/

University of Southern California Dissertations and Theses

/

Molecular basis of mouse epiblast stem cell and human embryonic stem cell self‐renewal

(USC Thesis Other)

Molecular basis of mouse epiblast stem cell and human embryonic stem cell self‐renewal

PDF

Transcript (if available)

Content
Molecular Basis of Mouse Epiblast Stem Cell and
Human Embryonic Stem Cell Self ‐renewal

By
Xing liang Zhou

A Dissertation Presented to the
FAC U LTY O F TH E GR A DUAT E SC H OOL of
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree of
DOCTOR OF PHILOS OP HY
(GENETIC , MOLECU LAR, A ND C ELLULAR BIO L OGY)

May 2017

Copyright 2017 Xingliang Zhou
ii

Dedication

To my parents for their unconditional support and faith in me

iii

Table of Contents
Chapter 1 Intr oduction ........................................ ............................................................... ................................................... 1
T he developmental origin of pluripot ent stem cells ............ ............................................................... ................. 1
Naïve and primed sta te pluripotency .......................................................................................... ............................... 3
Molecular mechanisms of primed s tat e pluripotency maintenance i n vitro ............................................. 8
The roles of β‐catenin in regula ti ng primed stat e pl uripotency ............................................................... .... 10
The roles of Hippo signaling pat hway in pluripotency regulati o n .............................................................. . 15
The search of novel c u lture cond ition s for pluripo tent ste m c e lls ............................................................ .. 19
Chapter 2 Cytoplasmic β‐ catenin regulates primed state pluripotency ......................................................... 25
CHIR/IWR‐1 media te m o u se EpiSC s el f‐renew a l through β‐catenin . ......................................................... 25
Cytoplasmic retentio n o f β ‐catenin is sufficient to m aint ain pr imed state PSC sel f‐renewal ........... 27
Subcellular l o calization is cruci al for β ‐catenin i n mediatin g self‐renewal ............................................. 36
TGFβ/Sm a d and FG F/MA PK sign a lin g p athwa y s ar e not the direct ta r g ets of CHIR/IWR‐1 in
mouse EpiSCs .................................................. ............................................................... .................................................... 42
Identific a tion of cytopl asmic β‐cateni n bindin g p a rtners under CHIR/IWR‐1 condition .................. 44
Chapter 3 The roles of YAP/TAZ i n pri m ed stat e pl u ripotency mai ntenance.............................................. 47
T AZ , but not Y AP, is a b inding p a r tne r of cytopla s mic β‐ca te nin in m o use Epi S Cs under
CHIR/IWR‐1 c ondition ......................................................................................................... .......................................... 47
Overexpr ession o f T A Z in d uces mous e EpiSC differ e ntiatio n in a β‐catenin i n dependent man n er
.............................................................. ............................................................... ............................................................... ...... 48
Nuclear translocation of TAZ induc es mouse EpiSC differ e ntiati o n ........................................................... 5 2
Overexpr ession and nucl ear transl ocation o f YAP i nduce mous e Ep iSC differentiation .................... 54
Reducing T A Z le v el in m o use EpiSCs s hows no ob v i ous phen oty p e . .......................................................... 56
T AZ/YAP are dispensable for mouse ESC self‐renewal ............ ............................................................... ......... 57
TAZ/ YA P ar e req uired fo r the prop er c onversion o f m o use ESC s to EpiSCs .......................................... 61
Cytoplasmic retentio n o f TAZ and/or Y AP prom ot e s m ouse E piSC s e lf‐ren e w al in th e ab sence o f
nuclear β‐catenin ............................................. ............................................................... .................................................. 66
T AZ , but not Y AP, is a b inding p a r tne r of cytopla s mic β‐ca te nin in hu man ES Cs under CH I R/IWR‐
1 condition .................................................................................................................. ......................................................... 73
Nuclear translocation of TAZ/ Y AP i n d uces hum a n ESC differe ntiat ion .................................................... 74
Cytoplasmic retentio n o f TAZ or YAP promotes hu m an ESC self‐ren ewal i n t h e absenc e o f nucle ar
β‐catenin ..................................................... ............................................................... ........................................................... 75
Chapter 4 Screen for novel human ESC culture conditions ...................................................................... ........... 78
CHIR/IWR‐1 plus ActA or bFGF ar e re q uired for se rum‐fre e cul tur e of mouse EpiSCs ..................... 78
iv

Mouse EpiSC s are se n sitiv e to ch emic al inhibition of TGFβ/Smad and FG F/ MAPK sign a ling u nder
CHIR/IWR‐1 c ondition ......................................................................................................... .......................................... 83
CHIR/IWR‐1 plus bFG F a nd/or ActA a re optim al f o r serum‐fr e e cul ture of human ESCs ................. 84
N2B27 is a c urrently a vailable optimal medium for serum‐free c u lture of human ESCs ................... 86
Small‐scale s creening of c hemica ls th a t boost human ESC self‐re newal .................................................. 91
Small molec u le activat or s of LIF/STA T3 sign a ling show syner g ist ic effect w i th LI F in m o use a n d
human ESCs .................................................... ............................................................... ...................................................... 95
Small molec u le activat or s of TGFβ/S m ad signalin g pro motes mouse EpiSCs and human ESCs .. 103
Chapter 5 Discussion and perspectives ......................... ............................................................... ............................. 110
Current m od el for CHIR/I W R‐1‐media t ed mo u se E piSC and h u m an ESC self‐renewal .................. 110
The role of c y toplasmic β ‐ cateni n in m ouse EpiSC and h um an E SC self‐renewal .............................. 112
Pleiotropic r o le o f TAZ a n d YAP in m ouse E piSC a nd hum an E SC se lf‐renewal .................................. 114
Dispensable roles of T AZ and YAP in mouse ESC self‐renewal .... .............................................................. 116
Insights from small scale chem ical screening on human ESCs .... ............................................................... 117
Conclusions and p erspectives .................................. ............................................................... ................................. 119
Materials an d Methods ........................................................................................................ ............................................. 122
References .................................................... ............................................................... .......................................................... 132

v

List of Figures and Tables
Figure 1 De v elopme ntal potency of c e lls in early mouse embr y os. ........................................................ 2
Figure 2 Sign a ling p athw a y s invol v ed i n the m a inte nance o f n aï v e state PS Cs in vitro ........................ 6
Figure 3 Con version betw een ESCs an d EpiSCs in c u lture....................................................................... 7
Figure 4 Sign a ling p athw a y s invol v ed i n the m a inte nance o f pri m ed stat e PS C s in vitro .................. 10
Figure 5 Th e cano n ical W nt sign a ling p athwa y . ..................................................................................... 12
Figure 6 Mo d ulation o f β ‐catenin m ai ntains pri me d state PSC s e l f‐ren e wal ....................................... 14
Figure 7 Th e core co mpo n ents o f ma mm alian Hip p o sign aling pathw ay ............................................ 16
Figure 8 Hippo sign a ling d irects the formatio n o f t r ophectoder m in early m ouse e mbr y o ............... 18
Figure 9 Los s of β‐cat eni n in Ct nn 1
‐/‐
mouse EpiSCs. ............................................................................. 26
Figure 10 β‐ catenin is required for moue EpiSC self‐renew a l und er CHIR/IWR‐1 condit ion. ........... 26
Figure 11 Subcellular dist ribution of Δ Nβ‐catenin‐ E R
T2
is contr o lled by 4‐O HT. ................................ 28
Figure 12 ΔNβ‐catenin‐ ER
T2
activ ate s Wnt tar get gen e s in res ponse to 4‐O HT. ................................. 29
Figure 13 Cytoplasmic rete ntion of ΔNβ‐catenin‐ E R
T2
mai ntai ns 46 C m ous e EpiSC self‐ renewal. ... 30
Figure 14 Expression of Δ Nβ‐catenin‐ E R
T2
in Ct nn b1
‐/‐
mouse EpiSCs. ............................................... 31
Figure 1 5 Δ Nβ‐catenin‐ E R
T2
activ ate s Wnt tar get gen e s in res ponse to 4‐O HT in Ctnn b 1
‐/‐
EpiSCs. 32
Figure 16 Cytoplasmic rete ntion of ΔNβ‐catenin‐ E R
T2
mai ntai ns Ctnn b 1
‐/‐
mouse EpiSC self‐
renewal. ...................................................................................................................................................... 33
Figure 17 Subcellular dist ribution of Δ Nβ‐catenin‐ E R
T2
is contr o lled by 4‐O HT in HE S2 hum an E SCs.
.................................................................................................................................................................... 34
Figure 18 Cytoplasmic rete ntion of ΔNβ‐catenin‐ E R
T2
mai ntai ns HES2 hu man ESC self‐renewal. ... 35
Figure 19 Bl ocking the i n t eracti on bet ween β‐cate n in a nd TCF f a mily transcription factors reduced
the transacti v ation acti vity of β‐catenin. ................................................................................................. 37
Figure 2 0 C y toplasmic tr anslocatio n of ΔNβ‐cat en in‐ER
T2
mut ants induces differentiati on in
Ctnnb 1
‐/‐
m o u se EpiSCs. ............................................................................................................................ 38
Figure 2 1 C y toplasmic tr anslocatio n of ΔNβ‐cat en in‐ER
T2
mut ants induces differentiati on in HES2
human ESCs. ............................................................................................................................................... 39
Figure 2 2 M e mbr a ne‐ b o u nd β‐cateni n is not req u i red for Mou s e Ep iSC s e lf‐renewal u n d er
CHIR/IWR‐1 condition. ............................................................................................................................. 41
Figure 23 TG Fβ/Smad an d FGF/MAP K pathw ays are n o t th e d i rect ta rgets of CH I R/IW R‐1. ........... 43
Figure 2 4 C a pture o f cyt o p lasmic β‐ca t enin i nterac ting p artners by c o ‐IP. ......................................... 45
Figure 2 5 TA Z is a β‐cate n in bi n ding partner i n mouse EpiSCs un der CHIR/ I WR‐1 condi t ion. ......... 48
Figure 2 6 TA Z o v ere x pres sion in duces differentiati on in 46C mouse EpiSCs under CHIR/ I WR‐1
condition. ................................................................................................................................................... 49
Figure 2 7 Ov erexpressio n of TA Z activates Hippo t arget gen e s i n 4 6C m ouse EpiSCs. ...................... 49
Figure 2 8 TA Z o v ere x pres sion in duces differentiati on in 46C mou se EpiSCs under ActA/ b FGF
condition. ................................................................................................................................................... 51
Figure 2 9 Ov erexpressio n of TA Z activates Hippo t arget gen e s i n Ctnn b 1
‐/‐
mouse EpiSCs. .............. 51
Figure 3 0 TA Z o v ere x pres sion in duces differentiati on in Ctnn b 1
‐/‐
mouse Epi SCs. ............................ 52
Figure 31 Expression of TAZ‐ER
T2
fusi on prot ein in CD1 mouse EpiSCs. ............................................. 53
Figure 32 TA Z‐ER
T2
acti v a t es Hippo t arget ge nes i n r esponse to 4 ‐OHT in CD 1 m o use Epi SCs. ........ 53
Figure 33 Nuclear trans lo cation of TA Z induce differentiation i n CD 1 mous e EpiSCs. ....................... 54
vi

Figure 3 4 Ov erexpressio n and n u clear translocation of YAP ind u c e mouse EpiSC dif f e re ntia tion. .. 55
Fig ure 3 5 Knock‐down e f f icie ncy of T AZ s hRNAs in CD1 m ouse E piS Cs. ............................................. 56
Figure 3 6 Kn ock‐down o f TAZ does n o t affect CD 1 mouse EpiSC sel f‐renewal under CH IR /IWR‐1
condition. ................................................................................................................................................... 57
Figure 3 7 Lo ss of T A Z pr o tein i n Wwt r 1
‐/‐
mous e E SCs. ........................................................................ 58
Figure 38 Frame‐shi f t mutations d et ected in Ww t r1 locus o f Wwtr 1
‐/‐
mo u se ESCs. ......................... 58
Figure 3 9 TA Z is dispensa b le for m ous e ESC self‐re newal under 2 i/LIF condition. ........................... 59
Figure 4 0 Lo ss of Y AP pr o tein i n Yap 1
‐/‐
mouse ESCs. ........................................................................... 60
Figure 41 Frame‐shi f t mutations d et ected in Ww t r1 locus o f Yap 1
‐/‐
mo u se ESCs. ............................ 60
Figure 4 2 YA P is dispensa b le for m ous e ESC self‐re newal under 2 i/LIF condition. ........................... 61
Figure 43 TAZ is required for th e prop er ESC‐to‐ Ep iSC conversion. .................................................... 62
Figure 44 TAZ is required for pr oper EpiSC self‐renewal under C HIR/IWR‐1 condition. .................. 63
Figure 45 Expression of TAZ‐ER
T2
fusi on prot ein in Wwtr1
‐/‐
mo u se ESCs. ........................................ 64
Figure 46 TA Z‐ER
T2
rescu es the defect of E SC‐to‐Ep iSC conv ersion in Wwtr1
‐/‐
mous e ESC s. .......... 64
Figure 4 7 YA P is required for t h e prop er ESC‐to‐ Ep iSC conv ersio n. .................................................... 65
Figure 48 YA P‐ER
T2
rescues the defect of E SC‐t o‐Ep iSC conversion in Yap1
‐/‐
mouse ESCs. .............. 66
Figure 4 9 Ov erexpressio n of TA Z‐ER
T2
a nd/or YAP‐E R
T2
is not sufficient to maint ain mo use EpiSC
self‐renew al. .............................................................................................................................................. 68
Figure 50 Cytoplasmic retention o f TA Z al one is no t suffici ent t o m aint ain mouse EpiSC self‐
renewal. ...................................................................................................................................................... 69
Figure 51 Expression of TAZ‐ER
T2
fusi on prot ein in Ctnn b 1
‐/‐
mouse E piSCs. .................................... 71
Figure 52 TA Z‐ER
T2
p romote s mouse E p iSC se lf ‐rene w a l i n the a b se nce of β ‐ca te nin. ...................... 72
Figure 53 YA P‐ER
T2
and TAZ‐ER
T2
pro motes mo use EpiSC self‐ r enewal in t h e absenc e o f β‐cate n in.
.................................................................................................................................................................... 72
Figure 54 TA Z‐ER
T2
p romote s mouse E p iSC se lf ‐rene w al when nuclear tr a nslocation o f β‐catenin is
blocked. ...................................................................................................................................................... 73
Figure 5 5 TA Z is a cytopl a smic β ‐caten in bindin g p artner i n hu m an ESCs un der CHIR/I WR‐1
condition. ................................................................................................................................................... 74
Figure 56 Nuclear trans location of TAZ or YA P ind u ces hu man ESC differentiation. ......................... 75
Figure 5 7 Ov erexpressio n of TA Z‐ER
T2
or Y AP‐E R
T2
p romotes hu man ESC self‐renewal w h e n
nuclear tr a nslocation of β‐catenin is blocked. ........................................................................................ 76
Figure 58 Serum is required for th e p r oper att ach m ent of m ou se EpiSCs o n gelati n‐coated plate. .. 79
Figure 59 Laminin and M a trigel f acilitate mous e EpiSC attachmen t in ser u m ‐fre e m e di um. ............ 80
Figure 6 0 CH IR/IWR‐1 is not su fficien t to m aint ain mou e E piSC s elf‐renew a l in serum‐ f ree medium.
.................................................................................................................................................................... 82
Figure 61 Su pplement o f ActA or bFG F on top of C H IR/IWR‐1 i s su fficie n t t o m aint ain mouse EpiSC
self‐renew al i n serum‐ fre e m edium......................................................................................................... 83
Figure 6 2 A 8 3‐01 a nd PD 03 blocks m ouse E piSC s elf‐ren e wal under CHIR/ IWR‐1 condi t ion. ......... 84
Figure 6 3 CH IR/IWR‐1 pl u s ActA or b F GF m aint ai ns hum an ES C self ‐ren ew al in ser u m‐free
medium,...................................................................................................................................................... 86
Figure 64 H 9 human ESCs cultured in mTeSR m e dium show typ i cal E SC col ony morphology. ......... 88
Figure 65 N2B27 is beneficial fo r s e rum‐free cul ture of human ESCs. ................................................. 89
Figure 6 6 N2 and B 27 sup plement ar e optimal am o n g th eir deri vat iv es for h uma n ESC cu l ture...... 91
Figure 6 7 Ch emicals th at i mproves hu man ESC self‐renewal i n N2B 27 m edium supple m e nted with
ActA/bFGF/ CHIR/IWR‐1. ......................................................................................................................... 93
vii

Figure 6 8 Ch emicals th at h amp e r hu m a n ESC sel f ‐renew al in N 2 B27 m ediu m suppleme n t e d with
ActA/bFGF/ CHIR/IWR‐1 .......................................................................................................................... 94
Figure 6 9 Ch emicals th at s how a potential clonal s e lection proc ess in N2B27 medium supplemented
with ActA/bFGF/CHIR/IWR‐1. ................................................................................................................ 95
Figure 7 0 CR M co mpoun d s show a sy n ergistic e ffe c t with L IF o n m ouse ESC self‐ren e wal . ............. 97
Figure 7 1 J A K inhibit o r b l ocks the e ffect of CRM o r LIF on pro m otin g mou s e ESC s e lf‐r enew al. ...... 98
Figure 7 2 CR M co mpoun d s show syn ergistic e ffect with L IF o n STA T3 activation. ............................ 98
Figure 73 CR M co mpoun d s cannot mai ntai n lon g ‐ t erm sel f ‐ren ewal of mouse ESCs. ...................... 100
Figure 7 4 CR M4 23F show strong eff ect on pro m oti n g mouse ESC sel f‐ren e wal . .............................. 100
Figure 7 5 CR M4 23F is a b le to m a intai n short‐term self‐ren ewal o f mo use ES Cs. ............................. 101
Figure 7 6 Th e E f fect o f C R M42 3 F on m ouse ESC self‐ren e wal i s de pendent on L I FR. ...................... 102
Figure 77 CR M compounds show neg a tiv e ef fe cts on rat ESC self‐r enew al. ...................................... 103
Figure 78 G4 promotes short‐term sel f‐renewal of mouse EpiSC s i n the pres ence o f bFG F . ........... 105
Figure 7 9 G 4 promot e s lo ng‐ter m self‐ r enewal of E 3 m ouse Epi SC in th e pre sence o f b FG F. ......... 106
Figure 8 0 G 4 selectivel y a ctivat e Sm ad2 in a dos e‐ a nd time‐dep endent man ner. ........................... 107
Figure 8 1 G 4 promot e s h u man ESC se lf‐renew al in N2B2 7 medi um su pplemented with
bFGF/CHIR/ IWR‐1. ................................................................................................................................. 108
Figure 8 2 G 4 an a logs pro motes hu man E S C se lf ‐rene w a l i n N2B2 7 m ediu m suppleme n t e d with
bFGF/CHIR/ IWR‐1. ................................................................................................................................. 109
Figure 8 3 C u rrent m odel f or CHIR/ IWR‐1 m e diate d primed sta t e PS C s e lf‐r enew al. ...................... 111

Table 1 Di ffe r ent typ e s o f PSCs derived from dev eloping mo use e mbry os............................................ 3
Table 2 Co m p arison amo n g differ ent types of PSCs from mouse and hu m a n ....................................... 7
Table 3 E x amples o f β‐cateni n in teracting partner candidates id entified b y Mass Sp ec analysis ..... 46
Table 4 Co mponents of N 2 a n d B2 7 su pplement ................................................................................... 90
Table 5 Prim er sequ e nces for P CR ......................................................................................................... 125
Ta ble 6 Tar g et seq uenc es of shRNAs a nd guide RNAs ......................................................................... 126
Table 7 Prim er sequ e nces for q PCR ....................................................................................................... 128
Table 8 Smal l molecules tested in this study ........................................................................................ 130
viii

Acknowledgement
I would like t o express my p rofound gratitude to m y Ph.D. mentor, Dr. Qi‐Long Ying, for his
long‐term support and guidance. D r. Ying has established himsel f as a wel l ‐ recognized s tem
cell biologist as w ell as a n outst a nd ing mentor. His pers ev eran c e a nd r igorousness insp ired
me w ith curiosity and confidence i n both s cience a nd l ife. H is criticis m and encouragemen t
guided m e through the progress o f t rial a nd e rror towards self‐ improvemen t. T his body o f
work, along with m y advanc ement of k nowl edge, sk ill, a nd v is ion , could not be a chiev ed
without him. I sincerely thank him for all his attention and efforts on me over the past years.
I would like t o acknowledge the current a nd p revio u s members of m y dissertatio n
committee, D rs. Robert E. Maxson, Justin I chida, D enis Evseenko , an d Leslie P. W einer, for
their cr i tic a l comments and help on my resear c h project and car eer planning.
I would like to thank USC/NICHD T32 Training Program in Develop mental B iology, Stem
Cells and R egen er atio n for support on scient ific tr a in ing and p rofessional devel o pment.
I would like t o express my d eep g ratefulness t o a ll the current and previous members of Ying
Lab. I t has been e xtr emely fort un ate for me t o pursue m y Ph.D. degree in t his grea t
laboratory. Especially, I would like t o thank Drs. S houdon g Ye, Hoon Kim and Chih‐I Tai for
their tremendous guidance and help, Drs. Chang Tong and Ping Li for their valuable
suggestions and encouragement, Dr s. Guanyi Huang, Dongbo Qiu an d Junfeng Jiang for their
insight f ul d iscussion, and Jean P au l Chadarevian, R uizhe Wang, Bryan Ruiz, and Se J ung Lee
for their persist e nt a ss ista nce.
ix

Abstract
Mouse epiblast stem cell (mEpiSC) and human embryonic stem cell (hESC) are primed state
pluripotent stem c ells w hose s elf‐ren e wal c a n be m aintained by two small molecules,
CHIR99021 ( CHIR) and IWR‐1, i n s e rum‐containing medium t hrough cytop l asm i c
stabiliza t io n and ret e ntion of β‐caten in. Th e underlying m echan ism, h owever, remains
largely unk n own. H er e I show t ha t cytoplasmic β‐catenin interac ts w ith and retains TAZ in
the cytoplasm. C ytoplasmic r etention o f TAZ/YA P promotes m EpiSC s elf‐renewal in t h e
absence of n uclear β‐catenin, whereas n u clear tra n slocation of TAZ/YAP leads to
differ ent i at ion. T AZ/YA P i s dispens a ble for naïve stat e mouse e mbryonic s tem cell ( m ESC)
self‐ren ewa l b ut r equir e d for the proper c onv e rsion of m ESCs t o m E p iSCs. The s e lf‐ren ewal
of hESCs, like that of mEpiSCs, is regulated by a similar mecha nis m i nvolving c ytoplasmic
retention of β‐catenin and TAZ. Based on CHIR and IWR‐1, I also i dentify a group of g rowth
factors and small molecules that a re b enef icial to h ESC self ‐re newal under serum‐free
conditio n. R esults f rom this s tudy d emonstr a te t hat transcripti on c o‐activators, such a s β ‐
caten i n a n d TAZ/YA P, c an e x e rt f unction a l roles in t he c y t oplas m. D iscoveries i n this s tudy
not only e xtends our understanding of the molecular mechanisms that r egulat e pluripotency
but also provide new hints for the future optimization of serum ‐free hESC c ulture c onditio n s.
1

Chapter 1 Introduction
The developmental origin of pluripotent stem cells
Mammalian development p rocess, f rom the fert ilized e gg to v arious types of somatic cells, is
accompanied by the progressive loss of developmental potential, termed potency, of
individual c ells ( Figure 1 ). D uring t h e early cleavag e d iv isions of the fertilized egg, each cell
in the embryo is able to contribute to the formation of a whole embryo[1]. Cells with such
developmental potential are toti p o tent. Upon t he f ormation o f the blastocyst, cells in the
embryo a re s epara t ed i nto two groups, the inner cell mass (ICM) and the trophectoderm (TE).
Cells i n the ICM are able t o give r ise to a ll types of c ells i n the newly formed animal, but not
the extr aembryonic t issue[2]. Ce lls w ith s u ch d evelopmental p o tential are pluripotent.
Following g astrulation, m ost cells i n the developing e mbryo und ergo f urther specifica tion ,
limiting t heir e mbryonic c ontr ibution w i thin t he der iva t ives o f their tissue of origin. Cells
with s uch development a l potential are multipotent. T h e o nly exc eptio n f or t his specificat ion
process is t he p rimordial germ c ell s ( PGCs) wh ich are unip otent i n the newly formed a nimal,
as they can only give rise to germ cells, but are able to regai n pluripotency g iven a ppropriate
cues, such as the forma t ion of z ygo t e.
2

Figure 1 Developmental potency of cells in early mouse embryos.
Pluripotent s t em c ells can be derived from th e I nn e r Cell M a ss o f e arly bl a sto c yst or the E piblast
from late bla s tocyst. Ada pted fro m A nna‐K aterin a Hadja n ton a kis L ab, M e morial Slo an Kett e ring
Cancer C ent e r.

Pluripoten t stem c ells ( PSCs) possess the abil ity to g ive r i se to any type of cell in the body,
including d e riv a t i ves from a ll three primar y germ l ayers , t hat is, ectoderm, mesoderm a nd
endoderm, in a ddit i on t o the g e rm line. Pluripotency i s n o t an i nnate property o f the zy gote,
but emerges in t he I CM o nly afte r the formation of t he p re‐impl antation b lastocyst, a nd
persists i n t h e epiblast u ntil t he p ost‐implanta tion e gg cylind er s tag e . Therefore, P SCs only
exis t tr ans i ently in vivo during earl y stag es o f embryogenesis. H owev er, when p ro vided w ith
proper c ulture c onditions, P SCs ar e capable of i ndefin ite in vitro self‐ren e wal while ret a in in g
3

their develo p mental p otent i al. Sev e ral types of P SCs have b een isola t ed a nd e stablished f rom
embryos at various developmental stages (Table 1). For example, e mbryonic s tem cells
(ESCs) c an b e establis hed from t he I CM o f pre‐implant a tion e mbr yos[3, 4 ], epiblast s tem
cells ( EpiSCs) can be e s t ablished f r o m the epiblast o f pos t ‐imp lantat ion embryos [5, 6 ], and
embryonic germ cells (EGCs) can be established from the PGCs at a l ater dev elopmental
stage[7‐10]. Alternatively, PSCs, such as induced pluripotent s tem cells ( iPSCs), can be
generated from somatic cells t hrough reprogramming[11, 12].
Table 1 Different types of PSCs derived from developing mouse embryos
Embryonic Day E0 to E2.0 E3.5 E5.25 E7.25 to E13.5
Developmental
Stage
Zygote to early
8‐cell embr yo
Pre‐impl antation
blastocyst
Pos t ‐impl a n t ation
egg c y linder
Late gastrula to
presomit e PGCs
Tissue of origin Blasomere ICM Epiblast PGCs
Potency in vivo Totipot e nt P luripotent P luripotent U nipotent
PSC Type ‐ ESCs E piSCs EGCs
Potency in vitro ‐ Pluripotent Pluripotent Pluripotent

Naïve and primed state pluripotency
PSCs p ossess two critical c harac t eristics, in definite s el f‐ren e wal ability and unlimited
differ ent i at ion potentia l. F rom the studies o f P S Cs in vivo and in vitro , several hallmarks o f
pluripotency h ave been e st ablished. First, P SCs expr ess high l e vels o f pluripotency f actor s
such as SSEA‐1 on the cell surface, a lkaline phosphatase (AP) i n the cytosol, a nd O ct4 and
Nanog in the nucleus. Second, PSCs form compact colonies with d efin ed b oundar ies during
in vitro c ulture. Third, P SCs can differentiate i nto cells f rom all thr ee primar y g e rm l ayers
and form e mbryoid bodies ( EBs) in vitro o r teratomas in vivo. Last and the most important,
PSCs a r e a b l e to r e‐ent e r embryon i c develop m ent when i ntroduced back to blastocyst and
4

contribute t o germline‐competent c himeras. A mong a ll the hallma rks described above, t he
ability to c ontribute t o g ermlin e‐competent chimera is w idely accepted as the “golden
stand a rd” for PSCs.
To date, d espite t hat PSCs f rom various s p ecies a nd developmental stages have been
reported, o n ly ESCs fro m mouse and rat bear a ll the hallmarks o f PSCs, especially the ability
to g ive ris e t o germline‐competent c himeric progen ies[1 3‐15], r epr e sen t in g the naïv e stat e
pluripotency. This pluripotent state is characterized by the ex pressio n o f naïv e pluripotency
markers, s uch as Rex1 and Nr0b1 , two active X c hromosomes i n female c ells, and the
toleranc e t o single c e ll dissociat ion. I n cont rast, EpiS Cs f rom mouse and rat, as well as
currently available ESCs from human and non‐human primates[16, 17], have most of the PSC
hallmarks but distinguish from mouse and rat ESCs by their inability to c ontribute t o
chimeras, r e present i ng t he p r i med stat e pluripotency. This p lur ipotent stat e is c ha racterized
by t he e xpr e ssion o f ea rly specific ation mark ers, s uch as Fgf5 and Brachyury , one active X
chromosome in female cells and the sensitivity to single cell d issociat ion. I n s p ite of t he
shared properties of pluripotency, naïve state PSCs are commonl y considered e quivalent t o
the cells in ICM or early stage p re‐implanta t ion ep iblast w hile p rimed state PSCs a re
commonly consider ed e quivalent t o the cells in late s tage p ost‐ impla n tation ep i blast[18].
Both naïve and primed state PSCs can be captured in culture und er p r o per condit ions in vitro,
despite that the culture conditi on s for each d iffer significant ly. Na ïve sta t e PS Cs, such a s
mouse ESCs, can be m aint ained in s erum‐containing medium s upp l e mented w ith leukemia
inhibitory factor (LIF)[19, 20]. It has been shown that LIF pro motes mouse ESC self‐renewal
through activat i on o f signal t r a nsd u cer and activator of t r a nsc ription 3 (STAT3)[21] w hile
5

B MP4 in s e ru m p r om otes m ou se ES C sel f‐ re ne wal th rou gh activat i o n of Smad1/inhibitor of
DNA bindin g 1(Id1)[22 ]. Altern at iv ely, n a ï ve s t a te P SCs ca n als o be m aint ained in serum‐free
conditio ns supplemented with t wo small m olecule inhibitors ( 2i) – C HIR99021 ( CHIR, an
inhibitor of g lycogen s y nthase k inase 3, o r G S K3) and PD0325901 ( PD03, a n inhibitor o f
mitogen‐ac tivated pro t ein kinase ( MAPK)/extracellular signal r e gulated kinase ( ERK)
kinase, or MEK)[13]. CHIR promotes self‐renewal through activat ion of W nt /β‐catenin
signal ing. P D03 contributes to s elf‐renewal through repressing the differ ent i at ion induced
b y MAPK si gnal in g. Sig nal i ng p a thways in volv ed in naïve stat e p luripotency maintenanc e in
vitro is summarized in Figure 2. In contrast, primed state PSCs, suc h as m ouse E piSCs and
currently available human ESCs, can be maintained in serum‐cont aining m ediu m
supplemented w i th t ra nsformin g gr owth f actor β (TGFβ) a nd f ibro blast growt h f actor 2
(FGF2, also known as b asic FGF, or bFGF)[5, 6, 9].

