Angiogenesis, Metastasis, and the Cellular...

Published OnlineFirst February 9, 2010; DOI: 10.1158/1541-7786.MCR-09-0326

Angiogenesis, Metastasis, and the Cellular Microenvironment Molecular
Cancer
Research
Myeloid Zinc Finger 1 Induces Migration, Invasion, and In vivoMetastasis through Axl Gene Expression in Solid Cancer
Giridhar Mudduluru1, Peter Vajkoczy2, and Heike Allgayer1

Abstract

Authors' AMolecularHeidelbergCharite Un

Note: Suppcer Resear

CorresponSurgery/MoCenter-DKUniversity H2226; Fax:heidelberg.

doi: 10.115

©2010 Am

www.aacr

Dow

Myeloid zinc finger 1 (MZF1) is a member of the SCAN domain family transcription factors that formdimers through their highly conserved SCAN motifs. Silencing of MZF1 inhibits cell proliferation, and abnor-mal expression of MZF1 results in cancer development. However, a potential role of MZF1 in metastasis re-mains unclear. Axl is a receptor tyrosine kinase and was first identified as a transforming gene in chronic myeloidleukemia. Axl overexpression induces proliferation, migration, and invasion and is highly expressed in differenthuman cancers. In this study, we show that overexpression of MZF1 induces migration and invasion in colo-rectal (Rko, SW480) and cervical (HeLa) cancer cells. In addition, we show that MZF1 binds to the Axl pro-moter, transactivates promoter activity, and enhances Axl-mRNA and protein expression in a dose-dependentmanner. In vitro, sh-RNA knockdown of Axl reduced MZF1-induced migration and invasion in HeLa and Rkocells (P = 0.05). Additionally, Rko cells overexpressing MZF1 showed increased tumor formation and liver me-tastasis in the chicken-embryo-metastasis assay in vivo. Furthermore, the expression of MZF1 and Axl was sig-nificantly higher in resected colorectal tumors compared with corresponding normal tissues (P = 0.02; P = 0.05),and MZF1 expression was positively correlated with Axl gene expression in tumor tissues (P < 0.01). Takentogether, this is the first study to show that MZF1 induces invasion and in vivo metastasis in colorectal andcervical cancer, at least in part by regulating Axl gene expression. Mol Cancer Res; 8(2); 159–69. ©2010 AACR.

Introduction

The myeloid zinc finger gene-1, MZF1, encodes a tran-scription factor belonging to the krüppel family of zinc fin-ger proteins. MZF1 contains 13 C2H2 zinc fingersarranged in two domains that are separated by a short gly-cine- and proline-rich sequence. Both of these DNA-bind-ing domains (domain 1, zinc finger 1-4; domain 2, zincfinger 5-13) select similar DNA-binding consensus se-quences, containing a core of four or five guanine residues:1-4, 5′-AGTGGGGA-3′; 5-13, 5′-CGGGnGAGGGG-GAA-3′ (1). Transcription of the gene leads to at leastthree transcripts of approximately 3, 7.5, and 9 kb. Allthree isoforms have two DNA-binding domains that inter-act with two different consensus sequences. Additionally,MZF1b and c encode a unique 257 amino acid amino ter-minus containing a SCAN box, or leucine-rich domain,

ffiliations: 1Department of Experimental Surgery Mannheim/Oncology of Solid Tumors, DKFZ and University of, Heidelberg, Germany and 2Department of Neurosurgery,iversitätsmedizin Berlin, University of Berlin, Berlin, Germany

lementary data for this article are available at Molecular Can-ch Online (http://mcr.aacrjournals.org/).

ding Author: Heike Allgayer, Department of Experimentallecular Oncology of Solid Tumors (German Cancer ResearchFZ-Heidelberg), Mannheim Medical Faculty, Ruprecht-Karls-eidelberg, 68167 Mannheim, Germany. Phone: 49-621-383-49-621-383-3809/-1938. E-mail: [email protected]

8/1541-7786.MCR-09-0326

erican Association for Cancer Research.

journals.org

on June 30, 2018. mcr.aacrjournals.org nloaded from

which facilitates protein interactions (2). Differential ex-pression of MZF-1 interferes with hematopoietic cell dif-ferentiation and proliferation (3-5) and is associated withthe antiapoptotic, oncogenic transformation of cell lines(6, 7). MZF1 transcriptionally regulates protein kinaseC α in human hepatocellular carcinoma (8), which is animportant family of signaling molecules that regulate cellproliferation, differentiation, transformation, and apoptosis.MZF1 antisense oligonucleotide–pretreated SK-Hep-1 hu-man hepatocellular carcinoma cells showed reduced proteinkinase C expression, and antisense-treated cells s.c. im-planted into nude mice exhibited lower tumor growth andprolonged formation time (9). However, an effect of MZF1on invasion and metastasis, and associated target genes, havenot yet been studied conclusively.Axl receptor tyrosine kinase was originally described as a

transforming gene isolated from chronic myeloid leukemiapatients (10). It belongs to the TAM (Tyro3, Axl, Mer) re-ceptor tyrosine kinase family, sharing Gas6 as a commonligand, the product of growth arrest–specific gene 6 (11).Gas6 binds to TAM family members with different affin-ities. Gas6 was first discovered as an upregulated gene ingrowth-arrested cells (12), and further studies with Gas6signaling suggested its role in cell survival (13, 14), prolif-eration (15, 16), stimulation of cell migration (17), andcell-cell adhesion through Axl (18). Intracellular signalingof Axl is also activated by its homophilic and heterophilicinteractions (19), mediated mainly by a multisubstratedocking site (20). Overexpression of Axl can transform fi-broblasts even in the absence of a ligand (21), this is at leastin part being mediated by Gas6/Axl-induced activation of

159

© 2010 American Association for Cancer Research.

http://mcr.aacrjournals.org/

Mudduluru et al.

