scieee Science in your language
[en] (orig)

Differential regulation of AMP-activated protein kinase in healthy and cancer cells explains why V-ATPase inhibition selectively kills cancer cells.

Abstract

The cellular energy sensor AMP-activated protein kinase (AMPK) is a metabolic hub regulating various pathways involved in tumor metabolism. Here, we report that vacuolar H+-ATPase (V-ATPase) inhibition differentially affects regulation of AMPK in tumor and non-tumor cells and that this differential regulation contributes to the selectivity of V-ATPase inhibitors for tumor cells. In non-malignant cells, the V-ATPase inhibitor archazolid increased phosphorylation and lysosomal localization of AMPK. We noted that AMPK localization has a pro-survival role, as AMPK silencing decreased cellular growth rates. In contrast, in cancer cells, we found that AMPK is constitutively active and that archazolid does not affect its phosphorylation and localization. Moreover, V-ATPase-independent AMPK induction in the tumor cells protected them from archazolid-induced cytotoxicity, further underlining the role of AMPK as a pro-survival mediator. These observations indicate that AMPK regulation is uncoupled from V-ATPase activity in cancer cells and that this makes them more susceptible to cell death induction by V-ATPase inhibitors. In both tumor and healthy cells, V-ATPase inhibition induced a distinct metabolic regulatory cascade downstream of AMPK, affecting ATP and NADPH levels, glucose uptake, and reactive oxygen species (ROS) production. We could attribute the pro-survival effects to AMPK's ability to maintain redox homeostasis by inhibiting ROS production and maintaining NADPH levels. In summary, the results of our work indicate that V-ATPase inhibition has differential effects on AMPK-mediated metabolic regulation in cancer and healthy cells and explain the tumor-specific cytotoxicity of V-ATPase inhibition.

Read accessible full text

Differential regulation of AMP-activated protein kinase in healthy and cancer cells explains why V-ATPase inhibition selectively kills cancer cells.

Author: Bartel, Karin,Müller, Rolf,von Schwarzenberg, Karin
Publisher: American Society for biochemistry and Molecular Biology
Year: 2019
DOI: 10.1074/jbc.RA119.010243
Source: https://repository.helmholtz-hzi.de/bitstream/10033/622016/1/Bartel%20et%20al%20plus%20supp_info.pdf
Tumo -speci ic e ec o V-ATPase inhibi ion
1
Di e en ial egula ion o AMP-ac i a ed p o ein kinase in heal hy and cance cells
explains why V-ATPase inhibi ion selec i ely kills cance cells
Ka in Ba el1, Rol Mülle 2, Ka in on Schwa zenbe g1
1Depa men o Pha macy, Pha maceu ical Biology, Ludwig-Maximilians-Uni e si y o Munich,
Munich, Ge many.
2Helmhol z Cen e o Pha maceu ical Resea ch Saa land, Helmhol z Cen e o In ec ion Resea ch and
Depa men o Pha macy a Saa land Uni e si y, Saa land Uni e si y Campus, Saa b ücken, Ge many.
Running i le: Tumo -speci ic e ec o V-ATPase inhibi ion
To whom co espondence may be add essed: Ka in on Schwa zenbe g, Depa men o Pha macy,
Pha maceu ical Biology, LMU Munich, Bu enand s asse 5-13, 81377 Munich, Ge many; Tel.: 0049-
89-218077165; Fax: 0049-89-218077170; E-mail: ka in. on.schwa [email p o ec ed]
Keywo ds: V-ATPase, AMPK, umo me abolism, ROS, apop osis, cance , a chazolid, umo
supp esso , glucose s a a ion, pH homeos asis
Abs ac
The cellula ene gy senso AMP-
ac i a ed p o ein kinase (AMPK) is a
me abolic hub egula ing a ious pa hways
in ol ed in umo me abolism. He e, we epo
ha acuola H+-ATPase (V-ATPase)
inhibi ion di e en ially a ec s egula ion o
AMPK in umo and non- umo cells and ha
his di e en ial egula ion con ibu es o he
selec i i y o V-ATPase inhibi o s o umo
cells. In non-malignan cells, he V-ATPase
inhibi o a chazolid inc eased phospho yla ion
and lysosomal localiza ion o AMPK. We
no ed ha AMPK localiza ion has a p o-
su i al ole, as AMPK silencing dec eased
cellula g ow h a es. In con as , in cance
cells, we ound ha AMPK is cons i u i ely
ac i e and ha a chazolid does no a ec i s
phospho yla ion and localiza ion. Mo eo e ,
V-ATPase–independen AMPK induc ion in
he umo cells p o ec ed hem om
a chazolid-induced cy o oxici y, u he
unde lining he ole o AMPK as a p o-
su i al media o . These obse a ions indica e
ha AMPK egula ion is uncoupled om V-
ATPase ac i i y in cance cells and ha his
makes hem mo e suscep ible o cell dea h
induc ion by V-ATPase inhibi o s. In bo h
umo and heal hy cells, V-ATPase inhibi ion
induced a dis inc me abolic egula o y
cascade downs eam o AMPK, a ec ing ATP
and NADPH le els, glucose up ake, and
eac i e oxygen species (ROS) p oduc ion.
We could a ibu e he p o-su i al e ec s o
AMPK’s abili y o main ain edox
homeos asis by inhibi ing ROS p oduc ion and
main aining NADPH le els. In summa y, he
esul s o ou wo k indica e ha V-ATPase
inhibi ion has di e en ial e ec s on AMPK-
media ed me abolic egula ion in cance and
heal hy cells and explain he umo -speci ic
cy o oxici y o V-ATPase inhibi ion.
