Ampka2 Regulates Bladder Cancer Growth Through SKP2-Mediated Degradation of P27 Stavros Kopsiaftis1,2, Katie L

Published OnlineFirst September 16, 2016; DOI: 10.1158/1541-7786.MCR-16-0111

Cell Cycle and Senescence Molecular Cancer Research AMPKa2 Regulates Bladder Cancer Growth through SKP2-Mediated Degradation of p27 Stavros Kopsiaftis1,2, Katie L. Sullivan1,2, Isha Garg1,2, John A. Taylor III3, and Kevin P. Claffey1,2,4

Abstract

AMP-activated protein kinase (AMPK) is the central metabolic proliferation and decreased p27 protein. The reduced p27 protein regulator of the cell and controls energy consumption based upon was determined to be dependent upon SKP2. Additionally, loss of nutrient availability. Due to its role in energy regulation, AMPK AMPKa2 in a xenograft and a chemical carcinogen model of has been implicated as a barrier for cancer progression and is bladder cancer resulted in larger tumors with less p27 protein suppressed in multiple cancers. To examine whether AMPK and high SKP2 levels. Consistent with the regulation observed regulates bladder cancer cell growth, HTB2 and HT1376 bladder in the bladder cancer model systems, a comprehensive survey cells were treated with an AMPK activator, 5-aminoimidazole-4- of human primary bladder cancer clinical specimens revealed carboxamide ribonucleotide (AICAR). AICAR treatment reduced low levels of AMPKa2 and p27 and high levels of SKP2. proliferation and induced the expression of p27Kip1 (CDKN1B), which was mediated through an mTOR-dependent mechanism. Implications: These results highlight the contribution of Interestingly, AMPKa2 knockdown resulted in reduced p27 AMPKa2 as a mechanism for controlling bladder cancer growth levels, whereas AMPKa1 suppression did not. To further deter- by regulating proliferation through mTOR suppression and mine the exact mechanism by which AMPKa2 regulates p27, induction of p27 protein levels, thus indicating how AMPKa2 HTB2 and HT1376 cells were transduced with an shRNA targeting loss may contribute to tumorigenesis. Mol Cancer Res; 14(12); AMPKa2. Stable knockdown of AMPKa2 resulted in increased 1182–94. 2016 AACR.

Introduction Due to its role in responding to cellular energy and controlling the switch from anabolic to catabolic processes in response to Bladder cancer is currently the fifth most commonly diagnosed low nutrients, AMPK has long been thought to function to cancer with average age of onset of 71 years old. It is also one of inhibit tumorigenesis. In many cancers, the activation of AMPK the most expensive cancers to treat due to lifelong surveillance results in decreased proliferation and survival of cancer cells and invasive procedures (1, 2). Adenosine monophosphate (10–12). Additionally, it has been reported that AMPK activa- activated protein kinase (AMPK) is a critical protein in the cell tion and/or AMPKa2 protein or mRNA is diminished in breast that signals to control proliferation and protein translation under cancer, hepatocellular carcinoma, and gastric cancer (13–17). cellular stress (3, 4). Once activated by an upstream kinase such Although there are little data on the selective functions of the as STK11, AMPK signaling negatively regulates MTOR (mTOR) AMPKa isoforms, emerging evidence suggests that the proteins to halt protein translation and proliferation when energy avail- may indeed have some selective functions. For example, ability is low (5–7). AMPK is composed of a, b, and g subunits. AMPKa2 has been shown to localize to the nucleus and can The a subunit, which is composed of two isoforms (PRKAA1/ phosphorylate nuclear targets, whereas AMPKa1 does not AMPKa1orPRKAA2/AMPKa2), functions as the catalytic sub- (18, 19). Additionally, Phoenix and colleagues demonstrated unit, while the b and g subunits function as a regulatory and that AMPKa2 / mouse embryonic fibroblasts that were trans- AMP-sensing subunit, respectively (8, 9). formed with H-RasV12 formed tumors in a xenograft, while the AMPKa1 / mouse embryonic fibroblasts did not (20). / 1Center for Vascular Biology, University of Connecticut Health Center, AMPKa2 vascular smooth muscle cells also displayed Farmington Connecticut. 2Department of Cell Biology, University of increased proliferation and a reduction in CDKN1B (p27) Connecticut Health Center, Farmington Connecticut. 3Department of Surgery, University of Connecticut Health Center, Farmington, Con- protein (21). All of these data demonstrate that AMPKa2may necticut. 4Neag Comprehensive Cancer Center, University of Connec- be more selective for controlling p27 protein levels and pro- ticut Health Center, Farmington Connecticut. liferation than AMPKa1. Note: Supplementary data for this article are available at Molecular Cancer The cell-cycle inhibitor protein p27 is a critical protein for Research Online (http://mcr.aacrjournals.org/). controlling cellular proliferation, and its loss has been heavily Corresponding Author: Kevin P. Claffey, University of Connecticut Health implicated in tumorigenesis. Low expression of p27 protein in Center, 263 Farmington Avenue, Farmington, CT 06030. Phone: 860-679- many different cancer tissues, including bladder cancer, has 8713; Fax: 860-679-1201 E-mail: [email protected] been shown to be associated with worse survival and invasive doi: 10.1158/1541-7786.MCR-16-0111 disease (22–24). Under normal conditions, most urothelial cells express p27 due to their slow proliferative rate. In a 2016 American Association for Cancer Research. chemical carcinogenesis model of bladder cancer, p27 / mice

