gas1 inhibitsmetastaticandmetabolicphenotypes in...

Metabolism

Gas1 InhibitsMetastaticandMetabolicPhenotypesin Colorectal CarcinomaQingguo Li1,2, Yi Qin2,3, Ping Wei2,4, Peng Lian1,2, Yaqi Li1,2, Ye Xu1,2, Xinxiang Li1,2,Dawei Li1,2, and Sanjun Cai1,2

Abstract

Growth arrest–specific 1 (Gas1) plays a critical role in growthsuppression. Previous study indicated that Gas1 was closelyassociated with survival in patients with colorectal cancer; how-ever, the underlying molecular mechanism remains unclear. Inthis study, we sought to determine the role of Gas1 in tumori-genesis and metastasis, and elucidate the possible mechanism.First, Gas1 was determined as a negative regulator of oncogenesisand metastasis in colorectal cancer. Mechanistically, Gas1 nega-tively regulated the aerobic glycolysis, a process that contributedto tumor progression and metastasis by providing energy sourceand building blocks for macromolecule synthesis. To furtherconsolidate the role of Gas1 in glycolysis, the impact of Gas1 inthe transcription of key glycolytic enzymes for glucose utilizationwas examined. As expected, GLUT4, HK2, and LDHB exhibited adecreased expression pattern. Consistentwith this observation, anin vivo subcutaneous xenograft mouse model also confirmed the

hypothesis that Gas1 is a negative regulator of glycolysis asreflected by the decreased 18FDG uptake in PET/CT system.Moreover, Gas1 negatively regulated the AMPK/mTOR/p70S6Ksignaling axis, a well-established cascade that regulates malignantcancer cell behaviors including proliferation, metastasis, andaberrant cancer metabolism. In the end, it was determined thatGas1 is a transcriptional target of FOXM1,whose role in colorectalcancer has been widely studied. Taken together, these studiesestablish Gas1 as a negative regulator in colorectal cancer.

Implications: Gas1 suppresses cell proliferation, invasion, andaerobic glycolysis of colorectal cancer both in vitro and in vivo.Mechanistically, Gas1 inhibited EMT and the Warburg effectvia AMPK/mTOR/p70S6K signaling, and Gas1 itself was directlyregulated by the transcription factor FOXM1. Mol Cancer Res; 14(9);830–40. �2016 AACR.

IntroductionColorectal cancer is one of the leadingmalignancies worldwide

and is the third cause of death in cancer patients (1). Althoughsignificant progress has been made in diagnosis and treatment ofcolorectal cancer, invasion, metastasis, and recurrence of thedisease are still challenging (2). Hence, there is an urgent needto better understand the genetic and biological characteristic ofcolorectal cancer, which will improve the efficacy of the treatmentof this disease, including surgical techniques, chemotherapymethods, and follow-up strategies.

Tumor invasion and metastasis are parts of a complicatedprocess inwhich the tumor grows, thendetaches from the primarysite and metastasizes to a distant organ. Previous research has

demonstrated that epithelial–mesenchymal transition (EMT)plays a key role in the early process of the metastasis of cancercells. This process involves the acquisition of the expression ofmesenchymal molecules, such as vimentin and N-cadherin,together with the loss of epithelial cell adhesion molecules suchas E-cadherin (3, 4).

Recent study indicated that metabolic reprogramming playscritical roles during the EMT process, and provides metabolicadvantage for EMT cells (5, 6). Normally differentiated cells relyprimarily on the oxidation of pyruvate in the mitochondria togenerate energy for cellular physiology; however, even with suf-ficient oxygen, rapidly growing cancer cells rely on aerobic gly-colysis to generate energy. This phenomenon is termed as theWarburg effect. The Warburg effect not only provides cancer cellswith ATP and nutrients, but also creates an acidic environmentthat leads to destruction of extracellular matrix and facilitatesmetastasis. Therefore, identifying key players synergistically reg-ulates the metastasis and glycolysis will provide powerful strat-egies in the diagnosis and treatment for colorectal cancer (7).

Aberrations of protein-coding genes, including both oncogenesand tumor suppressive genes have been widely accepted to playcritical roles in process of colorectal cancer. Previously, our studiesshowed that growth arrest–specific protein 1 (Gas1) could con-tribute to predicting metastasis or recurrence in stage II and IIIcolorectal cancer (8). However, the underlying mechanisms thatGas1 contributed to colorectal cancer oncogenesis andmetastasisremain elusive. Hence, in this study, we performed a series of invitro and in vivo studies and demonstrated Gas1 as a tumorsuppressor in colorectal cancer. Mechanistically, Gas1 negativelyregulated the EMT process, and was indicated as a negative

1DepartmentofColorectal Surgery, FudanUniversity ShanghaiCancerCenter, Shanghai, China. 2Department of Oncology, Shanghai MedicalCollege, Fudan University, Shanghai, China. 3Department of Pancre-atic Surgery, Fudan University Shanghai Cancer Center, Shanghai, P.R.China. 4Department of Pathology, Fudan University Shanghai CancerCenter, Shanghai, China.

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

Q. Li and Y. Qin contributed equally to this article.

Corresponding Authors: Sanjun Cai, Fudan University Shanghai Cancer Center,No. 270 Dong'an Road, No. 133 Lingling Road, Shanghai, China. Phone: 8602-1641-75590; Fax: 8602-1641-75590; E-mail: [email protected]; and DaweiLi, [email protected]

doi: 10.1158/1541-7786.MCR-16-0032

�2016 American Association for Cancer Research.

