) we e co-cul u ed on opposi e sides o anspa en memb anes o up o six weeks. The ea e ba ie unc ion, small molecule pe meabili y, localiza ion o RPE and endo helial cell ma ke p o eins, cellula fine s uc u e, and g ow h ac o sec e ion o we e e alua ed. A e co-cul u e, he RPE specific CRALBP and endo helial cell specific on Willeb and ac o we e app op ia ely localized. In addi ion, he gene al mo phology, pigmen a ion, and fine s uc u e o hESC-RPE cells we e unaffec ed. Co-cul u e inc eased he ba ie unc ion o hESC-RPE cells, de ec ed bo h wi h TEER measu emen s and cumula i e pe meabili y o FD4 –al hough he diffe ences a ied among he cell lines. Co- cul u ing significan ly al e ed VEGF and PEDF sec e ion, bu again he diffe ences we e cell line specific. The esul s o his s udy showed ha co-cul u e wi h hREC affec s hESC-RPE unc ionali y. In addi ion, co- cul u e e ealed d as ic cell line specificdiffe ences, mos no ably in g ow h ac o sec e ion. This model has he po en ial o be used as an in i o ou e blood- e inal ba ie model o d ug pe meabili y es ing. 1. In oduc ion Re ina is a pa o he cen al ne ous sys em and hus p o ec ed by a blood b ain ba ie analogue, blood e inal ba ie (BRB), consis ing o inne (iBRB) and ou e pa (oBRB). oBRB consis s o h ee laye s: e inal pigmen epi helial cells (RPE), B uch's memb ane and cho oidal endo helial cells (CEC). RPE cells ace he neu al e ina and a e connec ed o each o he by igh junc ions, o ming a selec i e ba ie . RPE has se e al i ally impo an unc ions such as phagocy osis o shed pho o ecep o ou e segmen s, i amin A me abolism, egula ion o he anspo o nu i i e subs ances, abso p ion o s ay ligh , and con ol o e inal ion balance [1,2]. On he o he side, oBRB is in con ac wi h he blood s eam ia he enes a ed CECs. In e ac ions be ween RPE and CEC a e i ally impo an o anspo ing nu ien s and wa e o he e ina, and emo ing he me abolic was e o he blood s eam [2,3]. Be ween hese wo cell ypes lies a laye ed ex acellula ma ix (ECM) s uc u e - B uch's memb ane [4]. Besides offe ing suppo and sepa a ing RPE om he CEC, B uch's memb ane i is also a playing g ound o g ow h ac o s sec e ed by he RPE, such as ascula endo helial g ow h ac o (VEGF) and pigmen epi helial de i ed g ow h ac o (PEDF). An imbalance in his homeos asis migh lead o dys egula ion in angios asis [5,6], o example o cho oidal neo ascula iza ion and age- ela ed macula degene a ion (AMD) [7]. Due o i s excellen p o ec i e p ope ies oBRB also ac s as a ba ie o many d ug molecules, making medical he apy o many e inal diseases demanding. This is he case in diabe ic e inopa hy (DR), AMD, pos e io u ei is and e ini is pigmen osa [8]. Fo many decades ocula d ug deli e y o pos e io segmen s o he eye has been assessed h p://dx.doi.o g/10.1016/j.yexc .2017.08.004 Recei ed 29 Ma ch 2017; Recei ed in e ised o m 31 July 2017; Accep ed 2 Augus 2017 ⁎ Co esponding au ho . 1 Au ho s con ibu ed equally o his wo k. Expe imen al Cell Resea ch 359 (2017) 101–111 A ailable online 08 Augus 2017 0014-4827/ © 2017 The Au ho (s). Published by Else ie Inc. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/BY-NC-ND/4.0/). MARK wi h in i o animal s udies, bu ecen ly due o hei high cos s and e hical conside a ions, al e na i e me hods, such as ex i o and in i o echniques ha e gained popula i y [9].Ex i o cul u es o isola ed issues ha e been used o e alua e mo phology, in eg i y and unc ion o oBRB [10–12].In i o models can p o ide a aluable p e-clinical es ing pla o m o disease modeling, es ing nu ien /pha maceu ical afficking, as well as d ug deli e y and a ge ing [13,14].In i o models based on p ima y RPE cell cul u es o animal o igin [14] ha e been widely used, al hough hey suffe om species- ela ed applic- abili y p oblems. The unc ionali y o oBRB has been ex ensi ely s udied only in RPE cells alone ( e e ed onwa ds as RPE solo-cul u e), gene ally using he immo alized RPE cell line ARPE-19 [14–17]. These RPE solo-cul u es, howe e , do no de elop cell-cell in e ac ions ypical o na i e human RPE, hus se e al oBRB co-cul u e models ha e been de eloped [7,18,19]. RPE cells ha e been co-cul u ed wi h endo helial cells in “con ac ing”cul u es as mix u es pe mi ing di ec con ac [20], sandwiched wi h ECM p o eins [18,21], o cul u ed on opposi e sides o amnio ic memb ane [22] o T answell inse s [23,24]. In addi ion, “non-con ac ing”co-cul u es ha e been pe o med whe e one cell ype is cul u ed on op o a cell cul u e inse and he o he cell ype on he bo om o he well pla e [23–26]. These a ious oBRB models ha e been gene a ed ei he wi h p ima y RPE cells, which ha e a limi ed a ailabili y, o wi h immo alized RPE cell lines which ha e mo phological and de elopmen al abno mali ies [9,27]. Thus, he e is a need o new oph halmic in i o models wi h close esemblance