1182 Mol Cancer Res; 14(12) December 2016

AMPKa2 Regulates Bladder Cancer Growth through p27

develop bladder cancer at a much earlier time point than their staining solution (50 mg/mL PI and 100 mg/mL RNase A diluted wild-type counter parts due to the critical role of p27 in in PBS) for 30 minutes before analysis by flow cytometry. controlling urothelial proliferation (25). These experiments demonstrate how crucial p27 is in controlling bladder cancer AICAR, compound C, and rapamycin treatments and cell growth. The exact mechanism for loss of p27 function immunoblot analysis in bladder cancer is currently unknown. One of the major HTB2 and HT1376 cells were plated and 18 hours later sub- mechanisms governing p27 regulation is S-phase kinase-asso- sequently re-fed and treated with 0.5 mmol/L AICAR, 5 nmol/L ciated protein 2 (SKP2) mediated degradation (26). Although rapamycin or 5 mmol/L compound C for a total of 24 hours. there are no studies correlating the expression status of SKP2 Whole-cell lysates were then harvested and analyzed by immu- and p27 in bladder cancer, independent studies have shown noblot as described previously (14). Immunoblots were visual- that p27 is downregulated in invasive bladder cancer and SKP2 ized using ECL reagents (Millipore) and developed on a Kodak is upregulated (27). Therefore, this study seeks to determine Multimodal Imager (2000MM). Signal was quantified by select- theimpactofAMPKa activation on bladder cancer prolifera- ing each band and measuring total photon counts using the tion and whether this is dependent on or mediated in part by Kodak imaging software. Immunoblots are representative of at SKP2-mediated degradation and control of p27 protein. least three distinct experiments. All blots signals are normalized to the b-actin representative for that experiment. Materials and Methods Immunofluorescence Cell lines and reagents HTB2 and HT1376 cells transduced with either shGFP or Cell lines were purchased and maintained according to the shAMPKa2 were plated in a 6-well dish containing glass cover ATCC. Cells obtained from ATCC were frozen down within 5 slips and allowed to grow for 48 hours. Cells were then fixed and passages of being received, and aliquots were not cultured for stained using a p27 antibody. Nuclei were visualized by staining more than 15 passages. Rapamycin and 5-aminoimidazole-4- with 40,6-diamidino-2-phenylindole (DAPI). Cover slips were carboxamide ribonucleotide (AICAR) were obtained from mounted with Prolong gold antifade reagent (Invitrogen) and Cayman Chemical. Compound C was obtained from EMD Milli- visualized on a Zeiss Axio Observer.Z1 with a axiocam MRm pore. Propidium iodide and RNase A were obtained from Invi- camera using Zen software at 400 magnification. trogen. Antibodies targeting AMPKa1 and AMPKa2 were Thr172 obtained from U.S. Biological, total AMPKa, pAMPKa , siRNA transfection SKP2, and pS6 were obtained from Cell Signaling Technology, Lipofectamine 2000 (Life Technologies) was used to transfect p27 was obtained from BD Pharmingen, KI67 was obtained from HTB2 or HT1376 cells with 100 nmol/L of either siLuciferase, DAKO and b-actin was obtained from Abcam. siAMPKa1, siAMPKa2 (Dharmacon), or siSKP2 (Cell Signaling Technology). Cells were lysed 72 hours post transfection and HTB2 and HT1376 stable cell transfection, selection, and further analyzed by immunoblot. maintenance HTB2 and HT1376 cells were transduced with lentivirus-con- Quantitative RT-PCR taining shGFP and shAMPKa2 in the LKO.1-puromycin vector RNA isolation and cDNA synthesis were carried out using (Sigma) as described previously (20). Briefly, HEK-293 FT cells the RNeasy mini kit (Qiagen) and iScript cDNA synthesis kit were plated and transfected the next day with the LKO.1 vector (Bio-Rad), respectively, according to the manufacturer pro- (shGFP or shAMPKa2) and vectors containing RRE, REV, and tocol. Primers are as follows: AMPKa1forward(50-AGGAG- VSV-G. Lentiviral particles were harvested 72 hours post transfec- AGCTATTTGATTATATCTGTAAGAATG-30), AMPKa1 reverse tion and used to transduce HTB2 and HT1376 cells. HTB2 and (50-ACACCAGAAAGGATCTGTTGGAA-30), AMPKa2forward HT1376 cells were transduced for 24 hours followed by 24 hours (50-CGGCTCTTTCAGCAGATTCTGT-30), AMPKa2 reverse (50- of recovery before being selected with puromycin for 72 hours. ATCGGCTATCTTGGCATTCATG-30), Cyclophilin A forward Stable expression was assessed post selection. (50-CTGGACCCAACACAAATGGTT-30), and Cyclophilin A reverse (50-CCACAATATTCATGCCTTCTTTCA-30). Quantita- Cell culture growth assays tive PCR was then performed with SYBR green fluorescence Cells were plated in a 24-well dish (20,000 cells/well) and on a Bio-Rad thermocycler with a MyIQ detection system. treated with AICAR the next day in triplicate. Live-cell counts were taken on day 0 and at indicated experimental times. Briefly, cells were trypsinized and triturated to form a single-cell suspen- Xenograft model sion. Cells, in triplicate, were counted using a Coulter Counter HT1376 shGFP and HT1376 shAMPKa2 were injected in the fl 6 (Beckman Coulter) and normalized to day 0 cell counts to ank (2 10 per site) region of 6- to 8-week-old male NSG mice. obtain percent growth. Graphs are representative data of one Tumors were measured using an external caliper twice a week over experiment out of a total of three distinct experiments. 34 days. The estimate ellipsoid volume equation (length width height 0.52) was used to calculate tumor volume. Cell-cycle analysis HTB2 and HT1376 cells (1 106) were plated in p60 dishes. BBN chemical carcinogenesis model The next day, the medium was aspirated and replaced with fresh Male wild-type and AMPKa2 / C57BL/6 mice (28) at the age medium or fresh medium containing 1 mmol/L AICAR. Cells of 8 weeks were supplemented with 0.05% N-Butyl-N-(4-hydro- were treated for a total of 24 hours and subsequently fixed in 70% xybutyl)nitrosamine (BBN) in their drinking water twice weekly ethanol. Fixed cells were incubated with propidium iodide (PI) for 20 weeks. After 20 weeks, mice were harvested and their

www.aacrjournals.org Mol Cancer Res; 14(12) December 2016 1183

Kopsiaftis et al.

1184 Mol Cancer Res; 14(12) December 2016 Molecular Cancer Research

AMPKa2 Regulates Bladder Cancer Growth through p27

Figure 2. Knockdown of AMPKa2 in bladder cancer cells leads to a reduction in p27 levels. A and B, HTB2 and HT1376 cells were transfected with siRNA (100 nmol/L) against luciferase, AMPKa1, or AMPKa2 and assessed by immunoblot 72 hours after transfection for AMPKa1, AMPKa2, total AMPK (a1/a2), p27, and b-actin.

bladders were removed, weighed, and processed for histologic count 72 hours after treatment. Treatment with AICAR led to a analysis. significant decrease in proliferation in both cell lines (Fig. 1A and B). In order to determine the mechanism by which AMPK Histologic assessment activation reduces proliferation, cell-cycle analysis was per- Bladder cancer tissue array (BL1002) was obtained from U.S. formed on HTB2 and HT1376 cells treated with 1 mmol/L Biomax. Immunohistochemistry was performed on paraffin AICAR for 24 hours. Cell-cycle analysis revealed that both cell sections and quantified as described previously (14). Staining lines presented with a significant increase in the proportion of fi fi was quanti ed from ve 40 images per section using Image cells in G1, indicating that AMPK activation controls prolifer- Pro Plus or through manual counting. Histoscore was deter- ation of bladder cancer cells through inducing G1 arrest (Fig. 1C mined based scoring from three independent reviewers. Sam- and D). To better understand the signaling mechanisms ples that contained no staining were scored as 0, while samples responsible for AMPK mediated suppression of proliferation, that contained staining were scored from a range of 1 to 3, both HTB2 and HT1376 cells were treated with 0.5 mmol/L where 3 is representative of the strongest staining. Samples were AICAR for 24 hours and immunoblots were performed. Both also given a percentage coverage for each score, and then the cells lines displayed potent activation of AMPK in response to numbers were multiplied to obtain a range of 0 to 300. Images AICAR as demonstrated by AMPKaThr172 phosphorylation and were taken on a Zeiss Axioplan 2 microscope with AxioCam ACACA (ACC) phosphorylation (direct substrate of AMPKa). MRc camera and AxioVision software (Oberkochen) at 400 Furthermore, both cell lines demonstrated inhibition of magnification. mTOR activation as assessed by a decrease in phosphorylation of ribosomal protein S6 (S6). Additionally, HTB2 cells dis- Statistical analysis played a 110% upregulation of p27 and HT1376 cells dis- Data from each experiment were represented as the mean played a 27% increase in p27 levels in this experiment when SEM. Statistical analysis was performed using the two-tailed normalized to b-actin in response to AICAR (Fig. 1E and F and Student t test or ANOVA with tests for equal variance where Supplementary Fig. S1A and S1B). AICAR treatment did not appropriate. Significance is represented as , P < 0.05; , P < alter CCNE1 (Cyclin E) or CDK2 levels and resulted in only 0.01; , P < 0.001. minor suppression of CCND1 (Cyclin D1) and CDK4 in HT1376 cells (Supplementary Fig. S2A and S2B). Taken together, these data demonstrate the importance of AMPK sig- Results naling in maintaining cellular proliferation of bladder cancer AMPK activation reduces proliferation in human bladder cells by regulating p27 protein levels and thus causing G1 cancer cells through mTOR-mediated control of p27 arrest. Because AMPK is a major negative regulator of mTOR In order to determine the effect of AMPK activation in blad- activity through activation of the upstream TSC1/2 complexes, der cancer, HTB2 (representative of low-grade bladder cancer) we next analyzed whether mTOR inhibition could replicate and HT1376 (representative of high-grade bladder cancer) the AMPK dependent regulation of p27 by using rapamycin. bladder cancer cell lines were treated with 0.5 mmol/L AICAR, HTB2 and HT1376 cells were treated with 5 nmol/L rapamycin an AMPK activator, and proliferation was determined by cell for 24 hours and immunoblot analysis was performed. In both

