Ci a ion: Aldaz, P.; Ma ín-Ma ín,
N.; Saenz-An oñanzas, A.;
Ca asco-Ga cia, E.; Ál a ez-Sa a,
M.; Elúa-Pinin, A.; Polla d, S.M.;
Law ie, C.H.; Mo eno-Vallada es, M.;
Samp ón, N.; e al. High SOX9
Main ains Glioma S em Cell Ac i i y
h ough a Regula o y Loop In ol ing
STAT3 and PML. In . J. Mol. Sci. 2022,
23, 4511. h ps://doi.o g/10.3390/
ijms23094511
Academic Edi o : Ch is ina Pipe i
Recei ed: 11 Ma ch 2022
Accep ed: 18 Ap il 2022
Published: 19 Ap il 2022
Publishe ’s No e: MDPI s ays neu al
wi h ega d o ju isdic ional claims in
published maps and ins i u ional a il-
ia ions.
Copy igh : © 2022 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license (h ps://
c ea i ecommons.o g/licenses/by/
4.0/).
In e na ional Jou nal o
Molecula Sciences
A icle
High SOX9 Main ains Glioma S em Cell Ac i i y h ough a
Regula o y Loop In ol ing STAT3 and PML
Paula Aldaz 1, Na alia Ma ín-Ma ín2, Ande Saenz-An oñanzas 1, Es e ania Ca asco-Ga cia 1,3,
Ma íaÁl a ez-Sa a 1, Alejand o Elúa-Pinin 4, S e en M. Polla d 5, Cha les H. Law ie 6,7 ,
Manuel Mo eno-Vallada es 1,4, Nicolás Samp ón1,4, Jü gen Hench 8, Robin Lo ell-Badge 9,
A kai z Ca acedo 2,8,10,11 and Ande Ma heu 1,3,7,*
1G oup o Cellula Oncology, Biodonos ia Heal h Resea ch Ins i u e, 20014 San Sebas ian, Spain;
[email p o ec ed] (P.A.); ande [email p o ec ed] (A.S.-A.);
[email p o ec ed] (E.C.-G.); ma ia.al a [email p o ec ed]g (M.Á.-S.);
[email p o ec ed] (M.M.-V.); nicolas.samp [email p o ec ed] (N.S.)
2Cen e o Coope a i e Resea ch in Biosciences (CIC bioGUNE), Basque Resea ch and Technology
Alliance (BRTA), Bizkaia Technology Pa k, 48160 De io, Spain; [email p o ec ed] (N.M.-M.);
[email p o ec ed] (A.C.)
3CIBER o F ail y and Heal hy Aging (CIBERFES), Ca los III Ins i u e, 28029 Mad id, Spain
4Donos ia Uni e si y Hospi al, 20014 San Sebas ian, Spain; [email p o ec ed]
5Cen e o Regene a i e Medicine & Edinbu gh Cance Resea ch UK Cen e, Ins i u e o Regene a ion and
Repai , Edinbu gh EH16 4UU, UK; [email p o ec ed]
6G oup o Molecula Oncology, Biodonos ia Heal h Resea ch Ins i u e, 20014 San Sebas ian, Spain;
[email p o ec ed]
7Ike basque, Basque Founda ion o Science, 48009 Bilbao, Spain
8Ins i u e o Pa hology, Uni e si y Hospi al Basel, 48009 Basel, Swi ze land; [email p o ec ed]
9The F ancis C ick Ins i u e, London NW1 1AT, UK; [email p o ec ed]
10 Biochemis y and Molecula Biology Depa men , Uni e si y o he Basque Coun y (UPV/EHU),
48940 Leioa, Spain
11 CIBER o Cance (CIBERONC), Ca los III Ins i u e, 28029 Mad id, Spain
*Co espondence: ande [email p o ec ed]; Tel.: +34-943006073
Abs ac :
Glioma s em cells (GSCs) a e c i ical a ge s o glioma he apy. SOX9 is a ansc ip ion
ac o wi h c i ical oles du ing neu ode elopmen , pa icula ly wi hin neu al s em cells. P e ious
s udies showed ha high le els o SOX9 a e associa ed wi h poo glioma pa ien su i al. SOX9
knockdown impai s GSCs p oli e a ion, con i ming i s po en ial as a a ge o glioma he apy. In his
s udy, we cha ac e ized he unc ion o SOX9 di ec ly in pa ien -de i ed glioma s em cells. No ably,
ansc ip ome analysis o GSCs wi h SOX9 knockdown e ealed STAT3 and PML as downs eam
a ge s. Func ional s udies demons a ed ha SOX9, STAT3, and PML o m a egula o y loop ha
is key o GSC ac i i y and sel - enewal. Analysis o glioma clinical biopsies con i med a posi i e
co ela ion be ween SOX9/STAT3/PML and poo pa ien su i al among he cases wi h he highes
SOX9 exp ession le els. Impo an ly, di ec STAT3 o PML inhibi o s educed he exp ession o
SOX9, STAT3, and PML p o eins, which signi ican ly educed GSCs umo igenici y. In summa y, ou
s udy e eals a no el ole o SOX9 ups eam o STAT3, as a GSC pa hway egula o , and p esen s
pha macological inhibi o s o he signaling cascade.
Keywo ds:
glioblas oma; glioma s em cell; SOX9; he apy; ansc ip ome; STAT3; PML; pha maco-
logical inhibi ion
1. In oduc ion
Glioblas oma, IDH wild ype, (GB) is a g ade IV di use glioma [
1
], and ep esen s
he mos common and agg essi e p ima y b ain umo class in adul s, wi h a median
su i al o 15 mon hs and a 5-yea su i al a e o less han 5% [
2
]. Con en ional ea men
consis s o su gical bulk emo al o he umo , ollowed up by combined adio he apy
In . J. Mol. Sci. 2022,23, 4511. h ps://doi.o g/10.3390/ijms23094511 h ps://www.mdpi.com/jou nal/ijms
In . J. Mol. Sci. 2022,23, 4511 2 o 20
and emozolomide (TMZ)-based chemo he apy [
3
]. Howe e , cu en he apy p o ocols
ha e low success, and pa ien s almos always elapse, o en in a mo e agg essi e manne .
This is pa ially due o a cellula hie a chy, in which a subpopula ion o glioma s em cells
(GSCs) con ibu e o umo elapse [
4
] and he apeu ic esis ance [
5
,
6
]. The e o e, s a egies
aiming a he e adica ion o GSCs a e po en ially p omising o imp o ing he p ognosis o
pa ien s wi h GB [7].
GSCs sha e p ope ies wi h neu al s em cells (NSCs). S udies in mouse models
and human umo specimens ha e demons a ed ha NSCs a e likely he cells o o igin
o human GB [
8
–
10
]. The e is inc easing e idence indica ing ha ansc ip ional and
epigene ic pa hways con olling no mal NSCs ac i i y also con ibu e o he egula ion
o GSCs [
6
,
11
]. T ansc ip ion ac o s ha go e n s em cell di e en ia ion can po en ially
unc ion as oncogenes by p omo ing he acquisi ion o ansc ip ional p og ams equi ed
o umo igenesis, including epigene ic dys egula ion [
12
]. Among hem, se e al membe s
o he SOX (
S
ex-de e mining egion Y (SRY)-b
OX
) amily o ansc ip ion ac o s ha e been
iden i ied o be essen ial o GB p opaga ion and GSC ac i i y [13–16].
SOX9 is a c i ical de elopmen al egula o ha plays an essen ial ole in he es ab-
lishmen and main enance o adul s em cells in a wide ange o issues, including he
cen al ne ous sys em [
17
]. S udies on SOX9 gain and loss o unc ion in cell cul u es and
mice models e ealed ha i main ains adul NSCs [
18
,
19
]. Mo eo e , SOX9 acili a es he
neoplas ic ans o ma ion o di e en cell ypes including NSCs [
20
–
22
] and exe s a p o-
oncogenic ac i i y by con olling cance s em cells (CSCs) in se e al umo ypes, including
GB [
14
,
23
]. In umo biopsies, SOX9 exp ession is gene ally ele a ed and co ela es wi h
poo p ognosis [
20
,
24
]. Al hough i is known ha he SOX9 ansc ip ion ac o plays a ole
in CSCs, he SOX9- ela ed molecula downs eam e ec o s in hese cells in human samples
emain poo ly unde s ood.
In his s udy, we p esen e idence ha SOX9 is a c i ical egula o o GSC main enance.
We e eal ha STAT3 and PML a e c i ical e ec o s egula ed by SOX9. We ound a gene ic
egula o y loop in ol ing SOX9, STAT3, and PML ha modula es GSC ac i i y. Hence,
pha macological inhibi ion o he SOX9–STAT3–PML pa hway may ep esen a no el
ea men s a egy o GB.
