Oncogene (2010) 29, 4205–4215 & 2010 Macmillan Publishers Limited All rights reserved 0950-9232/10 www.nature.com/onc ORIGINAL ARTICLE The RNA-binding zinc-finger tristetraprolin regulates AU-rich mRNAs involved in breast cancer-related processes

N Al-Souhibani1, W Al-Ahmadi1, JE Hesketh2, PJ Blackshear3 and KSA Khabar1

1Program in BioMolecular Research, King Faisal Specialist Hospital and Research Center, Riyadh, Saudi Arabia; 2Institute for Cell and Molecular Biosciences, Newcastle University, Newcastle upon Tyne, UK and 3Laboratory of Signal Transduction, NIEHS, National Institutes of Health, Research Triangle, NC, USA

Tristetraprolin (TTP or ZFP36) is a tandem CCCH zinc- Introduction finger RNA-binding protein that regulates the stability of certain AU-rich element (ARE) mRNAs. Recent work Breast cancer is the most common type of malignant suggests that TTP is deficient in cancer cells when cancer among women, with a high incidence and compared with normal cell types. In this study we found mortality rate, and comprises almost a fifth of all female that TTP expression was lower in invasive breast cancer cancers (McPherson et al., 2000). Cancer metastasis is cells (MDAMB231) compared with normal breast cell dependent on the tumor’s ability to degrade components lines MCF12A and MCF-10. TTP targets were probed of the extracellular matrix by different proteolytic using a novel approach by expressing the C124R zinc- enzymes (Liotta et al., 1980; Liotta, 1986; Bacac and finger TTP mutant that functions as dominant negative Stamenkovic, 2008). Alterations in the expression of and increases target mRNA expression. In contrast to many have been implicated in the invasiveness and wild-type TTP, C124R TTP was able to increase certain metastatic potential of malignant breast cancers (Liotta ARE-mRNA expressions in serum-stimulated breast et al., 1980; Liotta, 1986). Several products of the cancer cells. Using an ARE-gene microarray, novel AU-rich element mRNA (ARE-mRNA) family are targets of TTP regulation were identified, namely, overexpressed in cancer cells and such increased expres- urokinase plasminogen activator (uPA), uPA receptor sion correlates with increased invasion and metastasis. and matrix metalloproteinase-1, all known to have These include cyclooxygenase 2 (Soslow et al., 2000), prominent roles in breast cancer invasion and metastasis. urokinase plasminogen activator (uPA) (Nanbu et al., Expression of these targets was upregulated in tumori- 1994), uPA receptor (uPAR) (Roldan et al., 1990) and genic types, particularly in highly invasive MDAMB231. matrix metalloproteinase 1 (MMP1) (Fini et al., 1987). The mRNA half-lives of these TTP-regulated genes were The expression of ARE-mRNA gene products is increased in TTP-knockout embryonic mouse fibroblasts, regulated by RNA-binding , which are trans- as assessed using real-time polymerase chain reaction, acting factors that bind to AREs and regulate the whereas forced restoration of TTP by transfection led to a stability of ARE-mRNAs. Among those is tristetrapro- reduction in their mRNA levels. RNA immunoprecipita- lin (TTP), an RNA-binding protein that is rapidly and tion confirmed an association of TTP, but not C124R, transiently expressed in response to extracellular stimuli with these target transcripts. Moreover, TTP reduced, such as serum, growth factors, phorbol esters and whereas the mutant C124R TTP increased, the activity of insulin (DuBois et al., 1990; Lai et al., 1990) and binds reporter constructs fused to target ARE. As a result of to and destabilizes mRNAs that contain AREs in their TTP regulation, invasiveness of MDAMB231 cells was 30 untranslated regions (UTRs) (Carballo et al., 1998). reduced. The data suggest that TTP, in a 30 untranslated TTP is a tandem zinc-finger protein and consists of two region—and ARE-dependent manner, regulates an zinc-fingers that are necessary for TTP binding to important subset of cancer-related genes that are involved tumor necrosis factor (TNF)-a 30UTR, a major target in cellular growth, invasion and metastasis. for TTP (Lai et al., 2000). Mutation of either zinc-finger Oncogene (2010) 29, 4205–4215; doi:10.1038/onc.2010.168; by a single-point mutation (Cys to Arg), for example, published online 24 May 2010 C124R mutant, fails to bind TNF-a 30UTR and ARE (Lai et al., 1999). Well-established targets of TTP Keywords: posttranscriptional control; AU-rich elements; involved in cancer have been identified, namely, RNA-binding proteins; RNA stability; tristetraprolin cyclooxygenase 2 (Sawaoka et al., 2003; Young et al., 2009) and vascular endothelial growth factor (Essafi- Benkhadir et al., 2007). Recently, we have shown that a Correspondence: Dr KSA Khabar, Program in BioMolecular polyadenylation transcript variant of HuR is an ARE- Research, King Faisal Specialist Hospital and Research Center, mRNA to which wild-type TTP, but not the zinc-finger P3354, Riyadh 11211, Saudi Arabia. mutant (C124R), competes with HuR for binding to E-mail: [email protected] Received 19 November 2009; revised 3 April 2010; accepted 9 April 2010; HuR mRNA itself (Al-Ahmadi et al., 2009a). Further, published online 24 May 2010 TTP mRNA targets have been recently identified, TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4206 including immediate early response 3 (Lai et al., 2006) and interleukin (IL)-10 (Stoecklin et al., 2008). MCF10A ER- TTP mRNA and protein levels were recently found to

be significantly decreased in tumors of the thyroid, lung, ER+ ** ANOVA ovary, uterus and breast compared with nontransformed MCF12A tissue samples (Brennan et al., 2009). Furthermore, loss + of TTP expression was observed in several adenomas Cell line MCF7 ER and adenocarcinomas and correlates with increased

expression of HuR and cyclooxygenase 2 (Young et al., Tumor Normal 2009). As a result, TTP dysfunction may lead to MDAMB231 ** ER- abnormalities that contribute to cancer processes. Our aim was to investigate the role of TTP in global 0

regulation of ARE-gene expression and identify novel 20,000 40,000 60,000 80,000 100,000120,000140,000 TTP-regulated genes in the context of cancer, particu- Relative TTP expression (arbitrary units) larly breast cancer. Observations that the TTP zinc- finger mutant C124R can lead to an increase in TNF-a 1,500,000 mRNA (Lai et al., 2002) and HuR mRNA expression Control (Al-Ahmadi et al., 2009a) in an ARE-dependent manner TTP 1,250,000 C124R promoted us to further explore this dominant-negative feature. Thus, we used C124R as a biological probe for 1,000,000 finding a functional ARE-mRNA repertoire using microarray experiments. We showed that TTP binds to and destabilizes important cancer-related transcripts, 750,000 namely, MMP1, uPA and uPAR transcripts, in a 30UTR and ARE-dependent manner. We confirmed these 500,000 targets using different experimental approaches. We also showed that TTP can suppress the invasive ARE-GFP Fluorescence 250,000 potential of breast cancer cells. 0 0 25 50 75 100 125 150 175 200 plasmid concentration (ng) Results HEK293 MDAMB231

