ATM is down-regulated by N-Myc–regulatedmicroRNA-421Hailiang Hua,1, Liutao Dua, Gindy Nagabayashia, Robert C. Seegerb, and Richard A. Gattia,c,1

aDepartment of Pathology and Laboratory Medicine, and cDepartment of Human Genetics, David Geffen School of Medicine at the University of California,Los Angeles, CA 90095; and bDivision of Hematology-Oncology, and Saban Research Institute, Children's Hospital Los Angeles, Keck School of Medicine,University of Southern California, Los Angeles, CA 90027

Edited by George Klein, Karolinska Institute, Stockholm, Sweden, and approved December 8, 2009 (received for review July 13, 2009)

Ataxia-telangiectasia mutated (ATM) is a high molecular weightprotein serine/threonine kinase that plays a central role in themaintenance of genomic integrity by activating cell cycle check-points and promoting repair of DNA double-strand breaks. Little isknown about the regulatorymechanisms for ATMexpression itself.MicroRNAs are naturally existing regulators that modulate geneexpression in a sequence-specific manner. Here, we show that ahuman microRNA, miR-421, suppresses ATM expression by target-ing the 3′-untranslated region (3′UTR) of ATM transcripts. Ectopicexpression of miR-421 resulted in S-phase cell cycle checkpointchanges and an increased sensitivity to ionizing radiation, creatinga cellular phenotype similar to that of cells derived from ataxia-telangiectasia (A-T) patients. Blocking the interaction betweenmiR-421 andATM 3′UTRwith an antisensemorpholino oligonucleo-tide rescued the defective phenotype caused by miR-421 overex-pression, indicating that ATMmediates the effect ofmiR-421 on cellcycle checkpoint and radiosensitivity. Overexpression of the N-Myctranscription factor, an oncogene frequently amplified in neuro-blastoma, induced miR-421 expression, which, in turn, down-regulated ATM expression, establishing a linear signaling pathwaythat may contribute to N-Myc-induced tumorigenesis in neuroblas-toma. Taken together, our findings implicate a previously unde-scribed regulatory mechanism for ATM expression and ATM-dependent DNA damage response and provide several potentialtargets for treating neuroblastoma and perhaps A-T.

neuroblastoma | S-phase checkpoint | radiosensitivity | DNA repair

Ataxia-telangiectasia mutated (ATM) kinase plays a hier-archical regulatory role in the double-strand break (DSB)-induced DNA damage response in which ATM transduces a DSBdamage/repair signal to downstream effector machinery by phos-phorylating critical protein substrates (1–4). ATM mutations,which usually result in loss of ATM protein expression (5), lead tothe autosomal recessive progressive neurodegenerative diseaseataxia-telangiectasia (A-T) (6, 7). Both homozygotes and hetero-zygotes are at an increased risk for cancer (8). ATM has beenreported to be regulated by a transcription factor, E2F-1, (9) andtheATM gene is also reported to be subject to epigenetic silencingsuch as by methylation of the ATM promoter (10, 11), suggestingthat ATM can also be up-regulated at the transcriptional levelunder some circumstances. MicroRNAs regulate gene expressionthrough inhibition of translation or degradation of the targetedmRNA (12, 13). Physiological functions of microRNAs havebeen observed in normal and lineage-targeted development (14)as well as in the context of human cancers (15). In this study,we demonstrate that miR-421 targets the 3′-untranslated region(3′UTR) of ATM and down-regulates its expression, whereasmiR-421 expression is driven by theN-Myc transcription factor, anoncogene that is frequently amplified in neuroblastoma cells.

ResultsMiR-421 Suppresses ATM Expression by Targeting 3′UTR of ATM. Toexplore the possibility that microRNAs might regulate ATMexpression, we searched the 3′UTR of the human ATM gene for

microRNA-binding motifs using the MicroCosm Targets program(EMBL-EBI). Nine nucleotides at the 5′-end of hsa-miR-421(miR-421)were perfectly complementary to the target sequence inthe 3′UTR ofATM (including the “seed sequence” from positions2–8) (Fig. 1A). This suggested thatATMmight be a target formiR-421. To validate this in silico prediction, we cloned the ATM 3′UTR portion containing the miR-421 target site into a Renillaluciferase reporter construct (Fig. 1B) and established a luciferasereporter assay following cotransfection of reporter constructs withprecursor miR-421 (pre-miR-421) into HeLa cells. A significantreduction (30%) in the luciferase activity of the reporter constructcontaining the ATM 3′UTRwas observed in the presence of miR-421, whereas no changes were noted in the luciferase activity of theunmodified construct (pRL) with miR-421 expression (Fig. 1C).Deletion of six nucleotides of seed sequence (Δ6) led to the loss ofreduction in miR-421-mediated luciferase activity (Fig. 1C). Tofurther confirm thatATMis a target formiR-421, we examined theendogenous ATM protein level by immunoblot after transientlytransfecting pre-miR-421 intoHeLa cells. As shown in Fig. 1D, theATM expression level decreased as the concentration of trans-fected pre-miR-421 was increased. As an indication of ATM kin-ase activity (16), phosphorylation of SMC1 at the serine-966residue (pS966-SMC1) was measured following DNA damage by10-Gy irradiation (IR).A significant reduction in the pS966-SMC1was observed when pre-miR-421 was introduced into HeLa cellsfollowed by IR, as compared with the introduction of a non-relevant control pre-miR precursor (Fig. 1E). ATMmRNA levelswere measured by quantitative real-time PCR and were notdecreased in the presence of miR-421 (Fig. 1F), suggesting thatmiR-421 down-regulates ATM at a translational rather thantranscriptional level.