6

Figure 2 Signaling pathways involved in the maintenance of naïve state PSCs in vitro
Naïve st ate pluripotency c an b e mai n tained in vitro via activation of LIF/STAT3 pathway by LIF in
the presence o f s e rum. Alternati v e ly, it c an a lso be m aint ain e d via simultaneous activation of Wnt/β‐
catenin path way and in hibition of F GF /MAPK p a th way by tw o s mall m olecul e inhibit o rs ( 2i).

PSCs from the same species can interchange between the naïve st ate and the pr imed stat e i n
culture (Figure 3). The convers i on f rom naïve stat e to p rimed s tat e c an b e readil y achiev ed
by c hangin g the culture condition [23], fulfilling t he c rit e ria of a controlled differentiation
process. T h e r eversion f rom primed state t o na ïve sta t e, h oweve r, c annot be r ealized without
the help o f trans g en e[ 24‐27] o r complex application of c hemical s[28‐30], rep r esenting a n
authentic r e programming p roces s . Based on t he c omparison of t he ir d ev elopmental
potent ial a n d molecular propert i es ( Table 2 ) , the primed sta te PSCs are developmentally
more specified t ha n t h e naïv e st ate PSCs. It i s worth noting t h at, although c onventional
human ESCs are named as “ESCs”, they are highly similar to mous e EpiSCs i n al most e very
7

aspect ( see Table 2 for comparis on). C onsis t ent with t his notion, several recent reports have
shown that human PSCs with similar properties of mouse ESCs can b e esta blished from
currently available primed state h u man ESCs or human fibroblast s through reprogramming
under optimized culture conditions[31‐37]. Using the same condi tions, n aiv e ‐like huma n
PSCs can a lso be derived directly from human embryos[35, 37, 38].
Table 2 Comparison among different types of PSCs from mouse and human
Cell Type
Mouse
ESCs/iPSCs
Mouse
EpiSCs
Mouse
EGCs
Human
ESCs/iPSCs
Human
EGCs
Naïve State
Human PSCs
Pluripotent
State
Naïve Primed P rimed Primed P rimed Naïve
Colony
morphology
Dome Flat Dome Flat Flat Dome
Culture
Condition
LIF/B MP4
or 2i
Activin/FGF
SCF/FGF
/LIF
Activin/FGF
SCF/FGF
/Forskolin
2i a nd
other factors
1

X chromosome
inactivation
XaXa X aXi ‐ XaXi ‐ X aXa
Single ‐cell
dissociation
Toler a nt S ensitive S ensitive S ensitive S ensitive T oler ant
Teratoma Yes Yes Yes Yes Yes Yes
Chimera Yes No Y es N o ‐ Yes
Germline
transmission
Yes No N o ‐ ‐ ‐
Refs [3, 4] [ 5, 6] [7, 8] [ 16] [ 9, 1 0] [ 32‐ 37, 39]
1
Culture conditions used to deriv e a n d maint a in n aïve state hu man PSCs v ary among reports

ESCs EpiSCs
TGFβ/FGF2
Klf4, Gbx2, Tfcp2l1
or chemicals
Figure 3 Conversion between ESCs and EpiSCs in culture
ESC‐to‐EpiSC conversion can be achieved as a differentiation pr ocess by s witching c ultur e c onditio n.
Epi‐to‐ESC conversion can be achieved as a reprogramming proces s by f o rced express i on o f n aïve
state pluripo tency fact ors or complex use of chemicals.
8

Molecular mechanisms of primed state pluripotency maintenance in vitro
Mouse EpiSCs and conventional human ESCs are two examples of pr imed state P SCs in
culture. T hey util ize similar mo lecular mechanisms t o maintain pluripotency ( summariz ed
in F igur e 4 ) , as e v i den c ed b y their id ent i cal culture conditio n c onsisting of t ransforming
growth f actor β (TGFβ) a nd f ibroblast g r owth f actor (FGF) in t h e pr esence o f serum or s erum
replacemen t. T her e for e , mouse EpiSCs c ould s erve as a n u nparall eled tool f or t he studies o f
human ESCs, especiall y i n the asp e cts of u nc overing the molecul ar m echanisms governing
primed s tate p luripotency and di scovering novel and robust c ult ure conditio ns f or
mainta in in g human ESC self‐renewal.
TGFβ, or A c t ivin A ( Act A ), p romotes mouse E p iSC a n d human ESC s elf‐renew a l through the
activ a tion o f TGFβ/Smad signal ing pathway. T GFβ, o r ActA, binds to TGFβ type II receptor
(TGFβRII) o n the cell surface, w h i ch t rigg ers the forma t ion of TGFβRII and TGFβ t ype I
receptor ( TGFβRI, also k nown a s a c tivin A rec e ptor‐like k i nas e , or ALK) complex. Activated
TGFβRI i n TGFβRI/II complex acts a s a recept or t yrosin e kinas e (RTK) and pho s phorylates
Smad2/3, allowing heterodimer formation between Smad2/3 and Sma d4 i n the cytoplasm.
Smad2/4 or S mad3/4 h eterod imers then t ra nslocate i n t o the nucle us, where they a ct a s a
transcription factor to regulate the expression of various gene s in volved i n both
pluripotency m ain t en a n ce a nd l in eage c ommitmen t[40]. TGFβ/Smad signaling pathway
exerts cent r al roles in regulati n g m ou se Epi SC and h u m an ESC se lf‐renewal, as tr e atment of
TGFβ r eceptor inhibito rs l eads to r apid differ e nt iat i on i n both cells[41]. In human ESCs,
nuclear Smads bind t o the proximal p romoter regio n o f NANOG a nd directl y activ a te i ts
expression. In turn, Nanog expression blocks the endoderm diffe rentiation i nduc ed b y other
TGFβ/Smad target g enes a nd r einforces pluripotency[4 2, 43]. Sma d2/3 a lso control the
9

express i on o f other pluripotency g enes such as OCT4 and MYC, as revealed in ChIP‐seq
a n a lys i s[44, 45]. In addi t i on , Noda l, a n ot h e r membe r of T GFβ/A ctA f ami ly growth f a ctors, i s
also able to promote human ESC self‐renewal by inhibiting the d iffer e nt iation i nto
neuroectoderm[46].
FGF is another growth factor that promotes mouse EpiSC and huma n ESC s e lf‐renew al in
vitro . FGF signalin g is a ctivat ed b y a ligand‐r eceptor in ter a ction o n the cell surface that l eads
to auto‐phosphorylation of FGF r eceptor (FGFR) i n its int r acell ular d omain. P hosphorylated
FGFR serves as an RTK to activate multiple adaptor proteins tha t regulate d ifferent
downstream signaling pathways. FGF receptor substrate 2 (FRS2) and growth f actor
receptor‐bound 2 ( GRB2) are key mediators of p hosphatidylinosit id e 3‐kinases (PI3K)/Akt
and mitogen‐activated protein k i nase ( MAPK) signaling pathways downstream o f FGFs.
FGFR a lso a c tivates phospholipase C γ (PLCγ) and regulates PLCγ signaling in response to
FGF[47]. Like T GFβ signaling, F GF s ignaling i s vital for mouse EpiSC and human ESC self‐
renewal in culture, as addition of FGFR inhibitors leads to rap id d ifferentiation[28, 48] .
However, i n contrast t o the conserved transc riptional regulatio n of T GFβ/Smad signaling
pathway in p rimed state PSCs, th e downstream s ignal transductio n and t r a n script iona l
response to FGF is partially different between mouse EpiSCs and human ESCs. In mouse
EpiSCs, FGF activat e s MAPK sig n a ling p athw ay, which inhibits t h e differ ent i at ion towards
neuroectod erm and preven ts t he r evers i on t o w ards na iv e pluripot ent state [28]. In h uman
ESCs, FGF is shown to both cooperate with TGFβ/Smad signaling t o maintain N anog
express i on a nd a ct iva t e PI3 K /Akt p athway t o enhance cel l p rolif eration and surv ival[28, 48,
49]. FGF signaling al so c ontributes t o human ESC self‐ren ewal b y inhibiting n eural
induction[ 5 0 ].
10

Figure 4 Signaling pathways involved in the maintenance of primed state PSCs in vitro
Primed s tate p luripotency c a n be m aintained in v itro v ia s i m ult an eous a ctivatio n of T G F β/Smad
pathway by activin A and FGF/MAP K pa thway by FGF2 in the pre se nce of serum.

The roles of β‐catenin in regulating primed state pluripotency
Wnt sig n aling pathway plays a ce ntral and conserved rol e i n cel l prolifer atio n and lin e ag e
specific atio n during e mbryonic development [51]. Ca non i cal Wnt s ig naling p ath w ay h as a ls o
been i mplicated in c on trolling t he f a t e choice o f v a rious types of stem cells[52, 53]. In the
absence of Wnt ligand, cytoplasmic β‐catenin is actively regulated by the β‐catenin
destructio n complex ‐ a protein c omplex c on sisting of a denomato us p olyposis c oli (APC),
Axin1, casein kinase 1α (CK1α), and glycogen synthase kinase 3β ( GSK3β). Specific ally,
successive p hosphorylation o f β‐ catenin o n i ts N ‐terminus by C K1α and GSK3β leads to
proteasome‐dependent degr a dat i o n . Up on b in ding o f W n t l i gand to i t s receptor com pl e x on
11

the cell sur f ace, t he β‐ c aten in des truction c o m plex i s function ally i nhibited, leading to t he
stabiliza t io n and accumulation o f β‐caten i n in t he c y t oplasm. A ccumulated cytoplasmic β‐
catenin then t ranslocates into t h e n ucleus a nd b inds to T ‐cell factor ( TCF)/ lymphoid
enhancing factor 1 (LEF1) family transcription factors to regul ate gen e e xpress ion
(summariz e d in F igure 5). Many t a r get genes o f W nt/β‐cat e nin si gn a l in g pa t h way ha v e b e e n
iden tified i n a cell type‐specific manner. A mong direc t Wnt/β‐caten in t argets, Axin2 is a
negativ e f eedback reg u lator that b inds to β‐catenin i n the cyto plasm and blocks i ts n uclear
translocat io n[54].

12

Figure 5 The canonical Wnt signaling pathway.
In t he a bs ence o f Wnt ligand, cytoplasmic β‐catenin is p hosphor ylated b y the β‐catenin destruction
complex and degrad ed i n a pro t eas o me‐depe n de nt m ann e r. I n the presence of Wnt ligand, the β‐
catenin destruction complex is i nhibited, allowing n uclear tr a n slocation of β ‐catenin a nd a ctivatio n
of Wnt/β‐cat enin t arget genes. Adapt e d from Wray et al. (20 12) Trends in C e ll Biology[55].

The roles of W nt/β‐cat e nin signal ing pathway in p rimed state PS C fate c hoice have b een
controver sial among differ ent r e ports. S at o et a l.[56] r eported t hat activ a tion o f W n t
signal ing using BIO, a s pecific GSK 3 i nhibitor, mainta ins h u man ESC self‐renewal. Fernandez
et al.[57] showed that FZD7, a Wnt receptor, is required for th e normal e x p ansion o f
undifferentiated human ESCs. Other studies[58, 59] also demonstr a ted that a ctivation of
13

Wnt/β‐catenin s i gn alin g, b y either Wnt3a or G SK3 inhibit i on, contributes to human ESC self‐
renewal by p reventing epithelial ‐mesenchymal transition (EMT) a nd increa s ing human E S C
clonal survival. Conversely, Dravid et al.[60] and Cai et al.[6 1] r ep orted that W nt l igand
trea tment o n h uman ESCs trig gers c ell prolifer ation a n d differe ntiatio n i n a dose‐dependent
manner. Results from several studies[62, 63] on directed differ en tiat ion of h uman ESCs
further d e monstrated t hat activa tion o f Wn t sign aling, b y eithe r W n t3a or G S K 3 inhibitio n ,
facilitat e s t h e inductio n towards mesoendoderm. Recently, several studies[64, 65] also
revealed t hat r e press i o n o f Wn t/β‐caten i n sig n aling in h uman ES Cs, via chemica l i nhibitor s
or g enet ic a blation o f β‐catenin, d o es n ot h a m per self‐r enew al. Instead of promoting self‐
re ne wal, e ndoge n ous Wnt si gnalin g mediates s pontaneous d if f e ren tia t ion and generat e s an
equilibrium of l ineag e s pecific a tio n t owards distinct f ates[66, 67]. The exac t r o les of β‐
catenin in p rimed s t at e PSC s e lf‐renewal, as w ell as t he u nderl y i ng m olecular m echanisms,
remain largely obscure.
In a r ecent study[68], our lab discovered t hat the co mbined u se o f two small molec u le
inhibitors, CHIR99021 ( CHIR) and IWR‐1, c an a lso maintain m ouse EpiSC and human ESC
self‐ren ewa l w ithout e xogenous s upplement of T GFβ and FGF. C ons is ten t ly, a recent
report[69] described a simila r c ulture c onditio n c ont a in ing FGF 2 and IWR ‐ 1 for the
maint e na nce of p rimed state PS Cs including mouse EpiSCs, Rhesus macaque ESCs, and
human ESCs. Thes e r e ports provide n e w hin t s for the mechanist i c studies on the role of β‐
caten i n in p rimed st ate pluripotency regulatio n.
In our two small molecules culture system for mouse EpiSCs and human ESCs, CHIR i s a GSK3
inhibitor[ 7 0 ] that p ro motes the s t abiliza t ion of β‐caten in . IWR ‐1 i s a tanky r as e inhibitor[71 ]
14

that stabiliz e s Axin1/2 and incr eas e s their p ro tein l evels in the cytoplasm. Axin2, but not
Axin1, is a negative feedback regulator of the canonical Wnt si g n aling pathway. I n the
cytoplasm, Axin2 binds to β‐catenin and blocks its nuclear tran slocation[54, 7 2]. Thus, the
combined use of CHIR and IWR‐1 stabilizes and retains β‐catenin in cytoplasm[68] (Figure
6). As a transcription co‐activator, β‐catenin is well‐documented in regulating gene
express i on i n the nucleus. H ow c ytoplasmic β‐caten in c on trols stem cell fate c hoice, h owever ,
is currently unknown.

Figure 6 Modulation of β‐catenin maintains primed state PSC self ‐renewal
Cytoplasmic retentio n of β ‐catenin b y simultaneou s u se o f CHIR and IWR‐1 maintains mouse EpiSC
and hu man ESC self‐re n ewal in the p r esence o f se rum.

15

The roles of Hippo signaling pathway in pluripotency regulation
The Hippo signaling pathway is a key regulator of early embryog enesis, body p atter n ing,
tissue homeostasis, a nd o rgan s ize[73]. Yes‐associated p rotein (YAP, encoded by YAP1 ) and
transc riptio nal co‐activ ator w ith a PDZ‐binding domain ( TAZ, a l so k nown a s WW domain
containing t ranscr iptio n r egulator 1 , encoded by WWTR1 ) are well‐known tra n s c ription co‐
activ a tors i n the Hippo signal i ng p a t hway. Und e r normal c ondit ions, Y AP/TAZ l oc alize in t h e
nucleus where they b in d to T EA‐domain tra n scription fact or 1 ‐4 (TEAD1‐4) and regulate t he
expression of Hippo target genes. In response to extracellular or i ntracellular stimuli,
mammalian S TE20‐like p rotein k inase 1/2 (MST1/2) p hosphoryla tes and activates large
tumor suppressor kinase 1 /2 ( LATS 1/2). Activated LATS1/2 then p hosphoryla te Y AP/TAZ
in t he c yt oplasm, promoting th e r e tentio n of YA P /TAZ i n the cyt oplasm v ia 14‐3‐3
binding[74] (Figure 7). The Hippo signaling pathway is able to sense different types o f
stimuli such as mechanical forces, cytoskeletal changes, and G‐ protein c oupled r eceptor s
(GPCR)‐me diated e xtracellular si gnals[75]. Despite th at a l arg e b ody of st u dies h av e
ext e nded t he c omplexity of H ip po signalin g i n var i ous context s, the exact molecular
mechanisms through which various stimuli are transduced and con v e rged o n the regulatio n
of YAP/TAZ have not yet been d issected in d e t a il.
16

Figure 7 The core components of mammalian Hippo signaling pathway
In the absence of intrinsic or extrinsic stimuli (b, Hippo Off state), MST1/2 and LATS1/2 are not
activat e d, a l l owing nucl ear transloc ation o f Y A P /TAZ. Nuc l ear Y AP/ T AZ b inds t o TEA D 1‐ 4
transcription factors to r egul ate Hipp o target g en es. In t he p r esence o f intrinsic or e xtr i nsic s timuli
(a, Hippo O n state), activated MS T1/ 2 p hosphorylate L ATS1/ 2 a s well as t heir r egulat ory protei ns,
SAV1 a nd M OB1A/B, res p ectively. Activated LATS 1/2 then p h o sphorylate YAP/TAZ in the cytoplasm,
blocking t he n uclear t ranslocati on of Y AP/TAZ v ia p romoting t he binding between YAP/TAZ and 14‐
3‐3. Adapt ed fro m Jo hnso n et al. (2 0 1 3 ) Natur e Re v iews Dru g Dis covery[74].

The Hippo s ignaling p a t hway i s wel l ‐characterized a s the “ g ate‐ keeper ” of o rga n s ize during
development and h o meostasis[76]. YAP/ TAZ exert conserved roles in r egulating
progenitor/ s tem cell s e lf‐ren ewal a nd d iffer e nt iat i on i n multip le o rgans including l i ver,
17

pancreas, s a livar y g la nd, kidney, lung, hear t, i nt est i ne, skin, a nd n ervous sy stem[77‐79].
Nuclear YAP/TA Z (Hippo p athway k in ases i nactive st at e) p romote progenitor/stem cell
expans ion, w hereas c y t oplasmic r estr icted Y A P/TAZ (Hippo p athwa y kinases activ e s tate)
are r e quired f or t he m aint ena n ce o f proper t issue homeostasis. Upon t issu e damage o r
tumor formation, the subcellular distribution of YAP/TAZ is dra matically c hanged,
highlighting the importance of Hippo signaling pathway in maint aining n ormal tissue
homeostasis and adult tissue‐spec ific s tem cell po pulations. I n a ddition, several evidences
have demonstra ted th e importan ce o f Hippo signaling i n early em bryonic development, a s
the subcellular localization o f Y A P/ TAZ is e ssential for the pr oper sepa r atio n of T E from I CM.
During t he f ormatio n o f mouse blastocyst, blastomeres in t he m o rula s tage e mbryo form
tight junctions between each o th er[80]. Ther efore, c ells i n the outer layer acquire apical‐
basal polarity and subsequently activate the Hippo signaling pathway, which triggers
commitment tow a rds TE. Cells i n the innermost layer do n ot a cqu ire cellular polarity, which
preser ves p l uripotency in ICM[ 81, 82] (Figure 8 ).
18

Figure 8 Hippo signaling directs the formation of trophectoderm in early mouse
embryo
During the formation of blastocyst, blastomeres in the outer la yer of m or ula acq u ire apical‐bas al
polarity d uring compaction, l ea ding t o the nuclear a ccumul a tion of YAP/TAZ. Activation of Hippo
target g e n es i nduce the c o mmit m e n t t o wards trophectoderm i n t he se c ells. The i nner m o s t groups o f
blastomeres do n ot a c q uire a pica l ‐ basal polar i ty, keeping YAP/ TAZ in the cytoplasm. These
blastomeres give rise to the pluripotent cells in ICM. Adapted f r om V arel as ( 20 14) Dev e l opment [7 3] .

In contrast to early embryonic development in vivo , the fu nction o f Hippo s ignaling p athwa y
in PSC fate choice in vitro i s much l ess conclusive. Several studies r eported the function of
YAP in promoting mouse ESC self‐renewal under LIF/BMP4 conditio n. L ia n et a l.[83] f ound
that overexpression of YAP or a co nstitutively active fo rm of Y AP p romotes mouse ESC self ‐
renewal while knock‐down of YAP or TEAD induces differentiation. T a m m et al .[84] fu r th er
revealed t h a t YA P is r egulated b y Yes, a c y t o p lasmic t yr osine k in as e, downst r eam of L IF
stimulus. Activated YAP forms co mplex with T EAD2 a nd p romotes mouse ESC self‐renewal
through up ‐regulating multiple p luripotency factors including O ct4 and Nanog. A f ollow‐up
study from P ijuan‐Galit o e t al. [8 5] s howed that f actors i n ser um, such a s Inter‐ α‐inhibitor
(IαI) a lso promotes s el f‐renewal through Yes/YAP/TEAD s ignal i ng by triggering the auto‐
phosphorylation of Y es . However, a r ec ent stu dy from C hung e t a l.[8 6] r epor ted that Y AP i s
dispens a ble for mouse ESC sel f ‐r enew al b ut r equired for proper differ ent i at ion, a s Y A P
19

knock‐out (Yap1
‐/‐
) mouse ESCs c an b e established and properly m ainta i ned under
LIF/serum conditio n but exert e d defec t s during differentiation. I n addition t o directly
regulate P S C self‐ren e wal and differentiatio n, t he H ippo signal ing pathway als o c ooperates
with other s ignal i ng p a t hways to b oost self‐renewal, whic h is m ainly observed a s a funct i on
of TAZ in primed state human ESCs. Varelas et al.[87] reported that T AZ b inds to S mad2/4
complex following T G F β stimulation a n d enhances t h e n uclear r et ent i on o f Smad2/4
complex. I n the nucleus, T AZ p romotes th e association between Smad2/4 complex to
Media t or c omplex t o facilitat e T GFβ/Smad targ et g ene expres sion. In a following study,
Beyer et al.[88] further demonstrated that TAZ/YAP/TEAD, Smad2/3, and Oct4 forms a
complex in the nucleus of human ESCs. This complex binds to the e nhancer regions of b oth
pluripotency g en es a n d g enes r es ponsible f or m esoendo derm diffe r e nt iat i on, al lowing t h e
balanced r egulation of self‐ren e wa l and d i ffer ent i at ion. C onsis t e ntly, Qin et a l.[ 89] f ound
that L ATS2, by a ntagon izin g TA Z, i s a barrier o f human iPSC f ormation. Like TAZ in human
ESCs/iPSCs, Alarcon et al.[90] showed that YAP boosts BMP4/Smad 1 sig n aling in p romoting
mouse ESC self‐ren ewa l b y mediat ing Smad1‐depend ent transc ripti o n o f targ et g enes. The
deta iled m olecular m echanisms r e gard ing the control of YAP/TAZ t a rget g ene express i on
and the as sociatio n of YAP/ T A Z w ith other proteins r emain uncle ar i n the contex t of
pluripotency regulatio n.
The search of novel culture conditions for pluripotent stem cells
Currently available human ESCs and primed state mouse EpiSCs sh are a number o f similar
properties ( Table 2). Th e fac t t hat t w o dist inct stat e s of p lur ipotency c an b e captu r ed in vitro
as ESCs and EpiSCs h as stimulate d int e rest t o determin e w h ether the naïve state
pluripotency i n huma n can also b e maint a ined in vitro . Several groups h ave recently
20

reported the derivation and maintenance of naïve state human PSCs that resemble mouse
ESCs in various aspects. Gafni et al.[32] demonstrated that naïve human PSCs with the
molecular characterist ics and fu nc tional p rop e rties s i milar to those of m ouse E SCs can be
established and main t a in ed i n d e fined condit ions. I n tr ig uingly, as reported by the authors,
factors that m ainta i n n a ïve mouse ESCs such as L IF/2i and facto rs t hat maintain p r i med state
human ESCs s uch as T GFβ1/FGF2 are both r equired for the mainten ance of their naïve state
human PSCs, making i t difficult t o i nt erpret t heir r esul ts. Cha n et a l.[33] u sed a culture
conditio n c o ntaining 2 i, h uman L IF a nd D or somorphin (3i+LIF) t o der ive an d main tai n
Nanog‐positive human PSCs that share a similar gene expression p r ofile w i th
preimplant a tion h uman e piblast. T hey compared t he g en e express i on p rofile o f na ïve human
PSCs with cells in the equivalent stage human embryos, which pr ovides a n important metho d
for the eval uation o f es tablished human PSCs in vitro . Mor e over, Valamehr e t al.[36] r eported
that t he c o m bination o f 2i w ith human LIF and FGF2 c ould m ainta in a n a ï ve‐ l ike s t at e o f
pluripotency i n human PSCs. Similarly, Ware et al.[37] reported the acquirement and
maint e na nce of a n aïv e ‐like sta t e o f p luripoten c y with 2 i a n d F GF2 but without h u man LIF.
These two studies u t i lized FGF2 a s a n e ss en tial f actor for the maintenanc e of n aïve s tate
human PSCs. From w hat has been r evealed in t he m ouse, FGF2 p rom otes t he e xit from n aïv e
stat e pluripotency. Th erefor e, h uman P SCs ma inta in ed i n these c onditions likely s tay in a n
intermediate state between the naïve and the primed pluripotent s tates. M ost recently,
Theunissen et a l.[35] r eported tha t t he c ombinatio n o f 5 inhibitors (of MEK, GSK3, BRAF,
ROCK, and SRC) and human LIF (5i+ LIF) c ould m aintain naïve stat e human PS Cs w ith the
express i on o f Oct4‐GFP t ra nsgene driven by t he OCT4 d istal enhanc er, a hallmark of n aïv e
pluripotency. It is worth noting that human PSCs derived and ma intained i n 5i+LIF a re
21

reportedly r esponsive to A ctivin A , a self‐renewal‐promoting f actor for primed state PSCs.
Takashima et al.[34] r eported th at the combination of titrated 2i with LIF and PKC inhibitor
could maint a in a n aïv e ‐like human PSC population t hat sh ares t h e sa me t ra n script i on f a ct o r
control circuitry with naïve state mouse ESCs. All recent repor ts s ummarized above obtained
populations of human PSCs with mouse ESC‐like morphology. Howev er, the gen e e xpress ion
profiles, ep igen et ic sta tus, depen dency of e xtrac ellular stimul i and transc rip t ion factor
regulatory networks of these naïve state human PSC populations ar e d i stinct f rom each o ther.
Given that e piblast development in primates is more complicated t han that i n rodents, o ne
possible explanat ion for such diver sit ies b etween differ e nt n a ïve state human PSC
populations is t hat they stay in similar b ut d istinct in ter m edi at e pluripotency stat e s that
reflect successive but distinct developmental stages in vivo .
Pres ent clues strongly suggest t hat the properties o f PS Cs a re commonly shared a mong
differ ent sp ecies a nd t he m aint en ance o f plu r ipotency i s gover ned by a highly conserved
mechanism from m ouse t o human. F or e xample, 2i c ondition d evelo ped for mainta in in g
mouse ESC self‐ren ew al a lso applies o n ra t ESC culture[13‐15]. Moreover, all currently
reported n aïve s tate h u m an P SCs require LIF/ 2i[31‐37], the fact o r s for mouse and rat ESC
self‐ren ewa l , to m aint a i n pluripotency. The application o f c onventional human ESC culture
condition on the derivation of mouse EpiSCs further proves that PSCs in the same pluripotent
state, either naïve or primed, are highly similar to each other r eg ardl ess of t heir species o f
origin. Even t hough mouse EpiSCs a nd h uman E SCs are no t complet ely iden tical, f or e xample,
human ESCs a r e p o sitiv e for REX1 a nd n ega t iv e for FGF5[91], the c o m prehensiv e
comparison b etween m ouse EpiSCs and human ESCs i ndicat es t hat m ost charact e ris t ic s
observed i n primed st a te P SCs a r e highly c onserv ed b etween spec ies w hile a m inorit y of
22