TTP mRNA expression in normal and tumor breast cell TTP lines To compare the expression of endogenous TTP mRNA in normal versus tumor breast cell lines, four different β-actin mammary cell lines were used. Estrogen receptor (ER)

status was considered when selecting cell lines: one normal CC TTP CC TTP C12R C12R and one tumor cell line were chosen that were ER positive Vector Vector (MCF12A and MCF-7, respectively) and similarly one Figure 1 (a) TTP expression profile in normal and tumor breast normal and one tumor cell line that did not express ER at cell lines. Cells were serum starved (0.5% fetal bovine serum) detectable levels, namely, MCF10A and MDAMB231. overnight, then restimulated with 10% serum for 1 h. Total RNA was then extracted and reverse transcribed for quantitative PCR. The latter is well known for its increased invasiveness. Taqman expression assays specific for human TTP were performed To induce the expression of TTP mRNA, serum- and quantitation was normalized to human RPLP0 as the starved cell lines were stimulated with serum for 1 and endogenous control in the real-time PCR experiments. Values 2 h. TTP mRNA expression in all four cell lines showed shown are means±s.e.m. from three independent experiments, a pattern that is typical of TTP, in which mRNA **Po0.01 as determined by ANOVA. ER: estrogen receptor. (b) Response of ARE-GFP reporter in HEK293 cells transfected accumulation peaked at 1 h and rapidly decreased to with 25 ng ARE-GFP reporter plasmid and cotransfected with near-basal levels at 2 h (data not shown). A comparison various amounts of TTP, C124R (mutant TTP) or a vector plasmid of TTP mRNA expression levels at 1 h was carried out alone. GFP fluorescence was measured 24 h after transfection. (c) for the four cell lines (Figure 1a). Normal mammary cell Level of TTP protein in untransfected and transfected cells after serum induction. A total of 40 mg of total protein was loaded onto lines, particularly MCF10A, exhibited a high expression sodium dodecyl sulfate-PAGE gel and probed with an antibody of TTP mRNA, when compared with the highly invasive against TTP or b-actin as a loading control and transfection MDAMB231 cell line (Figure 1a). The ER-negative, efficiency was determined by measurement of cotransfected GFP. nontumorigenic MCF10A cell line showed the highest expression; it was 10-fold higher than MDAMB231 cells that exhibited the least TTP expression of all four cell varying mobilities were observed in both cell lines, lines (Figure 1a). Similarly, TTP protein expression was suggesting different phosphorylation states for TTP. much higher in MCF10A cells than in MDAMB231 When comparing tumor cell lines, there was a large cells (Figure 2d, upper panel). Several bands with difference (5.6-fold, P ¼ 0.02) between the noninvasive

Oncogene TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4207 ** ** 100,000 ** 100,000 * ** uPA * MMP1 10,000 10,000

1,000 1,000

100 100 relative uPA mRNA expression

10 relative MMP1 mRNA expression 10

MCF7 MCF7 MCF12A MCF10A MCF12A MCF10A MDAMB231 MDAMB231

100,000 ** * uPAR * MCF10 MDAMB231 50 kDa 10,000 TTP 36 kDa

uPAR 1,000 50 kDa

uPA 47 kDa

relative uPAR mRNA expression 100 40 kDa MCF7 ACTB MCF12A MCF10A MDAMB231 Figure 2 mRNA and protein expression profile of TTP-regulated candidate targets in normal and tumor breast cell lines. Real-time PCR experiments were performed using Taqman expression assays specific for human uPA (a), uPAR (b) and MMP1 (c), and normalized to human RPLPO as the endogenous control. Results are means±s.e.m. of three experiments. *Po0.05, **Po0.005. (d) Protein expression levels of TTP, uPA and uPAR in the ER-negative breast cell lines MCF10A (normal) and MDAMB231 (tumor). A volume of 60 mg of total protein was loaded onto sodium dodecyl sulfate-PAGE gel and probed with antibodies against human TTP, uPA, uPAR or b-actin as a loading control.

MCF-7 and the highly invasive MDAMB231 (Figure 1a). nonbinding mutant C124R exerts a dominant-negative These results suggest that TTP is deficient to a high degree effect on ARE-mRNA transcripts that results in their in the highly invasive breast tumor cell line. increased stabilization (Lai et al., 2002), and, as a result, overexpression of the mutant C124R TTP construct was Expression of wild-type and mutant TTP used subsequently in microarray experiments in an and identification of TTP-regulated genes attempt to more accurately probe ARE-mRNAs signals To investigate the gene expression profile effect of the that are potentially destabilized by TTP but may not be restoration of deficient TTP in the highly invasive breast evident because of very low signal intensities in the gene cell line, MDAMB231, we first determined the optimal expression microarray. The ARE-complementary DNA functional doses of expression levels of recombinant microarray (Khabar et al., 2004; Al-Ahmadi et al., wild-type and C124R mutant TTP inferred from a cell- 2009b) was used with RNA from HEK293 and based assay with EGFP reporter fused to ARE-30UTR MDAMB231 cells overexpressing recombinant TTP (Figure 1b). Despite the stability of green-fluorescent and the mutant TTP. There were no detectable basal protein (GFP), which might mask the actual decrease in levels of TTP in these tumor cell lines. Transfection of the ARE-GFP reporter brought on by TTP destabiliza- TTP or its mutant led to the restoration and expression tion of its mRNA, the effect of wild-type TTP on of TTP proteins (Figure 1c). The C124R mutant did destabilization of the ARE-GFP reporter was readily indeed function as dominant negative and increased the detectable and maximal at lower amounts with a 50% gene expression of a subset of ARE-mRNAs (Table 1) in reduction in fluorescence at 25 ng compared with which a proportion of genes (approximately one-third) control, non-ARE GFP. In contrast, the C124R mutant were reduced by wild-type TTP (Table 1, bold text). showed a fourfold increase in reporter activity using Several transcripts on the list seemed to be stabilized by the same amount of plasmid. The TTP zinc-finger wild-type TTP. A great majority of these transcripts

Oncogene TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4208 Table 1 Genes regulated by TTP in cancer cells Gene symbol Name Experiment/control ratio

C124R TTP s.e.m. WT TTP s.e.m.