Figure 1. AMPK activation reduces proliferation through mTOR-mediated control of p27. A and B, HTB2 (A)andHT1376(B) bladder cancer cells were treated with 0.5 mmol/L AICAR for 72 hours, and cell proliferation was assessed by direct cell count. C and D, cell-cycle analysis of HTB2 (C)andHT1376(D) cells treated with 1 mmol/L AICAR for 24 hours. E and F, HTB2 (E)andHT1376(F) were treated with 0.5 mmol/L AICAR for 24 hours, and immunoblot analysis for p-AMPKa, p-ACC, p-S6, and p27 was performed. b-Actin was used as a loading control. G and H, HTB2 (G)andHT1376(H) bladder cancer cells were treated with 5 nmol/L rapamycin for 24 hours and immunoblot for p-S6, p27, and b-actin was performed. I and J, HTB2 (I)andHT1376(J) were treated simultaneously with rapamycin (5 nmol/L) and compound C (5 mmol/L)for24hoursandimmunoblotforp27,p-S6,andb-actin was performed. Statistical signiﬁcance: , P < 0.05; , P < 0.01; , P < 0.001.

www.aacrjournals.org Mol Cancer Res; 14(12) December 2016 1185

Kopsiaftis et al.

Figure 3. AMPKa2 knockdown reduces p27 through a SKP2-mediated mechanism and results in a proliferative advantage. A and B, HTB2 (A)andHT1376(B) cells were transduced with shRNA to GFP or AMPKa2, and stable cell lines were created. HTB2 and HT1376 stable cell lines were assessed by immunoblot for AMPKa1, AMPKa2, total AMPK (a1/a2), and b-actin. C and D, Transduced cells (A and B) were assessed by qRT-PCR for AMPKa1and AMPKa2 mRNA levels. E and F, HTB2 shGFP, HTB2 shAMPKa2, HT1376 shGFP, and HT1376 shAMPKa2 cells were assessed by immunoblot analysis for p27, SKP2, and b-actin. G, HTB2 and HT1376 stable cell lines were stained with p27 (green) and DAPI (blue). H, Quantiﬁcation of staining in G. I and J, HTB2 and HT1376 stable cells were plated and cell counts were taken daily over 7 days. Graphs represented as a percentage of growth of day 0 cell count. Statistical signiﬁcance: , P < 0.05; , P < 0.01; , P < 0.001.

cell lines, rapamycin signiﬁcantly inhibited the phosphory- tary Fig. S1E and S1F). Taken together, these data demonstrate lation of ribosomal protein S6 (S6), indicating that mTOR that AMPK mediated control of p27 is dependent upon its activation was blocked. Additionally, rapamycin treatment ability to inhibit mTOR. induced the expression of p27 by 57% and 43% when normalized to b-actin in this experiment, consistent with the effects AMPK mediated control of p27 is dependent on the of AICAR on the HTB2 and HT1376 cells, respectively (Fig. 1G AMPKa2 isoform and H and Supplementary Fig. S1C and S1D). To determine AMPK is a heterotrimeric kinase composed of three subunits more speciﬁcally if AMPK mediated suppression of mTOR (a, b,andg), with the a subunit containing the kinase domain. is responsible for the observed upregulation of p27 protein In order to determine which of the two a subunits were and cellular growth, HTB2 and HT1376 cells were treated responsible for the observed regulation of p27, HTB2 cells were with rapamycin and compound C, an AMPK inhibitor. In both transfected with 100 nmol/L of siRNA against luciferase as a cell lines, compound C resulted in a downregulation of p27 control, AMPKa1, or AMPKa2, and analyzed by immunoblot. protein and an upregulation of p-S6. Additionally, rapamycin As depicted in Fig. 2A, transfection of siAMPKa1inHTB2 inhibition of mTOR resulted in reduced p-S6 and coordinate cells caused a 50% reduction of AMPKa1 protein, but did not upregulation of p27. The combination of compound C and alter the levels of the p27 protein. However, transfection of rapamycin demonstrated that by blocking AMPK activation siAMPKa2 resulted in a 55% reduction in AMPKa2 levels and a the upregulation of p27 caused by mTOR inhibition with caused a 45% reduction in p27 protein levels (Supplementary rapamycin alone was diminished indicating that effect is in Fig. S3A). Transfection of HT1376 cells with siAMPKa1 resulted part mediated through mTOR. (Fig. 1I and J and Supplemen- in a 48% reduction in AMPKa1 but also reduced p27 protein

1186 Mol Cancer Res; 14(12) December 2016 Molecular Cancer Research

AMPKa2 Regulates Bladder Cancer Growth through p27

Figure 4. Knockdown of SKP2 negates the effects of AMPKa2-mediated control of p27 and cell proliferation. A and B, HTB2 shGFP, HTB2 shAMPKa2, HT1376 shGFP, and HT1376 shAMPKa2 cells were transfected with siRNA (100 nmol/L) to luciferase or SKP2 and assessed for SKP2, p27, and b-actin levels by immunoblot 72 hours after transfection. C and D, Day 3 cell counts compared with day 0 cell counts of transduced cells transfected with siLuc or siSKP2. Statistical signiﬁcance: , P < 0.05.