MiR-421 Regulates Cell Cycle S-Phase Checkpoint and CellularRadiosensitivity. To determine the cellular functions of miR-421,we created an miR421-overexpressing HeLa stable cell line byinfecting the cells with an miR421-containing lentivirus andselecting a stable infectant with blasticidin (HeLa/miR-421) (Fig.2A). We also created a control stable infectant cell line withscrambled shRNA (HeLa/scram) (17). Real-time PCR detectedan ∼120-fold increase in the expression of mature miR-421 in theHeLa/miR-421 cells compared with the HeLa/scram control cells(Fig. 2B). Both ATMprotein expression andATM kinase activity,as indicated by the level of post-IR pS966-SMC1, were sig-nificantly reduced in the HeLa/miR-421 cells (Fig. 2C).

Author contributions: H.H. and R.A.G. designed research; H.H., L.D., and G.N. performedresearch; R.C.S. contributed new reagents/analytic tools; H.H. and R.A.G. analyzed data;and H.H., R.C.S., and R.A.G. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence may be addressed at: Department of Pathology and Labora-tory Medicine, David Geffen School of Medicine at UCLA, 675 Charles Young Drive, LosAngeles, CA 90095. E-mail: hhu@mednet.ucla.edu or rgatti@mednet.ucla.edu.

This article contains supporting information online at www.pnas.org/cgi/content/full/0907763107/DCSupplemental.

1506–1511 | PNAS | January 26, 2010 | vol. 107 | no. 4 www.pnas.org/cgi/doi/10.1073/pnas.0907763107

Dow

nloa

ded

by g

uest

on

July

4, 2

021

mailto:hhu@mednet.ucla.edumailto:rgatti@mednet.ucla.eduhttp://www.pnas.org/cgi/content/full/0907763107/DCSupplementalhttp://www.pnas.org/cgi/content/full/0907763107/DCSupplementalwww.pnas.org/cgi/doi/10.1073/pnas.0907763107

ATM regulates DNA damage-induced cell cycle checkpoints atG1-S and intra-S phase (18, 19). A hallmark of A-T cells is thefailure to arrest DNA synthesis in S-phase following DNA dam-age and continuous incorporation of nucleotides into DNAdespite damage (i.e., radioresistant DNA synthesis) (20, 21).Thus, we anticipated that miR-421 overexpression might regulateDNA damage-induced cell cycle S-phase checkpoints. To assessthis, HeLa/scram and HeLa/miR-421 cells were irradiated (10Gy) to introduce DNA damage and BrdU was used to followDNA incorporation. As expected, a reduction in the percentageof BrdU-positive cells in S-phase was observed for the HeLa/scram control cells (14.88% pre-IR vs. 12.56% post-IR), indi-cating a normal block in the DNA synthesis (Fig. 2D, Upper Leftand Right); in contrast, an increase in the percentage of BrdU-positive cells in S-phase was observed for HeLa/miR-421 cells(11.67% pre-IR vs. 15.38% post-IR) (Fig. 2D, Lower Left andRight), indicating that miR-421 overexpression overcomes theIR-induced DNA synthesis block and mimics the radioresistantDNA synthesis of A-T cells. The miR421-induced continuous

DNA synthesis was also seen with lower doses of IR at 2 and 5 Gy(Fig. S1A). We noticed that the pre-IR percentage of HeLa/miR-421 cells in S-phase was lower than that of control HeLa/scramcells, suggesting that miR-421 might regulate this cell cyclecheckpoint independent of DNA damage. Similar results wereobserved using a human breast cancer cell line, MDA-MB-231,when miR-421 was overexpressed (Fig. S2 A and B).A clonogenic assay was used to determine whether over-

expression ofmiR-421 affects cellular radiosensitivity. As shown in

Fig. 1. miR-421 suppresses ATM expression by targeting ATM 3′UTR. (A)Mature miR-421 sequences and recognition sites within 3′UTR of ATM. Theseed sequence of miR-421 is shown in the box. WT and del6 (Δ6) ATM-3′UTRtargets are also shown. (B) Constructs of Renilla luciferase [Luc; unmodifiedconstruct (pRL)-CMV] containing WT or 6-nt deleted (Δ6) ATM 3′UTR. (C)Luciferase (Luc.) activity of pRL and modified constructs containing WT ormutant (Δ6) 3′UTR. Luciferase constructs were cotransfected with pre-miR-CTL (control, 50 nM) or pre-miR-421 (50 nM) into HeLa cells. Renilla luciferaseactivity was measured 36 h after incubation and normalized to firefly luci-ferase. Asterisk indicates significant down-regulation of pre-miR-421 againstconstruct containing WT ATM 3′UTR. (D) Immunoblot of endogenous ATMexpression in HeLa cells 96 h after transfection of increasing amounts of pre-miR-421 (using pre-miR-CTL to compensate for equal amounts of total miRs).SMC1 served as a loading control for the blot. (E) Immunoblot of pS966-SMC1 in HeLa cells that were transiently transfected with pre-miR-CTL (100nM) or pre-miR-421 (100 nM) after 10-Gy IR to activate the DNA damageresponse. A WT lymphoblastoid cell line (LCL) served as a positive control forSMC1 and ATM protein. (F) Real-time PCR of ATM mRNA from HeLa cellstransfected with pre-miR-CTL (100 nM) or pre-miR-421 (100n M). Data werenormalized to the level of GAPDH mRNA, and the ratio of ATM/GAPDH inHeLa control cells was set to 1.

