Signal Transduction

Genotoxic Damage Activates the AMPK-a1Isoform in theNucleus viaCa2þ/CaMKK2Signalingto Enhance Tumor Cell SurvivalDiana Vara-Ciruelos1, Madhumita Dandapani1, Alexander Gray1, Ejaife O. Egbani2,A. Mark Evans2, and D. Grahame Hardie1

Abstract

Many genotoxic cancer treatments activate AMP-activatedprotein kinase (AMPK), but the mechanisms of AMPK activa-tion in response to DNA damage, and its downstream con-sequences, have been unclear. In this study, etoposide activatesthe a1 but not the a2 isoform of AMPK, primarily within thenucleus. AMPK activation is independent of ataxia-telangiec-tasia mutated (ATM), a DNA damage-activated kinase, and theprincipal upstream kinase for AMPK, LKB1, but correlates withincreased nuclear Ca2þ and requires the Ca2þ/calmodulin-dependent kinase, CaMKK2. Intriguingly, Ca2þ-dependent acti-vation of AMPK in two different LKB1-null cancer cell linescaused G1-phase cell-cycle arrest, and enhanced cell viability/survival after etoposide treatment, with both effects beingabolished by knockout of AMPK-a1 and a2. The CDK4/6

inhibitor palbociclib also caused G1 arrest in G361 but notHeLa cells and, consistent with this, enhanced cell survival afteretoposide treatment only in G361 cells. These results suggestthat AMPK activation protects cells against etoposide by lim-iting entry into S-phase, where cells would be more vulnerableto genotoxic stress.

Implications: These results reveal that the a1 isoform of AMPKpromotes tumorigenesis by protecting cells against genotoxicstress, which may explain findings that the gene encodingAMPK-a1 (but not -a2) is amplified in some human cancers.Furthermore, a1-selective inhibitors might enhance the anti-cancer effects of genotoxic-based therapies. Mol Cancer Res; 16(2);345–57. �2017 AACR.

IntroductionThe wild mandrake or May Apple (Podophyllum peltatum),

whichwasused in traditionalNativeAmericanmedicine, containsthe natural product podophyllotoxin (1). Podophyllotoxin itselfproved to be too toxic for human use, but the synthetic derivativeetoposide (also known as VP-16) was approved for cancer treat-ment in 1983. Etoposide acts by binding to topoisomerase II (2),an enzyme that can relax DNA supercoiling, insert or removeknots, and catenate or decatenate DNA. It catalyzes an ATP-dependent cycle in which both strands of one DNA helix arebroken, followed by passage through this break of a second helixand religation of the breaks in the first. By inhibiting the religationstep, etoposide and other anticancer agents, such as doxorubicin,create double-strand breaks. They can also cause trapping ofcomplexes in which tyrosine residues of the topoisomerase II

homodimer remain covalently attached to phosphate groups ofthe nucleotide at the 50 end of a DNA break, interfering withsubsequent DNA replication and transcription (3). As topoisom-erase II function is particularly crucial during S-phase when DNAis being replicated, rapidly proliferating cells (including tumorcells) aremore susceptible to cell death induced by etoposide thanquiescent cells.

TheAMP-activated protein kinase (AMPK) is a sensor of cellularenergy status expressed in essentially all eukaryotic cells, whichoccurs universally as heterotrimeric complexes comprising cata-lytic a subunits and regulatory b and g subunits (4–6). AMPK isactivated by phosphorylation of a conserved threonine residuewithin the activation loop of the kinase domain on the a subunit,usually referred to as Thr172. Thr172 is phosphorylated in vivo bya complex containing the tumor suppressor kinase LKB1, or byCa2þ/calmodulin-dependent protein kinase kinases (CaMKK),especially CaMKK2 (4–6). The LKB1 complex appears to beconstitutively active, but displacement of ATP by AMP at two ormore sites on the regulatory g subunit of AMPK leads to increasednet Thr172 phosphorylation. This occurs because AMP bindingtriggers conformational changes that promote Thr172 phosphor-ylation by LKB1 and inhibit Thr172 dephosphorylation by pro-tein phosphatases, with these effects being mimicked by ADP(7, 8); binding of AMP (but not ADP) also causes further allostericactivation (9). Because increases in the cellular ADP:ATP ratio(signifying energy deficit) are always accompanied by even largerincreases in the AMP:ATP ratio, these three effects allow AMPK toact as an ultrasensitive sensor of cellular energy status. TheCaMKK2–AMPK pathway appears to be activated solely byincreases in intracellular Ca2þ, which can be caused by addition

1Division of Cell Signalling & Immunology, College of Life Sciences, University ofDundee, Dow Street, Dundee, Scotland, United Kingdom. 2Centre for IntegrativePhysiology, University of Edinburgh, Hugh Robson Building, George Square,Edinburgh, Scotland, United Kingdom.

Note: Supplementary data for this article are available at Molecular CancerResearch Online (http://mcr.aacrjournals.org/).

D. Vara-Ciruelos and M. Dandapani are co-first authors of this article.

Corresponding Author: D. Grahame Hardie, University of Dundee, School ofLife Sciences, Dundee DD1 5EH, Scotland, UK. Phone: 44-1382-384253; Fax:44-1382-385507; E-mail: d.g.hardie@dundee.ac.uk

doi: 10.1158/1541-7786.MCR-17-0323

�2017 American Association for Cancer Research.

MolecularCancerResearch

www.aacrjournals.org 345

to cells of Ca2þ ionophores, such as A23187, or by hormones orcytokines that increase intracellular Ca2þ (4–6).

Each AMPK subunit exists as multiple isoforms (a1/a2, b1/b2,g1/g2/g3) encoded by multiple genes (PRKAA1/2, PRKAB1/2,PRKAG1/2/3). Intriguingly, although the PRKAA2 gene (encodinga2) is quite often mutated in human cancers, consistent with theidea that it helps to exert the tumor suppressor role of LKB1, thePRKAA1 and PRKAB2 genes (encoding a1 and b2) are frequentlyamplified instead, suggesting that their amplification may beselected for because they promote tumor formation (5, 10). Thereason for the distinct behaviors of these AMPK-encoding genes inhuman cancers has been unclear.

Fu and colleagues (11) reported that etoposide activatedAMPK, and claimed that the effect was dependent on the proteinkinase ataxia-telangiectasia mutated (ATM), because it appearedto be absent in ATM-deficient cells. ATM is a member of thephosphatidylinositol 3-kinase–like kinase (PIKK) family,with thegenetic disorder ataxia-telangiectasia being caused by loss-of-function mutations in the ATM gene. ATM is activated in normalcells by double-strand breaks in DNA, such as those induced byetoposide. Once activated, ATM autophosphorylates on Ser1981and phosphorylates downstream targets such as the histonevariant, H2AX, and Structural Maintenance of ChromosomesProtein-1 (SMC1); these events can be used as biomarkers bothfor ATM activation and for DNA damage (12). ATM also phos-phorylates LKB1 at Thr366 (13), and it was reported that activa-tion of AMPK by etoposide in prostate cancer cells was reduced byshRNA-mediated knockdown of either ATM or LKB1 (14), sug-gesting the existence of a kinase cascade from ATM to LKB1 toAMPK. Arguing against this, however, etoposide still activatedAMPK in the LKB1-null HeLa cell line (11), while ionizingradiation (which causes double-stranded DNA breaks and alsoactivates ATM) activates AMPK in another LKB1-null tumor line,A549 cells (15).

To address these discrepancies, we investigated the mecha-nism by which etoposide activates AMPK. We show that acti-vation of AMPK is restricted to the a1 isoform within thenucleus and does not require ATM or LKB1 but is dependenton CaMKK2 and triggered by increases in nuclear Ca2þ. We alsoshow that prior activation of either isoform enhances cellsurvival during etoposide treatment. These results not onlyprovide a potential explanation for the findings that the geneencoding AMPK-a1 (but not a2) is amplified in some humancancers, but also suggest that pharmacologic inhibitors ofAMPK might make tumor cells more sensitive to death inducedby DNA-damaging treatments, and might therefore be usefuladjuncts to chemotherapy or radiotherapy. If these inhibitorswere a1 selective, this might avoid potential systemic sideeffects caused by inhibition of a2 complexes.

Materials and MethodsMaterials

A23187, nocodazole, propidium iodide, etoposide, palboci-clib, STO609, and KU-5593were from Sigma-Aldrich. Antibodiesagainst actin (A5441) were from Sigma-Aldrich, against p-AMPK(pT172; A2535), p-ATM (pS1981; 4526), and total ATM (2631)were fromCell Signaling Technology, and against phospho-SMC1(S966-P; A300-050A) and total SMC1 (A300-055A) were fromBethyl Laboratories. Antibodies against AMPK-a1 and a2, thephosphorylated form of acetyl-CoA carboxylase (pACC), and

total ACC have been described previously (16, 17). Secondaryantibodies were from Li-Cor Biosciences.

