1
Scien i ic RepoR s | 7: 6016 | DOI:10.1038/s41598-017-06233-9
www.na u e.com/scien i ic epo s
Compa a i e p o eomic analysis o
human emb yonic s em cell-de i ed
and p ima y human e inal pigmen
epi helium
Heidi Hongis o1, An i Jylhä2, Janika Nä inen1,2, Jochen Rieck1, Tanja Ilma inen1, Zol án
Ve éb3, Ulla Aapola2, Roge Beue man2,4, Go an Pe o ski3,5, Hannu Uusi alo2,6 & Heli
Sko man1
Human emb yonic s em cell-de i ed e inal pigmen epi helial cells (hESC-RPE) p o ide an unlimi ed
cell sou ce o e inal cell eplacemen he apies. Clinical ials using hESC-RPE o ea diseases such
as age- ela ed macula degene a ion (AMD) a e cu en ly unde way. Human ESC-RPE cells ha e been
ho oughly cha ac e ized a he gene le el bu hei p o ein exp ession p o ile has no been s udied a
la ge scale. In his s udy, p o eomic analysis was used o compa e hESC-RPE cells di e en ia ed om
wo independen hESC lines, o p ima y human RPE (hRPE) using Isoba ic ags o ela i e quan i a ion
(iTRAQ). 1041 common p o eins we e p esen in bo h hESC-RPE cells and na i e hRPE wi h majo i y
o he p o eins simila ly egula ed. The hESC-RPE p o eome e lec ed ha o no mal hRPE wi h a
la ge numbe o me abolic, mi ochond ial, cy oskele al, and anspo p o eins exp essed. No signs o
inc eased s ess, apop osis, immune esponse, p oli e a ion, o e inal degene a ion ela ed changes
we e no ed in hESC-RPE, while impo an RPE speci ic p o eins in ol ed in key RPE unc ions such as
isual cycle and phagocy osis, could be de ec ed in he hESC-RPE. O e all, he esul s indica ed ha he
p o eome o he hESC-RPE cells closely esembled ha o hei na i e coun e pa s.
The e inal pigmen epi helium (RPE) is a mul i unc ional, pola ized epi helial cell laye be ween he neu osen-
so y e ina and he cho oid, which plays key oles in pho o ecep o unc ion and ision. The RPE cells anspo
nu ien s, was e p oduc s, ions and luid be ween he cho oidal blood supply and he sub e inal space. RPE also
phagocy izes shed pho o ecep o ou e segmen s (POS), abso bs sca e ed ligh , sec e es many impo an signal-
ling molecules and unc ions in he e inoid isual cycle1. This highly me abolically ac i e cell ype is exposed o
cons an ligh s imuli and high oxida i e s ess making i ulne able o oxida i e damage. Thus, abno mali ies in
RPE cell unc ion may lead o e inal degene a ion and pho o ecep o cell dea h. The RPE is he ocal poin o
many e inal degene a i e diseases such as age- ela ed macula degene a ion (AMD), he mos common cause
o blindness in he elde ly in wes e n coun ies. AMD is a mul i ac o ial, age-associa ed disease cha ac e ized by
accumula ion o insoluble d usen in he e ina, degene a ion o RPE and pho o ecep o s in he d y o m, and
cho oidal neo ascula iza ion in he exuda i e, we o m o he disease2.
T ea men op ions o he e inal degene a i e diseases such as AMD a e cu en ly e y limi ed and mos ly
only delay disease p og ession. Cellula ansplan a ion o eplace he a ec ed RPE is conside ed as a p omis-
ing he apeu ic s a egy o ea hese diseases. Macula ansloca ion and au ologous RPE ansplan a ion wi h
pe iphe al RPE ha e demons a ed he easibili y and e ec i eness o au ologous RPE cell eplacemen he apy in
AMD pa ien s, bu hese su gical p ocedu es ca y signi ican complica ions3. Many cell ypes ha e been es ed as
a sou ce o RPE ansplan a ion issue including oe al RPE4 and RPE cell lines5, 6. Issues ela ed o sca ce issue
1BioMediTech Ins i u e, Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland.
2Depa men o Oph halmology, Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland.
3S em Cells and Eye Resea ch Labo a o y, Depa men o Oph halmology, Facul y o Medicine, Uni e si y o Szeged,
Szeged, Hunga y. 4Singapo e Eye Resea ch Ins i u e and Duke-NUS School o Medicine, Singapo e, Singapo e.
5Cen e o Eye Resea ch, Depa men o Oph halmology, Oslo Uni e si y Hospi al and Uni e si y o Oslo, Oslo,
No way. 6Tampe e Uni e si y Hospi al Eye Cen e , Uni e si y o Tampe e, Tampe e, Finland. Co espondence and
eques s o ma e ials should be add essed o H.H. (email: heidi.m.hongis o@s a .u a. i)
Recei ed: 3 Augus 2016
Accep ed: 12 June 2017
Published: xx xx xxxx
OPEN
www.na u e.com/scien i ic epo s/
2
Scien i ic RepoR s | 7: 6016 | DOI:10.1038/s41598-017-06233-9
a ailabili y and cha ac e is ics o immo alized adul human cell lines, and he ac ha hey only weakly mimic
some o he na i e RPE cha ac e is ics a e in i o passaging, make hese cells subop imal o ea men o he
la ge popula ion o pa ien s7, 8.
Human plu ipo en s em cells (hPSC), including bo h human emb yonic s em cells (hESC) and human
induced plu ipo en s em cells (hiPSC) can be di e en ia ed o e inal cells, including pho o ecep o s and ma u e
and unc ional RPE cells9. Thei high capaci y o sel - enew and wide di e en ia ion po en ial makes hem an
excellen cell sou ce o bo h cellula models o esea ch pu poses as well as cell eplacemen he apy app oaches.