species‐specific tr aits m ay e xist. As similar m olecular m echanisms govern t h e m aintenance
of t he s ame pluripoten t sta t e among different species, st udies on existing PSCs will provide
valuable insights into the searc h of pluripotency in other spec ies.
As a m att e r of f act, r ec ent inv e stig ations f ocusing on t he m ole cular mechanisms o f mouse
ESC self‐renewal h av e provided a b roader a n d m ore deta iled u nde rs tand ing of p luripotency
maint e na nce. F or e xa mple, sever a l studies h ave shed l ight o n th e det a iled r elat ionship
between STAT3 activation and mouse ESC self‐renewal. Huang et al.[92] showed that STAT3
phosphorylation at T y r 705 and S e r727 have d istinct functions on mouse ESC self‐renewal.
Phosphorylation at T y r 705 is e ss ential f or L IF/STAT3‐ m ediated mouse ESC self‐renewal.
Phosphorylation at S er727 is d isp e nsable f or t he s elf‐renewal but b eneficial to t he o ptimal
prolifer atio n and surv ival o f mouse ESCs. In ter e stingly, b locking the phosphorylation at
Ser727 d i d not hamper s elf‐renew a l but reinforced i t by i nhibit ing neural differentia tion. Ta i
et a l.[93] d emonstrat e d that a ct iva t ion lev e l of S TAT3 s hou l d b e controlled within a n optimal
rang e to p r o mote ESC self‐ren e wa l. A ctiv atio n of L IF/STA T3 s ign al in g up‐regulates b oth self‐
renewal‐promoting genes and diff erentiatio n‐inducing g enes. O n one hand, the expression
of g enes t hat promote sel f ‐ren ewal s hould reach a certa i n t h res hold t o maint a in
pluripotency, as h ypoactiva t ion of S TA T3 f a i led t o p reven t m eso derm/endoderm
specific atio n in m ouse ESCs. O n t h e other hand, the exp r essio n of t he g enes t hat promote
differ ent i at ion canno t exceed a c ert a in l evel, as hyperact i v a ti on o f STAT3 l e d to trophoblast
differentiation and ES C death. C o llectively, t hese s tudies i mpl y that, rather t h a n simply
turning on and off LIF/STAT3 signal, it is the fine tuning of t his signal ing pathway that k eeps
the balance between s elf‐ren e wal and differ ent i at ion. S imilarly, recent studies have also
provided n ovel u nders t and i ng a bout t he r ole of W nt/β‐catenin signaling in mouse ESC self‐
23

renewal. Wray et al. have shown that TCF3 is a transcription re pr essor of s ev eral c rucial
pluripotency f acto rs[9 4]. Upon β‐ c aten in b in ding, the r e pressiv e effect o f TCF3 i s allev i at ed
and the exp r essio n l ev els of c ore pluripotency f actors, s u ch a s Oct4 and Nanog, increased.
Other studies identif ied s everal c ritical pluripotency g en es, s uch as Klf2 , Esrrb, and Tfcp2l1 ,
as dir ect d o wnstream t arg e ts o f Wnt/β‐catenin sig n aling in m ous e ESCs[26, 2 7, 95]. I t i s
important to n ot e th at Tfcp2l1 h as b een i dent ified a s the dow n str e am t a r get of b ot h
LIF/STAT3 and Wnt/β ‐ catenin sig n aling path ways, whic h indicates i ts c r i tical r o le a s the
intersec tio n o f thes e two major pathways t hat r e gulate n aïv e st ate pluripotency. It s hould
also b e noted that a ct ivatio n of W nt/β‐catenin signaling pathwa y al one, b y eith e r Wnt3 a or
GSK3 inhibitors, can only maintain short term (<1 week) ESC self‐r e newal. I n a long‐term
culture, s elf‐ren e wal is g radua lly o vercome with n on‐neural dif ferentiatio n [13]. It i s
therefore believ ed t ha t Wnt/β‐caten i n has dual e ffects on p luri potency: m aint aining sel f‐
ren e wal a n d inducin g differ e nt iat i on, highlighting t he n eed to clearly dissect t he differ e nt ial
effects of W nt/β‐catenin signalin g on self‐renewal and differen tiatio n.
The effo rt o f searchin g novel cu lture condit ions c ould b e devoted via two approaches. A
st ra ight ‐f orwa rd a pproa ch i s to s creen n ovel c ombinations of gr owth f actors a nd c hemicals,
trying t o id ent i fy p hen o typically o ptimized c u l ture c ondit i ons. Recent reports on naïve state
human PSCs w ell repr esen t this a pproach b y giving r ise to v ar io us c ombinations of f actors
that c an m aintain some h allmarks o f naïve st ate pluripotency. The discovery of distinct factor
combinatio ns p rovides grea t opportunit ies t o not only f ur ther u nders t and the m e chanisms
of p luripotency maintenance but also f urther o ptimize cu rrent c onditions with t he g oal to
apply stem c ells i n therapeut i c sett ings. However, t he l ack of me c h a n i sti c i llustra ti on f or
most o f the reported c onditio n s makes their p ot ent i al a pplicati o n u nreliable an d
24

unpredicta ble. A n alt e rnat ive app r oach i s to d esig n nov e l condi tions based on t he i n‐depth
understand ing o f p luripotency re gulation, trying to fulfill all the requirements for long‐term
PSC s e lf‐renewal. R e cent i nv est i gatio n s on L IF/STAT3 a nd W n t /β‐ cate ni n si gnali n g
pathways largely support this approach by revealing the details o f signal t ransduction in
regulating pluripotency. Current evidences imply the necessity to dissect the balance
between self‐renew al a nd differentia tion t r i gger e d by a c ert a in signaling pathway, fine‐
tuning i t to wards the maintenance of p luripotency. B oth approac hes have b een i nt ensively
effective, b ut n either alone is sat isfacto r y a n d sufficient. A s i n dic ated f rom phenotypic al
screen ing, t he c ross‐tal k b etween differ e nt signalin g p at hways that u nderlies t he differen t
combinations of growth factors and chemicals is an emerging phe nomenon t h at c alls f or
further inv e stig at ion. I nspired fro m m echanistic studies, targe ted sc reening for factors or
their combinations, such as those specifically activate STAT3 o n Ty r705 but not Ser727 o r
those specifically a ct iva t e th e pluripotency g roup b ut n ot t he lineage c o mmitment group of
Wnt/β‐catenin t a rget g enes, w i ll greatly ad va nce the sea r ch o f novel culture conditio ns f or
PSCs.
25

Chapter 2 Cytoplasmic β ‐catenin regulates primed state
pluripotency
CHIR/IWR ‐1 mediate mouse EpiSC self ‐renewal through β‐catenin
In a r ecent study[68], our lab discovered t hat the co mbined u se of CHIR and IWR‐1
(CHIR/IWR‐1 herein aft e r) c ould s ubstitute TG Fβ a nd F GF2 in m ain taining mouse EpiSC and
human ESC self‐ren ew al. CHIR i s a GSK3 i nh ibitor t hat promotes the stabiliza t ion of β‐
catenin. I WR‐1 i s a tankyrase inhibitor that s tabilizes Axin2 a nd i ncreases A x i n2 p rotein l evel
in the cytoplasm. Thus, the combined use of CHIR and IWR‐1 stab ilizes a nd r et ains β‐catenin
in c ytoplasm[68]. To d etermine w hether c yto p lasmic β‐catenin i s necessary for mouse EpiSC
self‐ren ewa l u nder C H I R/IWR‐1 condit ion, I g en erat ed β‐catenin knock‐out (Ctnnb1
‐/‐
)
mouse EpiSCs u nder A ctA/bFGF c ondit i on. The loss of β ‐catenin p rot ein i n Ctnnb1
‐/‐
mouse
EpiSCs was confirmed by western blot (Figure 9). Ctnnb1
‐/‐
m ouse E piSCs could be
continuously m ainta i ned under ActA/bFGF condit ion. W hen transfe rre d t o C H IR/IWR‐ 1
conditio n, h owever, Ctnnb1
‐/‐
mouse EpiSCs failed to expand (Figure 10), indicating that β‐
catenin is dispensable for mouse EpiSCs cultured under ActA/bFG F conditio n but required
under CHIR /IWR‐1 conditio n .
26

Figure 9 Loss of β‐catenin in Ctnn1
‐/ ‐
mouse EpiSCs.
Western blot a nal y sis of β ‐catenin i n Ctnnb1
+/+
and Ctnnb1
‐/‐
mouse EpiSCs. GAPDH is loading control.

Figure 10 β‐catenin is required for moue EpiSC self ‐renewal under CHIR/IWR ‐1
condition.
Represent a ti ve p hase c ontrast i m ages o f Ctnnb1
+/+
and Ctnnb1
‐/‐
m ous e E piSCs cul t ured u nder
ActA/bFGF o r CHIR/IWR‐1 conditio ns for 7 days.
27

Cytoplasmic retention of β‐catenin is sufficient to maintain primed state
PSC self ‐renewal
To t est whether cytopla s mic retention of β‐catenin i s suffic i en t to m aintain mouse EpiSC self‐
renewal, I u tilized a ΔNβ‐catenin‐ER
T2
f usion protein to a rtific ially control the subcellular
localization o f β‐catenin. D egrad a tion o f β‐catenin is c ontrolled by CK1α and GSK3β
phosphorylation sites in its N‐terminus. Truncation of the N‐te rminus m akes ΔNβ‐catenin
constitutively s tabilized[96], wh ile the fusio n o f ΔNβ‐catenin w i th t he m utant estrog en
receptor α (ERα) ligand binding domain (ER
T2
) makes its subcellular localization u nder t h e
control of 4‐hydroxytamoxifen (4‐OHT), a synthetic ERα ligand[9 7] ( Fig u re 11A). I
introduced a f loxed ΔN β‐catenin‐ER
T2
t ra nsgene i nto the 46C m ouse E piSCs. I n t h e absence
of 4 ‐OHT, ΔNβ‐catenin‐ER
T2
f usion protein remained i n t h e cytoplasm. A dminis trat ion of 4 ‐
OHT resulted i n the nuclear tra n slocation of ΔNβ‐catenin‐ER
T2
, as confirmed by western blot
(Figure 11 B). The subcellular dist ribution o f ΔNβ‐catenin‐ER
T2
f usion protein was f u rther
confirmed by i mmunofluorescent stain i ng i n the presen ce a nd a bsence o f 4‐O H T (Figure
11C). N uclear ΔNβ‐catenin‐ER
T2
w as a ble to a ctivat e th e β‐catenin‐depend ent tr anscript ion,
as i n dicat ed b y the in creased ex pression l evels of W n t /β‐cateni n targ ets Axin2 and Cdx1
(Figure 12) . 46C m ouse EpiSCs overexpres sing ΔNβ‐catenin‐ER
T2
c ould b e mainta in ed
continuously for more than 20 passages in basal medium without addition o f exogenous
cytokines or small molecules (Figure 13A). The addition of 4‐OH T resulted i n r a pid
differentiation, even in the presence of IWR‐1 (Figure 13B). Th ese results are likely
attr ibutable t o the presence o f the ΔNβ‐catenin‐ER
T2
transgene, as its excision by Cre
recombinas e was associat ed w ith the revers io n to a w ild‐type Ep i S C ‐ li ke phe n ot y pe (Fi gu r e
13C). Collectively, these results suggest that retention of sta bil i zed β‐catenin in t he
28

cytoplasm is sufficien t for mouse EpiSC self‐ r en ewal w hereas n u clear β‐catenin induces
mouse EpiSC differen tiation.

Figure 11 Subcellular distribution of ΔN β‐catenin ‐ER
T2
is controlled by 4 ‐OHT.
(A) Schemat i c illustration of t he d o m ains o f ΔN β‐catenin‐ER
T2
f usion pr otein. (B) W e stern blot
analysis o f Δ N β‐catenin‐ ER
T2
levels in the cytoplasm, the nucleus, or whole cell lysate in 46C‐ΔNβ‐
catenin‐ ER
T2
mouse EpiS Cs in the abs ence or pres ence of 1 µ M 4‐OHT. α‐ Tubul in a nd H istone H 4 are
loading control for cytoplasmic and nucl ear lysate, respectively. (C) Representative
immu noflu o r escent s t a i n ing o f ΔNβ‐catenin‐ E R
T2
i n 46C‐ΔN β‐catenin‐ER
T2
mouse EpiSCs in the
absenc e or p resence of 1 µM 4‐OH T. H oechst stains the nuclei.
29

Figure 12 ΔN β‐catenin ‐ER
T2
activates Wnt target genes in response to 4 ‐OHT.
qPCR an alysi s of the l evel s of W nt t arg ets ge nes, Axin2 and Cdx1 , in w ild‐type ( WT) 46C or 46C‐ΔNβ‐
catenin‐ ER
T2
mous e EpiS C s in the a bs ence (N T ) or presence o f 1 µM 4‐OHT.
30

Figure 13 Cytoplasmic retention of ΔN β‐catenin ‐ER
T2
maintains 46C mouse EpiSC self ‐
renewal.
(A) Representative phase contrast images from day 1 and day 62 cultur e of 46C‐ΔN β ‐catenin‐ER
T2

mouse EpiSCs in basal me dium only. (B) Represen tative phase con trast images o f 46C‐ΔNβ‐catenin‐
ER
T2
mouse EpiSCs cultured in basal medium supplemented with CHIR/I WR‐1 in the abse nce or
presence of 1 µM 4‐OHT for 3 days. (C) Representative phase con trast i m ages o f 46C‐ΔNβ‐catenin‐
ER
T2
mouse EpiSCs transfected with a transgene for Cre and cultured in basal medium with or
without CHIR/IWR‐1 for 7 days.
31

Since endog e nous w ild‐type β‐cat enin a nd e xo genous ΔNβ‐catenin‐ ER
T2
f usion pr otein b oth
exis t in t he 46C‐ΔNβ‐catenin‐ER
T2
E piSCs, i t is p ossible that e ndogenous β‐catenin, e ither
cytoplasmic or n uclear, contribu tes to t he o b s erved phenotype. To b etter in vestiga t e the
function o f ΔNβ‐cateni n ‐ER
T2
i n mouse EpiSC self‐renewal, I overexpress e d ΔNβ‐catenin‐
ER
T2
in Ctnnb1
‐/‐
mo u s e Epi SC s. The e x pre ssi on of ΔNβ‐ca t en i n ‐ER
T2
in Ctnnb1
‐/‐
Epi SC s was
confirmed by w estern b lot (Figure 14). U n der ActA/bFGF conditio n without 4‐ OHT, ΔNβ‐
catenin‐ER
T2
fusion protein remained in the cytoplasm. Upon 4‐OHT treatment , ΔNβ‐
catenin‐ER
T2
fusion protein translocated into the nucleus and activated β‐c atenin t arg e t
genes, as confirmed by qPCR analysis (Figure 15). Ctnnb1
‐/‐
m ouse E piSCs overexpressin g
ΔNβ‐catenin‐ER
T2
c ould b e maintain ed f or m ore than 2 0 passages without an y exogenou s
cytokines or s mall mol e cules (Fig ure 16A). W hen treated with 4 ‐OHT, these cells rapidly
differentiated even in the presence of ActA/bFGF (Figure 16B). Taken together, these r e sult s
demonstrate that cytoplasmic stabilization and retention of β‐c atenin i s both r eq uired and
sufficient to mainta i n m o use EpiSC self‐ren ewa l under CHI R/IWR‐ 1 condit ion.

Figure 14 Expression of ΔN β‐catenin ‐ER
T2
in Ctnnb1
‐/ ‐
mouse EpiSCs.
Western blot a nalysis of ΔNβ‐catenin‐ER
T2
expres sion levels i n Ctnnb1
‐/‐
m ouse E piSCs trans f ected
with empt y vector ( E V) or a tr ansgene for ΔNβ‐c a teni n‐ER
T2
(ΔNβER). GAPDH is loading control.
32

Figure 15 ΔN β‐catenin ‐ER
T2
activates Wnt target genes in response to 4 ‐OHT in
Ctnnb1
‐/ ‐
EpiSCs.
qPCR analysis of the levels of Wnt targets genes, Axin2 and Brachyury , in Ctnnb1
‐/‐
E piSCs transfected
with e mpty v ector or a t ransgene f or ΔNβ‐catenin ‐ER
T2
i n the absenc e (NT) o r presence o f 1 µM 4 ‐
OHT.
33

Figure 16 Cytoplasmic retention of ΔN β‐catenin ‐ER
T2
maintains Ctnnb1
‐/ ‐
mouse EpiSC
self ‐renewal.
(A) Representative phase contrast images from day 1 and day 47 cultur e of Ctnnb1
‐/‐
;ΔNβ‐cateni n ‐
ER
T2
m ouse E piSCs in b as al m edium only. (B) R e presentative p hase c o ntras t i mages o f Ctnnb1
‐/‐
;ΔNβ ‐
catenin‐ ER
T2
mouse EpiSCs cultured in basal medium supplemented with CHIR/IWR‐1 i n the absenc e
or presence of 1 µM 4‐O H T for 3 days.

Similarly, C HIR/IWR‐1 can substi t u te T GFβ/FGF2 in m aintaining human ESC self‐renewal.
To d etermine w hether c ytoplasmic r etentio n o f β‐catenin is a ble t o pr omote human ESC self‐
renewal, I o verexpressed ΔNβ‐catenin‐ER
T2
in HES2 human ESCs. The subcellular
localization o f ΔNβ‐catenin‐ER
T2
i n the pres ence a nd a bsence o f 4‐OHT was c o nfirmed by
western blot ( Figure 1 7). HES2‐ΔN β ‐catenin‐ER
T2
h uman E SCs could be m aintained for over
34

10 passages without exogenous cy tokines or s mall molec u les whil e r e ta ining an ESC i dent it y
(Figure 18A ). The addit i on of 4‐OHT, however, induced ra pid dif ferentiation (Fig u re 18B), a
phenotype consistent w ith mouse EpiSCs. To d etermine w hether E R
T2
d omain or t reatment
with 4‐OHT contributed to the observed phenotype, I also overex pressed EGFP‐ER
T2
in HES2
human ESCs and did not observe any difference between non‐treated and 4‐OHT‐treated
groups (Figure 18C). Taken together, these results demonstrate that cytoplasmic retention
of β‐catenin i s sufficient to m aint ain mouse EpiSC and human ES C self‐renew al, whereas
nuclear tra n slocation of β‐caten i n induces mouse EpiSC and huma n ESC differ en t i at ion.

Figure 17 Subcellular distribution of ΔN β‐catenin ‐ER
T2
is controlled by 4 ‐OHT in HES2
human ESCs.
Western blot a nalysis of ΔNβ‐catenin‐ER
T2
levels in the cytoplasm (Cyto.) and the nucleus (Nu.) in
HES2‐ΔNβ‐catenin‐ER
T2
hum an E SCs in the a bse n ce (NT ) or pre se nce of 1 µM 4 ‐ OHT . α‐T u buli n and
Oct4 are loading contr o l for cyto plas mic and nucl ear lys a te, re spectively.
35

Figure 18 Cytoplasmic retention of ΔN β‐catenin ‐ER
T2
maintains HES2 human ESC self ‐
renewal.
(A) Representative phase contrast images from day 1 and day 61 cultur e of H ES2‐ΔNβ ‐ catenin‐ ER
T2

human ESCs i n b a sal medium only. (B) R epresentat ive p h ase co ntr ast images o f HES 2 ‐ΔNβ‐catenin‐
ER
T2
human ESCs cultured in basal medium supplemented with CHIR/IWR ‐1 i n the absence (N o
Treatment) or presence of 1 µM 4‐OHT for 3 days. (C) Representa tive p hase c ontrast i mages of H ES2‐
EGFP‐ER
T2
human ESCs cultured in basal medium supplemented with CHIR/IWR ‐1 i n the abs e nce
(No T r ea tment) or pre s ence o f 1 µM 4 ‐OHT f or 3 da y s. S ca le ba r s , 20 0 µm.
36

Subcellular localization is crucial for β‐catenin in mediating self ‐renewal
Nuclear tra n slocation and subsequent i ntera c tion w ith TCF famil y transc riptio n facto r s are
the two critical steps i n the ac tivation o f the canonical Wnt s ignaling p athway [ 98]. Since
many of Wnt target genes have been shown to be responsible for lin e age commitmen t [64,
94, 99], cytoplasmic retention of β‐catenin may block the activ ation of these different ia tion‐
inducing g enes a nd t hus enhance self‐ren ewal. To i nvestigat e w h ether the inter a ctio n
between β ‐ catenin and TCF family t ra nscription f acto rs i s invol ved in m ouse E piSC
differentiation induced by the nuclear translocation of β‐caten in, I g e nerat e d ΔNβ‐catenin‐
ER
T2
m utants c arr y ing point mutations that i nterrupt t he i nter actio n between β‐catenin and
TCF family t ransc r iptio n f actors. S p ecific ally, A295W/I296W, K4 35E, a n d H 469A /K470A
mutations abrogate t h e a bility o f β‐catenin t o i nt eract with T C F3 [100], TCF 4 [101], and
TCF3/TCF4 /LEF1 [102 ], respec tiv e ly. I ov er expressed t h ese mutan ts i n Ctnnb1
‐/‐
mouse
EpiSCs and assessed the transactivation activity of these mutants by qPCR analysis of Axin2
and Brachyury expression before and after 4‐OHT treatment (Figure 19). 4 ‐hour treatment
of 4 ‐OHT i n Ctnnb1
‐/‐
;ΔNβ‐catenin‐ER
T2
m ouse E piSCs induced 48.28 ± 1 .99 fold i ncrease o f
Axin2 expression. The introduction o f A295W/ I 296W or K 435E mutation in to ΔNβ‐catenin‐
ER
T2
reduced its induction of Axin2 e xpress ion by 45.28 ± 0 .03% a nd 80.92 ± 0 .02 %,
respect i vel y , while the introduct i on o f H46 9 A/K47 0 A mutation e s sentially a bolished i ts
transactivation activity (Figure 19). Although TCF3 is the majo r member o f TCF tra n script ion
factor f amily involved i n Wnt sign a l ing in ESCs [94], the differ ent i al t r a nscript i ona l a ctivit ies
observed a mong ΔNβ‐caten in‐ER
T2
m utants i s potentially a ttr ibut able t o different levels a nd
activ i ties o f TCF fam i ly t ransc r ip tion f actors [99]. Ctnnb1
‐/‐
E piSCs overexpressing ΔNβ‐
catenin‐ER
T2
o r an y of i ts r ema i n e d undiffer ent i at ed i n bas a l medium o nly. The p r esence o f
37

4‐OHT, surprisin gly, r esulted in r apid differ e ntia tion o f all c ell lines overexpr essin g ΔNβ‐
catenin‐ER
T2
m utants, including the H46 9 A /K470 A mutant t hat lacks transacti vatio n
activ i ty ( Figure 20). These r e sults ind i cat e t hat nuclear tra n s location o f β‐catenin induces
mouse EpiSC differen tiation reg a r d less of the β‐catenin‐depend en t tran scri pti on statu s.

Figure 19 Blocking the interaction between β‐catenin and TCF family transcription
factors reduced the transactivation activity of β‐catenin.
qPCR a nalysis of Axin2 and Brachyury e xpression in Ctnnb1
‐/‐
m ouse E piSCs overexpressi ng w ild‐type
or mut ated Δ Nβ‐catenin‐ E R
T2
in the a bsence (N T ) or presence of 1 µM 4‐O H T for 4 h .
38

Figure 20 Cytoplasmic translocation of ΔN β‐catenin ‐ER
T2
mutants induces
differentiation in Ctnnb1
‐/ ‐
mouse EpiSCs.
Represent a ti ve p h a se c ontrast images o f Ctnnb1
‐/‐
m ous e E pi SCs over exp r essing i ndi c ated ΔNβ‐
catenin‐ ER
T2
mutants cultured in basal medium in the absence or presence of 1 µM 4‐OHT for 3 days.
Scale bars, 200 µ m .

To investigate whether the interaction between β‐catenin and TC F family t ranscription
factors is i n v olved in h uman ESC differ e nt iat i on i nduced b y the n uclear t ranslocation o f β‐
catenin, I overexpressed ΔNβ‐catenin‐ER
T2
m utants c arrying A295W/I296W, K435E, or
H469A/K470A m utation in H ES2 huma n ESCs. In b asal m edium only, HES2 human ESCs
overexpres sing a ny o f the ΔNβ‐catenin‐ER
T2
mutants expanded continuously and retained
their ESC i d e nt ity. I n th e presence o f 4‐OHT, h owever, all huma n ESC lines over expressing
ΔNβ‐catenin‐ER
T2
m utants d iffer e ntia ted (Fig ure 21). Th ese r e sults indica te t ha t nuclear
translocat io n of β‐catenin i nduces h uman ESC differentiation r egardless of the β‐catenin‐
depend ent transc riptio n status. Taken together, this f u r ther i m plies that r et ent i on o f β ‐
39

catenin in the cytoplasm is the key for β‐catenin‐mediated mous e EpiSC and human ESC self‐
renewal.

Figure 21 Cytoplasmic translocation of ΔN β‐catenin ‐ER
T2
mutants induces
differentiation in HES2 human ESCs.
Represent a ti ve p h a se c on trast im ages o f HES 2 h u m an E SCs ov erexpressing indicated ΔNβ‐catenin‐
ER
T2
mutants cultured in basal medium in the absence or presence of 1 µM 4‐OHT for 7 days. Scale
bars, 20 0 µ m .

A recent r eport has su ggested t hat β‐catenin forms a complex wi th O ct4 and E‐ Cadherin o n
the cytoplasmic membrane a nd t hi s complex is i nvolved in r egulating ESC self‐renewal and
differ ent i at ion[103]. β‐caten i n is r ecruited t o the cell membrane m ainly through binding t o
E‐Cadherin [104]. To i n v estigat e d whether membrane‐bound β ‐catenin a lso has a role i n the
maint e na nce of m ouse EpiSCs, I t ested the s e lf‐ren ewal a bility of E ‐Cadherin knock‐out
(Cdh1
‐/‐
) mouse EpiSCs u nder differ e nt c ond i tions. Cdh1
‐/‐
m ouse E piSCs were n ormally
mainta in ed u nder A ctA /bFGF c ondition. When t ransferr ed t o CHIR/ IWR‐1 condition, Cdh1
‐
40

/‐
m ouse EpiSCs expan ded n ormal l y and retained a n Ep iSC id ent i ty (Figure 22) , ind i cat i ng
that m embrane‐bound β‐catenin is l ikely not r e quired f or m ouse EpiSC self‐r en ewal u nder
CHIR/IWR‐1 condition.
41

Figure 22 Membrane ‐bound β‐catenin is not required for Mouse EpiSC self ‐renewal
under CHIR/IWR ‐1 condition.
(A) Representative phase contrast images of Cdh1
‐/‐
mouse EpiSCs cultured under ActA/bFGF or
CHIR/IWR‐1 c onditions f or 1 0 da ys. (B) Represent a tive p has e c on trast and Oct4 i mmunofluorescent
staining i mages o f Cdh1
‐/‐
mouse EpiSCs under CHIR/IWR‐1 condition. (C) qPCR analysis of the levels
of n aï ve a nd p rimed stat e pluripotenc y m arkers i n 46C mo use ESC s, CD1 m ouse E piSCs, a nd Cdh1
‐/‐

(E‐Cad KO) m ouse EpiSC s.
42

TGF β/Smad and FGF/MAPK signaling pathways are not the direct targets
of CHIR/IWR ‐1 in mouse EpiSCs
β‐catenin is well characterized as a cytoskeletal protein on the cell membrane and a
transcription co‐activator i n th e nu cleus [105]. The function o f β‐catenin in t he c ytoplasm,
however, remains largely unknown. I reasoned that cytoplasmic β ‐c atenin m ight p romote
mEpiSC self‐renewal by interacting with and/or modulating proteins tha t r eg ulate self‐
renewal and differentiation. T GFβ /Smad a nd F GF/MAPK signal ing p athways have b een w ell‐
character i z e d as k ey r egulators of m EpiSC self‐ren ewa l . I sough t to determine w hether
cytoplasmic β‐catenin inter a cts w i th r egulator y proteins i n the se t wo p athways t o p romote
self‐ren ewa l . To t his en d, I i n v est i g a ted wh eth e r thes e tw o sig nal i ng p athways are act i vated
by CHIR/IWR‐1 in CD1 mEpiSCs, a mEpiSC line derived from embryo nic day 5.75 (E5.75)
CD1 mouse embryos[68]. CD1 mouse EpiSCs were treated with CHIR or IWR‐1 or both
following serum starvation. The activation of TGFβ/Smad signaling and FGF/MAPK signaling
were assessed by western blot anal ysis of p‐Smad2/3 and p‐ERK1/ 2 levels b efore and after
trea tment (Figure 23). Interest ingl y, n either p ‐ S mad2/3 n or p ‐E RK1/2 level increased 1h o r
2h after CHIR and/or IWR‐1 treatment, indicating that TGFβ/Smad and FGF/MAPK signaling
are unlikely the direct targets of CHIR/IWR‐1 in mouse EpiSCs. Of note, p‐ERK1/2 level did
show a slight increase after 4h treatment, which worth further in vest igat ion. B oth pSmad2/3
and p‐ERK1/2 l evels increased 12h a ft er t reatment i n ev ery g r ou p, l ikely due to t h e
accumulation of autocrin e TGFβ and FGF ligands.
43

Figure 23 TGF β/Smad and FGF/MAPK pathways are not the direct targets of CHIR/IWR ‐1.
(A) Western blot a nalysi s phosphorylated S mad 2 ( p‐Smad2) a nd to tal Smad2 (t‐Smad2) levels i n CD1
mouse EpiSCs t reated w ith indicated factors for 1h a fter s erum starvation o vernight. (B) Western blo t
analysis of phosphorylated Smad2 (p‐Smad2) and total Smad2 (t‐S mad2) l e vels i n CD 1 mouse EpiS Cs
treated with i ndicated f ac tors f or 2 h or 4 h aft e r s e rum starvat io n o v erni ght. ( C) W est e rn b lot an alysis o f
phosphoryla t ed E rk 1/2 (p‐Erk1/2) a nd t otal E r k 1/2 (t‐Erk 1/2) l e v e ls i n CD 1 mo use EpiSCs t re ated
indicated fac t ors for 2 h o r 4 h a fter se r um starv ati on o v erni g ht . α‐Tub u lin i s loading con trol.
44

Identification of cytoplasmic β‐catenin binding partners under
CHIR/IWR ‐1 condition
Since no o bvious f unction of c y t oplasmic β‐caten in h as b een des cribed, I r e a s oned t ha t
cytoplasmic β‐catenin might prom ote mouse EpiSC self‐renew al t h rough inter a ction w i th
and modulation of other proteins that, in turn, have a key func tion o n regulating s elf‐renewal
and d i ffer entia t ion. T o gain a n un biased v iew of a ll the binding partners o f cy toplasmic β‐
catenin under CHIR/I WR‐1 c onditi on, I performed co‐immunoprecip itat ion ( c o‐IP) of
cytoplasmic β‐caten i n using a n tib o dies a gain st β‐caten in o r ERα i n 46C‐ΔNβ‐catenin‐ER
T2

mouse EpiSCs. The co‐IP protein p ool w as t hen subject to S DS‐PA GE separa t ion a n d west ern
blot a nalysis (Figure 24). I n the Co‐IP protein pool, I detecte d several well‐known β‐catenin
int e ract ing p r ot ei ns in cy top l asm , su ch as GS K 3β ( b u t not GS K3α ), A xin1, and Axin2 (Figure
24), which confirms t he s uccessful capture of β‐catenin interacting proteins in the cytoplasm.
Next, I a na lyzed the Co‐IP protein pool u sing h igh performance liquid c hro m atograph y
(HPLC) f ollowed by t andem mass spectrometry ( HPLC‐MS/MS) an d id ent i fied o ver 10 0
candida t es t hat poten t ially int er act with β‐catenin in the c yto plasm (Table 3).
45

Figure 24 Capture of cytoplasmic β‐catenin interacting partners by co ‐IP.
Western blot analysis of G SK3, Axin1, and Axi n2 in the co‐ IP pr ot ein p ool o f cytoplasmic β‐catenin i n
46C‐ΔNβ‐catenin‐ER
T2
mouse EpiSCs cultured und er CHI R/IW R‐1 conditio n. Lan e 1, 5 % I nput ; L a n e
2, IP; L ane 3, supernat a nt .