Cell cultureAll cell lines were from the European Collection of Authenti-

cated Cell Cultures (ECACC), except immortalized mouseembryo fibroblasts (MEF), which were a gift from Dr. BenoitViollet, INSERM (Paris, France; ref. 18). HeLa andG361 cells wererevalidated during the project by STR profiling (Public HealthEngland, certificate dated August 14, 2015); HEK-293 cells werepurchased fromECACC in January 2016.G361 cells were culturedin McCoy's 5A medium containing 1% (v/v) glutamine, 10%(v/v) FBS, and 1% (v/v) penicillin/streptomycin. HeLa and HEK-293 cells and MEFs were cultured in DMEM with 10% (v/v) FBSand 1% (v/v) penicillin/streptomycin.

AMPK assaysEndogenous AMPK isoforms were assayed by immunoprecip-

itate kinase assays as described previously (19, 20) using theAMARA peptide (21) as substrate. AMPK-a1 or a2 were assayedby immunoprecipitation with a1- or a2-specific antibodies,whereas total AMPK was assayed by immunoprecipitation withan equal mixture of a1- and a2-specific antibodies.

Construction of AMPK knockout cellsKnockout of AMPK-a1 anda2 (PRKAA1 and PRKAA2) inG361

andHeLa cells was carried out using the CRISPR-Cas9method, asdescribed previously for G361 cells (22).

Immunofluorescence microscopyCells were grown on glass coverslips in 6-well dishes until 40%

to 80% confluent and then treated with drugs or vehicle control.The coverslips were washed 3 times in PBS, fixed in 4% parafor-maldehyde in PBS for 20minutes, and washed again 3 times withPBS. The cells were then permeabilized in 0.2% Triton X-100 inPBS for 5 minutes and washed again 3 times with PBS. They werethen placed in blocking buffer (PBS with 0.2% fish skin gelatin)for 1 hour and stained with primary antibody (1:300–1:1,000 inblocking buffer) for 30minutes. After further washing in blockingbuffer, the coverslips were incubated with fluorescent secondaryantibody (1:200 dilution in blocking buffer) for 30minutes. Theywere then mounted on slides using Vectashield and sealed withnail varnish. The slides were imaged on the DeltaVision decon-volutionmicroscope and images acquired and deconvolved usingthe SoftWorx program.

Live cell imaging with Fluo4Cells were grown on glass bottom dishes (WillCo Wells)

until 40% to 80% confluent. They were then loaded with Fluo4-AM dye (1 mmol/L) plus Hoechst 33342 for 20 to 30 minutes.The cells were then washed with PBS, and fresh medium wasadded. The WillCo dish was mounted onto a stage preheated to37�C and enclosed with a lid to deliver 5% CO2 to the cells.Calcium flux was measured using the DeltaVision OMX systemwith fluorescence measured at 488 nm using a 63� objective.Image analysis and quantification was performed using ImageJsoftware.

siRNA knockdown of CaMKK2Reverse transfection of siRNA constructs against CaMKK2 was

carried out according to protocols detailed in the Ambion siRNA

Vara-Ciruelos et al.

Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research346

Starter Kit. Briefly, transfection reagent (siPORT) was diluted 1 in20 inOpti-MEMmediumand allowed to incubate for 10minutes.Meanwhile, cells were trypsinized for 5 minutes and the trypsinquenched with complete media. siRNA (Ambion, Life Technol-ogies) was diluted in Opti-MEM to a final concentration of 5nmol/L. siRNA and diluted transfection reagent were then mixedand incubated for 10 minutes. The siRNA- siPORT mixture wasadded to6-well plates, and cellswere seeded into the6-well plates.The plates were returned to the incubator for 24 hours, after whichtreatment with vehicle or drugs was carried out. Knockdown wasconfirmed by Western blotting.

Clonogenic and MTT survival assaysFor clonogenic survival assays, cells were seeded into 6-well

plates at equal density and treated at 40% to 80% confluence withvehicle control or different concentrations of drug for 24 hours.Cells were trypsinized in 0.5 mL of trypsin:EDTA for 5 minutesand then diluted with 1 mL of complete medium. The cells in thecontrol flask were counted using a hemocytometer under amicroscope, and 2,000 cells were seeded in triplicate into 10-cmdishes containing 10 mL of medium. The same volume of cellsuspension was aspirated from the treated flasks and seeded intriplicate into 10-cm dishes also containing 10 mL of medium.The dishes were placed at 37�C in an incubator for 10 to 15 days.On the last day, the medium was aspirated; the cells were fixedwith ice-coldmethanol for 10minutes and stainedwith 0.3%w/vmethylene blue in methanol for 10 minutes. The dishes werewashed with deionized water, and the number of colonies wascounted manually.

ForMTT assays, cells were seeded in 12-well plates until 40% to80% confluence and treated for 48 hours. The cells were thenincubated with 0.5% of 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphe-nyl-2H-tetrazolium bromide (MTT) for 3 hours. The growthmedium was aspirated off, and formazan crystals were dissolvedin 500 mL of acidified isopropanol. The intensity wasmeasured byspectrophotometry at 570 nm.

Cell-cycle analysisCells were seeded in 6-cm dishes until 40% to 80% confluence

and treated for 16 to 18 hours. Cells were trypsinized and the cellswere pelleted, transferred into 5 mL Falcon tubes, and washedwith PBS containing 1% FCS and 0.1 mmol/L EDTA. The cellswere then fixed with ice-cold 70% ethanol for >2 hours. The cellswere washed twice in the PBS buffer and stained with propidiumiodide (diluted 1 in 20 in the PBS buffer) with RNase added todigest RNA. Cell-cycle analysis was performed on theCalibur flowcytometer and analysis performed using FlowJo software.

Additional analytic proceduresSDS-PAGE was performed using precast Bis-Tris 4% to 12%

gradient polyacrylamide gels in the MOPS buffer system (Invi-trogen). Proteins were transferred to nitrocellulose membranes(Bio-Rad) using the Xcell II Blot Module (Invitrogen). Mem-branes were blocked for 1 hour in TBS containing 5% (w/v)nonfat dried skimmed milk. The membranes were probed withappropriate antibody (0.1–1 mg/mL) in TBS-Tween and 2%(w/v) nonfat dried skimmed milk. Detection was performedusing secondary antibody (1 mg/mL) coupled to IR 680 or IR800 dye, and the membranes were scanned using the Li-CorOdyssey IR imager.

Presentation of data and statistical analysisSignificances of differences were estimated using GraphPad

Prism 6 for Mac OSX, using Student t test, one-way, or two-wayANOVA as appropriate. Unless stated otherwise, Sidak multiplecomparison test was used for post hoc analysis. Numbers ofreplicates (n) refer to biological replicates, that is, the number ofindependent cell cultures analyzed. Significances of differencesare indicated as follows: �,P<0.05; ��,P<0.01; ���,P<0.001; ����,P < 0.0001 or †, P < 0.05; ††, P < 0.01; †††, P < 0.001; ††††, P <0.0001; ns, not significant.

ResultsAMPK activation by etoposide does not require LKB1

We initially studied the effects of etoposide in cell lines expres-sing LKB1, including the human embryonic kidney cell line, HEK-293. Surprisingly, we could not detect significant AMPKactivationin these cells using kinase assays, although it could be readilyobserved in two LKB1-null tumor cell lines, that is, HeLa (cervicalcancer) and G361 (melanoma) cells. We could also observeAMPK activation in immortalized MEFs, which do express LKB1(Fig. 1A).

In HeLa cells, AMPK activation following addition of 30mmol/L etoposide was quite slow and relatively modest in extent,becoming significant by 6 to 9 hours and reaching a maximum of2- to 3-fold after 18 hours. Infusion of high-dose etoposide inhumans results in peak plasma concentrations ranging from 50 to200mmol/L (23), andAMPKactivation after 16-hour treatment ofHeLa cells occurred at etoposide concentrations from 30 to 300mmol/L (Fig. 1B). AMPK activation was associated with increasedThr172 phosphorylation, by 1.8 � 0.3-fold (not significant), 2.2� 0.3-fold (P < 0.05), and 2.3� 0.3-fold (P < 0.05) at 30, 100, and300 mmol/L etoposide, respectively (mean � SEM, n ¼ 6), andwith phosphorylation of the ATM substrate SMC1, althoughphosphorylation of the latter appeared to be saturated at 30mmol/L (Fig. 1B). As expected, activation and Thr172 phosphor-ylation of AMPK was also observed following treatment of HeLacells with the Ca2þ ionophore A23187 (which activates theupstream kinase CaMKK2), although the degree of activation waslarger than that obtained with etoposide. However, A23187treatment did not lead to phosphorylation of SMC1, suggestingthat ATM was not activated by increasing cellular Ca2þ (Fig. 1B).Very similar effects of etoposide and A23187 were observed inG361 cells (Fig. 1C), where the effects of etoposide on Thr172phosphorylation were 2.1� 0.1-fold (not significant), 3.5 � 0.4-fold (P<0.001), and3.7�0.5-fold (P<0.001) at 30, 100, and300mmol/L, respectively (mean � SEM, n ¼ 4).