Fig. 2. miR-421 regulates cell cycle S-phase checkpoint and cellular radio-sensitivity. (A) Scheme of a U6 promoter-driven miR-421 cloned into a len-tiviral vector with two LTRs and a selection marker for blasticidin driven bySV40 promoter. (B) Real-time PCR of miR-421 expression in HeLa cells stablyoverexpressing scrambled shRNA (HeLa/scram) or miR-421(HeLa/miR-421).Data were normalized to an internal control RNU66, and the ratio of miR-421/RNU66 in HeLa/scram cells was set to 1. (C) Immunoblot of ATM andpSMC1 in HeLa/scram and HeLa/miR-421 cells with or without 10-Gy IR. Notethe reduction of both ATM and pSMC1 in miR421-overexpressing cells. (D)Analysis of IR-induced cell cycle S-phase checkpoint by FC. Stably over-expressing HeLa/scram and HeLa/miR-421 cells were treated with or without10 Gy. DNA synthesis at S-phase was labeled with BrdU. (Left Upper andLower) Results of one experiment representative of three independentexperiments. Box R5 indicates the percentage of BrdU+ S-phase cells pre- orpost-IR. (Right) Summary of change of BrdU+ cells pre- and post-IR for HeLa/scram and HeLa/miR-421 cells from three independent experiments, usingthe algorithm (R5+IR−R5−IR)/R5−IR × 100%. (E) HeLa/scram and HeLa/miR-421cells were irradiated at the indicated doses, and colony survival was meas-ured after 2 weeks. (F) Effect of miR-421 on proliferation of HeLa cells, asmeasured by cell population doublings with culture time. (G) Effect ofmiR-421 on IR-induced cell death, as measured by propidium iodide stainingFC. The percentage of propidium iodide-positive cells was normalized to theunirradiated cells in each group.

Hu et al. PNAS | January 26, 2010 | vol. 107 | no. 4 | 1507

GEN

ETICS

Dow

nloa

ded

by g

uest

on

July

4, 2

021

http://www.pnas.org/cgi/data/0907763107/DCSupplemental/Supplemental_PDF#nameddest=sfig01http://www.pnas.org/cgi/data/0907763107/DCSupplemental/Supplemental_PDF#nameddest=sfig02

Fig. 2E, the survival fractions of HeLa/miR-421 cells post-IR (1and 2 Gy) were significantly reduced relative to those of HeLa/scram control cells. MiR-421 overexpression did not alter theproliferation rate ofHeLa cells (Fig. 2F) but increased post-IR celldeath (Fig. 2G), which is consistent with the decreased survivalfraction in the clonogenic assay. A similar effect of miR-421 onradiosensitivity was observed withMDA-MB-231 cells (Fig. S2C).

Effects of miR-421 on S-Phase Checkpoint and Radiosensitivity AreATM Dependent. A singlemicroRNA is predicted tomodulate>200targets of protein expression (22).To further determinewhether theeffects ofmiR-421 on cell cycle checkpoints and radiosensitivity aremediated through ATM, we used an antisense morpholino oligo-nucleotide (AMO) to block the recognition sequence of ATM 3′UTR (Fig. 3A). Treatment of HeLa/miR-421 cells with ATM 3′UTR target site-specific AMO (AMO-ATM) resulted in the abro-gationofmiR421-mediateddown-regulationofATMexpression, asshown by both Western blot and ELISA (Fig. 3B). This effect wasnot observed when cells were treated with scrambled control AMO(AMO-scram) (Fig. S1B). Following IR, AMO-ATM treatment

also resulted in an increase of pS966-SMC1 in HeLa/miR-421 cells(Fig. 3C). Blocking with AMO-ATM further restored the S-phasecell cycle checkpoint and radiosensitivity of theHeLa/miR-421cells,as shown by radioresistant DNA synthesis assay and clonogenicsurvival assay (Fig. 3 D and E). Taken together, these AMO-ATMexperiments suggest that the effect of miR-421 on the cell cycle S-phase checkpoint and radiosensitivity is mediated through ATM.

Transcription Factor N-Myc Up-Regulates miR-421 Expression.HumanmiR-421 is located intergenically at chromosome Xq13. Inter-estingly, another microRNA, miR-374b, is located just 85 bpproximal to miR-421, forming a microRNA cluster that is drivenby a single promoter (Fig. 4A). The function of miR-374b is stillunknown. To determine which transcription factors might influ-ence miR-421 expression, we performed in silico analysis of the