MolecularCancerResearch

Mol Cancer Res; 14(9) September 2016830

on August 5, 2018. © 2016 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from

Published OnlineFirst July 11, 2016; DOI: 10.1158/1541-7786.MCR-16-0032

http://mcr.aacrjournals.org/

regulator of glycolysis. Further clinical and pathologic analysesdemonstrated that Gas1 expression level was negatively associ-ated with SUVmax value reflected by PET/CT imaging, supportingthe notion of Gas1 as a negative regulator in vivo. In the end, wedetermined that Gas1 negatively regulated the AMPK/mTOR/p70S6K signaling axis, and Gas1 itself was a downstream targetof FOXM1, a well-established player in colorectal cancer.

Materials and MethodsPatient information and tissue specimens

For PET/CT and Warburg study, the first study cohort included71 colorectal cancer patients who underwent radical surgerybetween January 2008 and December 2012 at Fudan UniversityShanghai Cancer Center (FUSCC). Preoperative 18F-FDGPET/CTexamination and histopathology confirmation of the presence ofcolorectal adenocarcinoma were conducted in all patients. Thedemographic and clinical characteristics of the patients are sum-marized in Supplementary Table S1.

For tissue microarray (TMA) based IHC study, colon cancertissues were obtained from 185 patients who underwent initialradical surgery, including 24 cases at stage I, 81 at stage II, and80 atstage III. All the patients had a histologic diagnosis of coloncancer. Detailed clinical characteristics of the patients are sum-marized in Supplementary Table S2.

None of these patients included in the study had receivedneoadjuvant therapy. All the subjects involved in this studyprovided written informed consent. This project was approvedby the Ethics Committee of FUSCC.

Construction of the TMA and IHC stainingConstruction of the TMA and IHC staining were performed as

describedpreviously (9, 10).Gas1 andFOXM1anti-human rabbitpolyclonal antibodies were used at a dilution of 1:100(AP11869a; Abgent) and 1:50 (sr-500; Santa Cruz Biotechnolo-gy), respectively; PBS was used alternately as negative control. Allimmunostainings were independently evaluated by two pathol-ogists and a consensus justification based on discussion wasrecorded. For clinicopathologic correlation analysis, we used afour-tiered scoring system (negative to 3þ), which took intoaccount the percentage of positive cells and staining intensity(8, 11). Gas1 was positively stained at cytoplasm and membranewhereas FOXM1was nucleolus stained.We separately interpreted0 and 1þ as "low expression," whereas 2þ and 3þ as "strongexpression."

Cell cultureThe human colon cancer cell lines HT-29, HCT116, SW480,

SW620, RKO, Colo205, Ls174T, and LoVo were originallyobtained from the ATCC. The cells were cultured in RPMI1640medium containing 10% FBS in a humidified 37�C incubatorsupplemented with 5% CO2.

Plasmids and the establishment of stable transfection cell linesGas1 full-length cDNA was cloned from HCT116 cDNA using

primers: 50- ATGGTGGCCGCGCTGCTGGGC -30 (forward prim-er) and 50- CTAAAAGAGCGGCCCAAGCAG -30 (reverse primer).PCR product was cloned into PCMV-N-flag vector, and thenFLAG-tagged Gas1was cloned into pCDH-CMV-MCS-EF1-Purovector to generate pCDH-Gas1 construct. pLKO.1 TRC cloningvector (Plasmid10878; Addgene) was used to generate constructs

expressing shRNAs against Gas1. The 21bp shRNA target againstGas1 were 50-GGGCTGTCTATTAGCATATTT-30 (1#) and 50-GCCATGTATGAAAGTCTC-30 (2#), respectively. pLKO.1-scram-ble shRNA (Plasmid1864; Addgene) with limited homology withany known sequences in the human was used as a negativecontrol. For establishment of stable cells that expressed Flag-Gas1,HCT116 and SW480 were transfected with the pCDH-Gas1expression vector and the control vector. For establishment ofknockdown stable cells, RKO and HT29 were transfected with thepLKO.1-shGas1 expression vector and pLKO.1-scramble. Trans-fected cells were selected using puromycin after the cells weretransfected with expression/knockdown vector or control plas-mids. These stable cells were all used for functional studies asbelow in the text.

Cell proliferation and clonogenic assayCell proliferation was assessed by CCK8 assay as described

previously (10). To determine clonogenic ability, 200 cells weretransplanted in each well of a 6-well dish and allowed to grow for14 days to form colonies. Cells were fixed with methanol andstained with 0.1% crystal violet. All the visible colonies werecalculated manually.

In vitro migration and invasion assaysMigration and invasion assaywere performedusing a Transwell

system (Costar) according to manuals. Chambers were incubatedat 37�C for 24 hours, and three duplicates were prepared for eachgroup. Successfully translocated cells were fixed and then stainedwith 0.2% crystal violet. The total cell numbers of five randomvisual fields were counted, and the average was calculated.

E-cadherin and vimentin immunofluorescenceCells were grown on coverslips, fixed in 4% paraformaldehyde

for 20 minutes, incubated in a blocking buffer (1% BSAand 0.25% Triton X-100 in PBS; pH 7.4), and probed with anE-cadherin antibody or vimentin antibody, then cells were incu-bated with Alexa Flour 594 TgG donkey anti-rabbit (1:500;Invitrogen) for an hour at room temperature. To detect nuclei,cells were costained with DAPI. Fluorescence images were photo-graphed with a confocal microscopy.

Glycolysis analysisGlucose Uptake Colorimetric Assay Kit (Biovision) and Lactate

Colorimetric Assay Kit (Biovision)were purchased to examine theglycolysis process in colon cancer cells according to the manu-facturer's protocol. Real-time PCR was performed to test expres-sion of glycolytic enzymes. All reactions were run in triplicate.

Cell apoptosis rate analysisCells stably transfected with PCDH-Gas1 and Gas1-shRNA

were used for this analysis. For apoptosis rate analysis, cells wereincubated with Annexin V-FITC (BD Biosciences) and propidiumiodide for 10 minutes at room temperature in the dark. Afterstaining, the cells were analyzed using a flow cytometer(CYTOMICS FC 500; Beckman Coulter).