46

Table 3 Examples of β‐catenin interacting partner candidates identified by Mass Spec analysis
Structure Metabolism Signaling Biosynthesis
Myosin
Glyceraldeh y de‐3‐phosphate
dehydrogenase
PP1β catalytic subunit 40S ribosomal protein
Tropomyosin ATP s y ntha se I GF2 mRNA‐binding pr o tein 1 E longation factor 1α
Actin‐r e lat e d protein Aden ylate k i nas e P KC/ C K2 s ubstrate 3 60S ribosomal protein
Tubulin
L‐threonine 3‐
dehydrogenase
Growth hormone‐regulated TBC
protein 1
RuvB‐like 2
F‐actin capping protein A lpha‐enolase
Guanine nucleotide‐bin ding
protein G(I)/G(S)/G(T) subunit β1
Nascent pol y peptide‐a s sociated
complex subunit α
Glypican‐4
Acyl‐coenzyme A
thioest e ras e
GTPase ERas
Guanine nucleotide‐bin ding
protein subunit beta‐2‐ l ike 1
Kera tin Glutathione S‐transfera se P 1 Ras‐related protein R‐ R as 5 '‐nucleotidase
Cadher in A lkalin e phosphatase GTP‐bindin g nuclear pr otein R an T HO complex subunit 4
Nucleolin Aldehyde dehydrogenase
cAMP‐dependen t protein kinas e
type Iα
NEDD4‐bin ding protein 1
α‐Catenin Trifunction a l enzyme R as‐related protein R‐ R as2

47

Chapter 3 The roles of YAP/TAZ in primed state pluripotency
maintenance
TAZ, but not YAP, is a binding partner of cytoplasmic β‐catenin in mouse
EpiSCs under CHIR/IWR ‐1 condition
YAP and TA Z are two t r anscr i ptio n co‐activat ors in t he H ippo si gnaling pathway. R ecently,
YAP/TAZ h a ve b een reported t o inter a ct w it h β‐catenin in v ar iou s context s . Among many
reported β‐catenin interacting partners, I focused on YAP/TAZ a s cand idat e proteins f or t wo
reasons. F irst, the interac t ion between β‐catenin and YAP/TAZ takes place in the
cytoplasm[106, 107]. S e cond, YAP and TAZ are involved i n the fa te c ontrol o f various types
of s tem cells, including human ESCs [87, 88, 108]. To d etermine whether YAP and TAZ are
binding partners of cytoplasmic β‐catenin in the context of CHI R/IWR‐1‐mediated E piSC
self‐renewal, I performed co‐IP of β‐catenin in CD1 mouse EpiSC s cultured u nder C HIR/IWR‐
1 condition. Proteins in β‐catenin co‐IP pool was then subject to w estern b lot analys is. A
strong p hys i cal int e rac t ion between β‐caten in a nd T A Z w a s detec t e d (Figure 25), whereas a
physical interaction between β‐catenin and YAP was barely detec table. T his result s uggests
that T AZ, but not YAP, i s a β‐catenin binding p a rtner in t h e c y toplasm of m ouse E piSCs under
CHIR/IWR‐1 condition.
48

Figure 25 TAZ is a β‐catenin binding partner in mouse EpiSCs under CHIR/IWR ‐1
condition.
Western blo t a n a lysis o f T AZ a nd Y A P i n t he co‐IP protein po ol of β‐catenin in CD1 mouse EpiSCs
cultured und er CHIR/IW R‐1 conditio n.

Overexpression of TAZ induces mouse EpiSC differentiation in a β‐catenin
independent manner
To c haract erize the fun c tion o f TA Z in m ouse E piSC s elf‐r e newal , I overexpr essed TAZ in 4 6 C
mouse EpiSCs c ultured under CHIR /IWR‐1 c onditio n . While mouse E piSCs transfected with
empty vect or e xpand e d and ret a in ed a n EpiSC iden tit y , those tra ns fec t ed w ith TA Z tran sgene
differ ent i at ed r apidly so that n o stable c ell line c ould b e established ( Figure 26). The
express i on levels of T A Z a nd its d irect downs t ream t a r get Ctgf [109] increas e d significan tly
in TAZ‐transfected cells as comp ared to empty vector tr a n sfect e d cells (Figure 27).
49

Figure 26 TAZ overexpression induces differentiation in 46C mouse EpiSCs under
CHIR/IWR ‐1 condition.
Representative phase contrast images of 46C mouse EpiSCs transf ected with empty vector or a
tra nsg ene f o r T A Z a nd cu ltured under CHIR/IWR‐ 1 conditio n for 3 d ays.

Figure 27 Overexpression of TAZ activates Hippo target genes in 46C mouse EpiSCs.
qPCR a na lysis of the leve l s of Taz and Hippo target gene, Ctgf , in 46C m ouse EpiSCs trans f ected with
empty vector or a tra nsgene f or T A Z .

To determine whether the differentiation phenotype induced by TAZ i s a universal
phenomenon o r unique t o the CHIR/IWR‐1 culture condition, I als o int r oduced e mpty v ecto r
50

or TAZ transgene into 46C mouse EpiSCs cultured under ActA/bFGF c ondition. Like t hose
under CHI R /IWR‐1 c ondit i on, mouse EpiSCs t ra nsfected w ith TAZ t ra nsg e ne r apidly
differ ent i at ed i n Act A /bFGF condition so t hat no stable cell line c ould b e established (Figur e
28). These results ind i cate t hat o v erexpress i on o f TAZ and subs equent a ctiva t ion of T AZ
targ et genes induce differ ent i at ion of mouse EpiSCs cultured in CHIR/IWR‐1.
Since β‐catenin is a b ind i ng p art n er o f TAZ (Fig ure 25) a nd n uc lear t ra nslocation o f β‐catenin
induces mEpiSC different ia tion ( Figures 13, 16), I invest iga t ed whether β‐catenin is
necess ary for TAZ‐ind u ced mEpiSC differen t iat i on. To t his end, I overexpres sed TAZ in
Ctnnb1
‐/‐
m ouse E piSCs. I ncreased e xpressio n l evels of b oth TAZ and its target g enes, Ctgf
and Axl [11 0 ], were c o n firmed b y qPCR a naly sis (Figure 29). Interes t ing ly, Ctnnb1
‐/‐
mouse
EpiSCs o ver e xpress ing TAZ tra n sg ene also differentiated r a pidly ( Figure 30). Taken together,
these resul t s demonst r ate tha t n uclear t ra nslocation o f TAZ pr o motes mo use EpiSC
differ ent i at ion in a β‐ca ten i n‐ind e p e ndent a n d culture conditio n‐independent ma nner.

51

Figure 28 TAZ overexpression induces differentiation in 46C mouse EpiSCs under
ActA/bFGF condition.
Representative phase contrast images of 46C mouse EpiSCs transf ected with empty vector or a
tra nsg ene f o r T A Z a nd cu ltured under ActA/bFG F condition for 3 days.

Figure 29 Overexpression of TAZ activates Hippo target genes in Ctnnb1
‐/ ‐
mouse
EpiSCs.
qPCR analysis of the levels of Taz and Hippo target gene, Ctgf and Axl, i n Ctnnb1
‐/‐
mouse EpiSCs
transfect e d with empt y v ector or a transgene for T AZ.
52

Figure 30 TAZ overexpression induces differentiation in Ctnnb1
‐/ ‐
mouse EpiSCs.
Represent a ti ve p h a se c ontrast images o f Ctnnb1
‐/‐
m ouse E piSCs transfected with e mpty v ector or a
tra nsg ene f o r T A Z a nd cu ltured under ActA/bFG F condition for 3 days.
Nuclear translocation of TAZ induces mouse EpiSC differentiation
To det ermine w hether m EpiSC differ ent i at ion is i nduced b y cytop lasmic o r nuclear TAZ, I
overexpres sed TA Z‐ER
T2
in CD1 mouse EpiSCs cultured under CHIR/IWR‐1 condition. In th e
absence o f 4 ‐OHT, TA Z‐ER
T2
exclusively stays in the cytoplasm. Therefore, I was able to
recover stable EpiSC lines overexpressing TAZ‐ER
T2
and confirm the presence of TAZ‐ER
T2

by w estern b lot (Figure 31). Upon t rea t men t w ith 4‐OHT, T AZ‐ER
T2
t ranslocated into t he
nucleus, a s evid enced by q PCR analysis o f TAZ and its targ et g e nes, Ctgf and Cyr61 [109]
(Figure 32). Und e r CHIR/IWR‐1 p lu s 4‐OHT condit ion, C D1‐TAZ‐ER
T2
EpiSCs d i ffe ren t i a ted
rapidly (Fig ure 33), co nfirming t h a t nuclear t r anslocat ion of T AZ and subsequent ac tiv a tio n
of its target genes induc e m ouse EpiSC differen tiat ion.
53

Figure 31 Expression of TAZ ‐ER
T2
fusion protein in CD1 mouse EpiSCs.
Western blot a nalysis of T AZ‐ER
T2
e xpression level in C D1 m ouse E piSCs transfected with e mpty
vector ( EV) or a pl a smid encoding TA Z‐ER
T2
fusio n protein.

Figure 32 TAZ ‐ER
T2
activates Hippo target genes in response to 4 ‐OHT in CD1 mouse
EpiSCs.
qPCR analysis of the levels of Taz and Hippo target s ge nes, Ctgf and Cyr61, in CD1 EpiSCs transfected
with empt y v ector or a plasmid enco d ing TAZ‐ ER
T2
in the absence (NT) or pr esence o f 1 µM 4 ‐OHT.
54

Figure 33 Nuclear translocation of TAZ induce differentiation in CD1 mouse EpiSCs.
Representative phase contrast images of CD1 mouse EpiSCs transf ect e d with e mpty v ector or a
plasmid encoding T AZ‐ER
T2
and cultured under CHIR/IWR‐1 condition in the absence or pres ence o f
1 µM 4 ‐OHT f or 3 d a ys.

Overexpression and nuclear translocation of YAP induce mouse EpiSC
differentiation
Since YAP and TAZ are both t ranscription c o‐activato rs i n the H ippo s ignaling p at hway, I also
investig at ed t he r ole of Y AP i n mouse EpiSC self‐r en ewal b y ove rexpr e ssing YAP i n CD1
mouse EpiSCs. While mouse EpiSCs o verexpres sing empty vector r emain ed u ndifferentiate d
under CHIR /IWR‐1 c on ditio n , those overexpr essin g YAP differ e nt i at ed r apidly so that n o
stable c ell line c ould b e est a blished (Figure 34A). T o d e termin e w h ether cyto plasmic or
nuclear YA P induces mouse EpiSC differ ent i ation, I o verexpr e sse d YAP‐ER
T2
in CD1 mouse
55

EpiSCs u nd er C HIR/I W R‐1 condition. S t a ble CD1‐Yap‐ER
T2
cell line was established in the
absence of 4 ‐OHT. These cells e xpanded normally u nder C HIR/IWR‐ 1 condit ion wh ile
reta in ing a n EpiSC i dentit y. U po n treatment with 4 ‐OHT, CD1‐Yap ‐ER
T2
E piSCs underwent
rapid differ ent i at ion (Figure 34B). S ince T A Z a nd YAP b oth act as transcription co‐activator
to r egulat e a shared set o f targ et g enes i n Hip p o signal in g pat hway, these result s imply that
overexpres sion a nd n uclear t rans l o cation o f TAZ/YAP induce m ous e EpiSC differen t iatio n
via act i vation of TAZ/ YAP target g enes.

Figure 34 Overexpression and nuclear translocation of YAP induce mouse EpiSC
differentiation.
(A) Representative phase contrast images of CD1 mouse EpiSCs transfect e d with e mp ty v ector or a
transgene for YAP and cultured und er CHIR/IW R‐1 condition for 3 d ays. (B) Repr esentative p has e
contrast i mages of C D1 m ouse E pi SC s transf ected a plas mid e n cod ing YAP‐ER
T2
and cultured under
CHIR/IWR‐1 condition i n t he a bsenc e o r presence o f 1 µM 4‐O HT f or 3 day s.
56

Reducing TAZ level in mouse EpiSCs shows no obvious phenotype
TAZ induces mouse EpiSC differentia t ion wh en o v e rexpressed and translocated i nto th e
nucleus. Since repressing a differentiation factor could contri bute t o self ‐renewal, I sought
to det erm i n e wh e th e r d own‐ regu l a tion o f TAZ is ben efic ial to mo use EpiSC self‐r enewal. To
this end, I introduced s crambled or TA Z shRNAs in t o CD1 mouse E piSCs under CHIR/IWR‐1
conditio n. K nock‐down effic iency w as a ss ess e d by q PCR analys is (Figure 3 5 ). S table mouse
EpiSC lines express i ng scrambled o r TAZ shRNA were t hen cultured in the presence or
absence of C HIR/IWR‐1 for 7 days . Surprisingly, no d iffer e nce w as o bserved between
scrambled and TAZ shRNA groups ( Figure 36), suggesting that dec reas ing TA Z level has no
obvious effect on mouse EpiSC sel f ‐ren ewal under C HIR/IWR‐1 con dition.

Figure 35 Knock ‐down efficiency of TAZ shRNAs in CD1 mouse EpiSCs.
qPCR analysis of Taz lev el in CD 1 EpiS Cs carrying d i ffer e nt T A Z s hRNAs.
57

Figure 36 Knock ‐down of TAZ does not affect CD1 mouse EpiSC self ‐renewal under
CHIR/IWR ‐1 condition.
Representative phase contrast images of CD1 mouse EpiSCs carrying scrambled or TAZ shRNAs
cultured in t h e absenc e o r prese nce of CHIR/IWR‐ 1 conditio n for 7 d ays.

TAZ/YAP are dispensable for mouse ESC self ‐renewal
To further investigate the role of TAZ in mouse EpiSC maintenance, I gene rated TAZ knock‐
out (Wwtr1
‐/‐
) mouse ESCs u sing C RISPR/Ca s9 t echnology. C omplete eliminat ion of TAZ
protein in Wwtr1
‐/‐
m ouse ESCs was confir med by w ester n b lot (Figure 37). Frame‐sh ift
mutations were s ubsequently dete cted w ithin the guid e RNA targ eting region in the genome
of each individual clone with null level of TAZ (Figure 38). Wwtr1
‐/‐
mouse ESCs expanded
efficiently under 2 i /LIF c onditio n a nd f orm e d typic a l compact E SC c olonies ( Figure 39) ,
indica ting t hat TA Z is d ispens ab le for mouse ESC self‐renewal.
58

Figure 37 Loss of TAZ protein in Wwtr1
‐/ ‐
mouse ESCs.
Western blot analysis of TAZ protein level in individual clones of Wwtr1
‐/‐
mouse ESCs. GAPDH is
loading c ontrol.

Figure 38 Frame ‐shift mutations detected in Wwtr1 locus of Wwtr1
‐/ ‐
mouse ESCs.
Geno mic DNA sequence a nd c orresp onding prot ein sequence o f the disrupted Wwtr1 g ene locus in
Wwtr1
‐/‐
m ESC Clon e 1 s howing a frame‐shift mut a tion o f 4‐bp d eletio n t hat leads to a p re‐matur e
stop codon i n the resulti ng prot e in s e q uence. G uide RNA seq ue nc e is hi g hlighted in r e d box.
59

Figure 39 TAZ is dispensable for mouse ESC self ‐renewal under 2i/LIF condition.
Represent a ti ve ph a se c ontrast images of Wwtr1
‐/‐
m ouse ESCs cultured und er 2i/L I F.

Since YA P exerts s imilar e ffects a s TA Z in i nducing mouse EpiSC differ e nt ia tion, I als o
generated YAP knock‐out (Yap1
‐/‐
) mouse ESCs using a similar approach. Complete
elimina t ion of YAP p rotein i n Yap1
‐/‐
m ouse ESCs was confirmed by w estern b lot (Figure 40).
Frame‐shift mutations were det ected within the guide RNA target in g reg i on i n the genome
of each indi vidu al cl one wi t h nu ll level of Y AP (Figure 41). In te re stin gly, Yap1
‐/‐
mou se ESCs
also e xpanded efficiently under 2i/LIF c ondition a nd f ormed typical compact ESC colonies
(Figure 42), ind i cat i ng Y AP i s dispensable for mouse ESC self‐r en ewa l . Taken tog e ther, thes e
results indicate t hat TAZ, a s we ll a s YAP, i s dispensable for mouse ESC self‐renewal,
consistent w ith a r e cent study[86 ].
60

Figure 40 Loss of YAP protein in Yap1
‐/ ‐
mouse ESCs.
Western blot analysis of YAP protein level in individual clones of Yap1
‐/‐
mouse ESCs. GAPDH is
loading c ontrol.

Figure 41 Frame ‐shift mutations detected in Wwtr1 locus of Yap1
‐/ ‐
mouse ESCs.
Geno mic DN A seq u enc e a nd c orr e sponding prot ein s e qu enc e o f the disrupted Yap1 g ene locus in
Yap1
‐/‐
m ous e E SCs showing frame‐shift mut a tions that l eads t o a pr e‐m at ure stop c od on i n t he
resulting protein sequence.
61

Figure 42 YAP is dispensable for mouse ESC self ‐renewal under 2i/LIF condition.
Represent a ti ve ph a se c ontrast images of Wwtr1
‐/‐
m ouse ESCs cultured und er 2i/L I F.

TAZ/ YAP are required for the proper conversion of mouse ESCs to EpiSCs
To c onver t Wwtr1
‐/‐
ESCs into EpiSCs, I transferred these cells to ActA/bFGF plus CHIR/IWR‐
1 culture co ndit ion, f ollowing a s ta ndard ESC‐t o ‐EpiSC c onversio n p ro tocol[23]. Surprisingly,
transf erre d Wwtr1
‐/‐
ESC s un der we n t ma ssi v e ce ll de a t h st a r tin g f r om da y 2 . A ll the
trans f erred cells d ied b y day 5. A mong t hree i ndiv idual clones tested, C11 and C12 failed t o
give rise to any mouse EpiSCs, while C1 gave rise to few colonies with poor mouse EpiSC
morphology (Figure 43). I then recovered mouse EpiSC colonies f ro m Wwtr1
‐/‐
C1 mouse
ESCs a nd t ested their self‐renew al a bility u nder d iffer e nt c onditions. In the presence of
CHIR/IWR‐1 o nly, Wwtr1
‐/‐
C1 mouse EpiSCs expanded poorly with compromised
morphology compared to wild‐type 46C mouse EpiSCs (Figure 44). When s upplemented
with ActA/bFGF in addition to CHIR/IWR‐1, Wwtr1
‐/‐
C 1 mouse EpiSCs s how e d healthier
morphology (Figure 44). The failed conversion from ESCs to EpiSCs and the compromised
self‐ren ewa l a bility o f conver ted mouse EpiSCs o bserved in Wwtr1
‐/‐
c lones suggest t hat TAZ
is nec essary for the proper conv ersion of mo use ESCs to EpiSCs.
62

Figure 43 TAZ is required for the proper ESC ‐to ‐EpiSC conversion.
Representative phase contrast images of Wwtr1
‐/‐
mouse ESCs cultured under 2i/LIF or
ActA/bFGF/ CHIR/IWR‐1 conditions for 5 d ays.
63

Figure 44 TAZ is required for proper EpiSC self ‐renewal under CHIR/IWR ‐1 condition.
Represent a ti ve p hase c on tras t images o f wild‐type 46C o r Wwtr1
‐/‐
m ouse E piSC ( clone 1 ) cultured
under CHIR/ I WR‐1 condi t ion with or without Act A/bFGF for 7 days .

Next, I o v e r e xpress ed T AZ‐ER
T2
in Wwtr1
‐/‐
mouse ESCs. Expression of TAZ‐ER
T2
in Wwtr1
‐
/‐
;TAZ‐ER
T2
mouse ESCs were confirmed by western blot (Figure 45). I trans f e rre d Wwtr1
‐
/‐
;TAZ‐ER
T2
mouse ESCs to ActA/bFGF/CHIR/IWR‐1 culture condition, followin g the same
protocol a s for Wwtr1
‐/‐
ESCs. The Wwtr1
‐/‐
;TAZ‐ER
T2
mouse ESCs were able to convert to
mouse EpiSCs a t h i gh e fficiency (Figure 46), in dicating t ha t TA Z‐ER
T2
c ompletely rescued the
ESC‐to‐EpiSC conversion defect in Wwtr1
‐/‐
E SCs. C ollect ively, t hese r esults i nd icate that
cytoplasmic TAZ is n ecessary fo r the proper conversion of mouse ESCs to mouse EpiSCs.
64

Figure 45 Expression of TAZ ‐ER
T2
fusion protein in Wwtr1
‐/ ‐
mouse ESCs.
Western blot a nal y sis of TAZ‐ER
T2
f usion protei n levels i n Wwtr1
‐/‐
mouse E SCs not trans fected (NT)
or transfected with a pl asmid encoding T AZ‐ER
T2
fusion protei n. GAPDH is l o ading contr o l.

Figure 46 TAZ ‐ER
T2
rescues the defect of ESC ‐to ‐EpiSC conversion in Wwtr1
‐/ ‐
mouse
ESCs.
Phase contr a st i mages o f Wwtr1
‐/‐
m ouse E SCs not transfect e d (NT) o r transfected wi th p lasmid
encoding TA Z ‐ER
T2
fusio n protein an d cultured u nder CHIR/IWR‐1 condit ion for 5 d ay s.

To i nv estig a te t he f unc t ion of YA P i n mouse EpiSC self‐renewal, I transferred Yap1
‐/‐
mouse
ESCs into ActA/bFGF/CHIR/IWR‐1 culture condition, following the s ame protocol a s for
65

Wwtr1
‐/‐
mouse ESCs. Like Wwtr1
‐/‐
mouse ESCs, Yap1
‐/‐
m ouse E SCs also u nderwent
massive c el l death starting f rom day 2. B oth two individu al c lo nes subjected to c onversion,
2C10 and 3C8, fail ed to give rise to any mouse EpiSCs (Figure 47). N ext, I introduced a Y AP‐
ER
T2
t ransg e ne i nto Yap1
‐/‐
mouse ESCs. The Yap1
‐/‐
;YAP‐ER
T2
mouse ESCs were able to
convert to mouse EpiSCs at h igh effic iency ( Figure 48), in dicat ing that Y AP‐ER
T2
completely
rescued th e ESC‐to‐EpiSC c onversion defect i n Wwtr1
‐/‐
ESCs. Taken together, these results
indica te t ha t YAP is nec essary f or the proper ESC‐to‐EpiSC conv ersio n .

Figure 47 YAP is required for the proper ESC ‐to ‐EpiSC conversion.
Representative phase contrast images of Yap1
‐/‐
mouse ESCs cultured under 2i/LIF or
ActA/bFGF/ CHIR/IWR‐1 conditions for 5 d ays.
66

Figure 48 YAP ‐ER
T2
rescues the defect of ESC ‐to ‐EpiSC conversion in Yap1
‐/ ‐
mouse
ESCs.
Phase contr a st i mages o f Wwtr1
‐/‐
m ouse E SCs not transfect e d (NT) o r transfected wi th p lasmid
encoding TA Z ‐ER
T2
fusio n protein an d cultured u nder CHIR/IWR‐1 condit ion for 5 d ay s.

Cytoplasmic retention of TAZ and/or YAP promotes mouse EpiSC self ‐
renewal in the absence of nuclear β‐catenin
Given tha t c ytoplasmic r eten tion o f TAZ can rescue t he ESC‐to‐E piSC c onversion and self‐
renewal defects of Wwtr1
‐/‐
mouse EpiSCs, I further investigated whether cytoplasmic
ret e nt ion of T AZ i s su ffic ien t to m ainta i n mouse EpiSC self‐ren ewal w ithout e xogenous
cytokines or small molecules. For this purpose, I examined the self‐ r en ewal a bility o f CD1‐
TAZ‐ER
T2
mouse EpiSCs in basal medium with or without CHIR/IWR‐1. In th e presenc e o f
CHIR/IWR‐1, both the CD1‐empty vector and CD1‐TAZ‐ER
T2
mouse EpiSCs remained
67

undiffer ent i ated. In b as al m edium without CH IR/IWR‐1, both C D1‐ empty vecto r a nd C D1‐
TAZ‐ER
T2
mouse EpiSCs differentiated after passaging (Figure 49), indic ating that
overexpres sion o f TAZ‐ER
T2
i s not suffic ien t t o m aint ain mouse EpiSC self‐r en ewal. Since
YAP exerts s imilar e ffects i n r e scuing t he ES C ‐to‐EpiSC convers ion in Yap1
‐/‐
mouse EpiSCs,
I reasoned t hat cytoplasmic retention of b oth YAP and TAZ might b e necess ary t o m aintain
mouse EpiSC self‐r en ewal. For t h is p urpose, I ov erex pressed Y A P‐ER
T2
and TAZ‐ER
T2

simultaneo usly i n CD 1 mouse EpiSCs u sing d ual drug s election. C D1‐YAP‐ER
T2
‐TAZ‐ER
T2

EpiSCs expanded normally in the presence of CHIR/IWR‐1 and reta in ed a n EpiSC iden tit y .
However, in basal medium without CHIR/IWR‐1, CD1‐YAP‐ER
T2
‐TAZ‐ER
T2
E piSCs
differentiated after continuous passaging (Figure 49), like CD1 ‐TAZ‐ER
T2
EpiSCs. These
results ind i cate t hat o v erexpress i on o f TAZ‐ ER
T2
a nd/or YAP‐ER
T2
i s not sufficient to
mainta in m ouse EpiSC self‐ren ewa l .
68

Figure 49 Overexpression of TAZ ‐ER
T2
and/or YAP ‐ER
T2
is not sufficient to maintain
mouse EpiSC self ‐renewal.
Represent a ti ve p h a se c o n trast im ages o f C D 1 mouse EpiS Cs t ransf ect e d with e mpty v ector or
transgenes f or T AZ‐ER
T2
and/or YAP‐ER
T2
a nd c ultured und e r bas a l medium in t he p resence or
absenc e of CHIR/IWR‐1 for 14 days.

To det ermine w heth er c ytoplasmic r etent i on o f TA Z is sufficient for the maintenance of
mEpiSCs, I a lso tested t he s elf‐ren e wal ability of Wwtr1
‐/‐
;TAZ‐ER
T2
mouse EpiSCs in basal
medium with or without CHIR/IWR‐1. Interestingly, Wwtr1
‐/‐
;TAZ‐ER
T2
mouse EpiSCs
remained u ndiffer e nt iated in t h e p resence of C HIR/IWR‐1 but dif fer e nt iat e d in b as a l
medium (Figure 50), indicating that cytoplasmic retention of TA Z alone is n ot suffic ien t t o
mainta in m ouse EpiSC self‐ren ewa l .
69

Figure 50 Cytoplasmic retention of TAZ alone is not sufficient to maintain mouse EpiSC
self ‐renewal.
Represent a ti ve p hase c ontrast i m ages o f Wwtr1
‐/‐
;TA Z ‐ER
T2
m ouse E piSCs cultured under b asal
medium or C H IR/IWR‐1 c ondition for 5 days.