o he unc ionali y and mo phology o he na i e issue [9,13]. RPE cells de i ed om human emb yonic s em cells (hESC) p o ide an unlimi ed sou ce o cells which closely esemble hei na i e coun e pa [28–34]. They ha e been shown o exp ess RPE specific genes and p o eins, become highly pigmen ed and pola ized, display se e al RPE specific unc ions such as PEDF sec e ion and phagocy osis o pho o ecep o ou e segmen s, and exp ess unc ional memb ane p o eins impo an o main aining ba ie p ope ies [35,36]. When ansplan ed o abbi s, hESC-de i ed RPE cell shee s es o e ERG and ou e nuclea laye (ONL) hickness o a sho pe iod o ime [37]. In his wo k we aimed o de elop a hESC-RPE based con ac ing oBRB co-cul u e model by u ilizing h ee diffe en hESC lines. The effec s o co-cul u e on hESC-RPE ba ie p ope ies we e e alua ed wi h ans-epi helial elec ical esis ance (TEER) measu emen s and pe meabili y es s, he mo phology wi h con ocal and elec on mic o- scopy and cellula unc ionali y wi h g ow h ac o sec e ion assays. The esul s o his wo k could pa e he way owa ds an in i o model ha could be use ul o d ug deli e y esea ch and e inal disease modeling. 2. Ma e ials and me hods 2.1. Cells 2.1.1. Endo helial cells Human Umbilical Vein Endo helial Cells (HUVECs) we e ex ac ed a BioMediTech, Uni e si y o Tampe e om he umbilical co ds acqui ed om scheduled Cesa ean sec ions acco ding o Hamil on e al. [38]. HUVECs we e cul u ed in HUVECmem: Gibco M199 medium (Li e Technologies, Ca lsbad, CA, USA) wi h 1% Gibco Glu aMAX™(Li e Technologies), 2% Penicillin-S ep omycin (Lonza G oup L d., Basel, Swi ze land), 10% e al bo ine se um (FBS Gold, PAA, GE Heal hca e L d., Li le Chal on , UK), 0.3% Endo helial Cell G ow h Supplemen (Fi s Link L d., Wol e hamp on, UK), 0.08% Fungizone ampho e izin B (Li e Technologies) and 0.1% hepa in sodium sal om po cine in es inal mucosa (Sigma-Ald ich, S . Louis, MO, USA). The HUVECs we e ou inely cul u ed on 5 μg/cm 2 collagen I (COLI, Sigma-Ald ich) coa ing, passaged a 80–90% confluency and used o expe imen s a passages 5–10. P ima y human e inal mic o ascula endo helial cells (hREC), ACBRI 181, a e a comme cially-a ailable cell line (Cell Sys ems, Ki kland, WA, USA). The cells we e g own on ECM A achmen Fac o ™(AF, Cell Sys ems). Fo he cul u e o ACBRI 181 a special CSC medium was used, consis ing o Comple e Se um-F ee Medium wi h Rocke Fuel and an ibio ic Bac-Off(Cell Sys ems), supplemen ed wi h 10% e al bo ine se um (FBS Gold) and 1% penicillin-s ep omy- cin (Camb ex Bio Science, Walke s ille, MD, USA). ACBRI181 we e used o expe imen s a passage 8. Be o e he co-cul u e expe imen s he p e e ence o cell cul u e subs a a o HUVEC and ACBRI181 cells was assessed by pla ing he cells on COLI, collagen IV (COLIV, Sigma-Ald ich) and AF. Polys y ene cell cul u e pla es we e coa ed wi h 5 μg/cm 2 o ECM by incuba ing o 1–3 h a 37 °C. The ea e he wells we e b iefly washed wi h phos- pha e buffe ed saline (DPBS, Lonza G oup L d.). Cells we e washed h ice wi h DPBS, dissocia ed wi h T ypsin-EDTA (Lonza G oup L d.), coun ed and passaged on op o coa ings a a densi y o 100,000 cells/ cm 2 . Cells we e cul u ed o wo days and fixed wi h 4% pa a o mal- dehyde (PFA, Sigma-Ald ich) solu ion and subjec ed o he indi ec immunofluo escence s aining (below). These p elimina y es s (Supplemen al Fig. 1) demons a ed ha he e was no ma ked diffe ence in he cell numbe on diffe en subs a a. Thus he COLIV which is ou inely used o hESC-RPE cell cul u e [39], can also be used o endo helial cells. 2.1.2. Human emb yonic s em cell de i ed e inal pigmen epi helial cells The hESC lines Regea11/013 (46, XY), Regea08/017 (46, XX) and Regea08/023 (46, XY) es ablished p e iously by ou g oup in Uni e si y o Tampe e, we e used [40]. The undiffe en ia ed hESCs we e cul u ed simila ly as p e iously desc ibed [41] a +37 °C in 5% CO 2 on human o eskin fib oblas eede cells (hFFs; 36,500 cells/cm 2 ; CRL-2429™; ATCC, Manassas, VA, USA) which we e mi o ically inac i a ed ei he wi h γ-i adia ion (40 Gy) o mi omycin C (10 μg/ ml, Sigma-Ald ich), in se um- ee condi ions in Knock-Ou Dulbecco's Modified Eagle Medium (KO-DMEM) con aining 20% Knock-Ou se um eplacemen (KO-SR), 2 mM Glu amax, 0.1 mM 2-me cap- oe hanol (all om In i ogen, Ca lsbad, CA, USA), 1% Minimum Essen ial Medium nonessen ial amino acids, 50 U/ml penicillin/s ep- omycin (bo h om Camb ex Bio Science), and 8 ng/ml human basic fib oblas g ow h ac o (bFGF; R & D Sys ems Inc., Minneapolis, MN). The cul u e medium was eplenished fi e imes a week. Undiffe en ia ed colonies we e passaged on o new eede cells ei he manually once a week o enzyma ically a en-day in e als wi h T ypLE Selec (In i ogen). Spon aneous RPE diffe en ia ion was induced by educing KO-SR concen a ion om 20% o 15% and emo ing bFGF