AMPK activation by etoposide occurs primarily within thenucleus

Figure 2A shows fluorescencemicrographs of HeLa cells treatedwith vehicle (DMSO), etoposide (100 mmol/L), or A23187. Thecells were fixed and labeled with antibodies recognizing ATMphosphorylated at Ser1981 (pATM) or AMPK phosphorylated atThr172 (pT172). Quantification of the nuclear and cytoplasmicfluorescence is shown in Fig. 2B, left (n ¼ 10). As expected, largeincreases (13-fold) in nuclear staining were obtained with theanti-pATM antibody after treatment with etoposide, but notA23187. In contrast, large increases in nuclear staining wereobtained with anti-pT172 antibody after treatment with eitheretoposide (13-fold) or A23187 (9-fold). Although A23187 also

Etoposide-Induced AMPK Activation Enhances Cell Survival

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 347

appeared to increase cytoplasmic pT172 staining in HeLa cells,this was not statistically significant. Similar results were obtainedwith G361 cells, although in these cells, A23187 did cause a

significant increase in cytoplasmic pT172 staining (Supplemen-tary Fig. S1; quantification in Fig. 2C, right). These results showthat etoposide activates AMPK primarily in the nucleus, in bothHeLa and G361 cells.

Using the anti-pT172 antibody to detect AMPK activation inHeLa cells, fluorescence could be observed in most nuclei aftertreatment for 18 hours with 100 mmol/L etoposide, but was stillobserved in some nuclei even at much lower etoposide concen-trations (250 nmol/L, Supplementary Fig. S2, arrows); theseinvariably corresponded with nuclei that were also positive forpATM, indicating the occurrence of DNA damage.

Because basal Thr172 phosphorylation and AMPK activity ismuch higher in the cytoplasm of cells that express LKB1, such asHEK-293 cells (24), we hypothesized that this might have beenmasking changes in the small pool of AMPK in the nucleus,explaining why we had not been able to initially detect effectsof etoposide in HEK-293 cells. Consistent with this, in controlHEK-293 cells treatedwithDMSO for 18hours, therewas a diffusecytoplasmic and nuclear fluorescence detected using anti-pT172antibodies, such that it was difficult to distinguish the twocompartments. Most cells also displayed at least one perinuclearpatch with more intense fluorescence; these appear to representthe Golgi apparatus, as they stained with antibodies against theGolgi maker, GM130 (data not shown). In cells treated with lowetoposide concentrations (3 mmol/L), the nuclei were morefluorescent than the cytoplasm in some cells, once again corre-sponding with cells where the signal for nuclear pATM wasincreased (Supplementary Fig. S3, arrows). Figure 2C showsquantification of results from a large number of cells (n ¼ 18).The mean absolute fluorescence appeared to increase in responseto etoposide in the nucleus and decrease in the cytoplasm, whileremaining constantwhenaveraged across thewhole cell, althoughthese effects were not statistically significant (Fig. 2C, left). How-ever, when the results were expressed as nuclear:cytoplasmicratios, there was a significant increase in response to 3 mmol/Letoposide (Fig. 2C, right). Thus, etoposide appears to activateAMPKwithin the nuclei ofHEK-293 cells, although this is difficultto detect because it can bemasked by the high basal activity in thelarger pool of AMPK in the cytoplasm.

AMPK activation by etoposide is specific for the a1 isoformBecause the sequences around Thr172 are identical in a1 and

a2, the anti-pT172 antibody does not distinguish the two AMPKcatalytic subunit isoforms. However, human a1 (predicted mass64,009 Da) is slightly larger than a2 (62,319 Da), and the twoisoforms can be resolved by SDS-PAGE. In Fig. 3A, we analyzedlysates from triplicate dishes of control and etoposide-treatedHeLa cells by dual label Western blotting using anti-pT172 anti-bodies labeledwith IRDye 680 (red), and eithera1- ora2-specificantibodies labeled with IRDye800 (green). In the upper blot(probed with anti-pT172 and anti-a1), a single band appearspredominantly green in the control lanes but becomes moreyellow (indicating increased Thr172 phosphorylation) after eto-poside treatment. In the lower blot (probed with anti-pT172 andanti-a2), two bands are evident, an upper red band (a1) thatbecomes more intense after etoposide treatment, and a lowergreen band (a2), whose intensity or color does not change. Theintensity in the pT172 (red) channel fora1 increased by 2.1� 0.1-fold (P < 0.0001, n¼ 3) and 2.0� 0.1-fold (P¼ 0.0001, n¼ 3) inthe top and bottom panels respectively, but was unchanged fora2. These results show that Thr172 phosphorylation in response

****

3

A

B

C

2

1AM

PK

Act

ivity

(rel

ativ

e to

con

trol)

DMSO ETOP

HeLa cells G361 cells MEFs

DMSO ETOPDMSO ETOP

****

***

***

Effect of etoposide on AMPK activity in various cell types

3

2

1

AM

PK

Act

ivity

(rel

ativ

e to

con

trol) 4

*********

Effect of etoposide/A23187 (HeLa cells)

Etoposide (µmol/L):A23187 (µmol/L):

0 0300100300 10000

pSMC1

SMC1

AMPK-α

pT172

1

2

3

4

**

****

******

Effect of etoposide/A23187 (G361 cells)

pSMC1

SMC1

AMPK-α

pT172

Etoposide (µmol/L):A23187 (µmol/L):

0 0300100300 10000

AM

PK

Act

ivity

(rel

ativ

e to

con

trol)

Figure 1.

Activation of AMPK by etoposide in different cell types. A,HeLa cells, G361 cells,and immortalized MEFs were incubated with 30 mmol/L etoposide for 18 hours(HeLa, G361) or with 20 mmol/L etoposide for 2 hours (MEFs). Cell lysates wereimmunoprecipitated with anti–AMPK-a1 and a2 antibodies for AMPK assay.Results are expressed relative to the mean activity (�SEM) in incubationswithout etoposide; results significantly different from the DMSO control by t testare indicated (HeLa and G361 cells, n ¼ 4; MEFs, n ¼ 3). B, HeLa cells wereincubated with increasing concentrations of etoposide, or with 10 mmol/LA23187, for 18 hours. Top, lysates were immunoprecipitated with a mixture ofanti–AMPK-a1 anda2 antibodies for AMPK assay. Results are expressed relativeto themean activity (�SEM) in incubationswithout etoposide or A23187 (n¼ 3).Results significantly different from the DMSO control by one-way ANOVA areindicated. Equal amounts of protein from lysates from two of the threeexperiments were analyzed by Western blotting using anti-pSMC1, total SMC1,anti-pT172 (phospho-AMPK), and amixture of anti-a1 anda2 antibodies.C,AsB,but using G361 cells.

Vara-Ciruelos et al.

Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research348

Effect of etoposide/A23187 on localization of pATM and pAMPK (HeLa cells)DAPI

A

B

C

anti-pT172anti-pATM Merge (pATM/pT172)

DM

SO

Eto

posi

deA

2318

7

****

***

nsns

pT172 (HeLa)12468

10121416

Fluo

resc

ence

(% c

ontro

l)

12468

10121416

Etoposide:A23187:

Nuclear Cytoplasmic

++

++

Fluo

resc

ence

(% c

ontro

l)

12

4

6

8

12

4

6

8

Etoposide:A23187:

Nuclear Cytoplasmic

++

++

pT172 (G361)****

***

****

ns

pATM (G361)****

nsnsns

Effect of etoposide/A23187 on localization of pATM and pAMPK (quantification)

**** pATM (HeLa)

nsns

ns

5

10

15

20

Fluo

resc

ence

inte

nsity

(arb

itrar

y un

its)

+ + +Etoposide:Nuclear Cytoplasmic Whole cell

Effect of etoposide on localization of pAMPK in HEK-293 cells (quantification)

0.5

1.0

pT17

2 flu

ores

cenc

e (n

ucle

us:c

ytop

lasm

)

+Etoposide:

****nsnsns

Figure 2.

Etoposide activates AMPK in thenucleus. A, Fluorescencemicrographs of HeLa cells treated for18 hours with DMSO (control), 100mmol/L etoposide, or 10 mmol/LA23187. Cells were fixed,permeabilized, and stainedwith 40 ,6-diamidino-2-phenylindole (DAPI,blue), anti-pATM labeled with FITC(green) and anti-pT172 labeled withTexas red. In the right-hand images,the green and red channels havebeen merged to assess thecolocalization of pATM and pT172.Arrows, nuclei prominently labeledwith anti-pT172 antibody. B,Quantification of nuclear andcytoplasmic staining with anti-pATMand anti-pT172 in HeLa cells (left)and G361 cells (right). Using ImageJsoftware, the total cell area wasdefined, as was the area of the nucleidefined by DAPI staining. The meanfluorescence intensity in the nucleusand the cytoplasm of the etoposide-or A23187-treated cells is expressedrelative to that of DMSO controls.Results are mean � SEM (n ¼ 10),and statistical significance ofdifferences from DMSO controls isindicated (ns, not significant). C,Quantification of effects ofetoposide on nuclear andcytoplasmic localization of pAMPK inHEK-293 cells; analysis as in B.Results are mean� SEM (n¼ 18); ns,not significant.