Since nuclear tran slocation of β ‐caten in i s sufficient t o ind u c e mouse EpiSC differ ent i at ion, I
investig at ed w hether c ytoplasmic r et ent i on o f TAZ ca n p r omote mouse EpiSC self‐renewal
in the absence of nuclear β‐catenin. First, I introduced TAZ‐ER
T2
f us ion prot ein into Ctnnb1
‐
/‐
mo u se Epi SC s. Ex pressi on le v e l of T A Z ‐ ER
T2
in Ctnnb1
‐/‐
;TAZ‐ER
T2
E piSCs was assessed by
western blot ( Figure 51). In b asal m edium supplemented with A ct A/bFGF, stab le Ctnnb1
‐/ ‐

mouse EpiSC lines over expr essin g e it her empty vector o r TA Z‐ER
T2
remained
undiffer ent i ated. In b as al m edium without exo g enous cytokines, interestingly, o nly Ctnnb1
‐
/‐
;TAZ‐ER
T2
mouse EpiSCs expanded and retained an EpiSC identity after pas saging ( Figure
70

52). The ad ditio n o f 4‐ OHT in Ctnnb1
‐/‐
;TAZ‐ER
T2
mouse EpiSCs culture in basal medium
induced rapid differentiation. Since YAP exerts similar function as T AZ i n r e gulating m ouse
EpiSC self‐r enew al, I also t ested w h ether cyto plasmic retent ion of both YAP and TAZ has
synergistic effects. I o verexpressed Y AP‐ER
T2
and TAZ‐ER
T2
in the Ctnnb1
‐/‐
mouse EpiSCs
and established stable o verexpre ssion cell line b y dual d rug se lect ion. I n basal medium w ith
or w ithout A ctA/bFGF s upplement, Ctnnb1
‐/‐
;YAP‐ER
T2
;Taz‐ER
T2
mouse EpiSC were able to
expand and retain an EpiSC identity (Figure 53), a phenotype similar to those observed in
Ctnnb1
‐/‐
;TAZ‐ER
T2
mouse

EpiSCs. These resul t s indica te t hat cytoplasmic reten t ion of T A Z
and/or Y AP p romotes self‐renewal a nd i s sufficient t o maintain mouse EpiSC self‐ren ewal
only in the absence of β‐catenin. It is worth noting that in a long‐term self‐renewal a ssay,
both Ctnnb1
‐/‐
;TAZ‐ER
T2
and Ctnnb1
‐/‐
;YAP‐ER
T2
;Taz‐E R
T2
mouse EpiSCs differentiated
gradually i n basal medium a fter 2 0 days o f continuous c ulture a nd p assaging, indicating t hat
cytoplasmic retention of TAZ only partially mimic the function of c yt oplasmic β‐caten in i n
mainta in in g mouse EpiSC self‐r en ewal.
Next, I tested whether c ytoplasmic retention of TAZ can promote mouse EpiSC sel f‐renewal
when nuclear translocation of β‐catenin is blocked by IWR‐1. To this end, I cultured CD1‐
TAZ‐ER
T2
m ouse EpiSC i n basal medium s u pplemented with eith e r CHIR/I WR‐ 1 or I WR‐1
alone. B oth wild‐type 46C and 46C‐ T AZ‐ER
T2
m ouse EpiSCs remain ed u ndifferen t ia ted in t h e
presenc e o f CHIR/IWR‐1, whereas the absenc e of C HIR/I W R‐1 induc ed d ifferentia tion o f
both c ell lines. I n the presenc e o f IWR‐1 only, 46 C‐TA Z‐ER
T2
mouse EpiSCs expanded and
reta in ed a n EpiSC iden tity w hile w ild‐type 4 6 C m ouse EpiSCs d i f fer e ntiated after passag i ng
(Figure 54) . These res u lts indica t e t hat cy to plasmic ret e nt ion of T AZ i s able t o promote
mouse EpiSC self‐renewal only in the absence of nuclear β‐catenin. In a long‐term self‐
71

renewal assay, 4 6C‐TAZ‐ER
T2
mouse EpiSCs differentiated gradually i n basal m e dium o r in
the presen ce o f IWR‐1, r espectiv ely, a fter 2 0 days o f continuou s culture and passaging.
Collectively , t hese data indica te t hat cytoplasmic retentio n of TAZ and/or YAP, in addition to
inhibiting TAZ/YAP target gene transcription, also promotes EpiSC self‐renewal in a β‐
catenin‐ind e pend ent m a nner.

Figure 51 Expression of TAZ ‐ER
T2
fusion protein in Ctnnb1
‐/ ‐
mouse EpiSCs.
Western bl ot a nalysis of T AZ‐ER
T2
f usion prot ein l e vels i n Ctnnb1
‐/‐
m ouse E piSCs trans f ected with
empty vector or plasmid encoding TAZ‐ER
T2
fusio n protein.
72

Figure 52 TAZ ‐ER
T2
promotes mouse EpiSC self ‐renewal in the absence of β‐catenin.
Representative p hase c ontrast im ages o f Ctnnb1
‐/‐
m ouse EpiSCs tr ansfected with e mpty
vector o r plasmid encoding T AZ‐ER
T2
f usion protein and cultured u nder A ct A/ bFGF, basal
medium or basal medium plus 1 µM 4‐OHT conditions fo r 4 days.

Figure 53 YAP ‐ER
T2
and TAZ ‐ER
T2
promotes mouse EpiSC self ‐renewal in the absence
of β‐catenin.
Representative p has e c ontras t images o f Ctnnb1
‐/‐
m ouse E piSCs transfected with p lasmids
encodin g YA P‐ER
T2
and TAZ‐ER
T2
f usion protein, r espectiv ely, a nd c ultured under
ActA/bFGF, basal medium or basal medium plus 1 µM 4‐OHT conditions for 4 days.
73

Figure 54 TAZ ‐ER
T2
promotes mouse EpiSC self ‐renewal when nuclear translocation
of β‐catenin is blocked.
Representative phase contrast images of wild‐type 46C or 46C‐TA Z‐ ER
T2
mEpiSCs c ul tured under
basal mediu m , IWR‐1, or CHIR/IWR‐ 1 conditio n f o r 9 d a ys.

TAZ, but not YAP, is a binding partner of cytoplasmic β‐catenin in human
ESCs under CHIR/IWR ‐1 condition
Next, I determined wh ether YA P a n d TA Z ar e the binding partn e rs o f cytoplasmic β‐catenin
in H9 human ESCs cultured under CHIR/IWR‐1 condition. From β‐ca tenin co‐IP assay
followed by w estern b lot analys is, a strong p hysical int e ractio n was detect ed b etween β‐
catenin and TAZ, but no detectable interaction was found betwee n β‐c a tenin and YAP (Figure
55). These find ings i dent ify TA Z as a b in ding p ar tn er o f cytop l asmic β‐catenin under
CHIR/IWR‐1 condition.
74

Figure 55 TAZ is a cytoplasmic β‐catenin binding partner in human ESCs under
CHIR/IWR ‐1 condition.
Western blot a nalysis of T AZ a nd Y AP i n th e c o ‐immunopr e cipitat ion protei n po ol f or c ytoplasmic β ‐
catenin in H 9 hu man ES Cs culture d under CHIR/ I WR‐1 condi t ion.

Nuclear translocation of TAZ/YAP induces human ESC differentiation
To characterize the role of TAZ and YAP in human ESC self‐renew al, I overexpressed TAZ‐
ER
T2
or YAP‐ER
T2
in H9 human ESCs. In mTeSR‐1 medium, H9‐TAZ‐ER
T2
and H9‐YAP‐ER
T2

human ESCs e xpanded normally a nd f orm e d typical h u man ESC colon ies. mTeSR‐1 is a
serum‐free hESC c ulture m edium su pplemented with b FGF and TGFβ[ 111]. Upon 4 ‐OHT
treatment, H 9‐TAZ‐ER
T2
and H9 YAP‐ER
T2
h uman E SCs underwent rapid differentiatio n
(Figure 56 ), i ndica t in g that n uclear t ran s location o f TAZ or Y AP induce human ESC
differ ent i at ion.
75

Figure 56 Nuclear translocation of TAZ or YAP induces human ESC differentiation.
Representative p hase c ontrast images o f H9‐YAP‐ER
T2
and H9‐TAZ‐ER
T2
human ESCs
cultured u nder m TeSR‐1 (No Treat m ent) o r m TeSR‐1 plus 1 µM 4 ‐OH T c onditions for 3 days.

Cytoplasmic retention of TAZ or YAP promotes human ESC self ‐renewal in
the absence of nuclear β‐catenin
I further tested whether cytoplasmic retentio n of TAZ and/or YA P can maintai n h uman ESC
self‐ren ewa l i n the abs e nce o f g ro wth factors and chemicals by tran sfer rin g H 9 ‐ TAZ‐ER
T2
,
H9‐YAP‐ER
T2
, and H9‐EGFP‐ER
T2
h uman ESCs into seru m‐free N2B 27 basal m e dium[112] .
All of t hree h uman ESC l ines d iffer e ntia ted in t he a bsence o f C HIR/IWR‐1 while self‐ren ewed
properly i n the presen ce o f CHIR/IWR‐1 p lus ActA/bFGF. S urprisingly, in the presence of
CHIR/IWR‐1 only, H9‐TAZ‐ER
T2
and H9‐YAP‐ER
T2
human ESCs w er e able t o expand a n d
76

retain human ESC identity. H9‐EGFP‐ER
T2
human ESCs differentiated after passaging (Figure
57). It is worth noting that in a long‐term self‐renewal assay, H9‐TAZ‐ER
T2
H9‐YAP‐ER
T2

human ESCs g radually d iffer e nt iat e d in N 2B2 7 s upplemented w i th CHIR/IWR‐1 only. These
results suggest t hat cytoplasmic retention of T AZ p romotes hESC self‐renewal in the absence
of nuclear β‐catenin.

Figure 57 Overexpression of TAZ ‐ER
T2
or YAP ‐ER
T2
promotes human ESC self ‐renewal
when nuclear translocation of β‐catenin is blocked.
Re pre s e nta tive pha se contra s t imag e s of H9 human ESCs transf ect e d with plasm ids e ncoding E G FP‐
ER
T2
, T A Z‐ER
T2
, or Y AP‐ER
T2
fusion proteins and cultured in N2B27 medium with or without
ActA/bFGF a n d/or CHIR/ IWR‐1 suppl eme n t for 12 days.

77

Collectively, o bservatio n s described above demonstrate that c yt oplasmic l ocalized β ‐catenin
promotes mouse EpiSC and human ESC through a similar mechanism of retaining TAZ in the
cytoplasm.

78

Chapter 4 Screen for novel human ESC culture conditions
CHIR/IWR ‐1 plus ActA or bFGF are required for serum ‐free culture of
mouse EpiSCs
The discovery of CHIR/IWR‐1 as a novel culture condition for mo use EpiSCs p resents a n
unpreceden ted opport unity to d evelop a n o v el c ulture c ondit i on for currently a vailable
human ESCs. For the po tent ial applications o f h u man ESCs a nd i P SCs in t herap e utic scena rios,
a chemically defin ed c ulture c ondition i s high ly p referr ed. How ever, all the tests regard ing
CHIR/IWR‐1 o n mouse EpiSCs w ere carr ied o ut i n serum‐contain ing m edium. T o facilita t e
the screening and tes t ing of m ouse EpiSCs in a serum‐free condi tio n , I tested t he self‐renewal
ability o f C D1 m ouse E piSCs in s erum‐free N2B27 m e dium u nder v a rious conditio ns.
Surprisingl y , CD1 mouse EpiSCs f ailed to attac h to the culture plate in N2B27 med i um, even
in the presence of CHIR/IWR‐1 plus ActA/bFGF (Figure 58). Never th eless, t he f loatin g cells
could survive and exp a nd i n aggr egat es, ind i cating a defect in attac h ment t o gelatin‐coated
culture surface. T o find a b ett e r culture surfac e for mouse Epi SCs in s erum‐free medium, I
test ed several t ypes o f extrac ellular matr ix ( ECM), including l aminin, fibron ectin, f etal
bovine ser um ( FBS), and Matr ig el . While mouse EpiSCs c ould a tta ch t o every ECM in t h e
presence of serum, only laminin‐ and Matrigel‐coated plate faci litated mouse EpiSC
attachment ( Figure 59) i n serum‐f ree N2B2 7 medium s upplemented with C HIR/IWR‐1 plus
ActA/bFGF. G iven t hat Matrig el i s commonly used i n feeder‐fr ee culture of human ESCs, I
further confirmed that M atrig e l could facilitate t he a ttachment o f CD1 mouse EpiSCs i n
N2B27 medium supplemented w i t h either Act A /bFGF or CHIR/IWR‐1.
79

Figure 58 Serum is required for the proper attachment of mouse EpiSCs on gelatin ‐
coated plate.
Representative phase contrast images of CD1 mouse EpiSCs cultur ed i n serum‐contai n i ng m edium
supplement e d with C HIR/IWR‐1 or s e r um‐fre e m e d ium supple ment e d with ActA/bFGF/ CHIR/IWR‐
1 for 6 days on gela tin‐coated plates.
80

Figure 59 Laminin and Matrigel facilitate mouse EpiSC attachment in serum ‐free
medium.
Represent a ti ve p hase con trast im ages o f CD1 m o us e EpiSCs cult u r ed in seru m‐contai nin g or seru m‐
free medium s upplemented with A ctA/bFGF/C HIR/IWR‐1 on gelatin‐, FBS‐, laminin (LMN)‐, or
fibron ectin ( F N)‐coate d p l ates for 6 da y s.
81

To d etermine w hether C HIR/IWR‐1 is s ufficient to m aint a i n mouse EpiSCs under serum‐free
conditio n, I c arr ied o ut a s elf‐ren e wal assa y using CD1 m o use E piSCs in s erum‐free N2B27
medium supplemented with different combinations of CHIR/IWR‐1 a nd/or ActA/bFGF.
While mouse EpiSCs c ultured unde r N2B27/ActA/bFGF con d ition exp anded normally, those
cultured u nder N 2B27 /CHIR/IWR‐1 conditio n died a fter p assaging (Figure 60) , indic a ting
that C HIR/I W R‐1 is n ot suffic ien t to m aint ain mouse EpiSC self‐renewal in the absence of
serum. Interestingly, the addition of either ActA or bFGF to N2 B27/CHIR/IWR‐1 r escued t he
defec t a nd m ainta i ned normal s elf‐renew a l of m ouse E piSCs (Figu re 6 1). These r e sults
indica te t ha t CHIR/IWR‐1 o nly partially substitute Act A / b FGF in maintaining mouse EpiSC
self‐renewal in serum‐free condition. For the optimal mouse Epi SC sel f‐ren e wal, s upplement
of ActA/CHIR/IWR‐1 or bFGF/CHI R/ IWR‐1 is the minima l requiremen t.
82

Figure 60 CHIR/IWR ‐1 is not sufficient to maintain moue EpiSC self ‐renewal in serum ‐
free medium.
Representative phase contrast images of CD1 mouse EpiSCs cultured in N2 B27 medium
supplemented with ActA/ bFGF o r CH IR/IWR‐1 o n Matrigel‐coated pl ates b e f ore and aft e r passagin g .
83

Figure 61 Supplement of ActA or bFGF on top of CHIR/IWR ‐1 is sufficient to maintain
mouse EpiSC self ‐renewal in serum ‐free medium.
Representative phase contrast images of CD1 mouse EpiSCs cultured in N2 B27 medium
supplemented with CHIR / IWR‐1 plus ActA or bFGF on Matrigel‐coat ed plat e s afte r passaging.

Mouse EpiSCs are sensitive to chemical inhibition of TGF β/Smad and
FGF/MAPK signaling under CHIR/IWR ‐1 condition
Since serum contains a variety of growth factors and cytokines, i t is p ossible that b asal l evels
of T GFβ and/or F GF l igands existing i n serum contribute t o the self ‐renewal u nder
CHIR/IWR‐1 c ondition. To i nvestigate t h i s possibility, I ap plie d sev e r a l chemical i nhibitors
of TGFβ/Smad and FGF/MAPK signaling pathways to CD1 mouse mouse EpiSCs under
CHIR/IWR‐1 o r Act A /bFGF condit ion. A s expec t ed, in s er um‐contain i ng m ediu m
supplemented Ac tA/bFGF, a ddition o f PD03[113] o r A83‐01[114], a TGFβRI inhibitor,
trigg e r e d differ ent i at ion (Figure 62). S u rprisingly, in s erum‐c ontaining mediu m
84

supplemented with CHIR/IWR‐1, both PD03 and A83‐01 induced rapi d mouse EpiSC
differ ent i at ion as w ell (Figure 62) , indic a ting t hat TGFβ /Smad signaling and F G F/MAPK
signal ing co ntribute t o t h e mouse EpiSC self‐renewal u nder t his c ondition. Collectively, t hese
results suggest that CHIR/IWR‐1 only substitute the exog enously supplemented ActA/bFGF
in serum‐co n taining medium f or t he maintenance of mouse EpiSC self‐renew a l.

Figure 62 A83 ‐01 and PD03 blocks mouse EpiSC self ‐renewal under CHIR/IWR ‐1
condition.
Representative phase contrast images of CD1 mouse EpiSCs cultur ed i n serum‐contai n i ng m edium
supplement e d with CHIR / IWR‐1 plus D MSO, A8 3 ‐ 01 or PD 0 3 f o r 3 d a ys.

CHIR/IWR ‐1 plus bFGF and/or ActA are optimal for serum ‐free culture of
human ESCs
A major aim of studies on CHIR/I WR‐1‐mediated mouse EpiSC and h uman ESC self‐ren ewal
is t o develo p a novel a n d robust s erum‐free c u lture condition f or h u m an ESCs and iPS C s fo r
85

both r esea rch and therapeut i c se tt ings. To f acilit ate t h e screen i ng a nd o pt imizat ion of
serum‐free culture conditio n, I t est e d the abilit y of C HIR/IWR‐ 1 to m ainta i n human ESC self‐
renewal in serum‐free medium. H9 human ESCs, routinely propagat ed i n mTeSR 1 m edium,
were s eeded into N 2B27 medium s up plemented with A ctA/bFGF, CHIR /IWR‐1, or b oth.
After continuous passaging, only ActA/bFGF plus CHIR/IWR‐1 cond ition supported proper
self‐renewal of human ESCs (Figure 63). In order to further und ers t a n d the effect o f each
factor o n human ESC self‐ren ewa l i n serum‐free m edium, I seeded H9 human ESCs into
N2B27 medium with different combinations of ActA, bFGF, CHIR, a nd I WR‐1. Interest ingly,
among all the tested c ombinations, o nly ActA/CHI R/IWR‐1 a n d bFG F/C H IR/IWR‐1
supported r obust expansion of h uman ESCs aft e r 3 pas s ages ( Figu re 63). Hu man ESCs i n
bFGF/IWR‐1 condition underwent massive differentiation in the f irst p assage b ut f ormed
several typical human ESC coloni es b y pass age 3, i n dicating a p otential c lonal selection
process. A fter 3 p assages, s light differences b etween A ctA/CHIR /IWR‐1 a nd
bFGF/CHIR/IWR‐1 were a lso observ ed, as c ells f ormed smaller and flatter colonies in
ActA/CHIR / IWR‐1 compared t o those in b FGF/CHIR/IWR‐1. I also p e rformed a 1‐factor
subtraction assay, i n which one factor i s subt racted f rom ActA/ bFGF/CHIR/IWR‐1 in e ach
group, t o determine the crit ical f actors i n this c ombinatio n . A fter p assaging, only
ActA/CHIR / IWR‐1 and bFGF/CHIR/IW R‐1 mainta in ed p roper ES C self‐ re newa l, wi t h ESC
colonies i n ActA/ C HI R/IWR‐1 bein g smaller and flatter. T herefor e, c onsistent results
indica te t ha t bFGF/CHIR/IWR‐1 is t he m in ima l r equired s u pplemen t for optimal culture of
human ESCs in N2B27 medium.
86

Figure 63 CHIR/IWR ‐1 plus ActA or bFGF maintains human ESC self ‐renewal in serum ‐
free medium,
Representative phase contrast images of H 9 human ESCs cultured in N 2B 27 medium sup p lement ed
with ActA/b F GF, CHIR/I WR‐1, or CH I R/IWR‐1 pl us ActA or b FGF for 18 d a ys.

N2B27 is a currently available optimal medium for serum ‐free culture of
human ESCs
In d etail e d comparisons, H 9 huma n ESCs i n N 2 B27/ActA/bFGF/CHIR/ IWR‐1 were s l i ghtly
differ ent to t hose i n mTeSR1, as c olonies i n mTeSR are more c om pact w ith better d e fined
boundaries ( Figure 64). The formula of m TeSR also c onta in s TGFβ a nd F GF2[115], but differs
from N2B27 in many aspects, indi cating t hat the basal m e dium a n d nutritional additives of
N2B27 mig h t contain factors that a ffect human ESC self‐r enew al. N 2B27 medium c onsist o f
DMEM/F12 medium w ith N2 s uppleme n t and N e uralBasal Medium w ith B27
87

supplement[112]. To i n v estigat e t h e e ffect o f N 2 and B2 7 supple ment o n human ESC self‐
renewal, I seeded H9 human ESCs in DMEM/F12 medium or N2B27 med ium, b oth
supplemented with A ctA/bFGF/CHIR/IWR‐1. After passaging, cells forme d c ompact
colonies u nder N 2B 27‐based conditio n but differ e ntia ted un der D MEM/F12‐based
conditio n (Figure 65A ), i ndicat in g that N 2 and B27 supplement i s beneficial t o human ESC
self‐renewal. I also s eeded H9 h uman ESCs in m TeSR basal medium o r N2B27 medium, both
supplemented with m TeSR supplement. After passaging, cells form ed c ompact c olonies i n
mTeSR basal medium w ith mTeS R supplement but different i a t ed i n N2B27 basal medium
with m TeS R supplement (Figure 6 5 B), indica t i ng t ha t N2 a nd B 27 su pplement might conta i n
factors that i nduce differentiatio n . Of n ote, d espite t hat the complete f ormula o f mTeSR has
been p ublished, componen ts i ncluded in m TeSR supplement i s not clearly labeled. T herefore,
it is possible that the overdosing of some shared factors betwe en m TeSR supplement a n d
N2/B27 supplement caused d ifferent ia tion i n the combinatio n of N2B27 medium a nd
mTeSR supplement. C ollectively, t hese r esults d emonstr a te t hat N2 and B27 supplements
are benefic i al t o human ESC self‐renew al w hile sever al f actors that p otent i al ly i nduces
differ ent i at ion exis t.
88

Figure 64 H9 human ESCs cultured in mTeSR medium show typical ESC colony
morphology.
Representative phase contrast images of H9 human ESCs cultured in m TeSR m edium or N 2B27
medium sup plement e d w i th ActA/bF GF/CHIR/IW R‐1 for 9 d a y s .
89

Figure 65 N2B27 is beneficial for serum ‐free culture of human ESCs.
(A) Representative phase contrast images of H9 human ESCs cultu red in N 2B 27 o r DM EM/ F 1 2
medium s up plement e d with ActA/bFGF/CHIR/IWR‐1 for 5 d a ys. (B) R epresentative p h ase contr a st
images of H9 human ESCs cultured in mTeSR basal medium or N2B27 m e d ium supple ment e d with
1X mTeSR supplement for 5 days.

N2 a nd B 27 supplements ( compared i n Table 4[116]) w e re d esig ned for the induction and
maint e na nce of c ells i n the neur al l ineag e . Therefore, t hey con tain s ome fac t ors, s uch as
retinoic a cid (RA), known (or potentially u nk nown) to c ause h uman ESC differentiation. In
an effort to further optimize N2B27 medium as a serum‐free basal medium for human ESC
self‐ren ewa l , I compared t he e ffect s o f sever a l N2/B27 deriva tives, including N2A and N2B
90

derived from N2 as well as B27‐RA, SM1 and SM1‐RA derived from B27, i n DMEM/F12‐based
medium supplemented with ActA/bFGF/CHIR/IWR‐1. H9 human ESCs fo rmed c ompact
colonies i n the combinations o f N 2 /B27, N 2A/B27, a nd N 2B/B27 bu t differentiated i n th e
combinatio ns c ontaining B27‐RA, SM1 or S M1‐RA (Figure 66). Diff erences between N2, N2A,
and N2B are minor, as colonies formed in N2B are slightly more compact than t hose i n N2
and N2A, and therefore negligible. Taken together, N2B27 basal m e dium i s an o ptimal
medium that is currently availab le for serum‐free culture of hu man ESCs.
Table 4 Components of N2 and B27 supplement
N2 and B27 shared B27 specific
Components Amount Component Amount
Insulin 4 µg/ml Catalase 2 .5 µg/ml
Transferrin 5 µg/ml Superoxide D ismutase 2.5 µg/ml
Selenium 0 .016 µg/m l DL‐α‐toco pherol acetate 2 .5 µg/ml
Proges tero ne 0.0063 µ g / m l DL‐α‐toco p h erol 1 µg/ml
Putrescine 16.1 µg/ml Glutathione (reduced) 1 µg/ml
Albumin, bovine 2.5 µg/ml Corticost e r o ne 0.02 µg/ml
Linoleic Acid 1 µg/ml Biotin 0 .1 µg/ml
Linolenic Acid 1 µg/ml L‐C a rnitine 2 µg/ml
D‐(+)‐galactose 15 µg/ml
Retinyl acet ate 0.1 µg/ml
Ethanolamine 1 µg/ml
T3 (triodo‐ 1 ‐thyrosin e ) 0.002 µg/m l
91

Figure 66 N2 and B27 supplement are optimal among their derivatives for human ESC
culture.
Represent a ti ve p hase contrast im a g e s of H9 huma n E SCs cultured under i n dicated co nditions for 10
days.

Small ‐scale screening of chemicals that boost human ESC self ‐renewal
As d escribed a bove, the comparis on b etween m TeSR an d N2B27/A c tA /bFGF/ CHIR/IWR‐1
shows the latter is s ub‐optimal f or h uman E SCs. H owever, the co mparison b etween d iff e rent
serum‐free basal medium available currently indicates N2B27 is optimal. T he c ontras t
between these two comparisons rais es t he p ossibility t h a t addit ional factors or c hemicals
are required for the optimal self‐renewal of human ESCs under
N2B27/ActA/bFGF/CH IR/IWR‐1 c ondi tion. To i dentif y factors and p otentially s ignaling
pathways t hat can boost human ESC self‐ren ewal, I carr ied o ut a small‐scale screen ing of
92

chemicals available in l ab, targ eting a majority o f the s i gnali ng p at hways that h ave been
shown to r egulate ESC self‐ren ew al ( complete l ist of c hemicals in Mater ials a nd Methods).
Among all the tested c hemicals, 53AH, S B590885, F orsko l in ( FSK) , LY294002, a nd W H4‐0‐
23 showed p ositiv e effect o n human ESC s e lf‐ r en ewal, as c ompact c olonies w ith well‐defin e d
b ou n daries rais e d from th e grou ps contai nin g t h e se ch e m i ca l s ( F igure 67). On the contrary,
a group of c hemicals t hat hamper s human ESC self‐renewal w as a l so i dent ified , including
Vertiporfin, S P600125, Go6976, Go6983, J AK I nhibitor I , IKK2 I n hibitor IV, Stat3 Inhibitor,
and CK2 Inh i bitor I V , as h uman E SCs rapidly differentiated o r d ied in g r o ups containing t hes e
chemicals (Figure 68 ). I t is w o r th n ot ing that t he c ell death o bserved during t he
adminis t ra t i on o f thes e chemicals may not necessarily a s s ociate w ith the inhibition o f th eir
targ ets, a s g e neral toxic i ty a nd o ff‐target effects could also cause differ ent i at ion or c ell death.
Additionall y , a third group of c he micals s tood o ut w ith the obs ervation t hat th e majority o f
human ESCs differentiated during the first passage while compac t human ESC colonies
emerged ag ain in l ater p assages (Figure 6 9 ). S uch chemicals inc lude I WP‐2, C59, a nd F H535,
which are all Wnt/β‐caten i n s i g n aling inh i bitors. This u nique p henotype i ndicat es a
potent ial clonal s election p roce ss. A s the act i vat i on l ev el o f Wnt/ β‐catenin s i gnaling is
largely heterogeneous in human ESC culture[66], application of Wnt signaling antagonists
might selec t ively expa nd those w ith low level of Wnt signaling ac tiv a tion.
93

Figure 67 Chemicals that improves human ESC self ‐renewal in N2B27 medium
supplemented with ActA/bFGF/CHIR/IWR ‐1.
Representative phase contrast images of H9 human ESCs cultured in m TeSR m edium or N 2B27
medium sup plement e d w i th ActA/bF GF/CHIR/IW R‐1 plus indicated ch emic als for 12 days.
94

Figure 68 Chemicals that hamper human ESC self ‐renewal in N2B27 medium
supplemented with ActA/bFGF/CHIR/IWR ‐1
Representative phase contrast images of H9 human ESCs cultured in m TeSR m edium or N 2B27
medium sup plement e d w i th ActA/bF GF/CHIR/IW R‐1 plus indicated ch emicals for indicated days.
95

Figure 69 Chemicals that show a potential clonal selection process in N2B27 medium
supplemented with ActA/bFGF/CHIR/IWR ‐1.
Representative phase contrast images of H9 human ESCs cultured in m TeSR m edium or N 2B27
medium sup plement e d w i th ActA/bF GF/CHIR/IW R‐1 plus indicated ch emicals for indicated days.