2. Resul s
2.1. High SOX9 Le els Co ela e wi h Lowe Pa ien Su i al
We i s cha ac e ized he exp ession o SOX9 in glioma g ade aking ad an age o
publicly a ailable TCGA and Remb and coho s. He ein, we ound ha SOX9 was highe
in g ade IV cases, linking i s high le els o umo malignancy (Figu e 1A). Then, we
analyzed he clinical ele ance o SOX9 in GB by s udying i s exp ession in a coho o
88 human GB pa ien s om Donos ia Uni e si y Hospi al and compa ed i wi h heal hy
b ain issue (Figu e 1B). The le els o SOX9 mRNA we e signi ican ly up egula ed in
GB, whe e mo e han 80% o samples (71 ou o 88 pa ien s) showed o e exp ession ( old
change highe han wo) (Figu e 1C). These esul s we e also con i med a he p o ein
le el since TMA da a om Donos ia Hospi al and Uni e si y Hospi al Basel showed ha
SOX9 exp ession was ele a ed in GB samples (Figu e 1D,E). Rema kably, su i al analysis
e ealed ha GB pa ien s wi h high SOX9 displayed signi ican ly poo e ou comes han
pa ien s wi h low exp ession. Thus, pa ien median su i al was educed om 24 o
13 mon hs in pa ien s wi h high SOX9 om Donos ia Hospi al (Figu e 1F) and om 8 o
3.5 mon hs in pa ien s om Hospi al Basel (Figu e 1G). These esul s sugges SOX9 as a
p ognos ic bioma ke in GB.
In . J. Mol. Sci. 2022,23, 4511 3 o 20
In . J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 3 o 21
Figu e 1. High le els o SOX9 co ela e wi h poo su i al:(A) SOX9 mRNA exp ession in g ade II,
III, and IV o glioma in TCGA and Remb and coho s; (B) SOX9 mRNA exp ession in GB pa ien s
om he Donos ia Uni e si y Hospi al (DUH; n = 88) ela i e o he mean exp ession in heal hy
b ain issue (n = 6); (C) pe cen age o pa ien s wi h “o e exp ession” and “no mal” exp ession o
SOX9 in he DUH coho ; (D) ep esen a i e images o TMAs o SOX9 s aining in GB samples om
he DUH (n = 47) and Uni e si y Hospi al Basel (UHB; n = 20) coho s; (E) pe cen age o pa ien s
wi h “SOX9+” (less han 60% o SOX9-posi i e cells) and “SOX9++” (equal o o mo e han 60%) ex-
p ession om TMAs shown in “C”; (F) Kaplan–Meie su i al analysis in GB pa ien s om he
DUH coho (n = 47; p = 0.003) and (G) UHB coho (n = 20; p = 0.018) based on SOX9 p o ein exp es-
sion le els de e mined by TMA.
2.2. SOX9 Up egula ion Inc eases Tumo igenic Capaci y o GSCs
We p e iously demons a ed ha SOX9 was o e exp essed in oncosphe es de i ed
om con en ional and pa ien -de i ed GSCs, bo h in i o and in i o [23]. In o de o
u he asce ain he molecula and biological p ocesses con olled by SOX9 in GSCs, we
i s o e exp essed SOX9 in GNS166 pa ien -de i ed cells and also in U373-MG cells (Fig-
u e 2A) and pe o med in i o and in i o expe imen s. In i o, SOX9 o e exp ession
p omo ed inc eased cell g ow h (Figu e 2B). In i o, we obse ed an enhanced umo
o ma ion capaci y when SOX9-up egula ed GNS166 cells we e injec ed o ho opically in
immunode icien NOD-SCID mice. This led o a signi ican dec ease in mice su i al (p =
0.004), wi h an o e all li espan o 35 ± 7.8 days, compa ed wi h 116 ± 11.5 days o median
su i al in con ols (Figu e 2C). In addi ion, we injec ed con ol and SOX9-up egula ed
U373-MG cells subcu aneously in Foxn1nu/Foxn1nu nude mice, which no ably inc eased
hei umo igenic po en ial. Thus, 87.5% o mice inocula ed wi h SOX9-up egula ed U373-
MG cells de eloped umo s, compa ed wi h he 12.5% o mice injec ed wi h con ol cells.
Tumo s gene a ed by U373-MG cells wi h o e exp ession o SOX9 appea ed ea lie (day
53 s. day 84), and hey had g ea e olume (Figu e 2D,E). Immunohis ochemis y anal-
yses also showed ha hese umo s had highe SOX9 and inc eased p oli e a i e capaci y
measu ed by Ki67 exp ession (Figu e 2F). Addi ionally, SOX2 le els we e also inc eased
Figu e 1.
High le els o SOX9 co ela e wi h poo su i al: (
A
)SOX9 mRNA exp ession in g ade II,
III, and IV o glioma in TCGA and Remb and coho s; (
B
)SOX9 mRNA exp ession in GB pa ien s
om he Donos ia Uni e si y Hospi al (DUH; n= 88) ela i e o he mean exp ession in heal hy b ain
issue (n= 6); (
C
) pe cen age o pa ien s wi h “o e exp ession” and “no mal” exp ession o SOX9
in he DUH coho ; (
D
) ep esen a i e images o TMAs o SOX9 s aining in GB samples om he
DUH (n= 47) and Uni e si y Hospi al Basel (UHB; n= 20) coho s; (
E
) pe cen age o pa ien s wi h
“SOX9
+
” (less han 60% o SOX9-posi i e cells) and “SOX9
++
” (equal o o mo e han 60%) exp ession
om TMAs shown in “C”; (
F
) Kaplan–Meie su i al analysis in GB pa ien s om he DUH coho
(n= 47; p= 0.003) and (
G
) UHB coho (n= 20; p= 0.018) based on SOX9 p o ein exp ession le els
de e mined by TMA.
2.2. SOX9 Up egula ion Inc eases Tumo igenic Capaci y o GSCs
We p e iously demons a ed ha SOX9 was o e exp essed in oncosphe es de i ed
om con en ional and pa ien -de i ed GSCs, bo h
in i o
and
in i o
[
23
]. In o de o u -
he asce ain he molecula and biological p ocesses con olled by SOX9 in GSCs, we i s
o e exp essed SOX9 in GNS166 pa ien -de i ed cells and also in U373-MG cells (Figu e 2A)
and pe o med
in i o
and
in i o
expe imen s.
In i o
, SOX9 o e exp ession p omo ed
inc eased cell g ow h (Figu e 2B).
In i o
, we obse ed an enhanced umo o ma ion
capaci y when SOX9-up egula ed GNS166 cells we e injec ed o ho opically in immunode-
icien NOD-SCID mice. This led o a signi ican dec ease in mice su i al
(p= 0.004)
, wi h
an o e all li espan o 35
±
7.8 days, compa ed wi h 116
±
11.5 days o median su i al
in con ols (Figu e 2C). In addi ion, we injec ed con ol and SOX9-up egula ed U373-MG
cells subcu aneously in Foxn1
nu
/Foxn1
nu
nude mice, which no ably inc eased hei umo i-
genic po en ial. Thus, 87.5% o mice inocula ed wi h SOX9-up egula ed U373-MG cells
de eloped umo s, compa ed wi h he 12.5% o mice injec ed wi h con ol cells. Tumo s
gene a ed by U373-MG cells wi h o e exp ession o SOX9 appea ed ea lie (day 53 s.
day 84), and hey had g ea e olume (Figu e 2D,E). Immunohis ochemis y analyses also
In . J. Mol. Sci. 2022,23, 4511 4 o 20
showed ha hese umo s had highe SOX9 and inc eased p oli e a i e capaci y measu ed
by Ki67 exp ession (Figu e 2F). Addi ionally, SOX2 le els we e also inc eased sugges ing
en ichmen o s emness ac i i y (Figu e 2F). Taken oge he , ou esul s show ha high
SOX9 le els endo se umo igenic capaci y o GSCs.
In . J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 4 o 21
sugges ing en ichmen o s emness ac i i y (Figu e 2F). Taken oge he , ou esul s show
ha high SOX9 le els endo se umo igenic capaci y o GSCs.
Figu e 2. SOX9 up egula ion leads o inc eased umo igenici y: (A) ep esen a i e immunoblo o
SOX9 exp ession in indica ed glioma cells in ec ed wi h pWXL GFP (“Con ol”) and pWPXL-Sox9
(“SOX9”) plasmids. β-ac in was used as loading con ol; (B) quan i ica ion o cell g ow h in con ol
and SOX9 o e exp ession in U373-MG and GNS166 cells a day 5; (C) Kaplan–Meie cu es ep e-
sen ing NOD-SCID mice su i al a e s e eo ac ic injec ion o GNS166 con ol and SOX9 GNS166
cells (n = 4).; (D) umo olume a indica ed ime poin s a e subcu aneous injec ion o con ol and
SOX9 U373-MG cells in immunode icien mice (n = 4); (E) ep esen a i e image o umo s in bo h
lanks ( igh lank: injec ion o SOX9-o e exp essed cells; le lank: con ol cells). Tumo s a e indi-
ca ed by a ows; (F) ep esen a i e images o SOX9, SOX2, and Ki67 IHC s aining om subcu ane-
ously gene a ed umo s in “Con ol” and “SOX9” g oups. Scale ba co esponds o 100 µm. * p <
0.05, ** p < 0.01.