Etoposide-Induced AMPK Activation Enhances Cell Survival

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 349

to etoposide occurs exclusively with the a1 isoform. To confirmthis, we immunoprecipitated complexes from cell lysates ofHeLa,G361, and A549 cells using a1- or a2-specific antibodies andmeasured kinase activity. The results (Fig. 3B–D) showed that 30to 300 mmol/L etoposide caused progressive activation of a1complexes in all three cell types. InHeLa andA549 cells, etoposidefailed to activate a2 complexes (Fig. 3B and D), while in G361cells, the activity of a2 complexes was too low tomeasure reliably(Fig. 3C). As a positive control, the Ca2þ ionophore A23187markedly activated both a1 and a2 complexes in HeLa andA549 cells, and a1 complexes in G361 cells. We also performedWestern blots in A549 cells (Fig. 3E), which showed, similar to theresults in HeLa and G361 cells (Fig. 1B and C), that pSMC1phosphorylation was increased by etoposide but not A23187,whereas there were small increases in Thr172 phosphorylation inresponse to etoposide and a larger increase in response toA23187,correlating with the changes in AMPK activity. Interestingly,phosphorylation of the downstream target of AMPK, ACC, didnot appear to increase in response to etoposide, although it clearlyincreased in response to A23187 (Fig. 3E).

AMPK activation by etoposide requires CaMKK2 but not ATMAs AMPK activation by etoposide in HeLa, G361, and A549

cells occurs in the absence of LKB1, we examined whether it

required the other well-established upstream kinase for AMPK,CaMKK2. Figure 4A shows that the CaMKK2 inhibitor STO609had no effect on etoposide-induced phosphorylation of SMC1(which is catalyzed by ATM), but did inhibit basal AMPKactivity and Thr172 phosphorylation, as well as the increasesinduced by etoposide (although at the concentration used itdid not block the effects of etoposide completely). Figure 4Aalso shows that (as in A549 cells, Fig. 3E), etoposide did notincrease phosphorylation of the classical AMPK target ACC. AsSTO-609 is not completely selective for CaMKK2 (17), we alsoassessed the requirement for CaMKK2 via siRNA knockdown.Western blots show that we obtained effective CaMKK2 knock-down that blocked AMPK activation and Thr172 phosphory-lation in response to both etoposide (Fig. 4B) and A23187 (Fig.4C). As expected, CaMKK2 knockdown did not affect etopo-side-induced phosphorylation of the ATM target SMC1 (Fig.4B), although it prevented phosphorylation of ACC in responseto A23187 (Fig. 4C).

To examine the involvement ofATM,weused theATM-selectiveinhibitor KU-55933 (25). As expected, KU-55933 blocked thephosphorylation of the ATM target SMC1 in response to etopo-side. However, although it appeared to cause a modest reductionin basal activity and Thr172 phosphorylation of AMPK, it did notprevent the increased AMPK activity and Thr172 phosphorylation

pT172 (red)AMPK-α1 (green)

pT172 (red)AMPK-α2 (green)

DMSO Etoposide (30 µmol/L)

Effect of etoposide on phosphorylation of AMPK-α1/α2 (HeLa cells)

1

2

35

15

A

B C

D E

AM

PK

Isof

orm

act

ivity

(rel

ativ

e to

con

trol)

Effect on AMPK activity (HeLa cells)

10

**

****

*******

****

nsnsns

DMSO ETOP30 µmol/L

ETOP300 µmol/L

ETOP100 µmol/L

A2318710 µmol/L

α1 α1α1α1α1 α2 α2α2α2α2

Effect on AMPK-α1 activity (G361 cells)

1

2

3

4

6

5

DMSO ETOP30 µmol/L

ETOP300 µmol/L

ETOP100 µmol/L

A2318710 µmol/L

AM

PK

-α1

Act

ivity

(rel

ativ

e to

con

trol)

α1 α1α1α1α1

****

************

1

2

3468

Effect on AMPK activity (A549 cells)

DMSO ETOP10 µmol/L

ETOP300 µmol/L

ETOP100 µmol/L

A2318710 µmol/L

α1 α2

AM

PK

Isof

orm

act

ivity

(rel

ativ

e to

con

trol)

α1 α2α1 α2α1 α2α1 α2 α1 α2ETOP

30 µmol/L

pSMC1

pT172

tubulinACC

pACC

DMSO ETOP10 µmol/L

ETOP300 µmol/L

ETOP100 µmol/L

A2318710 µmol/L

ETOP30 µmol/L

Effect on phosphorylation (A549 cells)

SMC1

AMPK-α

******

****

****

****

****

ns nsns ns

Figure 3.

Activation of AMPK by etoposide isspecific for the a1 isoform. A, HeLacells were treated in triplicate withDMSO (control) or with 30 mmol/Letoposide for 18 hours. Lysates wereprepared and equal protein loadingsanalyzed by SDS-PAGE andWesternblotting using anti-pT172 labeledwith IRDye 680 (red), together witheither anti–AMPK-a1 (top) or anti–AMPK-a2 (bottom) labeled withIRDye 800 (green). B, HeLa cellswere treated as in Fig. 1B, except thata1- or a2-containing complexeswere immunoprecipitatedseparately before AMPK assay. C, AsB, but using G361 cells; the activity ofa2-containing complexes was toolow to measure reliably. D, As B, butusing A549 cells. In B–D, results aremean� SEM (n¼ 4 inB, n¼ 6–9 inC,n ¼ 7–10 in D); statistical significnceof differences from DMSO controls isindicated. E, Western blots oftriplicate samples from theexperiment shown in D.

Vara-Ciruelos et al.

Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research350

caused by etoposide (Fig. 4D). Thus, in these cells, the regulationof AMPK by etoposide is ATM independent.

AMPK activation is associated with increases in nuclear [Ca2þ]As activation and Thr172 phosphorylation of AMPK in response

to etoposide required CaMKK2, we suspected that it might beassociated with increases in nuclear Ca2þ concentration. Figure 5Ashows a time course ofAMPKactivation after additionof etoposidein HeLa cells. After a lag of around 3 hours, AMPK activationbecame evident by 6 hours and increased further up to 18 hours.We used the Ca2þ-sensitive dye fluo-4, administered to cells in theform of the acetoxymethyl ester fluo-4 AM, to estimate changes inCa2þ within the nucleus and cytoplasm at 3-hour intervals. Theratio of fluorescence at the time of measurement and at time zero(F/F0), was used as an index of [Ca2þ]. Interestingly, slow, pro-gressive increases in nuclear [Ca2þ] were observed, which followeda similar time course to changes in AMPK activity, becomingsignificant at 6 hours and continuing to increase up to about 18hours. Cytoplasmic Ca2þ did not exhibit such significant changesover the same time points (Fig. 5B).

A23187 enhances cell survival during etoposide treatment viaAMPK activation

Using the CRISPR-Cas9 system, we generated HeLa and G361cells with single (a1KO or a2KO) or double (DKO) knockouts ofthe AMPK catalytic subunits. Figure 6A shows that a1 and/or a2were absent from the knockout HeLa cells as expected. Interest-ingly, a2 expression was markedly upregulated in a1KO HeLacells.Wewere unable to detecta2byWestern blotting in theG361cells (Fig. 6B), consistent with our failure to detect it using kinaseassays (see above). Also consistent with this, phosphorylation of

the AMPK target ACC in response to the AMPK activator H2O2

(26) was abolished in G361 cells by knockout of a1 only,although in HeLa cells, this required knockout of both a1 anda2 (Fig. 6A and B).

We next used two types of cell survival assay to assess the effectof prior AMPK activation by A23187 on cell death induced byetoposide treatment of HeLa cells. Using MTT assays and treat-ment for 48 hours (Fig. 6C), 10 or 30 mmol/L etoposide reducedcell survival to 21% and 7% of the DMSO control, respectively.However, prior treatment with A23187 for 6 hours to activateAMPK significantly enhanced cell survival to 47% and 14% ofcontrol, respectively. Similar protective effects of A23187 wereseen ina1KOanda2KOcells, although the basal levels of survivalappeared to be slightly higher in these cells. However, the pro-tective effect of A23187 treatment was completely abolished inDKO cells at both etoposide concentrations. Similar results wereobtained using themore sensitive clonogenic survival assays (Fig.6D), althoughmuch lower concentrations of etoposide (100 and250 nmol/L) and shorter incubation times (18 hours) were usedin these assays. One other difference was that clonogenic survivalwas significantly lower in DKO cells at both 100 and 250 nmol/Letoposide even in cells not treated with A23187, an effect not seenusing the MTT assays. Figure 6E confirms using anti-pSMC1 blotsthat ATM was activated to a similar extent at all four concentra-tions of etoposide used in Fig. 6C and D and that this wasunaffected by single or double knockouts of the two AMPKcatalytic subunit isoforms.

We suspected that the protection against cell death provided byAMPKactivationmight be due toG1 cell-cycle arrest, whichwouldreduce entry of cells into S phase, where they would be morevulnerable toDNAdamage induced by etoposide. To confirm that

AM

PK

Act

ivity

(% c

ontro

l)

**

*****

pSMC1SMC1pT172

AMPK-α1

Actin

AMPK-α2pACC

ACC

Etoposide:STO-609:

100

200

A B

C D+++ ++++ +

Effect of STO-609 on etoposide activation

pT172AMPK-α1

GAPDH

AMPK-α2pACC

ACC

A23187: +++ +Scrambled siRNA CaMKK2 siRNA

Effect of CaMKK2 knockdown on A23187 activation

200

400

600

800

AM

PK

Act

ivity

(% c

ontro

l)

**** ****

ns

Effect of CaMKK2 knockdown on etoposide activation

100

200

AM

PK

Act

ivity

(% c

ontro

l)

CaMKK2pSMC1

SMC1pT172

AMPK-α1AMPK-α2

Etoposide: +++ +Scrambled siRNA CaMKK2 siRNA

** ***

ns

100

Effect of KU-55933 on etoposide activation

50

150

Etoposide:KU-55933:

pSMC1

SMC1

pT172AMPK-α

Actin

+ ++++ +++

AM

PK

Act

ivity

(% c

ontro

l) *

* ns

*

***

Figure 4.