Fig. 3. ATMmediates the effect of miR-421 on cell cycle S-phase checkpointand radiosensitivity. (A) Schematic working model of ATM 3′UTR thatwas targeted by an antisense AMO. AMO-ATM was designed to match themiR-421 recognition site of ATM 3′UTR and specifically block the down-regulation of ATM by impeding the binding of mature miR-421. (B) (Left)Immunoblot of ATM expression in HeLa/scram and HeLa/miR-421 cells trea-ted with or without AMO-ATM (2 μM) for 5 days. The fold change in ATMexpression is shown below the immunoblot. (Right) ELISA was also used todetermine ATM concentration. (C) Immunoblot of pSMC1 in HeLa/scram andHeLa/miR-421 cells treated with AMO-scram (2 μM) or AMO-ATM (2 μM) for 5days, followed by 10-Gy IR. The fold change in pSMC1 level is shown belowthe immunoblot. Note the increase of pSMC1 in HeLa/miR-421 cells treatedwith AMO-ATM. (D) Analysis of cell cycle S-phase checkpoint after treatmentof AMO. HeLa/scram and HeLa/miR-421 cells were treated with AMO-scram(2 μM) or AMO-ATM (2 μM) for 5 days and irradiated with increasing doses ofradiation (2, 5, and 10 Gy). DNA synthesis was monitored by BrdU incorpo-ration and analyzed by FC. The percentage of BrdU+ S-phase cells at the startpoint (unirradiated) was arbitrarily set to 50%, and all other data werenormalized to this point. This plot is representative of three independentexperiments. The arrow indicates that AMO-ATM treatment rescues thedefect of HeLa/miR-421 cells. (E) Colony survival fraction with exposureto AMO. HeLa/scram and HeLa/miR-421 cells were treated with AMO-scram(2 μM) or AMO-ATM (2 μM) for 5 days, and 500 cells were plated in triplicate;cells were irradiated with increasing doses of radiation, and surviving colo-nies were scored after 2 weeks. The survival fraction at each radiation dosewas normalized to that of the nonirradiated control. The arrow indicatesthat the AMO rescued the radiosensitivity of HeLa/miR-421 cells.

Fig. 4. miR-421 is up-regulated by N-Myc overexpression in HeLa cells. (A)Chromosomal location of miR-374b/miR-421 cluster on chromosome Xq13,sharing the same promoter. The promoter region (1 kb), containing an E-box(5′-CACGTG-3′), was cloned into luciferase construct pGL3-basic to createpGL3-PR421 and drives the transcription of firefly luciferase (Luc). (B) Luci-ferase activity of the miR-421 promoter. Luciferase constructs [pGL3-PR421and unmodified construct (pRL)-CMV] were cotransfected, with vector(Vec) or N-Myc, into HeLa cells. Firefly luciferase activity was measured 24 hand 48 h after incubation and normalized to Renilla luciferase activity. (C)Real-time PCR of endogenous miR-421expression in HeLa cells transientlytransfected with vector or N-Myc. Data were normalized to RNU66. (D)Immunoblot of ATM expression in HeLa cells transiently transfected withincreasing amounts of N-Myc (using vector to compensate for equal amountsof total DNA). The fold change in ATM protein expression is shown belowthe blot. (E) ELISA measurement of ATM concentrations in HeLa cells tran-siently transfected with N-Myc. The asterisk indicates significant inhibition ofATM by N-Myc overexpression. (F) ELISA to determine ATM concentration inHeLa cells transiently transfected with the indicated DNA constructs (vectoror N-Myc) and anti-miR-CTL (50 nM) or anti-miR-421(50 nM) inhibitors. (G)Immunoblot of ATM expression in HeLa cells transiently transfected withindicated DNA constructs (vector or N-Myc) and anti-miR-CTL (50 nM) or anti-miR-421(50 nM) inhibitor. Only the top band corresponds to ATM.

1508 | www.pnas.org/cgi/doi/10.1073/pnas.0907763107 Hu et al.

Dow

nloa

ded

by g

uest

on

July

4, 2

021

http://www.pnas.org/cgi/data/0907763107/DCSupplemental/Supplemental_PDF#nameddest=sfig02http://www.pnas.org/cgi/data/0907763107/DCSupplemental/Supplemental_PDF#nameddest=sfig01www.pnas.org/cgi/doi/10.1073/pnas.0907763107

promoter region (including 2 kb upstream of the miR-374b stemloop) using the transcription factor binding site program CON-SITE (Materials and Methods). This identified a binding site forN-Myc (an E-box) at −85 nucleotides relative to the miR-374bstem loop (Fig. 4A). To validate this prediction, we cloned a 1-kbDNA fragment of the promoter region into a firefly luciferasereporter construct and examined the effect of N-Myc on miR-421promoter-driven luciferase activity. Overexpression of N-Myc inHeLa cells activated miR-421 promoter-driven luciferase activity24 and 48 h after transfection (Fig. 4B). Consistent with theluciferase assay, the expression level of endogenous mature miR-421 was also increased by overexpression of N-Myc in the HeLacells, as measured by microRNA real-time PCR (Fig. 4C). Mostinterestingly, ATM protein expression was reduced in HeLa cellsthat were transiently transfected with N-Myc, as detected byimmunoblotting and ELISA (Fig. 4 D and E), strongly suggestingthat N-Myc stimulates miR-421 expression, which, in turn, down-regulates ATMexpression. Anti-miR-421 inhibitor is a chemicallymodified antisense oligonucleotide designed specifically to bindto and inhibit endogenous miR-421 molecule. Cotransfection ofanti-miR-421 inhibitor into HeLa cells along with N-Myc con-struct restored the ATM expression that was suppressed inN-Myc-transfected cells, further confirming that miR-421 medi-ates N-Myc-induced ATM down-regulation (Fig. 4 F and G). Wealso noticed that anti-miR-CTL relieved the ATM expression tosome extent, which might be caused by the nonspecific binding ofanti-miR-CTL to the endogenous miR-421 (Fig. 4 F and G).