levels by 38%. However, transfection with siAMPKa2 resulted mulation, immunofluorescence for p27 was performed on in a 51% reduction in AMPKa2 levels and a 51% reduction in HTB2 shAMPKa2, HT1376 shAMPKa2, and their respective p27 protein levels (Fig. 2B and Supplementary Fig. S3B). Taken control cells (Fig. 3G). Quantification of the percentage of together, these data suggest that AMPKa2 may be more selective nuclear p27 cells revealed that there were significantly fewer in regulating the levels of p27 in bladder cancer cells. p27-positive cells in the shAMPKa2 cell lines (Fig. 3H). Consistent with the downregulation of total p27 and nuclear AMPKa2 regulates cellular proliferation in bladder cancer p27, the HTB2 and HT1376 shAMPKa2 cells displayed a through p27 significant increase in proliferation as assessed by direct cell To determine if AMPKa2 is indeed responsible for the count over 7 days (Fig. 3I and J). Further analysis of HTB2 and AMPK mediated control of cell proliferation, bladder cancer HT1376 cells transduced with a different shRNA targeting cells that express the AMPKa2 protein were transduced with AMPKa2confirmed these results (Supplementary Fig. S4C lentiviral-shRNA to AMPKa2 or GFP (as a control). The HTB2 andS4DandS5A–S5I). Together, these data demonstrate and HT1376 stably selected clones were analyzed by immu- the importance of AMPKa2 in controlling bladder cell prolif- noblot. HTB2 shAMPKa2 cells displayed a 56% reduction in eration by regulating the levels of p27 potentially through AMPKa2 levels and a 13% reduction in total AMPK alpha 1 SKP2-mediated mechanisms. and 2 (tAMPKa1/a2) levels (Fig. 3A). HT1376 shAMPKa2 cells displayed a 46% reduction in AMPKa2 levels and a 28% AMPK-mediated control of p27 is SKP2 dependent reduction in tAMPKa1/a2 levels (Fig. 3B). Quantitative real- Due to the observed upregulation of SKP2 in the AMPKa2 time PCR analysis of these cells also revealed corresponding knockdown cell lines, cells were transfected with 100 nmol/L reductions in AMPKa2 mRNA in both HTB2 (77%) and siRNA to either luciferase as a control or SKP2 for 72 hours HT1376 (54%) cells (Fig. 3C and D). Further analysis of the to determine if AMPKa2-mediated control of p27 is indeed AMPKa2 knockdown cells revealed that p27 protein levels mediated through SKP2. Knockdown of SKP2 in both HTB2 were downregulated by 50% in the HTB2 shAMPKa2 cells and HT1376 indeed relieved the repression of p27 that (Fig. 3E) and 41% in the HT1376 shAMPKa2 cells (Fig. 3F). In occurred in the AMPKa2 knockdown cells (Fig. 4A and B). In order to determine more specifically what could be causing the order to determine if this affect was biologically important, downregulation of p27, SKP2 levels were analyzed. SKP2 is a proliferation assays were performed simultaneously with subunit of the F-box ubiquitin ligase complex, which targets immunoblot analyses. In both HTB2 and HT1376 cells, knock- p27 (29). In both the HTB2 shAMPKa2 and the HT1376 down of SKP2 significantly negated the proliferative advantage shAMPKa2 cells, SKP2 levels were upregulated by 40% and of the shAMPKa2 cells (Fig. 4C and D). Furthermore, knock- 81%, respectively (Fig. 3E and F). In both cell lines, shAMPKa2 down of SKP2 in HTB2 and HT1376 cells transduced with the did not affect the cell-cycle proteins Cyclin D1, Cyclin E, or second shRNA clones confirmed these results (Supplementary CDK2, however, did induce a moderate increase in CDK4 Fig. S6A–S6D). These data signify that the effect that AMPKa2 (Supplementary Fig. S4A and S4B). Furthermore, to determine has on p27 and its role in controlling cellular proliferation is if the downregulation of p27 also affected its nuclear accu- dependent on SKP2 regulation.

www.aacrjournals.org Mol Cancer Res; 14(12) December 2016 1187

Kopsiaftis et al.

250 )

A 3 shGFP 200 shAMPKa2

150

100

Tumor volume (mm Tumor 0 01020 30 40 Day B shGFP shAMPKa2

200

150 p-AMPK 100

50 (arbitrary units) p-AMPK Intensity

0 shGFP shAMPKa2 Figure 5. C 80 Knockdown of AMPKa2inHT1376 60 cells promotes tumor growth in a xenograft model. A, HT1376 shGFP KI67 40 and HT1376 shAMPKa2(n ¼ 9/

20 group) cells were injected into the ﬂank of NSG mice. Tumor volume 0 Percent Ki67-positive cells was measured twice weekly. B–F, shGFP shAMPKa2 pAMPK (B), KI67 (C), p-S6 (D),

150 p27 (E), and SKP2 (F) D immunohistochemical staining representative images and 100 p-S6 corresponding quantiﬁcation (n ¼ 9/ group). Statistical signiﬁcance: 50 , P 0.05; , P 0.01; , P 0.001.

p-S6 Intensity < < < (arbitrary units)

0 shGFP shAMPKa2

E 50

p27 30

20 Percent P27- positive cells 10

0 shGFP shAMPKa2

F 40

SKP2 20 positive cells

Percent SKP2- 10

0 shGFP shAMPKa2

Knockdown of AMPKa2 results in larger tumor formation in a fixed for immunohistochemical analysis. In order to establish xenograft model if the same signaling cascade was occurring in the xenograft To determine whether AMPKa2anditsregulationofp27 model that occurred in vitro, the levels of phosphorylated could affect overall tumorigenicity, the bladder cancer cells, AMPK were assessed by immunohistochemistry in the shGFP HT1376 shGFP and HT1376 shAMPKa2 cells were injected and the shAMPKa2 tumors. Indeed, knockdown of AMPKa2 (2 106 cells) into the flank region of Nod Scid gamma (NSG) significantly decreased p-AMPKThr172 levels by 23% in the mice and tumors were allowed to grow for 34 days. On day 34, xenograft model (Fig. 5B). Consistent with the decreased AMPK HT1376 shAMPKa2 tumors were 56% larger (P < 0.05) than activity in the shAMPKa2 tumors, it was also observed that the shGFP control tumors (Fig. 5A) as determined by external there was a concomitant activation of mTOR by 48% in these caliper measurements and were subsequently collected and tumors as determined by p-S6 levels (Fig. 5D). Additionally,

1188 Mol Cancer Res; 14(12) December 2016 Molecular Cancer Research

AMPKa2 Regulates Bladder Cancer Growth through p27

A 100

0 Bladder weight at harvest - - Wild-type AMPKa2 / BCWild-type AMPKa2-/- 100

p-AMPK 60 40

(arbitrary units) 20

Figure 6. Staining intensity/area 0 Genetic deletion of AMPKa2 results in WT AMPKa2-/- larger tumors in a BBN carcinogen– D E 150 induced bladder tumor model. A, Bladder weights of wild-type and AMPK 2/ mice after 20 weeks of 100 a t-AMPKa continuous BBN supplementation in drinking water (n ¼ 20/group). 50 B, Representative images of p-AMPKa (arbitrary units) staining. C, quantification of p-AMPKa Staining intensity/area 0 AMPKa2-/- staining in wild-type and AMPKa2 / F G WT bladders. D, Representative images 60 of tAMPKa staining. E, quantification of tAMPKa staining in wild-type 40 and AMPKa2 / bladders. F, KI67 representative images of KI67 staining. 20 G, Quantification of KI67 staining in wild-type and AMPKa2 / bladders. 0 Percent Ki67-positive cells H, Representative images of p-S6 Wild-type AMPKa2-/- fi staining. I, quanti cation of p-S6 H I 50 staining in wild-type and AMPKa2/ 40 bladders. J, Representative images of p27 staining. K, Quantification of p27 p-S6 30 / staining in wild-type and AMPKa2 20 bladders. L, Representative images of

(arbitrary units) 10 SKP2 staining. M, Quantiﬁcation of Staining intensity/area 0 SKP2 staining in wild-type and -/- WT AMPKa2 AMPKa2 / bladders. Statistical J K signiﬁcance: , P < 0.05. 40

30 p27 20

10 Percent positive nuclear P27 cells

0 WT AMPKa2-/- L M 150

100 SKP2

50 (arbitrary units)

SKP2 Staining intensity 0 WT AMPKa2-/- knockdown of AMPKa2 in the xenograft model signiﬁcantly control of cellular proliferation in vitro is also seen in an in vivo increased proliferation (45% vs. 57%), as assessed by KI67 model of bladder cancer. staining(Fig.5C).Furthermore,p27levelsweresigniﬁcantly diminished by 55% (33% vs. 14%) in the shAMPKa2 group Genetic deletion of AMPKa2 results in larger bladder tumors in when compared with the shGFP group (Fig. 5E), with a corre- a chemical carcinogenesis model sponding increase in SKP2 levels (19% vs. 32%; Fig. 5F). Taken In order to better establish the impact that AMPKa2has together, these data demonstrate that the AMPKa2-mediated on bladder tumorigenesis in an endogenous system, AMPKa2

www.aacrjournals.org Mol Cancer Res; 14(12) December 2016 1189

Kopsiaftis et al.