Etoposide effects on AMPK requireCaMKK2 but not ATM. A, HeLa cellswere treatedwith 30mmol/L etoposidefor 18 hours with or withoutpretreatment for 1 hour with 25 mmol/LSTO-609. Total AMPK wasimmunoprecipitated and assayed (top;mean� SEM, n¼ 3) andWestern blotsof duplicate dishes of cells analyzedusing various antibodies (bottom). B,HeLa cellswere treatedwith scrambledsiRNA or siRNA targeted at CaMKK2for 24 hours, prior to treatment with orwithout 30 mmol/L etoposide for 18hours; other analyses as in A. C, HeLacells were treated with scrambledsiRNA or siRNA targeted at CaMKK2for 24 hours, prior to treatment with orwithout 10 mmol/L A23187 for 1 hour;other analyses as in A. D, HeLa cellswere treatedwith 30mmol/L etoposidefor 18 hours with or withoutpretreatment for 1 hour with 10 mmol/LKU-55933. Total AMPK wasimmunoprecipitated and assayed (top;mean� SEM, n¼ 3) andWestern blotsof duplicate dishes of cells analyzedusing various antibodies (bottom).

Etoposide-Induced AMPK Activation Enhances Cell Survival

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 351

A23187 induced G1 arrest in HeLa cells, we treated them with orwithout A23187 in the presence of nocodazole (which preventscells that have already traversed the G1:S boundary from progres-sing through into the subsequent G1 phase). Figure 6F shows thatA23187 caused a >10-fold increase in G1:G2 ratio in WT cells, asexpected for an agent causing G1 arrest. Similar results wereobtained in a1KO or a2KO cells, showing that either isoform iscapable of causing cell-cycle arrest. However, a significant increasein G1:G2 ratio was not observed in DKO cells.

Other treatments causing G1 arrest provided similar protectionagainst etoposide

To test whether other treatments that caused G1 arrest wouldprovide protection against etoposide-induced cell death, weinitially used palbociclib (also known as PD332991), an inhib-itor of the cyclin-dependent kinases CDK4/CDK6 (27, 28).Interestingly, palbociclib did not cause cell-cycle arrest in HeLa

cells (Fig. 7A). However, in G361 cells, palbociclib from 0.1 to10 nmol/L caused a progressive increase in the proportion ofG361 cells in G1 phase, with corresponding decreases in theproportions in S and G2–M phases (Fig. 7B). Tellingly, palbo-ciclib did not protect against cell death induced by etoposide inclonogenic survival assays in HeLa cells (Fig. 7C), but providedmarked protection in G361 cells (Fig. 7D). We have shownpreviously that A23187 treatment of G361 cells causes a G1

arrest that is abolished in AMPK DKO cells (22). Supplemen-tary Figure S4 confirms that, similar to HeLa cells, prior treat-ment with A23187 protected G361 cells against cell deathinduced by etoposide in clonogenic survival assays, with theeffect being abolished in DKO cells.

To confirm that another agent that caused cell-cycle arrestwould protect HeLa cells against etoposide, we used aphidicolin,which causes arrest in early S-phase by inhibiting DNA polymer-ase-a (29).Unlike palbociclib, aphidicolin caused amarked arrestat the G1–S boundary (Fig. 7E) and also provided significantprotection against cell death induced by etoposide in clonogenicsurvival assays (Fig. 7F).

DiscussionWe have confirmed that the DNA-damaging agent etoposide

activates AMPK in several different cell lines as reported pre-viously (11, 14), but also present new findings that AMPKactivation by etoposide occurs primarily in the nucleus and isspecific for complexes containing the a1 isoform of the catalyticsubunit even when a2 was also expressed, as in HeLa and A549cells. As these cells do not express LKB1, our results rule out arole for that kinase. We show instead that the effect is catalyzedby the alternate upstream kinase CaMKK2, as it was reducedeither by the CaMKK inhibitor STO-609, or by knocking downCaMKK2 using siRNA. The correlation between AMPK activa-tion and nuclear [Ca2þ] (Fig. 5) suggests that activation ofAMPK by etoposide may be mediated by increases in nuclear[Ca2þ] rather than any intrinsic change in CaMKK2 activity,although the source of the increased nuclear Ca2þ, and themechanism by which it is released in response to etoposideand/or DNA damage, remains unclear.

We were initially unable to detect significant Thr172 phos-phorylation or activation of AMPKby etoposide inHEK-293 cells,which express normal levels of LKB1. However, we subsequentlyobserved increased Thr172 phosphorylation in the nuclei ofHEK-293 cells by immunofluorescencemicroscopy. We believe that wecould not detect changes in global Thr172 phosphorylation orAMPK activation in these cells because the high basal levels ofthese parameters in the cytoplasm (due to the high basal activityof LKB1) were obscuring changes in the small pool of AMPKwithin the nucleus.

Itwas previously reported that AMPKactivationby etoposide ina prostate cancer cell line (C4-2) was reduced by knocking downATM or LKB1 using RNAi (14). As ATM had been previouslyshown to phosphorylate LKB1 at Thr366 (13), it was proposed(14) that there was a kinase cascade from ATM to LKB1 to AMPK.However, Thr366 phosphorylation has been reported to have noeffect on either the activity or the localization of LKB1 in G361cells (13).Moreover, thismechanism cannot explain the effects inthe cells we studied because: (i) in HeLa cells, etoposide activatedAMPK via amechanism that required CaMKK2, but not LKB1; (ii)although etoposide activated ATM in HeLa cells, the ATM

2 4 6 8 10 12 15 20

50

100

150

200

A

B

AM

PK

Act

ivity

(% in

itial

)Time course of AMPK activation by etoposide

Time after addition (hr)

Nuclear/cytoplasmic [Ca2+] after addition of etoposide

** ** ***

Rat

io F

/F0

(nuc

leus

)

1.0

1.5 ******

********

****

Rat

io F

/F0

(cyt

opla

sm)

Time after addition (hr)3 6 9 12 15 18 21 24

1.0

1.5

NUCLEUS

CYTOPLASM

Etoposide

Control

Control

Etoposide

Figure 5.

The effect of etoposide on AMPK correlates with an increase in nuclearCa2þ. A, Time course of AMPK activation in HeLa cells during treatmentwith 30 mmol/L etoposide; results are mean � SEM (n ¼ 3); valuessignificantly different from zero time value are indicated. B, Time course ofchanges in nuclear and cytoplasmic Ca2þ in HeLa cells during treatmentwith 30 mmol/L etoposide. Results (F/Fo, i.e., fluorescence as a ratio ofmean fluorescence at time zero) are expressed as mean � 95%confidence intervals (n ¼ 12) with significant differences betweenetoposide-treated and control indicated at each time point.

Vara-Ciruelos et al.

Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research352

inhibitor KU-55933 did not prevent increased AMPK activity andThr172 phosphorylation in response to etoposide, despite block-ing phosphorylation of a known ATM target, SMC1.

One interesting question raised by our study is why AMPKactivation by etoposide is restricted to nuclear a1 complexes.LKB1 is only catalytically active when it forms a complex with the

0.5

1.0

1.5

2.0

Cel

l cyc

le (G

1:G

2 ra

tio) HeLa cells

A23187: +++ +WT α1KO α2KO DKO

Cell cycle arrest by A23187 in AMPK KO cells

***

****

****

ns

Characterization of CRISPR knockout HeLa cells

AMPK-α1AMPK-α2

ACCpACC

H2O2: + + ++WT α2KOα1KO DKO

Characterization of CRISPR knockout G361 cells

AMPK-α1

ACCpACC

H2O2: + + ++WT α2KOα1KO DKO

AMPK-α2

20

40

60

80

Col

ony

surv

ival

(% c

ontro

l)A23187: +

Etoposide (100 nmol/L)

+++WT α1KO α2KO DKO

+

Etoposide (250 nmol/L)

+++WT α1KO α2KO DKO

*******

****

****

††††

†††

††

†††††††

Protection against etoposide by A23187 (clonogenic)Protection against etoposide by A23187 (MTT assay)

20

40

60

Cel

l sur

viva

l (%

con

trol)

A23187: +

Etoposide (10 µmol/L)

+++WT α1KO α2KO DKO

+

Etoposide (30 µmol/L)

+++WT α1KO α2KO DKO

HeLa cells

ns

ns

***

HeLa cells

****

****

****

***

****

*ns

ns

pSMC1SMC1

GAPDH

WT α1KO α2KO DKO WT α1KO α2KO DKODMSO Etoposide

100

nmol/

L30

µm

ol/L

10 µ

mol/

L25

0 nm

ol/L

pSMC1SMC1

GAPDH

pSMC1SMC1

GAPDH

pSMC1SMC1

GAPDH

SMC1 phosphorylation in AMPK KO cells

pT172A B

C D

E F

pT172

[Eto

posi

de]

Figure 6.