Al eady in he 1920ies O o Wa bu g
disco e ed ha umo cells show an al e ed
me abolism by using glycolysis as he main
ene gy sou ce e en in he p esence o oxygen,
named ae obic glycolysis o he Wa bu g
e ec . Since hen many modi ica ions in
oncogenes and umo supp esso s like
HIF1alpha, Ak , Ras o p53 could be di ec ly
connec ed o egula e he exp ession and
ac i i y o impo an componen s o umo
me abolism, p esen ing umo me abolism as
one hallma k o cance (1). Bu despi e
ongoing esea ch o iden i y he di e en
aspec s o cance me abolism, he me abolic
al e a ions ha a e c i ical o umo
p og ession emain la gely unknown.
Recen ly, he highly conse ed ene gy senso
AMP ac i a ed p o ein kinase (AMPK) came
in o ocus as me abolic hub showing o
egula e many di e en pa hways in ol ed in
umo me abolism. I belongs o a amily o
se ine/ h eonine kinases and consis s o a
ca aly ic α subuni and egula o y β and γ
subuni s (2). AMPK is ac i a ed by a a ie y
o me abolic o oncogenic s ess condi ions
h p://www.jbc.o g/cgi/doi/10.1074/jbc.RA119.010243The la es e sion is a
JBC Pape s in P ess. Published on Oc obe 11, 2019 as Manusc ip RA119.010243
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
Tumo -speci ic e ec o V-ATPase inhibi ion
2
like nu ien s a a ion o hypoxia and di ec s
he cell owa ds me abolic changes ha
p oduce ATP and es o e ene gy homeos asis
(3). By sensing he AMP:ATP a io i can
inc ease ca abolic p ocesses ha gene a e ATP
like a y acid oxida ion and glycolysis and
inhibi anabolic p ocesses ha consume ATP
such as p o ein and lipid syn hesis. Based on
he inding ha LKB1, he ups eam kinase
ac i a ing AMPK, is equen ly inac i a ed in
umo cells AMPK has been his o ically
p oposed as a umo supp esso . This
assump ion was con i med by showing ha
gene ic loss o AMPK accele a es umo
g ow h o an expe imen al model o
lymphangioma and by he ac ha
pha macological ac i a ion o AMPK leads o
g ow h inhibi ion o se e al umo cell lines
(4,5). Howe e , accumula ing e idence shows
ha AMPK has a dual ole in cance and can
ha e a p o- umo igenic ole as shown o
di e en umo s, especially unde s ess
condi ions like glucose dep i a ion o
oxida i e s ess (6,7).
In e es ingly, i was ecen ly disco e ed ha
he acuola H+-ATPase (V-ATPase) is
needed o ac i a ion o AMPK du ing glucose
s a a ion by o ming a complex wi h he
sca olding p o ein AXIN1 and LKB1 (8). The
V-ATPase is a highly conse ed mul isubuni
p o on pump loca ed a endolysosomal
memb anes o mos euka yo ic cells. I is
esponsible o main aining pH homeos asis
and is essen ial o in acellula a icking and
ecep o ecycling (9). I has been in oduced
as a p omising an i- umo a ge as inhibi ion
o he V-ATPase by se e al d ugs like
concanamycin o he myxobac e ial
compound a chazolid leads o apop osis
induc ion, inhibi ion o mig a ion and in asion
in a a ie y o umo cells in i o and in i o
wi hou a ec ing non-malignan cells (10,11).
Fu he mo e, i was also e ealed la ely ha
he V-ATPase has addi ional unc ions besides
jus egula ing pH and endocy osis. Apa om
being in ol ed in AMPK egula ion i was
shown o play a ole in mTOR-media ed amino
acid sensing and leads o an induc ion o
glycolysis (12,13), e ealing o play a ole in
me abolism.
The impo an ole o V-ATPase o AMPK
homeos asis and he con o e sial discussion
o AMPK ac i i y in umo con ex p omp ed
us o in es iga e he connec ion o V-ATPase
inhibi ion and AMPK ac i a ion in umo cell
su i al in de ail.
In his s udy we used a chazolid as a highly
po en ool o speci ically block he V-ATPase
and ound a di e en ial e ec on AMPK
ac i a ion in umo and non- umo cells which
esul s in di e en me abolic egula ion and
sensi i i y o apop osis induc ion. In non-
umo cells, AMPK is mos ly inac i e,
ea men wi h a chazolid howe e , lead o a
p o ound ac i a ion o AMPK ha ing a
p o ec i e e ec agains oxida i e s ess
induced by he d ug. Tumo cells on he
con a y, showed a cons i u i e ac i a ion o
AMPK i espec i e o a chazolid ea men ,
ye V-ATPase independen ac i a ion o
AMPK also p o ec ed om apop osis
induc ion. We p opose ha AMPK egula ion
in umo cells is uncoupled om V-ATPase
unc ion dep i ing hem om AMPK-
media ed p o ec ion and ende s hem mo e
sensi i e o cy o oxici y induced by V-ATPase
inhibi o s. Hence, dis inc AMPK egula ion
in cance and non-malignan cells accoun s o
he umo cell speci ici y o V-ATPase
inhibi o s.