Encou aging esul s ha e shown ha ansplan ed hESC-de i ed RPE cells (hESC-RPE) can media e unc ional
pho o ecep o escue in he Royal College o Su geons (RCS) a model o e inal degene a ion10–12. Mo eo e ,
ongoing phase I/II clinical s udies ha e ecen ly demons a ed ha i is possible o sa ely implan hESC-RPE
o end s age pa ien s wi h AMD and o he e inal degene a i e diseases13. Simila s udies wi h au ologous
hiPSC-RPE ha e also been ini ia ed in Japan14, al hough suspended la e o one o he wo pa ien s due o sa e y
conce ns ega ding genomic s abili y o he pa ien ’s hiPSCs15.
Ou esea ch g oup, along wi h many o he s, has shown ha he hESC-RPE s uc u e, unc ion, and physi-
ology closely esembles ha o hei na i e coun e pa s wi h a high a e o pigmen a ion, polygonal, cuboidal
epi helial cell mo phology, cellula ine s uc u e, and exp ession o many RPE signa u e genes and p o eins16–22.
In addi ion he cells show epi helial in eg i y and unc ionali y wi h he abili y o phagocy ose POS and sec e ion
o g ow h ac o s18. Howe e , la ge scale compa a i e s udies o he p o eome, he o al p o ein complemen o
a genome, o he hESC-RPE cells a e lacking, while he p o eome o hRPE, especially ela ed o AMD has been
ex ensi ely s udied7, 23–25. S ingen cha ac e iza ion o hPSC-RPE using s anda dized and quan i iable me hods
is impo an o gain con idence ha hey a e sa e and adequa e eplacemen s o diseased RPE i hey a e u ilised
in clinical se ings.
In his s udy, we ha e pe o med a compa a i e p o eomic analysis o in i o di e en ia ed hESC-RPE cells o
p ima y hRPE using Isoba ic ags o ela i e quan i a ion (iTRAQ). iTRAQ quan i ica ion is based on inco po-
a ion o s able iso opes in o pep ides ollowed by liquid ch oma og aphy ac iona ion and andem mass spec-
ome y analysis. iTRAQ-based p o eomics has p e iously been used o example o s udy ea ly s age hESC
di e en ia ion26 and changes in isual cycle p o eins in mice RPE in conjunc ion wi h RPE65 gene deli e y27. We
ha e also p e iously success ully applied he echnique o cha ac e iza ion o co neal epi helial cells de i ed om
hPSC28. In his s udy, he esul s showed ha he o e all spec um o p o ein exp ession o he hESC-RPE cells
and he p ima y RPE we e simila .
Resul s
A la ge numbe o common p o eins we e exp essed and egula ed a simila le el in hESC-RPE
and hRPE. iTRAQ p o eomics was pe o med o compa e p o ein exp ession o hESC-RPE o p ima y hRPE.
The expe imen al se up and wo k low o he s udy is shown in Fig.1a. Th ee eplica e samples o wo hESC lines
Regea08/017 and Regea08/023 di e en ia ed o RPE cells ( e e ed o as 08/017 and 08/023 hESC-RPE), we e
compa ed o hRPE samples om h ee di e en cada e ic human dono eyes. Diges ed pep ides we e labelled
wi h iTRAQ eagen s as shown in Fig.1a, inco po a ing one biological sample o 08/017, one sample o 08/023,
and one sample o hRPE o each o he analysis eac ions. The labelled, combined samples we e subjec ed o h ee
LC-MS/MS analyses, which we e conduc ed in wo echnical eplica es each. Fo da a analysis he eplica e esul s
we e combined as shown o allow ela i e quan i a i e compa ison o he p o ein exp ession in hESC-RPE and
hRPE.
The iTRAQ p o eomics allows he iden i ica ion and quan i ica ion o p o eins ha a e exp essed in all sam-
ples included in each labelling (Analysis 1.1–3.2 in Fig.1). 2643 unique p o eins (Unip o Accession) exp essed
by bo h he hESC-RPE and hRPE we e iden i ied in a leas one echnical eplica e sample (Fig.1b). The numbe
o p o eins p esen a e s ingen da a p ocessing and il e ing o ensu e da a quali y a e shown in Fig.1b. 1041
unique p o eins exp essed in a leas wo biological eplica e samples o bo h 08/017 and 08/023 hESC-RPE, and
in he hRPE dono sample we e included o inal da a analysis. A comple e lis o he p o eins is p o ided in
Supplemen a y TableS1. Log2 scale was used o analysing he da a. As p o ein exp ession o he hESC-RPE we e
compa ed o he p o ein exp ession in hRPE, o e wo- old di e ences in p o ein exp ession (co esponding
o old change alues >1 o <−1 on log2 scale) we e conside ed signi ican o e exp ession (up egula ion) o
unde exp ession (down egula ion). Majo i y o he p o eins we e exp essed a simila le els in hESC-RPE and
hRPE. 82% o he p o eins we e simila ly exp essed in 08/017 hESC-RPE and 81% in 08/023 hESC-RPE, espec-
i ely, when compa ed o hRPE (Fig.2a and c). Only 7% o he p o eins we e up egula ed in 08/017 and 8% in
08/023 hESC-RPE cells, while 11% we e down egula ed in bo h cell lines compa ing o hRPE. Pa hway analysis
o he o e - and unde exp essed p o eins in he hESC-RPE compa ed o hRPE was conduc ed o bo h 08/017
and 08/023 o inspec i en iched ac i e pa hways a ise om he di e en ially exp essed se o p o eins. IPA
(Ingenui y Pa hway Analysis, Qiagen) co e analysis o he canonical pa hways indica ed ha p o eins associa ed
wi h mi ochond ial dys unc ion, oxida i e phospho yla ion, and pho o ansduc ion pa hways we e down egu-
la ed in hESC-RPE (Fig.2b and d). The log2 old change alues o he di e en ially exp essed p o eins o bo h
08/017 and 08/023 a e p esen ed in Supplemen a y TableS2, and isualised in Supplemen a y Fig.S1. O e all,
mos ly homogenous da a wi h simila exp ession pa e ns ac oss he mass spec ome y analyses ( h ee analyses
wi h wo echnical eplica es each) was p oduced, as shown in Supplemen a y Fig.S2. Va ia ions in he exp ession
le els be ween he h ee analyses we e hough o esul om he h ee indi idual hRPE dono samples used o
compa ison in each analysis. A compa ison o a pooled sample om he h ee hRPE dono s o a single dono
(Dono 1) was done o e alua e in e -dono a iabili y in p o ein exp ession. The esul s showed mos ly simila
exp ession wi h only 2% o he p o eins di e en ially exp essed be ween pooled dono sample and Dono 1, indi-
ca ing ha he di e ences be ween he dono s we e sub le (Supplemen a y Fig.S3).