C4BPA Complement component 4 binding protein a 9.14 2.69 5.3 1.00 TCF7L2 Transcription factor 7-like 2 (T-cell specific, HMG-box) 6.06 2.99 2.18 0.39 ELAVL2 ELAV (embryonic lethal, abnormal vision, Drosophila)-like 2 5.36 2.23 2.53 0.47 TRAF1 TNF receptor-associated factor 1 4.36 1.07 0.44 0.04 PLAUR Plasminogen activator, urokinase receptor 4.34 1.70 0.55 0.02 BTN3A3 Butyrophilin, subfamily 3, member A3 3.99 1.10 1.7 0.15 GATA3 GATA binding protein 3 3.96 1.39 2 0.36 PHLDA1 Pleckstrin homology-like domain, family A, member 1 3.55 1.41 2.16 0.54 NR3C1 Nuclear receptor subfamily 3, group C, member 1 3.44 0.39 1.9 0.29 SUMO1 Ubiquitin-like 1 (sentrin) 3.26 0.61 2.17 0.49 NLGN1 Neuroligin 1 3.24 0.36 2 0.19 PPAP2B Phosphatidic acid phosphatase type 2B 2.96 0.39 2.03 0.40 IGFBP6 Insulin-like growth factor binding protein 6 2.92 0.59 2.3 0.55 IL1A Interleukin 1, a 2.84 0.46 1.4 0.22 CCL2 Chemokine (C–C motif) ligand 2 2.83 0.83 1.6 0.30 BHMT Betaine-homocysteine methyltransferase 2.74 0.34 1.7 0.30 CSF1 Colony stimulating factor 1 (macrophage) 2.66 0.69 0.86 0.04 PLAU Plasminogen activator, urokinase 2.61 0.21 0.52 0.04 SERINC3 Tumor differentially expressed 1 2.57 0.37 0.73 0.06 MEF2C MADS box transcription enhancer factor 2, polypeptide C 2.49 0.43 1.1 0.15 BHLHE41 Basic helix-loop-helix domain containing, class B, 3 2.49 0.38 1.7 0.14 JUN V-jun avian sarcoma virus 17 oncogene homolog 2.43 0.14 0.61 0.06 CXCR4 Chemokine (C–X–C motif), receptor 4 (fusin) 2.42 0.24 1.8 0.18 NFKBIA Nuclear factor of j light polypeptide gene enhancer in B-cells inhibitor, a 2.32 0.23 0.79 0.03 PROSC Proline synthetase co-transcribed (bacterial homolog) 2.29 0.16 1.2 0.21 IRF1 Interferon regulatory factor 1 2.28 0.23 0.61 0.03 MMP1 Matrix metalloproteinase 1 (interstitial collagenase) 2.04 0.07 0.67 0.04 SRP54 Signal recognition particle 54 kDa 2.03 0.05 1.7 0.42 IL10 Interleukin 10 1.70 0.18 0.49 0.01 COL3A1 Collagen, type III, a 1 1.55 0.15 0.37 0.03 KIF21A Kinesin family member 21A 1.30 0.09 0.36 0.08 GABARAPL1 GABA(A) receptor-associated protein like 1 1.30 0.10 0.43 0.03 KIAA0090 KIAA0090 protein 1.10 0.07 0.44 0.04 CHN1 Chimerin (chimaerin) 1 1.10 0.10 0.45 0.04 HEY1 Hairy/enhancer-of-split related with YRPW motif 1 1.10 0.10 0.48 0.06 IL13RA1 Interleukin 13 receptor, a 1 1.10 0.10 0.46 0.05 EDN2 Endothelin 2 1.10 0.13 0.41 0.04 RSAD2 Viperin 1.08 0.12 0.34 0.10 ATP1B3 ATPase, Na þ /K þ transporting, b 3 polypeptide 1.05 0.08 0.47 0.05 TNFAIP3 Tumor necrosis factor, a-induced protein 3 1.04 0.10 0.37 0.03 THBS1 Thrombospondin 1 1.04 0.07 0.43 0.07 LRRC16A Leucine-rich repeat containing 16 1.02 0.13 0.51 0.03 CYCS Cytochrome c 1.00 0.12 0.43 0.05 VTA1 6 open reading frame 55 0.98 0.08 0.32 0.05 OSGIN2 Chromosome 8 open reading frame 1 0.98 0.08 0.46 0.04 SRI Sorcin 0.98 0.10 0.36 0.08 PHACTR2 open reading frame 56 0.97 0.11 0.42 0.03 PDGFB Platelet-derived growth factor b polypeptide 0.96 0.13 0.43 0.05 ATP1B1 ATPase, Na þ /K þ transporting, b 1 polypeptide 0.93 0.05 0.48 0.06 HK3 Hexokinase 3 (white cell) 0.93 0.11 0.43 0.06 MAP2K6 Mitogen-activated protein kinase kinase 6 0.93 0.05 0.40 0.02 RAP1A RAP1A, member of RAS oncogene family 0.93 0.11 0.42 0.09 FGF7 Fibroblast growth factor 7 (keratinocyte growth factor) 0.92 0.14 0.51 0.04 GABRE g-Aminobutyric acid (GABA) A receptor, epsilon 0.90 0.09 0.47 0.03 NFIX Nuclear factor I/X (CCAAT-binding transcription factor) 0.90 0.07 0.39 0.04 GPR64 G protein-coupled receptor 64 0.90 0.09 0.37 0.10 LTA Lymphotoxin alpha (TNF superfamily, member 1) 0.89 0.11 0.42 0.04 DENND4B DENN/MADD domain containing 4B 0.89 0.07 0.40 0.08 IFIT5 Retinoic acid- and interferon-inducible protein (58 kDa) 0.88 0.08 0.46 0.05 RASGRP1 RAS guanyl releasing protein 1 (calcium and DAG-regulated) 0.88 0.05 0.48 0.03 HIST1H4C H4 histone family, member G 0.87 0.17 0.42 0.03 WNT11 Wingless-type MMTV integration site family, member 11 0.84 0.10 0.46 0.09 STAR Steroidogenic acute regulatory protein 0.83 0.15 0.43 0.06

Abbreviations: TNF, tumor necrosis factor; TTP, tristetraprolin. Genes in bold are those that were down-regulated by TTP and upregulated by C124R.

Oncogene TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4209 represent non-ARE genes, and thus their increase is not TTP a primary effect of TTP but most likely due to secondary 10 ** or unknown mechanisms. 9 8 Expression profile of MMP1, uPA and uPAR mRNA 7 in normal and tumor breast cells Within the list of potential TTP targets identified from 6 microarray analysis, there are several genes that code for 5 products that have established roles such as invasion 4 and metastasis in breast cancer, namely, MMP1, uPA 3 Mutant TTP and uPAR; therefore, we focused on them for further

normalized uPA/RPLPO 2 investigation. We compared expression of these tran- scripts in the same cell lines that were analyzed for TTP 1 0 expression. MDAMB231 cells exhibited the highest α α expression of uPA with a 300-fold increase over ER- IgG TTP IgG TTP Precipitating Antibody positive tumor MCF7 cells and a 10- to 20-fold greater expression than that in normal cell lines (Figure 2a). 4 TTP However, uPA expression in both normal cell lines was * higher than that in MCF7 cells. Similarly, uPAR mRNA expression in MDAMB231 cells was 30-fold 3 higher than that in MCF7 cells and 3- to 5-fold higher than that in normal cells lines, which, in turn, showed a 6- to 12-fold increase in uPAR mRNA expression over 2 Mutant TTP MCF7 cells (Figure 2b). The expression profiles of uPA and uPAR mRNAs are quite similar (Figures 2a and b), indicating a coordinated expression of these two genes. 1