Resul s
V-ATPase inhibi ion ac i a es AMPK in non-
umo cells
To es whe he inhibi ion o V-
ATPase leads o AMPK ac i a ion we ea ed
di e en umo (MDA-MB-231, MCF7, T24,
HUH7) and non- umo cells (HEK293,
MCF10A, HMLE) wi h a chazolid and
analyzed phospho yla ion o AMPK on
Th 172. We ound ha all umo cells ha e
cons i u i ely ac i a ed AMPK and ha
a chazolid had no e ec on he ac i a ion le el
(Fig. 1A and S1A). As onishingly howe e , V-
ATPase inhibi ion lead o a p o ound
ac i a ion o AMPK in all used non- umo
cells, which ypically showed a low basal
AMPK ac i a ion (Fig. 1B and S1B). To es
whe he his e ec is speci ic o V-ATPase
inhibi ion o a gene al s ess induced e ec we
s a ed he cells o glucose and subsequen ly
analyzed AMPK ac i a ion. As shown in
Figu es 1C and S1C, he e is an ac i a ion o
AMPK in all es ed cell lines and no di e ence
be ween umo and non- umo cells. These
esul s sugges ha he di e en ial e ec seen
on AMPK phospho yla ion is no induced by a
gene al s ess esponse bu speci ic o V-
ATPase inhibi ion. To u he con i m
a chazolid-induced AMPK ac i a ion we
analyzed phospho yla ion o ace yl CoA
ca boxylase (ACC). As a downs eam a ge o
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
Tumo -speci ic e ec o V-ATPase inhibi ion
3
AMPK ACC is a good ma ke o AMPK
ac i i y. He e we could show he same e ec
han wi h AMPK – an ac i a ion in non- umo
HEK293 cells and no change in MDA-MB-
231 cells (Fig. S2). To make su e ha V-
ATPase is inhibi ed in all cell lines o he same
ex end we used a pH sensi i e lyso acke and
could show a simila e ec by a chazolid
ea men (Fig. S3). To u he analyze AMPK
s a us in umo and non- umo issue we used
pa ien de i ed b eas - umo samples and
no mal b eas issue om he same pa ien s,
espec i ely, and analyzed ac i i y o AMPK
ia Wes e n Blo . In e es ingly, we ound ha
umo cells showed an inc eased le el o o al
AMPK and also a sligh ly highe le el o
phospho yla ed AMPK (Fig. 1D and S1D),
indica ing ha AMPK ac i a ion is bene icial
o cance cells.
In ol emen o V-ATPase inhibi ion in
AMPK ac i a ion
I is known ha V-ATPase in e ac s
wi h he sca olding p o ein AXIN1 in non-
umo cells s a ed o glucose a he lysosome
and ha his in e ac ion is equi ed o ac i a e
AMPK (8). The e o e, we es ed i a chazolid
leads o an in e ac ion o he V-ATPase and
AXIN1 in umo cells. We ound ha
ea men wi h a chazolid leads o an inc eased
in e ac ion o AXIN1 and he V-ATPase in he
non- umo cell line HEK293 bu no in he
b eas cance cell line MDA-MB-231 (Fig.
2A). A simila e ec could be shown o
colocaliza ion o AXIN1 and lysosomes (Fig.
S4). Fu he mo e, using lysosomal
ac iona ion we could show ha AMPK is
inc easingly loca ed a he lysosome in
a chazolid ea ed HEK293 cells bu no in
MDA-MB-231 cells (Fig. 2B). Fo AMPK
ac i a ion i is essen ial ha also LKB1 is
loca ed a he lysosome. We inally ound ha
a chazolid leads o a ans e o LKB1 o he
lysosome in HEK293 cells in con as o umo
cells whe e no di e ence be ween ea ed and
non- ea ed cells could be obse ed, as shown
by con ocal mic oscopy (Fig. 2C). These
esul s u he sugges ha he e is a di e ence
in AMPK ac i a ion by V-ATPase inhibi ion
in umo and non- umo cells.
V-ATPase inhibi ion al e s AMPK ela ed
me abolic pa ame e s
AMPK is ac i a ed due o an inc eased
a io o AMP:ATP, he e o e we analyzed he
e ec o V-ATPase inhibi ion on ATP con en .
We ound ha umo cells we e only sligh ly
a ec ed bu he non- umo cell lines HEK293
and MCF10A showed a signi ican educ ion
o ATP le el a e a chazolid ea men as
shown by a luminescen assay (Fig 3A).
Fu he mo e, classical AMPK ac i a ion is
induced by an inc eased le el o AMP.
The e o e we analyzed changes in AMP by
HPLC (Fig. S5). In e es ingly we could no
ind a clea inc ease nei he in non- umo
HEK293 cells, no in MDA-MB-231 cells –
which migh indica e a di e en ac i a ion
mode by V-ATPase inhibi ion. Bu we ound
an inc eased glucose up ake and a sligh bu
no signi ican up egula ion o Glucose
ecep o 1 (GLUT1) in HEK293 cells – which
can be a consequence o AMPK ac i a ion. In
con as , he umo cell lines showed only a
sligh e ec (MDA-MB-231) o e en a
dec ease (MCF7) in glucose up ake (Fig. 3B,
C) showing a dis inc me abolic egula ion o
umo and non- umo cells by V-ATPase
inhibi ion.
AMPK ac i a ion p o ec s non- umo cells
om a chazolid induced cy o oxici y
As AMPK ac i a ion is qui e
con o e sially discussed as ei he umo
supp esso o oncogene we analyzed i
a chazolid induced AMPK phospho yla ion is
p o-su i al o p o-apop o ic. The e o e, we
silenced AMPK in HEK293 cells (Fig. 4B)
and ea ed he cells wi h a chazolid. We ound
ha HEK293 cells wi h silenced AMPK
showed a dec eased g ow h a e a e
a chazolid ea men compa ed o con ol cells
(Fig. 4A) which sugges s a p o ec i e ole o
AMPK ac i i y.
I AMPK ac i a ion p o ec s cells om
a chazolid induced cy o oxici y, hen u he
inducing AMPK in umo cells should
dec ease a chazolid induced apop osis
induc ion. The e o e, we ea ed MDA-MB-
231 cells wi h he AMP analogue AICAR in
combina ion wi h a chazolid which inc eased
phospho yla ion o AMPK (Fig. 4C and S6A).
In e es ingly we ound ha apop osis
induc ion was signi ican ly educed by
combining a chazolid wi h AICAR compa ed
o a chazolid ea ed cells alone (Fig 4D). This
was accompanied by a dec eased clea age o
PARP-1 con i ming a dec eased apop osis
induc ion and a p o ec i e ole o AMPK
(Fig. 4E and S6B). To con i m ha his e ec
is eally AMPK dependen we epea ed he
expe imen wi h a silenced AMPKα1 and
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
Tumo -speci ic e ec o V-ATPase inhibi ion
4
ound ha AICAR does no p o ec MDA-
MB-231 cells om a chazolid induced
apop osis con i ming he p o ec i e ole o
AMPK (Fig 4F and S6C).