Western blot analysisWestern blotting assay was done as described previously (10).

Briefly, total proteins were isolated by lysing cells in ice-cold RIPAbuffer containing protease and phosphatase inhibitors (Roche).Total proteins were separated by SDS-PAGE gel and blotted onto

Gas1 Expression in Colorectal Cancer

www.aacrjournals.org Mol Cancer Res; 14(9) September 2016 831




polyvinylidene difluoride membranes (Bio-Rad). After blockedwith 5% nonfat milk, the membranes were probed with primaryantibodies, anti-Gas1 rabbit polyclonal antibody (1:1,000 dilu-tion; Abcam), anti-FOXM1 rabbit polyclonal antibody (1:1,000dilution; Santa Cruz Biotechnology), anti-E-cadherin, N-cad-herin, Snail rabbit polyclonal antibody (1:1,000 dilution;Abcam), and anti-Flag mAb (1:10,000 dilution; Clone M2). Afterbeing thoroughly washed, membranes were further incubatedwith corresponding secondary antibodies. Finally, the bandswerevisualized using enhanced chemoluminescence (Pierce; ThermoScientific).

RNA isolation and qRT-PCR analysisTotal RNA from the tissues and cells was extracted using

TRIzol reagent (Invitrogen). RNA quality and concentrationwere determined using the Nanodrop 2000 system (ThermoFisher Scientific). The expression status of target genes andb-actin were determined by qRT-PCR using an ABI 7900HTReal-Time PCR system (Applied Biosystems) using a PowerSYBR Green PCR Master Mix (Invitrogen). All reactions wererun in triplicate. All RT-PCR primers were displayed in Sup-plementary Table S3.

Luciferase assaysFor the luciferase assays, theGas1 promoterwas cloned into the

pGL3 basic vector (Promega). Then, HCT116 and RKO cells (8�103 cells/well) were cultured in 94-well plates and cotransfectedwith the pGL3-Gas1, pGL3-control, or pGL3-Gas1(mutated),pcDNA3.1-FOXM1/pcDNA-control, and Renilla plasmid usingLipofectamine 3000 (Invitrogen). Forty-eight hours after trans-fection, cells were lysed using 20 mL of passive lysis buffer. Next, adual-luciferase assay was carried out as directed by the manufac-turer (Promega). The ratio of firefly to Renilla luciferase activitywas used to express luciferase activities. All experiments wereperformed in triplicate. Data are represented as mean � SD.

Chromatin immunoprecipitation assayHCT116 and RKO cells were prepared for a chromatin immu-

noprecipitation (ChIP) assay using ChIP Assay Kit (Millipore)according to the manufacturer's protocol. The resulting precipi-tated DNA samples were analyzed using PCR to amplify twopotential binding region of the Gas1 promoter with the primers1# 50-GTGGTGATCAAGACCCAAAGACAG-30(forward primer)and 50-TAAGGAGGCTCGGATATGCAGCCC-30 (reverse primer)and 2# 50-GGAGAAAGGAGAAAGCGGGCAGGC-30 (forwardprimer) and 50-TGGCTTCACTCGGCGGCAGCTTC-30 (reverseprimer). The PCR products were resolved electrophoretically ona 1.5% agarose gel and visualized using Goodview staining.

Xenografted nude mice modelTo evaluate in vivo tumorigenesis, five nude mice (male, 4–8

weeks old Balb/C athymic nude mouse) were prepared forHCT116 cells implantation transfected with pCDH-Gas1 orpCDH-vector and RKO cells transfected with PLKO.1-ShGas1 orPLKO.1-scramble. Cells were injected subcutaneously into theright/left forelimbs of nude mice. After 4 weeks, all the injectedmice were euthanatized. Tumor xenografts were harvested andweighted. Tumor volume (TV) was calculated weekly for 4 weeksaccording to the formula: TV (mm3)¼ length�width2� 0.5. Allanimal experiments were performed according to guidelines forthe care and use of laboratory animals and were approved by

Institutional Animal Care and Use Committee of FudanUniversity.

Statistical analysisData were analyzed using SPSS 21.0 statistical package (SPSS).

Based on requirements, either the x2 or Fisher exact test wasapplied to assess the correlations between gene expression andvarious histopathologic features. The Transwell and CCK8 resultswere analyzedbyone-wayANOVAor independent sample t test. AP value < 0.05 was considered statistically significant.

ResultsGas1 inhibited the viability of colon cancer cells in vitro

To assess the role of Gas1 in colon cancer viability andtumorigenic potential, we first examined the endogenousexpression level of Gas1 in six colon cancer cells (Supplemen-tary Fig. S1), and then used lentivirus-mediated overexpressionof Gas1 in HCT116 and SW480 cells (which exhibited thelowest endogenous Gas1 expression) and silencing of Gas1 inHT29 and RKO cells (which exhibited the highest endogenousGas1 expression). Overexpression and knockdown efficiencywere verified by RT-PCR and Western blotting (Fig. 1A and B).The effect of Gas1 on tumor cell growth was measured byCCK-8 assay and the results demonstrated that knockdown ofGas1 significantly enhanced cells proliferation, whereas over-expression of Gas1 decreased the cell viability with statisticalsignificance (P < 0.05; Fig. 1C). Next, we observed a significantincrease in colony formation capacity in cells transfected withGas1-shRNA. Conversely, there was apparent reduction incolony formation ability in pCDH-Gas1–transfected cells(P < 0.05; Fig. 1D). These observations collectively suggest thatGas1 expression may regulate cell cycle or apoptosis on tumorgrowth. Gas1 is a cell-cycle arrest protein. Then, we proceededto use flow cytometer to investigate apoptosis, and found thatknockdown of Gas1 reduced cell apoptosis, whereas overex-pression of Gas1 induced apoptosis (P < 0.05; Fig. 1E), suggest-ing that Gas1 regulated the viability of colon cancer cellthrough both G1 phase arrest and cell apoptosis.