Subsequently, we examined for the protein expression of normalized uPAR/RPLPO uPA and uPAR in ER-negative invasive MDAMB-231 0 cells in comparison with its normal counterpart, the IgG α TTP IgG α TTP normal ER-negative breast cell line MCF10A, as these Precipitating Antibody cell lines exhibit the most differential TTP expression both at the mRNA (Figure 1a) and protein level TTP (Figure 2d, upper panel). Protein levels of uPA and 12 *** uPAR are almost undetectable in normal breast cells compared with high expression in the invasive cell line 10 (Figure 2d, lower panels). TTP was clearly observed in 8 abundant amounts in MCF10A compared with the invasive MDAMB231 (Figure 2d). 6 The highest expression of MMP1 (Figure 2c) is clearly 4 evident in the invasive ER-negative MDAMB231 tumor Mutant TTP cell line, with an approximately 2000-fold higher 2 expression compared with the noninvasive ER-positive normalized MMP1/RPLPO MCF7 tumor cell line and an approximately 1000-fold 0 increase compared with both ER-positive and ER- IgG α TTP IgG α TTP negative normal cell lines. Precipitating Antibody Figure 3 Real-time PCR monitoring of endogenous candidate mRNAs (uPA, uPAR and MMP1) associated with TTP or mutant Binding characteristics of TTP and mutant TTP TTP precipitated with anti-TTP antibody from MDAMB231 cells. to uPA, uPAR and MMP1 mRNAs Normal rabbit immunoglobulin G control was used as an antibody To determine whether the target mRNAs were actually control with both TTP and mutant TTP samples. Quantitation of bound to TTP in vitro, we performed an immunopre- associated mRNA was performed using real-time PCR and cipitation assay of TTP-associated mRNAs using normalized to a housekeeping mRNA, RPLPO. MDAMB231. We found that uPA, uPAR and MMP1 mRNAs were bound to immunoprecipitated TTP uPA, uPAR and MMP1 are regulated by TTP protein with a 9-fold, 3-fold and 11-fold increase, in MDAMB231 cells respectively, in their anti-TTP-bound RNA precipitate To determine the effect of TTP on endogenous levels of samples over those with immunoglobulin G control uPA, uPAR and MMP1 mRNA, TTP expression was (Figure 3a–c). The nonbinding zinc-finger mutant of restored in MDAMB231 cells by transfection of wild- TTP gave results similar to immunoglobulin G control, type TTP vector. The time-course curve of uPA gene thereby confirming the nonbinding behavior of this expression shows that uPA mRNA levels peak at 2 h, mutant protein (Figure 3a–c). returning to baseline levels at 8 h after stimulation of

Oncogene TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4210 serum (Figure 4a). The maximum effect of TTP over- reduction in uPA mRNA expression compared with expression on uPA mRNA was at 2 h compared with the cells transfected with vector alone; however, as the control. Introduction of TTP resulted in a 45% transfection efficiency of MDAMB231 cells was ap- proximately 30% as determined by GFP, the actual reduction in mRNA transcripts in completely trans- 2.0 fected cells is likely to have been greater. The time- Control course induction curve of MMP1, but not that of uPAR, TTP uPA was similar to that of uPA, showing a transient pattern 1.5 of expression (Figures 4b and c). The reduction in uPAR and MMP1 mRNA was evident as a result of TTP expression (Figures 4b and c). These results indicate that ** restoration of TTP expression in MDAMB231 cells 1.0 * * results in a significant reduction in uPA, uPAR and MMP1 mRNAs. uPA/RPLPO

0.5 The uPA, uPAR and MMP1 mRNA transcripts and protein levels are increased in TTP knockout cells 0.0 Next, we decided to evaluate endogenous levels of uPA, 01248 uPAR and MMP1 mRNAs in mouse embryo fibroblast Time (hrs) cells that lack TTP compared with wild-type cells. Serum induction experiments in TTP þ / þ and TTPÀ/À 2.0 Control cells resulted in an obvious increase in the mRNA levels TTP of all three transcripts (ANOVA, Po0.001) in the uPAR absence of TTP when compared with wild-type cells 1.5 (Figure 5). The patterns of expression were similar for uPA and uPAR in these cell lines, as their mRNAs increase to a maximum at 2 h after induction, after which they decline reaching near-baseline levels after 6 h 1.0 * * (Figures 5a and b). Maximum mRNA levels were achieved in both cell lines 2 h after serum induction in uPAR/RPLPO both uPA and uPAR, with a 40% increase in mRNA 0.5 expression because of TTP knockdown (Figures 5a and b). The expression profile of MMP1 in these cell lines was quite different, showing a continuous increase in 0.0 expression, which reached a maximum at 4 h for TTPÀ/À 01248 and at 6 h for TTP þ / þ cells (Figure 5c). The greatest Time (hrs) difference between the two cell lines is evident at 2 h after induction. In the absence of TTP, MMP1 mRNA 3.5 levels were 2.5-fold higher compared with those in Control TTP þ / þ mouse fibroblasts (Figure 5c). At the protein 3.0 TTP level, western blotting showed greater protein levels of MMP1 2.5 uPA, uPAR and MMP1 in TTP knockout cells over wild-type cells (Figure 5d). In fact, TTP þ / þ cells did not 2.0 show any detectable levels of these proteins. * * 1.5 The mRNA half-life changes of the TTP-regulated MMP1/RPLPO 1.0 transcripts Actinomycin D time course experiments were performed þ / þ 0.5 to determine the decay rate of target mRNAs in TTP and TTPÀ/À fibroblasts. We fit these data to a one-phase 0.0 exponential model, a model that has the best fit (40.9 01248 goodness of fit) for mRNA decay, which tends to Time (hrs) plateau after a certain period of time. This method has Figure 4 TTP regulation of candidate target mRNAs in been recently used by us and others (Vlasova et al., 2008; MDAMB231 cells. MDAMB231 cells were transfected with TTP Al-Ahmadi et al., 2009a). All tested target mRNAs were plasmid or vector control, serum starved overnight, then serum unstable with apparent half-lives of less than 1 h (0.5– induced for 8 h, after which total RNA was extracted. Endogenous 0.6 h) in the TTP intact wild-type cells (Figure 6). uPA, uPAR and MMP1 mRNA expression levels were measured using real-time PCR and normalized to human RPLPO. The Stability was significantly increased by at least 14-fold in results are means±s.e.m. of five independent experiments. the absence of TTP (Figure 6). In addition, MMP1 *Po0.05, **Po0.005 (Student’s t test). mRNA was highly stabilized in TTPÀ/À cells (Figure 6c).

Oncogene TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4211 uPA uPAR MMP1 4.5 10 ** ** 10 +/+ *** 9 9 TTP 4.0 -/- 3.5 ** * 8 8 TTP 3.0 7 7 6 2.5 6 ** 5 5 2.0 4 4 1.5 3 MMP1/B-actin 3 1.0 2 2 0.5 1 1 Normalized uPAR/B-actin Normalized uPA/B-actin 0.0 0 0 01246 01246 01234567 Time (hrs) Time (hrs) Time (hrs)

TTP+/+TTP-/- TTP+/+TTP-/- 44 kDa 52 kDa TTP 36 kDa MMP1 34 kDa 40 kDa uPA β-actin

35 kDa uPAR

Figure 5 TTP regulation of TTP-regulated mRNAs in TTP þ / þ and TTPÀ/À mouse fibroblasts. Cells were serum starved, then induced with 10% serum for 6 h, after which total RNA was extracted. Endogenous uPA (a), uPAR (b) and MMP1 (c) mRNA levels were measured using real-time PCR and normalized to mouse b-actin. Results are means±s.e.m. of five independent experiments. *Po0.05, **Po0.005. (d) Protein expression levels of TTP, uPA, uPAR and MMP1 in TTP þ / þ and TTPÀ/À mouse fibroblasts. A volume of 60 mg of total protein was loaded onto sodium dodecyl sulfate-PAGE gel and probed with antibodies against mouse TTP, uPA, MMP1 or b-actin as a loading control.