N-Myc/miR-421/ATM Pathway in Neuroblastoma Cells. The N-Mycgene is frequently amplified in human neuroblastoma cells andis used as a prognostic marker for neuroblastoma (23, 24). Tofurther explore the N-Myc/ATM relation, we examined ATMexpression in seven human neuroblastoma cell lines: Four celllines (CHLA-134, CHLA-136, LA-N-1, and LA-N-5) are N-Mycamplified, whereas the other three (CHLA-15, CHLA-90, andCHLA-255) are not N-Myc amplified. We noted a low level of N-Myc expression in CHLA-90 cell lines compared with the othertwo cell lines, CHLA-15 and CHLA-255, with undetectable N-Myc expression (Fig. 5A). We found that the ATM expressionlevels were significantly lower in the four N-Myc-amplified celllines compared with those of three N-Myc-nonamplified cell lines(Fig. 5A), suggesting that N-Myc might negatively regulate ATMexpression through miR-421. To confirm the interaction of N-Myc, miR-421, and ATM in neuroblastoma cells, we selected LA-N-1 (N-Myc+) and CHLA-255(N-Myc−) for the followingexperiments. i) Chromatin immunoprecipitation (ChIP) showedthat the in vivo binding of N-Myc to miR-421 promoter onlyoccurred in the N-Myc-amplified LA-N-1 cells and not in the N-Myc-nonamplified CHLA-255 cells (Fig. 5B). ii) Consistent withthe in vivo binding of N-Myc, endogenous miR-421 expressionwas up-regulated (∼2-fold) in LA-N-1 cells (Fig. 5C). We alsoexaminedmiR-421 expression in the other five neuroblastoma celllines. The miR-421 levels were significantly higher in the four N-Myc-amplified cell lines (Fig. S3A). We noticed that the miR-421level in CHLA-90 was higher than that in the other two N-Myc-nonamplified cell lines CHLA-15 and CHLA-255 (Fig. S3A). Thismight be caused by the low-level expression of N-Myc in CHLA-90 (Fig. 5A). A similar expression pattern for miR-374b wasobserved in these neuroblastoma cell lines (Fig. S3B), supportinga model that the miR-421 and miR-374b cluster is driven by thesame promoter (Fig. 4A). iii) Treatment of LA-N-1 cells withAMO-ATM, which is complementary to the miR-421 bindingsites at ATM 3′UTR, and with anti-miR-421 inhibitor, which iscomplementary to miR-421, led to an increase in ATM expression(Fig. 5D). As expected, AMO-ATM treatment did not change themiR-421 expression level, whereas anti-miR-421 inhibitor down-regulated miR-421 expression (Fig. 5E), suggesting two differentmechanisms for AMO-ATM and anti-miR-421 on the abrogation

of miR-421-mediated down-regulation of ATM expression.Finally, the increase of ATM expression by AMO-ATM wasfurther confirmed by the flow cytometry phospho-SMC1 (FC-pSMC1) assay, which was recently developed to measure ATM

Fig. 5. N-Myc negatively regulates ATM via miR-421 in neuroblastoma cells.(A) Immunoblot of ATM expression in N-Myc-amplified (CHLA-134, CHLA136,LA-N-5, and LA-N-1) or -nonamplified (CHLA-15, CHLA-90, and CHLA-255)neuroblastoma cell lines. We observed some N-Myc expression in CHLA-90,although it is an N-Myc-nonamplified cell line; ATM expression was relativelylower in this cell line when compared with CHLA-15 or CHLA-255. A-T lym-phoblastoid cells (AT-LCL) and WT lymphoblastoid cells (WT-LCL) are neg-ative and positive controls, respectively, for ATM expression. In total, 100 μgof total protein for all neuroblastoma cells and only 25 μg of total proteinfor AT-LCL and WT-LCL were loaded; SMC1 served as a loading control.(B) ChIP PCR assay detects the in vivo binding of N-Myc protein to the miR-421 promoter DNA. A PCR fragment of expected size (246 bp) was seenin the N-Myc-amplified (amp.) LA-N-1 cells immunoprecipitated with thespecific anti-N-Myc antibody (lane 5) but not without antibody or withnonspecific mouse IgG (lanes 2 and 3). No signal was seen in the N-Myc-nonamplified CHLA-255 cells immunoprecipitated with no antibody, non-specific mouse IgG, or specific anti-Myc antibody (lanes 7–9). PCR with inputDNA was used as a positive control. (C) Real-time PCR of endogenous miR-421 in the N-Myc-amplified LA-N-1 cells and the N-Myc-nonamplified CHLA-255 cells. RNU66 was used as an internal control. (D) Immunoblot of ATMexpression in CHLA-255 and LA-N-1 cells treated with AMO-scram (4 μM) orAMO-ATM (4 μM) for 5 days or in L-AN-1 cells transfected with anti-miR-CTL(100 nM) or anti-miR-421 inhibitor (100 nM) for 96 h. The fold change inATM expression is shown below. (E) Real-time PCR of miR-421 expression inLA-N-1 cells treated with AMO-scram (4 μM) or AMO-ATM (4 μM) for 5 daysor transfected with anti-miR-CTL (100 nM) or anti-miR-421 (100 nM) inhibitorfor 4 days. RNU66 was used as an internal control. (F) FC-SMC1 detection ofIR-induced ATM-dependent phosphorylation of SMC1 in AMO-treated neu-roblastoma cells. LA-N-1 and CHLA-255 cells were treated with AMO-scram(4 μM) or AMO-ATM (4 μM) for 5 days and subjected to 10-Gy IR. The pSMC1level is indicated by the fluorescence intensity. The filled peaks represent thecells without IR, and unfilled peaks represent post-IR cells. This panel isrepresentative of three independent experiments. (G) Linear signalingpathway in which N-Myc up-regulates miR-421 expression and miR-421, inturn, down-regulates ATM expression by targeting its 3′UTR.