om he hESC cul u e medium. This modified medium is called RPEbasic. The hESC colonies we e manually dissec ed and cul u ed in suspension in low cell bind six-well pla es (Nalgene, NUNC, Tokyo, Japan), as floa ing agg ega es (emb yoid bodies). The emb yoid bodies we e allowed o ma u e o 48–109 days un il sufficien pigmen a ion appea ed, changing he medium h ee imes a week. The pigmen ed a eas o floa ing agg ega es we e manually selec ed, cu and subsequen ly dissocia ed wi h T ypsin–EDTA. Acqui ed single cell suspensions we e fil e ed h ough 100 µm BD Falcon cell s aine (BD Biosciences, San Jose, USA), and cells we e pla ed on o well pla es coa ed wi h human COLIV (5 g/cm 2 ) o expand cell numbe s and pu i y he cell popula- ion. A e expansion o his passage 2 o 65‒349 days, he hESC-RPEs we e dissocia ed wi h T ypsin–EDTA, fil e ed h ough a s aine and coun ed o be pla ed o expe imen s, in which he hESC-RPE cells we e in passage 3. Cells sec e e se e al soluble ac o s o he cul u e medium. The human e al RPE cell [42] and ma u e hESC-RPE cell [43] condi ioned media ha e p e iously been shown o induce neu onal cell diffe en ia- ion [42,43]. As in he na i e e inal issue, he endo helial cells ecei e soluble queues om RPE. The e o e, p io he co-cul u e expe imen s we wished o es he abili y o HUVEC (always be ween he passage 7 H. Sko man e al. Expe imen al Cell Resea ch 359 (2017) 101–111 102 and passage 9), and ACBRI181 (always passage 8) cells o g ow in esh RPEbasic medium and RPEbasic condi ioned by ma u e hESC-RPE cells. The effec s we e compa ed o he cul u es whe e bo h endo helial cell ypes we e g own in hei own media. Cells we e cul u ed o 7 days and new es media we e changed e e y day. This p elimina y es sugges ed ha endo helial cells p e e condi ioned RPEbasic (Supplemen al Fig. 1) o e he esh RPEbasic medium. The e o e all co-cul u es we e main ained in co-cul u e medium –1:1 mix u e o hESC-RPE condi ioned RPEbasic and esh RPEbasic. 2.2. T ea men o polye hylene e eph hala e subs a a o he cell cul u e In o de o ha e equal cul u e condi ions o bo h cell ypes he anspa en polye hylene e eph hala e (PET) memb anes (Millipo e 12 Well Millicell, 1.0 µm po e diame e ; Me ck Millipo e, Bille ica, MA, USA) we e emo ed om he suppo ing hanging cul u e inse s wi h a scalpel and placed wi hin Minushee (pa no. 1300, Minucells and Minu issue-Ve iebs GmbH, Bad Abbach, Ge many) ca ie s, which we e cleaned using E OH (fi s 10 min in 70% E OH, hen 5 min in 99.5% E OH ollowed wi h d ying) be o e use. Minushee cons uc s we e coa ed wi h COLIV (5 µg/cm 2 ), o 1 h a +37 °C, wi h he memb anes flipped a e 30 min o ensu e ha he coa ing was equal on bo h sides. A e coa ing, he memb anes we e washed once wi h 1xPBS. The PET memb ane was always o ien ed simila ly wi hin he Minushee ca ie , o ensu e ha all samples we e uni o m. All wo k was ca ied ou unde he lamina flow hood. The pilo es s e ealed ha he endo helial cells had a he low adhe ence on he COLIV coa ed PET memb ane. The e o e we inc eased he amoun o eac i e binding si es wi h plasma ea men . PET memb anes wi h Minushee ca ie s we e placed on o a 12-well pla e, wi h he un ea ed side acing up. The plasma ea men was ca ied ou wi h O 2 gas in Pico, Diene Elec onic (Ebhausen, Ge many) using cons an 0.4 mBa p essu e and 50 W powe o 1 min o wi h Reac i e Ion E che (RIE Ad anced Vacuum Scandina ia AB, Sweden) wi h pa ame e s 30 scm, 30 W 30 m o , 1 min). The ea e Minushee cons uc s we e coa ed simi- la ly as wi hou plasma ea men . The seconda y pilo es exhibi ed ha plasma ea men inc eased he numbe o endo helial cells on PET memb ane a e 9–10 days o cul u e (Supplemen al Fig. 2). 2.3. Cell pla ing and ma u a ion o co-cul u e Human RECs and hESC-RPE cells we e co-cul u ed wi h he “con ac ing”me hod [23]. Bo h cell ypes we e pla ed immedia ely a e coa ing o he Minushee cons uc s wi h COLIV. Fi s , he endo helial cells (100,000 cells/cm 2 ) we e pla ed in 150 μl o CSC medium, and we e allowed o adhe e a 37 °C, 5% CO 2 o 4 h. The ea e he Minushee ca ie s we e flipped, he CSC medium was emo ed and he hESC-RPE cell suspension (100,000 cells/cm 2 ) was applied on o he o he side o he memb anes in RPEbasic. Then, addi ional 500 μl o he co-cul u e medium was added o he con- s uc s. A e 4 h o incuba ion i.e. a e hESC-RPE cells had a ached o he memb ane su ace, an addi ional 1.5 ml o he co-cul u e medium was added. The co-cul u e samples we e allowed o ma u e o 4–6 weeks (Regea08/017 36d ± 7d; Regea08/023 32d ± 4d and Regea11/013 34d ± 4d), changing he medium h ee imes a week. The co-cul u e medium was used du ing he fi s h ee weeks, and pu e RPEbasic o he emaining ime. The co-cul u es we e analyzed once he hESC- RPE cells egained hei pigmen a ion. 2.4. T ans-epi helial elec ical esis ance The ba