These findings indicate that TTP regulates the decay of Discussion uPA, uPAR and MMP1 mRNA transcripts. TTP is deficient in many cancer types, including breast cancer; thus, knowledge on TTP induction and regula- TTP effect on uPA, uPAR and MMP1 GFP reporter tion of its mRNA targets may lead to new insights into activity the role of TTP deficiency in cancer processes. In this Further evidence of TTP regulation of uPA mRNA as study, we have identified, through microarray analysis, one of the newly identified targets of TTP in this study novel targets of TTP that have prominent roles in cancer was provided by experiments using an EGFP reporter invasion and metastasis. It has been shown previously construct fused with either TNF-a, uPA, uPAR ARE or that TTP destabilizes TNF-a, cyclooxygenase 2, vascu- the MMP1 30UTR region (Figure 7a). These constructs lar endothelial growth factor, IL-2 and IL-10 transcripts were cotransfected with TTP or mutant TTP (C124R) (Lai et al., 1999; Sawaoka et al., 2003; Essafi-Benkhadir into HEK293 cells. Fluorescence of GFP was measured et al., 2007; Stoecklin et al., 2008; Young et al., 2009). 24 h after transfection (Figure 7b) and showed that This work shows TTP regulation of key cancer-related reporter activity of TNF-a showed a nearly 80% genes, specifically, uPA, MMP1 and uPAR, in an reduction, whereas that of uPA and uPAR was reduced important malignant breast cancer cell model. by approximately 60% in cells transfected with TTP The detection of transcripts with low expression compared with control cells. MMP1, the entire 30UTR levels, and thus low signals in microarray data, can be region of which was used, showed a 20% reduction as a problematic if they are further downregulated by result of TTP transfection. C124R, on the other hand, experimental treatments (Asyali et al., 2004). Therefore, increased the reporter activity of both TNF-a and uPA in this work, we took advantage of a nonbinding TTP ARE (Figure 7b), indicating the opposite (dominant- mutant that behaves as dominant negative to probe for negative) effect. The results indicate that TTP regulates targets of TTP-mediated mRNA decay. Comparing the the expression of uPA, uPAR and MMP1 mRNAs in an effects of overexpression of TTP and the TTP mutant on ARE-dependent manner. ARE-mRNA expression (Table 1) led to a list of potential targets in which effects of both wild-type and TTP effect on invasiveness of MDAMB231 cells mutant TTP are consistent. Some of the transcripts that To examine the effect of TTP on the invasiveness of appear in this list are established targets of TTP, such as MDAMB231 cells, we transfected these cells with TTP IL-10 (Stoecklin et al., 2008) and colony-stimulating or C124R plasmids and cotransfected with luciferase factor (Carballo et al., 2000; Blackshear, 2002). However, construct. The cells were then overlaid on matrigel. uPA, uPAR and MMP1 are all novel targets of TTP Luciferase was measured in cells that had invaded the regulation; real-time PCR, immunoprecipitation and use matrigel to the bottom chamber. There was a marked of TTP knockout cells, all confirmed them as TTP targets. reduction in the invasiveness of TTP-transfected cells The uPA, uPAR and MMP1 genes were the most compared with either C124R-transfected cells or control significant candidates for TTP regulation, because they cells (Figure 7c). TTP mutant had a minimal effect on all share a common functional role; they are key factors the ability of cells to invade the matrigel. in invasion and metastasis in various types of cancer,

Oncogene TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4212 Y=Span*exp(-K*X) + Plateau BamHI XbaI pA Prom EGFP 3’UTR +/+ 100 TTP -/- TTP uPA ARE: cactgtctcagtttcactttcacatagatgtccctttcttggccagttatccc Ttccttttagcctagttcatccaatcctcactgggtggggtgaggaccactcctgtacactg 90 aatatttatatttcactatttttatttatatttttgtaattttaaat TNF ARE: ttgtgattatttattatttatttattatttatttatttacagatgaatgtatttatttgggagat apparent t1/2=7 uPAR ARE:tgttgttgttattaattaatattcatattatttattttatacttacataaagattttgtac 80 MMP1 3’UTR

ANOVA*** 70 apparent t1/2=0.5 TNFα 140 - ARE ANOVA*** 60 ANOVA*** uPA mRNA remaining (%) 120 uPA uPAR MMP1 50 100 0123456 * 80 ActD (hrs) 60 **

40 ** 100 ** 20 apparent t =7 0 90 1/2 GFP Fluorescence % of PCR3.1 control TTP TTP TTP TTP TTP PCR3.1 C124RPCR3.1 C124RPCR3.1 C124RPCR3.1 C124RPCR3.1 C124R 80 100

70 75 apparent t1/2=0.6

uPAR mRNA remaining (%) 60 50

50 0123456 25 * ActD (hrs)

invasion (% of PCR3.1 control) 0 PCR3.1 TTP C124R 100 Figure 7 TTP regulation of expression of EGFP reporter activity apparent t1/2>10 and invasion assays (a) EGFP reporter constructs were fused with 80 uPA ARE, TNF ARE, uPAR ARE or MMP1 30 UTR regions (cloned in BamHI/xbaI sites in stable growth hormone 30UTR (upper panel). (b) HEK293 cells were cotransfected with reporters 60 and TTP, C124R (mutant TTP) or a vector plasmid alone. GFP fluorescence was measured 24 h after transfection. (c) TTP modulation of invasion. MDAMB231 cells were transfected with 40 TTP, C124R or PCR3.1 control plasmid and cotransfected with luciferase construct. At 24 h after transfection, the cells were reseeded in invasion chambers in serum-free medium and incubated apparent t1/2=0.7 overnight. Invaded cells were lysed and luciferase was measured in

MMP1 mRNA remaining (%) 20 illuminometer. Transfection efficiency was monitored and found comparable with the three constructs. 0 0123456 leads to the degradation of the extracellular matrix . ActD (hrs) MMP1 (also called interstitial collagenase) is a metal- Figure 6 TTP-regulated mRNA decay curve in TTP þ / þ and loenzyme that also degrades basement membrane TTPÀ/À mouse fibroblasts. Cells were serum starved overnight, components, namely, Type III collagen, in the extra- induced with 10% serum for 1 h, then treated with actinomycin D cellular matrix, contributing to the invasive and meta- (5 mgmlÀ1) for indicated times. Total RNA was extracted and used static processes (Liotta et al., 1980; Liotta, 1986) and for real-time PCR. (a) uPA; (b) uPAR; (c) MMP1. The results are increased expression of MMP1 results in connective means þ s.e.m. of five independent experiments. The one-phase exponential decay model is described in Materials and methods. tissue destruction that can lead to many pathological conditions (Vincenti and Brinckerhoff, 2002). Posttranscriptional regulation of MMP1, uPA and including breast cancer (Quax et al., 1990; Christensen uPAR mRNA expression takes place because of the et al., 1996; Iwata et al., 1996; Kleiner and Stetler- presence of AREs in the 30UTR of these transcripts Stevenenson, 1999; Fisher et al., 2000; Dass et al., 2008). (Roldan et al., 1990; Nanbu et al., 1994). An earlier uPA is a serine protease that, when bound to its receptor report revealed that the AUUUA motif present in the uPAR, converts plasminogen to plasmin, which, in turn, 30UTR of the collagenase gene destabilized its mRNA