Altered Gas1 expression affected colon cancer cell migrationand invasion in vitro

To assess the role of Gas1 on migration and invasion of coloncancer cells, Transwell migration and invasion assays were per-formed in Gas1 overexpressed or silenced colon cancer cells. Theresults demonstrated that downregulation of Gas1 expressionstrongly promoted themigration of RKO andHT29 cells, whereasforced exogenous expression of Gas1 attenuated the migrationcapacity ofHCT116 and SW480 cells (Fig. 1F). This result was alsoconfirmed by invasion assay (Fig. 1G). Because of low invasionability, SW480 and HT29 were not used for Transwell invasionanalysis.

Gas1 inhibited the EMT process of colon cancer cellsTo determine whether it was EMT that mediated the inhibitory

effect of Gas1 on migration and invasion, we examined theexpression of several EMT markers. As expected, overexpressionofGas1 increased the expression level of E-cadherin anddecreasedthe levels of vimentin, N-cadherin, and Snail, whereas knock-down of Gas1 reduced the expression level of E-cadherin andincreased the levels of vimentin, N-cadherin, and Snail (Fig. 2A).

Li et al.

Mol Cancer Res; 14(9) September 2016 Molecular Cancer Research832




These results were further confirmed by immunofluorescence(Fig. 2B).

Effects ofGas1on aerobic glycolysis in colon cancer cells in vitroIt is well accepted that tumor formation and progression

require glucose metabolism transformation accordingly. Cancercells exhibit a shift of glucose metabolism to less efficient glyco-lytic pathways in response to regional hypoxic stress. Under suchstress, cancer cells rely on glycolysis to fuel its malignant prop-erties (5, 12, 13). To determine the effect of altered Gas1 expres-sion on aerobic glycolysis in colon cancer cells, we calculated theglucose utilization, and lactate concentrations of the stably trans-fected cells. Overexpression of Gas1 strongly decreased the glu-cose utilization, and lactate concentrations in HCT116 andSW480 cells, whereas knockdown of Gas1 expression increasedthe glucose utilization, and lactate concentrations in RKO andHT29 cells (Fig. 3A). Glycolysis is a multistep enzymatic reactioninvolved with a series of rate-limiting enzymes (Fig. 3B). To assessthe effect of Gas1 on the expression of the rate-limiting enzymesthat involved in glycolysis, we carried out qRT-PCR to examine thetranscriptional levels of these enzymes. As shown in Fig. 3C andD,both mRNA and protein levels of GLUT4, HK2, and LDHB were

significantly reduced in Gas1 overexpression cells, whereas theirexpressions were significantly increased in Gas1-silenced cells.Taken together, these results validatedGas1 as a negative regulatorof glycolysis.

Gas1 is negatively associated with cancer cell growth,tumorigenesis, and Warburg effect in vivo

To confirm the suppressive role of Gas1 in cancer cellgrowth, tumorigenesis, and Warburg effect in vivo, we per-formed tumorigenesis assays in nude mice by subcutaneousinjection two paired colon cancer cells, HCT116-vector/HCT116-Gas1 and RKO-Scramble/ShGas1 1# cells. The resultsshowed that tumors derived from HCT116 cells stably expres-sing exogenous Gas1 were significantly smaller and lighterthan tumors derived from control cells (P < 0.05; Fig. 4A1),whereas tumor derived from Gas1 knockdown RKO cells weresignificantly larger and heavier than the corresponding controlgroup (P < 0.05; Fig. 4B1).

High uptake of 18F-FDG by tumors has been suggested to be areflection of the Warburg effect, and the PET/CT imaging systemwas developed on the basis of the glycolysis thesis as a powerfuldiagnostic means. All mice underwent evaluation with 18F-FDG

Figure 1.

Gas1 associates with viability and tumorigenic ability of colon cancer cells. Efficiency of Gas1 overexpression and knockdown in colon cancer cell lines was measuredby Western blot analysis (A) and RT-PCR (B). Influence of Gas1 expression on viability of colon cancer cells was measured by CCK-8 assay (C) and cloneformation (D). E, Gas1 expression–induced cell apoptosis. Gas1 expression could effectively influence cell migration (F), and cell invasion (G). The results areexpressed as the mean � SD of three independent experiments.






PET/CT scan before being sacrificed. The results showed that theSUVmax were higher in HCT116-Vector and RKO-shGas1 groupthan that of their paired group (Fig. 4A2 and B2). IHC stainingalso indicated that the expression levels of GLUT4, HK2, andLDHBwere lower inGas1-overexpressedHCT116 cells (Fig. 4A3).Conversely, therewere higherGLUT4,HK2, and LDHBexpressionlevels in tumor samples derived from Gas1-silenced RKO cells

(Fig. 4B3). These results were also confirmed in a cohort ofpatients with colon cancer. Patients with high Gas1 expressionalways exhibited low SUVmax and low GLUT4, HK2, and LDHBexpression levels, whereas lowGas1 expression groupwas accom-panied with high SUVmax and high GLUT4, HK2, and LDHBexpression. The difference were of statistical significance (Fig.4C and Supplementary Table S1; P < 0.05).

Figure 2.

Gas1 inhibits EMT in colon cancer. A,Western blot analysis of phenotypicmarkers including E-cadherin,N-cadherin, Vimentin, and Snail in Gas1overexpression and knockdown cells.b-Actin was used as control. B,immunofluorescence analysis ofphenotypic marker includingE-cadherin, Vimentin. The red signalrepresents the staining ofcorresponding protein, and the bluesignal represents the nuclear DNAstaining by DAP.