P o-su i al ole o AMPK esul s om
di e en s ess esponse and main aining
edox homeos asis
As one known p o-su i al ole o
AMPK is he con ol o edox homeos asis by
p e en ing ROS accumula ion and
main aining NADPH le els, we analyzed i
a chazolid ea men leads o a dis u bance o
edox homeos asis. As shown in Fig. 5A, V-
ATPase inhibi ion leads o an induc ion o
ROS in umo cells bu a he o a dec ease o
ROS in non- umo cells. Fu he mo e,
a chazolid lead o an inc ease o
NADPH/NADP a io in HEK293 cells while
MDA-MB-231 cells showed a dec eased a io
(Fig. 5B) sugges ing a dis u bed edox balance
in umo cells. Ac i a ion o AMPK keeps he
edox balance by dec easing a y acid
syn hesis (7). The e o e, we analyzed he
ansc ip ion o FASN and SCD1 – wo majo
playe s in he a y acid syn hesis pa hway – in
umo and non- umo cells. In e es ingly we
ound ha a chazolid ea men lead o a s ong
up egula ion o bo h enzymes in umo cells
bu no in non- umo cells (Fig. 5C). To
connec hese e ec s wi h he p o-su i al
e ec o AMPK we ea ed MDA-MB-231
cells wi h he ROS sca enge Ti on and
analyzed apop osis induc ion (Fig 5D). We
ound ha combina ion o a chazolid wi h
Ti on dec eased apop osis induc ion. Fu he
suppo ing his hypo hesis is ha he
combina ion o AICAR and a chazolid
dec eases ROS p oduc ion in MDA-MB-231
cells (Fig. 5E). Finally, silencing o AMPK in
HEK293 cells leads o an up egula ion o ROS
a e a chazolid ea men con i ming ha
AMPK ac i a ion is esponsible o he
p o ec ion o ROS induced apop osis ia V-
ATPase inhibi ion (Fig. 5F).
Discussion
This wo k p o ides e idence ha V-
ATPase inhibi ion by a chazolid leads o a
di e en ial me abolic egula ion in umo and
non- umo cells which esul s in an inc eased
sensi i i y o umo cells owa ds he
ea men . Ou majo indings a e ha V-
ATPase inhibi ion leads o AMPK ac i a ion
only in heal hy cells which p o ec s hem om
a chazolid-induced cy o oxici y. In umo
cells, as depic ed in Figu e 6, his p o ec ion is
missing which esul s in inc eased apop osis
induc ion caused by a dis inc e ec on AMPK
downs eam cascade including ATP, glucose
up ake, NADPH le el and ROS p oduc ion.
As hese e ec s could be ab oga ed by
ac i a ing AMPK in umo cells independen ly
o V-ATPase, a no el ole o AMPK in V-
ATPase inhibi ion induced cy o oxici y was
un a eled. This p o ides a no el in e es ing
insigh in o he egula ion o s ess esponses
in di e en cell ypes and a be e
unde s anding o he mechanism o ac ion o
V-ATPase inhibi ion ega ding me abolism.
One majo p oblem in cance he apy is o ind
compounds ha speci ically a ge cance cells
and lea e non-malignan cells una ec ed and
he eby minimizing oxic side e ec s and
make chemo he apy mo e bea able o cance
pa ien s. In ensi e esea ch has led o he
in oduc ion o se e al compounds ha a ge
e.g. g ow h ac o ecep o s ha a e
o e exp essed in umo s like ce uximab
(EGFR) (14), as uzumab (HER2) (15) o
ima inib which a ge s he bc -abl y osinkase,
a umo -speci ic mu a ion (16). Bu despi e he
success expe ienced wi h hese compounds
selec i i y o chemo he apeu ics is s ill a
majo challenge.
The V-ATPase is a p omising an i- umo
a ge and i s inhibi ion was shown o ha e a
p onounced e ec on umo cells, bu lea es
non- umo cells almos una ec ed as we and
o he s could al eady show (11,17). We could
e en in oduce mo e umo and non- umo
cells lines o p o e a clea di e ence in
sensi i i y owa ds V-ATPase inhibi ion in
umo e sus non- umo cells (Fig. S7). The
an i- umo e ec o V-ATPase inhibi ion was
mos ly con ibu ed o he dis u bance o
endocy o ic ecycling p ocesses and se e al
publica ions by ou g oup and o he s could
elucida e se e al unde lying mechanisms
(11,13,18,19). In ecen yea s howe e , i
became ob ious ha he V-ATPase also plays
a pi o al ole in nu ien sensing. Zoncu e al.
showed ha he V-ATPase is ine i able o
mTOR-media ed amino acid sensing a he
lysosome, by building a complex be ween
mTORC1, he V-ATPase and Ragula o (12).
In e es ingly, Zhang e al. ound ha he V-
ATPase is no only equi ed o mTOR
egula ion bu also o AMPK ac i a ion unde
glucose dep i a ion. Thei g oup imp essi ely
showed ha he V-ATPase-Ragula o complex
se es as a docking si e o LKB1-media ed
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
Tumo -speci ic e ec o V-ATPase inhibi ion
5
AMPK ac i a ion wi h he aid o he
sca olding p o ein AXIN1, by o ming V-
ATPase-Ragula o /LKB1-AMPK complex a
he lysosome (8). Finally, McGui e e al.
showed ha glucose s a a ion leads o an
inc eased V-ATPase assembly which is
p eceded by AMPK ac i a ion (20). These
indings s eng hen he hypo hesis ha V-
ATPase is in ol ed in me abolic egula ion o
he cell.