www.na u e.com/scien i ic epo s/
3
Scien i ic RepoR s | 7: 6016 | DOI:10.1038/s41598-017-06233-9
hESC-RPE showed a no mal RPE p o ein exp ession p o ile. The PANTHER (P o ein ANalysis
Th ough E olu iona y Rela ionships) Classi ica ion Sys em and Gene On ology (GO) e m mapping we e used
o examine speci ic p o ein g oups o in e es . PANTHER classi ica ion o he exp essed p o eins acco ding o
biological p ocess, classi ied he p o eins in o 13 g oups as shown in Fig.3. The la ges g oup o p o eins was me -
abolic p ocesses which included 407 p o eins. Mos o hese p o eins (369) ell unde he subclass o p ima y me -
abolic p ocesses. Cellula p ocesses included 380 p o eins and con ained p o eins in ol ed in cell communica ion
Figu e 1. Wo k low o he s udy. (a) S udy design o he iTRAQ p o eomics. Human ESC-RPE samples om h ee
independen di e en ia ion expe imen s consis ed o h ee eplica e samples each. Human RPE samples we e
collec ed om h ee cada e ic dono eyes. To al p o ein was ex ac ed and diges ed wi h ypsin. Pep ides we e
labeled wi h iTRAQ 4-plex labels 114–117 shown in colo s, and analyzed as echnical duplica es wi h Nano-RPLC-
T iple TOF ins umen a ion. Fo inal da a analyses he esul s we e pooled oge he om all six LC-MS/MS analyses
sepa a ely o 08/017 nad 08/023 hESC-RPE. (b) A Venn diag am illus a ing he da a p ocessing be o e da a analyses
om o al numbe o p o ein hi s o 1041 p o eins accep ed o inal da a analyses.
www.na u e.com/scien i ic epo s/
4
Scien i ic RepoR s | 7: 6016 | DOI:10.1038/s41598-017-06233-9
(76), cell cycle (47), cellula componen mo emen (19), cy okinesis (11) and g ow h/p oli e a ion (1). Cellula
componen o ganisa ion and biogenesis was he hi d la ges g oup wi h 154 p o eins. Localisa ion wi h 123 p o-
eins included p o eins in ol ed in anspo . A subclass esponse o s ess, he majo subclass unde esponse o
s imulus, included 34 p o eins. Immune sys em p ocesses included 23 p o eins.
To analyse he speci ic p o ein g oups mo e ca e ully, GO e m mapping was used o iden i y p o eins con-
nec ed o speci ic GO e ms (Supplemen a y TableS3). The p o ein exp ession p o ile o hESC-RPE cells was
ound o closely esemble ha o na i e RPE. A la ge numbe o mi ochond ial p o eins we e exp essed a simila
le el (85% o 08/017 and 86% o 08/023) o sligh ly down egula ed, in he hESC-RPE cells compa ed o p ima y
hRPE (Fig.4a). All oge he 139 p o eins in ol ed in cellula s ess we e iden i ied (Fig.4b). O hese 78% we e
simila ly egula ed in 08/017 hESC-RPE and 79% in 08/023 hESC-RPE compa ed o hRPE. P o eins ha ell in o
he ca ego y cellula s ess included many p o eins in ol ed in immune esponses, such as majo his ocompa i-
bili y complex MHC class I an igen (F agmen , HLA-A) and CD59 an igen ha we e simila ly egula ed. O he
p o eins o in e es in he class included galec in-3-binding p o ein (LGALS3BP), clus e in (CLU), annexin A1
(ANXA1) and −5 (ANXA5), apolipop o eins A (APOA) and -E (APOE), c ys allin alpha A (CRYAA), c ys allin
alpha B (CRYAB), cys a in C, ca alase (CAT), galec in-1, supe oxide dismu ase 1 (SOD1) and −2 (SOD2), and
p o ein S100. A la ge numbe o cy oskele al p o eins (189) we e also mos ly simila ly egula ed in he hESC-RPE
and hRPE (83% o 08/017 and 80% o 08/023) (Fig.4c). Cy oskele al ke a ins: −8, −10, −18, and −19 could
be de ec ed and o hese −8, 10, −18 we e up egula ed in bo h hESC-RPE lines as we e ez in (EZR), ib illin-1
Figu e 2. P o ein exp ession di e ences be ween hESC-RPE and hRPE. The p opo ions o up egula ed (g een)
and down egula ed ( ed) p o eins acco ding o ela i e exp ession le els o he p o eins compa ed o hRPE o
(a) 08/017 hESC-RPE and (c) 08/023 hESC-RPE. Fo he majo i y o p o eins in bo h RPE lines, he e we e no
di e ences in exp ession le el be ween hESC-RPE cells and hRPE (g ey). Numbe o p o eins included = 1041.
De aul se ings o IPA we e used o conduc co e pa hway analysis o he o e - and unde exp essed p o eins
(188 p o eins o 08/017 and 195 p o eins o 08/023). Ten en iched pa hways wi h lowes p- alues o he
di e en ially exp essed p o eins in (b) 08/017 and (d) 08/023 hESC-RPE compa ed o hRPE. To al numbe o
genes/p o eins in each pa hway shown on op o each ba and pe cen age p esen in he da a analysed on he
y-axis (le ). Up- and down egula ion o he p o eins in he da a is shown in ed and g een colou s espec i ely,
and p- alue (o ange line) is shown in -log scale, y-axis ( igh ). Figu es in (b,c) c ea ed wi h IPA (Qiagen)
so wa e.
www.na u e.com/scien i ic epo s/
5
Scien i ic RepoR s | 7: 6016 | DOI:10.1038/s41598-017-06233-9
(FBN1), ib illin-2 (FBN2), and ilamin A (FLNA). P oli e a ion and cell cycle p o eins also showed mos ly sim-
ila egula ion in hESC-RPE and hRPE (Fig.4d).