Li et al.





Gas1 inhibits EMT andWarburg effect via AMPK activation andmTOR pathway inhibition

AMPK plays critical roles in inducing EMT and Warburg effect(14–16). Therefore,Western blottingwas performed to determinewhether AMPK, mTOR, and p70s6k were involved in Gas1-mediated EMT and Warburg effect. As expected, overexpressionof Gas1 increased the expression of AMPK, and phosphorylatedAMPK, and inhibited the expression of mTOR, pmTOR, p70S6K,and phosphorylated p70S6K (p-p70S6K) in HCT116 and SW480cells. Conversely, knockdown of Gas1 inhibited AMPK activationand enhanced mTOR, pmTOR, p70S6K, and p-p70S6K expres-sion (Fig. 5).

Gas1 is directly regulated by the transcription factor FOXM1To explore the transcriptional regulation of Gas1 expression in

colorectal cancer, we analyzed transcription factor binding sitesspanning the 2.0 kb upstream of transcription starting site. Two

Forkhead box family transcription factor binding elements(AAACAA) were identified (Fig. 6A). FOXM1 was selected forfurther investigation due to its well-established role in colorectalcancer (17, 18). To validate the hypothesis, we first tested theimpact of FOXM1 on Gas1 transcription in colon cancer cells andindicated FOXM1 as a negative regulator of Gas1 expression (Fig.6B and C). Then, we obtained a luciferase reporter construct(pGL3–Gas1–Luc) containing a segment of the human Gas1promoter and examined the effect of FOXM1 on the promoteractivity. Dual luciferase assay indicated that FOXM1 inhibited theGas1 promoter activity in a dose-dependent manner (Fig. 6D).ChIP assay was performed to demonstrate that FOXM1 occupiedthe promoter region of Gas1 spanning from �415 to �424 (Fig.6E). To confirm that this site mediated the Gas1 response toFOXM1, mutations of the selected sequence were introduced bychanging the sequence (TTTGTTTGTT) to (CCCACCCACC),which completely abolished the putative responsive site. This

Figure 3.

The Gas1 regulates glucose metabolism of colon cancer cells. A, forced expression of Gas1 impaired glycolysis whereas knockdown of Gas1 expressionpromoted glycolysis in colon cancer cells. B, glycolysis is a multistep enzymatic reaction involved with a series of rate limiting enzymes. C and D, overexpressionof Gas1 inhibited the transcriptional level (C) and protein level (D) of key glycolytic enzymes in colon cancer cells (� , ��P < 0.05).






Figure 4.

Gas1 is negatively associated with cancer cell growth, tumorigenesis, andWarburg effect in vivo. A1, B1, representative photographs of animals during the course ofstudy. Colon cancer cellswere paired subcutaneously injected into forelimbs in nudemice. The left sideswereHCT116-Gas1 inA1 andRKO-scramble inB1,whereas theright sides were HCT116-vector in A1 and RKO-shGas1 in B1. Tumors weight and tumor volumes were measured on the indicated days. A2, B2, representativephotographs of PET/CT of animals. All mice inA1 andB1 underwent evaluationwith 18F-FDG PET/CT scan (laying facedown) before sacrificed. The values of SUVmax

were higher in HCT116-vector and RKO-scramble group than paired groups (P < 0.05). A3, B3, immunohistochemical study showed weaker LDHB, GLUT4,HK2, and Ki67 staining in HCT116-Gas1 and RKO-shGas1 groups than their control groups. C, patients with high Gas1 expression always exhibited low SUVmax and lowGLUT4, LDHB, and HK2 expression levels, whereas weak Gas1 expression was accompanied with high SUVmax and high GLUT4, LDHB, and HK2 expression.

Li et al.





mutated version of the Gas1 promoter was cloned into the pGL3-basic-luciferase reporter. A dual-luciferase reporter assay showedthat mutation of the putative FOXM1-binding site in the Gas1promoter completely abolished the FOXM1 responsiveness of theconstruct (Fig. 6F), demonstrating that the 50-TTTGTTTGTT-30 sitewithin the Gas1 promoter mediated the FOXM1 response. TMA-based IHC staining showed a significant reverse association ofFOXM1andGas1 expression in colon cancer patients (Fig. 6G andSupplementary Table S2). Taken together, these findings sug-gested transcription factor FOXM1 as a functional regulator ofGas1 in colon cancer.

DiscussionIt has been reported that metastasis occurs in nearly 50%

colorectal cancer patients after curative colectomy, which is themajor cause of death in colorectal cancer patients (19). Studiesof prognosis for patients with colorectal cancer and of prog-nostic factors to predict the risk of metastasis for individualcolorectal cancer patients are intriguing and could affect clinicalpractice. Established biomarkers such as KRAS, BRAF, and EGFRhave already been proven to play significant roles in prognosisand selection of patients for personalized therapy. As such,there has been a great deal of effort to improve the care ofpatients and understand the biology of colorectal cancer. Ourprevious study has shown that Gas1 may serve as a novelprognostic biomarker involved in the pathogenesis and metas-tasis of colon cancer (8). However, the underlying mechanismof Gas1 still remains elusive. To this end, we sought to identifythe role of Gas1 in tumorigenesis and metastasis and providethe possible mechanism. Accumulating studies have reportedthat Gas1 are low expressed in various cancers and regulatedcell growth arrest and apoptosis (20–23). Two recent studiesreported that Gas1 expression was associated with drug resis-tance in non–small cell lung cancer (24) and gastric cancer(25). Gas1 also was identified as a novel melanoma metastasissuppressor gene (26). In consistent with these observations, weconfirmed that Gas1 could also regulated cell growth and

apoptosis in colorectal cancer. Importantly, we found Gas1strongly correlated with invasion ability of colorectal cancer byinhibiting EMT and Warburg effect.