A B

Adjacent non-tumor 300 a 2

200

100 Relative AMPK Bladder tumor Histoscore (0-300) 0 Adjacent non-tumor Tumor

Adjacent non-tumor D

300

200

AMPK a 2 100 Bladder tumor

Histoscore (0-300) 0 Adjacent non-tumor Bladder tumor

EF GH

600 600 200 25

20 150 400 400 15 100 RPKM RPKM 10 200 200 50 5 (arbitrary units) p27 Expression 0 0 0 0 % SKP2 Positive cells Adjacent non-tumor Bladder tumor Adjacent non-tumor Bladder tumor Adjacent non-tumor Bladder tumor Adjacent non-tumor Bladder tumor

I AMPKα2 p27 SKP2

Adjacent non-tumor

Bladder tumor

Figure 7. AMPKa2 suppression in bladder cancer correlates with p27. A, Representative 40 images of adjacent non-tumor and bladder tumor tissue stained for AMPKa2. Images represent an example of low AMPKa2 expression and high AMPKa2 expression in bladder tumor samples. B, Quantiﬁcation of AMPKa2 expression in adjacent non-tumor and bladder tumor samples (n ¼ 10 adjacent non-tumor and n ¼ 40 tumor). C, Representative 40 images of patient-matched adjacent non-tumor and bladder tumor tissue stained for AMPKa2. (Continued on the following page.)

1190 Mol Cancer Res; 14(12) December 2016 Molecular Cancer Research

AMPKa2 Regulates Bladder Cancer Growth through p27

knockout mice were utilized in a chemical carcinogenesis Immunostaining for p27 in adjacent non-tumor and bladder model. Age-matched wild-type and AMPKa2 / mice were tumor tissue revealed that p27 was significantly suppressed by given N-Butyl-N-(4-hydroxybutyl)nitrosamine (BBN) at a con- 73% in bladder cancer (Fig. 7G). In order to determine if the centration of 0.05% for 20 weeks in drinking water. After decrease in p27 in bladder cancer was due to an increase in 20 weeks of exposure, mice were euthanized, and the bladders SKP2 expression in bladder cancer, the levels of SKP2 were harvested were weighed as a surrogate marker for tumor assessed and found to be significantly elevated in bladder burden. AMPKa2 / bladders were significantly larger (44%, cancer (Fig. 7H). In order to recapitulate the results from the P < 0.05) than the wild-type control bladders (Fig. 6A). Addi- in vitro and in vivo studies demonstrating that lack of AMPKa2 tionally, bladder tumors from AMPKa2 / displayed a signif- results in an increase in SKP2 and a decrease in p27, the icant 52% reduction (P < 0.05) in AMPK activation as assessed expression of AMPKa2, SKP2, and p27 was assessed in the by p-AMPKThr172 when compared with wild-type mice (Fig. 6B same tissue sections. This analysis revealed that there was a and C), but only a 19% reduction in tAMPKa1/a2(Fig.6Dand significant negative correlation between p27 and SKP2, as E). AMPKa2 / bladder tumors also displayed a 33% increase expected, because SKP2 functions to target p27 for degradation, in mTOR activation as assessed by p-S6 (Fig. 6H and I). and that there was a significant correlation between AMPKa2 Consistent with the in vitro cell proliferation and xenograft and p27, further demonstrating that AMPKa2 contributes to model, AMPKa2 / tumors demonstrated a significant increase the regulation of p27 in bladder cancer (Fig. 7I). Together, in proliferation as assessed by KI67 staining and quantification these data demonstrate for the firsttimethatAMPKa2protein (Fig.6FandG).Furthermore,AMPKa2 / bladder tumors had levels are reduced in bladder cancer when compared with non- significantly fewer p27-positive cells (P < 0.05; Fig. 6J and K) cancerous tissue and that this reduction is correlated with an and a significant increase in SKP2 expression (P < 0.05; Fig. 6L increase in SKP2 expression with a concomitant reduction in and M), when compared with the wild-type control bladder p27 protein levels. tumors. Together, these data demonstrate that AMPKa2func- tions to control bladder cancer growth through regulating SKP2-mediated degradation of p27. Discussion AMPK is the central metabolic regulator of cells and regulates AMPKa2 is suppression in human bladder cancer correlates the critical anabolic and proliferative pathways largely through its with decreased p27 inhibition of mTOR (6, 30, 31). Many reports have suggested that AMPKa expression has been reported to be altered in many AMPK may function to inhibit tumorigenesis because of its different cancers (13–15). In order to determine if AMPKa2 involvement in regulating cell growth through canonical path- levels are altered in bladder cancer, immunohistochemistry ways, such as protein translation and cell growth through regu- was performed on a tissue array obtained from Biomax Inc. lating mTOR (8, 32–35). Although many reports have demon- which consisted of 10 adjacent non-tumor tissue samples and strated that AMPK can function to reduce tumor cell growth, there 40 bladder tumor samples. AMPKa2 protein levels were sup- have also been some reports demonstrating the opposite. For pressed by 59% in bladder tumor tissue when compared with instance, mouse embryonic fibroblasts null for both AMPKa1 adjacent non-tumor tissue as control, and was statistically and AMPKa2 that are transformed with H-RasV12 grew signif- significant (Fig. 7A and B). In order to rule out patient vari- icantly slower than their wild-type counter parts in a xenograft ability in expression of AMPKa2, 15 adjacent non-tumor blad- model (36). Also, in glioblastoma, it has been demonstrated der and patient matched bladder tumor tissue was immunos- that AMPKa activation facilitated tumor growth and that tained for AMPKa2. Analysis of high-grade patient matched AMPKa activation was localized to areas of the tumor with high samples confirmed the results of the bladder tissue array, proliferation, further indicating how AMPKa can function to indicating that AMPKa2proteinissignificantly suppressed by promote tumor growth (37). This creates a situation where it is 68% in bladder cancer (Fig. 7C and D). Furthermore, analysis necessary to determine the impact of AMPK activation in differ- of AMPKa2 mRNA from The Cancer Genome Atlas (TCGA) ent tumor types to avoid utilizing it as a therapeutic strategy in database revealed that AMPKa2wassignificantly suppressed at tumor types where it may promote growth. the mRNA level by 82% in bladder cancer (Fig. 7E). To avoid Although there are many reports on the relevance of AMPK patient variation, AMPKa2 mRNA expression was assessed signaling and its anti-growth effects in different tumor types, few in patient-matched samples from the TCGA dataset, and this studies have addressed the relevance of AMPK signaling in revealed that AMPKa2mRNAissignificantly suppressed by bladdercancer(14,20,38–40). In bladder cancer specifically, 76% in bladder cancer (Fig. 7F). Because AMPKa2wasfound activation of AMPK has been shown to result in translational to be suppressed in bladder cancer, it was important to deter- inhibition through a decrease in mTOR activation (41). Addi- mine if the expression of p27 and/or SKP2 were altered in tionally, Shorning and colleagues demonstrated that in a model bladder cancer. The expression of both p27 and SKP2 was of bladder cancer in which LKB1 / (an upstream kinase of assessed in bladder cancer using the array from Biomax Inc. AMPK) and PTEN / mice were utilized, mTOR was drastically