160


mitogen-activated protein kinase/extracellular signal-regu-lated kinase and phosphoinositide 3-kinase signaling. Fol-lowing its identification in myeloid leukemia, Axloverexpression has been reported in several types of humansolid cancers (22-27). Axl expression correlates with poorprognosis in acute myeloid leukemia (28), gastric (29), andadvanced lung adenocarcinoma patients (27), and with ad-hesion, motility, and invasiveness of osteosarcoma cells(30). Transfection of dominant-negative Axl lacking the ki-nase domain into Axl-expressing glioblastoma cells reducesmotility and the loss of cell-to-cell interaction (31). Axl isrequired for normal migration of GnRH neurons from theolfactory placode to the hypothalamus (32), and stableshRNA knockdown of Axl reduced blood vessel formationand functional circulation in a mouse model of angiogen-esis (33).The Axl gene has a GC-rich promoter, lacking TATA

and CAAT boxes and harboring multiple transcriptionalstart sites, these including the major one at −169, and asecond major one at −460 bp upstream of the translationstart site. Axl is constitutively regulated by Sp1/Sp3 tran-scription factors and further controlled by CpG islandmethylation within the Sp motifs and their flanking re-gions (34). Interestingly, however, this GC-rich promoteradditionally harbors multiple MZF1 binding motifs.Therefore, the aim of the present study was to investigate(a) if MZF1 is able to induce migration, invasion, and me-tastasis in vitro and in vivo in solid cancers; (b) if Axl geneexpression is regulated by MZF1; (c) if MZF1-induced mi-gration and invasion is mediated through Axl; and (d) ifthis is of biological relevance in resected patient tumors.The results here presented are the first to show thatMZF1 induces migration, invasion, and Axl gene expressionin colorectal and cervical cancer cells; that it induces tumorformation and liver metastasis in vivo; and that MZF1 andAxl are being coexpressed in colorectal cancer tissues.

Experimental Procedures

Cell Culture and ReagentsThe human cervical cancer cell line HeLa and the colon

cancer cell lines Rko and SW480 were purchased from theAmerican Type Culture Collection and cultured in MEMEagle, MEM, and L-15 medium, respectively, containing10% heat-inactivated fetal bovine serum, at 37°C in an at-mosphere of 5% CO2.

Cell Migration and Invasion AssayHeLa, Rko, and SW480 cells were transfected either

with mock or MZF1 expression vector. After 24 h, cellswere trypsinized, and 3 × 105 cells were plated on Boydenchambers either coated with 10 μg Matrigel per well (forinvasion assays) or uncoated (for in vitro migration assays)in serum-free medium containing 0.1% bovine serumalbumin. Ten percent fetal bovine serum medium in thelower chamber served as chemoattractant. After 20 h, non-invading cells were removed with cotton swabs. Invadedcells were trypsinized and counted using the ATP-lumines-

Mol Cancer Res; 8(2) February 2010

on June 30, 2018. mcr.aacrjournals.org Downloaded from

cence–based motility invasion assay (Promega) as previous-ly described by Leupold JH et al. (35).

RNA Isolation and Reverse Transcriptase-PCRTotal RNA was extracted by the RNeasy Midi kit (Qia-

gen) according to the manufacturer's instructions. Tenfresh frozen human colorectal tumor and correspondingnormal tissues were obtained from the tumor banks ofthe department, which is approved by the institutionalethics committee. Tissues were homogenized and totalRNA was isolated using Trizol (Invitrogen) according tothe manufacturer's protocol. One microgram of totalRNA (either from cells or tissues) was reversely transcribedby random hexamer primer using SuperScript II reversetranscriptase (Invitrogen). The single-strand cDNA wasamplified by PCR using specific primer pairs with standardconditions. Glycer-aldehyde-3-phosphate dehydrogenasewas used as an internal control. The PCR products wereanalyzed by electrophoresis on a 2% ethidium bromide–stained agarose gel. The sequences of the sense (S) andantisense (AS) primers used for either semiquantitative, orSybr green quantitative, PCR were as follows: MZF1 (S,5′-AGTGTAAGCCCTCACCTCC-3′, and AS, 5′-GGGTCCTGTTCACTCCTCAG-3′), glycer-aldehyde-3-phosphate dehydrogenase (S, 5′-GTCTTCACCACCATG-GAGAA-3′, and AS, 5′-ATCCACAGTCTTCTGGGT-GG-3′), and Axl (S, 5′-GGTGGCTGTGAAGACGA-TGA-3′, and AS, 5′-CTCAGATACTCCATGCCACT-3′).

Preparation of Protein Extracts and ImmunoblottingProtein lysates were prepared as previously described by

Mudduluru et al. (34). Briefly, for immunoblotting, sam-ples (40 μg/lane) were boiled for 5 min, separated throughSDS-PAGE, and transferred to polyvinylidene difluoridemembranes. After transfer, the membranes were blockedwith 5% nonfat milk containing TBS with 0.1% Tweenfor 3 h at room temperature, and then probed with prima-ry antibodies against either MZF1 (gift from Dr. DorothyTuan, the Department of Biochemistry and Molecular Bi-ology, Medical College of Georgia, Augusta, GA; ref. 36),Axl (#Santa cruz-1096), or Actin (#Santa cruz-1616-R)overnight at 4°C. After three washes with TBS with0.1% Tween, blots were incubated with secondary antibo-dies conjugated with horseradish peroxidase. After finalwashes with TBS with 0.1% Tween, the membranes wereexposed to film after enhanced chemiluminescence (Amer-sham Biosciences).

Generation of Luciferase Reporter Constructs andExpression VectorsLuciferase reporter constructs were generated as de-

scribed in Mudduluru et al. (34). Expression vectors con-taining human MZF1 (37) were gifts from Dr. Tara Sander(Department of Surgery, Children's Research InstituteMedical College of WI). Sh-RNA vectors of sh-Axl#RHS4430-98903492 (Mature Sense sequence: GATCC-TAAGCATCTAAGTT) and nonsilencing control(#RHS4346) were purchased from Open Biosystems.

Molecular Cancer Research



MZF1 Induces Axl Expression in Colorectal Cancer Cells


Site-Directed Mutation AnalysisMutant constructs were generated by the QuikChange

XL site-directed mutagenesis kit from Stratagene usingAxl1 (−1,276 to +7) as a template. Successful incorporationof the mutations was confirmed by automated sequencing.

Transfection and Dual-Luciferase AssaysFor luciferase assays, 0.13 × 106 Rko or 0.05 × 106 HeLa

cells were plated in 24 -well plates. Luciferase reporter plas-mids (500 ng of each) with or without MZF1-expressingplasmids were transfected in regular medium using Effec-tine (Qiagen). For all luciferase reporter assays, pRL-TK(50 ng, Renilla Luciferase; Promega) was cotransfectedand luminescence was measured to normalize transfectionefficiency. Cells were washed twice with PBS and lysedwith 100 μL passive lysis buffer (Promega) for 20 min.Twenty microliters of cell lysate were measured throughthe Dual-Luciferase reporter assay system (Promega) ac-cording to the manufacturer's instructions. Luciferase as-says were performed after 24 h of transfection. Axl(−1,276 to +7) mock-treated luciferase activity was set to100%, and activity of the other constructs was calculatedand plotted as a percentage. Assays for all samples weredone in triplicate, the error bars representing the SD.

Preparation of Nuclear Extracts and ElectrophoreticMobility Shift AssayPreparation of nuclear extracts and electrophoretic mo-

bility shift assays (EMSA) was done as previously described(34). In brief, nuclear extracts (5 μg) were incubated in abuffer [25 mmol/L HEPES (pH 7.9), 0.5 mmol/L EDTA,0.5 mmol/L DTT, 0.05 mol/L NaCl, and 4% (v/v) glyc-erol] with 50,000 cpm of the [33P] ATP-end–labeled(phage T4 polynucleotide kinase, ICN #151935) oligonu-cleotide containing region −578/−538 of the Axl promoterfor 20 min in the absence, or presence, of a 50-fold excessof wild-type, or 50-/100-fold excess of mutated, unlabeledoligonucleotides to show specificity of the binding at roomtemperature. A total of 0.5 μg of polydeoxyinosinic-deox-ycytidylic acid was present in each reaction to block unspe-cific binding. The reactions were subjected to gelelectrophoresis on a 5% polyacrylamide gel containing5% glycerol in 0.25× Tris borate-EDTA buffer. For super-shift analysis, 2 or 4 μg of MZF1 antibody were added tothe reactions 20 min after nuclear extracts and oligonucle-otide had been incubated. Supershift reactions were incu-bated for 60 min at 4°C and electrophoresed at 4°C toensure complex stability. The gels were dried and analyzedby autoradiography.

Chorioallantoic Membrane AssayThe chorioallantoic membrane CAM assay was done as

previously described (38, 39). Tumors were weighed usingelectronic balance and the average of tumor weight was cal-culated among groups. Genomic DNA from embryonic li-vers was prepared using the Puregene DNA purificationsystem. Quantification of human cells and DNA copy

www.aacrjournals.org


number, and Fluorogenic Taqman QPCR, was done as de-scribed by Leupold et al. (35).

Statistical AnalysisStatistical analysis was done using the SPSS version 14.0

(SPSS). Student's t test was done to determine the statisti-cally significant differences between groups. P values of≤0.05 were considered significant.

Results

MZF1 Induces Migration and InvasionTo determine if MZF1 regulates migration and invasion

in HeLa, Rko, and SW480, cells (showing low or no endog-enous expression of MZF1, and no methylation of the Axlpromoter; ref. 34) were either mock transfected withpCDNA3-vector or with pCDNA3-MZF1, respectively.After 24 h, cells were either studied in Boyden chambersfor migration or in 10 μg Matrigel-coated chambers forinvasion. MZF1-expressing cells showed significantly in-creased migration and invasion (P < 0.05) in three cells lineswhen compared with mock vector–expressing cells (Fig. 1Aand B). Real-time PCR analysis was done, determining anoverexpression of MZF1 (Supplementary Fig. S1A-C), andthe results confirmed a significant overexpression in MZF1-transfected cells.

MZF1 Transactivates Axl Promoter Activity andInduces Axl Gene ExpressionA TRANSFAC database search on the Axl promoter re-

vealed that it harbors multiple putative MZF1 bindingmotifs (Fig. 2A). To determine the ability of MZF1 to reg-ulate Axl promoter activity, MZF1 and a luciferase reporterdriven by the −1,276/+7 upstream region of the Axl gene,which reflected maximum Axl promoter activity (34), werecotransfected into HeLa and Rko cells. The peptidyl argi-nine deiminase type I [PADI 1 (-195 to +84 bp)] promoterwas transfected as a positive control (40), which is regulat-ed by MZF1 (Supplementary Fig. S2A). MZF1 transacti-vated the Axl promoter in a dose-dependent manner inboth of the cell lines (Fig. 2B). To show that MZF1-in-duced Axl promoter activity results in an increase of Axltranscript and protein, HeLa cells were transfected withincreasing concentrations of MZF1-expressing vector.After 48 hours, Axl mRNA and protein were found in-duced when compared with mock controls (Fig. 2C andD). Taken together, these data indicate that MZF1 trans-activates the Axl promoter and positively regulates Axl geneexpression.

MZF1 Binding Motifs Regulate Axl Promoter ActivityTo analyze the functional relevance of the predicted pu-

tative MZF1 binding motifs for Axl promoter activity,three individual MZF1 binding motifs at −563/−557 bp,−479/−473 bp, −80/−72 bp, and combinations of all threeof these motifs were mutated. The Sp1 binding motif re-ported previously (34) also harbors a putative MZF1 motif

Mol Cancer Res; 8(2) February 2010 161



Mudduluru et al.

162


at −512/−504 bp; however, we have reported that this mo-tif regulates promoter activity explicitly through the bind-ing of Sp1 family members (34). Hence, the particularmotif was not included in this study. Mutation of all threeof the MZF1 motifs tested reduced MZF1-inducible Axlpromoter activity (Fig. 3A). In particular, the MZF1 mu-tation at −479/−473 bp most significantly reduced consti-tutive, but especially MZF1-induced Axl promoter activity(Fig. 3A), and other combined mutations did not furtherreduce either constitutive or MZF1-induced promoteractivity compared with this mutation (data not shown;Fig. 3A). These results suggest that the MZF1 bindingmotif at −479/−473 bp is the most important motif forAxl transactivation upon MZF1 overexpression.

MZF1 Nuclear Proteins Bind to the Axl PromoterTo show that the MZF1-induced Axl gene expression is

dependent on its binding to the promoter, gel shift andsupershift analysis were done using nuclear extracts eitherwith mock or MZF1-transfected HeLa cells. EMSA anal-ysis was done with oligonucleotides corresponding to theMZF1 binding motif at −563/−557 bp (100% consensus



motif ) and the functionally most relevant −479/−473 bp(93% consensus) motif of the Axl promoter. To determinethe specific binding of MZF1 to these motifs, a 50-foldexcess of unlabeled wild-type, or a 50- and 100-fold excessof an oligo nucleotide mutated for the binding of MZF1,were used for cold competition. As shown in Fig. 3B, thecompetition with wild-type oligo nucleotides completelyabolished the specific binding of transcription factors,whereas with the mutated oligos as competitors, the forma-tion of specific complexes stayed intact. To show the spe-cific binding of MZF1 to this motifs, a comparativesupershift analysis with an MZF1-specific antibody wasdone with nuclear extracts of HeLa cells made to overex-press MZF1. Overexpression of MZF1 intensified thebinding of especially one specific complex observed inthe EMSA (Fig. 3C, down arrows). With MZF1 antibody,a supershift of this specific band was seen (Fig. 3C, up ar-rows). The MZF1 binding motif of the ERCC1 promoter(−324/−298 bp) was used as a positive control (41), whichshowed a significant increase in MZF1 binding withMZF1-overexpressing nuclear extracts of HeLa cells (Sup-plementary Fig. S2B). These results suggest that MZF1

FIGURE 1. Overexpression of MZF1 induces migration and invasion. A and B, migration and invasion assay after overexpression of MZF1 in HeLa,Rko, and SW480 cells. Twenty-four hours after transfection, cells were plated on Boyden chambers with Matrigel (10 ng/well) for invasion and withoutMatrigel for migration assay. Cells were collected after 20 h from the top and migrated/invaded cells from the bottom of the chambers and were countedusing the ATP luminescence–based motility invasion assay. Mock-treated cells served as control. The experiment was done thrice independently, withtriplicates each time. Columns, mean of one of the three independent experiments; bars, SD. *, statistically significant changes. P < 0.05.






binds to the major MZF1 consensus motifs of the Axlpromoter.

Axl Mediates MZF1-Induced Migration and InvasionTo show that Axl is one of the main mediators of

MZF1-induced migration and invasion, sh-RNA expres-sing constructs targeting Axl, or nonsilencing constructsas controls, were cotransfected with the MZF1 expressionconstruct into either HeLa or Rko cells. After 24 h, cellswere either studied in Boyden chambers for migration orin Matrigel-coated chambers for invasion. As expected,MZF1- and nonsilencing vector-expressing cells showedinduced migration (Fig. 4A) and invasion (Fig. 4B) inboth of the cell lines when compared with mock-trans-fected cells. However, Axl-sh-RNA– and MZF1-expres-sing cells showed significantly less migration (Fig. 4A)and invasion (Fig. 4B) when compared with MZF1-and nonsilencing vector-expressing cells; under similarconditions, they could not change cell proliferation signif-



icantly (data not shown). This was paralleled by the ob-servation that MZF1 overexpression induced theexpression of Axl protein in the both of the cell lines(Fig. 4C and D). In addition, Axl mRNA was inducedby MZF1 in HeLa, Rko, and SW480 cells as expected(see supplementary Fig. S1). These results suggested thatMZF1-induced migration and invasion is at least in partmediated through Axl.

MZF1 Enhances Tumor Formation and MetastasisIn vivoOur previous results suggested that MZF1 induces

migration and invasion in vitro. To study the effect ofMZF1 overexpression on in vivo tumor formation andmetastasis, we used the chicken-embryo-metastasis assay(CAM), which is able to differentiate primary tumorgrowth, intravasation, and metastasis in a highly sensitiveand quantitative in vivo setting (35). Rko cells trans-fected with pCDNA3-vector, or with pCDNA3-MZF1,

FIGURE 2. Overexpression of MZF1 induces Axl gene expression. A, schematic representation of the 5′ flanking region of Axl promoter with putative Sp1and MZF1 binding sites. Forward arrow, the Major (MTSS) and the second major (SMTSS) transcription start sites; +1, the translation start site. B, 500 ngof Axl (−1,276 to +7) luciferase reporter construct and MZF1 (500 ng) expressing plasmid were cotransfected along with 50 ng of a Renilla luciferaseplasmid. The firefly luciferase activity was normalized with Renilla luciferase activity. Columns, mean of three independent experiments done inquadruplicates; bars, SD. C, real-time PCR analysis of MZF1 and Axl upon overexpression of MZF-1 in HeLa cells. D, Western blot analysis of MZF1, Axl,and β-Actin (loading control) upon MZF1 overexpression in HeLa cells. Relative protein expression of MZF1 and Axl obtained by densitometric analysis isrepresented as bar graphs. Specific MZF1 or Axl band intensities were normalized with β-Actin. Data represents one of three independent experiments.




Mudduluru et al.

164


were inoculated on the upper CAM of 10-day old chick-en embryos, and chicken embryos were grown for 7more days. On day 17, the in vivo primary tumors wereresected from the upper CAM and were weighed.MZF1-transfected cells developed significantly higher tu-mor weights than vector-transfected cells in this model(P = 0.014; Fig. 5A). At the same time, livers were har-vested from the embryos, and the number of metasta-sized cel ls quantif ied with real-t ime Alu-PCR.Consequently, MZF1-transfected cells showed signifi-



cantly higher metastasis than vector-transfected cells inthe liver (Table 1; P = 0.05; Fig. 5C). These two observa-tions clearly implicate the ability of MZF1 to induce tumorformation and metastasis in vivo.

Axl Is Coexpressed with MZF1 in Resected ColorectalTumorsFinally, to support the in vivo relevance of the me-

chanisms identified, 10 resected colorectal carcinoma tis-sues of patients and corresponding normal mucosae were

FIGURE 3. MZF1 binds and transactivates the Axl promoter region. A, Axl (−1,276/+7), the promoter construct showing maximum activity, was mutatedat diverse MZF1 binding motifs as indicated. One microgram of wild-type/mutant Axl promoter constructs along with Renilla luciferase plasmid weretransfected into HeLa. Axl wild-type promoter activity was set to 100% (first bar), and activity of the mutated and/or MZF1 cotransfected constructs wasplotted as a percentage of this value. The most significant reduction of constitutive and MZF1-induced activity was seen with the mutation at −479/−473 bp,compared with wild-type activity. All experiments were done in triplicate and the results were averaged and plotted as shown. Columns, mean; bars,SD; *, P < 0.05. B, EMSA was done with HeLa nuclear extract and a γ-p33–labeled oligonucleotide corresponding to the MZF1 binding motifs at−563/−557 bp and −479/−473 bp of the Axl promoter. For competition, a 50-fold excess of unlabeled wild-type or a 50- and 100-fold excess of mutantoligos preventing MZF1 binding was used. Bottom, the wild-type and a mutated oligonucleotide sequence. Specific MZF1 complexes are indicatedwith parenthesis. C, supershift analysis done for the same MZF1 binding motifs (−564/−558 bp and −479/−473 bp) with mock and MZF1-overexpressingHeLa nuclear extracts. Two and 4 μg of MZF1 antibody were used for supershift analysis; IgG served as negative control. Up arrows and parenthesis,specific supershifted complexes. Down arrows, the specific disappearance of the original band due to the antibody binding. Experiments were done thrice.Data represents one of the independent experiments.






investigated for endogenous expression of MZF1 and Axl.We observed that 8 patients of 10 showed an increase ofMZF1 gene expression in resected tumors compared withnormal tissues (Fig. 6A) and 7 patients of 10 also showed ahigher Axl gene expression in tumors than normal tissues(Fig. 6B). The differences between the mean Axl andMZF1 gene expression were significant (P = 0.05, P =0.02) between resected tumor and corresponding normaltissue (Fig. 6C). Furthermore, MZF1 gene expressionwas positively and significantly correlated with Axl gene ex-pression in the resected tumor tissues (Rs = 0.87; P < 0.01;Fig. 6D).

Discussion

This is the first comprehensive study to show that anoverexpression of MZF1 transcription factor leads to thetransactivation of Axl promoter activity and gene expres-sion, resulting in an enhanced migratory, invasive, andin vivo metastatic potential of colorectal and cervicalcancer cells. The in vivo relevance of this mechanismis supported by the observation that the expression of



MZF1 and Axl is significantly increased in resected co-lorectal carcinomas compared with normal tissues, andthat the observed expression of both correlates signifi-cantly in resected colorectal tumors. Our findings sug-gest that the induction of Axl, invasion, and metastasisby MZF1 is playing an important role in tumor forma-tion and progression in solid cancers.MZF1 was originally identified as one of the tran-

scriptional regulators during hemopoietic development(42). In later studies, it was found that overexpressionof MZF1 inhibits apoptosis and promotes oncogenesis(6, 7), and silencing of MZF1 with antisense oligonu-cleotides was shown to reduce tumor growth (9). How-ever, the mechanisms by which MZF1 exerts anoncogenic potential are unknown except that it can reg-ulate protein kinase C α expression (8). Moreover, datasuggesting a tumorigenic potential of MZF1 have beencontroversial (43). In the present study, we confirm thatoverexpression of MZF1 is tumorigenic, and for the firsttime, we report that MZF1 induces tumor growth andmetastasis through the transactivation of Axl, thus add-ing to the molecular targets and mechanisms that can

FIGURE 4. Axl mediates MZF1-induced migration and invasion. A and B, HeLa or Rko cells were cotransfected with the vectors, as indicated, and after24 h, cells were seeded with and without Matrigel in quadruplicates for invasion and migration assays, respectively. Mock-treated cells served as acontrol. C and D, overexpression of MZF1, sh-RNA knockdown of Axl, and MZF1-induced Axl expression were confirmed by Western blot analysis. β-Actinserved as internal control. Relative protein expression of MZF1 and Axl obtained by densitometric analysis is represented as bar graphs. Specific MZF1or Axl band intensities were normalized β-actin. Values are represented as relative percentages (mock-treated samples served as 100%). The experimentwas done thrice with triplicates each time. Columns, mean of one of the independent experiments; bars, SD. **, P < 0.05; *, P = 0.05.




Mudduluru et al.

166


mediate an oncogenic potential of MZF1 in epithelial-derived tissues. Axl has been reported as a transforminggene (21) that is overexpressed in several tumors (22-27), and the Gas6/Axl signaling pathway is known toinduce cell proliferation, antiapoptosis, migration, inva-sion, and angiogenic processes, which are most likelymediated through Ras, Src, mitogen-activated protein ki-nase/extracellular signal-regulated kinase, phosphoinosi-tide 3-kinase/Akt, and NF-κB signaling pathways (14,15, 19, 29, 31, 32, 44-49). However, still further studiesare needed to enhance the understanding of the down-stream components of the Gas6/Axl signaling axis in tu-mor formation. It is known that an overexpression ofthe Axl gene can lead to the activation of Axl by autop-hosphorylation (50), this in turn being able to activatethe Axl signaling axis. This is the most likely mechanismto explain our observation that upon MZF1 stimulation,Axl is significantly overexpressed, this enhancing migra-tion and invasion. For the latter, the presence of Axl hasbeen shown to be essential in our present study becausean sh-RNA knockdown strategy on Axl reduced MZF1-inducible migration and invasion considerably. Thein vivo relevance of MZF1-induced Axl overexpression,this hypothetically leading to Axl autoactivation and ac-tivation of the Axl signaling axis, is supported by theobservation that MZF1 is able to induce tumor growthand metastasis in the CAM in vivo model, and especiallyalso by the observation in resected patient tissues thatprimary colorectal cancers express significantly higheramounts of endogenous MZF1 than corresponding nor-mal tissues, this being significantly associated with a par-allel increase of endogenous Axl gene expression.MZF1 has been initially described as a transcription

factor important for the development of cells of the he-mopoietic origin. In addition, correspondingly, many he-mopoietic cell–specific genes such as CD34, lactoferrin,



and myeloperoxidase are regulated by MZF1 (4, 51,52), whereas very few genes of cell types of different or-igin (e.g., keratinocytes, osteoblasts) such as PADI 1 andN-cadherin are transactived by MZF1 (40, 53). Axl isexpressed in certain leukemia-type cells, but also in di-verse solid tumors (22-27), and excitingly, we found thatthe Axl gene promoter harbors multiple MZF1 motifs.Among them, mutation of the MZF1 motif at −479/−473 bp reduced promoter activity most significantly,and additional combination mutations did not showconsiderably more reduction of activity, showing thatthis motif is the most relevant for Axl transactivationby MZF1. Furthermore, promoter analysis indicated thatthe upstream region between −1,276 to +7 of the Axlgene, which we have characterized previously as themaximum Axl promoter (34), containing all putativeMZF1 binding motifs, showed induced promoter activityupon MZF1 overexpression in cultured colorectal andcervical cancer. Deletion of MZF1 binding consensus

©

Table 1. Number of mock and MZF1 cells

Mock (no. of cells)

Molecul

2010 American Association for Ca

MZF1 (no. of cells)

644 ± 59
438 ± 35 335 ± 39 452 ± 31 348 ± 90 461 ± 59 12 ± 3 727 ± 49
936 ± 129
498 ± 94 25 ± 7 637 ± 32
376 ± 92
598 ± 28 331 ± 42 689 ± 30 335 ± 06
FIGURE 5. MZF1 increases the potential of Rko cells to be tumorigenic and metastatic in vivo. Either mock- or MZF1-transfected Rko cells (2 × 106)were inoculated on the upper CAM of chicken eggs (n = 10). The tumor weights were measured and DNAs of the livers were extracted 7 d after inoculation.A, average of tumor weights from MZF1-transfected compared with mock vector-transfected cells. B, two tumors from two independent eggs fromeach group (mock- and MZF1-transfected samples) are represented. The scale is indicated in centimeters. C, the genomic DNA was analyzed by Real-timeAlu-PCR to determine the number of metastasized Rko cells in livers. The individual cell number per egg is given in the adjacent box. The CAM assaywas done twice with independent transfections. Columns, mean of two independent experiments; bars, SD. **, P < 0.014; *, P = 0.05.

ar Cancer Research

ncer Research.




sequences in the −1,276 to +7 Axl promoter luciferaseconstruct did not show a significant reduction of spon-taneous constitutive promoter activity, which is not sur-prising because we have already shown that constitutiveAxl gene expression is regulated by GC-rich zinc fingerbinding transcription factors Sp1/Sp3 (34). Therefore,from the present study, it can be concluded that over-expressed MZF1 binds to the Axl promoter, inducingAxl promoter activity and gene expression.Taken together, this is the first comprehensive analy-

sis of the ability of MZF1 to induce migration, inva-sion, tumor formation, and metastasis in vivo incultured colorectal and cervical cancer. Our results sug-gest that MZF1 plays an important role in inducingthese phenomena through the transactivation of Axl insolid tumors. Therefore, further studies aiming at a bet-ter understanding of the pathways that lead to an in-duced MZF1/Axl gene expression in solid tumors andalso the screening of a larger series of patients are clear-ly warranted.



Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Erika Hillerich, Dr. Kumaraswamy Regalla, Paolo Ceppi, AlexandraKörner, and Dr. Laura D. Nelson for the excellent help and critical appraisal of themanuscript.

Grant Support

H. Allgayer was supported by Alfried Krupp von Bohlen und Halbach Founda-tion (Award for Young Full Professors), Essen, Hella-Bühler-Foundation, Heidel-berg, Dr. Ingrid zu Solms Foundation, Frankfurt/Main, the Hector Foundation,Weinheim, Germany, the FRONTIER Excellence Initiative of the University ofHeidelberg, and the Walter Schulz Foundation, Munich, Germany.

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

Received 7/29/09; revised 12/11/09; accepted 12/30/09; published OnlineFirst2/9/10.

References
1. Morris JF, Hromas R, Rauscher FJ III. Characterization of the DNA-
binding properties of the myeloid zinc finger protein MZF1: two inde-pendent DNA-binding domains recognize two DNA consensus se-quences with a common G-rich core. Mol Cell Biol 1994;14:1786–95.

2. Sander TL, Haas AL, Peterson MJ, Morris JF. Identification of a novelSCAN box-related protein that interacts with MZF1B. The leucine-rich SCAN box mediates hetero- and homoprotein associations.J Biol Chem 2000;275:12857–67.

FIGURE 6. MZF1 and Axl expression are significantly correlated in resected colorectal tissues. Total RNA was extracted from fresh-frozen humancolorectal cancer samples using Trizol. A and B, real-time PCR results comparing MZF1 and Axl expression in tumor compared with normal tissues.The expression levels of MZF-1 and Axl in normal tissues were normalized to 1. C, differences between normal and tumor tissues in all 10 patients analyzedfor average MZF1/Axl expression (*, P ≤ 0.05; Mann-Whitney test). D, Spearman correlation analysis shows a significant positive correlation betweenMZF1 expression and Axl gene expression in tumor tissues; Rs, Spearman correlation cofficent.




Mudduluru et al.

168


3. Bavisotto L, Kaushansky K, Lin N, Hromas R. Antisense oligonucleo-tides from the stage-specific myeloid zinc finger gene MZF-1 inhibitgranulopoiesis in vitro. J Exp Med 1991;174:1097–101.

4. Perrotti D, Melotti P, Skorski T, Casella I, Peschle C, Calabretta B.Overexpression of the zinc finger protein MZF1 inhibits hematopoi-etic development from embryonic stem cells: correlation with nega-tive regulation of CD34 and c-myb promoter activity. Mol Cell Biol1995;15:6075–87.

5. Robertson KA, Hill DP, Kelley MR, et al. The myeloid zinc finger gene(MZF-1) delays retinoic acid-induced apoptosis and differentiation inmyeloid leukemia cells. Leukemia 1998;12:690–8.

6. Hromas R, Morris J, Cornetta K, et al. Aberrant expression of themyeloid zinc finger gene, MZF-1, is oncogenic. Cancer Res 1995;55:3610–4.

7. Hromas R, Boswell S, Shen RN, et al. Forced over-expression of themyeloid zinc finger gene MZF-1 inhibits apoptosis and promotes on-cogenesis in interleukin-3-dependent FDCP.1 cells. Leukemia 1996;10:1049–50.

8. Hsieh YH, Wu TT, Tsai JH, Huang CY, Hsieh YS, Liu JY. PKCα ex-pression regulated by Elk-1 and MZF-1 in human HCC cells. Bio-chem Biophys Res Commun 2006;339:217–25.

9. Hsieh YH, Wu TT, Huang CY, Hsieh YS, Liu JY. Suppression of tu-morigenicity of human hepatocellular carcinoma cells by antisenseoligonucleotide MZF-1. Chin J Physiol 2007;50:9–15.

10. Janssen JW, Schulz AS, Steenvoorden AC, et al. A novel putativetyrosine kinase receptor with oncogenic potential. Oncogene 1991;6:2113–20.

11. Nagata K, Ohashi K, Nakano T, et al. Identification of the productof growth arrest-specific gene 6 as a common ligand for Axl,Sky, and Mer receptor tyrosine kinases. J Biol Chem 1996;271:30022–7.

12. Schneider C, King RM, Philipson L. Genes specifically expressed atgrowth arrest of mammalian cells. Cell 1988;54:787–93.

13. Melaragno MG, Cavet ME, Yan C, et al. Gas6 inhibits apoptosis invascular smooth muscle: role of Axl kinase and Akt. J Mol Cell Car-diol 2004;37:881–7.

14. van Ginkel PR, Gee RL, Shearer RL, et al. Expression of the receptortyrosine kinase Axl promotes ocular melanoma cell survival. CancerRes 2004;64:128–34.

15. Stenhoff J, Dahlbäck B, Hafizi S. Vitamin K-dependent Gas6 acti-vates ERK kinase and stimulates growth of cardiac fibroblasts. Bio-chem Biophys Res Commun 2004;319:871–8.

16. Sainaghi PP, Castello L, Bergamasco L, Galletti M, Bellosta P, AvanziGC. Gas6 induces proliferation in prostate carcinoma cell lines ex-pressing the Axl receptor. J Cell Physiol 2005;204:36–44.

17. Fridell YW, Villa J, Jr., Attar EC, Liu ET. GAS6 induces Axl-mediatedchemotaxis of vascular smooth muscle cells. J Biol Chem 1998;273:7123–6.

18. McCloskey P, Fridell YW, Attar E, et al. GAS6 mediates adhesion ofcells expressing the receptor tyrosine kinase Axl. J Biol Chem 1997;272:23285–91.

19. Budagian V, Bulanova E, Orinska Z, et al. A promiscuous liaison be-tween IL-15 receptor and Axl receptor tyrosine kinase in cell deathcontrol. EMBO J 2005;24:4260–70.

20. Braunger J, Schleithoff L, Schulz AS, et al. Intracellular signaling ofthe Ufo/Axl receptor tyrosine kinase is mediated mainly by a multi-substrate docking-site. Oncogene 1997;14:2619–31.

21. Burchert A, Attar EC, McCloskey P, Fridell YW, Liu ET. Determinantsfor transformation induced by the Axl receptor tyrosine kinase. On-cogene 1998;16:3177–87.

22. Craven RJ, Xu LH, Weiner TM, et al. Receptor tyrosine ki-nases expressed in metastatic colon cancer. Int J Cancer1995;60:791–7.

23. Berclaz G, Altermatt HJ, Rohrbach V, Kieffer I, Dreher E, AndresAC. Estrogen dependent expression of the receptor tyrosine kinaseaxl in normal and malignant human breast. Ann Oncol 2001;12:819–24.

24. Meric F, Lee WP, Sahin A, Zhang H, Kung HJ, Hung MC. Expressionprofile of tyrosine kinases in breast cancer. Clin Cancer Res 2002;8:361–7.

25. Hutterer M, Knyazev P, Abate A, et al. Axl and growth arrest specific



gene 6 are frequently overexpressed in human gliomas and predictpoor prognosis in patients with glioblastoma multiforme. Clin CancerRes 2008;14:130–8.

26. Tsou AP, Wu KM, Tsen TY, et al. Parallel hybridization analysis ofmultiple protein kinase genes: identification of gene expression pat-terns characteristic of human hepatocellular carcinoma. Genomics1998;50:331–40.

27. Shieh YS, Lai CY, Kao YR, et al. Expression of axl in lung adenocar-cinoma and correlation with tumor progression. Neoplasia 2005;7:1058–64.

28. Rochlitz C, Lohri A, Bacchi M, et al. Axl expression is associated withadverse prognosis and with expression of Bcl-2 and CD34 in denovo acute myeloid leukemia (AML): results from a multicenter trialof the Swiss Group for Clinical Cancer Research (SAKK). Leukemia1999;13:1352–8.

29. Wu CW, Li AF, Chi CW, et al. Clinical significance of AXL kinase fam-ily in gastric cancer. Anticancer Res 2002;22:1071–8.

30. Nakano T, Tani M, Ishibashi Y, et al. Biological properties and geneexpression associated with metastatic potential of human osteosar-coma. Clin Exp Metastasis 2003;20:665–74.

31. Vajkoczy P, Knyazev P, Kunkel A, et al. Dominant-negative inhibitionof the Axl receptor tyrosine kinase suppresses brain tumor cellgrowth and invasion and prolongs survival. Proc Natl Acad Sci U SA 2006;103:5799–804.

32. Allen MP, Linseman DA, Udo H, et al. Novel mechanism for gonad-otropin-releasing hormone neuronal migration involving Gas6/Arksignaling to p38 mitogen-activated protein kinase. Mol Cell Biol2002;22:599–613.

33. Holland SJ, Powell MJ, Franci C, et al. Multiple roles for the receptortyrosine kinase axl in tumor formation. Cancer Res 2005;65:9294–303.

34. Mudduluru G, Allgayer H. The human receptor tyrosine kinase Axlgene—promoter characterization and regulation of constitutive ex-pression by Sp1, Sp3 and CpG methylation. Biosci Rep 2008;28:161–76.

35. Leupold JH, Asangani I, Maurer GD, Lengyel E, Post S, Allgayer H.Src induces urokinase receptor gene expression and invasion/intra-vasation via activator protein-1/p-c-Jun in colorectal cancer. MolCancer Res 2007;5:485–96.

36. Yu X, Zhu X, Pi W, et al. The long terminal repeat (LTR) of ERV-9human endogenous retrovirus binds to NF-Y in the assembly of anactive LTR enhancer complex NF-Y/MZF1/GATA-2. J Biol Chem2005;280:35184–94.

37. Sander TL, Morris JF. Characterization of the SCAN box encodingRAZ1 gene: analysis of cDNA transcripts, expression, and cellularlocalization. Gene 2002;296:53–64.

38. Asangani IA, Rasheed SA, Nikolova DA, et al. MicroRNA-21 (miR-21)post-transcriptionally downregulates tumor suppressor Pdcd4 andstimulates invasion, intravasation and metastasis in colorectal can-cer. Oncogene 2008;27:2128–36.

39. Nikolova DA, Asangani IA, Nelson LD, et al. Cetuximab attenuatesmetastasis and u-PAR expression in non-small cell lung cancer: u-PAR and E-cadherin are novel biomarkers of cetuximab sensitivity.Cancer Res 2009;69:2461–70.

40. Dong S, Ying S, Kojima T, et al. Crucial roles of MZF1 and Sp1 in thetranscriptional regulation of the peptidylarginine deiminase type Igene (PADI1) in human keratinocytes. J Invest Dermatol 2008;128:549–57.

41. Yan QW, Reed E, Zhong XS, Thornton K, Guo Y, Yu JJ. MZF1 pos-sesses a repressively regulatory function in ERCC1 expression. Bio-chem Pharmacol 2006;71:761–71.

42. Hromas R, Collins SJ, Hickstein D, et al. A retinoic acid-responsivehuman zinc finger gene, MZF-1, preferentially expressed in myeloidcells. J Biol Chem 1991;266:14183–7.

43. Gaboli M, Kotsi PA, Gurrieri C, et al. Mzf1 controls cell proliferationand tumorigenesis. Genes Dev 2001;15:1625–30.

44. Goruppi S, Ruaro E, Varnum B, Schneider C. Gas6-mediated survivalin NIH3T3 cells activates stress signalling cascade and is indepen-dent of Ras. Oncogene 1999;18:4224–36.

45. Bellosta P, Zhang Q, Goff SP, Basilico C. Signaling through the ARKtyrosine kinase receptor protects from apoptosis in the absence ofgrowth stimulation. Oncogene 1997;15:2387–97.






46. Fridell YW, Jin Y, Quilliam LA, et al. Differential activation of the Ras/extracellular-signal-regulated protein kinase pathway is responsiblefor the biological consequences induced by the Axl receptor tyrosinekinase. Mol Cell Biol 1996;16:135–45.

47. Demarchi F, Verardo R, Varnum B, Brancolini C, Schneider C. Gas6anti-apoptotic signaling requires NF-κB activation. J Biol Chem2001;276:31738–44.

48. Korshunov VA, Mohan AM, Georger MA, Berk BC. Axl, a receptortyrosine kinase, mediates flow-induced vascular remodeling. CircRes 2006;98:1446–52.

49. Nielsen-Preiss SM, Allen MP, Xu M, et al. Adhesion-related ki-nase induction of migration requires phosphatidylinositol-3-ki-nase and ras stimulation of rac activity in immortalized



gonadotropin-releasing hormone neuronal cells. Endocrinology2007;148:2806–14.

50. Bellosta P, Costa M, Lin DA, Basilico C. The receptor tyrosine kinaseARK mediates cell aggregation by homophilic binding. Mol Cell Biol1995;15:614–25.

51. Hromas R, Davis B, Rauscher FJ III, et al. Hematopoietic transcrip-tional regulation by the myeloid zinc finger gene, MZF-1. Curr TopMicrobiol Immunol 1996;211:159–64.

52. Morris JF, Rauscher FJ III, Davis B, et al. The myeloid zinc finger gene,MZF-1, regulates the CD34 promoter in vitro. Blood 1995;86:3640–7.

53. Le Mée S, Fromigué O, Marie PJ. Sp1/Sp3 and the myeloid zinc fin-ger gene MZF1 regulate the human N-cadherin promoter in osteo-blasts. Exp Cell Res 2005;302:129–42.




2010;8:159-169. Published OnlineFirst February 9, 2010.Mol Cancer Res Giridhar Mudduluru, Peter Vajkoczy and Heike Allgayer Metastasis through Axl Gene Expression in Solid Cancer

In vivoMyeloid Zinc Finger 1 Induces Migration, Invasion, and

Updated version

10.1158/1541-7786.MCR-09-0326doi:

Access the most recent version of this article at:

Material

Supplementary

http://mcr.aacrjournals.org/content/suppl/2010/02/09/1541-7786.MCR-09-0326.DC1

Access the most recent supplemental material at:

Cited articles

http://mcr.aacrjournals.org/content/8/2/159.full#ref-list-1

This article cites 53 articles, 25 of which you can access for free at:

Citing articles

http://mcr.aacrjournals.org/content/8/2/159.full#related-urls

This article has been cited by 8 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] at

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

Permissions

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

.http://mcr.aacrjournals.org/content/8/2/159To request permission to re-use all or part of this article, use this link

on June 30, 2018. © 2010 American Association for Cancer Research. mcr.aacrjournals.org Downloaded from


http://mcr.aacrjournals.org/lookup/doi/10.1158/1541-7786.MCR-09-0326

http://mcr.aacrjournals.org/content/suppl/2010/02/09/1541-7786.MCR-09-0326.DC1

http://mcr.aacrjournals.org/content/8/2/159.full#ref-list-1

http://mcr.aacrjournals.org/content/8/2/159.full#related-urls

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

mailto:[email protected]

http://mcr.aacrjournals.org/content/8/2/159


Angiogenesis, Metastasis, and the Cellular...

Documents

Transcript of Angiogenesis, Metastasis, and the Cellular...