Small molecule activators of LIF/STAT3 signaling show synergistic effect
with LIF in mouse and human ESCs
The applica t ion of h uman E SCs or i PSCs i n therapeut i c sett ings o f ten requir es a c ulture
system t ha t conta i ns n o animal‐derived c omponent ( xen o ‐free). Therefore, it is beneficial if
a chemical, such a s CHIR, can su bstitute a g ro wth factor, such as Wnt, in culture medium to
achieve the same r egulatory effect o n cer t ain signal ing p a thway s. T o date, the majority o f
chemicals used i n ESC culture are inhibitors t hat serv e as a n a nt agon ist of c er tain sign a lin g
pathways. Among them, some c hemi cals, such a s CHIR, serve as a n agonist by inhibiting a
96

negative regulator of that signaling pathway. Very few chemical a ctiv ators that d irectly serv e
as an agonist has been developed and tested in the context of E SC m aintenance. CRM
compounds are small molecule a ctiva t ors of L IF/STAT3 s ignal i ng p a thway that h ave been
function ally v alida t ed i n chondrocytes, repr es ent i ng t he r are b ut u seful group of c hemical
activ a tors. To i nves tig a te t heir p otent i al a pplicatio n in ESC maintenance, I test ed t he e ffect
of these chemicals on mouse an d human ESC self‐renewal.
In mouse ESCs, LIF maintains self‐renewal in serum‐containing m ediu m via the ac tiva tion o f
LIF/STAT3 signal ing p a thway[21]. To determine t he e ffec t of C RM compounds on mouse ESC
self‐ren ewa l , I carried o ut a c olony forma t ion assay using 4 6 C mouse ESCs i n serum‐
containin g m edium supplemented with e ith e r LIF or C RM c ompounds . While AP+ colonies
were observed in both LIF and CRM compound groups, CRM compound g roups showed
signific antl y few e r colonies tha n LIF group (Figure 70), indica ting that CRM compounds a r e
not as r obust as L IF i n mainta in in g mouse ESC self‐renewal. The a d ditio n of J A K I nhibitor I ,
a JAK inhibitor that b locks LIF/STAT3 signaling[117], resulted in n o AP+ colony i n either L I F
or C RM compoun d gr oup (Fi gur e 7 1 ) , indicating t hat CR M compoun d s promotes mouse ESC
self‐ren ewa l t hrough t he a ctiva t io n of S TAT3 signaling. To t est whether CRM compounds
have synergistic effect with LIF, I also performed colony formation assay in serum‐
containing m edium with d iffer e nt c ombinations of L IF a nd C RM c o mpounds. S urprisingl y,
more AP+ colonies formed in groups with both LIF and CRM compounds than the group with
LIF only (Figure 70), indicating a synergistic effect of LIF an d CRM c o mpounds. T o further
understand the mechanism of by which CRM compounds boost mouse ESC sel f ‐ren ewal, I
looked i n t o the STA T 3 a c tivat i on l evels in 4 6 C ESCs tr eat e d wit h LIF and/or C RM c ompo unds
by western blot. After 30‐min treatment, the combination of LIF and CRM compounds
97

induced a h i gh l evel o f phosphorylation o n T y r 705 of S TAT3 ( pY705‐STAT3) while LIF alone
only i nduced a m oderate lev e l of p Y705‐STAT3 (Figure 72). Similar pattern was observed
for S727 p h osphoryla t ion (pS727‐STAT3), i ndicating that b oost L IF‐mediated m ouse E SC
self‐ren ewa l by increas i ng the a ctivatio n level of STAT3.

Figure 70 CRM compounds show a synergistic effect with LIF on mouse ESC self ‐
renewal.
AP staini n g of 46C m ous e ESCs c u ltured under i ndicated condit io ns for 7 da y s.
98

Figure 71 JAK inhibitor blocks the effect of CRM or LIF on promoting mouse ESC self ‐
renewal.
AP staini n g of 46C m ous e ESCs c u ltured under i ndicated condit io ns for 7 da y s.

Figure 72 CRM compounds show synergistic effect with LIF on STAT3 activation.
Western blot a nalysis of t he l ev els of p hosphorylated STAT o n T yr70 5 (pY 7 0 5 ) a nd Se r 72 7 (pS72 7 )
and total S T AT3 (tS T AT3) i n 46C m o use ESCs t re ated w ith in dicated factors for 30 min after serum
starvatio n o vernight. GA PDH is loadi ng co n trol.

Next, I t es ted whether CRM compounds can m a int a in l ong ‐ term sel f‐renewal of m ouse E SCs.
Both LIF and/or CRM compounds promoted 46C ESC self‐renewal in short‐term c ulture.
99

After continuous p assaging, howe ver, ESCs in g roups that c ontai n CRM compound
differ ent i at ed ( Figure 73). Sinc e CRM compounds can boost the a ctivat ion of L IF/STAT3
signal ing, t his differ en tiat ion phenotype is p otent i ally d ue t o t oxicit y and off‐t a rget e ffect o f
CRM compound. To t his end, I t ested the effec t o f CRM423F, an a nalog of C RM423 with
improved s pecific i ty a nd m inimal t oxicity shown in c hon drocyte, o n mouse ESC self‐ren ewa l .
Interestingly, CRM423F, when used alone, showed a strong ability in promoting ESC self‐
renewal similar to that of LIF in colony formation assays (Figu re 7 4). However, n o obvious
synergistic effect between CRM423F and LIF was observed. In a l ong‐term s elf‐renewal assay,
both C RM423F a nd L IF w ere able t o maintain E SC s elf‐renewal in sho r t‐term c ulture ( Figure
75). After passaging, however, E SCs in C R M 423F g roup d ifferenti ated, indicating t hat
CRM423F i s still sub‐optimal for maintaining long‐term mouse ES C self‐ren ew al. Structural
studies have shown that CRM compo unds a c tivates STAT3 signaling by directly binding to
glycoprote in 130 (gp130) /LIF re c e ptor (LIFR) on ce ll su rf a ce . I n ano t her s e lf‐renewal a ssay
using Hyg34, a L IFR
‐/‐
m ouse E SC l ine, n either L IF n or C RM423F w as a ble to p romote E S C
self‐renewal ( Figure 7 6), confirming t hat CR M423F p ro motes self ‐ r en ewal t h r ough L IFR‐
mediat ed a ctivat ion of S TAT3 s ign a ling.
100

Figure 73 CRM compounds cannot maintain long ‐term self ‐renewal of mouse ESCs.
Represent a ti ve p h a se c ontrast images o f 46C m ous e E SCs cultured i n bas a l medium s up plement e d
with LI F or C R M comp ou nds for 16 days.

Figure 74 CRM423F show strong effect on promoting mouse ESC self ‐renewal.
AP staini n g of 46C m ous e ESCs c u ltured under i ndicated condit io ns for 7 da y s.
101

Figure 75 CRM423F is able to maintain short ‐term self ‐renewal of mouse ESCs.
Representative phase contrast images of 46C mouse ESCs cultured basal medium supplemented with
indicated factors for 6 days.
102

Figure 76 The Effect of CRM423F on mouse ESC self ‐renewal is dependent on LIFR.
Represent a ti ve p h a se c ontrast images o f Hyg34 (LIFR
‐/‐
) m o use ES Cs c ultured bas a l m e diu m
supplement e d with indicat ed factors for 6 d ays.

LIF/STAT3 signal ing h a s minimal effect o n ra t ESC sel f ‐r enew al. Since CRM compounds are
able t o boost LIF/STAT3 signaling in mouse ESCs, I also explore d their e ffect on r a t a n d
human ESC self‐ren ew al. Rat ESCs r equire s erum‐free c o ndit ion f o r p roper self‐renew al.
Since BMP4 i s the facto r i n serum that c ooperates w ith LIF in mainta i ning m ouse ESC self‐
renewal, I t ested CRM compounds in N 2B27 medium s upplemented with BMP4 in a colony
formatio n assay. A P staining a ft er 14 days o f culturing (Figure 77) showed that 2i is the
optimal conditio n for r a t ESC self‐r enewal. Add i tion o f CRM compounds on top of 2i reduced
the number of AP+ colonies while addition of LIF/BMP4 on top of 2 i has negligible e ffect,
indica ting t he p otent i al t oxicit y an d off‐tar g et e ffect of C RM compounds. I n the ab sence of 2 i,
LIF/BMP4 with or without CRM compounds cannot maintain rat ESC self‐renew al.
103

Collectively, these results shows that CRM compounds are not be ne f i ci al t o rat ESC se lf ‐
renewal. Human ESCs are primed state PSCs in which LIF/STAT3 si g n aling is n o t a ctiv ated .
Increa sed s e nsitiv ity t o L IF i n human ESCs, however, h ave been shown to associated with
some naïv e ‐like proper ties.

Figure 77 CRM compounds show negative effects on rat ESC self ‐renewal.
AP staini n g of D Ac8 r a t ESCs culture d in N2B 27 medium und er ind icated conditions fo r 14 days.

Small molecule activators of TGF β/Smad signaling promotes mouse
EpiSCs and human ESCs
Inspir ed b y the effect o f CRM compounds, I a lso invest igat ed t h e effect o f small molecule
activ a tors o f TGFβ/Smad signaling in t he c ontext o f mouse Epi S C and human ESC self‐
ren e wal. C onven t ion a l culture condit ion for mouse EpiSCs a nd h u man ESCs r equires
simultaneo us a ctivat ion of T GFβ/Smad signa l ing and FGF/MAPK sig naling[5, 6, 16]. G4 i s a
small molecule T GFβ agonist, d ev eloped a nd t est e d in c hondrocyt es, that directl y binds to
104

TGFβR on the cell surface. To test their effect on mouse EpiSC self‐r enew al, I ca rried o ut a
long‐term self‐ren ewa l a ssay using E3 m ouse E piSCs in s eru m‐con taining mediu m
supplemented with bFGF. After culture and passaging for 11 days , typical compact EpiSC
colonies were observed in both ActA and G4 groups, with various c oncentrations of G 4, w hile
differ ent i at ed c ells w er e observed i n bFGF o nly group (Figure 7 8). After 29 d a ys o f culture,
typical compact EpiSCs remained in the groups containing ActA, 10µ M G4, or 1 µM G 4 (Figure
79), indicating that G4 is able to substitute ActA in maintaining long‐term self‐renewal of
mouse EpiSCs. To u nderstand the mechanism by w hich G 4 promotes self‐ren ew al, I looked
into t he a c t iva t ion l e v e ls o f Smad2/3 in E 3 EpiSCs t r e ated w ith ActA or G4. After 1h
treatment, both ActA a nd G 4 were able to trigger phosphorylation of Smad2 (p‐Smad2) in a
dose‐depen dent m ann e r (Figure 80A). H owever, p‐Smad3 was only o bserved in A ctA ‐
treated sample but not in G4‐treated samples (Figure 80A), indi cating t hat G4 i s selectiv ely
activ a ting S mad2 i n mouse EpiSCs. Simila rly, s elect i ve a c t iv ati on of Smad2 wa s a l so
observed i n CD1 EpiSCs i n a time‐depend e nt m anner (Figure 80B), f urther c onfirming the
potency and selectivity of G4 in mouse EpiSCs. To expand the di sc overy from G 4 to h uma n
ESCs, I also carried out a self‐renewal assay using H9 human ES Cs i n N2B2 7 medium
supplemented with b FGF/CHIR/IWR‐1 . After 1 0 days of c ulture, ty pical compact human ESC
colonies were observed in groups containing 5µM or 1µM G4 (Figu re 81), indicating that G4
is a ble to substitute Act A i n mainta ining human ESC self‐r enew a l. C ollectively, t hese r esults
demonstrate that G4, as a TGFβ agonist, is able to promote mous e EpiSC and human ESC sel f ‐
renewal through selectiv e a cti va ti on of Sma d2.
105

Figure 78 G4 promotes short ‐term self ‐renewal of mouse EpiSCs in the presence of
bFGF.
Represent a ti ve p hase c ontrast images of E3 mouse EpiSCs culture d in b asal m edium sup p lement e d
with indicat ed fact o rs fo r 11 days.
106

Figure 79 G4 promotes long ‐term self ‐renewal of E3 mouse EpiSC in the presence of
bFGF.
Represent a ti ve p hase c ontrast images of E3 mouse EpiSCs culture d in b asal m edium sup p lement ed
with indicat ed fact o rs fo r 29 days.
107

Figure 80 G4 selectively activate Smad2 in a dose ‐ and time‐dependent manner.
(A) Western blot a nalysis of p hosph o rylated S m ad 2 (p‐Smad2), t o tal S m ad2, p hosphoryl a ted Smad 3
(p‐Smad3), and total Samd3 levels in E3 mouse EpiSCs treated with indicated factors for 1h after
serum star vation overni ght. (B) W e stern blot a nalysis of p hosph orylated S mad2 ( p‐Smad2), total
Smad2, phosphorylated Smad3 (p‐Smad3), and total Samd3 levels i n CD1 mouse EpiSCs t reated w ith
indicated factors for i n dicated ti me aft er serum starvation overnight.
108

Figure 81 G4 promotes human ESC self ‐renewal in N2B27 medium supplemented with
bFGF/CHIR/IWR ‐1.
Representative phase contrast images of H 9 human ESCs cultured in N 2B 27 m edium sup p lement e d
with indicat ed fact o rs fo r 10 days.

In addition to G4, I also examined the effects of several G4 an alogs on h uman E SC s elf‐
renewal. T h e se a nalogs h ave shown stronger e ffect on a ctivating T GF β/Smad signaling with
lower toxicity in chondrocytes. I carried out a self‐renewal assay of H9 human ESCs using G4
and its anal ogs in N 2B27 medium s upplemented with b FGF/CHIR/IWR‐1. After 11 days o f
cu l t u r e, h um an ESC col on ies wer e ob served i n grou ps contai n i n g G4, CIM61, or B13 (Figur e
82). Taken together, these observations suggest that G4 is the optimal small mo l e cule T GFβ
agonis t for mouse EpiSC and human ESC culture, w hereas C IM61 an d B13 are als o suitable
for human ESC cultur e as alter na tives.
109

Figure 82 G4 analogs promotes human ESC self ‐renewal in N2B27 medium
supplemented with bFGF/CHIR/IWR ‐1.
Representative phase contrast images of H 9 human ESCs cultured in N 2B 27 m edium sup p lement e d
with indicat ed fact o rs fo r 11 days.

110

Chapter 5 Discussion and perspectives
Current model for CHIR/IWR ‐1 ‐mediated mouse EpiSC and human ESC
self ‐renewal
Conventional c ulture c ondition f or m ouse E piSCs and currently a vailable h uman E SCs
requires su pplement of A ctA a n d bFGF i n ser u m‐contain ing mediu m. A recent study from
our lab dis c overed t h a t supplement of C HIR and IWR‐1 can substitute ActA and bFGF in
serum‐containing medium to maintain mouse EpiSC and human ESC s elf‐renewal [68]. Such
effect o f CHIR/IWR‐1 o n maint a in ing s e lf‐renew al i s mediat ed b y the cytoplasmic
localization of β‐catenin, which is realized by simultaneous in hibition o f GSK3β by C HIR and
stabilization of Axin2 by IWR‐1. Cytoplasmic stabilization of β ‐catenin m aint ains m ouse
EpiSC and human ESC self‐ren ew al t hrough t wo m echanisms: t he i n hibition o f Wnt/β‐
caten i n tar g et g enes, which blocks differ e n t iat i on, a n d the cyt opla smic r egula t ion of i ts
binding par t ner s , whic h promotes s elf‐renew a l. T A Z i s a β‐caten in b inding p ar tner i n the
cytoplasm of mouse EpiSCs and human ESCs under CHIR/IWR‐1 condi tion. Nuclear
localization o f TAZ and the subsequent a ctivat ion of H ippo t arg et g enes l eads to
differ ent i at ion. C y t oplasmic r et en t i on o f TAZ promotes m ouse E p iSC and human ESC sel f ‐
ren e wal in a β‐caten in‐ i ndep enden t m ann e r via two poten t ial mec han i sms: t he i n h ibition of
Hippo t arg e t genes that i nduce differentiatio n , and/or t he m odu lation o f other yet unknown
factors tha t r egulates pl u ripotency (Figure 83).
111

Figure 83 Current model for CHIR/IWR ‐1 mediated primed state PSC self ‐renewal.
In the presence of CHIR and IWR‐1, β‐catenin is stabilized, by inhibition of GSK3 in t he β ‐catenin
destruction c o mplex, a nd r etained in t he c ytoplas m , by e levated level of Axin2. In addition to blocking
β‐catenin‐dependent tran scription, c ytoplasmic β ‐c ateni n a lso retains its binding partner, TAZ, in the
cytoplasm, p romoting s el f‐renew a l through inhibition of H ippo target genes as well as other yet
unknow n mechanisms.

CHIR/IWR‐1‐mediated mouse EpiSC and human ESC self‐renewal is s till dependent on
T GFβ/Sma d a n d/or FGF/MA P K si gn a l in g, t hough on ly a ba sa l le ve l of a ctiva t io n is r equired.
Basal levels of TGFβ and FGF ligands in serum under serum‐conta ining condit io ns, as w ell as
exogenously supplement e d Act A o r bFGF u nder serum‐free conditio ns, fulfill thi s
requirement. CHIR/IWR‐1, through cytoplasmic stabilization of β ‐catenin, la rg ely reduces
the need for exogenous TGFβ and FGF ligands but does not abolish the depen dence o n
TGFβ/Smad and FGF/MAPK s ign a ling. Two p o tent ial mechanistic ex p lanatio n s exis t. O n on e
hand, cytoplasmic β‐catenin or i ts b inding p artners mig h t elevate the activation levels of
112

TGFβ/Smad and/or F GF/MAPK signaling to c omparable levels a s those under ActA/bFGF
conditio n, r epresent ing a model th at CHIR/IWR‐1 and ActA/bFGF c on verge on the s ame set
of sig nalin g p athway s and ta rg et g en es t o maint a in self‐renew a l. On the other hand,
cytoplasmic β‐catenin or i ts b inding p artner s might modulate k e y proteins, in signaling
pathways o ther t han TGFβ/Smad and FGF/MAPK, that c ontribute to pluripotency
maint e na nce, r epres e n t ing a model that C HIR/IWR‐1 and Act A /bFGF u tilize p arallel
me cha n i sms t ha t con t rol di st in ct groups of si g n a li n g pa t hwa ys and t a r get genes t o m aint ain
self‐ren ewa l . Therefore, it is o f great interest t o analyze the binding partn ers of cytoplasmic
β‐catenin, as revealed by Mass Spec analysis (Table 3), and ide n t ify key proteins th at m edia te
self‐ren ewa l u nder C HIR/IWR‐1 condit ion. I n add i t i on, comprehen sive a nalysis o f
differ ent i al g en e exp r essio n p r o files und e r Act A /bFGF and CHIR/ IWR‐1 condit io ns
(GSE31461)[68], as r eveal e d by R NA‐seq a nalys i s, a nd i dentifica tion o f k e y signaling
pathways a nd b iologic a l processes highly r ep resented i n these p rofiles, c ould a lso provide
insight f ul in f ormatio n .
The role of cytoplasmic β‐catenin in mouse EpiSC and human ESC self ‐
renewal
Wnt/β‐catenin s i gnaling pathway has been ex t ens i vely st u died in th e con tex t of h u m a n ESC
self‐renewal[56‐61, 64, 65]. Accumulating e v i denc es s ug gest t ha t Wnt/β‐catenin signaling
exerts opposing e ffects b y promoting both self‐renew a l an d diff er ent i ation o f mouse EpiSCs
and human ESCs. This i s consist e n t w ith the observat ion that t w o groups o f distinct W nt
targ et g enes are induc ed upon Wn t/β‐catenin sign aling activ a tio n i n m ou se ESC s, for m i ng a
balance between sel f‐r e new a l and differ ent i at ion. T herefore, to a chiev e optimal self‐renew al,
113

it i s critical to shift the balance towards the g r oup of g enes that m aintain pluripotency. This
can be r ealized by t wo a pproaches, fine‐tun in g the modulation o f Wnt/β‐catenin pathway
and simultaneous regulation of Wnt and other signaling pathways . Current o bservations
suggest that CHIR/IWR‐1 mediate mouse EpiSC and human ESC self‐ renew a l through bot h
approaches. On one hand, CHIR and IWR‐1 keep Wnt/β‐catenin signaling in a u nique st atus
that β‐catenin is neither degraded nor actively regulating transcription. The concentrations
of CHIR and IWR‐1 represents a carefully set balance. Higher CH IR c oncentration o r lower
IWR‐1 concentr at ion leads to β‐c at enin t arg e t gene a ct iva t ion, which leads to dif ferentiation,
whereas lower CHIR concentration or higher IWR‐1 concentration leads to e nhanced
degrad at ion and ev entually l oss of c ytoplasmic s tabilized β‐cat en in, which also l eads to
differ ent i at ion. O n the other hand, cytoplasmic β‐catenin, i ns tea d of r egulating Wnt/β ‐
caten i n sig n aling path way, i nt era c ts with a v a riety o f cy t oplas mic protein s to r egulate other
signaling pathways that are critical for mouse EpiSC and human ESC sel f ‐ren ewal. TAZ is o ne
example of β‐catenin b inding p a r tn ers in t he c y t oplasm. Retent ion of TAZ by β‐catenin in the
cytoplasm not only i nhibits Hippo t arget genes to b lock d iffer entiation but also positively
regulates self‐renewal through a currently unknown mechanism. T he f act that c ytoplasmic
ret e nt ion o f T A Z a nd/o r YAP can n o t fully r ec a p itulate the funct io n of c ytoplasmic r et en tion
of β‐catenin i ndicat es t h a t other crit ical b inding p artn ers of β‐catenin also e xist i n t h e context
of CHIR/IWR‐1‐mediated self‐renewal. The observation that cytop lasmic r eten t i on o f TAZ
cannot control the subcellular lo calization o f β‐catenin furthe r suggests other β‐catenin
binding pa r t ner s work in par allel or upstream of TAZ.
Canonical Wnt signaling pathway often s e rv es a s a parad i gm f or the mechanist i c studies o f
signal ing pathways on the molecular level. The mode of action o f β‐catenin in c a n onical W nt
114

signal ing p a thway, i . e . regulation o f protein stabilit y in c yto plasm and activation o f
transcription in the nucleus, has been the “central dogma” for the functionality of v arious
factors inv o lved i n sig n al t r a nsdu ction. Ev i dences i n this stud y shows that c ytoplasmic β ‐
catenin, which has been dismissed previously as “proteins to be d egraded” o r “proteins to b e
translocat ed”, c ould p lay import ant roles in s elf‐renewal through a novel mode of action. In
addition t o blocking W nt t arget genes, t he l oc alizat ion o f β ‐catenin in the cytoplasm is crucial
to self‐renewal as β‐catenin translocation into the nucleus wit hout a ctiv ation o f W nt t ar get
genes r e sulted i n rapid differ ent i at ion. T he f un ction of c yt opl asmi c β‐cate ni n on m ai ntai ni ng
self‐ren ewa l i s consistent w ith and further extends the underly ing mechanism of r ecen t
observations t hat Wnt inhibition p r o motes mouse EpiSC and human E SC s elf‐renewal[66‐69,
118]. My f inding t hat th e outcome o f β ‐catenin s tabilizatio n , e ither by W nt l ig and stimulation
or G SK3 in hibition, is dictat e d by t he l ocalization o f β‐cateni n with in t he c ell, p otentially
explains s everal p revious controversial reports[56, 60, 6 1, 64‐ 67]. This n ovel m ode of a ctio n
of β‐catenin opens up a new avenue for future studies on the Wn t sig n aling pathway by
illuminating how β‐catenin inter a c t s with a nd r egulates f a c tors in th e cytoplasm.
Pleiotropic role of TAZ and YAP in mouse EpiSC and human ESC self ‐
renewal
The role of Hippo signaling pathway has been implicated in controlling the fate choice
process of various types of stem cells[77‐79]. Although several r eport s h ave shed l ight o n its
role i n regu lating P S C self‐renewal [83‐86, 90, 119], espec i ally human ESC self‐renewal[87,
88, 108, 1 20], conclusions from t h e m are not consistent ‐ e ven paradoxical in s ome cases.
This m ight b e due to t h e c omplexit y of t he H ippo signalin g p ath way, i ncluding i t s u pstream
115

stimuli, m ediat i ng f act o rs, and va rious regul a tory m ech a nisms. Using genetic and cellular
approaches, I have directly demonstra t ed t h a t nuclear translocation of TAZ or YAP and
subsequent activation of Hippo target genes, independent of ups tre a m Hi ppo si gn a l in g
regulations, induce differentiation in mouse EpiSC and human ES Cs. The cy toplasmic
ret e nt ion of T AZ a nd/ o r YAP, i nd epend e nt o f other reg u latory f a ctors, p romotes mouse
EpiSC and h u man ESC self‐ren ewal. Observat io ns i n th is st u dy t h erefo r e est a blishes TA Z and
YAP as critical regulators of primed state PSC self‐renewal. Th e differen t iatio n i nducin g
effect o f nuclear TA Z/ YAP is l ikel y attr ibuted t o activat i on o f Hippo pathway target genes,
among which many a r e k ey r egulators in l in eage s p e cification. T he s elf‐renewal promoting
effect o f cyt o plasmic TAZ/YA P can be p artially e xplain ed b y inh ibit ion of H ippo t arget genes.
As cytoplasmic retention of TAZ/YAP is able to promote short‐te rm s elf‐renewal of m ouse
EpiSCs and human ESCs, it is possible that, in addition to inhibiting Hippo target genes,
cytoplasmic TAZ/YAP interact w it h other cur r ently unknown factors to positively regulate
self‐ren ewa l .
Sever a l previous r ep orts h ave established TAZ as a s elf‐renewal f actor as n uclea r
translocat io n of T AZ p r o motes human ESC self‐ren ewal u nder T GFβ / F GF2 condition[87, 88].
Nuclear TA Z, i n respo n se t o TGFβ s timulation, binds to S mad2/4 complex and enhance
TGFβ/Smad2 signaling by both promoting the nuclear retention of S mad2/4 c omplex[87]
and triggering the interaction among Smad2/4, Oct4, and TAZ to mediate the selectiv e
activ a tion o f TGFβ/Smad2 t arget genes that p ro motes self‐renewa l[88]. However, i t is w orth
noting that, nuclear translocation of TAZ, as well as the interaction between TAZ and
Smad2/4 complex, o n l y happens w hen Sma d 2 is p hos p horylated (pS m ad2) u pon TGFβ
signal ing activation[87]. In m ouse E piSCs under CHIR/IWR‐1 c ond it ion, p Smad2 is n ot
116

detectable, indicating t hat TGFβ signaling is not massively act iv ated. In t his scenario, my
observations suggest that TAZ is retained in the cytoplasm via int e rac t ion with c yt oplasmic
β‐catenin. F orced nuclear t r ansloca t i on of T A Z u n d e r C HIR/IW R‐1 c ondition, instead o f
promoting TAZ/Smad 2/4 interac t io n, c aus e s the as sociat ion betw een TAZ and TEAD to
activ a te H ippo t arget genes that p romotes differ ent i at io n, sugg esting t hat th e outcome of
TAZ nuclea r translocat ion is l argely c ulture c ondit i on‐ and sig nal i ng c ontext‐d e pend ent.
Observatio ns from the cytoplasmic retentio n of TAZ and its roles in promoting mouse EpiSC
and human ESC self‐renewal i n th e absence o f T GFβ under CHIR/I W R‐1 condition further
expand ed o ur c urrent u ndersta n d i ng o n the r e gulatory e ffec t s of TAZ in primed state PSC
self‐ren ewa l .
Dispensable roles of TAZ and YAP in mouse ESC self ‐renewal
In addition to primed state PSCs, YAP has also been established as an essential factor for
mouse ESC self‐ren ewa l u nder L IF /serum o r LIF/BMP4 c onditio n s. Knock‐down o f YAP in
mouse ESC s c ause differentiatio n [ 83], wherea s activat i on o f YAP b y Yes downstream o f LIF
signal ing p r omotes m ouse ESC s elf‐ren e wal [84]. The fa ct t hat Yap1
‐/‐
, as w ell as Wwtr1
‐/‐
,
mouse ESCs c an b e est a blished an d maint a in ed p roperly under 2i/ L I F conditio n, h owever,
demonstra t es t ha t YA P and TA Z ar e disp ensab l e for mouse ESC self‐renewal. In support of
my observations, a recent report [86] has also revealed the disp ens a bility of YAP and TA Z in
mouse ESC self‐renewal. As a matter of fact, Y AP and TAZ are no t activated in I CM b ut i n TE
during t he f ormation o f the blastocyst. D i srupting L AST1/2 a c t i vity i n a developing
blastocyst l eads to i ncr e ased n uclear l ocalizat ion of YAP/ T AZ a nd c o n sequent in crease o f TE
specification. Consistently, simultaneous loss of YAP and TAZ i n early embr yos causes
defects in cell fate determination and embryonic lethality at t he m orula stage[81].
117