EMT is considered to be critical for invasive and metastaticprogression in cancer. The process of EMT is associated withthe downregulation of epithelial markers and aberrant upre-gulation of mesenchymal markers. These processes are initi-ated by zinc finger transcriptional repressors such as Snail,which suppresses E-cadherin expression (27). In this study, weprovided evidence that Gas1 was a critical negative regulator ofEMT and thereby inhibited the progression and metastasis ofcolorectal cancer. Knockdown of Gas1 induced downregula-tion of epithelial markers and upregulation of mesenchymalmarkers and Snail. Moreover, overexpression of Gas1 resultedin diminished invasion and metastasis of colon cancer cellsaccompanied with upregulation of epithelial markers anddownregulation of mesenchymal markers. Furthermore, inTMA-based IHC study, we found that Gas1 expression wasnegatively correlated with advanced tumor stage (Supplemen-tary Table S2).

The Warburg effect, a hallmark of cancer cells, has beenhighlighted in recent decades (7). Coding and noncoding genesmay regulate a number of metabolic enzymes, and the aberrantlyexpressed components might provide a growth advantage forcancer cells. The vast majority of studies onmetabolic reprogram-ming have been performed in the setting of neoplastic transfor-mation. Considerably, little is known about metabolic repro-gramming in the context of metastatic transformation. Somestudies have pointed out that metastasis is closely related tometabolism transformation. In triple-negative breast cancer, lossof FBP1 by Snail-mediated epigenetic repression provides meta-bolic advantage for highly metastatic cells (28). However, thecorrelation between metabolism and metastasis in colorectalcancer has seldom been reported. Because of the role of Gas1 inproliferation and metastasis, we questioned whether the impactwas the result of metabolism transformation, because metabo-lism not only provided cancer cells with energy supply and needfor macromolecule synthesis but also created an acidic

Figure 5.

Gas1 regulates AMPK/mTOR/p70s6Ksignal pathway. Overexpression of Gas1increased expression of AMPK andphosphorylated AMPK and inhibitedmTOR, pmTOR, p70S6K, andphosphorylated p70S6K expression inHCT116 and SW480 cells. Conversely,knockdown of Gas1 inhibited AMPKactivation and enhance mTOR, pmTOR,p70S6K, and p-p70S6K expression.






environment that facilitates breakdown of extracellular matrixthat facilitates metastasis. As expected, knockdown of Gas1increased glucose uptake and lactate secretion, whereas overex-

pression of Gas1 decreased glucose uptake and lactate secretion.The altered metabolism induced by Gas1 may be required forcancer cell growth and metastasis.

Figure 6.

Gas1 is directly regulated by the transcription factor FOXM1. A, two related sequence within a 2000 bp amplicon in the Gas1 promoter as a putative responsiveelement able to bind FOXM1 protein. B and C, levels of Gas1, and FOXM1 were determined by Western blot analysis and RT-PCR. HCT116 and RKO cellswere transfected with FOXM1 and an empty vector for 48 hours. FOXM1 could significantly decrease Gas1 expression in both mRNA and protein levels(� , P < 0.05). D, luciferase activity of Gas1-luc construct after transfection of FOXM1 plasmid in HCT116 and RKO cells. Luciferase activities were dose dependentlydecreased after being cotransfected with FOXM1 than with control vector (�, P < 0.05). E, FOXM1 directly bound to the promoter of Gas1 (Site 1#). ChIPassaywas performed in HCT116 andRKO cells transfectedwith a vector expressing FOXM1. RT-PCRwas performed usingprimers specific for site 1# and 2#of theGas1promoter (� , P < 0.05; �� , P > 0.05). F, mutation of the putative FOXM1-binding site in the Gas1 promoter completely abolished the FOXM1 responsivenessof the construct (� ,P>0.05).G,Gas1 expressionwas inversely correlatedwithFOXM1 inpatientswith colorectal cancer. Theexpression levels ofGas1 andFOXM1 in 185patient's samples were determined using IHC. The representative imagines were shown that IHC staining of Gas1 and FOXM1 in colorectal cancer (P < 0.001).

Li et al.





Moreover, we found that Gas1 inhibited EMT and Warburgeffect via AMPK activation andmTOR pathway inhibition. AMPKis a sensor of cellular energy level, which is triggered in conditionsof low intracellular ATP after various stresses such as hypoxia andnutrient deficiency. AMPK stimulation serves as a metaboliccheckpoint to inhibit ATP consuming processes, leading to a stopof cell growth and proliferation (29, 30), and thus reverses EMT(8, 31). Stimulation of AMPKactivity requires phosphorylation ofthe alpha subunit at Thr172 in the activation loop by upstreamkinases (29, 32). The fact that Gas1 can induce increasing levels ofAMPK phosphorylation at Thr172 supported that Gas1 inducesAMPK activity. A large body of evidence has revealed that mTORsignaling is one of the major downstream pathways regulated byAMPK. Recent studies have revealed that AMPK suppressesmTORactivity and the phosphorylation of p70S6K activity (33). Theresults of this study revealed that Gas1 inhibited EMT and War-burg effect through the AMPK/mTOR/P70S6K pathway. Based onthe decisive role of AMPK/mTOR/P70S6K on cell-cycle and nutri-ent sensing (34–36), in combination with our observation of theinhibitory role of Gas1 in proliferation, we speculated that Gas1was connected to the activationofAMPK/mTOR/S6Kaxis andhadan impact on the total and active states of AMPK/mTOR/S6Ksignaling pathway.