ie p ope ies o he solo-cul u ed and co-cul u ed samples we e e alua ed wi h TEER measu emen s a he end o he ma u a ion pe iod: Each sample, cul u ed on he PET inse placed o he Minushee ca ie (Fig. 1A), was cu ou (Fig. 1A). The cul u e inse was cu o wo pieces (Fig. 1B), and clamped o P2307 slide (Physiologic Ins umen s, San Diego, CA, USA) (Fig. 1AB), placed in o a cus om-made eflon chambe [36] filled wi h 1xPBS, and TEER was measu ed wi h he Millicell ERS-2 Vol ohmme e (Me ck Millipo e) (Fig. 1C). TEER alues (Ωcm 2 ) we e calcula ed by sub ac ing he TEER o a simila ly ea ed subs a a-coa ed PET memb ane wi hou cells, and by mul iplying he esul by he su ace a ea. TEER alues we e ob ained om fi e indi idual diffe en ia ion expe imen s wi h all h ee diffe en hESC lines wi h mul iple pa allel samples, and wo echnical eplica es. 2.5. Pe meabili y es s A e he TEER measu emen , he samples in slide s P2307 we e placed in o he Ussing Chambe de ice (EM-CSYS-8 (Physiologic Fig. 1. The illus a ion how cells cul u ed o 4–6 weeks on PET fil e inse clamped o Minushee ca ie was ans e ed o ba ie unc ion es s. A e he 4–6 weeks o cell cul u e, he fil e inse wi h cells was cu ou om he Minushee ca ie (A). In o de o ha e wo echnical eplica es, he inse was cu o wo pieces (see a ows), and each one was placed o one P2307 slide (B). The ansepi helial elec ical esis ance was measu ed in a cus om-made P2307 slide chambe (in he inse in C) wi h Millicell ERS-2 Vol ohmme e (C). The sample was hen placed o EM-CYS-8 Ussing chambe wi hou he elec odes (D) o he pe meabili y assay. A e he assay he same cell samples we e p ocessed o immunofluo escence o elec on mic oscopy. H. Sko man e al. Expe imen al Cell Resea ch 359 (2017) 101–111 103 ins umen s)) wi h P2300 EasyMoun Diffusion Chambe s o measu e he ba ie p ope ies o he co-cul u es (Fig. 1D). The pe meabili y assessmen s we e ca ied ou in 2.25 ml o RPEbasic wi hou KO-SR ( he ea e e e ed o as Ussing medium) bo h on he dono and accep o side. On he dono side 1 mg/ml o 4 kDa fluo escein iso hiocyana e–dex an (FD4, Sigma-Ald ich) was used as he es molecule. Blank samples we e collec ed p io o he es om bo h dono and accep o media. The pH in chambe s was kep cons an wi h CO 2 gas (5% CO 2 , 10% O 2 , 85% N 2 ) bubbling. A e 1 h, 2 h, 3 h and 4 h, wo pa allel 100 μl samples we e aken om he accep o side o he chambe , and 200 μl o esh medium was added o balance he emo ed olume. Samples we e analyzed wi h Wallac Vic o 2 1420 mul ilabel coun e spec opho ome e (Tu ku, Finland). The cumula i e pe meabili y i.e. he cumula i e amoun o fluo es- cence appea ing in he accep o solu ion e sus ime, was calcula ed om he a e age alue o he wo pa allel abso bance measu emen alues o he accep o (basal side) om which sub ac ing he alue o pu e Ussing medium (i.e. blank). This alue was hen di ided by he o al olume o accep o side. The dilu ion effec o he added esh medium a e e e y hou . This was calcula ed acco ding o Pi kanen e al. [44]. Cumula i epe meabili y Faccep o Fpu e medium medium ol ml =(− ) ()*) F is he emi ed fluo escence alue. Accep o is in he beginning de oid o fluo escence, si ua ed in chambe basal o hESC-RPE, apical o EC. Pu e medium is esh medium wi hou fluo escence. (*) in equa ion is he dilu ion effec o esh medium a e e e y hou . The cumula i e amoun pe mea ed was calcula ed om cumula i e pe meabili y by di iding by he spec opho ome e alue o he exposu e medium and mul iplied by 100%. Cumula i e pe meabili y Cumula i e pe meabili y F dono %= *100 The diffusion o he solu es ac oss cell cul u es was cha ac e ized by calcula ing he appa en pe meabili y coefficien (Papp, cm/sec) ac- co ding o Pi kanen e al. [44], cumula i e pe meabili y o FD4 a e e e y hou was plo ed, and he slope (kx+b) o he linea po ion, was di ided wi h he su ace a ea (0.031 cm 2 ) mul iplied wi h he spec o- pho ome ic alue (i.e. maximum in ensi y) o he p obe molecule in he dono solu ion. Pe meabili y coe icien kx b su ace a ea cm F dono =(+) (2)* Dono in he equa ion is he maximum in ensi y om he Ussing medium ha con ains 1 mg/ml o FD4. 