(Continued.) D, Quantification of patient-matched samples (n ¼ 15). E, AMPKa2 mRNA in 19 adjacent non-tumor and 164 bladder tumor samples. (Represented as reads per kilobase of transcript per million mapped reads, RPKM.) F, AMPKa2 mRNA in 15 patient-matched adjacent non-tumor and bladder tumor samples (represented as RPKM). G, Quantification of p27 staining from array (n ¼ 10 adjacent non-tumor and 40 bladder tumor). H, Quantification of SKP2 staining from array (n ¼ 10 adjacent non-tumor and 40 bladder tumor). I, representative 40 images of AMPKa2, p27, and SKP2 staining in adjacent non-tumor and bladder tumor samples. Statistical significance: , P < 0.05; , P < 0.001. The data presented in this figure are generated in part by data gathered by the TCGA Research Network.

www.aacrjournals.org Mol Cancer Res; 14(12) December 2016 1191

Kopsiaftis et al.

upregulated in tumorigenesis and this was blocked by rapamy- proliferation. Knockdown of SKP2 in both the control and cin treatment (42). These data suggest an importance of AMPK AMPKa2 knockdown cells caused an increase in p27 protein signaling in bladder cancer. We therefore assessed the effect of levels. Consistent with these data, knockdown of SKP2 reversed AMPK activity on bladder cancer cells. The AMPK activator the effects that AMPKa2 knockdown had on cellular proliferation, AICAR was used to determine the impact of AMPK signaling indicating that SKP2 is indeed responsible for the observed effects. on bladder cancer cell proliferation. This revealed that activation Although further research is warranted to determine the direct of AMPK causes an upregulation of p27 and a reduction in impact that p27 expression on bladder cancer growth, these data proliferation. Two different reports have demonstrated that suggest that a mechanism of proliferative control exists in bladder AMPK activation causes phosphorylation of p27 and induces cancer, involving the cross-talk between metabolic and cell-cycle p27 stability (43, 44). In order to determine if AMPK control of control pathways. bladder cancer growth is due to a direct phosphorylation of p27 To further understand the impact that AMPKa2 may have on or due to downstream signaling, bladder cancer cells were bladder cancer cell growth in vivo, a xenograft model utilizing treated with rapamycin, an mTOR inhibitor, and compound HT1376 shGFP and HT1376 shAMPKa2 was used. Consistent C, an AMPK inhibitor. This revealed that AMPK regulation of with the in vitro data, HT1376 shAMPKa2 cells resulted in larger p27 was at least partially mTOR dependent. tumors when compared with their control. Additionally, AMPK AMPK is a heterotrimeric protein composed of a, b, and g activation was reduced in the HT1376 shAMPKa2, which resulted subunits. The a subunit is responsible for the kinase activity of the in concomitant mTOR activation, an increase in proliferation, a protein and exists as one of two isoforms, a1ora2. There have decrease in nuclear p27, and an increase in SKP2 staining. been conflicting reports to date as to whether the a isoforms Although the xenograft model showed an increase in tumor function redundantly or whether they have additional specificor growth in vivo when AMPKa2 was reduced, a genetic chemical selective functions (32). A chemical genetic screen of AMPKa2 model of bladder cancer was used to further investigate the substrates defined several possible targets related to cell mito- importance of AMPKa2 in bladder tumorigenesis. In this model, sis (19). Additionally, it has been demonstrated that only wild-type and AMPKa2 / mice received BBN in drinking water AMPKa2 / mouse embryonic fibroblasts formed tumors upon to induce bladder cancer formation and represents a model that H-Ras transformation in a xenograft model, suggesting that more closely mimics that of human bladder cancer. In this model, AMPKa1 and AMPKa2 may indeed have selective functions the genetic deletion of AMPKa2 resulted in larger tumor forma- (14). Furthermore, knockdown of only AMPKa2 and not AMPKa1 tion and increased proliferation. Furthermore, the deletion of in vascular smooth muscle cells resulted in a decrease of p27 AMPKa2 resulted in decreased nuclear p27 staining and increased protein and conferred a proliferative advantage to cells (21). To SKP2 staining. Although the deletion of AMPKa2 may have address the question as to whether AMPKa1, AMPKa2, or both additional effects on tumorigenesis beyond its contribution to isoforms can be responsible for the regulation of p27 protein levels, increasing mTOR, the results are consistent with the in vitro data siRNA specific to each isoform was utilized and p27 levels were and the xenograft model showing that loss of AMPKa2 contri- evaluated. Consistent with the study in vascular smooth muscle butes to a reduction in p27 protein with a concomitant increase in cells, only the knockdown of AMPKa2 resulted in a reduction of SKP2 expression. p27 protein, indicating that the AMPKa2 isoform is a highly This study demonstrates for the first time the impact that specific mechanism of bladder cancer proliferative control. AMPK, more specifically AMPKa2, has on bladder cancer To better understand the impact that AMPKa2 may have in growth through p27 regulation. AMPK has long been impli- controlling bladder cancer cell growth, HTB2 and HT1376 cells cated as a potential tumor suppressor-like protein in cancer due were transduced with shRNA to AMPKa2. Although there was a to its multifaceted role in controlling cellular metabolism, and significant decrease in AMPKa2 protein level after transduction, recently it has been reported that AMPKa2levelsarealteredin there was only a modest decrease in total AMPKa1/a2, indicating cancer. It has been reported that AMPKa2 protein is suppressed that the AMPKa1 isoform may be the major isoform expressed in in breast cancer and hepatocellular carcinoma and suppressed bladder cancer. Despite being the minor isoform, knockdown of in early stages of gastric cancer at the RNA level (13, 14, 17, 38). AMPKa2 had profound effects on increasing cellular proliferation These data clearly implicate AMPKa2specifically as an impor- and p27 protein levels. Consistent with decreased p27 protein tant protein in regulating cancer cell growth. To our knowledge, levels, AMPKa2 knockdown resulted in decreased p27 nuclear there are no published studies of AMPK regulation in human accumulation. AMPKa2 knockdown also had an effect on CDK4 bladder cancer to date. This study demonstrates that AMPKa2 protein levels in HT1376 cells; however, this was not observed in mRNA and protein is suppressed in bladder cancer when HTB2 cells and could be potentially attributed to the higher the compared with non-tumor urothelium. Previous reports have loss of Rb protein observed in HT1376 cells (45). In order to demonstrated that the AMPKa2 promoter is methylated and specifically address what mechanism may be controlling AMPK- suppressed in hepatocellular carcinoma (17). Because AMPKa2 mediated regulation of p27 protein, SKP2 levels were assessed due mRNA is suppressed in bladder cancer, it is plausible that to its widely reported role in functioning as an E3 ubiquitin ligase AMPKa2 is suppressed due to methylation, although further targeting p27 and several reports demonstrating that activation of studies are required to validate this notion. Furthermore, our AMPK and/or inhibition of mTOR results in a decrease in SKP2 study demonstrates that p27 protein is downregulated in blad- protein levels (24, 46–49). Not surprisingly, knockdown of der cancer and is correlated with AMPKa2expressioninthe AMPKa2 in bladder cancer cells resulted in a prominent upregu- bladder. This further demonstrates the importance of AMPKa2 lation of SKP2 protein levels. Given that AMPKa2 knockdown in in regulating p27 protein levels in normal urothelium and in bladder cancer cells resulted in an upregulation of SKP2 protein, bladder cancer cells, where loss promotes proliferation and siRNA to SKP2 was utilized to determine if SKP2 indeed was growth of tumors. AMPK is a multifaceted tumor suppressor- responsible for the downregulation of p27 and the increased cell like protein and our study demonstrates how it is involved in