2.3. SOX9 Down egula ion Dec eases S emness and Tumo igenic Capaci y in GSCs
To u he add ess he impac o SOX9 on he egula ion o GSCs, we knocked down
SOX9 in GNS166 cells and con en ional cells U373-MG and U251-MG (wi h in e media e
and high basal le els o SOX9, espec i ely). Da a om immunoblo ing expe imen s con-
i med an e ec i e educ ion in SOX9 le els accompanied by a ma ked diminishmen o
SOX2 exp ession (Figu e 3A). In his con ex , we de ec ed a signi ican dec ease o mo e
han 50% in cell g ow h in U373-MG and U251-MG cells (Figu e 3B), as well as a educ ion
in he numbe o phospho-His one H3 (P-H3)-posi i e cells in GNS166 cells (Figu e 3C) in
SOX9 knockdown cells. This impai men in cell p oli e a ion was accompanied by a sig-
ni ican inc ease in senescence measu ed by cy oplasmic SA-β-gal ac i i y (Figu e 3D) and
also by a dec ease in apop osis [25]. Mo eo e , SOX9 knockdown cells also p esen ed
ma kedly educed s emness ac i i y measu ed by dec eased colony o ma ion abili y (Fig-
u e 3E), lowe numbe o so aga oci (Supplemen a y Ma e ial Figu e S1A), inc eased
exp ession o di e en ia ion ma ke s (Figu e 3F), and educed umo sphe e o ma ion by
mo e han 60% in bo h p ima y and seconda y cul u e condi ion (Figu es 3G and S1B).
We pe o med in i o expe imen s and ound a no ably educed umo o ma ion
and p og ession capaci y o shSOX9 U373-MG cells in Foxn1nu/Foxn1nu nude mice. Thus,
only 33% o mice inocula ed wi h shSOX9 cells de eloped umo s, compa ed wi h 85% o
con ols, and hey o med smalle umo s (Figu e 3H). In line wi h his, shSOX9 umo s
exp essed lowe le els o SOX9 and Ki67 han con ol umo s (Figu e 3I). Simila ly, SOX2
Figu e 2.
SOX9 up egula ion leads o inc eased umo igenici y: (
A
) ep esen a i e immunoblo
o SOX9 exp ession in indica ed glioma cells in ec ed wi h pWXL GFP (“Con ol”) and pWPXL-
Sox9 (“SOX9”) plasmids.
β
-ac in was used as loading con ol; (
B
) quan i ica ion o cell g ow h
in con ol and SOX9 o e exp ession in U373-MG and GNS166 cells a day 5; (
C
) Kaplan–Meie
cu es ep esen ing NOD-SCID mice su i al a e s e eo ac ic injec ion o GNS166 con ol and
SOX9 GNS166 cells (n= 4); (
D
) umo olume a indica ed ime poin s a e subcu aneous injec ion
o con ol and SOX9 U373-MG cells in immunode icien mice (n= 4); (
E
) ep esen a i e image o
umo s in bo h lanks ( igh lank: injec ion o SOX9-o e exp essed cells; le lank: con ol cells).
Tumo s a e indica ed by a ows; (
F
) ep esen a i e images o SOX9, SOX2, and Ki67 IHC s aining
om subcu aneously gene a ed umo s in “Con ol” and “SOX9” g oups. Scale ba co esponds o
100 µm. * p< 0.05, ** p< 0.01.
2.3. SOX9 Down egula ion Dec eases S emness and Tumo igenic Capaci y in GSCs
To u he add ess he impac o SOX9 on he egula ion o GSCs, we knocked down
SOX9 in GNS166 cells and con en ional cells U373-MG and U251-MG (wi h in e media e
and high basal le els o SOX9, espec i ely). Da a om immunoblo ing expe imen s
con i med an e ec i e educ ion in SOX9 le els accompanied by a ma ked diminishmen o
SOX2 exp ession (Figu e 3A). In his con ex , we de ec ed a signi ican dec ease o mo e han
50% in cell g ow h in U373-MG and U251-MG cells (Figu e 3B), as well as a educ ion in he
numbe o phospho-His one H3 (P-H3)-posi i e cells in GNS166 cells (Figu e 3C) in SOX9
knockdown cells. This impai men in cell p oli e a ion was accompanied by a signi ican
inc ease in senescence measu ed by cy oplasmic SA-
β
-gal ac i i y (Figu e 3D) and also by
a dec ease in apop osis [
25
]. Mo eo e , SOX9 knockdown cells also p esen ed ma kedly
educed s emness ac i i y measu ed by dec eased colony o ma ion abili y (Figu e 3E),
lowe numbe o so aga oci (Supplemen a y Ma e ials Figu e S1A), inc eased exp ession
o di e en ia ion ma ke s (Figu e 3F), and educed umo sphe e o ma ion by mo e han
60% in bo h p ima y and seconda y cul u e condi ion (Figu es 3G and S1B).
In . J. Mol. Sci. 2022,23, 4511 5 o 20
In . J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 5 o 21
exp ession was also educed (Figu e 3I). Addi ionally, o ho opic injec ion o shSOX9
GNS166 cells e ealed a signi ican inc ease in mice su i al (p = 0.009) passing om ~100
days in con ols o almos 200 days in shSOX9 (Figu e 3J). Taken oge he , ou esul s poin
ou ha SOX9 is a ele an media o o malignan pheno ypes o GB by modula ing GSCs
capaci ies.
Figu e 3. SOX9 knockdown impai s umo igenici y: (A) ep esen a i e immunoblo o SOX9 and
SOX2 exp ession in glioma cells wi h SOX9 down egula ion (shSOX9-1; sh1) and con ols (pLKO).
β-ac in was used as loading con ol (n = 3); (B) quan i ica ion o cell g ow h in con ol and sh1 U373-
MG, U251-MG cells a day 5 (n = 3); (C) quan i ica ion o P-H3
+
GNS166 cells in ec ed wi h pLKO o
sh1 plasmids (n = 3); (D) quan i ica ion o SA-β-gal
+
cells in sh1 and con ol cells (n = 3); (E) quan i-
ica ion o numbe o colonies o med om glioma cell lines in ec ed wi h pLKO o sh1 plasmids (n
= 3); (F) exp ession o Tuj1 and p27
KIP
in GNS166 cells in ec ed wi h pLKO o sh1 plasmids (n = 2);
(G) quan i ica ion o p ima y umo sphe es a e 10 days in cul u e (n = 3). Quan i ica ions a e ex-
p essed ela i e o pLKO; (H) umo olume o subcu aneous umo s gene a ed by pLKO- and sh1-
in ec ed U373-MG cells (n = 4); (I) SOX9, SOX2, and Ki67 IHC s aining om subcu aneously gene -
a ed umo s in “pLKO” and “sh1” g oups. Scale ba co esponds o 100 µm; (J) Kaplan–Meie cu es
ep esen ing mice su i al a e s e eo ac ic injec ion o pLKO and sh1 GNS166 cells. * p < 0.05, ** p
< 0.01, *** p < 0.001.
2.4. T ansc ip omic Analysis Re eals STAT3 and PML as Media o s o SOX9 Ac i i y in GSCs
We add essed he molecula mechanisms unde lying SOX9 unc ion in GB. Fo ha
pu pose, we compa ed exp ession mic oa ay da a om SOX9 knockdown GNS166 cells
wi h con ol GNS166 cells. Clus e analysis showed di e ences in gene exp ession be-
ween bo h pheno ypes being he le els o SOX9 gene he mos dec eased alida ing he
expe imen al app oach (Figu e 4A and Supplemen a y Ma e ials Table S1). Gene
Figu e 3.
SOX9 knockdown impai s umo igenici y: (
A
) ep esen a i e immunoblo o SOX9 and
SOX2 exp ession in glioma cells wi h SOX9 down egula ion (shSOX9-1;sh1) and con ols (pLKO).
β
-ac in was used as loading con ol (n= 3); (
B
) quan i ica ion o cell g ow h in con ol and sh1
U373-MG, U251-MG cells a day 5 (n= 3); (
C
) quan i ica ion o P-H3
+
GNS166 cells in ec ed wi h
pLKO o sh1 plasmids (n= 3); (
D
) quan i ica ion o SA-
β
-gal
+
cells in sh1 and con ol cells (n= 3);
(
E
) quan i ica ion o numbe o colonies o med om glioma cell lines in ec ed wi h pLKO o sh1
plasmids (n= 3); (
F
) exp ession o Tuj1 and p27
KIP
in GNS166 cells in ec ed wi h pLKO o sh1 plasmids
(n= 2); (
G
) quan i ica ion o p ima y umo sphe es a e 10 days in cul u e (n= 3). Quan i ica ions
a e exp essed ela i e o pLKO; (
H
) umo olume o subcu aneous umo s gene a ed by pLKO- and
sh1-in ec ed U373-MG cells (n= 4); (
I
) SOX9, SOX2, and Ki67 IHC s aining om subcu aneously
gene a ed umo s in “pLKO” and “sh1” g oups. Scale ba co esponds o 100
µ
m; (
J
) Kaplan–Meie
cu es ep esen ing mice su i al a e s e eo ac ic injec ion o pLKO and sh1 GNS166 cells. * p< 0.05,
** p< 0.01, *** p< 0.001.
We pe o med
in i o
expe imen s and ound a no ably educed umo o ma ion
and p og ession capaci y o shSOX9 U373-MG cells in Foxn1
nu
/Foxn1
nu
nude mice. Thus,
only 33% o mice inocula ed wi h shSOX9 cells de eloped umo s, compa ed wi h 85% o
con ols, and hey o med smalle umo s (Figu e 3H). In line wi h his, shSOX9 umo s
exp essed lowe le els o SOX9 and Ki67 han con ol umo s (Figu e 3I). Simila ly, SOX2
exp ession was also educed (Figu e 3I). Addi ionally, o ho opic injec ion o shSOX9
GNS166 cells e ealed a signi ican inc ease in mice su i al (p= 0.009) passing om
~100 days in con ols o almos 200 days in shSOX9 (Figu e 3J). Taken oge he , ou esul s
poin ou ha SOX9 is a ele an media o o malignan pheno ypes o GB by modula ing
GSCs capaci ies.