HeLa and G361 cells are protected against etoposide-induced cell death by prior AMPK activation using A23187. A, Parental HeLa cells (WT) or HeLa cells withCRISPR-Cas9 knockouts of AMPK-a1 (a1KO), AMPK-a2 (a2KO) or both (DKO) were treated in duplicate with 1 mmol/L H2O2 for 10 minutes. Cell lysateswere then analyzed byWestern blotting using the indicated antibodies.B,AsA, but usingG361 cells.C, Survival ofWT, single, or double knockout HeLa cells assessedusing MTT assays, following treatment with 10 or 30 mmol/L etoposide for 18 hours, with or without prior treatment with 10 mmol/L A23187 for 6 hours.Results aremean� SEM (n¼ 12).Where appropriate, statistical significance is indicated by asterisks for differences between treatmentswith andwithout A23817; ns,not significant. D, Survival of WT, single, or double knockout HeLa cells assessed using clonogenic assays, following treatment with 100 or 250 nmol/Letoposide for 18 hours, with orwithout prior treatmentwith 10mmol/LA23187 for 6 hours. Results are numbers of colonies counted (mean� SEM; n¼4) expressed asa percentage of colony numbers in controls without etoposide. Where appropriate, statistical significance is indicated by asterisks for differences betweentreatments with and without A23817, and by daggers (†) for differences between WT and DKO; ns, not significant. E, Phosphorylation of SMC1 in HeLa cells treatedwith or without different concentrations of etoposide as in C and D; F, Cell-cycle arrest in WT, single, or double knockout HeLa cells treated with or without3 mmol/L A23187 for 6 hours and then 70 ng/mL nocodazole for a further 18 hours. The cells were then fixed, stained, and analyzed by flow cytometry to determineDNA content and hence cell-cycle phase. Results are expressed as ratios of cells in G1:G2 phase (mean� SEM; n¼ 4). Asterisks indicate the significance of differencesbetween cells treated with and without A23187; ns, not significant.

Etoposide-Induced AMPK Activation Enhances Cell Survival

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 353

PBC does not protect HeLa cells against etoposide

20

40

60

80

100

PBC: + ++Etoposide

(100 nmol/L)Etoposide

(250 nmol/L)DMSO

Col

ony

surv

ival

(% c

ontro

l)

PBC does not cause cell-cycle arrest in HeLa cells

0PBC (µmol/L): 5.01.00.50.1 10

PBC protects G361 cells against etoposide

20

40

60

80

100

Col

ony

surv

ival

(% c

ontro

l)

PBC: + ++Etoposide

(100 nmol/L)Etoposide

(250 nmol/L)DMSO

****

***

G1 G2S

20

40

60

80

A B

C D

E F

Cel

l-cyc

le p

hase

(%)

20

40

60

80

0PBC (µmol/L): 5.01.00.50.1 10

****

********

****

****

************

***

********

****

****

***

PBC does cause cell-cycle arrest in G361 cells

Cel

l-cyc

le p

hase

(%)

G1 G2S

0Aph (µmol/L): 10510.5 50

20

40

60

80

Cel

l-cyc

le p

hase

(%)

Aphidicolin causes cell-cycle arrest in HeLa cells

************

********

****

*

************

****

****

Aphidicolin protects against etoposide in HeLa cells

20

40

60

80

100

Col

ony

surv

ival

(% c

ontro

l)

Aph (5 µmol/L): + ++Etoposide

(100 nmol/L)Etoposide

(250 nmol/L)DMSO

****

****

Figure 7.

Protection against cell death using the CDK4/6 inhibitor palbociclib (PBC) correlateswith its ability to cause a G1 cell-cycle arrest.A,HeLa cells were treatedwith theindicated concentrations of palbociclib for 6 hours and then with nocodazole (70 ng/mL) for 18 hours, after which cells were fixed, stained with propidiumiodide, and subjected to cell-cycle analysis by flow cytometry. Results (mean � SEM; n ¼ 4) show the proportion of cells in each cell-cycle phase; none of thedifferences between palbociclib treatment and controls were significant. B, As A, but using G361 cells; asterisks show statistically significant differences fromcontrolswithout palbociclib.C,Clonogenic survival of HeLa cells after treatmentwith orwithout palbociclib (3 mmol/L for Hela cells; 500 nmol/L forG361) for 6 hours,followed by treatment for a further 18 hours with vehicle (DMSO), or 100 or 250 mmol/L etoposide. Results are numbers of colonies counted (mean � SEM;n¼ 4) expressed as a percentage of survival in controls without etoposide. None of the effects of palbociclib were significant.D, As C, but using G361 cells; asterisksrepresent statistical significance of effects of palbociclib. E, Effect of different concentrations of aphidicolin (Aph) on the cell cycle of HeLa cells, analyzed asin A and B. Asterisks indicate significant differences from the control without aphidicolin for each cell-cycle phase. F, Effect of 5 mmol/L aphidicolin on clonogenicsurvival of HeLa cells treated with etoposide, analyzed as in C and D. Asterisks indicate significant differences from controls without etoposide.

Vara-Ciruelos et al.

Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research354

accessory subunits STRAD and MO25 (30), and association withthese subunits causes its exclusion from the nucleus (31), imply-ing that LKB1 may only be able to activate cytoplasmic AMPK.Consistent with this, using a FRET-based reporter that can mon-itor AMPK activity in real time in single cells (32), AMPK activa-tion by 2-deoxyglucose (an inhibitor of glycolysis that activatesAMPK by increasing cellular AMP, and is therefore LKB1 depen-dent; ref. 33) occurred only in the cytoplasm, whereas activationby A23187 (a Ca2þ ionophore that works via the Ca2þ/CaMKK2pathway independently of AMP; ref. 33) occurred initially in thecytoplasm and then in the nucleus. These results suggest thatCaMKK2must be present in the nucleus, as if the effect was due totranslocation of activated AMPK, it should work equally well aftertreatment with 2-deoxyglucose or A23187. There is also evidencethata2-containing complexes are activated primarily by the AMP-and LKB1-dependent mechanism rather than the Ca2þ/CaMKK2mechanism in vivo. For example, conditional LKB1 knockout inskeletal and cardiacmuscle prevented activation ofa2-containingcomplexes in response to contraction in skeletal muscle or ische-mia in cardiac muscle, while having little or no effect on theactivity of a1-containing complexes (34, 35). The reasons for thisapparent selectivity of upstream kinases for a1 or a2-containingcomplexes in vivo remain unclear, because the LKB1:STRAD:MO25 complex (24) and CaMKK2 (36) phosphorylate andactivate both a1- or a2-containing complexes in cell-free assays,while treatment with the Ca2þ ionophore A23187 can activate a2complexes in intact cells (Fig. 2B). One possibility is that thisisoform selectivity is due to different subcellular locations of theupstream kinases and the AMPK complexes containing the dif-ferent a isoforms. Both a1 and a2 contain well-defined nuclearexport sequences (37), while nuclear localization sequences areless well defined, although a short conserved basic sequence ina2(also present in a1) has been proposed to fulfill that role (38). Ifa1 complexes were more abundant than a2 complexes in thenuclei of unstimulatedHeLaorG361 cells, thismight explainwhyonly the former were activated by etoposide. Arguing against thispossibility, however, is evidence that it is a2 rather than a1 that isenriched in the nucleus (39), although that was obtained usingdifferent cell types. Interestingly, etoposide treatment did notcause phosphorylation of ACC at the AMPK site Ser79, unlikeA23187 (Figs. 3E and 4A and C). ACC phosphorylation is auniversally used biomarker for AMPK activation, and etoposideis the first AMPK activator we have studied that does not trigger it.The obvious explanation is that etoposide activates AMPK pri-marily in the nucleus, from which ACC (a large cytoplasmicprotein) is excluded.

Another novel finding in our study was that prior elevation ofintracellular Ca2þ for 6 hours using the Ca2þ ionophoreA23187 protected against cell death induced by etoposide,both in short-term MTT assays and in clonogenic survivalassays. In both cases, the effect was AMPK dependent, becauseit was eliminated in HeLa and G361 cells when both catalyticsubunit isoforms (a1 and a2) were knocked out using theCRISPR-Cas9 system. Single knockouts did not abolish theeffect in HeLa cells, suggesting that either isoform is capableof exerting the protective effect, although we have found thatonly a1 complexes are activated by etoposide. A caveat here,however, is that knockout of AMPK-a1 caused a marked upre-gulation of AMPK-a2 expression (Fig. 6A), raising the possi-bility that overexpressed AMPK-a2 might perform functionsthat endogenous levels might not. A notable difference between

the MTT and the clonogenic survival assays is that effects ofetoposide on cell survival were observed in clonogenic assays atmuch lower concentrations (100–250 nmol/L, as opposed to10–30 mmol/L for MTT assays). This is a common finding andmay occur because a low level of DNA damage, which may bequite difficult to detect, is nevertheless sufficient to preventclonal growth of single cells.