Converging e vid e nc es f rom in vitro and in vivo studies suggest that YAP/TAZ are beneficial
but not required f or t he m aint ena n ce o f naïv e pluripotency. It is l ikely that t he i n dispens able
roles of YAP described in previous studies are specific to LIF/ serum culture condition.
Compariso n o f the requirement of YAP/TAZ between different mouse ESC culture conditio n s
also i mplies t hat mouse ESCs c ultured in 2 i/LIF better mimic th e pluripotent cells in vivo in
the context of Hippo signaling r egul ation.
Insights from small scale chemical screening on human ESCs
Comparison between ActA/bFGF and ActA/bFGF plus CHIR/IWR‐1 cond itio ns i n serum‐free
medium reveals the high potency of CHIR/IWR‐1 in boosting mouse EpiSC and human ESC
self‐renewal. However, the failu re o f maint a in ing mouse EpiSC a nd h uman ESC self‐renewal
by C HIR/IWR‐1 alone under feeder ‐free s e rum‐free conditio n impl ies t he c ooperatio n
between CHIR/IWR‐1 and other factors present in serum or feeder ‐condit i o n ed m ed ium.
Addition o f ActA a nd/or bFGF o n top of C HIR/IWR‐1 enables mouse EpiSC and human ESC
maint e na nce under feeder‐fr ee s e r u m‐free condit ion, es t ablishin g ActA/bFGF/ CHIR/IWR‐1
as a p ot en tially r obust combina t ion for th e dev e lopment of n ove l culture conditions.
Nevertheless, c omparison between mTeSR and N2B27/ActA/bFGF/CHIR /IWR‐1 s uggests
that o ther f actors t hat contr i bute t o optimal human ESC maintenance a lso exist.
N2B27/ActA/bFGF/CHIR/IWR‐1 c onditio n t rigg ers higher g row t h rat e and better
selectiv ity on self‐ren e wing v s differentiating colonies, where as m TeSR maintains compact
colony m orphology and optimal cell viability. I t will be o f gre a t b en efit i f further o p timizat i on
of human ESC culture condition could combine the advantages of both m TeSR and
N2B27/ActA/bFGF/CHIR/IWR‐1 by achieving higher growth rate and better cel l v iability,
118

especially f rom low density s eed ing or e ven single c ell, w hile ret a ining the ty pical colony
morphology and high s e lectiv ity o n self‐renewal vs different ia t ion.
The dev e lopment of n ovel c ulture c ondit i on c an b e divided in to tw o aims, optimization o f
basal medium a nd s creen ing for novel combinat ions o f supplemental factors. On one hand,
N2B27 is the currently available o ptimized s erum‐free medium f o r human ESC culture.
However, components in N2B27 include several differentiation fa ctors and potentially
redundant factors. C areful c omparison between N 2B27 and o t her c hemically d ef ined
medium ( CDM), such a s TeSR‐E8[121 ], could bring ins i ghts f or f u r t her optimization a nd
simplification o f basal medium f rom N2B27. O n the other hand, s ear c h for novel
combinatio ns o f factor s that o ptimally m aintain human ESC self‐ renewal can be f acilit ated
by c hemica l library screening us ing a reliabl e r eporter line, such as Oct4‐ or Nanog‐GFP.
A lt h ough s ma ll sca le , t he scre e n i n g of che mi ca ls a va i la b le i n l ab p rovided several promising
hits t hat morphologically i mprove s human ESC self‐renewal u nder
N2B27/ActA/bFGF/CH IR/IWR‐1 c ondi tion. B a sed on c urrent i nform a t ion, f urth er t itr a t i on
of o ptimal c oncentrations and trails for different combinations would lead to a better
combinatio n of s upplemental fact ors. I t is a lso possible that m ore rounds of c hemical library
screening based on the optimal factors in the previous round co uld accelera te t he
optimization p rocess. I n addition , the trails o f small mole cule a ctivators for LIF/STAT3 and
TGFβ/Smad signaling open u p a new av en ue f or t he d evelopment o f c hemically defin ed
culture conditio n. A lthough seve ral small mo lecule a ctivators, especially C RM423 and G4,
are highly potent in promoting PSC self‐renewal, the underlying t oxicity and potent ial off‐
target effect makes them sub‐optimal for long‐term maintenance of PSCs. Also, as FGF/MAPK
signal ing is a lso a crit ical p athway f or h uman ESC m aintenanc e , i t will be b eneficial to
119

develop a small molecule activator of this pathway with high sp ecificity. A ft er a p roof‐of‐
concept, t hese s mall molecule a ctivators r e quire fur t her chemic al m odifica t ion and tes t ing
before optimal application in human ESC culture.
Conclusions and perspectives
In summary, my study establishes cytoplasmic localization of β‐ cate ni n as t he c e n tral
mediator o f mouse EpiSC and human ESC self‐ r en ewal u n d er C HIR/I WR‐1 condition. This i s
realiz ed b y not only i nhibit ion of W nt/β‐catenin t a rg et g enes b ut a lso cytoplasmic
int e ract ion between β‐caten i n a n d TAZ. My data demo n stra tes th a t, u nder C HIR/IWR‐1
conditio n, T AZ p romotes s e lf‐ren ewal w hen ret a in ed i n the cy top lasm b y β‐catenin and
induces differentiatio n when t r a ns located into t he n ucleus. Var ious m echanisms by w hich
TAZ promotes s elf‐ren e wal may exist, i ncluding t he i nhib ition o f Hippo target genes and t he
int e ract ion with o ther c ytoplasmic p roteins. I n addition t o TAZ , oth e r binding p a rtners o f
cytoplasmic β‐catenin may exist, which awaits further investiga tion t o reach a
comprehensive under s tand ing of t he m olecular m echanisms b y w hic h CHIR/IWR‐1
mediates mouse EpiSC and human ESC self‐renewal. Based on CHIR/ IWR‐1, m y observations
reveal ActA/bFGF/CHIR/IWR‐1 in N2B27 medium on Matrigel as the optimal combination
f or se r u m‐f r e e cult ur e of huma n ESC s. Re su lt s f r om a small‐sca l e chemical screening a s well
as t rails on sever al small molecule a ctivat ors of L IF/STAT3 a nd T GFβ/Sma d signaling
provide hin t s for the further op tim i zation of a novel human ESC culture conditio n .
Over t he p a st decad e s, v arious c ombinations of m edia, c y tokin e s, growth factors and small
molecules have b een e xplored to e stablish p luripotency in c ultu re. Distilled from t hese
empirical observat ion s, mediat in g factors and signal ing pathway s important to t he
120

maint e na nce of p luripotency ar e t h en i den t ified and in tensiv ely studied, with the hope of
fully elucidating the molecular mechanism that governs pluripot ent stem c ell fate c hoice.
From facts and clues collected thus far, it is well defined to conclude t hat the m e chanism s
govern ing t h e na ïve pluripotent st ate and the primed p luripoten t state are largely distinct ‐
even o ppo site i n a n u mber o f cases. H owever, th e mechanisms g ov e r ni ng t he same
pluripotent state, e ith e r na ïve or p rimed, a re l argely c onserv e d among species. These
observations n ot o nly reflect th e essential molecular mechanism s for pluripotency c ontrol
but also provide invaluable insights for the future optimizatio n of c ulture c onditions in
clinical settings as well as the potential attempts to derive n ew c ell types from o ther species .
Rather than simply switching signals on and off, pluripotency i s controlled by a r egulatory
networ k fo rmed f rom the int e ra ction betw een v ariou s signaling p athways[1 22]. Known
and/or unknown factors are not likely working alone, but rather intersect and cooperate
with e ach other to a chieve the o v ert effect o n pluripotency. Th er efore, l arg e ‐scale h igh‐
throughput a nalys i s of g ene expres sion p rofile a nd e pig e n e tic s ta tus as w ell as screen ing for
cDNA, guid e RNA or c hemical co mp ound l ibraries w ill be d esirable ways to generate a
genome‐, e pigenome‐, t ranscriptome‐, a nd p roteome‐wide v iew o f t h e pluripotency n etwo rk
for future s tudies. Mor e over, a n al ysis o f gen e e xpress ion profile and regulatory network in
early stage embryos[123‐125] w il l serve as a valuable guide and reference for the search of
naïv e a n d primed st a te p luripotency in vitro .
Despite the rapid prog ress in b et t e r understa nding the p l uripot ency a nd c ell fat e c hoice in
recen t y ea r s, several c r itic al q uestions a nd c hallenges r emain to b e further addressed. I f
iPSC‐based c ell‐replacement ther apies a re t o be r ealiz e d, i t is o f great importance t o
121

establish s a fe a nd r o b ust cultur ing methods for patient‐deriv e d i PSCs. If g enetically
manipulated animals are to b e utilized a s diseas e models, it i s a lso critical t o develop
methods to d er ive an d maint a in P SCs from l ives tock a nimals a nd non‐human primat es.
Discoveries made t hus far from m ouse, rat and human PSC studies indicate that the core
factors and molecular basis for pluripotency m aint ena n ce a re c o mmonly shared a mong
differ ent cell types and species. A better understand ing o f h ow p luripotency is e stablished
and main ta ined w ill provide a w i der insight for future p rogress in both scientific and
translational applications.

122

Materials and Methods
Cell culture
The medium for CF1 and DR4 feeders (MEF medium) was DMEM (Invit rogen) s upplemented
with 10% fetal bovine serum (HyClone). The basal medium for routine maintenance of
mouse ESCs and mouse EpiSCs was DMEM supplemented with 10% feta l bovine s erum, 1%
non‐essential amino acids (Invit rogen), 2mM L‐glutamine (Sigma) , 1mM sodium p yruvate
(Sigma) and 0.1mM β‐mercaptoet hanol (Sigma). S erum‐free N2B2 7 m edium was prepa r ed
by m ixing 500 ml o f DMEM/F12 (Invitrog e n) w ith 500 ml o f Neurob asa l Medium (Invit rogen)
and add i ng 5 ml of N2 Supplemen t (Invitrogen), 10 ml of B27 Sup plement (Inv it rogen), 5ml
of 0 .2M L‐glutamine (I n v itrog e n), a n d 1ml of 0 . 1 M β‐mercaptoeth ano l ( Sigma). Mouse ESCs,
mouse EpiSCs, rat ESCs and human ESCs were cultured on feeder p lates or 0 .1% gelatin‐
coated plates at 37°C in 5% CO 2 .
46C m ouse ESCs[112] w ere maintain ed i n basal medium s upplemented with 10ng/ml LIF
(Petrot e ch o r prepared i n‐house). Yap1
‐/‐
, Wwtr1
‐/‐
, and Hyg34[126] m ouse E SCs were
mainta in ed i n basal medium suppl emented with 10ng/ ml L IF, 3µM C HIR99021 and 1µ M
PD0325901. Ctnnb1
‐/‐
m ouse ESCs[68, 127] w ere maintained o n feeder i n bas a l medium
supplemented with 1 0ng/ml L IF a nd 1 µM P D0325901. C HIR99021 a nd PD0325901 w ere
synthes i zed in the Div is ion of Sig na l Transduct ion Ther ap y, Uni versity of Dundee, UK.
CD1 mouse EpiSCs w er e der i ved fr om e piblasts o f E5.75 CD1 mouse embryos as previously
described[68]. 46C m ouse EpiSCs were converted from 46C mouse E SCs in b as al m edium
supplemented with 2 0 n g/ml a ctiv in A ( Pepro Tech), 20ng / ml b asic fi broblast g rowth facto r
(bFGF) (PeproTech), and 2.5µM IWR‐1 (Sigma). Both CD1 and 46C m ouse EpiSCs were
123

routinely mainta in ed o n gelatin‐coated p lat e i n basal medium su pplemented with 1 .5µ M
CHIR99021 a nd 2 . 5 µM I WR‐1. Ctnnb1
‐/‐
mouse EpiSCs were converted from Ctnnb1
‐/‐
mouse
ESCs o n CF‐1 f eed er p la tes in b asal m edium supplemented with 2 0 ng /ml activin A, 20ng/ml
bFGF a nd 10µM Y‐27632 (Sigma). A fter c onversion, Ctnnb1
‐/‐
mouse EpiSCs were maintained
on g elatin‐coated plat e in b asal m edium supplemented with 20n g/ml activin A, 20ng/ml
bFGF, and 5µM Y‐27632. Cdh1
‐/‐
mouse EpiSCs w ere routinely main ta ined o n g e latin plat e in
basal medium s upple m ented with either 20 ng/ml activ i n A and 2 0 ng/ml bFGF or 1.5µM
CHIR99021 a nd 2 .5µM I WR‐1. Co nvers i on o f Yap1
‐/‐
and Wwtr1
‐/‐
mouse ESCs into mouse
EpiSCs was carried out on gelatin‐coated plate in basal medium sup p l e mented w ith 20ng/ml
activin A, 20ng/ml bFGF, 1.5µM CHIR, and 2.5µM IWR‐1. E3 mouse EpiSCs[28] w ere
routinely mainta in ed o n gelatin‐coated p lat e i n basal medium su pplemented with 1 .5µ M
CHIR99021 and 2.5µM I WR‐1.
HES2 human ESCs were cultured on feeder plates in DMEM/F12 (Inv itrog e n) supplemented
with 20% f etal b ovine s e rum, 1 % non‐essen t ia l amino acids, 2 mM L‐glutamine, 1 mM s odium
pyruvate, 0. 1mM β‐mercaptoethano l and 20ng/ml b FG F. W hen necess ary, H ES2 human ESC s
were passaged as cell clumps using CTK solution as previously d escribed [59 ]. H9 h uman
ESCs w ere cultured o n Matrig el ( Cornin g)‐coated plates i n mTeSR ‐1 ( STEMCELL
Technologies). When necessary, H9 human ESCs were passaged as c ell c l umps u sing
Accutase ( I n vit r ogen) or G entl e Cell Dissociation Reag ent (STEM CELL Technologies) in
accordance with manufacturer’s instructions. For self‐renewal t ests i n the absence o f
exogenous cytokines or chemicals, HES2 human ESCs were cultured i n the maint e na nce
medium w ithout b FGF supplement. H 9 human ESCs w er e cultured i n serum‐free N2B27
medium.
124

DAc8 rat ESCs[14] were cultured on feeder plates in N2B27 mediu m supp lemented w ith
3µM CHIR99021 and 1µM PD0325901. W hen passaging, both f loating and attac h ed r at ESC
colonies w ere collected w ithin the culture medium i n a 1 5 ‐ml c o nical tube. After removal of
culture medium, rat ESCs w ere d i ssociat ed i nto single c ells b y treatment with 1 X trypsin
solution, consisting o f 0.025% t ryp s in ( Inv i trogen), 1 mM EDTA ( Inv i t r ogen), a nd 1 % chicken
serum (Sigma) in 1 X PBS, b efore serum‐containing MEF medium w as u sed to sto p
dissociat io n process. S ingle cells w ere then p elleted and resus pended i n fr esh culture
medium before plating on fee ders a t the ratio of 1:3.
F or test of b a sal m e di u m and EC M coa tin g o n m ou se Ep iSCs and h u man ESCs, culture plates
were coated with 0.1% gelatin, 15µg/ml human fibronectin (Invitrogen), or FBS at 37°C for
20min, or 5 µg/ml Natural Mouse Laminin (Inv itrog e n) a t 37°C for 2h. Modified N2B27 media
was prepared by adding 2.5 ml of N2 Supplement, N2 Supplement‐A (N2A, STEMCELL
Technologies), or N 2 Supplement‐B (N2B, STEMCELL Technologies) as well as 5ml of B27
Supplement, B 27 minus Vitamin A (B27‐RA, Invitrogen), N euroCult SM1 Neuronal
Supplement (STEMCELL Technologies ), or Ne u roC u lt™ SM1 Su pplemen t Witho u t Vitamin A
(SM1‐RA, S TEMCELL Technologies ) into 500 ml o f DMEM/F12 (Invitr og en) supplemented
with 5ml of 0.2M L‐glut a mine (Invitrogen), and 1ml of 0.1M β‐me rcaptoethanol (Sigma).
Plasmid construction
The P i ggyB a c (PB) t ra n s poson‐base d vectors were o btained from SBI (PB Transposon vector
#PB511B‐1, PB T rans posase v ector #PB200PA‐1). To e nsure a hig h express i o n l ev el o f
trans g en e in ESCs, t he CMV ‐MCS ‐EF1 ‐pac d ual promoter c assette in P B Transposon v ector
was replaced b y a CAG‐MCS ‐IRES ‐pac or CAG‐MCS ‐IRES ‐hph c assette u sing Nhe I and Cla I. T o
125

construct t h e expr ession v ector s f or ΔNβ‐catenin‐ER
T2
, co ding s equence (CDS) for the 8 9 aa
N‐terminal t runcated ΔNβ‐catenin‐ER
T2
f usion protein was amplified from p CAG‐ΔNβ‐
catenin‐ER
T2
‐IRES‐hygr o plasmid by P CR u sing Q 5 High‐Fidelity DNA Polymeras e (New
England Biolabs). Floxed ΔNβ‐catenin‐ER
T2
plasmid was constructed by insertion of the
ΔNβ‐catenin‐ER
T2
c ass e tte into t he p CAG‐loxP‐IRE S‐pur o ‐STOP‐loxP‐EGFP‐pA v ector using
Xho I and Pac I. P B‐based ΔNβ‐catenin‐ER
T2
plasmid was constructed by insertion of the ΔNβ‐
catenin‐ER
T2
cassette into the PB‐CAG‐MC S‐IRES‐puro vector u sing Bgl II and Xho I. T o
generate t he e xpress io n vec t ors for mutant ΔNβ‐catenin‐ER
T2
, C D S f o r ΔNβ‐catenin w ith
A295W/I296W, K435E or H 469A/K470A m utatio ns w ere amplified f rom r espective pCAG‐
β‐catenin
mu t
‐IRES‐hygro plasmids and inserted i nto PB‐CAG‐MCS‐IRES‐puro v ec tor in f rame
wi th the C DS f or ER
T2
cloned from pCAGGS‐Cre‐ER
T2
pla sm i d usi n g Bgl II, Eco RI and Xho I. T o
generate the expression vector for YAP or TAZ, CDS for YAP or TAZ was amplified by PCR
from CD1 mouse EpiSC cDNA pool and inserted into PB‐CAG‐MCS‐IRE S‐puro e xpress ion
vector u sing Bgl II, a nd XhoI. To generate the expression vectors for YAP‐ER
T2
, TAZ‐ER
T2
, and
EGFP‐ER
T2
, CDS for YAP, T A Z , and EGFP w ere amplified b y PCR then i nserte d into P B‐CAG‐
MCS‐IRES‐puro or PB‐CAG‐MCS‐IRES‐hygro vector in frame with ER
T2
CDS using Bgl II, EcoRI
and Xho I. P rimer s e quences a r e listed in the Table 5.
Table 5 Primer sequences for PCR
Target Forward Sequence (5’ – 3’) Reverse Sequence (5’ – 3’)
ΔNβ‐catenin A GGGCTCAGAGGGTCC GAG CAGGTCAGTATCAAACCAGGCC
ER
T2
T C T GC T GGA GA CA T GAGA GC TG TCAAGCTGTGGCAGGGAAAC
TAZ ATGAATCCGTCCTCGGTGC TTACAGCCAGGTTAGAAAGGGC
YAP A TGGAGCCCGCGCAAC C TA TA ACCACGTGAG A AAGCTTTCTT
EGFP A TGGTGAGCAAGGGCGAGG TTA CTTGTACAGCTCGTCCATGCC
Wwtr1 exo n 2 CCT CCTCCTCTG A CTT G CAC ATAGGA CT GCTGGCGG AGAT
Yap1 exon 1 T CTGTCTCAGTTGGGACG CC T ACCCTTACCTGTCGCGAGT
Yap1 exon 2 ATTTGGTTGC C G TGACATAC T TCAAAGGAGGACTGCCGGA
126

For RNA inter f er ence i n mouse EpiSCs, short hairpin RN As ( shRNA s) s pecificall y targeting
mouse TAZ mRNA w ere select ed a nd c loned into a p LKO. 1‐TRC vect o r (Addgene #10878)
according to Addgene’ s pLKO.1 protocol. Target sequenc e of shRN As are listed in Table 6.
Table 6 Target sequences of shRNAs and guide RNAs
shRNA Targeting sequence (5’ – 3’)
TAZ shRNA #1 G CCTG C CA TGAGCA CA GATA T
TAZ shRNA #2 G AGGATTA GGATGCGTCAAG A
T A Z shRNA #3 C AGCC GAAT C T C G CAAT GAAT
scrambled shRNA AATTCTCCGAACGTG T CACG T
guide RNA Targeting sequence (5’ – 3’)
Wwtr1 gRNA #1 A CT CA TCA GGCGGCCA CCCG
Wwtr1 gRNA #2 G CAAGTCATCCACG T CACGC
Yap1 gRNA #1 A CGA C CTGGTGGCCGG CCGG
Yap1 gRNA #2 A CCAGGTCGTGCA CGT CCGC
Yap1 gRNA #3 G AGTGAGCTCGAACATGCTG

Transfection and viral infection
For trans g ene ov erexpressio n , 5× 10
5
m ouse EpiSCs or 1 ×10
6
human ESCs were transfected
with 4 µg pCAG‐Floxed vectors, or 2µg PB Transposon vector toge ther w ith 1µg PB
Tran sposas e vecto r , using Lipofectamin e LTX (Invit rogen) i n acc ordance with
manufactur er’s i nstruct i on. Tran sfected cells w ere selected u nd er 1 µg/ m l puromycin (Sigma)
or 100µg/ ml h ygromycin (Sigma) for 1 week s tarting 24h a fter t r ansfectio n . For RNA
interferenc e , 2×10
6
H E K 293FT (Invitrog e n) c ells w ere transfected w i th 2 µ g pLKO.1 shRNA
vector together with 1.25 g pxPAX2 and 0.75µg VSV‐G packaging v ec tors u sin g
Lipofectamine LTX. S upernata nt c ontain in g viral pa rticles was c ollected 48h a fter
transfection and purified through 0.45µm filter (Millipore). In the presence of 10µg/ml
polybrene (Sigma), 5×10
5
mouse EpiSCs were infected with lentivirus‐containing
127

supernat an t for 16h. Infect ed c ells w ere sel e cted u nder 1 µg/ml puromycin for 1 week
st a rti n g 24h a f te r i nf e c ti on .
Generation of Yap1
‐/‐
and Wwtr1
‐/‐
mESCs
Expressio n v ector pX 330‐U6‐Chime ric_BB‐CBh‐hSpCas9 for Cas9 a nd guide RNA was
obtained f r o m Addgen e (#422 30). To f acilit at e selec t ion, a P 2 A ‐ pac fragment w as i nser ted
in‐frame with Cas9 CDS using Fse I and Eco RI. Guide RNA candidat es specifically t argetin g
mouse Yap1 exon 1 or exon 2 or Wwtr1 e xon 2 were selected and t h eir p otent i al o ff‐targ et
sites were analyzed using NCBI BLAST tool. Annealed oligonucleo tid e s carry ing 20bp guide
RNA sequence w ere inserted i nto pX330‐puro v ector using Bbs I. T arget sequenc e s of g uide
RNAs are listed in Table 6. To perform a knock‐out in mESCs, 5× 10
5
46C mESCs were
trans f ected with 2 µg p X330‐puro v ectors c arry ing guide RNAs. Tr an sfec ted cells w er e
selected u nder 1 µg/ml puromycin for 3 days s tarting 2 4 h after t ra nsfect ion, a fter w hich
merely no c e ll survived in non‐tra n sfect e d gro u p. Trans fected c ells w ere then p a ssaged and
seeded a t 1×10
3
c ells p er w ell on 6 ‐well plates i n the absence of p uromycin. I ndividual
colonies were picked 10‐14 days after seeding and expanded. Loss of YAP or TAZ protein
was confirmed by w estern b lot. F rom the individual c lo nes that sh ow c omplete loss of Y AP
or T AZ p rot ein, genomic DNA was extr acted using Gen t ra P uregen e Genomic DNA Isolation
Kit (Q iag e n ) . Targ et r egions o f gR NA i n Yap1 exon 1, Yap1 exon 2 or Wwtr1 exon 2 were
amplified b y PCR usin g primers listed i n Ta ble 5 and sequenced to detect frame‐shift
mutations.
Real ‐time quantitative PCR
128

Total RNA was extracted from c ultured cells u sing Q uick‐RNA M in iPrep Kit (Z ymo Research).
cDNA w as s ynthes ized f rom 1µg total RNA using iScript cDNA S ynt hesis Kit (Bio‐Rad)
according to manufacturer’s ins t r u ct ion. q PC R assays w e r e ca r r i ed o ut using iTa q Univ ersal
SYBR Green Supermix (Bio‐Rad) on ABi 7900HT Fast Real‐time Syst em ( Ther mo F isher).
Gene e xpression l evels were n or malized to t hat of Gapdh . Pr imer sequences a re l isted in
Table 7.
Table 7 Primer sequences for qPCR
Target Forward Sequence (5’ – 3’) Reverse Sequence (5’ – 3’)
Gapdh TGAAGCAGGCATCTGAGGG CGAAGGTGGAAGAGTGGGAG
Oct4 GAAGCAG A AGAGGATCACCTTG TTCTTAAG GCTGAGCTGCAAG
Rex1 TCACTGTGCTGCCTCCAAGT GGGCACTGATCCGCAAAC
Nanog TCCAGAAGAGGGCGTCAGAT CAAATCCCAGCAACCACATG
Fgf5 GCAGCCCACGGGTCAA CGGTTGCTCGGACTGCTT
Axin2 GGGGGAAAACACAG CTTACA TTGACTGG G TCGCTTCTCTT
Brachyury CCGGTGCTGAAGGTAAATGT CCTCCATTGAGCTTGTTGGT
Taz TCCCCACAACTCCAGAAGAC C AAAGTCCCGAGGTCAACAT
Ctgf AGCGGTGAGTCCTTCCAAAG TTCCAG T CGGTAGGCAGCTA
Axl CGGAGCCTGAGACAATCTTC G CATGGACTGCATCTGAGAA
Cyr61 CACTGAAG AGGCTTCCTGTCTT A GGACGCA CTTCA C AG ATCC

Western bolt
Total prot ein was ext r a c ted by d irect lysis of c ells o n culture p l a te in RIPA b u ffe r ( Teknov a )
supplemented with H alt Protease I nhibitor C o c ktail (Ther m o Fish er), 20mM sodium f luoride
(Sigma), a nd 100mM sodium o rthova nad a te ( Sigma). Protein concen trat ion was measured
using Pierce BCA Protein Assay Kit (Thermo Fisher). 10‐15µg total protein was then
separated by S DS‐PA G E and transf erred o n to P VDF membrane ( Bio‐R ad) for ant i body
probing. T a r get bands were v isu a lized usin g Pierc e ECL S ubstrat e or S uperSignal W est
Femto Subs trate (Thermo Fisher) on a Fluor C h e m E i ma ge r (P r ote i n Si mple ) or a C he mi Doc
129

XPRS+ imager (Bio‐Rad). Primary antibodies used are as follows: β‐Catenin ( 1:2000, B D
Biosciences , 14/β‐Catenin), ER α (1:1000, S anta C ruz, M C‐ 20), pS mad2 ( Ser465/ 467, 1 :1000,
Cell Signaling, 1 38D4), S mad2 ( 1:1000, C ell Sig n aling, D 43B4), pSmad3 ( Ser423/ 425, 1 :1000,
Cell Signaling, C 25A9) , S mad3 ( 1:1000, C ell Signaling, C 67H9), Smad2/3 (1:1000, C ell
Signaling, D 7G7), Smad4 (1:1000, C ell Signaling, D 3M6U), p ‐ERK1 /2 ( Thr202/Tyr204,
1:1000, C ell Signaling, 1 97G2), ERK1/2 ( 1:1000, C ell Signaling) , TAZ ( 1:1000, B D Biosciences ,
M2‐616), YAP (1:1000, C ell Signal ing, D 8H1X), α ‐Tubulin ( 1:4000 , Invitrogen, B‐5‐1‐20) a nd
GAPDH (1:1000, Cell Signaling, 14C10).
Co ‐immunoprecipitation and Mass Spec analysis
Cell ex tract was prepa r ed f rom 5 ×10
6
mouse EpiSCs or human ESCs using IP Lysis Buffer
(50mM Tris‐HCl p H 7.5, 150mM NaCl, 0.1% N P‐40, 5 mM E DTA, 1 X Hal t Protease I nhibit or
Cocktail, 20mM s odium fluoride, and 100mM sodium o rthovanad a te). Supe rn a tan t f rom ce ll
extr act wa s incubated with a n t ibodies a g a inst β‐cat enin o v e rn ig ht a t 4°C followed by
incubation w ith Prot ein A/G PLUS‐Agarose ( Santa Cruz) for 2h a t 4 °C. The beads were t hen
washed f iv e t i mes w i t h I P Lys i s Buffer and resuspended in S DS loading buffer. IP protein
pool was then separated by SDS‐P AGE and vis u alized b y Coomassie B lue staining f ollowing
a protocol f rom USC Proteomics C ore Facility. Protein b ands of int e r e st w ere separat e d from
SDS‐PAGE gel with a sc alpel. Ma ss Spec analysis was carried out in US C Proteomics Core.
Immunofluorescent staining
Cells w ere fixed in 4 % paraformal dehyde f or 20 min and incubate d a t 37° C in b locking buffer
(5% BSA and 0.2% T ritonX‐100 in 1 X PBS) f or 1 h . Cells w ere inc ubated i n the presenc e o f
primary an tibodies a t 4°C ov ern i g h t and then w ashed thr ee times in 1XPBS for 5 min each.
130

Cells w ere t h en i ncubated w ith Alexa Fluor 488 (Invitrog e n, 1:1 000) s econdary a ntibody at
37°C for 1 h. N uclei were s tained w ith Hoechst 33342 (Invitroge n, 1:10000). The primary
antibodies a nd d ilutio ns u sed were O ct4 (sc‐5279; Santa Cruz, 1 :200) and ERα (1:100, S anta
Cruz, MC‐20).
Small molecule screening and testing
For the sma ll‐scale screening of c hemicals a v a ilable i n lab, 5 × 10
4
human ESCs were seeded
in M atr i gel‐ coated 12‐ well plates i n mTeSR overn i ght to a llow a ttachment. The next day,
medium was changed to N2B27 supplemented with 20ng/ml ActA, 20n g/ml b FGF, 1 .5µM
CHIR99021, and 2. 5µM IWR‐1. S mall molecules were u sed as l isted i n Table 8. Medium wa s
changed ev ery oth e r day while cells w ere passaged a s n ecessary . For the test of small
molecule a ctiva t ors of L IF/STAT3 p athway, 5× 10
3
m ouse ESCs wer e s eeded i n g e latin‐coat ed
12‐well plates i n routin e main ten a nce mediu m o vernigh t t o allow attachment. The next day,
medium w as c hanged t o serum‐co n ta ining basal medium w ith or w it hout 10ng/ ml L IF. Sma l l
molecules were u sed as l isted in Table 8. For the test of small m olecule activato rs o f
TGFβ/Smad pathway, 1×10
4
m ouse EpiSCs or 2×10
4
human ESCs were seeded in 24‐well
plates i n respectiv e r o u tine m aint ena n ce m edium overn i ght to a l low attachment. The nex t
day, medium was chan ged to N2B27 supplement e d with differ e nt co mbinations o f 20ng/ml
bFGF, 1.5µM CHIR99021, a nd 2 .5µM I WR‐1. Small mole cules were u s ed a s listed i n Table 8.
Medium was changed every other d ay while c ells were passaged as n ecessar y .
Table 8 Small molecules tested in this study
Compound Action on target Final conc. (µM)
53AH Tank yras e inhibitor 1
JW55 Tank yras e inhibitor 1‐2.5
C59 Porcn inhibitor 0.5
131