In . J. Mol. Sci. 2022,23, 4511 6 o 20
2.4. T ansc ip omic Analysis Re eals STAT3 and PML as Media o s o SOX9 Ac i i y in GSCs
We add essed he molecula mechanisms unde lying SOX9 unc ion in GB. Fo ha
pu pose, we compa ed exp ession mic oa ay da a om SOX9 knockdown GNS166 cells
wi h con ol GNS166 cells. Clus e analysis showed di e ences in gene exp ession be-
ween bo h pheno ypes being he le els o SOX9 gene he mos dec eased alida ing he
expe imen al app oach (Figu e 4A and Supplemen a y Ma e ials Table S1). Gene On ology
analysis e ealed JAK2-media ed signaling among he op pa hways signi ican ly al e ed
in esponse o SOX9 knockdown, oge he wi h in e e on signaling, in e leukin 17 (IL17)
signaling, o g ow h ho mones (Figu e 4B). Janus kinase 2 (JAK2) is a y osine kinase ha
ac i a es he signal ansduce and ac i a o o ansc ip ion 3 (STAT3) ansc ip ion ac o ,
which has been in ol ed in he p og ession o mos ypes o cance s including GB, which
is equi ed o GSC g ow h and sel - enewal [
26
]. In addi ion, ac i a ed STAT3 le els a e
known o co ela e wi h p omyelocy ic leukemia (PML) gene exp ession in se e al umo
models [
27
], including b eas cance , whe e i ac s as an ups eam egula o o PML [
28
]. In
his con ex , we hypo hesized ha STAT3 and PML could be pa o a egula o y pa hway
ha egula es GSC ac i i y. We alida ed ansc ip omic esul s and de ec ed educed le els
o STAT3 ( o al STAT3 and p-STAT3), as well as PML in shSOX9 GNS166 cells (Figu e 4C).
Simila esul s we e ob ained in SOX9-silenced U251-MG cells (Figu e 4C). Acco dingly,
mode a ely highe le els o STAT3, p-STAT3, and PML we e de ec ed in SOX9 o e exp ess-
ing U373-MG cells (Figu e 4D). These da a sugges ha SOX9 egula es STAT3 and PML
exp ession in GB.
We analyzed SOX9, STAT3/p-STAT3, and PML exp ession in a se o es ablished
glioma cell lines (Figu e 4E) and pa ien -de i ed cells cul u es (Figu e 4F,G) by immunoblo .
We obse ed a posi i e co ela ion among SOX9, p-STAT3, and PML p o ein le els in all
cell lines excep U373-MG cells (Figu e 4E–G). This co ela ion was also iden i ied in umo -
sphe es om U87-MG and U251-MG cells, whe e SOX9, p-STAT3, and PML exp ession was
inc eased (Figu e 4H). In addi ion, simila esul s we e ob ained in subcu aneous umo s
gene a ed om U373-MG seconda y umo sphe es in which he numbe o posi i e cells
and he in ensi y o s aining is highe o SOX9, STAT3, and PML, compa ed wi h U373-MG
pa en al cells (Figu e 4I) [
23
]. In line wi h his, he exp ession o STAT3 and PML also was
highe in g ade IV clinical biopsies om publicly a ailable TCGA and Remb and coho s
(Figu e 4J).
To u he ein o ce he link be ween SOX9, STAT3, and PML, we examined biop ic
issue. Rema kably, he associa ion was u he con i med in GB samples om he Hospi al
Basel coho , whe e we ound a signi ican posi i e co ela ion be ween SOX9 wi h PML
and p-STAT3, as well as be ween PML and p-STAT3 (Figu e 4K). To ein o ce his esul ,
we compa ed SOX9 and STAT3, as well as PML and STAT3 mRNA exp ession in The
Cance Genome A las (TCGA) coho (n= 580) [
29
] and again ound a posi i e co ela ion
(
ρ
Spea man = 0.61 and 0.68, espec i ely) (Figu e 4L). Toge he , hese esul s e eal ha
SOX9, STAT3, and PML exp ession co ela e in GB biopsies.
In . J. Mol. Sci. 2022,23, 4511 7 o 20
In . J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 7 o 21
Figu e 4. SOX9 exp ession co ela es wi h STAT3 and PML: (A) clus e analysis o GNS166 cells
a e ansc ip omic analysis; (B) op canonical pa hways al e ed in shSOX9 (sh1) GNS166 cells com-
pa ed o con ols (n = 3); (C) ep esen a i e immunoblo s o PML, p-STAT3, STAT3 and SOX2 ex-
p ession in sh1 and pLKO GNS166 and U251-MG cells (n = 3); (D) ep esen a i e immunoblo o
PML, p-STAT3, and STAT3 exp ession in SOX9 and con ol U373-MG cells (n = 3); (E) ep esen a i e
immunoblo o SOX9, p-STAT3, and PML exp ession in indica ed glioma cell lines (n = 3); (F) ep e-
sen a i e immunoblo o p-STAT3 and STAT3 exp ession in GNS166, GNS179, and GB1 pa ien -
de i ed glioma s em cells; (G) ep esen a i e immunoblo o SOX9 and PML p o ein le els in
Figu e 4.
SOX9 exp ession co ela es wi h STAT3 and PML: (
A
) clus e analysis o GNS166 cells a e
ansc ip omic analysis; (
B
) op canonical pa hways al e ed in shSOX9 (sh1) GNS166 cells compa ed
o con ols (n= 3); (
C
) ep esen a i e immunoblo s o PML, p-STAT3, STAT3 and SOX2 exp ession in
sh1 and pLKO GNS166 and U251-MG cells (n= 3); (
D
) ep esen a i e immunoblo o PML, p-STAT3,
In . J. Mol. Sci. 2022,23, 4511 8 o 20
and STAT3 exp ession in SOX9 and con ol U373-MG cells (n= 3); (
E
) ep esen a i e immunoblo
o SOX9, p-STAT3, and PML exp ession in indica ed glioma cell lines (n= 3); (
F
) ep esen a i e
immunoblo o p-STAT3 and STAT3 exp ession in GNS166, GNS179, and GB1 pa ien -de i ed glioma
s em cells; (
G
) ep esen a i e immunoblo o SOX9 and PML p o ein le els in pa ien -de i ed GSCs
(23) (n= 3); (
H
) ep esen a i e immunoblo o SOX9, p-STAT3, and PML exp ession in pa en al (“-”)
and oncosphe es (GSCs) om indica ed cells; (
I
) ep esen a i e images o SOX9, PML, and p-STAT3
IHC s aining om umo s a e subcu aneous injec ion o pa en al and umo sphe es om U373-MG
cells. Scale ba co esponds o 50 µm; (J)STAT3 and PML mRNA exp ession in g ades II, III, and IV
o glioma in TCGA and Remb and coho s; (
K
) TMA o p-STAT3, PML, and SOX9 in GB human
samples om he Hospi al Basel coho (n= 20). Chi-squa ed (
χ2
) es was used o compa e equency
da a and showed s a is ically signi ican di e ences be ween g oups (p< 0.01 o 0.0001, espec i ely)
wi h high and low s aining o he indica ed p o eins; (
L
) co ela ion o STAT3 wi h SOX9 and PML
mRNA exp ession in GB samples om he TCGA coho (n= 580) (
ρ
Spea man = 0.61 and 0.68,
espec i ely). N.A., no a ailable. ** p< 0.01, *** p< 0.001.
2.5. STAT3 Regula es GSC Ac i i y and I s Pha macological Inhibi ion Reduces Tumo igenici y
To in es iga e he SOX9–STAT3–PML axis as a po en ial he apeu ic a ge , we knocked
down STAT3 in GNS166 and U251-MG cells using wo di e en shRNA cons uc s (sh41 and
sh43) and analyzed he esul ing pheno ypes. Immunoblo con i med he educed le els
o STAT3 and p-STAT3, as well as SOX9 and PML (Figu e 5A). STAT3 silencing led o a
signi ican educ ion o mo e han 50% in cell p oli e a ion in bo h models (Figu e 5B), which
was associa ed wi h an inc eased numbe o senescen cells (Figu e 5C). Mo eo e , STAT3
knockdown impai ed he umo sphe e o ma ion o U251-MG glioma cells (Figu e 5D).
These esul s con i m o e lapping pheno ypes in SOX9 and STAT3 knockdown glioma
cells in i o.
We nex mo ed o
in i o
assays, which u he con i med hese esul s. Thus, im-
munode icien mice subcu aneously injec ed wi h sh41 o sh43 U251-MG cells did no o m
umo s (sh43) o o med hem a a low pe cen age and wi h less olume (sh41) han con ol
mice (Figu e 5E). In addi ion, he limi ing dilu ion assay e ealed ha STAT3 knockdown
led o a educed numbe o umo -ini ia ing cells (Figu e 5F). O e all, ou da a e eal ha
STAT3 knockdown p oduces equal pheno ypes o simila se e i y as hose obse ed a e
SOX9 down egula ion.