These findings to some extent turn the original view on therole of AMPK in cancer on its head. Because AMPK is imme-diately downstream of, and activated by, the tumor suppressorLKB1, because it inhibits cell growth and proliferation andswitches off the growth-promoting target-of-rapamycin com-plex-1 (TORC1) when activated (5, 40), and because use of theAMPK-activating drug metformin is associated with a lower riskof cancer in diabetics (41), it had been widely assumed thatAMPK was a tumor suppressor. Although AMPK may indeedinitially suppress the development of rapidly growing tumorsand there may therefore be selection pressure for the LKB1–AMPK pathway to be downregulated (42), complete loss offunction of AMPK in human cancers appears to be rare. In fact,there is increasing evidence that AMPK can under some circum-stances enhance the growth of tumor cells, perhaps by protect-ing solid tumors against the environmental and nutritionalstresses that occur before their new blood supply has been fullyestablished (5, 10, 43, 44). For example, a double knockout ofa1 and a2 isoforms in immortalized MEFs prevented theirgrowth as xenografts in immunodeficient mice (18). However,while knocking out a2 alone accelerated growth of H-RasV12–transformed MEFs in vivo, knocking out a1 completely pre-vented growth, suggesting that a1 but not a2 is required fortumor growth in vivo (45). These results are also consistent withresults of recent data mining from the human cancer genomeprojects, which revealed that although the PRKAA1 gene(encoding a1) is frequently amplified in human cancers (sug-gesting that this is a genetic change for which positive selectionhas occurred), the PRKAA2 gene (encoding a2) undergoes quitefrequent mutations instead, more consistent with it being atumor suppressor (5, 10). The results shown in Fig. 6 providethe novel finding that AMPK activation using A23187 in LKB1-null tumor cells protects them against cell death induced by theDNA-damaging agent, etoposide and that this is abolished inthe absence of AMPK. As etoposide treatment activates AMPKon its own, the presence of AMPK should be sufficient to protectthe cells against etoposide, even in the absence of A23187. Thiswas indeed the case in clonogenic survival assays in HeLa cells,where cell death was enhanced in the DKO cells not treatedwith A23187 (Fig. 6D). These results suggest that an AMPKinhibitor (particularly if selective for a1 complexes) might be auseful adjunct to treatment with etoposide, and perhaps othercancer therapies that damage DNA, such as doxorubicin orradiotherapy.

What is the mechanism by which AMPK protects tumor cellsagainst the effects of DNA-damaging agents such as etoposide?AMPK activation using the pharmacologic activator 5-aminoimi-dazole-4-carboxamide ribonucleoside (46), or by glucose depri-vation or overexpression of a mutant (T172D) AMPK kinasedomain (47), causes cell-cycle arrest inG1 phase.Wehave recentlyconfirmed that these effects of AMPK activators, which are asso-ciated with increased expression of the cyclin-dependent kinaseinhibitor p21 (CDKN1A), are AMPK dependent as they wereabolished by a double knockout of AMPK (22). A cell-cycle arrest

Etoposide-Induced AMPK Activation Enhances Cell Survival

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 355

in G1 phase would limit the entry of cells into S-phase, where theyare particularly vulnerable to the generation of double-strandedDNA breaks induced by etoposide while DNA is being replicated.Supporting this proposal, palbociclib, a potent and selectiveinhibitor of the G1 cyclin-dependent kinases CDK4/CDK6 (27,28), caused G1 arrest in G361 cells while not arresting HeLa cellsand, correlating with this, prior treatment of G361 but not HeLacells with palbociclib provided marked protection against celldeath induced by etoposide. In contrast, aphidicolin caused aG1–S phase cycle arrest in HeLa cells and also protected themagainst cell death induced by etoposide. Sensitivity of differenttumor cell lines to palbociclib has been shown to be inverselycorrelated with the expression of the CDK inhibitor p16(CDKN2A; refs. 48, 49). This may explain why the effect of theinhibitor differs between these two cell lines, as p16 is expressed atmuch higher levels in HeLa than in G361 cells (50).

In summary, DNA-damaging treatments such as etoposide(11) and ionizing radiation (15) have been previously reportedto activate AMPK, and it was suggested that the effect ofetoposide occurred via a kinase cascade from ATM to LKB1 toAMPK (14). However, we show that the effect of etoposide isindependent of ATM and LKB1 and involves instead an increasein nuclear Ca2þ that causes activation of CaMKK2. AMPKactivation is restricted to the a1 isoform and occurs within thenucleus, so that phosphorylation of ACC, the classical markerfor AMPK activation, is not observed. Moreover, prior AMPKactivation using A23187 in two different LKB1-null tumor celllines protects against cell death induced by etoposide, and thiswas AMPK dependent because the effect was abolished whenboth isoforms of AMPK were knocked out. Taken together withfindings that AMPK-a1 is required for growth of transformedMEFs as tumors in vivo (45), and that the gene encoding AMPK-a1 is frequently amplified in human tumors (5, 10), thissuggests that AMPK inhibitors, and particularly selective inhi-bitors of AMPK-a1, might be useful adjuncts to cytotoxic drugs,and perhaps also radiotherapy, in human cancer.

Disclosure of Potential Conflicts of InterestD.G.Hardie reports receiving a commercial research grant fromSchool of Life

Sciences at Dundee (7.2 M GBP over 5 years from a consortium of threecompanies (GlaxoSmithKline, Boehringer, Merck KGaA; see www.lifesci.dundee.ac.uk/search/site/DSTT; the authors' laboratory is associated with this, butcurrently receives no direct funding from it). No potential conflicts of interestwere disclosed by the other authors.

Authors' ContributionsConception and design: M. Dandapani, A.M. Evans, D.G. HardieDevelopment of methodology: D. Vara-Ciruelos, M. Dandapani, A. Gray,A.M. EvansAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): D. Vara-Ciruelos, M. Dandapani, E.O. Egbani,A.M. EvansAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): D. Vara-Ciruelos, M. Dandapani, E.O. Egbani,A.M. Evans, D.G. HardieWriting, review, and/or revision of the manuscript: D. Vara-Ciruelos,M. Dandapani, A. Gray, A.M. Evans, D.G. HardieAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): D. Vara-Ciruelos, A.M. Evans, D.G. HardieStudy supervision: A.M. Evans, D.G. Hardie

AcknowledgmentsD. Vara-Ciruelos was supported by a Programme Grant (C37030/A15101)

awarded to D.G. Hardie by Cancer Research UK, M. Dandapani by a ClinicalPhD studentship from theWellcome Trust, and A. Gray by a Senior InvestigatorAward (097726) awarded to D.G. Hardie by the Wellcome Trust. D.G. Hardiewas also indirectly supported by the pharmaceutical companies supporting theDivision of Signal Transduction Therapy at Dundee (Boehringer-Ingelheim,GlaxoSmithKline, and Merck KGaA). O.A. Egbani and A.M. Evans were sup-ported by a Programme Grant (081195) from the Wellcome Trust.

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received June 21, 2017; revised October 11, 2017; accepted November 1,2017; published OnlineFirst November 13, 2017.

References1. Slevin ML. The clinical pharmacology of etoposide. Cancer 1991;67:

319–29.2. Hande KR. Etoposide: four decades of development of a topoisomerase II

inhibitor. Eur J Cancer 1998;34:1514–21.3. Pommier Y, Leo E, Zhang H, Marchand C. DNA topoisomerases and

their poisoning by anticancer and antibacterial drugs. Chem Biol 2010;17:421–33.

4. Carling D. AMPK signalling in health and disease. Curr Opin Cell Biol2017;45:31–7.

5. Ross FA, MacKintosh C, Hardie DG. AMP-activated protein kinase: acellular energy sensor that comes in 12 flavours. FEBS J 2016;283:2987–3001.

6. Oakhill JS, Scott JW, Kemp BE. AMPK functions as an adenylatecharge-regulated protein kinase. Trends Endocrinol Metab 2012;23:125–32.

7. Oakhill JS, Steel R, Chen ZP, Scott JW, Ling N, Tam S, et al. AMPK is a directadenylate charge-regulated protein kinase. Science 2011;332:1433–5.

8. Xiao B, Sanders MJ, Underwood E, Heath R, Mayer FV, Carmena D, et al.Structure of mammalian AMPK and its regulation by ADP. Nature 2011;472:230–3.

9. Ross FA, Jensen TE, Hardie DG. Differential regulation by AMP and ADP ofAMPK complexes containing different gamma subunit isoforms. Biochem J2016;473:189–99.

10. Monteverde T, Muthalagu N, Port J, Murphy DJ. Evidence of cancerpromoting roles for AMPK and related kinases. FEBS J 2015;282:4658–71.

11. Fu X, Wan S, Lyu YL, Liu LF, Qi H. Etoposide induces ATM-dependentmitochondrial biogenesis through AMPK activation. PLoS One 2008;3:e2009.

12. Cremona CA, Behrens A. ATM signalling and cancer. Oncogene 2014;33:3351–60.

13. Sapkota GP, Deak M, Kieloch A, Morrice N, Goodarzi AA, Smythe C, et al.Ionizing radiation induces ataxia telangiectasia mutated kinase (ATM)-mediated phosphorylation of LKB1/STK11 at Thr-366. Biochem J 2002;368:507–16.