Zhang e al showed ha an in ac V-ATPase is
needed o build he complex o
LKB1/AXIN1/AMPK and V-ATPase and ha
inhibi ing he V-ATPase wi h concanamycin
mimicks glucose s a a ion and he e o e
ac i a es AMPK in HEK293 cells (8). I was
also shown ha ano he V-ATPase inhibi o ,
ba ilomycin, can ac i a e AMPK in myoblas s
(21). Consis en wi h hei indings we could
show ha inhibi ing he V-ATPase unc ion
pha macologically wi h a chazolid lead o an
ac i a ion o AMPK in HEK293 and o he
non- umo cells, o an inc ease in he binding
o V-ATPase o AXIN1 and o an inc eased
lysosomal localiza ion o he AMPK-LKB1
complex. In e es ingly, in umo cells
a chazolid had no in luence on AMPK
phospho yla ion and binding o V-ATPase o
AXIN1 was a he dec eased a e a chazolid
ea men , hough all umo cells al eady
showed a phospho yla ed AMPK in con ols.
This inding is consis en wi h o he g oups
showing high cons i u i e AMPK ac i i y in
umo cells like p os a e cance me as asis o
glioblas oma umo s (22,23).
Di e en ial ac i a ion was no due o a s ess
esis ance o o he esis ance mechanisms in
umo cells as glucose dep i a ion lead o an
inc eased ac i a ion o AMPK in all cell lines.
The e o e his e ec is speci ic o V-ATPase
inhibi ion and sugges s a di e en ial
egula ion o V-ATPase inhibi ion induced
s ess esponses in umo and non- umo cells.
In e es ingly, we did no ind an inc ease o
AMP, nei he in umo no non- umo cells.
This e ec migh suppo he special ole o
V-ATPase induced AMPK ac i a ion.
Rega ding his dis inc e ec we wan ed o
in es iga e i he ac i a ion o AMPK in non-
umo cells plays a ole o a chazolid-induced
cy o oxici y.
Looking a he ole o AMPK in umo con ex
makes clea ha AMPK egula ion is highly
complex and depends on he cell ype, he
se e i y and he kind o he s ess signal and i s
ole as umo supp esso o oncogene is
discussed con o e sially (24).
In he umo con ex AMPK has long been
seen as an an i- umo a ge which o igina ed
om he disco e y o he AMPK ups eam
kinase LKB1, he loss o which esul s in he
de elopmen o Peu z-Je ghe synd ome ha
comes along wi h an inc eased cance isk
(25). Fu he mo e, he use o - o en unspeci ic
- AMPK-ac i a o s, we e shown o inhibi
umo g ow h in a a ie y o s udies (26-28).
Ye , in ecen yea s i became ob ious ha
AMPK ac i a ion can e y well be
umo igenic and can e en be essen ial o
umo p og ession as shown o glioblas oma,
p os a e cance (23,29,30) o myc
o e exp essing umo s (31). In ha ega d,
mos da a showing a umo -supp essi e ole
o AMPK ely on pha macological s udies,
whe eas expe imen s using gene ic ools
mos ly suppo a p o-su i al ole o AMPK.
Se e al oncogenes like S c o Myc we e
shown o ac i a e AMPK (31,32) and silencing
o an AMPK subuni impai s he abili y o
cells o o m umo s in i o (33). In addi ion,
loss o he umo supp esso ollicullin also
ac i a es AMPK (34). All hese da a sugges a
con ex speci ic ole o AMPK ac i i y.
In iguingly, we ound ha silencing
AMPK1α in HEK293 cells ea ed wi h
a chazolid lead o a dec eased g ow h a e
compa ed o wild ype cells sugges ing a p o-
su i al unc ion o AMPK ac i a ion.
Mo eo e , inc easing AMPK ac i i y in
MDA-MB-231 b eas cance cells by he
AMPK ac i a o AICAR p o ec ed he cells
om a chazolid-induced cy o oxici y – an
e ec ha could be diminished by silencing
AMPK1α. Hence ou indings suppo a p o-
su i al ole o AMPK ac i a ion a e V-
ATPase inhibi ion. Impo an ly, ea ing non-
umo cells wi h a combina ion o a chazolid
and AICAR had no e ec on apop osis
induc ion (Fig. S8).
A e showing ha induc ion o AMPK by
a chazolid has a p osu i al e ec we wan ed
o elucida e he mechanism behind i . AMPK
leads o a a ie y o me abolic changes
acili a ing su i al in pe iods o poo
nu i ion o o he me abolic s esses. The bes
known p o-su i al unc ions o AMPK
ac i a ion a e au ophagy, mac omolecule
impo , edox homeos asis and enhanced
glycolysis (35-37). In ega d o glucose
me abolism, we ound an inc eased glucose
up ake and sligh ly highe GLUT1 exp ession
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om

Tumo -speci ic e ec o V-ATPase inhibi ion
6
selec i ely in non- umo cells and an
un a ying o e en dec eased glucose up ake in
umo cells which co ela es wi h he
di e en ial ac i a ion o AMPK upon V-
ATPase inhibi ion. This is in cohe ence wi h
he da a o o he s, showing an AMPK
dependen inc eased up ake o glucose and
glucose ecep o exp ession o p omo e b eas
cance g ow h (38) and p o ec ion om
oxida i e s ess (39). Conce ning edox
homeos asis, Jeon e al showed ha AMPK
ac i a ion p o ec s om ROS p oduc ion and
main ains NADPH le els cons an du ing
glucose dep i a ion leading o cell su i al (7)
– he e o e p o iding e idence ha AMPK
ac i a ion p omo es cell su i al du ing
ene gy s ess. Addi ionally AMPK was shown
o be esponsible o he an i-oxidan e ec s o
es e a ol (40) and essen ial o he edox
balance in leukemia cells whe e AMPK
inhibi ion esul s in educed leukemic
p og ession (41). Inc eased ROS le els ha e
been epo ed o be a ea u e o V-ATPase
inhibi ion in umo cells (42,43) which we
could con i m in his s udy o MDA-MB-231
and MCF-7 cells using a chazolid as V-
ATPase inhibi o . ROS p oduc ion in non-
umo cells on he o he hand was no a ec ed.