Since es ablished pha macological STAT3 inhibi o s could po en ially be ansla ed
o clinical p ac ice, we explo ed he e ec o STAT3 inhibi ion on he SOX9–STAT3–PML
axis and GB p og ession. We es ed he speci ic STAT3 inhibi o STX-0119, which p e en s
he dime iza ion o STAT3 and i s subsequen binding o DNA. We ound ha inc easing
doses (25, 50, and 100
µ
M) o his d ug signi ican ly educed iabili y in glioma and GSC
cells (Figu e 5G). In addi ion, he umo sphe e o ma ion abili y a e was signi ican ly
diminished in all cell cul u es (Figu e 5D). Rema kably, he pha macological inhibi ion
o STAT3 educed SOX9 and PML exp ession le els, as well as he exp ession o SOX2, a
s em cell ma ke (Figu e 5I), mimicking he e ec s o gene silencing o STAT3. We ha e
p e iously desc ibed ha SOX2 con ols SOX9 le els [
23
], so i migh be o in e es o
un a el whe he i plays a ole in he SOX9–STAT3–PML axis.
In . J. Mol. Sci. 2022,23, 4511 9 o 20
In . J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 9 o 21
Figu e 5. Gene ic and pha macological STAT3 inhibi ion in glioma cells: (A) ep esen a i e im-
munoblo s o p-STAT3, STAT3, PML, SOX9, and SOX2 in U251-MG and GNS166 cells in STAT3
knockdown (“sh41” and “sh43”) and con ol (“pLKO”) cells (n = 3); (B) quan i ica ion o cell g ow h
in sh41 and sh43 cells compa ed wi h con ols (n = 2); (C) quan i ica ion o SA-β-gal+ cells in STAT3
knockdown cells (n = 2); (D) quan i ica ion o umo sphe es (p ima y and seconda y) in sh41 and
sh43 U251-MG cells ela i e o pLKO condi ion (n = 2); (E) umo olume ep esen a ion a indica ed
ime poin s a e subcu aneous injec ion o U251-MG cells (n = 4); (F) ELDA plo o limi ing dilu ion
assay o sh41, sh43, and con ol U251-MG cells; (G) cy o oxici y exhibi ed by indica ed glioma cells
a e ea men wi h inc easing doses o STX-0119 o 72 h (n = 6); (H) quan i ica ion o umo sphe es
in indica ed cells a e ea men wi h 50 and 100 µM STX-0119 (n = 3). Calcula ions we e ela i e o
un ea ed cells; (I) ep esen a i e immunoblo o p-STAT3, STAT3, PML, SOX9, and SOX2
Figu e 5.
Gene ic and pha macological STAT3 inhibi ion in glioma cells: (
A
) ep esen a i e im-
munoblo s o p-STAT3, STAT3, PML, SOX9, and SOX2 in U251-MG and GNS166 cells in STAT3
knockdown (“sh41” and “sh43”) and con ol (“pLKO”) cells (n= 3); (
B
) quan i ica ion o cell g ow h
in sh41 and sh43 cells compa ed wi h con ols (n= 2); (
C
) quan i ica ion o SA-
β
-gal
+
cells in STAT3
knockdown cells (n= 2); (
D
) quan i ica ion o umo sphe es (p ima y and seconda y) in sh41 and
sh43 U251-MG cells ela i e o pLKO condi ion (n= 2); (
E
) umo olume ep esen a ion a indica ed
ime poin s a e subcu aneous injec ion o U251-MG cells (n= 4); (
F
) ELDA plo o limi ing dilu ion
assay o sh41, sh43, and con ol U251-MG cells; (
G
) cy o oxici y exhibi ed by indica ed glioma cells
a e ea men wi h inc easing doses o STX-0119 o 72 h (n= 6); (
H
) quan i ica ion o umo sphe es
in indica ed cells a e ea men wi h 50 and 100
µ
M STX-0119 (n= 3). Calcula ions we e ela i e o
un ea ed cells; (
I
) ep esen a i e immunoblo o p-STAT3, STAT3, PML, SOX9, and SOX2 exp ession
in U251-MG cells a e ea men wi h 50 and 100
µ
M STX-0119 o 72 h (n= 3). * p< 0.05, ** p< 0.01,
*** p< 0.001.
In . J. Mol. Sci. 2022,23, 4511 16 o 20
in cul u e, sphe es we e coun ed using a ligh mic oscope. Fo seconda y (2
y
) gene -
a ion, sphe es we e mechanically and enzyma ically disagg ega ed wi h accu ase (Li e
Technologies, Ca lsbad, CA, USA) and hen main ained in cul u e o ano he 10 days.
4.7. Senescence-Associa ed β-Galac osidase S aining
To e alua e cellula senescence, he senescence-associa ed
β
-galac osidase (SA-
β
-gal)
ac i i y was measu ed using a comme cial s aining ki (9860S, Cell Signaling, Dan e s, MA,
USA) acco ding o he manu ac u e ’s p o ocol.
4.8. Cell Viabili y Assay
Cells we e seeded in 96-well pla es a a densi y o 2.5
×
10
3
cells pe well and ea ed
24 h la e wi h he indica ed concen a ions o STX-0119 (Sigma-Ald ich) o 72 h o a senic
ioxide (ATO; Sigma-Ald ich) o 48 h, in sex uplica es. As a con ol, cells we e ea ed
wi h he indica ed sol en o each d ug. Cells we e hen incuba ed wi h 0.25 mg/mL
3-(4,5-dime hyl hiazol-2-yl)-2,5-diphenyl e azolium b omide (MTT; Sigma-Ald ich) o
3 h. The o mazan p oduced by iable cells was dissol ed in 150
µ
L o dime hyl sul oxide
(DMSO; Sigma-Ald ich), and he abso bance was de e mined a 570 nm in a mic opla e
eade (Mul iskan Ascen The mo Elec on Co po a ion, Wal ham, MA, USA).
4.9. Quan i a i e Real-Time PCR
To al RNA was ex ac ed using T izol (Li e Technologies). Re e se ansc ip ion was
pe o med om 1
µ
g o o al RNA using andom p ime s and he High-Capaci y cDNA
Re e se T ansc ip ion Ki (Li e Technologies) acco ding o he manu ac u e ’s guidelines.
Quan i a i e eal- ime PCR (qPCR) was hen ca ied ou using Powe SYBR
®
G een Mas e
Mix (The mo Scien i ic, Wal ham, MA, USA), 10 mM o p ime s, and 20 ng o cDNA in
an ABI PRISM 7300 he mocycle (Applied Biosys ems, Wal ham, MA, USA). The
∆∆
CT
me hod was used o ela i e quan i ica ion o gene exp ession, using GAPDH as he
e e ence gene. P ime sequences a e a ailable upon eques .
4.10. Immunoblo
Immunoblo s we e pe o med as p e iously desc ibed [
50
]. The ollowing p ima y
an ibodies we e used: abbi polyclonal an i-SOX9 (1:2000 dilu ion; Millipo e, Bu ling on,
MA, USA), mouse monoclonal an i-STAT3 (1:1000 dilu ion; Cell Signaling), abbi mon-
oclonal an i-phospho STAT3 (Ty 705) (1:1000 dilu ion; Cell Signaling), abbi polyclonal
an i-PML (1:1000 dilu ion; Be hyl labo a o ies, Mon gome y, TX, USA), abbi polyclonal
an i-SOX2 (1:500 dilu ion; Millipo e), and mouse monoclonal an i-
β
ac in (1:2000 dilu ion;
Sigma-Ald ich). P ima y an ibodies we e de ec ed wi h ho se adish pe oxidase (HRP)-
linked an ibodies: Goa an i- abbi o goa an i-mouse (San a C uz Bio echnology, Dallas,
TX, USA). P o ein de ec ion was pe o med using he NOVEX
®
ECL sys em (In i ogen,
Ca lsbad, CA, USA).
4.11. Immuno luo escence
A o al o 2
×
10
4
cells we e seeded in Lab-Tek II Chambe Slides (The moFishe ),
ixed a e 24 h wi h 4% o malin o 10 min a oom empe a u e (RT), and blocked wi h
phospha e-bu e ed saline (PBS) supplemen ed wi h 0.3% T i on X-100 (Sigma-Ald ich)
and 5% FBS o 1 h a RT. Cells we e hen incuba ed o 2 h a RT wi h mouse monoclonal
an i-His one H3 (1:1000 dilu ion; Abcam, Camb idge, UK) and wi h Alexa Fluo
®
488
abbi an i-mouse IgG (H + L) seconda y an ibody (1:500 dilu ion; A-11059, In i ogen)
o 1 h a RT. A e ha , slides we e washed and moun ed wi h a Vec ashield Moun ing
Medium wi h DAPI (Vec o Labo a o ies, Bu lingame, CA, USA). Images we e ob ained
wi h a Nikon Eclipse 80i mic oscope.