14. Luo L, Huang W, Tao R, Hu N, Xiao ZX, Luo Z. ATM and LKB1 dependentactivation of AMPK sensitizes cancer cells to etoposide-induced apoptosis.Cancer Lett 2013;328:114–9.

15. Sanli T, Rashid A, Liu C, Harding S, Bristow RG, Cutz JC, et al. Ionizingradiation activates AMP-activated kinase (AMPK): a target for radio-sensitization of human cancer cells. Int J Radiat Oncol Biol Phys2010;78:221–9.

16. WoodsA, Salt I, Scott J,HardieDG,CarlingD. The a1and a2 isoformsof theAMP-activated protein kinase have similar activities in rat liver but exhibitdifferences in substrate specificity in vitro. FEBS Lett 1996;397:347–51.

17. Hawley SA, Pan DA, Mustard KJ, Ross L, Bain J, Edelman AM, et al.Calmodulin-dependent protein kinase kinase-beta is an alternativeupstream kinase for AMP-activated protein kinase. Cell Metab 2005;2:9–19.

18. Laderoute KR, Amin K, Calaoagan JM, Knapp M, Le T, Orduna J, et al.50-AMP-activated protein kinase (AMPK) is induced by low-oxygen and

Mol Cancer Res; 16(2) February 2018 Molecular Cancer Research356

Vara-Ciruelos et al.

glucose deprivation conditions found in solid-tumor microenvironments.Mol Cell Biol 2006;26:5336–47.

19. Hardie DG, Salt IP, Davies SP. Analysis of the role of the AMP-activatedprotein kinase in the response to cellular stress. Methods Mol Biol2000;99:63–75.

20. Towler MC, Fogarty S, Hawley SA, Pan DA, Martin D, Morrice NA, et al. Anovel short splice variant of the tumour suppressor LKB1 is required forspermiogenesis. Biochem J 2008;416:1–14.

21. Dale S,WilsonWA, EdelmanAM,HardieDG. Similar substrate recognitionmotifs for mammalian AMP-activated protein kinase, higher plant HMG-CoA reductase kinase-A, yeast SNF1, and mammalian calmodulin-depen-dent protein kinase I. FEBS Lett 1995;361:191–5.

22. Fogarty S, Ross FA, Vara Ciruelos D, Gray A, Gowans GJ, Hardie DG. AMPKcauses cell cycle arrest in LKB1-deficient cells via activation of CAMKK2.Mol Cancer Res 2016;14:683–95.

23. Hande KR, Wedlund PJ, Noone RM, Wilkinson GR, Greco FA, Wolff SN.Pharmacokinetics of high-dose etoposide (VP-16-213) administered tocancer patients. Cancer Res 1984;44:379–82.

24. Hawley SA, Boudeau J, Reid JL, Mustard KJ, Udd L, Makela TP, et al.Complexes between the LKB1 tumor suppressor, STRADa/b andMO25a/bare upstream kinases in the AMP-activated protein kinase cascade. J Biol2003;2:28.

25. Hickson I, Zhao Y, Richardson CJ, Green SJ, Martin NM, Orr AI, et al.Identification and characterization of a novel and specific inhibitor ofthe ataxia-telangiectasia mutated kinase ATM. Cancer Res 2004;64:9152–9.

26. Auciello FR, Ross FA, Ikematsu N, Hardie DG. Oxidative stress activatesAMPK in cultured cells primarily by increasing cellular AMP and/or ADP.FEBS Lett 2014;588:3361–6.

27. Fry DW, Harvey PJ, Keller PR, Elliott WL, MeadeM, Trachet E, et al. Specificinhibition of cyclin-dependent kinase 4/6 by PD 0332991 and associatedantitumor activity in human tumor xenografts. Mol Cancer Ther 2004;3:1427–38.

28. Wang J, Gray NS. SnapShot: Kinase Inhibitors I. Mol Cell 2015;58:708.29. Ikegami S, Taguchi T,OhashiM,OguroM,NaganoH,ManoY. Aphidicolin

prevents mitotic cell division by interfering with the activity of DNApolymerase-alpha. Nature 1978;275:458–60.

30. Zeqiraj E, Filippi BM, Deak M, Alessi DR, van Aalten DM. Structure of theLKB1-STRAD-MO25 complex reveals an allosteric mechanism of kinaseactivation. Science 2009;326:1707–11.

31. Boudeau J, Baas AF, Deak M, Morrice NA, Kieloch A, Schutkowski M, et al.MO25a/b interact with STRADa/b enhancing their ability to bind, activateand localize LKB1 in the cytoplasm. EMBO J 2003;22:5102–14.

32. Tsou P, Zheng B, Hsu CH, Sasaki AT, Cantley LC. A fluorescent reporter ofAMPK activity and cellular energy stress. Cell Metab 2011;13:476–86.

33. Hawley SA, Ross FA, Chevtzoff C, Green KA, Evans A, Fogarty S, et al. Use ofcells expressing gamma subunit variants to identify diverse mechanisms ofAMPK activation. Cell Metab 2010;11:554–65.

34. Sakamoto K, McCarthy A, Smith D, Green KA, Hardie DG, Ashworth A,et al. Deficiency of LKB1 in skeletal muscle prevents AMPK activation andglucose uptake during contraction. EMBO J 2005;24:1810–20.

35. SakamotoK, ZarrinpashnehE, BudasGR, Pouleur AC,Dutta A, Prescott AR,et al. Deficiency of LKB1 in heart prevents ischemia-mediated activation ofAMPKalpha2 but not AMPKalpha1. Am J Physiol Endocrinol Metab2006;290:E780–E8.

36. Oakhill JS, Chen ZP, Scott JW, Steel R, Castelli LA, Ling N, et al. beta-Subunit myristoylation is the gatekeeper for initiating metabolic stresssensing by AMP-activated protein kinase (AMPK). Proc Natl Acad Sci U S A2010;107:19237–41.

37. Kazgan N, Williams T, Forsberg LJ, Brenman JE. Identification of a nuclearexport signal in the catalytic subunit of AMP-activated protein kinase. MolBiol Cell 2010;21:3433–42.

38. Suzuki A, Okamoto S, Lee S, Saito K, Shiuchi T, Minokoshi Y. Leptinstimulates fatty acid oxidation and peroxisome proliferator-activatedreceptor alpha gene expression in mouse C2C12 myoblasts by changingthe subcellular localization of the alpha2 form of AMP-activated proteinkinase. Mol Cell Biol 2007;27:4317–27.

39. Salt IP, Celler JW, Hawley SA, Prescott A, Woods A, Carling D, et al. AMP-activated protein kinase - greater AMP dependence, and preferentialnuclear localization, of complexes containing the a2 isoform. Biochem J1998;334:177–87.

40. Hardie DG, Schaffer BE, Brunet A. AMPK: an energy-sensing pathway withmultiple inputs and outputs. Trends Cell Biol 2016;26:190–201.

41. Evans JM, Donnelly LA, Emslie-Smith AM, Alessi DR, Morris AD. Metfor-min and reduced risk of cancer in diabetic patients. BMJ 2005;330:1304–5.

42. HardieDG.AMP-activated protein kinase: a key regulator of energy balancewith many roles in human disease. J Intern Med 2014;276:543–59.

43. Jeon SM, HayN. The double-edged sword of AMPK signaling in cancer andits therapeutic implications. Arch Pharm Res 2015;38:346–57.

44. Hardie DG. Molecular Pathways: Is AMPK a friend or a foe in cancer? ClinCancer Res 2015;21:3836–40.

45. Phoenix KN, Devarakonda CV, Fox MM, Stevens LE, Claffey KP. AMPKal-pha2 suppresses murine embryonic fibroblast transformation and tumor-igenesis. Genes Cancer 2012;3:51–62.

46. Imamura K, Ogura T, Kishimoto A, Kaminishi M, Esumi H. Cell cycleregulation via p53 phosphorylation by a 50-AMP activated protein kinaseactivator, 5-aminoimidazole- 4-carboxamide-1-beta-d- ribofuranoside, ina human hepatocellular carcinoma cell line. Biochem Biophys ResCommun 2001;287:562–7.

47. Jones RG, Plas DR, Kubek S, Buzzai M, Mu J, Xu Y, et al. AMP-activatedprotein kinase induces a p53-dependent metabolic checkpoint. Mol Cell2005;18:283–93.

48. Konecny GE, Winterhoff B, Kolarova T, Qi J, Manivong K, Dering J, et al.Expression of p16 and retinoblastoma determines response to CDK4/6inhibition in ovarian cancer. Clin Cancer Res 2011;17:1591–602.

49. Katsumi Y, Iehara T,MiyachiM, Yagyu S, Tsubai-Shimizu S, Kikuchi K, et al.Sensitivity of malignant rhabdoid tumor cell lines to PD 0332991 isinversely correlated with p16 expression. Biochem Biophys Res Commun2011;413:62–8.

50. Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S,et al. The Cancer Cell Line Encyclopedia enables predictive modelling ofanticancer drug sensitivity. Nature 2012;483:603–7.

www.aacrjournals.org Mol Cancer Res; 16(2) February 2018 357

Etoposide-Induced AMPK Activation Enhances Cell Survival