Impo an p o eins o RPE unc ion and cha ac e is ics we e exp essed in hESC-RPE cells.
P o eins in ol ed in he pe cep ion o ligh s imulus we e iden i ied and 13 o hese 23 p o eins we e down-
egula ed in he 08/017 hESC-RPE and 11 p o eins in he 08/023 hESC-RPE compa ed o he p ima y RPE,
while he es we e simila ly exp essed (Fig.5a). Impo an RPE-speci ic isual cycle p o eins iden i ied included
o example e inal-speci ic ATP-binding casse e anspo e (ABCA4), e inoid isome ohyd olase (RPE65),
Figu e 3. Dis ibu ion o p o eins. A pie cha showing dis ibu ion o he p o eins acco ding o hei biological
p ocess using Pan he Classi ica ion Sys em. The p o eins we e classi ied by hei co esponding gene names. 852
o he 1041 p o eins we e classi ied in o one o mo e o he 13 g oups.
Figu e 4. P o eins in ol ed in mi ochond ial unc ions, s ess, cy oskele on, and p oli e a ion. Rela i e
exp ession o (a) mi ochond ial, (b) cellula s ess- ela ed, (c) cy oskele al, and (d) p oli e a ion and cell cycle
p o eins in bo h hESC-RPE cell lines compa ed o hRPE. Each do ep esen s a single p o ein, p esen ed as
mean old changes on a log2 scale, whe e y = 0 deno es equal p o ein exp ession. Red dashed lines ma k he
bo de s o o e and unde exp ession (2- old di e ence in exp ession). Fo p o ein names and exac exp ession
alues in each class see Supplemen a y TablesS4–S7.

www.na u e.com/scien i ic epo s/
6
Scien i ic RepoR s | 7: 6016 | DOI:10.1038/s41598-017-06233-9
e inaldehyde-binding p o ein 1 (RLBP1), and e inol dehyd ogenase 11 (RDH11). The e we e 46 p o eins
in ol ed in ex acellula ma ix (ECM) and cell adhesion. Among hose we e in eg in subuni s αV (ITGAV)
and h ee be a chains −1 (ITGB1), −4 (ITGB4) and −8 (ITGB8), o which he in eg in ß-4 was up egula ed in
bo h hESC-RPE lines. In eg in binding alin-1 (TLN1) was sligh ly up egula ed as was in e cellula adhesion
molecule 1 (ICAM-1) and galec in-3 binding p o ein (LGALS3BP). Neu al cell adhesion molecule 1, (NCAM-
1) was down egula ed in he hESC-RPE, while i onec in (VTN) was among he simila ly exp essed p o eins
(Fig.5b). P o eins in ol ed in melanogenesis included pigmen epi helium-de i ed ac o (SERPINF1) ha was
simila ly exp essed in hESC-RPE and hRPE, while y osinase- ela ed p o ein 1 (TYRP1), was up egula ed in
08/017 hESC-RPE. A o al o 98 p o eins in ol ed in phagocy osis, p o ein deg ada ion and au ophagy we e
iden i ied. These included p o easome subuni s and 26 S p o easome p o eins. In e es ing p o eins in ol ed
in phagocy osis included in eg in alpha V, pa o he αVβ5 in eg in, and i s ligands e aspanin (CD81, ag-
men ), and lac adhe in (MFG-E8), and hese p o eins we e exp essed a simila le el as compa ed o hRPE. Also,
CD44, CD36 (sca enge ecep o class B, membe 2), and ca hepsin-D we e exp essed a simila le el in he
hESC-RPE and hRPE. T anspo e s and ion channels (90 p o eins in o al) exp essed included chlo ide in acel-
lula channel p o eins 1 and −6 (CLIC1, and CLIC6). Na+K+- anspo ing ATPase subuni s alpha-1 and be a-1
(ATP1A1, ATP1B1) we e simila ly egula ed. Aquapo in 1 ( agmen ) (AQP1) was up egula ed in hESC-RPE.
Vol age-dependen anion-selec i e channel p o eins 1 (VDAC1) and −2 (VDAC2) we e down egula ed in
hESC-RPE while VDAC3 was exp essed a simila le el. Amino acid anspo e (SLC1A3) was down egula ed
in hESC-RPE, while la ge neu al amino acids anspo e small subuni 1 (SLC7A5) was up egula ed (Fig.5b).
Discussion
Ul ima ely, p o eins exp essed by a cell de e mine he pheno ypic exp ession o genomic in o ma ion29. The p o-
eome is a dynamic en i y cha ac e izing he cell ype and changing wi h he physiological s a us o he cell.
Pionee ing p o eomic p o iling s udies ha e e ealed he p o ein exp essions in human RPE/cho oid complex
Figu e 5. P o eins in ol ed in impo an RPE unc ions. (a) Rela i e mean exp ession le els o he 23 p o eins
in ol ed in pe cep ion o ligh s imulus and (b) o selec ed impo an RPE p o eins. P o ein exp ession le els
shown on log2 scale. Red dashed lines ma k he bo de s o o e exp ession and unde exp ession (o e 2- old
di e ence in exp ession). E o ba s deno e s anda d de ia ion. Fo he ull p o ein names and exp ession alues
see Supplemen a y TableS8. *Quan i ica ion based on less han wo pep ides.