In . J. Mol. Sci. 2022,23, 4511 17 o 20
4.12. In Vi o Ca cinogenesis Assays
Fo subcu aneous injec ion, cells we e esuspended in PBS. B ie ly, 1
×
10
6
cells pe
condi ion we e injec ed subcu aneously in o bo h lanks o 8-week-old Foxn1
nu
/Foxn1
nu
nude mice. Ex e nal calipe s we e used o measu e umo size; umo olume was hen
es ima ed by he ollowing o mula: V = L
×
W
2×
0.5 (L = umo leng h and W = umo
wid h). Fo xeno ansplan a ion, GSCs we e injec ed s e eo ac ically in o he on al co ex
o 6–8-week-old NOD-SCID mice. B ie ly, GSCs we e disagg ega ed wi h accu ase and
esuspended in PBS. Then, 1
×
10
4
o 1
×
10
5
cells in a inal olume o 1
µ
L we e injec ed
in o he pu amen using a s e eo axic appa a us (Kop Ins umen s). Fo umo ini ia ion
assays, U251-MG cells a a densi y o 5
×
10
4
o 5
×
10
5
cells pe condi ion we e injec ed
subcu aneously in o bo h lanks o 8-week-old Foxn1
nu
/Foxn1
nu
nude mice. To calcula e he
numbe o ini ia ing cells, we used Ex eme Limi ing Dilu ion Analysis (ELDA) so wa e
(h p://bioin .wehi.edu.au/so wa e/elda/, accessed on 10 Ma ch 2022).
All animal handling and p ocedu es we e pe o med acco ding o he e hical guide-
lines es ablished by he Animal Expe imen a ion E hics Commi ee o Biodonos ia Heal h
Resea ch Ins i u e (CEEA14/016) and conduc ed in con o mi y wi h he Eu opean Union
ecommenda ions o animal expe imen a ion speci ied in he Di ec i e 2010/63/EU.
4.13. Immunohis ochemis y
Tumo s gene a ed in mice we e dissec ed, ixed in 10% o malin o 48 h, and em-
bedded in pa a in. Fou mic ome e - hick sec ions we e s ained wi h hema oxylin–eosin
(H&E) using a Va is ain Gemini ES machine (The moFishe , Wal ham, MA, USA). Fo
immunohis ochemis y (IHC), sec ions we e ehyd a ed and hea ed in ci a e bu e pH 6
o 10 min o an igen e ie al. Endogenous pe oxidase was blocked wi h 5% hyd ogen
pe oxide in me hanol o 15 min. Sec ions we e incuba ed wi h he ollowing p ima y
an ibodies: abbi polyclonal an i-SOX9 (1:1000 dilu ion; Millipo e), mouse monoclonal
an i-phospho-STAT3 (Ty 705) (1:100 dilu ion; Cell Signaling), mouse monoclonal an i-PML
(PG-M3) (1:200 dilu ion; San a C uz Bio echnology), abbi polyclonal an i-SOX2 (1:500
dilu ion; Millipo e), and abbi polyclonal an i-Ki67 (1:1000 dilu ion; Abcam). Sec ions
we e hen incuba ed wi h MACH 3 Rabbi HRP-Polyme (BioCa e Medical). S aining was
de eloped wi h 3,3’-diaminobenzidine (DAB) (Sp ing Bioscience, Pleasan on, CA, USA).
IHC images we e ob ained wi h a Nikon Eclipse 80i mic oscope.
4.14. T ansc ip ome Analysis
Exp ession mic oa ays we e pe o med om 0.5
µ
g o RNA om SOX9-silenced
GNS166 cells and con ol GNS166 cells using he Gene Chip Human Genome U133 Plus 2.0
a ay (A yme ix). Da a we e no malized by Robus Mul ia ay A e age (RMA) using
A yme ix
®
Exp ession Console
™
so wa e V1.1. Di e en ial gene exp ession analysis was
ca ied ou wi h Genesp ing so wa e GeneSp ing GX 14.9 (Agilen , San a Cla a, CA, USA).
Pa hway analysis was pe o med wi h he In e ac i e Pa hway Analysis (IPA
®
) so wa e
(Ingenui y Sys ems, Redwood ci y, CA, USA, h ps://www.qiagenbioin o ma ics.com/
p oduc s/ingenui y-pa hway-analysis, accessed on 10 Ma ch 2022). The da a discussed
in his publica ion ha e been deposi ed in NCBI’s Gene Exp ession Omnibus and a e
accessible in GSE181035.
4.15. Ch oma in Immunop ecipi a ion Assay
ASimpleChIP Enzyma ic Ch oma in IP Ki (Cell Signaling) was used o he ch oma in
immunop ecipi a ion (ChIP) assay. In de ail, U251-MG cells ha bo ing a doxycycline-
inducible HA-PMLIV plasmid we e g own in 150 mm dishes o 3 days and c oss-linked
wi h 35% o maldehyde o 10 min a RT. Then, cells we e incuba ed o 5 min a RT a e he
addi ion o glycine, washed wi h PBS, and sc aped in o PBS wi h 100
µ
M phenylme hylsul-
onyl luo ide (PMSF). Pelle ed cells we e lysed and nuclei ha es ed. Nuclea lysa es we e
diges ed wi h mic ococcal nuclease o 20 min a 37
◦
C and hen sonica ed in 500
µ
L aliquo s
on ice o 3 pulses o 15 s using a B anson sonica o . Lysa es we e cla i ied, and ch oma in
In . J. Mol. Sci. 2022,23, 4511 18 o 20
was s o ed a
−
80
◦
C un il used. HA-Tag polyclonal an ibody (Cell Signaling), abbi
polyclonal an i-PML (Be hyl labo a o ies), and no mal abbi IgG polyclonal an ibody (Cell
Signaling) we e incuba ed o e nigh a 4
◦
C wi h o a ion. A e ha , p o ein G-conjuga ed
magne ic beads we e added and incuba ed o 2 h a 4
◦
C. Washes and elu ion o ch oma in
we e la e pe o med and DNA quan i ica ion was ca ied ou using a Viia7 Real-Time PCR
Sys em (Applied Biosys ems) wi h SYBR G een and p ime s ha ampli y he p edic ed
PML binding egion o SOX9 p omo e (ch 17:70117013-70117409). P ime sequences used
we e o wa d 50-ccggaaac c gcag-30and e e se 50-cggcgagcac aggaag-30.
4.16. Da a A ailabili y S a emen
The au ho s con i m ha he da a suppo ing he indings o his s udy a e a ailable
wi hin he a icle and i s Supplemen a y Ma e ials. The mic oa ay da a ha suppo he
indings o his s udy a e openly a ailable in NCBI’s Gene Exp ession Omnibus eposi o y.
4.17. Da a Analysis
Da a a e p esen ed as he mean
±
s anda d e o o he mean (SEM) wi h he numbe
o expe imen s (n) in pa en heses. Mean alues om quan i a i e a iables we e compa ed
using S uden ’s - es . Log- ank es was pe o med o Kaplan–Meie su i al analyses.
Co ela ions we e calcula ed using he Spea man coe icien . Chi-squa ed (X
2
) es was
used o compa e equency da a. As e isks (*, ** and ***) indica e s a is ical signi icance
(p< 0.05, p< 0.01 and p< 0.001, espec i ely).
Supplemen a y Ma e ials:
The ollowing suppo ing in o ma ion can be downloaded a : h ps:
//www.mdpi.com/a icle/10.3390/ijms23094511/s1.
Au ho Con ibu ions:
P.A., N.M.-M., A.S.-A. and E.C.-G. pe o med he expe imen s in glioma cells;
A.E.-P. helped wi h s e eo axic expe imen s
in i o
in mice; S.M.P. gene a ed pa ien -de i ed cells;
A.E.-P., M.M.-V., J.H. and N.S. collec ed clinical da a and samples om pa ien s, e alua ed hem, and
pe o med he clinical expe imen s; C.H.L. pe o med he ansc ip omic analysis; M.Á.-S. helped o
w i e he manusc ip ; R.L.-B. and A.C. coo dina ed he expe imen s, analyzed esul s, and helped o
di ec he p ojec ; A.M. di ec ed he p ojec , con ibu ed o da a analysis, ob ained unds, and w o e
he manusc ip . All au ho s ha e ead and ag eed o he published e sion o he manusc ip .
Funding: This esea ch ecei ed no ex e nal unding.
Ins i u ional Re iew Boa d S a emen : No applicable.
In o med Consen S a emen : No applicable.
Da a A ailabili y S a emen :
The da a ha suppo he indings o his s udy a e included in he
manusc ip and Supplemen a y Ma e ials. Mic oa ay da a a e openly a ailable in GEO da abase
wi h access numbe GSE181035.
Acknowledgmen s:
P.A. and A.S.-A. we e ecipien s o p edoc o al ellowships om he AECC
ounda ion and Ca los III Ins i u e (ISCIII), espec i ely. M.Á.-S. holds a Sa a Bo ell pos doc o al
con ac om he ISCIII (CD19/00154). E.C.-G. was a ecipien o a S op Fuga de Ce eb os pos doc o al
ellowship and holds a Miguel Se e con ac om he ISCIII (CP19/00085). We hank he His ology
Pla o m o he Biodonos ia Heal h Resea ch Ins i u e, The Neu o-Oncology Commi ee o Donos ia
Uni e si y Hospi al, and Basque Biobank o hei help. This esea ch was suppo ed by g an s
om ISCIII and FEDER Funds (CP16/00039, DTS16/00184, PI16/01580, DTS18/00181, PI18/01612,
CP19/00085), and he Indus y and Heal h Depa men s o he Basque Coun y.