1192 Mol Cancer Res; 14(12) December 2016 Molecular Cancer Research

AMPKa2 Regulates Bladder Cancer Growth through p27

controlling bladder cancer growth through regulation of p27; limiting the capacity of bladder cancer growth because the however, further research is warranted to determine if the inhibition of SKP2 would lead to an increase in p27 protein effects of AMPKa2onbladdercancergrowtharemediated levels. through mTOR regulation or solely through the regulation of p27. It has been reported that p27 is often dysregulated in Disclosure of Potential Conflicts of Interest several types of cancer and that low expression is often linked to No potential conflicts of interest were disclosed. a more severe phenotype (24, 50). In bladder cancer specifically, it has been demonstrated that p27 / mice are much more susceptible to tumorigenesis in the BBN model and that Authors' Contributions low p27 expression in human bladder cancer has been linked to Conception and design: S. Kopsiaftis, K.P. Claffey metastatic disease and poor outcome (23, 25). It is therefore Development of methodology: S. Kopsiaftis, K.P. Claffey likely that a reduction of AMPKa2, which results in a reduction Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Kopsiaftis, J.A. Taylor III, K.P. Claffey of p27, would have a profound impact on bladder cancer Analysis and interpretation of data (e.g., statistical analysis, biostatistics, growth. computational analysis): S. Kopsiaftis, K.L. Sullivan, I. Garg, K.P. Claffey The direct targeting of AMPKa2 in bladder cancer would Writing, review, and/or revision of the manuscript: S. Kopsiaftis, K.L. Sullivan, result in the repression of mTOR activity and would result in J.A. Taylor III, K.P. Claffey an increase in p27 protein, which could limit bladder cancer Study supervision: K.P. Claffey growth (Graphical Abstract). However, due to the loss of AMPKa2 protein expression in bladder cancer, it would not Acknowledgments be feasible to target AMPKa2 as an effective mechanism for The authors would like to thank Roderick Franczak and Shilpa treating bladder cancer. Additionally, mTOR inhibitors are Choudhary for their technical assistance and sample preparation. The authors currently being used in the clinic with mixed success due to would also like to thank Benoit Viollet for providing the AMPKa2 / mice. off-target effects and toxicity despite their promise in preclin- The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked ical models. The results from this study demonstrate that an advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate inhibitor of SKP2 could be an effective way of treating bladder this fact. cancer patients because inhibiting SKP2 would repress the proliferative advantage that the loss of AMPKa2confersto Received April 5, 2016; revised August 12, 2016; accepted August 29, 2016; bladder cancer. This would be a more direct mechanism for published OnlineFirst September 16, 2016.

References 1. Degraff DJ, Cates JM, Mauney JR, Clark PE, Matusik RJ, Adam RM. 14. Fox MM, Phoenix KN, Kopsiaftis SG, Claffey KP. AMP-activated protein When urothelial differentiation pathways go wrong: implications kinase alpha 2 isoform suppression in primary breast cancer alters AMPK for bladder cancer development and progression. Urol Oncol 2013; growth control and apoptotic signaling. Genes Cancer 2013;4:3–14. 31:802–11. 15. Wang H, Zheng LY, Yang W, Wu FQ, Wang C, Yu L, et al. Prognostic 2. Rosenberg JE, Hahn WC. Bladder cancer: modeling and translation. Genes significance of AMPK activation and therapeutic effects of metformin in Dev 2009;23:655–9. hepatocellular carcinoma. Clin Cancer Res 2013;19:5372–80. 3. Carling D, Viollet B. Beyond energy homeostasis: the expanding role of 16. KimYH,LiangH,LiuX,LeeJS,ChoJY,CheongJH,etal.AMPKalpha AMP-activated protein kinase in regulating metabolism. Cell Metab modulation in cancer progression: multilayer integrative analysis 2015;21:799–804. of the whole transcriptome in Asian gastric cancer. Cancer Res 2012; 4. Fay JR, Steele V, Crowell JA. Energy homeostasis and cancer prevention: the 72:2512–21. AMP-activated protein kinase. Cancer Prev Res 2009;2:301–9. 17. Lee CW, Wong LL, Tse EY, Liu HF, Leong VY, Lee JM, et al. AMPK promotes 5. Liang J, Mills GB. AMPK: a contextual oncogene or tumor suppressor? p53 acetylation via phosphorylation and inactivation of SIRT1 in liver Cancer Res 2013;73:2929–35. cancer cells. Cancer Res 2012;72:4394–404. 6. Steinberg GR, Kemp BE. AMPK in health and disease. Physiol Rev 18. Salt I, Celler JW, Hawley SA, Prescott A, Woods A, Carling D, et al. AMP- 2009;89:1025–78. activated protein kinase: greater AMP dependence, and preferential nuclear 7. Dazert E, Hall MN. mTOR signaling in disease. Curr Opin Cell Biol localization, of complexes containing the alpha2 isoform. Biochem J 2011;23:744–55. 1998;334):177–87. 8. Shackelford DB, Shaw RJ. The LKB1–AMPK pathway: metabolism and 19. Banko MR, Allen JJ, Schaffer BE, Wilker EW, Tsou P, White JL, et al. growth control in tumour suppression. Nat Rev Cancer 2009;9:563–75. Chemical genetic screen for AMPKalpha2 substrates uncovers a network 9. Viollet B, Athea Y, Mounier R, Guigas B, Zarrinpashneh E, Horman S, et al. of proteins involved in mitosis. Mol Cell 2011;44:878–92. AMPK: lessons from transgenic and knockout animals. Front Biosci 20. Phoenix KN, Devarakonda CV, Fox MM, Stevens LE, Claffey KP. 2009;14:19–44. AMPKalpha2 suppresses murine embryonic fibroblast transformation 10. Hadad SM, Hardie DG, Appleyard V, Thompson AM. Effects of metformin and tumorigenesis. Genes Cancer 2012;3:51–62. on breast cancer cell proliferation, the AMPK pathway and the cell cycle. 21. Song P, Wang S, He C, Wang S, Liang B, Viollet B, et al. AMPKalpha2 Clin Transl Oncol 2014;16:746–52. deletion exacerbates neointima formation by upregulating Skp2 in vascu- 11. Kim J, Kundu M, Viollet B, Guan KL. AMPK and mTOR regulate lar smooth muscle cells. Circ Res 2011;109:1230–9. autophagy through direct phosphorylation of Ulk1. Nat Cell Biol 22. Shariat SF, Zlotta AR, Ashfaq R, Sagalowsky AI, Lotan Y. Cooperative effect 2011;13:132–41. of cell-cycle regulators expression on bladder cancer development and 12. Zhuang Y, Miskimins WK. Cell cycle arrest in Metformin treated breast biologic aggressiveness. Mod Pathol 2007;20:445–59. cancer cells involves activation of AMPK, downregulation of cyclin D1, and 23. Rabbani F, Koppie TM, Charytonowicz E, Drobnjak M, Bochner BH, requires p27Kip1 or p21Cip1. J Mol Signal 2008;3:18. Cordon-Cardo C. Prognostic significance of p27Kip1 expression in bladder 13. Hadad SM, Baker L, Quinlan PR, Robertson KE, Bray SE, Thomson G, et al. cancer. BJU Int 2007;100:259–63. Histological evaluation of AMPK signalling in primary breast cancer. BMC 24. Bloom J, Pagano M. Deregulated degradation of the cdk inhibitor p27 and Cancer 2009;9:307. malignant transformation. Semin Cancer Biol 2003;13:41–7.

www.aacrjournals.org Mol Cancer Res; 14(12) December 2016 1193

Kopsiaftis et al.