www.na u e.com/scien i ic epo s/
7
Scien i ic RepoR s | 7: 6016 | DOI:10.1038/s41598-017-06233-9
as well as in i o cul u ed RPE cells7, 25, 30–32, changes in RPE p o eome associa ed wi h he p og essi e s ages o
AMD23, 24, 33, 34, and in esponse o oxida i e s ess35. The p o ein con en o d usen and lipo uscin has been elu-
cida ed wi h p o eomic app oaches36. To ou bes knowledge, la ge scale analyses o hPSC-RPE p o eome ha e
no been p e iously pe o med. The objec i e o his s udy was o compa e he p o ein exp ession p o ile o RPE
cells di e en ia ed om hESCs o he na i e p ima y hRPE p o eome. We chose a sample se up wi h a di e en
biological eplica e o bo h hESC-RPE and hRPE included in each analysis, in o de o obse e ue biological
di e ences be ween he samples. In he quan i a i e 4-plex iTRAQ analyses, 1041 common p o eins we e p esen
in bo h hESC-RPE cell lines and p ima y hRPE samples. Impo an ly, majo i y o hese p o eins (>80%) we e
simila ly egula ed in he hESC-RPE and na i e hRPE, showing ha he hESC-RPE cells esembled hei na i e
coun e pa s. The exp ession pa e ns o he wo hESC-RPE lines we e also simila , showing ha ou di e en ia-
ion p o ocol yielded homogenous RPE cells wi h low cell line speci ic a ia ion in p o ein exp ession.
The p o ein p o ile o he hESC-RPE cells esembled ha o na i e p ima y hRPE. A la ge numbe o me a-
bolic and espi a o y p o eins we e exp essed e lec ing he highly me abolically ac i e na u e o he RPE. The
pa hway analysis o he di e en ially exp essed p o eins indica ed mi ochond ial dys unc ion and down egula ed
oxida i e phospho yla ion in he hESC-RPE, e lec ing educed ene gy equi emen s in he in i o cul u ed
hESC-RPE cells compa ed o na i e cells. This is consis en wi h some p e ious published esul s on he p o-
eome o in i o cul u ed RPE cells7. Apa om single up egula ed p o eins, no signs o inc eased s ess, apop-
osis, immune esponse, p oli e a ion, o e inal degene a ion- ela ed changes we e obse ed in he hESC-RPE.
Cellula s ess ma ke s and many p o eins conside ed as bioma ke s o AMD we e simila ly egula ed:
galec in-3-binding p o ein (LGALS3BP) which was sligh ly up egula ed in hESC-RPE has been implica ed as
AMD ma ke , bu also in ol ed in a achmen and sp eading o RPE cells37. Clus e in (CLU) and apolipop o ein
E (APOE) a e majo cons i uen s o d usen36, and we e simila ly exp essed in hESC-RPE and hRPE. C ys allin
alpha A (CRYAA), down egula ed in hESC-RPE and c ys allin alpha B (CRYAB) ha was simila ly egula ed, a e
p o ec i e o oxida i e inju y and bo h d usen cons i uen s as well as AMD bioma ke s34, 38. Cys a in-C (CST3)
is abundan ly sec e ed by he RPE39, while no much is known abou he unc ion o cys a in-B (CSTB) in he
RPE. Annexins (ANXA1, ANXA5) in ol ed in many p ocesses including in lamma ion and apop osis40, we e
also simila ly egula ed as we e me abolic p o eins such as p ohibi in (PHB), supe oxide dismu ases 1 and −2
(SOD1, SOD2), and glu a hione pe oxidase (GPX1) ha ha e been implica ed as oxida i e s ess bioma ke s41–43.
While in lamma ion- ela ed p o ein in e cellula adhesion molecule 1 (ICAM-1) was up egula ed, MHC class
I an igen ( agmen , HLA-A) was simila ly egula ed in bo h hESC-RPE lines and MHC class II an igen was
no de ec ed. Ou unpublished esul s, in acco dance wi h published da a by o he s44, show ha in he s anda d
cul u e condi ions, ou hESC-RPE cells exp ess MHC class I bu no MHC class II an igens wi hou exposu e o
p oin lamma o y cy okines. The memb ane complemen egula o y p o ein CD59 was also simila ly egula ed.
P o ein S100, and mac ophage mig a ion inhibi o y ac o (MIF) ha e p e iously been shown o pa o macula
and pe iphe al RPE p o eome25.
As highly di e en ia ed monolaye o pola ized cells, he RPE cells main ain hemsel es ou side he cell cycle.
Du ing in i o cell cul u e and ce ain pa hologies, RPE cells unde go dedi e en ia ion (epi helial- o-mesenchyme
ansi ion, EMT) o become ac i ely p oli e a ing cells acqui ing a ib oblas ic mo phology. Dedi e en ia ed RPE
cells exp ess cell cycle ma ke s and cy oke a ins such as cy oke a in −8, −18, and −1945. Cy okele al c osslinke s
ac in binding ilamin-A (FLNA) and ez in (EZR), which is localised o apical mic o illi in he RPE, we e up eg-
ula ed in he hESC-RPE. Ez in has a ole in POS phagocy osis bu i s o e exp ession has also been linked EMT46.
Fib illins −1 and −2 a e s uc u al p o eins o he ECM mic o ib ils, in ol ed in a a ie y o unc ions including
eye de elopmen and mo phogenesis47. O e exp ession o he cy oskele al emodelle s and cy oke a ins -8 and -18
in bo h hESC-RPE lines and cy oke a in-19 in 08/023 hESC-RPE, mos p obably e lec s he in i o na u e o he
cells. The in i o hPSC-RPE di e en ia ion in ol es se ial passaging a e which he cells p oli e a e o colonize
he cell cul u e su ace and hen egain hei epi helial pheno ype. The hRPE cells used o compa ison in his
s udy we e no in i o cul u ed o a oid he cul u e induced changes o hei cellula pheno ype and p o eome. No
signs o ans e o mo e ac i e s a us o cell cycle o p oli e a ion was de ec ed in he hESC-RPE cells.
Impo an ly, highly RPE-speci ic p o eins such as he isual cycle p o eins we e iden i ied in he hESC-RPE.