Oncogene TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4213 and that mutation of AREs led to increased stability of The increased expression of all three transcripts in its mRNA transcript (Vincenti et al., 1994). In addition, MDAMB231 cells is expected, given the highly invasive ARE-mediated destabilization of uPA and uPAR nature of these cells. What was odd, however, was the mRNAs has been reported, although the mediators of fact that uPA and uPAR exhibited a higher expression this destabilization have not yet been identified (Nanbu in both ER-positive and ER-negative normal cell lines et al., 1994; Shetty et al., 1997; Wang et al., 1998). In compared with noninvasive MCF-7 cells. However, this fact, a defect in the ARE-mediated decay of uPA was not the case for MMP1, which was only highly mRNA has been implicated in the stability of uPA in expressed in the invasive-type cell line. Regardless, the certain cell lines, such as MDAMB231 breast cancer findings indicate a relationship between the deficiency of cells (Nanbu et al., 1997). The present results suggest TTP in tumor and metastatic breast cancer cells and the that TTP is a mediator of uPA, uPAR and MMP1 high expression of MMP1, uPA and uPAR. mRNA turnover, which can lead to a reduction in the In summary, MMP1, uPA and uPAR have been invasiveness of breast cancer cells. shown here to be definite targets of TTP-ARE-mediated Overexpression studies have revealed that HuR, an degradation. The data suggest that TTP, in a 30UTR- ARE-binding protein that stabilizes mRNA transcripts, and ARE-dependent manner, regulates an important stabilizes its own mRNA (Al-Ahmadi et al., 2009a), in subset of cancer-related genes that are involved in addition to mRNA transcripts of both uPA and uPAR cellular growth, invasion and metastasis. TTP regulation (Tran et al., 2003). TTP reduces HuR mRNA levels, and of MMP1, uPA and uPAR mRNA has obvious competes with HuR-mediated stabilization of HuR implications for regulating the invasive and metastatic mRNA and HuR targets (Al-Ahmadi et al., 2009a). properties of certain cancer cells. The finding that TTP Thus, TTP action may involve both direct and indirect, regulates an important subset of ARE-gene products HuR-modulated actions. Both HuR expression levels highly visible in breast cancer, especially invasion and and cytoplasmic localization are increased in many metastasis, has promising implications toward under- cancer types (de Silanes et al., 2003; Denkert et al., 2004; standing the role of TTP in regulating cancer processes. Mrena et al., 2005; Heinonen et al., 2007). Concurrently, a loss of TTP expression has been reported in various cancers, such as breast and prostate cancers, as well as in colorectal adenocarcinomas and cancer cell lines, and Materials and methods TTP suppression has been correlated with a negative Cell culture and transfection prognosis (Brennan et al., 2009; Young et al., 2009). HEK293, MDAMB231 and MCF-7 cell lines were obtained Transcriptional silencing resulting from hypermethyla- from ATCC and cultured in Dulbecco’s Modified Eagle tion of CpG islands present in the TTP promoter has Medium (DMEM, Invitrogen, Carlsbad, CA, USA), supple- been suggested as a mechanism for TTP deficiency in mented with 10% fetal bovine serum and antibiotics. The cancer (Young et al., 2009). A number of miRNAs, such normal cell lines of MCF12A and MCF10A cells were cultured as miR-29a, are overexpressed in tissues of breast cancer in a 1:1 mixture of Ham’s F12 and DMEM, supplemented patients and target the mRNAs of proteins including with 0.01 mg mlÀ1 bovine insulin, 20 ng mlÀ1 epidermal TTP, leading to its suppression (Gebeshuber et al., growth factor and 500 ng mlÀ1 hydrocortisone. The TTP þ / þ À/À 2009). Furthermore, in cell lines and tissues in which and TTP mouse embryonic fibroblasts were previously TTP expression is ubiquitous, hyperphosphorylation of described (Lai et al., 2006) and were grown in DMEM. For transfection experiments, cells were transfected with CMV. TTP by p38 or ERK has an antagonistic role on the hTTP.tag, C124RCMV.hTTP.tag or pBS þ (Stratagene, La destabilizing activity of TTP, leading to the increased Jolla, CA, USA) plasmid using Lipofectamine 2000 (Invitrogen, mRNA stability of factors such as vascular endothelial Carlsbad, CA, USA), according to the manufacturer’s instruc- growth factor and IL-8 that promote the progression of tions. After 4–5 h, transfection media were replaced with regular tumors (Essafi-Benkhadir et al., 2007; Suswam et al., media and incubated overnight. Serum starvation of cells was 2008). Our results are consistent with published data carried out the following day. showing that TTP mRNA and protein expression are deficient in tumor breast cell lines compared with Western blotting normal cells (Brennan et al., 2009; Young et al., 2009). Equal amounts of protein samples were subjected to electro- Thus, dual TTP deficiency and HuR overexpression may phoresis on 10% polyacrylamide–sodium dodecyl sulfate lead to a synergistic increased expression of target gels (Protogel, National Diagnostics, Atlanta, GA, USA) or mRNAs, such as uPA, uPAR and MMP1, and NuPAGE 10%, followed by transfer to nitrocellulose mem- amplifying further associated cancer processes such as branes (Hybond ECL, Amersham Biosciences, Little Chalfont, invasion and metastasis. UK). Membranes were hybridized with primary antibody to MMP1, uPA and uPAR mRNAs were expressed goat anti-TTP (1/500) (Santa Cruz Biotech, Santa Cruz, CA, highly in breast cancer cell lines, especially in estrogen USA) or goat anti-uPA (1 mg/ml, human) (Abcam, Cambridge, receptor-negative and highly invasive MDAMB231 UK, ab40841) or goat anti-uPAR (1 mg/ml, human) (Abcam, ab3129), or to mouse anti-uPA (1/250, mouse), sheep anti- cells, and thus ER status may have a role in their MMP1 (1/500, Abcam, ab8480) or b actin (1/1000), followed expression. The ER-positive status is correlated with by secondary HRP-conjugated antibody. Signal detection was decreased invasiveness and metastasis. On the other performed with ECL western blotting detection reagents hand, MMP1 mRNA was virtually undetectable in (Amersham, Piscataway, NJ, USA). Protein molecular weight normal mammary cell lines. markers were used to verify the size of the proteins.