2.6. Indi ec immunofluo escence A e he pe meabili y es , indi ec immunofluo escen s aining was pe o med in o de o e alua e cell mo phology, s uc u e, and pola i y o he samples. All samples we e fi s washed wi h 1×PBS, fixed wi h 4% pa a o maldehyde (Sigma-Ald ich) o 10 min a RT, washed ou imes wi h 1×PBS, and pe meabilized wi h 0.1% T i onX- 100 (Sigma-Ald ich) in 1×PBS o 10 min a RT. A e ha , he samples we e again washed epea edly wi h PBS, and 3% bo ine se um albumin (BSA, Sigma-Ald ich) in PBS was added o he samples o 1–1.5 h a RT, o o e nigh a 4 °C, o block he non-specific binding si es. Then, he samples we e incuba ed o 1–1.5 h a RT o o e nigh a 4 °C wi h he p ima y an ibodies in 0.5% BSA-PBS. The endo helial cells we e s ained wi h abbi -an i- on Willeb and ac o ( W , 1:500, Daco, Glos up, Denma k); hESC-RPE cells we e s ained wi h mouse-an i- CRALBP (1:500, Abcam, Camb idge, UK). The igh junc ions we e s ained wi h mouse-an i-ZO-1 (1:250 In i ogen), and filamen ous ac in was s ained wi h amani a phallo oxin Phalloidin 680 (1:200, Sigma-Ald ich). The p ima y an ibody incuba ions we e ollowed by 4× washes wi h 1×PBS. The seconda y an ibodies, donkey-an i-mouse and goa -an i- abbi (bo h om Li e Technologies) we e bo h used a dilu ion o 1:1000 in 0.5% BSA-PBS. The seconda y an ibodies we e incuba ed o 1–1.5 h a RT. Finally, he samples we e washed ×4 wi h 1×PBS moun ed be ween wo co e glasses wi h Vec ashield®moun - ing medium wi h 4′,5-diamidino-2-phenylindole (DAPI) (Vec o Labo a o ies Inc.). The isualiza ion and imaging o he s ained samples was ca ied ou wi h LSM 780 con ocal mic oscope (Ca l Zeiss, Jena, Ge many) using magnifica ion o 40× and esolu ion o 1200×1200. 2.7. Enzyme-linked immunoso ben assay The concen a ions o sec e ed PEDF and VEGF we e de e mined using Chemikine PEDF Sandwich ELISA Ki (Millipo e, Temecula, CA, USA) and Quan ikine VEGF ELISA Ki (R & D Sys ems), espec i ely, acco ding o he manu ac u e 's ins uc ions. 2.8. T ansmission elec on mic oscopy A e 4–6 weeks o co-cul u e, he cons uc s we e fixed o 2 h a RT wi h 2% glu a aldehyde (Elec on Mic oscopy Sciences, Ha field, PA, USA) in 0.1 M phospha e buffe , and washed wi h 0.1 M phospha e buffe . The samples we e pos -fixed wi h 1% osmium e oxide (Ladd Resea ch, Willis on, VT, USA) o 1 h a RT. The samples we e hen dehyd a ed h ough an o de ed ace one se ies: 70% ace one, 94% ace one, and absolu e ace one (J.T. Bake ; A an o Pe o mance Ma e ials, B.V. De en e , Ne he lands). The ea e he samples we e imp egna ed wi h 1:1 mix u e o absolu e ace one and epoxy esin (Ladd Resea ch) o 1.5 h a RT. The embedding o pu e epoxy esin was done o e nigh a RT, and polyme iza ion o epoxy esin o 48 h a 60 °C. Thin sec ions we e s ained wi h 1% u anyl ace a e o 30 min and wi h 0.4% lead ci a e (Fluk, S einheim, Swi ze land) o 5 min. Samples we e examined and imaged wi h JEM-2100F TEM (Jeol L d., Tokyo, Japan). The leng h o mic o illi, cell hickness, and basal lamina hickness we e measu ed wi h Image J Image P ocessing and Analysis So wa e om he TEM images. 2.9. S a is ical analyses The s a is ical significance o TEER, pe meabili y assay and ELISA we e analyzed wi h PASW S a is ics, e sion18 wi h wo ailed Mann- Whi ney U. The image analysis om TEM images was calcula ed wi h R s a is ical compu ing and g aphics p og am ( e sion 3.3.1) (R Founda ion o S a is ical Compu ing, Vienna, Aus ia Founda ion o R Founda ion o S a is ical Compu ing, Vienna, Aus ia a is ical Compu ing, Vienna, Aus ia) wi h wo-sample Wilcoxon ank sum es . The numbe o eplica es is indica ed in he figu e legends. Resul s we e conside ed significan a p < 0.05. 2.10. E hical issues The ex ac ion o HUVECs om he umbilical co ds om ull e m scheduled Cesa ean sec ions has he suppo i e s a emen om he E hics Commi ee o Pi kanmaa Hospi al Dis ic (Mie inen/R13019). The Na ional Au ho i y o Medicolegal Affai s Finland has gi en app o al o s udy human emb yos (Dn o 1426/32/300/05) and we ha e a suppo i e s a emen om he E hics Commi ee o Pi kanmaa Hospi al Dis ic o de i e hESC lines om ea ly s age emb yos ha a e no used in in e ili y ea men s (Sko man/R05116). No new hESC lines we e gene a ed in his s udy. H. Sko man e al. Expe imen al Cell Resea ch 359 (2017) 101–111 104 Fig. 2. Rep esen a i e con ocal o ho-images a e 4–6 weeks o hESC-RPE solo-cul u es (A-C, J-L) and hESC-RPE and ACBRI181 co-cul u es (D-I, M-R) (n= 4, all wi h 2 biological and 2 echnical eplica es). Samples we e immunos ained wi h RPE specific ma ke CRALBP (g een) and endo helial cell specific on Willeb and ac o ( WF, ed) (A-I) o igh junc ion ma ke ZO-1 (g een) o filamen ous ac in binding phalloidin ( ed) (J-R). Nuclei we e coun e s ained wi h DAPI (blue). Images D) Regea08/017, E) Regea08/023 and F) Regea11/0113 om he hESC-RPE le el and G,H, and I om he endo helial le el o he same co-cul u es. Simila ly images M) Regea08/017, N) Regea08/023 and O) Regea11/0113 om he hESC- RPE le el and P,Q and R om he endo helial le el o he same co-cul u es. Scale ba s 10 µm. H. Sko man e al. Expe imen al Cell Resea ch 359 (2017) 101–111 105 3. Resul s 3.1. Cell mo phology Exp ession and localiza ion o cell ype specific p o eins was e alua ed using indi ec immunofluo escence a e 4–6 weeks o co- cul u e o hESC-RPE cells and ACBRI181 hRECs. Localiza ion o CRALBP, he RPE-specific p o ein, was simila in solo-cul u ed hESC-RPEs (Fig. 2A, B, C) and co-cul u ed hESC-RPEs (Fig. 2D, E, F). Also he endo helial cell specific WF labeling localized o hRECs (Fig. 2G, H, I) in co-cul u es. As hRECs we e isualized wi h he con ocal mic oscope h ough he pigmen ed hESC-RPE cells and h ough he 1 µm po es in he PET memb ane he in ensi y o he label is sligh ly ain e on one side. The ZO-1 igh junc ion p o ein had a simila appea ance bo h in solo-cul u ed (Fig. 2J, K, L) and co- cul u ed (Fig. 2M, N, O) hESC-RPEs. The co-cul u e o hESC-RPE cells appea ed o effec on epi helial cul u e: he endo helial cell solocul u e had ewe islands han en- do helial cells co-cul u ed wi h hESC-RPE cells (Supplemen al Fig. 3). Also he endo helial cells cul u ed wi h Regea08/017 and Regea08/023 we e mo e confluen han hose cul u ed wi h Regea11/013 (Supplemen al Fig. 4 and Supplemen al Fig. 5). The endo helial cells and hESC-RPEs migh in e ac wi h each o he as seen in a con ocal image whe e endo helial cell has sen a WF posi i e p ocess hough he 1 µm po es o PET fil e inse (Supplemen al Fig. 6). In addi ion, endo helial cells in e ac ed wi h each o he by o ming capilla y-like s uc u es which we e isible in all cul u es (whi e a ows in Supplemen al Fig. 4) and (whi e a owhead in Supplemen al Fig. 5). Cell mo phology and pigmen a ion o hESC-RPE cells we e un- affec ed by he co-cul u e, as seen om he low magnifica ion elec on mic oscopy images (Fig. 3A). Du ing ma u a ion and pola iza ion he RPE cells a e known o adop a mo e columna shape, hus he hESC- RPE cell heigh was quan ified (Fig. 3B) o solo- and co-cul u ed hESC-RPEs om TEM pho og aphs (Table 1). We ound ha co- cul u e significan ly dec eased he cell heigh o Regea08/017 hESC- RPE line (p=0.0014), bu no Regea08/023 (p=0.331) o Regea11/013 (p=0.364) hESC-RPE cell lines. Du ing cell ma u a ion and pola iza ion he b ush bo de is de eloped on he apical side o RPE cells. Mic o illi leng h was quan ified om TEM images o he solo- and co-cul u ed hESC-RPE (Fig. 4A-C), (Table 2). In Regea08/023 hESC-RPE cells he b ush bo de was significan ly sho e in solo-cul u es han in co-cul u es (p < 0.001). In Regea11/013 he mic o illi we e also sligh ly sho e in solo-cul u es han in co-cul u es bu his diffe ence did no each s a is ical significance (p = 0.058). Finally, due o se e al ou lie s in Regea08/017 hESC-RPE, he diffe ences we e no s a is ically signifi- can (p = 0.896). I should be no ed ha ul a hin sec ions (app ox. 70 nm) do no cap u e he ull leng h o he mic o illi bu a he a p ojec ion o hem in specific laye s o he c oss-sec ion. Howe e , quali a i ely, one can es ima e ha bo h he leng h and he numbe o mic o illi a e inc eased in co-cul u es. The cell-cell in e ac ion can affec ECM o ganiza ion and deposi- ion. In all solo-cul u es and co-cul u es he deposi ed ECM had a highly o ganized s uc u e (Fig. 4D,E). Quan i a ion o he diffe ences would no gi e eliable esul s. The e o e, only he ECM hickness was measu ed om TEM images o he solo- (Fig. 4D) and co-cul u ed hESC-RPEs (Fig. 4E), (Table 3). The co-cul u e significan ly inc eased he ECM deposi ion unde nea h he Regea11/013 hESC-RPE (p=0.020). The diffe ences we e no s a is ically significan o Regea08/017 (p=0.443) o Regea08/023 (p=0.068) hESC-RPE cells. 3.2. Ba ie p ope ies The in eg i y o ma u e cell cul u es was fi s analyzed ia TEER measu emen immedia ely a e cu ing he inse om he Minushee ca ie and placing i o he P2307 slide . This also ensu ed ha he cells we e placed igh ly on he slide . A e he 4–6 weeks solo-cul u e, ACBRI181 had eached he TEER o 3.9 Ωcm 2 (Fig. 5A). Co-cul u e wi h ACBRI181 endo helial cells significan ly inc eased TEER o Regea08/023 hESC-RPE (p=0.013). Fo he o he wo cell lines, al hough TEER was highe in co-cul u e, he diffe ences we e no significan due o high a ia ion (p=0.059 o Regea08/017 and p=0.310 o Regea11/013). When all he solo-cul u es o he h ee cell lines we e compa ed o all he co-cul u es, he o e all diffe ence in TEER was s a is ically significan (p = 0.002), (Fig. 5B). The TEER o p is ine PET fil e inse wi hou cells (0.3 cm 2 ) analyzed wi h Millicell ERS-2 Vol ohmme e p io cu ing, was 45 Ωcm 2 −54 Ωcm 2 , whils he same clamped inse (0.031 cm 2 ) wi h he same Millicell ERS-2 Vol ohmme e was 25 Ωcm 2 −29 Ωcm 2 . These esul s indica e ha he TEER was sligh ly comp omised by cu ing and clamping. The igh ness o he cul u es was also e alua ed by measu ing fluo escen dye pe meabili y in Ussing chambe o 4 h (Fig. 5C). The cumula i e pe meabili y o he emp y PET memb ane inc eased om 0.036% o 1.94% in his ime. Fo he ACBRI181 endo helial cells in solo-cul u e he inc ease was om 0.26% o 1.47%. Co-cul u e wi h ACBRI181 cells significan ly dec eased he cumula i e pe meabili y o Regea08/023 hESC-RPE (p=0.013). The diffe ences o he o he wo cell lines we e no significan (p=0.221 o Regea08/017 and p=0.445 o Regea11/013). Simila ly o TEER alues, when all solo-cul u es o he h ee cell lines we e compa ed o all he co-cul u es, he dec ease in cumula i e pe meabili y was s a is ically significan (p = 0.018) (Fig. 5D). The co-cul u e wi h ACBRI181 cells also dec eased he pe meabili y coefficien (10 −6 cm/sec) bu ha was no s a is ically significan Regea08/017 hESC-RPE (p = 0.17), Regea08/023 (p = 0.11) o Regea11/013 (p=0.73) (Fig. 5E). Bu when all solo-cul u es o he h ee cell lines we e compa ed o all he co-cul u es, he dec ease in cumula i e pe meabili y was s a is ically significan (p = 0.036) (Fig. 5F). 