The RPE laye is he p inciple si e o 11-cis e inal egene a ion in he isual cycle. RLBP1 and RPE65 a e impo -
an biochemical RPE ma ke s p esen in hESC-RPE and hiPSC-RPE18. The down egula ion o exp ession o
hese p o eins is expec ed o he in i o di e en ia ed and cul u ed hESC-RPE as compa ed o na i e issue,
since he hESC-RPE cells ha e no been in con ac wi h he neu al e ina o e inoids. I has been p e iously
shown ha hiPSC-RPE cells possess he machine y o p ocess e inoids and can exhibi a unc ional isual cycle
in i o and in i o. Human iPSC-RPE cells main ain RPE65 p o ein exp ession du ing cul u e in con as o
p ima y cells and a e incuba ion wi h all- ans e inol, hiPSC-RPE elease 11-cis e inaldehyde in o he cul u e
media48.
ECM and cell adhesion- ela ed p o eins iden i ied included in eg in subuni s αV and h ee be a chains: 1, 4,
and 8. In eg in ß-4, pa o he p ima y laminin ecep o in eg in α6ß4, was up egula ed in bo h hESC-RPE lines.
Laminin is an impo an componen o B uch’s memb ane and he epi helial cell basemen memb ane. In eg in
α6ß4 p omo es a achmen o RPE o B uch’s memb ane49 and is hus impo an o in eg a ion o hESC-RPE
cell ansplan s. Abso p ion o s ay ligh by he abundan melanin g anules is ano he impo an unc ion o
he RPE. P o eins in ol ed in melanin biosyn hesis such as y osinase- ela ed p o ein 1 (TYRP1) was de ec ed
as being up egula ed in he hESC-RPE. High pigmen a ion s a us and sec e ion o e ina p o ec i e pigmen
epi helium-de i ed ac o , PEDF (SERPINF1) by ou hESC-RPE cells has been shown p e iously18.
RPE cells a e among he mos ac i e phagocy es in he body. Ci cadian shedding o POS ac i a es a synch o-
nized phagocy ic esponse by RPE cells e e y mo ning50. Many phagocy osis ela ed p o eins we e de ec ed in
hESC-RPE. The in eg in α β5 is in ac equi ed o POS binding o he RPE51. In eg in α β5 complexes wi h
he e aspanin CD81, an in e ac ion ha p omo es pa icle binding. Engul men ecep o CD36 a he apical,
www.na u e.com/scien i ic epo s/
8
Scien i ic RepoR s | 7: 6016 | DOI:10.1038/s41598-017-06233-9
phagocy ic su ace o RPE cells and lac adhe in (Milk a globule-EGF ac o 8) a e also in ol ed in POS ecog-
ni ion. Ca hepsin-D (CTSD) is a majo p o eoly ic enzyme in phagocy ic cells ha diges s opsin. Ez in is known
o bind o lysosomal-associa ed memb ane p o ein 1 (LAMP-1) and CD44 du ing he o ma ion o phagocy ic
acuoles50, 52.
The RPE also unc ions as a ba ie be ween he neu al e ina and he enes a ed cho oid pe o ming a a ie y
o ec o ial anspo unc ions (wa e , ions, me aboli es, nu ien s and was e p oduc s) be ween he wo compa -
men s1. Impo an anspo e s and ion channels we e iden i ied in he hESC-RPE such as Na+K+- anspo ing
ATPase subuni s alpha-1, and be a-1 and −2 (ATP1A1, ATP1B1) localized a he apical su ace o he RPE as
shown p e iously o ou hESC-RPE18. Aquapo in 1 ( agmen ) (AQP1), channel p o ein ha media es wa e
anspo , is apically exp essed bo h in human oe al RPE p ima y cul u es and in hESC-RPE cells19. P o eomic
analyses o he plasma memb ane a e subcellula ac iona ion could p o ide u he insides in o exp ession o
anspo e and channel p o eins by he hPSCs.
A speci ic no mal RPE gene exp ession signa u e o 154 genes has been p oposed o alida ion o RPE-like
cells de i ed om hPSC53. In ha s udy, only eigh p o eins we e analysed a p o ein le el. The changes a he
mRNA le el o en do no lead o co esponding changes a he p o ein le el and la ge scale, p o ein le el s udies o
he hPSC-RPE a e he e o e also impo an o alida ion. The iTRAQ me hod is limi ed o he de ec ion o only
hose p o eins ha a e p esen in all samples o each iTRAQ labelling. Thus p o eins exp essed in he hESC-RPE
only, o solely in he p ima y hRPE sample could no be de ec ed wi h his me hod. In addi ion, a 30 kDa il e
was used in he expe imen al se up o educe noise and ion supp ession caused by he small molecules in he
samples. P o eins unde he limi such as leci hin e inol acyl ans e ase (LRAT), a 25 kDa isual cycle p o ein
and an impo an RPE ma ke , could no be de ec ed. In he iTRAQ, quan i ica ion o low abundancy p o eins can
cause miss quan i ica ion a ising om backg ound ions54, and he e o e ca e ul da a checking is impo an espe-
cially o p o eins wi h ex eme old change alues. Da a dependen analysis (DDA) mode is equi ed o iTRAQ
label in ensi y based quan i ica ion and causes missing alue gene a ion in he da a as numbe o eplica e uns
inc ease. This is a well-documen ed phenomenon in ITRAQ analyses55 and was also obse ed in ou da a wi h
661 p o eins de ec ed in only one analysis un i.e. one biological eplica e (Supplemen a y Table S1, p. 12–19).
Despi e o hese limi a ions o he me hod, iTRAQ enables mul iplexing, ha is compa ing ou di e en samples
in a single mass spec ome y expe imen . Highly eliable compa a i e, quan i a i e da a can hus be p oduced
especially wi h samples ha ha e simila p o ein exp ession pa e ns56. Al hough he 1041 p o eins de ec ed only
p esen s a ac ion o he o al RPE p o eome o o e 4000 indi idual p o eins de ec ed om RPE-cho oid sam-
ples in o he s udies25, he da a p esen ed p o ides new, impo an , and quan i a i e in o ma ion on exp ession o
o e 1000 common p o eins be ween hESC-RPE and hRPE. The idea o his s udy was o quan i a i ely compa e
he p o eome o hESC-RPE o hRPE, no o accumula e he whole hESC-RPE p o eome, which would equi e
a much la ge sample size and a di e en MS me hodology. New me hods like sequen ial window acquisi ion
o all heo e ical agmen ion spec a (SWATH), migh enable analysis up o 4000 p o eins om RPE samples.