Forkhead box proteins are a family of transcription factors thatplay important roles in regulating the expression of genesinvolved in cell growth, proliferation, differentiation, and metas-tasis (37). Recent studies demonstrated that FOXM1 plays animportant role in regulating EMT (17, 18, 38, 39) and Warburgeffect (40). Our data here demonstrated that the Gas1 promoterwas sufficient to allow FOXM1-dependent inactivation of geneexpression in transactivation assays and led to the identificationofa bona fide FOXM1 responsive element within a 424 bp fragmentof the Gas1 promoter, which contained a DNA sequence to whichFOXM1 can directly bind. Negative correlation between FOXM1and Gas1 was found in colorectal cancer cells and tissues. Col-

lectively, our study indicated that Gas1 was transcriptionallyregulated by FOXM1.

In summary, this study provided critical insight into therole of the Gas1 in colorectal cancer progression and identi-fied that Gas1 played important roles by inhibiting bothWarburg effect and EMT processes. Therefore, Gas1 may bea new biomarker and a therapeutic target for the treatment ofcolorectal cancer.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: P. Wei, D. Li, S. CaiDevelopment of methodology: Q. Li, Y. QinAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): Q. Li, Y. Qin, P. Lian, Y. Li, Y. Xu, X. Li, D. Li, S. CaiAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): P. Lian, Y. LiWriting, review, and/or revision of the manuscript: Q. Li, Y. Qin, P. WeiAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): Q. Li, Y. Qin, P. Wei, Y. Xu, X. LiStudy supervision: P. Wei, D. Li, S. Cai

AcknowledgmentsThis research was supported by the National Science Foundation of

China (Nos. 81372646 and 81101586) and National Key Basic ResearchProgram of China (2014CBA02002). The funders had no role in the studydesign, data collection and analysis, decision to publish, or preparation ofthe manuscript.

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicatethis fact.

Received January 28, 2016; revised June 4, 2016; accepted June 24, 2016;published OnlineFirst July 11, 2016.

References1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA Cancer J Clin

2015;65:5–29.2. Zhan J,NiuM,WangP, ZhuX, Li S, Song J, et al. ElevatedHOXB9expression

promotes differentiation and predicts a favourable outcome in colonadenocarcinoma patients. Br J Cancer 2014;111:883–93.

3. Zeisberg M, Neilson EG. Biomarkers for epithelial-mesenchymal transi-tions. J Clin Invest 2009;119:1429–37.

4. Desnoyers G, Frost LD, Courteau L, Wall ML, Lewis SM. Decreased eIF3eexpression can mediate epithelial-to-mesenchymal transition throughactivation of the TGFbeta signaling pathway. Mol Cancer Res 2015;13:1421–30.

5. Cha YH, Yook JI, Kim HS, Kim NH. Catabolic metabolism during cancerEMT. Arch Pharm Res 2015;38:313–20.

6. Kondaveeti Y, Guttilla Reed IK, White BA. Epithelial-mesenchymal tran-sition induces similar metabolic alterations in two independent breastcancer cell lines. Cancer Lett 2015;364:44–58.

7. Vander Heiden MG, Cantley LC, Thompson CB. Understanding the War-burg effect: the metabolic requirements of cell proliferation. Science2009;324:1029–33.

8. Jiang Z, Xu Y, Cai S. Down-regulated GAS1 expression correlates withrecurrence in stage II and III colorectal cancer. HumPathol 2011;42:361–8.

9. Li D, Peng Z, Tang H, Wei P, Kong X, Yan D, et al. KLF4-mediated negativeregulation of IFITM3 expression plays a critical role in colon cancerpathogenesis. Clin Cancer Res 2011;17:3558–68.

10. Li Q, Wu J, Wei P, Xu Y, Zhuo C, Wang Y, et al. Overexpression of forkheadBox C2 promotes tumor metastasis and indicates poor prognosis in colon

cancer via regulating epithelial-mesenchymal transition. Am J Cancer Res2015;5:2022–34.

11. Fromowitz FB, Viola MV, Chao S, Oravez S, Mishriki Y, Finkel G, et al. rasp21 expression in the progression of breast cancer. Hum Pathol 1987;18:1268–75.

12. Moreno-Sanchez R, Rodriguez-Enriquez S, Marin-Hernandez A, Saa-vedra E. Energy metabolism in tumor cells. FEBS J 2007;274:1393–418.

13. Warburg O. On the origin of cancer cells. Science 1956;123:309–14.14. Lee JH, Kim JH, Kim JS, Chang JW, Kim SB, Park JS, et al. AMP-activated

protein kinase inhibits TGF-beta-, angiotensin II-, aldosterone-, highglucose-, and albumin-induced epithelial-mesenchymal transition. Am JPhysiol Renal Physiol 2013;304:F686–97.

15. Chou CC, Lee KH, Lai IL, Wang D, Mo X, Kulp SK, et al. AMPK reverses themesenchymal phenotype of cancer cells by targeting the Akt-MDM2-Foxo3a signaling axis. Cancer Res 2014;74:4783–95.

16. Li X, Lu Y, LuH, Luo J, Hong Y, Fan Z. AMPK-mediated energy homeostasisand associated metabolic effects on cancer cell response and resistance tocetuximab. Oncotarget 2015;6:11507–18.

17. Zhang N, Wei P, Gong A, Chiu WT, Lee HT, Colman H, et al. FoxM1promotes beta-catenin nuclear localization and controls Wnt target-gene expression and glioma tumorigenesis. Cancer Cell 2011;20:427–42.

18. Li D, Wei P, Peng Z, Huang C, Tang H, Jia Z, et al. The critical role ofdysregulated FOXM1-PLAUR signaling in human colon cancer progressionand metastasis. Clin Cancer Res 2013;19:62–72.






19. Van Cutsem E, Nordlinger B, Adam R, Kohne CH, Pozzo C, Poston G, et al.Towards a pan-European consensus on the treatment of patients withcolorectal liver metastases. Eur J Cancer 2006;42:2212–21.