3.3. G ow h ac o sec e ion Finally, we assessed how he co-cul u e o hESC-RPE cells wi h ACBRI181 endo helial cells affec s g ow h ac o exp ession. ACBRI181 cells did no sec e e de ec able PEDF (Fig. 6A), as expec ed, and only sec e ed 0.56 ng/ml VEGF (Fig. 6B). PEDF sec e ion was significan ly highe in co-cul u ed Regea08/017 (p = 0.00004) and Regea08/023 (p = 0.000005) hESC-RPE cells, bu no in Regea11/013 hESC-RPE. VEGF exp ession, on he o he hand, was significan ly highe in co-cul u ed Regea08/017 and Regea11/13 hESC-RPE (p=0.032 and p=0.014, espec i ely), bu lowe o co-cul u ed Regea08/023 hESC-RPE (p=0.025). 4. Discussion In i o modeling o heal hy unc ional RPE equi es a pola ized and ma u e monolaye o RPE cells cha ac e ized by s ong ba ie p ope ies and low pe meabili y [23]. In addi ion, hese RPE cell cul u es should unc ion simila ly o na i e cells. Cell cul u e me hods should be consis en and efficien , yielding high quali y cell popula ions wi h li le a ia ion be ween ba ches. P ima y cells ex ac ed om issue end o diffe om isola e o isola e and he yield is always a he low [9]. Immo alized RPE cell lines p o ide a high yield, bu hey lack some o he cha ac e is ics, gene, and p o ein exp ession, c i ical o cell unc ions no mally obse ed in RPE [45]. Human ESCs offe a solu ion o he limi a ions o p ima y and immo alized cell lines – hey p o ide an unlimi ed sou ce o RPE cells, which ha e many unc ionally impo an cha ac e is ics common o na i e RPE cells [28–33]. P e iously, hESC-RPE cells ha e been used o in i o modeling o oBRB in solo-cul u es [28,36,46]. To ou knowledge ou s udy is he fi s o demons a e he use o hESC-RPE co-cul u ed wi h hRECs as a po en ial in i o model o oBRB. In ou ea lie s udies, we ha e shown ha diffe en ia ed hESC-RPE H. Sko man e al. Expe imen al Cell Resea ch 359 (2017) 101–111 106 cell cul u es adop cobbles one mo phology, exp ess RPE specific p o eins and localize hem co ec ly in solo-cul u es [41,47]. The p esen s udy shows ha co-cul u ing hESC-RPE cells wi h endo helial cells yield simila cobbles one mo phology and localiza ion o he RPE- specific p o ein CRALBP. Also, exp ession and localiza ion o WF in hRECs was unal e ed in hese co-cul u es. The e is no published co esponding da a, bu he esul s a e in acco dance wi h he da a p e iously epo ed in ARPE-19 co-cul u ed wi h HUVECs [22,48]. ARPE-19 cell line, howe e , has impo an dissimila i ies wi h human p ima y RPE cell cul u es, mos impo an ly low o lacking pigmen a- ion and pola iza ion [22,49]. In ou s udy, co-cul u ing hESC-RPE cells wi h ACBRI181 did no affec ei he cell pigmen a ion o o e all cell mo phology. Co-cul u ed hESC-RPE had high amoun s o pigmen- a ion, isible igh junc ions, hick b ush bo de and highly o ganized ECM. Al hough he e we e diffe ences in cell heigh , mic o illi leng h and ECM hickness, hese all we e wi hin he ange o expec ed biological a ia ion. The queues om endo helial cells al e RPE cell unc ions [3], and biological solu es hey sec e e modi y RPE ba ie unc ion in co- cul u es [22,50]. Human CECs and RECs sec e e diffe en solu es [51], and o he pu pose o an oBRB model hCECs would likely be he op imal choice. Howe e , acco ding o ou cu en knowledge hRECs a e comme cially a ailable, while hCECs a e no . P ima y bo ine RPE ha e p e iously shown o ha e highe TEER in solo-cul u e han in con ac ing and non-con ac ing co-cul u es wi h endo helial cells [50], bu human RPE and p ima y EC co-cul u es had highe TEER han RPE solocul u e [52]. On he o he hand, co-cul u ing ARPE-19 wi h HUVEC o 4 weeks on opposi e sides o polyes e memb anes inc eased TEER [48]. In ou s udy, co-cul u ing hESC-RPEs wi h hRECs clea ly inc eased TEER, al hough he diffe ence was only significan o one o he h ee cell lines. When he igh ness o he cul u es was e alua ed by assessing he leakage o flux o small molecules om dono o accep o side o he cul u e, all ou cell lines demons a ed dec eased pe meabili y in co-cul u es. Simila findings Fig. 3. A) The ansmission elec on mic og aphs a e 4–6 weeks o hESC-RPE solo-cul u es o hESC-RPE and ACBRI181 co-cul u es (n= 3 wi h 2 biological eplica es). O de om op o bo om: hESC-RPE cells, PET memb ane, and in co-cul u es hRECs. F om le o igh : Regea08/017 solo-cul