These di e ences came along wi h a dec eased
NADPH/NADP le el and syn hesis o wo
majo enzymes o lipid syn hesis selec i ely in
umo cells. Impo an ly he a chazolid
induced dis u bance in edox homeos asis was
o e come by AMPK ac i a ion in non- umo
cells, while umo cells ailed o ac i a e
AMPK upon V-ATPase inhibi ion. Simila
e ec s we e also epo ed o Myc-posi i e
melanoma whe e AMPK ac i a ion could
p o ec he cells om ROS induced apop osis
(44). We ound ha sca enging ROS by Ti on
dec eased a chazolid induced apop osis in
umo cells and combina ion wi h AICAR
dec eased ROS p oduc ion, whe eas AMPK
silencing esul ed in inc eased ROS
p oduc ion a e V-ATPase inhibi ion e en in
non- umo cells. This poin s o he impo ance
o he edox homeos asis main ained by
AMPK unde me abolic s ess which is
impo an o alle ia ing cy o oxic e ec s o
V-ATPase inhibi ion.
We p opose ha V-ATPase inhibi ion by
a chazolid leads o a s ess esponse in non-
umo cells in which AMPK sus ains cellula
homeos asis like ene gy supply and edox
homeos asis. In umo cells howe e , V-
ATPase does no lead o an ac i a ion o
AMPK-media ed me abolic changes
supposedly due o a de ec i e egula ion o
AMPK which makes he cells mo e ulne able
o V-ATPase induced me abolic s ess (Fig. 6).
These esul s demons a e he signi icance o
he V-ATPase o umo me abolism, gi e a
i s insigh in o he mechanisms o umo -
speci ic e ec s o V-ATPase inhibi ion and
unde line he impo ance o unde s and
mechanis ic di e ences o umo and non-
umo cells.
Expe imen al p ocedu e
Cell Cul u e and Compounds
MDA-MD-231, HEK293, T24 and
MCF-7 cells we e ob ained om DSMZ
(Heidelbe g, Ge many) and MDA-MB-231
and HEK293 cells we e cul u ed in DMEM
medium supplemen ed wi h 10% FCS. MCF-
7 cells we e main ained in RPMI1640 medium
supplemen ed wi h 10% FCS, Insulin and 1%
non-essen ial amino acids. MCF10A cells
we e pu chased om ATCC and cul u ed in
DMEM-F12 supplemen ed wi h 5% ho se
se um, 100 mg/ml epide mal g ow h ac o , 10
mg/ml insulin, 1 mg/ml hyd oco isone, 1
mg/ml chole a oxin and 1%
penicillin/s ep a idin. HMLE cell we e a kind
gi o D . Ch is ina Scheel (Helmhol z Cen e
Munich) and cul u ed in MEGCM medium
ob ained om P omoCell (Heidelbe g,
Ge many).
S a a ion expe imen s we e pe o med in
glucose- ee DMEM supplemen ed wi h 10%
o dialyzed FCS.
A chazolid was p o ided by P o . Di k
Menche, Uni e si y o Bonn, AICAR was
pu chased om Toc is bioscience (B is ol,
UK).
Pa ien -de i ed b eas umo samples and
heal hy b eas issue samples we e a kind gi
om he non-p o i o ganiza ion PATH
biobank (Augsbu g, Ge many) (45). The
collec ion o samples was app o ed by he
local e hics commi ee in Bonn and all pa ien s
p o ided w i en in o med consen . The
sample collec ion was ca ied ou acco ding o
he decla a ion o Helsinki.
Immunoblo ing
Cells we e lysed in a bu e con aining
EDTA 2 mM, NaCl 137 mM, Glyce ol 10 %,
Na4P2O7 2 mM, T is-Base 20 mM, T i on X-
100 1 % and Na-Glyce olphospha e 20 mM a
pH 7,5. Fo immunop ecipi a ion cells we e
lysed and incuba ed wi h an an ibody agains
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
Tumo -speci ic e ec o V-ATPase inhibi ion
7
AXIN1 o e nigh . Subsequen ly he lysa e
was incuba ed wi h P o ein Aga ose A beads
(Sigma Ald ich, Tau ki chen, Ge many),
washed wi h lysa e bu e and be a-
me cap oe hanol con aining sample bu e was
added. Equal amoun s o p o ein we e
sepa a ed on a SDS-PAGE and ans e ed o
a ni ocellulose memb ane. The memb anes
we e blocked wi h 5% milkpowde and hen
p obed wi h p ima y an ibodies.
Fo loading con ol 2,2,2-T ichlo oe hanol
(TCE) was added o he polyac ylamide gel
be o e polyme iza ion. The yp ophans o
p o eins unde go an ul a iole ligh -induced
eac ion wi h TCE which p oduces ligh in he
isible ange so ha p o eins can be de ec ed
(46). In some cases ac in was used as loading
con ol.
Following an ibodies we e used: pAMPK
Th 172, AMPK, AMPK1alpha, pACC Se 79,
ACC, PARP1 (Cell Signaling Technology,
Dan e s, MS), GLUT1 (No us Biologicals),
ac in, ab7, AXIN1, V-ATPase D, (San a C uz
Bio echnology, Dallas, TX), HRP-goa -an i-
abbi (Bio-Rad, Munich, Ge many) and HRP-
goa -an i-mouse (San a C uz Bio echnology).