Con lic s o In e es : The au ho s decla e no con lic o in e es .
In . J. Mol. Sci. 2022,23, 4511 19 o 20
Re e ences
1.
Louis, D.N.; Pe y, A.; Rei enbe ge , G.; on Deimling, A.; Figa ella-B ange , D.; Ca enee, W.K.; Ellison, D.W. The 2016 Wo ld
Heal h O ganiza ion Classi ica ion o Tumo s o he Cen al Ne ous Sys em: A summa y. Ac a Neu opa hol. 2016,131, 803–820.
[C ossRe ] [PubMed]
2.
Os om, Q.T.; Bauche , L.; Da is, F.G.; Del ou , I.; Fishe , J.L.; Lange , C.E.; Pekmezci, M.; Schwa zbaum, J.A.; Tu ne , M.C.; Walsh,
K.M.; e al. The epidemiology o glioma in adul s: A “s a e o he science” e iew. Neu o Oncol.
2014
,16, 896–913. [C ossRe ]
[PubMed]
3.
S upp, R.; Mason, W.P.; an den Ben , M.J.; Welle , M.; Fishe , B.; Taphoo n, M.J.B.; Belange , K.; B andes, A.A.; Ma osi, C.;
Bogdahn, U.; e al. Radio he apy plus concomi an and adju an emozolomide o glioblas oma. N. Engl. J. Med.
2005
,352,
987–996. [C ossRe ] [PubMed]
4.
Chen, J.; Li, Y.; Yu, T.-S.; McKay, R.M.; Bu ns, D.K.; Ke nie, S.G.; Pa ada, L.F. A es ic ed cell popula ion p opaga es glioblas oma
g ow h a e chemo he apy. Na u e 2012,488, 522–526. [C ossRe ]
5.
Bao, S.; Wu, Q.; McLendon, R.E.; Hao, Y.; Shi, Q.; Hjelmeland, A.B.; Dewhi s , M.W.; Bigne , D.D.; Rich, J.N. Glioma s em cells
p omo e adio esis ance by p e e en ial ac i a ion o he DNA damage esponse. Na u e 2006,444, 756–760. [C ossRe ]
6.
Lan, X.; Joe g, D.; Ca alli, F.; Richa ds, L.; Nguyen, L.; Vanne , R.; Guilhamon, P.; Lee, L.; Kushida, M.; Pellacani, D.; e al. Fa e
mapping o human glioblas oma e eals an in a ian s em cell hie a chy. Na u e 2017,549, 227–232. [C ossRe ]
7.
La hia, J.D.; Mack, S.C.; Mulkea ns-Hube , E.E.; Valen im, C.L.; Rich, J.N. Cance s em cells in glioblas oma. Genes De .
2015
,29,
1203–1217. [C ossRe ]
8.
Alcan a a Llaguno, S.; Chen, J.; Kwon, C.H.; Jackson, E.L.; Li, Y.; Bu ns, D.K.; Pa ada, L.F. Malignan as ocy omas o igina e om
neu al s em/p ogeni o cells in a soma ic umo supp esso mouse model. Cance Cell 2009,15, 45–56. [C ossRe ]
9.
Alcan a a Llaguno, S.R.; Wang, Z.; Sun, D.; Chen, J.; Xu, J.; Kim, E.; Pa ada, L.F. Adul Lineage-Res ic ed CNS P ogeni o s
Speci y Dis inc Glioblas oma Sub ypes. Cance Cell 2015,28, 429–440. [C ossRe ]
10.
Lee, J.H.; Lee, J.E.; Kahng, J.Y.; Kim, S.H.; Pa k, J.S.; Yoon, S.J.; Lee, J.H. Human glioblas oma a ises om sub en icula zone cells
wi h low-le el d i e mu a ions. Na u e 2018,560, 243–247. [C ossRe ]
11.
Pa el, A.P.; Ti osh, I.; T ombe a, J.J.; Shalek, A.K.; Gillespie, S.M.; Wakimo o, H.; Cahill, D.P.; Nahed, B.V.; Cu y, W.T.; Ma uza,
R.L.; e al. Single-cell RNA-seq highligh s in a umo al he e ogenei y in p ima y glioblas oma. Science
2014
,344, 1396–1401.
[C ossRe ] [PubMed]
12. Su à, M.L.; Riggi, N.; Be ns ein, B.E. Epigene ic ep og amming in cance . Science 2013,339, 1567–1570. [C ossRe ] [PubMed]
13. Su a, M.L.; Rheinbay, E.; Gillespie, S.M.; Pa el, A.P.; Wakimo o, H.; Rabkin, S.D.; Be ns ein, B.E. Recons uc ing and ep og am-
ming he umo -p opaga ing po en ial o glioblas oma s em-like cells. Cell 2014,157, 580–594. [C ossRe ] [PubMed]
14.
de la Rocha, A.M.; Samp on, N.; Alonso, M.M.; Ma heu, A. Role o SOX amily o ansc ip ion ac o s in cen al ne ous sys em
umo s. Am. J. Cance Res. 2014,4, 312–324.
15.
Ca en, H.; S icke , S.H.; Buls ode, H.; Gag ica, S.; Johns one, E.; Ba le , T.E.; Polla d, S.M. Glioblas oma S em Cells Respond o
Di e en ia ion Cues bu Fail o Unde go Commi men and Te minal Cell-Cycle A es . S em Cell Rep.
2015
,5, 829–842. [C ossRe ]
16.
MacLeod, G.; Bozek, D.A.; Rajakulend an, N.; Mon ei o, V.; Ahmadi, M.; S einha , Z.; Kushida, M.M.; Yu, H.; Cou inho, F.J.;
Ca alli, F.M.G.; e al. Genome-Wide CRISPR-Cas9 Sc eens Expose Gene ic Vulne abili ies and Mechanisms o Temozolomide
Sensi i i y in Glioblas oma S em Cells. Cell Rep. 2019,27, 971–986.e9. [C ossRe ]
17.
P i che , J.; A hwal, V.; Robe s, N.; Hanley, N.A.; Hanley, K.P. Unde s anding he ole o SOX9 in acqui ed diseases: Lessons
om de elopmen . T ends Mol. Med. 2011,17, 166–174. [C ossRe ]
18.
Cheng, L.-C.; Pas ana, E.; Ta azoie, M.; Doe sch, F. miR-124 egula es adul neu ogenesis in he sub en icula zone s em cell
niche. Na . Neu osci. 2009,12, 399–408. [C ossRe ]
19.
Sco , C.E.; Wynn, S.L.; Sesay, A.; C uz, C.; Cheung, M.; Gomez-Ga i o, M.V.; Boo h, S.; Gao, B.; Cheah, K.S.E.; Lo ell-Badge, R.;
e al. SOX9 induces and main ains neu al s em cells. Na . Neu osci. 2010,13, 1181–1189. [C ossRe ]
20.
Swa ling, F.J.; Sa o , V.; Pe sson, A.I.; Chen, J.; Hacke , C.S.; No hco , P.A.; G imme , M.R.; Lau, J.; Chesle , L.; Pe y, A.; e al.
Dis inc neu al s em cell popula ions gi e ise o dispa a e b ain umo s in esponse o N-MYC. Cance Cell
2012
,21, 601–613.
[C ossRe ]
21.
Ma heu, A.; Collado, M.; Wise, C.; Man e ola, L.; Cekai e, L.; Tye, A.J.; Caname o, M.; Bujanda, L.; Schedl, A.; Cheah, K.S.; e al.
Oncogenici y o he de elopmen al ansc ip ion ac o Sox9. Cance Res. 2012,72, 1301–1315. [C ossRe ] [PubMed]
22.
Guo, W.; Keckeso a, Z.; Donahe , J.L.; Shibue, T.; Tischle , V.; Reinha d , F.; Weinbe g, R.A. Slug and Sox9 coope a i ely de e mine
he mamma y s em cell s a e. Cell 2012,148, 1015–1028. [C ossRe ] [PubMed]
23.
Ga os-Regulez, L.; Aldaz, P.; A izabalaga, O.; Moncho-Amo , V.; Ca asco-Ga cia, E.; Man e ola, L.; Ma heu, A. mTOR inhibi ion
dec eases SOX2-SOX9 media ed glioma s em cell ac i i y and emozolomide esis ance. Expe Opin. The . Ta ge s
2016
,20,
393–405. [C ossRe ] [PubMed]
24.
Wang, L.; He, S.; Yuan, J.; Mao, X.-G.; Cao, Y.; Zong, J.; Tu, Y.; Zhang, Y. Oncogenic ole o SOX9 exp ession in human malignan
glioma. Med. Oncol. 2012,29, 3484–3490. [C ossRe ]
25.
Aldaz, P.; O aegi-Uga emendia, M.; Saenz-An oñanzas, A.; Ga cía-Puga, M.; Mo eno-Vallada es, M.; Flo es, J.M.; Ge o ska, D.;
A auzo-B a o, M.J.; Samp ón, N.; Ma heu, A.; e al. SOX9 p omo es umo p og ession h ough he axis BMI1-p21(CIP). Sci. Rep.
2020,10, 357. [C ossRe ]
In . J. Mol. Sci. 2022,23, 4511 20 o 20
26.