IWP ‐2 Porcn inhibitor 2.5
FH535 β‐catenin/ TCF & PPARγ/δ antag o nist 1
Go6976 PKCα/β inh i bitor 1
Go6983 pan‐PKC inhibitor 1‐5
SU6656 Src inhibitor 10
A83 ‐01 ALK4/5/7 inhibitor 0 .5‐1
SB590885 B‐Raf inhibitor 1
WH4 ‐0 ‐23 Lck and Src inhibitor 1
LY294002 PI3K inhibitor 1 0
SB203580 p38 MAPK i nhibitor 5
SP600125 JNK inhibitor 10
Forskolin (FSK) cyclic AMP agonist 1 2
Vertiporfin (VP) Inhibits YAP–TEAD interaction 1
DMH1 ALK2 inhibitor 0 .5
Pifithrin α p53 inhibitor 5‐10
JAK Inhibitor I pan‐JAK inh i bitor 5
Stat3 inhibitor STAT3 inhibitor 0.25
SHP inhibitor SHP1/2 inh ibitor 5
IKK ‐2 Inhibitor IV IKK‐2 inhib i tor 1
CK2 inhibitor IV casein k in a se 2 inhibit o r 5
CRM398 STAT3 activator 1‐10
CRM423 STAT3 activator 1‐10
CRM515 STAT3 activator 1‐10
CRM423F STAT3 activator, CRM 423F analog 1‐10
G4 Smad2/3 activato r 4‐10
CIM61 Smad2/3 activato r 4‐10
B13 Smad2/3 activato r 4‐10
M12 Smad2/3 activato r 4‐10
O15 Smad2/3 activato r 2‐5

132

References
1. Tarkowski, A.K. and J. Wroblewska, Development of blastomeres of mouse eggs isolated at the
4‐ and 8‐cell stage. J Embryol Exp Morphol, 1967. 18(1): p. 155‐80.
2. Johnson, M.H. and C.A. Ziomek, The foundation of two distinct cell lineages within the mouse
morula. Cell, 1981. 24(1): p. 71‐80.
3. Evans, M.J. and M.H. Kaufman, Establishment in culture of pluripotential cells from mouse
embryos. Nature, 1981. 292(5819): p. 154‐6.
4. Martin, G.R., Isolation of a pluripotent cell line from early mouse embryos cultured in medium
conditioned by teratocarcinoma stem cells. Proc Natl Acad Sci U S A, 1981. 78(12): p. 7634‐8.
5. Brons, I.G., et al., Derivation of pluripotent epiblast stem cells from mammalian embryos.
Nature, 2007. 448(7150): p. 191‐5.
6. Tesar, P.J., et al., New cell lines from mouse epiblast share defining features with human
embryonic stem cells. Nature, 2007. 448(7150): p. 196‐9.
7. Matsui, Y., K. Zsebo, and B.L. Hogan, Derivation of pluripotential embryonic stem cells from
murine primordial germ cells in culture. Cell, 1992. 70(5): p. 841‐7.
8. Resnick, J.L., et al., Long‐term proliferation of mouse primordial germ cells in culture. Nature,
1992. 359(6395): p. 550‐1.
9. Shamblott, M.J., et al., Derivation of pluripotent stem cells from cultured human primordial germ
cells. Proc Natl Acad Sci U S A, 1998. 95(23): p. 13726‐31.
10. Shamblott, M.J., et al., Human embryonic germ cell derivatives express a broad range of
developmentally distinct markers and proliferate extensively in vitro. Proc Natl Acad Sci U S A,
2001. 98(1): p. 113‐8.
11. Takahashi, K. and S. Yamanaka, Induction of pluripotent stem cells from mouse embryonic and
adult fibroblast cultures by defined factors. Cell, 2006. 126(4): p. 663‐76.
12. Okita, K., T. Ichisaka, and S. Yamanaka, Generation of germline‐competent induced pluripotent
stem cells. Nature, 2007. 448(7151): p. 313‐7.
13. Ying, Q.L., et al., The ground state of embryonic stem cell self‐renewal. Nature, 2008. 453(7194):
p. 519‐23.
14. Li, P., et al., Germline competent embryonic stem cells derived from rat blastocysts. Cell, 2008.
135(7): p. 1299‐310.
15. Buehr, M., et al., Capture of authentic embryonic stem cells from rat blastocysts. Cell, 2008.
135(7): p. 1287‐98.
16. Thomson, J.A., et al., Embryonic stem cell lines derived from human blastocysts. Science, 1998.
282(5391): p. 1145‐1147.
17. Thomson, J.A., et al., Isolation of a primate embryonic stem cell line. Proc Natl Acad Sci U S A,
1995. 92(17): p. 7844‐8.
18. Nichols, J. and A. Smith, Naive and primed pluripotent states. Cell Stem Cell, 2009. 4(6): p. 487‐
92.
19. Smith, A.G., et al., Inhibition of pluripotential embryonic stem cell differentiation by purified
polypeptides. Nature, 1988. 336(6200): p. 688‐90.
20. Gough, N.M., et al., Molecular cloning and expression of the human homologue of the murine
gene encoding myeloid leukemia‐inhibitory factor. Proc Natl Acad Sci U S A, 1988. 85(8): p. 2623‐
7.
21. Niwa, H., et al., Self‐renewal of pluripotent embryonic stem cells is mediated via activation of
STAT3. Genes Dev, 1998. 12(13): p. 2048‐60.
133

22. Ying, Q.L., et al., BMP induction of Id proteins suppresses differentiation and sustains embryonic
stem cell self‐renewal in collaboration with STAT3. Cell, 2003. 115(3): p. 281‐92.
23. Tosolini, M. and A. Jouneau, From Naive to Primed Pluripotency: In Vitro Conversion of Mouse
Embryonic Stem Cells in Epiblast Stem Cells. Methods Mol Biol, 2016. 1341: p. 209‐16.
24. Guo, G., et al., Klf4 reverts developmentally programmed restriction of ground state
pluripotency. Development, 2009. 136(7): p. 1063‐9.
25. Tai, C.I. and Q.L. Ying, Gbx2, a LIF/Stat3 target, promotes reprogramming to and retention of the
pluripotent ground state. J Cell Sci, 2013. 126(Pt 5): p. 1093‐8.
26. Ye, S., et al., Embryonic stem cell self‐renewal pathways converge on the transcription factor
Tfcp2l1. EMBO J, 2013. 32(19): p. 2548‐60.
27. Martello, G., P. Bertone, and A. Smith, Identification of the missing pluripotency mediator
downstream of leukaemia inhibitory factor. EMBO J, 2013. 32(19): p. 2561‐74.
28. Greber, B., et al., Conserved and divergent roles of FGF signaling in mouse epiblast stem cells and
human embryonic stem cells. Cell Stem Cell, 2010. 6(3): p. 215‐26.
29. Zhou, H., et al., Conversion of mouse epiblast stem cells to an earlier pluripotency state by small
molecules. J Biol Chem, 2010. 285(39): p. 29676‐80.
30. Illich, D.J., et al., Distinct Signaling Requirements for the Establishment of ESC Pluripotency in
Late‐Stage EpiSCs. Cell Rep, 2016.
31. Hanna, J., et al., Human embryonic stem cells with biological and epigenetic characteristics
similar to those of mouse ESCs. Proc Natl Acad Sci U S A, 2010. 107(20): p. 9222‐7.
32. Gafni, O., et al., Derivation of novel human ground state naive pluripotent stem cells. Nature,
2013. 504(7479): p. 282‐6.
33. Chan, Y.S., et al., Induction of a human pluripotent state with distinct regulatory circuitry that
resembles preimplantation epiblast. Cell Stem Cell, 2013. 13(6): p. 663‐75.
34. Takashima, Y., et al., Resetting transcription factor control circuitry toward ground‐state
pluripotency in human. Cell, 2014. 158(6): p. 1254‐69.
35. Theunissen, T.W., et al., Systematic identification of culture conditions for induction and
maintenance of naive human pluripotency. Cell Stem Cell, 2014. 15(4): p. 471‐87.
36. Valamehr, B., et al., Platform for induction and maintenance of transgene‐free hiPSCs resembling
ground state pluripotent stem cells. Stem Cell Reports, 2014. 2(3): p. 366‐81.
37. Ware, C.B., et al., Derivation of naive human embryonic stem cells. Proc Natl Acad Sci U S A,
2014. 111(12): p. 4484‐9.
38. Guo, G., et al., Naive Pluripotent Stem Cells Derived Directly from Isolated Cells of the Human
Inner Cell Mass. Stem Cell Reports, 2016. 6(4): p. 437‐46.
39. Wang, W., et al., Rapid and efficient reprogramming of somatic cells to induced pluripotent stem
cells by retinoic acid receptor gamma and liver receptor homolog 1. Proc Natl Acad Sci U S A,
2011. 108(45): p. 18283‐8.
40. Shi, Y. and J. Massague, Mechanisms of TGF‐beta signaling from cell membrane to the nucleus.
Cell, 2003. 113(6): p. 685‐700.
41. Mahmood, A., et al., Enhanced differentiation of human embryonic stem cells to mesenchymal
progenitors by inhibition of TGF‐beta/activin/nodal signaling using SB‐431542. J Bone Miner Res,
2010. 25(6): p. 1216‐33.
42. Xu, R.H., et al., NANOG is a direct target of TGFbeta/activin‐mediated SMAD signaling in human
ESCs. Cell Stem Cell, 2008. 3(2): p. 196‐206.
43. Vallier, L., et al., Activin/Nodal signalling maintains pluripotency by controlling Nanog
expression. Development, 2009. 136(8): p. 1339‐49.
44. Brown, S., et al., Activin/Nodal signaling controls divergent transcriptional networks in human
embryonic stem cells and in endoderm progenitors. Stem Cells, 2011. 29(8): p. 1176‐85.
134

45. Mullen, A.C., et al., Master transcription factors determine cell‐type‐specific responses to TGF‐
beta signaling. Cell, 2011. 147(3): p. 565‐76.
46. Vallier, L., D. Reynolds, and R.A. Pedersen, Nodal inhibits differentiation of human embryonic
stem cells along the neuroectodermal default pathway. Dev Biol, 2004. 275(2): p. 403‐21.
47. Lanner, F. and J. Rossant, The role of FGF/Erk signaling in pluripotent cells. Development, 2010.
137(20): p. 3351‐60.
48. Li, J., et al., MEK/ERK signaling contributes to the maintenance of human embryonic stem cell
self‐renewal. Differentiation, 2007. 75(4): p. 299‐307.
49. Greber, B., H. Lehrach, and J. Adjaye, Fibroblast growth factor 2 modulates transforming growth
factor beta signaling in mouse embryonic fibroblasts and human ESCs (hESCs) to support hESC
self‐renewal. Stem Cells, 2007. 25(2): p. 455‐64.
50. Greber, B., et al., FGF signalling inhibits neural induction in human embryonic stem cells. EMBO
J, 2011. 30(24): p. 4874‐84.
51. Clevers, H. and R. Nusse, Wnt/beta‐catenin signaling and disease. Cell, 2012. 149(6): p. 1192‐
205.
52. Clevers, H., K.M. Loh, and R. Nusse, Stem cell signaling. An integral program for tissue renewal
and regeneration: Wnt signaling and stem cell control. Science, 2014. 346(6205): p. 1248012.
53. Merrill, B.J., Wnt pathway regulation of embryonic stem cell self‐renewal. Cold Spring Harb
Perspect Biol, 2012. 4(9): p. a007971.
54. Jho, E.H., et al., Wnt/beta‐catenin/Tcf signaling induces the transcription of Axin2, a negative
regulator of the signaling pathway. Mol Cell Biol, 2002. 22(4): p. 1172‐83.
55. Wray, J. and C. Hartmann, WNTing embryonic stem cells. Trends Cell Biol, 2012. 22(3): p. 159‐68.
56. Sato, N., et al., Maintenance of pluripotency in human and mouse embryonic stem cells through
activation of Wnt signaling by a pharmacological GSK‐3‐specific inhibitor. Nat Med, 2004. 10(1):
p. 55‐63.
57. Melchior, K., et al., The WNT receptor FZD7 contributes to self‐renewal signaling of human
embryonic stem cells. Biol Chem, 2008. 389(7): p. 897‐903.
58. Ullmann, U., et al., GSK‐3‐specific inhibitor‐supplemented hESC medium prevents the epithelial‐
mesenchymal transition process and the up‐regulation of matrix metalloproteinases in hESCs
cultured in feeder‐free conditions. Mol Hum Reprod, 2008. 14(3): p. 169‐79.
59. Hasegawa, K., et al., A method for the selection of human embryonic stem cell sublines with high
replating efficiency after single‐cell dissociation. Stem Cells, 2006. 24(12): p. 2649‐60.
60. Dravid, G., et al., Defining the role of Wnt/beta‐catenin signaling in the survival, proliferation,
and self‐renewal of human embryonic stem cells. Stem Cells, 2005. 23(10): p. 1489‐501.
61. Cai, L., et al., Promoting human embryonic stem cell renewal or differentiation by modulating
Wnt signal and culture conditions. Cell Res, 2007. 17(1): p. 62‐72.
62. Bone, H.K., et al., A novel chemically directed route for the generation of definitive endoderm
from human embryonic stem cells based on inhibition of GSK‐3. J Cell Sci, 2011. 124(Pt 12): p.
1992‐2000.
63. Nakanishi, M., et al., Directed induction of anterior and posterior primitive streak by Wnt from
embryonic stem cells cultured in a chemically defined serum‐free medium. FASEB J, 2009. 23(1):
p. 114‐22.
64. Davidson, K.C., et al., Wnt/beta‐catenin signaling promotes differentiation, not self‐renewal, of
human embryonic stem cells and is repressed by Oct4. Proc Natl Acad Sci U S A, 2012. 109(12): p.
4485‐90.
65. Lian, X.J., et al., Interrogating Canonical Wnt Signaling Pathway in Human Pluripotent Stem Cell
Fate Decisions Using CRISPR‐Cas9. Cellular and Molecular Bioengineering, 2016. 9(3): p. 325‐
334.
135

66. Blauwkamp, T.A., et al., Endogenous Wnt signalling in human embryonic stem cells generates an
equilibrium of distinct lineage‐specified progenitors. Nat Commun, 2012. 3: p. 1070.
67. Kurek, D., et al., Endogenous WNT signals mediate BMP‐induced and spontaneous
differentiation of epiblast stem cells and human embryonic stem cells. Stem Cell Reports, 2015.
4(1): p. 114‐28.
68. Kim, H., et al., Modulation of beta‐catenin function maintains mouse epiblast stem cell and
human embryonic stem cell self‐renewal. Nat Commun, 2013. 4: p. 2403.
69. Wu, J., et al., An alternative pluripotent state confers interspecies chimaeric competency. Nature,
2015. 521(7552): p. 316‐21.
70. Ring, D.B., et al., Selective glycogen synthase kinase 3 inhibitors potentiate insulin activation of
glucose transport and utilization in vitro and in vivo. Diabetes, 2003. 52(3): p. 588‐95.
71. Chen, B., et al., Small molecule‐mediated disruption of Wnt‐dependent signaling in tissue
regeneration and cancer. Nat Chem Biol, 2009. 5(2): p. 100‐7.
72. Li, V.S., et al., Wnt signaling through inhibition of beta‐catenin degradation in an intact Axin1
complex. Cell, 2012. 149(6): p. 1245‐56.
73. Varelas, X., The Hippo pathway effectors TAZ and YAP in development, homeostasis and disease.
Development, 2014. 141(8): p. 1614‐26.
74. Johnson, R. and G. Halder, The two faces of Hippo: targeting the Hippo pathway for regenerative
medicine and cancer treatment. Nat Rev Drug Discov, 2014. 13(1): p. 63‐79.
75. Yu, F.X. and K.L. Guan, The Hippo pathway: regulators and regulations. Genes Dev, 2013. 27(4):
p. 355‐71.
76. Zhao, B., K. Tumaneng, and K.L. Guan, The Hippo pathway in organ size control, tissue
regeneration and stem cell self‐renewal. Nat Cell Biol, 2011. 13(8): p. 877‐83.
77. Ramos, A. and F.D. Camargo, The Hippo signaling pathway and stem cell biology. Trends Cell
Biol, 2012. 22(7): p. 339‐46.
78. Hiemer, S.E. and X. Varelas, Stem cell regulation by the Hippo pathway. Biochim Biophys Acta,
2013. 1830(2): p. 2323‐34.
79. Mo, J.S., H.W. Park, and K.L. Guan, The Hippo signaling pathway in stem cell biology and cancer.
EMBO Rep, 2014. 15(6): p. 642‐56.
80. Cockburn, K. and J. Rossant, Making the blastocyst: lessons from the mouse. J Clin Invest, 2010.
120(4): p. 995‐1003.
81. Nishioka, N., et al., The Hippo signaling pathway components Lats and Yap pattern Tead4
activity to distinguish mouse trophectoderm from inner cell mass. Dev Cell, 2009. 16(3): p. 398‐
410.
82. Home, P., et al., Altered subcellular localization of transcription factor TEAD4 regulates first
mammalian cell lineage commitment. Proc Natl Acad Sci U S A, 2012. 109(19): p. 7362‐7.
83. Lian, I., et al., The role of YAP transcription coactivator in regulating stem cell self‐renewal and
differentiation. Genes Dev, 2010. 24(11): p. 1106‐18.
84. Tamm, C., N. Bower, and C. Anneren, Regulation of mouse embryonic stem cell self‐renewal by a
Yes‐YAP‐TEAD2 signaling pathway downstream of LIF. J Cell Sci, 2011. 124(Pt 7): p. 1136‐44.
85. Pijuan‐Galito, S., C. Tamm, and C. Anneren, Serum Inter‐alpha‐inhibitor activates the Yes tyrosine
kinase and YAP/TEAD transcriptional complex in mouse embryonic stem cells. J Biol Chem, 2014.
289(48): p. 33492‐502.
86. Chung, H., et al., Yap1 is dispensable for self‐renewal but required for proper differentiation of
mouse embryonic stem (ES) cells. EMBO Rep, 2016. 17(4): p. 519‐29.
87. Varelas, X., et al., TAZ controls Smad nucleocytoplasmic shuttling and regulates human
embryonic stem‐cell self‐renewal. Nat Cell Biol, 2008. 10(7): p. 837‐48.
136

88. Beyer, T.A., et al., Switch enhancers interpret TGF‐beta and Hippo signaling to control cell fate in
human embryonic stem cells. Cell Rep, 2013. 5(6): p. 1611‐24.
89. Qin, H., et al., Transcriptional analysis of pluripotency reveals the Hippo pathway as a barrier to
reprogramming. Hum Mol Genet, 2012. 21(9): p. 2054‐67.
90. Alarcon, C., et al., Nuclear CDKs drive Smad transcriptional activation and turnover in BMP and
TGF‐beta pathways. Cell, 2009. 139(4): p. 757‐69.
91. De Los Angeles, A., et al., Accessing naive human pluripotency. Curr Opin Genet Dev, 2012.
22(3): p. 272‐82.
92. Huang, G., et al., STAT3 phosphorylation at tyrosine 705 and serine 727 differentially regulates
mouse ESC fates. Stem Cells, 2014. 32(5): p. 1149‐60.
93. Tai, C.I., E.N. Schulze, and Q.L. Ying, Stat3 signaling regulates embryonic stem cell fate in a dose‐
dependent manner. Biol Open, 2014. 3(10): p. 958‐65.
94. Wray, J., et al., Inhibition of glycogen synthase kinase‐3 alleviates Tcf3 repression of the
pluripotency network and increases embryonic stem cell resistance to differentiation. Nat Cell
Biol, 2011. 13(7): p. 838‐45.
95. Martello, G., et al., Esrrb is a pivotal target of the Gsk3/Tcf3 axis regulating embryonic stem cell
self‐renewal. Cell Stem Cell, 2012. 11(4): p. 491‐504.
96. Zhu, A.J. and F.M. Watt, beta‐catenin signalling modulates proliferative potential of human
epidermal keratinocytes independently of intercellular adhesion. Development, 1999. 126(10): p.
2285‐98.
97. Feil, R., et al., Regulation of Cre recombinase activity by mutated estrogen receptor ligand‐
binding domains. Biochem Biophys Res Commun, 1997. 237(3): p. 752‐7.
98. MacDonald, B.T., K. Tamai, and X. He, Wnt/beta‐catenin signaling: components, mechanisms,
and diseases. Dev Cell, 2009. 17(1): p. 9‐26.
99. Yi, F., et al., Opposing effects of Tcf3 and Tcf1 control Wnt stimulation of embryonic stem cell
self‐renewal. Nat Cell Biol, 2011. 13(7): p. 762‐70.
100. Graham, T.A., et al., Crystal structure of a beta‐catenin/Tcf complex. Cell, 2000. 103(6): p. 885‐
96.
101. Graham, T.A., et al., Tcf4 can specifically recognize beta‐catenin using alternative conformations.
Nat Struct Biol, 2001. 8(12): p. 1048‐52.
102. Jin, L., et al., Direct interaction of tumor suppressor CEACAM1 with beta catenin: identification of
key residues in the long cytoplasmic domain. Exp Biol Med (Maywood), 2008. 233(7): p. 849‐59.
103. Faunes, F., et al., A membrane‐associated beta‐catenin/Oct4 complex correlates with ground‐
state pluripotency in mouse embryonic stem cells. Development, 2013. 140(6): p. 1171‐83.
104. Orsulic, S., et al., E‐cadherin binding prevents beta‐catenin nuclear localization and beta‐
catenin/LEF‐1‐mediated transactivation. J Cell Sci, 1999. 112 ( Pt 8): p. 1237‐45.
105. Stepniak, E., G.L. Radice, and V. Vasioukhin, Adhesive and signaling functions of cadherins and
catenins in vertebrate development. Cold Spring Harb Perspect Biol, 2009. 1(5): p. a002949.
106. Azzolin, L., et al., Role of TAZ as mediator of Wnt signaling. Cell, 2012. 151(7): p. 1443‐56.
107. Azzolin, L., et al., YAP/TAZ incorporation in the beta‐catenin destruction complex orchestrates
the Wnt response. Cell, 2014. 158(1): p. 157‐70.
108. Ohgushi, M., M. Minaguchi, and Y. Sasai, Rho‐Signaling‐Directed YAP/TAZ Activity Underlies the
Long‐Term Survival and Expansion of Human Embryonic Stem Cells. Cell Stem Cell, 2015. 17(4):
p. 448‐61.
109. Lai, D., et al., Taxol resistance in breast cancer cells is mediated by the hippo pathway
component TAZ and its downstream transcriptional targets Cyr61 and CTGF. Cancer Res, 2011.
71(7): p. 2728‐38.
137

110. Verma, A., et al., Targeting Axl and Mer kinases in cancer. Mol Cancer Ther, 2011. 10(10): p.
1763‐73.
111. Ludwig, T.E., et al., Derivation of human embryonic stem cells in defined conditions. Nat
Biotechnol, 2006. 24(2): p. 185‐7.
112. Ying, Q.L., et al., Conversion of embryonic stem cells into neuroectodermal precursors in
adherent monoculture. Nat Biotechnol, 2003. 21(2): p. 183‐6.
113. Barrett, S.D., et al., The discovery of the benzhydroxamate MEK inhibitors CI‐1040 and PD
0325901. Bioorg Med Chem Lett, 2008. 18(24): p. 6501‐4.
114. Tojo, M., et al., The ALK‐5 inhibitor A‐83‐01 inhibits Smad signaling and epithelial‐to‐
mesenchymal transition by transforming growth factor‐beta. Cancer Sci, 2005. 96(11): p. 791‐
800.
115. Ludwig, T.E., et al., Feeder‐independent culture of human embryonic stem cells. Nat Methods,
2006. 3(8): p. 637‐46.
116. Podratz, J.L., et al., Myelination by Schwann cells in the absence of extracellular matrix assembly.
Glia, 1998. 23(4): p. 383‐8.
117. Pedranzini, L., et al., Pyridone 6, a pan‐Janus‐activated kinase inhibitor, induces growth
inhibition of multiple myeloma cells. Cancer Res, 2006. 66(19): p. 9714‐21.
118. Sumi, T., et al., Epiblast ground state is controlled by canonical Wnt/beta‐catenin signaling in the
postimplantation mouse embryo and epiblast stem cells. PLoS One, 2013. 8(5): p. e63378.
119. Meyn, M.A., 3rd, et al., SRC family kinase activity is required for murine embryonic stem cell
growth and differentiation. Mol Pharmacol, 2005. 68(5): p. 1320‐30.
120. Hsiao, C., et al., Human pluripotent stem cell culture density modulates YAP signaling. Biotechnol
J, 2016. 11(5): p. 662‐75.
121. Chen, G., et al., Chemically defined conditions for human iPSC derivation and culture. Nat
Methods, 2011. 8(5): p. 424‐9.
122. Dunn, S.J., et al., Defining an essential transcription factor program for naive pluripotency.
Science, 2014. 344(6188): p. 1156‐60.
123. Xue, Z., et al., Genetic programs in human and mouse early embryos revealed by single‐cell RNA
sequencing. Nature, 2013. 500(7464): p. 593‐7.
124. Yan, L., et al., Single‐cell RNA‐Seq profiling of human preimplantation embryos and embryonic
stem cells. Nat Struct Mol Biol, 2013. 20(9): p. 1131‐9.
125. Tang, F., et al., Tracing the derivation of embryonic stem cells from the inner cell mass by single‐
cell RNA‐Seq analysis. Cell Stem Cell, 2010. 6(5): p. 468‐78.
126. Dani, C., et al., Paracrine induction of stem cell renewal by LIF‐deficient cells: a new ES cell
regulatory pathway. Dev Biol, 1998. 203(1): p. 149‐62.
127. Brault, V., et al., Inactivation of the beta‐catenin gene by Wnt1‐Cre‐mediated deletion results in
dramatic brain malformation and failure of craniofacial development. Development, 2001.
128(8): p. 1253‐64.

Abstract (if available)

Abstract Mouse epiblast stem cell (mEpiSC) and human embryonic stem cell (hESC) are primed state pluripotent stem cells whose self-renewal can be maintained by two small molecules, CHIR99021 (CHIR) and IWR-1, in serum-containing medium through cytoplasmic stabilization and retention of β-catenin. The underlying mechanism, however, remains largely unknown. Here I show that cytoplasmic β-catenin interacts with and retains TAZ in the cytoplasm. Cytoplasmic retention of TAZ/YAP promotes mEpiSC self-renewal in the absence of nuclear β-catenin, whereas nuclear translocation of TAZ/YAP leads to differentiation. TAZ/YAP is dispensable for naïve state mouse embryonic stem cell (mESC) self-renewal but required for the proper conversion of mESCs to mEpiSCs. The self-renewal of hESCs, like that of mEpiSCs, is regulated by a similar mechanism involving cytoplasmic retention of β-catenin and TAZ. Based on CHIR and IWR-1, I also identify a group of growth factors and small molecules that are beneficial to hESC self-renewal under serum-free condition. Results from this study demonstrate that transcription co-activators, such as β-catenin and TAZ/YAP, can exert functional roles in the cytoplasm. Discoveries in this study not only extends our understanding of the molecular mechanisms that regulate pluripotency but also provide new hints for the future optimization of serum-free hESC culture conditions.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Role of beta-catenin in mouse epiblast stem cell, embryonic stem cell self-renewal and differentiation

PDF

Role of STAT3 phosphorylation in mouse embryonic stem cell self-renewal and differentiation

PDF

Pleotropic potential of Stat3 in determining self-renewal, apoptosis, and differentiation in mouse embryonic stem cells

PDF

Investigating the role of STAT3 in mouse and rat embryonic stem cell self-renewal and differentiation

PDF

The role of ERK1/2 in mouse embryonic stem cell fate control

PDF

Derivation and characterization of human embryonic stem (hES) cells and human induced pluripotent stem (hiPS) cells in clinical grade conditions

PDF

Epigenetic plasticity of cultured female human embryonic stem cells and regulation of gene expression and chromatin by PR-SET7 mediated H4K20me1

PDF

Characterization of human embryonic stem cell derived retinal pigment epithelial cells for age-related macular degeneration

PDF

Novel roles for Maf1 in embryonic stem cell differentiation and adipogenesis

PDF

Exploring stem cell pluripotency through long range chromosome interactions

PDF

β-catenin couples self-renewal, induction and epithelial morphogenesis and patterning at the outset of mammalian nephrogenesis

PDF

Roles of Klf4 in embryonic stem cells

PDF

The function of BS69 in mouse embryogenesis and embryonic stem cell differentiation

PDF

Harnessing the power of stem cell self-renewal pathways in cancer: dissecting the role of BMI-1 in Ewing’s sarcoma initiation and maintenance

PDF

Role of the bone marrow niche components in B cell malignancies

PDF

The role of Prkci in stem cell maintenance and cell polarity using a 3-D culture system

PDF

Molecular role of EZH2 overexpression in Colorectal Cancer progression

PDF

Interaction of epigenetics and SMAD signaling in stem cells and diseases

PDF

Identification and characterization of adult stem cells in the oral cavity

PDF

Polycomb repressive complex 2 subunit stabilizes NANOG to maintain self-renewal in hepatocellular carcinoma tumor-initiating stem-like cells