Building a whole p o eome lib a y o he hPSC-RPE is a a ge o ou u u e s udies.
To conclude, we we e able o de ec 1041 common p o eins be ween he wo hESC-RPE samples and p ima y
hRPE and demons a ed ha he hESC-RPE di e en ia ion yields high quali y RPE cells. The hESC-RPE cells di -
e en ia ed in i o, in se um- ee condi ions and in he absence o con ac wi h hei na u al niche, s ill esembled
he hRPE exp essing RPE speci ic unc ional p o ein machine y. No p o ein abno mali ies indica i e o pa hological
beha iou o he hESC-RPE was de ec ed which is impo an conside ing he u u e clinical applica ions. P o eomic
app oaches enabling absolu e, label- ee quan i ica ion such as SWATH echnique could be used in he u u e o
e en mo e p ecisely compa e he hPSC-RPE p o eome o na i e issue. As he hPSC-RPE also p o ide unlimi ed
oppo uni ies o model e inal degene a i e diseases in i o, quan i a i e, compa a i e p o eomics opens possibil-
i ies o unde s and p o ein le el disease mechanisms and disco e new he apeu ic a ge s o de as a ing diseases
such as AMD, and also aid in alida ion o hPSC-RPE o clinical applica ions.
Me hods
Human emb yonic s em cell-de i ed RPE cells. The wo hESC lines p e iously de i ed a ou labo-
a o y a Uni e si y o Tampe e we e used in his s udy: Regea08/017 (46;XX) and Regea08/023 (46;XY)57. The
undi e en ia ed hESCs we e cul u e in hESC cul u e medium consis ing o KnockOu DMEM supplemen ed
wi h 20% KnockOu Se um Replacemen (ko-SR, The mo Fishe Scien i ic), 2 mM Glu amax, 0.1 mM 2-me cap-
oe hanol, 1% Non-essen ial amino acids, 50 U/ml penicillin/s ep omycin, and 8 ng/ml human basic ib oblas
g ow h ac o (bFGF) on inac i a ed human o eskin ib oblas (CRL-2429; Ame ican Type Cul u e Collec ion)
eede cells. Undi e en ia ed s em cell colonies we e passaged wi h T ypLE Selec on o esh eede cell laye
e e y 10 days. Bo h cell lines a e ou inely ka yo yped and cha ac e ized o hei sel - enewal and di e en ia ion
capaci ies as well as absence o mycoplasma.
The hESCs we e di e en ia ed o RPE as p e iously desc ibed18. B ie ly, hESCs we e manually cu o suspen-
sion and cul u ed as loa ing cell agg ega es in RPE basic medium consis ing o same eagen s as hESC medium
excep o 15% ko-SR and no bFGF. Fo en ichmen , pigmen ed cells we e isola ed and seeded o human placen al
collagen IV (5 µg/cm2). Fo u he ma u a ion he cells we e seeded o collagen IV (5 µg/cm2) coa ed PET cell
cul u e inse s (Millipo e) and cul u ed o 129 days (Regea08/017) o 139 days (Regea08/023). Regea08/017 a
passage le els 33, 34 and 70, and Regea08/023 a passage le els 44, 61, and 70 we e used o gene a ing RPE cells.
Cells om h ee independen di e en ia ion expe imen s ( eplica es 1–3) we e used o he s udy. Each biological
eplica e consis ed o h ee cell cul u e inse s. The hESC-RPE cells we e collec ed wi h T ypLE Selec , washed
wice wi h PBS, pelle ed by cen i uga ion and s o ed as d y pelle s a −80 °C un il p o ein ex ac ion.
Human RPE cells. P ima y hRPE om h ee di e en dono s we e used. The dono eyes had no mac-
oscopical o mic oscopical signs o e inal degene a ion. Fo human e inal pigmen epi helium isola ion,
www.na u e.com/scien i ic epo s/
9
Scien i ic RepoR s | 7: 6016 | DOI:10.1038/s41598-017-06233-9
ollowing enuclea ion, he eye bulb was washed wi h 5% po idone iodine (Be adine), he an e io segmen was
cu app oxima ely 3.5 mm pos e io o he limbus and he an e io segmen , he i eous and neu al e ina we e
emo ed. The eyecup was hen illed wi h RPE cell cul u e medium (DMEM/F12, supplemen ed wi h 10% FCS,
1% an ibio ics-an imyco ics and 1% L-Glu amin) and he cells collec ed by a special glass cell sc ape o gen ly
emo e he RPE om B uch’s memb ane while con inuously ba hed in hei own cell cul u e medium. The RPE
was collec ed as a pelle by cen i uga ion (1000 pm, 10 min). The RPE pelle was washed wice wi h PBS, ol-
lowed by cen i uga ion a e each wash, and shipped o u he analysis a −20 °C.
P o ein ex ac ion. Human ECS-RPE pelle s we e suspended in 200 µl and hRPE pelle s in 400 µl o RIPA
lysis bu e , supplemen ed wi h 0.1% HALT p o ein inhibi o cock ail, and dissocia ed using a VWR disposable
pes le. The samples we e hen o exed o 2 min, incuba ed in an ul asonic ba h o 5 min, and on ice o 25 min.
Dissocia ed samples we e cen i uged a 14 800 pm o 15 min a 4 °C, supe na an s ans e ed o clean ubes and
hei p o ein concen a ions de e mined using he DC P o ein Assay Ki II (Bio-Rad Labo a o ies) acco ding o
he manu ac u e ’s ins uc ions. Bo ine se um albumin (BSA; Bio-Rad Labo a o ies) was used as a s anda d, and
all samples analysed in duplica es. 15 µg o o al p o ein/hESC-RPE inse we e used and pooled oge he o each
biological hESC-RPE eplica e (1–3) o bo h cell lines (45 µg in o al). Equal amoun (45 µg) was used o each
hRPE dono sample (1–3) (Fig.1). Addi ional pooled hRPE sample was p epa ed by pooling 15 µg o p o ein om
each hRPE dono (45 µg o p o ein in o al). P o eins we e p ecipi a ed by adding six olumes o cold ace one,
ollowed by quick o exing and o e nigh incuba ion a −20 °C. The ollowing day, samples we e cen i uged a
14 800 pm o 15 min a 4 °C. The ace one was disca ded and e apo a ed o 10 min, and he pelle s subjec ed o
p o ein diges ion.