She y, M.M.; Ree es, A.; Wu, J.K.; Coch an, B.H. STAT3 is equi ed o p oli e a ion and main enance o mul ipo ency in
glioblas oma s em cells. S em Cells 2009,27, 2383–2392. [C ossRe ]
27.
Hubacko a, S.; K ejciko a, K.; Ba ek, J.; Hodny, Z. In e leukin 6 signaling egula es p omyelocy ic leukemia p o ein gene
exp ession in human no mal and cance cells. J. Biol. Chem. 2012,287, 26702–26714. [C ossRe ]
28.
Ma in-Ma in, N.; Pi a, M.; U ose ic, J.; Aldaz, P.; Su he land, J.D.; Fe nandez-Ruiz, S.; Ca acedo, A. S a i ica ion and
he apeu ic po en ial o PML in me as a ic b eas cance . Na . Commun. 2016,7, 12595. [C ossRe ]
29.
B ennan, C.W.; Ve haak, R.G.; McKenna, A.; Campos, B.; Noushmeh , H.; Salama, S.R.; Da idsen, T. The soma ic genomic
landscape o glioblas oma. Cell 2013,155, 462–477. [C ossRe ]
30.
Lallemand-B ei enbach, V.; Jeanne, M.; Benhenda, S.; Nas , R.; Lei, M.; Pe es, L.; de The, H. A senic deg ades PML o PML-
RARalpha h ough a SUMO- igge ed RNF4/ubiqui in-media ed pa hway. Na . Cell Biol. 2008,10, 547–555. [C ossRe ]
31.
Adam, R.C.; Yang, H.; Rockowi z, S.; La sen, S.; Nikolo a, M.; O is ian, D.S.; Polak, L.; Kadaja, M.; Asa e, A.; Zheng, D.; e al.
Pionee ac o s go e n supe -enhance dynamics in s em cell plas ici y and lineage choice. Na u e
2015
,521, 366–370. [C ossRe ]
[PubMed]
32.
Wang, Z.; Xu, X.; Liu, N.; Cheng, Y.; Jin, W.; Zhang, P.; Wang, X.; Yang, H.; Liu, H.; Tu, Y. SOX9-PDK1 axis is essen ial o glioma
s em cell sel - enewal and emozolomide esis ance. Onco a ge 2017,9, 192–204. [C ossRe ] [PubMed]
33.
Gao, J.; Zhang, J.-Y.; Li, Y.-H.; Ren, F. Dec eased exp ession o SOX9 indica es a be e p ognosis and inhibi s he g ow h o glioma
cells by inducing cell cycle a es . In . J. Clin. Exp. Pa hol. 2015,8, 10130–10138. [PubMed]
34. Rahaman, S.O.; Ha bo , P.C.; Che no a, O.; Ba ne , G.H.; Vogelbaum, M.A.; Haque, S.J. Inhibi ion o cons i u i ely ac i e S a 3
supp esses p oli e a ion and induces apop osis in glioblas oma mul i o me cells. Oncogene
2002
,21, 8404–8413. [C ossRe ]
[PubMed]
35.
Tan, M.S.Y.; Sandana aj, E.; Chong, Y.K.; Lim, S.W.; Koh, L.W.H.; Ng, W.H.; Tang, C. A STAT3-based gene signa u e s a i ies
glioma pa ien s o a ge ed he apy. Na . Commun. 2019,10, 3601. [C ossRe ] [PubMed]
36.
Bi ne , P.; Toumangelo a-Uzei , K.; Na che , S.; Guen che , M. STAT3 y osine phospho yla ion in luences su i al in glioblas-
oma. J. Neu ooncol. 2010,100, 339–343. [C ossRe ] [PubMed]
37.
I o, K.; Be na di, R.; Mo o i, A.; Ma suoka, S.; Saglio, G.; Ikeda, Y.; Rosenbla , J.; A igan, D.E.; Te uya-Felds ein, J.; Pandol i, P.P.
PML a ge ing e adica es quiescen leukaemia-ini ia ing cells. Na u e 2008,453, 1072–1078. [C ossRe ]
38.
Ca acedo, A.; Weiss, D.; Leliae , A.K.; Bhasin, M.; de Boe , V.; Lau en , G.; Adams, A.C.; Sund all, M.; Song, S.J.; I o, K.; e al. A
me abolic p osu i al ole o PML in b eas cance . J. Clin. In es ig. 2012,122, 3088–3100. [C ossRe ]
39.
Amodeo, V.; Deli, A.; Be s, J.; Ba esaghi, S.; Zhang, Y.; Richa d-Lond , A.; Ellis, M.; Roshani, R.; Vou i, M.; Gala o i, S.; e al. A
PML/Sli Axis Con ols Physiological Cell Mig a ion and Cance In asion in he CNS. Cell Rep. 2017,20, 411–426. [C ossRe ]
40.
Ashizawa, T.; Miya a, H.; Iizuka, A.; Komiyama, M.; Oshi a, C.; Kume, A.; Akiyama, Y. E ec o he STAT3 inhibi o STX-0119 on
he p oli e a ion o cance s em-like cells de i ed om ecu en glioblas oma. In . J. Oncol. 2013,43, 219–227. [C ossRe ]
41.
Akiyama, Y.; Nonomu a, C.; Ashizawa, T.; Iizuka, A.; Kondou, R.; Miya a, H.; Sugino, T.; Mi suya, K.; Hayashi, N.; Nakasu, Y.;
e al. The an i- umo ac i i y o he STAT3 inhibi o STX-0119 occu s ia p omo ion o umo -in il a ing lymphocy e accumula ion
in emozolomide- esis an glioblas oma cell line. Immunol. Le . 2017,190, 20–25. [C ossRe ] [PubMed]
42.
Fuh, B.; Sobo, M.; Cen, L.; Josiah, D.; Hu zen, B.; Cisek, K.; Bhasin, D.; Regan, N.; Lin, L.; Chan, C.; e al. LLL-3 inhibi s STAT3
ac i i y, supp esses glioblas oma cell g ow h and p olongs su i al in a mouse glioblas oma model. B . J. Cance
2009
,100,
106–112. [C ossRe ] [PubMed]
43.
Jensen, K.V.; Cseh, O.; Aman, A.; Weiss, S.; Luchman, H.A. The JAK2/STAT3 inhibi o pac i inib e ec i ely inhibi s pa ien -
de i ed GBM b ain umo ini ia ing cells
in i o
and when used in combina ion wi h emozolomide inc eases su i al in an
o ho opic xenog a model. PLoS ONE 2017,12, e0189670. [C ossRe ] [PubMed]
44.
Zhou, W.; Cheng, L.; Shi, Y.; Ke, S.Q.; Huang, Z.; Fang, X.; Chu, C.-W.; Xie, Q.; Bian, X.-W.; Rich, J.N.; e al. A senic ioxide
dis up s glioma s em cells ia p omo ing PML deg ada ion o inhibi umo g ow h. Onco a ge
2015
,6, 37300–37315. [C ossRe ]
[PubMed]
45.
Ve haak, R.G.W.; Hoadley, K.A.; Pu dom, E.; Wang, V.; Wilke son, M.D.; Mille , C.R.; Ding, L.; Golub, T.; Jill, P.; Alexe, G.; e al.
In eg a ed genomic analysis iden i ies clinically ele an sub ypes o glioblas oma cha ac e ized by abno mali ies in PDGFRA,
IDH1, EGFR, and NF1. Cance Cell 2010,17, 98–110. [C ossRe ] [PubMed]
46.
Wang, Z.; Sun, D.; Chen, Y.-J.; Xie, X.; Shi, Y.; Taba , V.; B ennan, C.W.; Bale, T.A.; Jayewick eme, C.D.; Laks, D.R.; e al. Cell
Lineage-Based S a i ica ion o Glioblas oma. Cance Cell 2020,38, 366–379.e8. [C ossRe ] [PubMed]
47.
Gangoso, E.; Sou hga e, B.; B adley, L.; Rus, S.; Gal ez-Cancino, F.; McGi e n, N.; Güç, E.; Kapou ani, C.-A.; By on, A.; Fe guson,
K.M.; e al. Glioblas omas acqui e myeloid-a ilia ed ansc ip ional p og ams ia epigene ic immunoedi ing o elici immune
e asion. Cell 2021,184, 2454–2470.e26. [C ossRe ]
48.
Cappe , D.; Jones, D.T.W.; Sill, M.; Ho es ad , V.; Sch imp , D.; S u m, D.; Koelsche, C.; Sahm, F.; Cha ez, L.; Reuss, D.E.; e al.
DNA me hyla ion-based classi ica ion o cen al ne ous sys em umou s. Na u e 2018,555, 469–474. [C ossRe ]
49.
Polla d, S.M.; Yoshikawa, K.; Cla ke, I.D.; Dano i, D.; S icke , S.; Russell, R.; Bayani, J.; Head, R.; Lee, M.; Be ns ein, M.; e al.
Glioma s em cell lines expanded in adhe en cul u e ha e umo -speci ic pheno ypes and a e sui able o chemical and gene ic
sc eens. Cell S em Cell 2009,4, 568–580. [C ossRe ]
50.
Ma heu, A.; Kla , P.; Se ano, M. Regula ion o he INK4a/ARF locus by his one deace ylase inhibi o s. J. Biol. Chem.
2005
,280,
42433–42441. [C ossRe ]