Isola ion o Lysosomes
Lysosomes we e isola ed as
p e iously desc ibed (47). B ie ly, 1,5 x 106
cells we e seeded in 10 cm dishes and we e
ea ed as indica ed. Subsequen ly medium
was emo ed and 250 µl homogeniza ion
bu e was added o he cells. The cell
suspension was ans e ed o a glas g inding
essel and homogenized using a po e
homogenize . Nex , he homogena e was
cen i uged 15 min a 14.000 g a 4°C. The
supe na an was ans e ed o a 2 ml
polyca bona e cen i uga ion ube and an equal
olume o 16 mM CaCl2 o p ecipi a e he
lysosomes was added. Then he homogena e
was cen i uged a 25.000 g o 15 min. Nex ,
supe na an is disca ded and he pelle
esuspended in one olume o washing bu e
and again cen i uged. Now he pelle
con aining lysosomes is esuspended in 20 µl
o Washing bu e and used o Wes e n blo
analysis.
T ans ec ion
Cells we e ans ec ed wi h
GenaxxoFec (Genaxxon Bioscience,
Müns e , Ge many) acco ding o he
manu ac u e ’s ins uc ions. B ie ly,
AMPKα1 was silenced using ON-TARGET-
Plus SMARTpool siRNA om Dha macon
(Schwe e, Ge many) and non- a ge ing
siRNA as a con ol and we e ea ed as
indica ed 48 h a e ans ec ion.
Immunocy ochemis y
Cells we e seeded on µ-slides 8-well
ibidiT ea (IBIDI, Ma ins ied, Ge many) and
we e ea ed as indica ed. Cells we e ixed and
s ained as desc ibed be o e (48) and analyzed
wi h a Zeiss LSM 510 Me a con ocal
mic oscope (Jena, Ge many). Following
an ibodies we e used: LKB1 (Cell Signaling
Technology, Dance s, MS), Lamp1
(De elopmen al S udies Hyb idoma Bank
Iowa, IA) AlexaFluo 488-goa -an i- abbi ,
AlexaFluo 543-goa -an i-mouse (In i ogen,
Wal ham, MA).
Apop osis Assay
Subdiploid DNA con en was
de e mined acco ding o Nicole i e al (49).
B ie ly, cells we e ea ed as indica ed,
ha es ed, pe meabilized wi h Na iumci a e
con aining T i on X-100, s ained wi h 25
µg/ml p opidiumiodide and analyzed by low
cy ome y (Beck on Dickinson, Heidelbe g,
Ge many). Subdiploid cells le o he G1-peak
we e conside ed as apop o ic.
Cy o oxici y assay
Cells we e seeded in a 96 well pla e,
ea ed as indica ed and g ow h inhibi ion was
analyzed wi h he CellTi e Blue assay
(P omega, Madison, WI) acco ding o
manu ac u e ’s ins uc ions.
qRT-PCR analysis
To al mRNA was isola ed om cell
cul u e samples using Qiagen RNeasy Mini
Ki (Qiagen, Hilden, Ge many) acco ding o
he manu ac u e ’s ins uc ions. Fo cDNA
syn hesis, he High Capaci y cDNA Re e se
T ansc ip ion Ki (Applied Biosys ems, Fos e
Ci y, CA) was used. qRT-PCR was pe o med
wi h he AB 7300 RealTime PCR sys em, he
TaqMan Gene Exp ession Mas e Mix
(Applied Biosys ems) and he SYBR G een
PCR Mas e Mix (Applied Biosys ems)
acco ding o he manu ac u e 's ins uc ions.
All designed p ime s we e pu chased om
Me abion (Ma ins ied, Ge many).
Glucose up ake
Cells we e ea ed as indica ed.
Subsequen ly cells we e incuba ed wi h 100
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
Tumo -speci ic e ec o V-ATPase inhibi ion
8
µM 2-NBDG (2-[N-(7-Ni obenz-2-Oxa-1,3-
Diazol-4-yl)Amino]-2-Deoxy-D-Glucose)
(In i ogen, Wal ham, MA) in HANKS bu e
o 30 min a 37°C. Then cells we e ha es ed
and changes in glucose up ake we e measu ed
by low cy ome y.
ROS Measu emen
Cells we e ea ed as indica ed,
ha es ed and incuba ed wi h 10 µM o 2′,7′-
dichlo o luo escein diace a e (Sigma Ald ich,
Tau ki chen, Ge many) o 30 min a 37°C.
A e a washing s ep wi h PBS, eac i e
oxygen species p oduc ion was analyzed by
low cy ome y.
ATP
Cells we e seeded in a 96-well pla e
and ea ed as indica ed. ATP le el was
de e mined using he CellTi e -Glo assay
(P omega) acco ding o he manu ac u e ’s
ins uc ion and measu ed in a luminome e
(Be hold Technologies, Bad Wildbad,
Ge many) acco ding o manu ac u e ’s
ins uc ions.
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
Tumo -speci ic e ec o V-ATPase inhibi ion
9
Acknowledgemen : We would like o hank Be nade e G ohs and Ri a Soche o hei excellen wo k
and D . Simone Mose o doing HPLC analysis. This wo k was suppo ed by he DFG g an FOR 1406
SCHW 1781/1-1.
Con lic o in e es : The au ho s decla e ha hey ha e no con lic s o in e es wi h he con en s o
his a icle
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om

a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
Ka in Ba el, Rol Mülle and Ka in on Schwa zenbe g
explains why V-ATPase inhibi ion selec i ely kills cance cells
Di e en ial egula ion o AMP-ac i a ed p o ein kinase in heal hy and cance cells
published online Oc obe 11, 2019J. Biol. Chem.