P o ein diges ion, iTRAQ labelling, and mass spec ome y analysis. P o ein diges ion, label-
ling and mass spec ome y was done as desc ibed in de ail p e iously28. 30 kDa size exclusion il e s (Pall
Co po a ion) we e used in sample p epa a ion. Samples we e labelled wi h he isoba ic iTRAQ ags as shown in
Fig.1 and pooled oge he o he mass spec ome y analysis. Labelled samples we e analysed by Nano-RPLC-
T ipleTOF ins umen a ion using Eksigen ekspe ™ 425 NanoLC coupled o high speed T ipleTOF™ mass
spec ome e (Sciex 5600+). A capilla y RP-LC column (cHiPLC® Ch omXP C18-CL, 3 µm pa icle size, 120 Å,
75 µm i.d × 15 cm; Eksigen , Conco d, Canada) was used o LC sepa a ion o pep ides.
Da a p ocessing. P o ein pilo so wa e e sion 4.0.8085 (Ab Sciex) was used o analyse MS/MS da a
sea ched agains he UniP o /Swiss-P o p o ein da abase o p o ein iden i ica ion. Some impo an se ings
in he Pa agon sea ch algo i hm in p o ein pilo we e con igu ed as ollows. Sample ype: iTRAQ 4plex (pep ide
labelled), Cys-alkyla ion: MMTS, Diges ion: T ypsin, Ins umen : T ipleTOF 5600+, Sea ch e o : ho ough ID.
False disco e y a e (FDR) analysis was pe o med in he P o ein pilo and FDR <1% was se o p o ein iden i i-
ca ion. Only pep ides wi h 99% con i ma ion we e included in iden i ica ion and quan i ica ion. Sha ed pep ides
we e excluded om quan i ica ion ha was pe o med using P o einPilo so wa e wi h s aigh a e age o pep-
ides (p o eins wi h dis inc i e pep ides, n = 1964, Fig.1b). Da a p ocessing included log2- ans o ma ion and
cen al endency no maliza ion. G aphical p esen a ion o he da a no maliza ion is shown in Supplemen a y
Fig.S4. Ranking o p o eins (measu e o quali y) was c ea ed by aking he mean o he ankings in di e en
uns. Duplica e p o eins wi h he same gene names we e me ged so ha only one p o ein name was kep . I
bo h SwissP o and T embl duplica e p o eins we e p esen , SwissP o was chosen (numbe o p o eins a e
duplica e check, n = 1729, Fig.1b). The spec al peaks o p o eins wi h o e 4- old di e ence (log2 old change
alue >2 o <−2) in exp ession in ei he hESC-RPE line compa ing o hRPE we e manually checked, and mis-
quan i ied p o eins we e omi ed. These included pho o ansduc ion ela ed p o eins such as opsin, eco e in,
hodopsin, od ou e segmen memb ane p o ein 1, and pe iphe in-2 ha we e p esen in high abundance in he
human dono samples bu misquan i ied due o noise ions in he hESC-RPE. The manual checking con i med
ha hese pho o ecep o speci ic p o eins we e no p esen in he hESC-RPE (da a no shown). The da a was
il e ed so, ha only p o eins which we e iden i ied in a leas wo biological eplica e samples o bo h hESC-RPE
lines and human RPE (i.e. in wo o he h ee analysis uns), we e conside ed eliable and included in da a anal-
yses. P o eins based on de ec ion in only one biological eplica es (one analysis un, n = 661) we e il e ed ou .
Finally, p o eins wi h ei he echnical o biological CV >100 (n = 27), we e excluded om he inal da a analysis.
The p o eins il e ed ou a e shown in Supplemen a y TableS1. P- alues, which a e shown in he Supplemen a y
TableS1 and Supplemen a y Fig.S1, we e ob ained using a one sample - es , whe e he means o echnical ep-
lica es we e aken p io o es ing. Unadjus ed p- alues we e p esen ed (as no signi ican changes we e seen o
08/017 hESC-RPE o hRPE, and only se en p o eins (Unip o IDs: O15537, B7ZB41, J3KPF3, Q6FHV6, Q76LA1,
V9HWH2, X5DNM4) eached signi icance in old change o 08/023 wi h adjus ed he p- alues).
Classi ica ion in o speci ic g oups o in e es was done by GO mapping wi h he GO e ms connec ed o each
p o ein’s UniP o accession using UniP o .ws package in R (Ma c Ca lson. UniP o .ws: R In e ace o UniP o
Web Se ices. R package e sion 2.10.2.). The GO e ms o in e es as well as hei o sp ing e ms, i.e. mo e
speci ic e ms, we e included in he GO e m mapping p ocess due o he loose hie a chical s uc u e o GO
anno a ion. The GO e ms used a e lis ed in Supplemen a y TableS3. P ocessing o he da a, GO mapping, box-
plo s, and olcano plo s o isualiza ion we e c ea ed using R so wa e e sion 3.2.3 (R Co e Team, Founda ion
o S a is ical Compu ing, Vienna, Aus ia, h ps://www.Rp ojec .o g/). QIAGEN’s Ingenui y® Pa hway Analysis
(IPA®, QIAGEN Redwood Ci y, www.qiagen.com/ingenui y) was used o conduc canonical pa hway analyses
using he di e en ially exp essed (log2 old change >1 o <−1) p o eins in 08/017 and 08/023 hESC-RPE com-
pa ed o hRPE. The PANTHER Classi ica ion Sys em Ve sion 11.0 ( elease da e July 15, 2016) was used o classi y
he p o eins acco ding o hei in ol emen in biological p ocesses (h p://pan he db.o g/). Cha s we e c ea ed
in Mic oso Excel 2013.