20. Zamorano A, Lamas M, Vergara P, Naranjo JR, Segovia J. Transcriptionallymediated gene targeting of Gas1 to glioma cells elicits growth arrest andapoptosis. J Neurosci Res 2003;71:256–63.

21. Wang H, Zhou X, Zhang Y, Zhu H, Zhao L, Fan L, et al. Growth arrest-specific gene 1 is downregulated and inhibits tumor growth in gastriccancer. FEBS J 2012;279:3652–64.

22. Jimenez A, Lopez-Ornelas A, Estudillo E, Gonzalez-Mariscal L, GonzalezRO, Segovia J. A soluble form of GAS1 inhibits tumor growth andangiogenesis in a triple negative breast cancer model. Exp Cell Res2014;327:307–17.

23. Segovia J, Zarco N. Gas1 is a pleiotropic regulator of cellular functions:from embryonic development tomolecular actions in cancer gene therapy.Mini Rev Med Chem 2014;14:1139–47.

24. Zhang YW, Zheng Y, Wang JZ, Lu XX, Wang Z, Chen LB, et al. Integratedanalysis of DNA methylation and mRNA expression profiling revealscandidate genes associated with cisplatin resistance in non-small cell lungcancer. Epigenetics 2014;9:896–909.

25. Zhao L, Pan Y, Gang Y, Wang H, Jin H, Tie J, et al. Identification of GAS1 asan epirubicin resistance-related gene in human gastric cancer cells with apartially randomized small interfering RNA library. J Biolo Chem 2009;284:26273–85.

26. Gobeil S, Zhu X, Doillon CJ, Green MR. A genome-wide shRNA screenidentifies GAS1 as a novel melanoma metastasis suppressor gene. GenDevel 2008;22:2932–40.

27. Thiery JP, Acloque H, Huang RY, Nieto MA. Epithelial-mesenchymaltransitions in development and disease. Cell 2009;139:871–90.

28. Dong C, Yuan T, Wu Y, Wang Y, Fan TW, Miriyala S, et al. Loss of FBP1 bySnail-mediated repression provides metabolic advantages in basal-likebreast cancer. Cancer Cell 2013;23:316–31.

29. Shaw RJ.LKB1 and AMP-activated protein kinase control of mTOR signal-ling and growth. Acta Physiol 2009;196:65–80.

30. Zadra G, Batista JL, Loda M. Dissecting the dual role of ampk incancer: from experimental to human studies. Mol Cancer Res 2015;13:1059–72.

31. Lin H, Li N, He H, Ying Y, Sunkara S, Luo L, et al. AMPK inhibits thestimulatory effects of TGF-beta on Smad2/3 activity, cell migrationand epithelial to mesenchymal transition. Mol Pharmacol 2015;88:1062–71.

32. Shaw RJ, Kosmatka M, Bardeesy N, Hurley RL, Witters LA, DePinho RA,et al. The tumor suppressor LKB1 kinase directly activates AMP-activatedkinase and regulates apoptosis in response to energy stress. Proc Natl AcadSci U S A 2004;101:3329–35.

33. Inoki K, Zhu T, GuanKL. TSC2mediates cellular energy response to controlcell growth and survival. Cell 2003;115:577–90.

34. Lage R, Dieguez C, Vidal-Puig A, Lopez M. AMPK: a metabolic gaugeregulating whole-body energy homeostasis. Trends Mol Med 2008;14:539–49.

35. Hardie DG, Ross FA, Hawley SA. AMPK: a nutrient and energy sensorthat maintains energy homeostasis. Nat Rev Mol Cell Biol 2012;13:251–62.

36. Kim I, He YY. Targeting the AMP-Activated protein kinase for cancerprevention and therapy. Front Oncol 2013;3:175.

37. Myatt SS, Lam EW. The emerging roles of forkhead box (Fox) proteins incancer. Nat Rev Cancer 2007;7:847–59.

38. Li L, Li Z, KongX, XieD, Jia Z, JiangW, et al.Down-regulationofmicroRNA-494 via loss of SMAD4 increases FOXM1 and beta-catenin signaling inpancreatic ductal adenocarcinoma cells. Gastroenterology 2014;147:485–97 e18.

39. Huang C, Xie D, Cui J, Li Q, Gao Y, Xie K. FOXM1c promotes pancreaticcancer epithelial-to-mesenchymal transition and metastasis via upregula-tion of expression of the urokinase plasminogen activator system. ClinCancer Res 2014;20:1477–88.

40. Cui J, Shi M, Xie D, Wei D, Jia Z, Zheng S, et al. FOXM1 promotes thewarburg effect and pancreatic cancer progression via transactivation ofLDHA expression. Clin Cancer Res 2014;20:2595–606.


Li et al.




2016;14:830-840. Published OnlineFirst July 11, 2016.Mol Cancer Res Qingguo Li, Yi Qin, Ping Wei, et al. Carcinoma

Inhibits Metastatic and Metabolic Phenotypes in ColorectalGas1

Updated version

10.1158/1541-7786.MCR-16-0032doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2016/08/31/1541-7786.MCR-16-0032.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/14/9/830.full#ref-list-1

This article cites 40 articles, 12 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/14/9/830.full#related-urls

This article has been cited by 1 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mcr.aacrjournals.org/content/14/9/830To request permission to re-use all or part of this article, use this link



http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-16-0032

http://mcr.aacrjournals.org/content/suppl/2016/08/31/1541-7786.MCR-16-0032.DC1

http://mcr.aacrjournals.org/content/14/9/830.full#ref-list-1

http://mcr.aacrjournals.org/content/14/9/830.full#related-urls

http://mcr.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mcr.aacrjournals.org/content/14/9/830


gas1 inhibitsmetastaticandmetabolicphenotypes in...

Documents