10.1074/jbc.RA119.010243Access he mos upda ed e sion o his a icle a doi:
Ale s:
When a co ec ion o his a icle is pos ed• When his a icle is ci ed•
o choose om all o JBC's e-mail ale sClick he e
a Helmhol z Zen um ue In ek ions o schung - BIBLIOTHEK- on No embe 14, 2019h p://www.jbc.o g/Downloaded om
Supplemen al Ma e ial and Me hods
Cell lines, an ibodies
Panc3.027 panc ea ic cance cells we e ob ained om ATCC and cul i a ed in RPMI1640
media supplemen ed wi h 10% FCS and 10 mg/ml insulin. HepaRGTM cells we e ob ained om The mo
Fishe Scien i ic. Cells we e pla ed and main ained in Williams’ medium E supplemen ed wi h
Glu aMAXTM and HepaRG™ Thaw, Pla e, & Gene al Pu pose Medium Supplemen (Thaw, Pla e, &
Gene al Pu pose Wo king Medium) pu chased om The mo Fishe Scien i ic. Hepa ocellula
ca cinoma cell line HepG2 was pu chased om Ge man Resea ch Cen e o Biological Ma e ial
(DSMZ) and cul u ed in DMEM, 10% FCS.
pACC Se 79 and ACC an ibodies we e om Cell Signaling Technology, Dan e s, MA.
Lyso acke s aining
Se e al umo and non- umo cells we e seeded in IBIDI slides and ea ed wi h a chazolid o
4h. Subsequen ly a pH sensi i e Lyso acke (In i ogen, Wal ham, MA) was added o he wells and
cells we e measu ed a e an incuba ion ime o 30 min a he con ocal mic oscope. Acidic lysosomes
show a ed s aining whe eas alkaline lysosomes show no s aining. Pic u es ep esen one pic u e ou o
h ee expe imen s.
De e mina ion o AMP
2 x 106 cells we e seeded in 10 cm dishes and ea ed wi h a chazolid o 20 hou s. Subsequen ly
hey we e sc aped o and washed wice wi h PBS. The pelle was esuspended in 6 % pe chlo ic acid
and incuba ed o 20 min. A e ho oughly o exing he lysa e was cen i uged o 15 min a 14 000
pm. The supe na an was neu alized wi h 2M KOH and again cen i uged. The supe na an was used
o HPLC measu emen . AMP was quan i ied using e e se-phased HPLC (Agilen 1260 In ini y II LC
sys em wi h a 1260 In ini y Degasse , a 1100 Se ies qua e na y pump and 1100 Se ies diode a ay
de ec o ) and sepa a ed on a C18 column (20EC-C18 4 µm 46 x 150 mm a 15°C p o ec ed by
Phenomenex ODS 4 x 3 mm i.d. p e-column) o 10 min a a low- a e o 0,8 ml/min. The mobile-phase
was i luo ace ic acid.

1.5
1
0.5
0
co
a ch 10nM
co a ch
pAMPK/AMPK ( old induc ion)
MDA-MB-231
MCF-7
T24
pAMPK/AMPK ( old induc ion)
co a ch
HEK293
co
a ch 10nM
MCF10A
HMLE
4
3
2
1
0
*
**
pAMPK/AMPK ( old induc ion)
4
3
2
1
0
MDA-MB-231
MCF7
HEK293
co -gluc
*
*
*
2
1.5
1
0.5
0N T
pAMPK/AMPK ( old induc ion)
*
B
C D
A
Figu e S1: Quan i ica ion o AMPK Wes e nblo s. The umo cell lines MCF-7, MDA-MB-231 (MDA)
and T24 (A) and he non- umo cell lines HEK293, MCF10A and HMLE (B) we e ea ed wi h 10 nM
a chazolid (a ch) o 24 hou s and phospho yla ion o AMPK a Th 172 and o al AMPK p o ein we e
analyzed by Wes e n Blo . (C) HEK293, MCF-7 and MDA-MB-231 we e incuba ed in glucose- ee media
o 24 hou s and phospho yla ion o AMPK was analyzed. (D) Phospo yla ion o AMPK was analyzed in
pa ien -de i ed b eas cance issue and no mal b eas issue om he same pa ien . Quan i ica ion ep esen s
h ee di e en expe imen s. *p<0.05, s uden s - es .
HEK293 MDA
pACC
(Se 79)
ACC
LC
- + - + a ch
300 kDa -
200 kDa -
300 kDa -
200 kDa -
2
1.5
1
0.5
0
co a ch
2
1.5
1
0.5
0
pACC old induc ion
pACC/ACC
pACC/LC
pACC/ACC
pACC/LC
co a ch
pACC old induc ion
Figu e S2: A chazolid leads o phospho yla ion o ACC. HEK293 non- umo and MDA-MB-231 umo
cell lines we e ea ed wi h 10 nM a chazolid o 24 h. Phospho yla ion o ACC a Se 79 was analyzed by
wes e n blo . Quan i ica ion was done o pACC e sus ACC and pACC e sus loading con ol (LC) which
showed simila esul s.
HEK293 MDA
70 kDa -
50 kDa -
HEK293
HMLE
HUH7
MCF7
MDA-MB-231
T24
Co A ch 10nM
Hoechs
Lyso acke
Figu e S3: A chazolid leads o alkaliza ion o lysosomes. Di e en umo and non- umo cell lines we e
ea ed wi h 10 nM a chazolid o 24 h. Subsequen ly cells we e s ained wi h a pH sensi i e Lyso acke and
analyzed by con ocal mic oscopy. The scale ba is 25 µm, apa om MCF-7 (10 µm).
Lamp1
AXIN1 me ge
co
co
a ch
a ch
HEK293
MDA
Figu e S4: Colocaliza ion o AXIN1 and Lamp1 is shown a e a chazolid ea men in HEK293 cells
and cons i u i ely exis en in MDA-MB-231 cells. Cololocaliza ion o Lamp1 and AXIN1 was de e mined
in HEK293 and MDA-MB231 cells a e ea men wi h 10 nM a chazolid o 24 h by con ocal s aining. One
ep esen a i e image ou o h ee independen expe imen s is shown.