25. Hikosaka A, Ogawa K, Sugiura S, Asamoto M, Takeshita F, Sato SY, 39. Zakikhani M, Dowling R, Fantus IG, Sonenberg N, Pollak M. Metformin is et al. Susceptibility of p27 kip1 knockout mice to urinary bladder an AMP kinase-dependent growth inhibitor for breast cancer cells. Cancer carcinogenesis induced by N-butyl-N-(4-hydroxybutyl)nitrosamine Res 2006;66:10269–73. may not simply be due to enhanced proliferation. Int J Cancer 2008; 40. Dowling RJ, Zakikhani M, Fantus IG, Pollak M, Sonenberg N. 122:1222–8. Metformin inhibits mammalian target of rapamycin-dependent 26. Nickeleit I, Zender S, Kossatz U, Malek NP. p27kip1: a target for tumor translationinitiationinbreastcancercells.CancerRes2007;67: therapies? Cell Div 2007;2:13. 10804–12. 27. Elsherif E, Elbaky TA, Elserafy F, Elkady N, Dawood M, Gaber MA, et al. 41. Liu Z, Li X, Simoneau AR, Jafari M, Zi X. Rhodiola rosea extracts and beta-cateninandSKP2proteinsaspredictorsofgradeandstageof salidroside decrease the growth of bladder cancer cell lines via inhibition of non-muscle invasive urothelial bladder carcinoma. Chin Clin Oncol the mTOR pathway and induction of autophagy. Mol Carcinog 2012;51: 2016;5:6. 257–67. 28. Viollet B, Andreelli F, Jorgensen SB, Perrin C, Geloen A, Flamez D, et al. The 42. Shorning BY, Griffiths D, Clarke AR. Lkb1 and Pten synergise to AMP-activated protein kinase alpha2 catalytic subunit controls whole- suppress mTOR-mediated tumorigenesis and epithelial–mesenchymal body insulin sensitivity. J Clin Invest 2003;111:91–8. transition in the mouse bladder. PLoS One 2011;6:e16209. 29. Bochis OV, Irimie A, Pichler M, Berindan-Neagoe I. The role of Skp2 and its 43. Liang J, Shao SH, Xu ZX, Hennessy B, Ding Z, Larrea M, et al. The energy substrate CDKN1B (p27) in colorectal cancer. J Gastrointestin Liver Dis sensing LKB1-AMPK pathway regulates p27(kip1) phosphorylation medi- 2015;24:225–34. ating the decision to enter autophagy or apoptosis. Nat Cell Biol 2007; 30. Hardie DG. AMP-activated protein kinase: an energy sensor that regulates 9:218–24. all aspects of cell function. Genes Dev 2011;25:1895–908. 44. Short JD, Dere R, Houston KD, Cai SL, Kim J, Bergeron JM, et al. 31. Towler MC, Hardie DG. AMP-activated protein kinase in metabolic control AMPK-mediated phosphorylation of murine p27 at T197 promotes and insulin signaling. Circ Res 2007;100:328–41. binding of 14-3-3 proteins and increases p27 stability. Mol Carcinog 32. Dasgupta B, Chhipa RR. Evolving lessons on the complex role of AMPK in 2010;49:429–39. normal physiology and cancer. Trends Pharmacol Sci 2016;37:192–206. 45. Goodrich DW, Chen Y, Scully P, Lee WH. Expression of the retinoblastoma 33. Hardie DG. AMP-activated protein kinase: a cellular energy sensor with a gene product in bladder carcinoma cells associates with a low frequency of key role in metabolic disorders and in cancer. Biochem Soc Trans 2011; tumor formation. Cancer Res 1992;52:1968–73. 39:1–13. 46. Hsu JD, Kao SH, Ou TT, Chen YJ, Li YJ, Wang CJ. Gallic acid induces G2/M 34. Fogarty S, Ross FA, Vara Ciruelos D, Gray A, Gowans GJ, Hardie DG. AMPK phase arrest of breast cancer cell MCF-7 through stabilization of p27(Kip1) causes cell cycle arrest in LKB1-deficient cells via activation of CAMKK2. attributed to disruption of p27(Kip1)/Skp2 complex. J Agric Food Chem Mol Cancer Res 2016;14:683–95. 2011;59:1996–2003. 35. Zadra G, Batista JL, Loda M. Dissecting the dual role of AMPK in cancer: 47. Vervoorts J, Luscher B. Post-translational regulation of the tumor suppres- from experimental to human studies. Mol Cancer Res 2015;13:1059–72. sor p27(KIP1). Cell Mol Life Sci 2008;65:3255–64. 36. Laderoute KR, Amin K, Calaoagan JM, Knapp M, Le T, Orduna J, et al. 50- 48. FanGH,WangZM,YangX,XuLP,QinQ,ZhangC,etal.Resveratrol AMP-activated protein kinase (AMPK) is induced by low-oxygen and inhibits oesophageal adenocarcinoma cell proliferation via AMP- glucose deprivation conditions found in solid-tumor microenvironments. activated protein kinase signaling. Asian Pac J Cancer Prev 2014;15: Mol Cell Biol 2006;26:5336–47. 677–82. 37. Rios M, Foretz M, Viollet B, Prieto A, Fraga M, Costoya JAAMPK activation 49. Totary-Jain H, Sanoudou D, Dautriche CN, Schneller H, Zambrana L, by oncogenesis is required to maintain cancer cell proliferation in astrocytic Marks AR. Rapamycin resistance is linked to defective regulation of Skp2. tumors. Cancer Res 2013;73:2628–38. Cancer Res 2012;72:1836–43. 38. Zheng L, Yang W, Wu F, Wang C, Yu L, Tang L, et al. Prognostic significance 50. Philipp-Staheli J, Payne SR, Kemp CJ. p27(Kip1): regulation and function of AMPK activation and therapeutic effects of metformin in hepatocellular of a haploinsufficient tumor suppressor and its misregulation in cancer. carcinoma. Clin Cancer Res 2013;19:5372–80. Exp Cell Res 2001;264:148–68.

1194 Mol Cancer Res; 14(12) December 2016 Molecular Cancer Research

AMPKα2 Regulates Bladder Cancer Growth through SKP2-Mediated Degradation of p27

Stavros Kopsiaftis, Katie L. Sullivan, Isha Garg, et al.

Mol Cancer Res 2016;14:1182-1194. Published OnlineFirst September 16, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-16-0111

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2016/09/15/1541-7786.MCR-16-0111.DC1

Cited articles This article cites 50 articles, 19 of which you can access for free at: http://mcr.aacrjournals.org/content/14/12/1182.full#ref-list-1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/14/12/1182. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.