Oncogene TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4214 Determinations of half-life by the one-phase exponential decay used in transfection of HEK293 seeded at a density of 3.5  104 model cells per well in 96-well plates. A total of 30 ng of reporter Serum-stimulated wild-type TTP and knockout MEFs were construct was used and 50 ng of TTP, C124R or PCR3.1 PCR treated with actinomycin D (5 mg/ml) for increasing periods of products was cotransfected using Lipofectamine 2000. Fluor- time. Total RNA was extracted and subjected to reverse escence was measured 24 h after transfection using a Zenith transcription quantitative PCR as in Supplementary Methods. 3100 bottom read fluorescence reader (Anthos Labtec, The one-phase exponential decay curve analysis (GraphPad Eugendorf, Austria). Prism, La Jolla, CA, USA) was used to assess mRNA decay kinetics. The equation Y ¼ Span  exp (ÀK  X) þ Plateau describes the kinetics of mRNA decay, where X denotes time Invasion assays and Y the mRNA level (expressed as % remaining after MDAMB231 cells were seeded in six-well cell culture plates actinomycin D), where Y begins equal starts at Span þ Plateau and incubated overnight. Cells were transfected with 0.5 mgof and decays to Plateau with a rate constant K. Half-life is TTP, C124R or PCR3.1 plasmid and cotransfected with 0.6932/K. Span and Plateau are expressed in the same units as 0.25 mg of luciferase PCR product and incubated overnight. in the y axis. K is expressed as the inverse of the units used by The cells were reseeded onto the invasion chamber in serum- 5 the x axis. This model is suitable for mRNA determinations in free media at a density of 3  10 cells per well (24-well BD which mRNA levels plateau without dropping to zero and BioCoat Invasion Chamber, BD Biosciences, MA, USA) and with high regression coefficients, for example, 40.9. incubated for 24 h. Membranes were removed from the inserts and incubated with luciferase lysis buffer for 15 min, then luciferase dye was added to the lysate and luciferase was Immunoprecipitation of TTP-associated mRNAs measured. MDAMB231 cells were seeded in 100  20 mm culture dishes at a density of 3.6  106 cells per dish and treated as described above. RNA labeling and array hybridization Cells lysed with polysome lysis buffer (100 mM KCl, 5 mM MgCl2, 10 mM HEPES, pH 7.0, 0.5% Nonidet P-40) containing RNAse In Supplementary Methods. inhibitors and ethylenediaminetetraacetic acid-free proteinase 1 inhibitors and then centrifuged at 14 000 r.p.m. at 4 C. A total of Quantitative reverse transcription–PCR 100 ml of lysate was incubated for 1–2 h with 100 mlofpreswollen In Supplementary Methods. A/G agarose beads (Santa Cruz) precoated with 20 mgofanti- TTP or normal rabbit immunoglobulin G (Santa Cruz). The beads were washed with NT2 buffer (50 mM Tris–HCl (pH 7.4), 150 mM NaCl, 1 mM MgCl2 and 0.05% NP-40) and then Conflict of interest incubated in 100 ml of NT2 buffer containing 20 units of RNAse-free DNAse I to remove genomic DNA contamination, The authors declare no conflict of interest. washed, then incubated in 100 ml NT2 buffer containing 0.1% sodium dodecyl sulfate and 0.5 mg/ml proteinase K (15 min, 55 1C) to digest the protein bound to the beads. RNA was subject to TaqMan real-time reverse transcription-PCR as described in Acknowledgements Supplementary Methods. This study is a part of the Newcastle University doctoral thesis Reporter gene studies requirements of NA-S. NA-S was supported by a PhD EGFP reporter constructs containing 30UTR sequences scholarship from King Khalid Foundation, Riyadh. We harboring either the 140 base uPA ARE region, 61 base acknowledge Dr Wi S Lai for TTP plasmids and for the uPAR ARE region, the entire 30UTR region of MMP1, 64 knockout MEF line. We also acknowledge Mr Maher Al-Saif base TNF-a ARE or a control lacking an ARE region were for his technical assistance.

References

Al-Ahmadi W, Al-Ghamdi M, Al-Haj L, Al-Saif M, Khabar KS. lin is suppressed in many cancers, altering tumorigenic phenotypes (2009a). Alternative polyadenylation variants of the RNA binding and patient prognosis. Cancer Res 69: 5168–5176. protein, HuR: abundance, role of AU-rich elements and auto- Carballo E, Lai WS, Blackshear PJ. (1998). Feedback inhibition of regulation. Nucl Acids Res 37: 3612–3624. macrophage tumor necrosis factor-{alpha} production by tristetra- Al-Ahmadi W, Al-Haj L, Al-Mohanna FA, Silverman RH, Khabar prolin. Science 281: 1001–1005. KS. (2009b). RNase L downmodulation of the RNA-binding Carballo E, Lai WS, Blackshear PJ. (2000). Evidence that tristetra- protein, HuR, and cellular growth. Oncogene 28: 1782–1791. prolin is a physiological regulator of granulocyte-macrophage Asyali MH, Shoukri MM, Demirkaya O, Khabar KSA. (2004). colony stimulating factor messenger RNA deadenylation and Assessment of reliability of microarray data and estimation of signal stability. Blood 95: 1891–1899. thresholds using mixture modeling. Nucl Acids Res 32: 2323–2335. Christensen L, Wiborg Simonsen AC, Heeqaard CW, Moestrup SK, Bacac M, Stamenkovic I. (2008). Metastatic cancer cell. Annu Rev Andersen JA, Andreasen P. (1996). Immunohistochemical localiza- Pathol: Mech Dis 3: 221–247. tion of urokinase-type plasminogen activator, type-1 plasminogen- Blackshear PJ. (2002). Tristetraprolin and other CCCH tandem zinc- activator inhibitor, urokinase receptor and alpha(2)-macroglobulin finger proteins in the regulation of mRNA turnover. Biochem Soc receptor in human breast carcinomas. Int J Cancer 66: 441–452. Trans 30: 945–952. Dass K, Ahmad A, Azmi AS, Sarkar SH, Sarkar FH. (2008). Evolving Brennan CM, Kuwano Y, Alkharouf N, Blackshear PJ, Gorospe M, role of uPA/uPAR system in human cancers. Cancer Treat Rev 34: Wilson GM. (2009). The mRNA-destabilizing protein tristetrapro- 122–136.