u e; Regea08/017+ACBRI101 co-cul u e; Regea08/023 solo- cul u e; Regea08/023+ACBRI181 co-cul u e; Regea11/013 solo-cul u e; and Regea11/013+ACBRI181 co-cul u e. Scale ba s 10 µm. B) The heigh o hESC-RPE cells was measu ed om he mic og aphs using Image J analysis so wa e and p esen ed as box plo s. The hicke lines indica e medians. S a is ical significance p < 0.05 is indica ed wi h *. Table 1 The hESC-RPE cell heigh s measu ed om he TEM pho og aphs. Cul u e numbe o samples o al numbe o measu emen s ange median s a is ical signi icance solo- s. co-cul u e Regea 08/017 2 6 18.00 µm - 20.09 µm 19.54 µm p=0.0014 Regea 08/017+ACBRI181 9 28 10.12 µm - 17.82 µm 14.65 µm Regea 08/023 10 23 6.67 µm - 18.49 µm 9.16 µm p=0.3311 Regea 08/023+ACBRI181 9 32 7.88 µm - 12.33 µm 9.27 µm Regea 11/013 2 8 15.37 µm - 19.28 µm 16.57 µm p=0.3638 Regea 11/013+ACBRI181 5 10 10.50 µm - 19.99 µm 17.95 µm H. Sko man e al. Expe imen al Cell Resea ch 359 (2017) 101–111 107 ha e been epo ed p e iously wi h ARPE-19 and HUVEC co-cul u es [22] and p ima y bo ine [50] RPE o p ima y human RPE [52] and EC co-cul u es. The pe mea ion coefficien (P app ) o he FD4 h ough co- cul u es was 10 imes highe (2.93 + 10 −6 cm/s) han hough bo ine RPE-cho oid (2.36 × 10 −7 cm/s) [44], bu s ill he co-cul u e was 2 imes igh e han hESC-RPE solocul u es (5.64 + 10 −6 cm/s). The inc ease in TEER and dec ease in pe meabili y indica es ha hRECs imp o e he ba ie unc ions o RPE a leas in i o. This is also impo an o ecognize when hese in i o assays a e used o pe meabili y and anspo analysis o d ugs o o he molecules. The balance o PEDF and VEGF le els is impo an o e inal homeos asis as i s imula es endo helial cell mig a ion and RPE ma u a ion [53,54]. In he de eloping e ina PEDF is a diffe en ia ing ac o o RPE cells. Howe e , in ma u e e ina PEDF is a ophic ac o Fig. 4. The ansmission elec on mic og aphs a e 4–6 weeks o hESC-RPE solo-cul u es o hESC-RPE and ACBRI181 co-cul u es (n= 3 wi h 2 biological eplica es). A) Rep esen a i e images o he apical b ush bo de o solo-cul u ed hESC-RPE cells, and (B) hESC-RPE+ ACBRI181 co-cul u es. Scale ba 1 µm. (C) The leng h o apical mic o illi o hESC-RPE cells measu ed om he mic og aphs using Image J analysis so wa e and p esen ed as box plo s (C). The hicke lines indica e medians. S a is ical significance was analyzed bu no ound. D) Rep esen a i e image o ECM unde nea h solo-cul u ed hESC-RPE cells, and (E) hESC-RPE+ ACBRI181 co-cul u es. Scale ba 1 µm. (F) The heigh o hESC-RPE cells was measu ed om he mic og aphs using Image J analysis so wa e and p esen ed as box plo s. The hicke lines indica e medians. S a is ical significance p < 0.05 is indica ed wi h *. Table 2 The mic o illus leng hs in hESC-RPE cells measu ed om he TEM pho og aphs. Cul u e numbe o samples o al numbe o measu emen s ange median s a is ical signi icance solo- s. co-cul u e Regea 08/017 2 7 0.78 µm - 2.02 µm 1.20 µm p=0.8956 Regea 08/017+ACBRI181 13 51 0.71 µm - 8.49 µm 1.16 µm Regea 08/023 11 36 0.49 µm - 1.92 µm 1.03 µm p=0.0002 Regea 08/023+ACBRI181 11 54 0.75 µm - 2.15 µm 1.39 µm Regea 11/013 5 14 0.50 µm - 2.86 µm 1.395 µm p=0.0580 Regea 11/013+ACBRI181 5 14 0.86 µm - 2.47 µm 1.53 µm Table 3 The ECM hickness unde nea h hESC-RPE cells measu ed om he TEM pho og aphs. Cul u e Numbe o samples To al numbe o measu emen s Range Median S a is ical signi icance solo- s. co-cul u e Regea 08/017 2 6 0.97−1.25 µm 1.13 µm p = 0.4434 Regea 08/017+ ACBRI181 11 40 0.23−5.67 µm 1.04 µm Regea 08/023 14 54 0.31−0.82 µm 0.58 µm p = 0.0676 Regea 08/023+ ACBRI181 13 51 0.48−2.24 µm 1.21 µm Regea 11/013 7 30 0.48−3.02 µm 1.09 µm p = 0.020 Regea 11/013+ACBRI181 5 20 0.33−4.86 µm 2.22 µm H. Sko man e al. Expe imen al Cell Resea ch 359 (2017) 101–111 108 Fig. 5. A) The ba ie p ope ies o a e 4–6 weeks o hESC-RPE solo-cul u es o hESC-RPE and ACBRI181 co-cul u es (n= 3, wi h 2 biological and 2 echnical eplica es. Da a a e exp essed as mean +/-SD). (A) TEER alues om indi idual cell lines as solo-cul u es o endo helial cells, solo-cul u es o hESC-RPEs, and hESC-RPE+ACBRI181 co-cul u es, and(B) pooled cell lines, ** p < 0.01; *** p < 0.001. Cumula i e pe meabili y % o 4 kDa Fi c dex an (FD4, Sigma-Ald ich, Ge many) o solo- and co-cul u es a e he 4–6 weeks cul u e (n= 3, wi h 2 biological and 2 echnical eplica es) in indi idual cell lines in solo- and co-cul u es om 60 min o 240 min (C), and (D) in pooled cell lines a he 240 min ime-poin . Da a a e exp essed as mean+/-SD, * p < 0.05; *** p < 0.001. Pe meabili y o FD4 exp essed as pe meabili y coefficien (10 −6 cm/s) (n= 3, wi h 2 biological and 2 echnical eplica es. Da a a e exp essed as mean +/-SD) (E) om indi idual cell lines and in pooled cell lines, * p < 0.05. H. Sko man e al. Expe imen al Cell Resea ch 359 (2017) 101–111 109