Hu et al. PNAS | January 26, 2010 | vol. 107 | no. 4 | 1509

GEN

ETICS

Dow

nloa

ded

by g

uest

on

July

4, 2

021

http://www.pnas.org/cgi/data/0907763107/DCSupplemental/Supplemental_PDF#nameddest=sfig03http://www.pnas.org/cgi/data/0907763107/DCSupplemental/Supplemental_PDF#nameddest=sfig03http://www.pnas.org/cgi/data/0907763107/DCSupplemental/Supplemental_PDF#nameddest=sfig03

kinase activity in A-T patients and carriers (25). As shown in Fig.5F, treatment of LA-N-1 cells with AMO-ATM caused a post-IRshift in pSMC1, whereas treatment of CHLA-255 cells showed nochange in the post-IR pSMC1 shift. These observations arecompatible with the model that AMO-ATM could increase ATMexpression in N-Myc-amplified neuroblastoma cells and are alsoconsistent with the immunoblot results of ATM expression asshown in Fig. 5D.Because c-Myc shares a conserved E-box binding site (5′-CA-

CGTG-3′) with N-Myc (Fig. 4A), we were prompted to determinewhether c-Myc functions in a manner similar to N-Myc in up-regulating miR-421 expression. As shown in Fig. S4A, cotrans-fection of c-Myc with the miR-421 promoter construct into HeLacells resulted in a significant increase in miR-421 promoter-drivenluciferase activity, as did cotransfection of N-Myc. EndogenousmiR-421 expression in HeLa cells was similarly increased ∼1.5-fold after transfection of c-Myc (Fig. S4B).

DiscussionTaken together, our experiments suggest a previously undescribedmechanism of ATM regulation in which a noncoding small RNA,miR-421, down-regulates ATM expression through targetingATM 3′UTR. This substantially expands our understanding ofATM functions in cellular physiology, such as cell cycle check-point, radiosensitivity, and other ATM-mediated cellular func-tions. For example, microRNA profiling study has revealed thatmiR-421 is up-regulated in germinal center centroblast B cells(26), where physiological DNAdamage occurs frequently becauseof somatic hypermutation and class switch recombination (27).The miR421-mediated ATM down-regulation in centroblastsmight contribute to the escape of centroblast B cells from DNAdamage-induced cell cycle checkpoints and allow centroblasts todevelop into memory B cells or plasma cells. A recent reportcorroborates this concept in whichATR (ATMandRad3-related)kinase is transiently silenced by a transcription repressor Bcl-6 ingerminal center B cells (28). Interestingly, miR-421 expression isalso up-regulated in diffuse large B-cell lymphoma cell lines (29),suggesting that this newly identified miR421–ATM interactionmight be involved in the progression of diffuse large B-cell lym-phoma. It is known that about 10% of cases have overexpressionof c-Myc, a result of the c-myc translocation into the Ig locus (27).We have established that miR-421 expression is up-regulated by

the transcription factor N-Myc, establishing a linear signalingpathway (N-Myc → miR-421 → ATM) in such a manner that theoncogene N-Myc negatively regulates the tumor suppressor ATM(Fig. 5G). Because the ATM-driven DNA damage response isthought to be a physiological barrier in early human tumorigenesis(30–33), our findings add that miR421-mediated ATM down-reg-ulation may contribute to N-Myc-induced tumorigenesis in neuro-blastoma. The finding that the up-regulation of miR-421 can alterthe cellular radiosensitivity suggests that treatment of proliferatingcancer cells with miR-421-inducing agents might sensitize them forradiotherapy. Conversely the finding that exposure of neuro-blastoma cells to AMO-ATM increases ATM expression impliesthatAMO-ATMholds therapeutical potential forN-Myc-amplifiedneuroblastomas, perhaps by enhancing ATM-dependent apoptosisin response to DNA damage (34, 35) or driving nondividing dif-ferentiated neuronal cells to reenter S-phase (36). Lastly, the sup-pression of ATM bymiR-421 introduces two possible pathogeneticmechanisms forA-T:Amutation in theATM3′UTRmight enhancethe binding of miR-421, or a mutation of miR-421 might result inmiR-421 overexpression, both leading to the down-regulation ofATM expression. Such disease-causing mutations of microRNA-binding sites in the 3′UTR of the target genes have been reported(37). However, no suchmutations have been observed to date in A-T patients. Our findings also suggest thatmiR-421 could function asa modifier gene, contributing to the A-T phenotype and perhaps tothe variability of disease onset and progression.

Materials and MethodsCell Culture, miRNA Precursors, miRNA Inhibitors, AMO and Transfection.Neuroblastoma cell lines LA-N-1 and LA-N-5 were cultured in RPMI 1640with 15% (vol/vol) FBS and streptomycin/penicillin, and CHLA-15, CHLA-90,CHLA-134, CHLA-136, and CHLA-255 were cultured in Iscove’s Modified Dul-becco’s Medium with 15% (vol/vol) FBS and streptomycin/penicillin. Theprecursor miR-421, pre-miR-CTL, anti-miR negative control 1, and anti-miR-421 inhibitor were purchased from Applied Biosystems. Antisense AMO wassynthesized based on the ATM 3′UTR target sequence and conjugated withnonpeptide chemicals that are used to deliver AMO to cells (Gene-Tools). Thesequence of AMO-ATM is 5′-ATCAACAGATATAAACAGCAGG. A standardcontrol AMO (AMO-scram)was also purchased fromGene-Tools. N-Myc and c-Myc plasmidswere obtained fromOrigene andOpen Biosystems, respectively.All transfectionsweredonewith Lipofectamine2000 (Invitrogen) according tothe provided protocols. The M4 lentiviral vector expressing miR-421 wasgenerated by standard methods as detailed in SI Text.

RNA Extraction and Real-Time Quantitative PCR. Total RNA from cultured cellswas extracted by the mirVana miRNA isolation kit (Applied Biosystems). Taq-Man microRNA expression assays (Applied Biosystems) were usedto quantitate mature miR-421 expression according to the provided protocol.RNU66 or U6 expression assay was used as an internal control for miR-421expression.ATMmRNAquantificationweremeasured by real-time PCR basedon TaqMan Gene Expression Assays (Applied Biosystems), as previouslydescribed (38). Glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) mRNAwas used as an internal control to normalize ATMmRNA level. Real-time PCRquantitation of transcripts was expressed as ATM/GAPDH ratios.

Luciferase Reporter Assays and Transcription Factor Analysis. Cells were trans-fected with appropriate reporter vectors and then harvested and lysed forluciferase assays by a Dual-Glo assay kit (Promega) according to the manu-facturer’s protocol. The detailed protocol is described in SI Text.

The transcription factor binding site analysis was done using the CONSITEdatabase. A 2-kb genomic sequence upstream of miR-421 was used as theanalyzing template, and the cutoff value for transcription factors was set to99%; N-Myc transcription factor was the top candidate.

ATM-ELISA. An ELISA was used to determine the relative ATM expression incells and performed as previously described (39). More details on the ATM-ELISA are provided in SI Text.

BrdU Incorporation Assay. To analyze the S-phase checkpoint, cells takenat 70% confluence were irradiated with the indicated dose and incubated for20 h. BrdU was added to cells and incubated for 2 h. Cells were collected bytrypsinization and centrifugation. Cells were subject to the BrdUflow stainingaccording to themanufacturer’s protocol for BrdU FlowKits (BD Pharmingen).Three independent experiments were performed.

Clonogenic Survival Assay and Propidium Iodide-Staining Cell Death. MiR-421-expressing stable and control shRNA cells were plated at 500 cells per wellonto a six-well dish in triplicate and then incubated for 24 h to allow settling.Cells were treated with a series of IR doses (0, 1, 2, and 5 Gy) and grown for 2weeks before stainingwith 1% crystal violet. Clumps containingmore than 50cells were scored as “colony-positive”wells and counted by the Quantify Oneprogram in the VersaDoc Imaging System (Bio-Rad). To generate a radiationsurvival curve, the surviving fraction at each radiation dose was normalizedto that of a nonirradiated control. All experiments were repeated at leasttwice. For IR-induced cell death, cells were treated with 10-Gy radiation andstained with propidium iodide after 48 h of incubation to assess the numberof dead cells. Samples were analyzed by a FACScan Analytic Flow Cytometer(Becton Dickinson). Three independent experiments were done. Cell deathwas normalized to nonirradiated control cells.

FC-pSMC1 Assay and ChIP Assay. An FC-pSMC1 assay was performed as pre-viously described (25), and a ChIP assay was performed as previouslydescribed (40). More detailed protocols for Fc-pSMC1 and ChIP assays areprovided in SI Text.

Statistics. The Student’s t -test was used to evaluate the significant differenceof two groups of data in all the pertinent experiments. A P value

ACKNOWLEDGMENTS. We thank Dr. John Colicelli for M4 lentiviral vector andDr. Matteo Pellegrini and Aliz Raksi for microRNA target predictions andanalyses. This work was supported by Grant NS052528 from the National

Institutes of Health, the Ataxia-Telangiectasia Medical Research Foundation(Los Angeles, CA), and the Ataxia-Telangiectasia Ease Foundation (NewYork, NY).

1. Shiloh Y (2003) ATM and related protein kinases: Safeguarding genome integrity. NatRev Cancer 3:155–168.

2. Lavin MF (2008) Ataxia-telangiectasia: From a rare disorder to a paradigm for cellsignalling and cancer. Nat Rev Mol Cell Biol 9:759–769.

3. Matsuoka S, et al. (2007) ATM and ATR substrate analysis reveals extensive proteinnetworks responsive to DNA damage. Science 316:1160–1166.

4. Bakkenist CJ, Kastan MB (2003) DNA damage activates ATM through intermolecularautophosphorylation and dimer dissociation. Nature 421:499–506.

5. Concannon P, Gatti RA (1997) Diversity of ATM gene mutations detected in patientswith ataxia-telangiectasia. Hum Mutat 10:100–107.

6. Gatti RA, et al. (1988) Localization of an ataxia-telangiectasia gene to chromosome11q22-23. Nature 336:577–580.

7. Savitsky K, et al. (1995) A single ataxia telangiectasia gene with a product similar toPI-3 kinase. Science 268:1749–1753.

8. Swift M, Reitnauer PJ, Morrell D, Chase CL (1987) Breast and other cancers in familieswith ataxia-telangiectasia. N Engl J Med 316:1289–1294.

9. Berkovich E, Ginsberg D (2003) ATM is a target for positive regulation by E2F-1.Oncogene 22:161–167.

10. Roy K, Wang L, Makrigiorgos GM, Price BD (2006) Methylation of the ATM promoterin glioma cells alters ionizing radiation sensitivity. Biochem Biophys Res Commun 344:821–826.

11. Kim WJ, Vo QN, Shrivastav M, Lataxes TA, Brown KD (2002) Aberrant methylation ofthe ATM promoter correlates with increased radiosensitivity in a human colorectaltumor cell line. Oncogene 21:3864–3871.

12. He L, Hannon GJ (2004) MicroRNAs: Small RNAs with a big role in gene regulation.Nat Rev Genet 5:522–531.

13. Bartel DP (2009) MicroRNAs: Target recognition and regulatory functions. Cell 136:215–233.

14. Stefani G, Slack FJ (2008) Small non-coding RNAs in animal development. Nat Rev MolCell Biol 9:219–230.

15. Calin GA, Croce CM (2006) MicroRNA-cancer connection: The beginning of a new tale.Cancer Res 66:7390–7394.

16. Kitagawa R, Bakkenist CJ, McKinnon PJ, Kastan MB (2004) Phosphorylation of SMC1 isa critical downstream event in the ATM-NBS1-BRCA1 pathway. Genes Dev 18:1423–1438.

17. Hu H, et al. (2008) Integration of transforming growth factor beta and RAS signalingsilences a RAB5 guanine nucleotide exchange factor and enhances growth factor-directed cell migration. Mol Cell Biol 28:1573–1583.

18. Zhou BB, Elledge SJ (2000) The DNA damage response: Putting checkpoints inperspective. Nature 408:433–439.

19. Kastan MB, Bartek J (2004) Cell-cycle checkpoints and cancer. Nature 432:316–323.20. Houldsworth J, Lavin MF (1980) Effect of ionizing radiation on DNA synthesis in ataxia

telangiectasia cells. Nucleic Acids Res 8:3709–3720.

21. Painter RB (1981) Radioresistant DNA synthesis: An intrinsic feature of ataxiatelangiectasia. Mutat Res 84:183–190.

22. Krek A, et al. (2005) Combinatorial microRNA target predictions. Nat Genet 37:495–500.

23. Seeger RC, et al. (1985) Association of multiple copies of the N-myc oncogene withrapid progression of neuroblastomas. N Engl J Med 313:1111–1116.

24. Brodeur GM (2003) Neuroblastoma: Biological insights into a clinical enigma. Nat RevCancer 3:203–216.

25. Nahas SA, Butch AW, Du L, Gatti RA (2009) Rapid flow cytometry-based structuralmaintenance of chromosomes 1 (SMC1) phosphorylation assay for identification ofataxia-telangiectasia homozygotes and heterozygotes. Clin Chem 55:463–472.

26. Basso K, et al. (2009) Identification of the human mature B cell miRNome. Immunity30:744–752.

27. Klein U, Dalla-Favera R (2008) Germinal centres: Role in B-cell physiology andmalignancy. Nat Rev Immunol 8:22–33.

28. Ranuncolo SM, et al. (2007) Bcl-6 mediates the germinal center B cell phenotype andlymphomagenesis through transcriptional repression of the DNA-damage sensor ATR.Nat Immun 8:705–714.

29. Lawrie CH, et al. (2008) MicroRNA expression in lymphocyte development andmalignancy. Leukemia 22:1440–1446.

30. Bartkova J, et al. (2006) Oncogene-induced senescence is part of the tumorigenesisbarrier imposed by DNA damage checkpoints. Nature 444:633–637.

31. Bartkova J, et al. (2005) DNA damage response as a candidate anti-cancer barrier inearly human tumorigenesis. Nature 434:864–870.

32. Di Micco R, et al. (2006) Oncogene-induced senescence is a DNA damage responsetriggered by DNA hyper-replication. Nature 444:638–642.

33. Gorgoulis VG, et al. (2005) Activation of the DNA damage checkpoint and genomicinstability in human precancerous lesions. Nature 434:907–913.

34. Herzog KH, Chong MJ, Kapsetaki M, Morgan JI, McKinnon PJ (1998) Requirement forAtm in ionizing radiation-induced cell death in the developing central nervoussystem. Science 280:1089–1091.

35. Tian B, Yang Q, Mao Z (2009) Phosphorylation of ATM by Cdk5 mediates DNAdamage signalling and regulates neuronal death. Nat Cell Biol 11:211–218.

36. Herrup K, Yang Y (2007) Cell cycle regulation in the postmitotic neuron: Oxymoron ornew biology? Nat Rev Neurosci 8:368–378.

37. Sethupathy P, Collins FS (2008) MicroRNA target site polymorphisms and humandisease. Trends Genet 24:489–497.

38. Mitui M, et al. (2009) Functional and computational assessment of missense variantsin the ataxia-telangiectasia mutated (ATM) gene: Mutations with increased cancerrisk. Hum Mutat 30:12–21.

39. Du L, et al. (2009) Nonaminoglycoside compounds induce readthrough of nonsensemutations. J Exp Med 206:2285–2297.

40. Beierle EA, et al. (2007) N-MYC regulates focal adhesion kinase expression in humanneuroblastoma. J Biol Chem 282:12503–12516.

Hu et al. PNAS | January 26, 2010 | vol. 107 | no. 4 | 1511

GEN

ETICS

Dow

nloa

ded

by g

uest

on

July

4, 2

021