Oncogene TTP regulation of AU-rich mRNAs in breast cancer lines N Al-Souhibani et al 4215 de Silanes IL, Fan J, Yang X, Zonderman AB, Potapova O, Pizer ES McPherson K, Steel CM, Dixon JM. (2000). ABC of breast diseases: et al. (2003). Role of the RNA-binding protein HuR in colon breast cancer—epidemiology, risk factors, and genetics. Br Med J carcinogenesis. Oncogene 22: 7146–7154. 321: 624–628. Denkert C, Weichert W, Pest S, Koch I, Licht D, Kobel M et al. Mrena J, Wiksten J-P, Thiel A, Kokkola A, Pohjola L, Lundin J et al. (2004). Overexpression of the embryonic-lethal abnormal vision-like (2005). Cyclooxygenase-2 is an independent prognostic factor in protein HuR in ovarian carcinoma is a prognostic factor and is gastric cancer and its expression is regulated by the messenger RNA associated with increased cyclooxygenase 2 expression. Cancer Res stability factor HuR. Clin Cancer Res 11: 7362–7368. 64: 189–195. Nanbu R, Menoud PA, Nagamine Y. (1994). Multiple instability- DuBois RN, McLane MW, Ryder K, Lau LF, Nathans D. (1990). A regulating sites in the 30 untranslated region of the urokinase-type growth factor-inducible nuclear protein with a novel cysteine/ plasminogen activator mRNA. Mol Cell Biol 14: 4920–4928. histidine repetitive sequence. J Biol Chem 265: 19185–19191. Nanbu R, Montero L, D’Orazio D, Nagamine Y. (1997). Enhanced Essafi-Benkhadir K, Onesto C, Stebe E, Moroni C, Pages G. (2007). stability of urokinase-type plasminogen activator mRNA in Tristetraprolin inhibits Ras-dependent tumor vascularization by metastatic breast cancer MDA-MB-231 cells and LLC-PK, cells inducing vascular endothelial growth factor mRNA degradation. down-regulated for protein kinase C correlation with cytoplasmic Mol Biol Cell 18: 4648–4658. heterogeneous nuclear ribonucleoprotein C. Eur J Biochem 247: Fini ME, Plucinska IM, Mayer AS, Gross RH, Brinckerhoff CE. 169–174. (1987). A gene for rabbit synovial cell collagenase: member of a Quax PHA, van Leeuwen RTJ, Verspaget HW, Verheijen JH. (1990). family of metalloproteinases that degrade the connective tissue Protein and messenger RNA levels of plasminogen activators and matrix. Biochemistry 26: 6157–6165. inhibitors analyzed in 22 human tumor cell lines. Cancer Res 50: Fisher JL, Field CL, Zhou H, Harris TL, Henderson MA, Choong PF. 1488–1494. (2000). Urokinase plasminogen activator system gene expression is Roldan AL, Cubellis MY, Masucci MT, Behrendt NL, Lund LR, increased in human breast carcinoma and its bone metastases – a Dano K et al. (1990). Cloning and expression of the receptor for comparison of normal breast tissue, non-invasive and invasive human urokinase plasminogen activator, a central molecule in cell carcinoma and osseous metastases. Breast Cancer Res Treat 61:1–12. surface plasmin dependent proteolysis. EMBO J 9: 467–474. Gebeshuber CA, Zatloukal K, Martinez J. (2009). miR-29a suppresses Sawaoka H, Dixon DA, Oates JA, Boutaud O. (2003). Tristetraprolin tristetraprolin, which is a regulator of epithelial polarity and binds to the 30-untranslated region of cyclooxygenase-2 mRNA. A metastasis. EMBO J Rep 10: 400–405. polyadenylation variant in a cancer cell line lacks the binding site. J Heinonen M, Fagerholm R, Aaltonen K, Kilpivaara O, Aittomaki K, Biol Chem 278: 13928–13935. Blomqvist C et al. (2007). Prognostic role of HuR in hereditary Shetty S, Kumar A, Idell S. (1997). Posttranscriptional regulation of breast cancer. Clin Cancer Res 13: 6959–6963. urokinase receptor mRNA: identification of a novel urokinase Iwata H, Kobayashi S, Iwase H, Masaoka A, Fujimoto N, Okada Y. receptor mRNA binding protein in human mesothelioma cells. Mol (1996). Production of matrix metalloproteinases and tissue inhibi- Cell Biol 17: 1075–1083. tors of metalloproteinases in human breast carcinomas. Jpn J Soslow RA, Dannenberg AJ, Rush D, Woerner BM, Khan KN, Cancer Res 87: 602–611. Masferrer J et al. (2000). COX-2 is expressed in human pulmonary, Khabar KS, Al-Haj L, Al-Zoghaibi F, Marie M, Dhalla M, Polyak SJ colonic, and mammary tumors. Cancer 89: 2637–2645. et al. (2004). Expressed gene clusters associated with cellular Stoecklin G, Tenenbaum SA, Mayo T, Chittur SV, George AD, sensitivity and resistance towards anti-viral and anti-proliferative Baroni TE et al. (2008). Genome-wide analysis identifies interleukin- actions of interferon. J Mol Biol 342: 833–846. 10 mRNA as target of tristetraprolin. J Biol Chem 283: 11689–11699. Kleiner DE, Stetler-Stevenenson W. (1999). Matrix metalloproteinases and metastasis. Cancer Chemother Pharmacol 43: S42–S51. Suswam E, Li Y, Zhang X, Gillespie GY, Li X, Shacka JJ et al. Lai WS, Carballo E, Strum JR, Kennington EA, Phillips RS, (2008). Tristetraprolin down-regulates interleukin-8 and vascular Blackshear PJ. (1999). Evidence that tristetraprolin binds to AU- endothelial growth factor in malignant glioma cells. Cancer Res 68: rich elements and promotes the deadenylation and destabilization of 674–682. tumor necrosis factor alpha mRNA. Mol Cell Biol 19: 4311–4323. Tran H, Maurer F, Nagamine Y. (2003). Stabilization of urokinase Lai WS, Carballo E, Thorn JM, Kennington EA, Blackshear PJ. and urokinase receptor mRNAs by HuR is linked to its cytoplasmic (2000). Interactions of CCCH zinc finger proteins with mRNA. accumulation induced by activated mitogen-activated protein Binding of tristetraprolin-related zinc finger proteins to kinase-activated protein kinase 2. Mol Cell Biol 23: 7177–7188. AU-rich elements and destabilization of mRNA. J Biol Chem 275: Vincenti MP, Brinckerhoff CE. (2002). Transcriptional regulation of 17827–17837. collagenase (MMP-1, MMP-13) genes in arthritis: integration of Lai WS, Kennington EA, Blackshear PJ. (2002). Interactions of complex signaling pathways for the recruitment of gene-specific CCCH zinc finger proteins with mRNA. Non-binding tristetrapro- transcription factors. Arthritis Res 4: 157–164. lin mutants exert an inhibitory effect on degradation of AU-rich Vincenti MP, Coon CI, Lee O, Brinckerhoff CE. (1994). Regulation of element-containing mRNAs. J Biol Chem 277: 9606–9613. collagenase gene expression by IL-1 requires transcriptional and Lai WS, Parker JS, Grissom SF, Stumpo DJ, Blackshear PJ. (2006). post-transcriptional mechanisms. Nucl Acids Res 22: 4818–4827. Novel mRNA targets for tristetraprolin (TTP) identified by global Vlasova IA, Tahoe NM, Fan D, Larsson O, Rattenbacher B, analysis of stabilized transcripts in TTP-deficient fibroblasts. Mol Sternjohn JR et al. (2008). Conserved GU-rich elements mediate Cell Biol 26: 9196–9208. mRNA decay by binding to CUG-binding protein 1. Mol Cell 29: Lai WS, Stumpo DJ, Blackshear PJ. (1990). Rapid insulin-stimulated 263–270. accumulation of an mRNA encoding a proline-rich protein. J Biol Wang GJ, Collinge M, Blasi F, Pardi R, Bender JR. (1998). Chem 265: 16556–16563. Posttranscriptional regulation of urokinase plasminogen activator Liotta LA. (1986). Tumor invasion and metastases—role of the receptor messenger RNA levels by leukocyte integrin engagement. extracellular matrix: Rhoads Memorial Award lecture. Cancer Res Proc Natl Acad Sci USA 95: 6296–6301. 46: 1–7. Young LE, Sanduja S, Bemis-Standoli K, Pena EA, Price RL, Dixon Liotta LA, Tryggvason K, Garbisa S, Hart I, Foltz CM, Shafie S. DA. (2009). The mRNA binding proteins HuR and tristetraprolin (1980). Metastatic potential correlates with enzymatic degradation regulate cyclooxygenase 2 expression during colon carcinogenesis. of basement membrane collagen. Nature 284: 67–68. Gastroenterology 136: 1669–1679.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene