JNK and PI3K differentially regulate MMP-2 and MT1-MMP mRNA and protein in response to actin cytoskeleton reorganization in endothelial cells

 doi:10.1152/ajpcell.00300.2005 291:579-588, 2006. First published May 3, 2006;Am J Physiol Cell Physiol

Eric Ispanovic and Tara L. Haas cytoskeleton reorganization in endothelial cells MT1-MMP mRNA and protein in response to actin JNK and PI3K differentially regulate MMP-2 and

You might find this additional information useful...

63 articles, 36 of which you can access free at: This article cites http://ajpcell.physiology.org/cgi/content/full/291/4/C579#BIBL

including high-resolution figures, can be found at: Updated information and services http://ajpcell.physiology.org/cgi/content/full/291/4/C579

can be found at: AJP - Cell Physiologyabout Additional material and information http://www.the-aps.org/publications/ajpcell

This information is current as of September 14, 2006 .  

http://www.the-aps.org/.American Physiological Society. ISSN: 0363-6143, ESSN: 1522-1563. Visit our website at a year (monthly) by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2005 by the

is dedicated to innovative approaches to the study of cell and molecular physiology. It is published 12 timesAJP - Cell Physiology

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

JNK and PI3K differentially regulate MMP-2 and MT1-MMP mRNA andprotein in response to actin cytoskeleton reorganization in endothelial cells

Eric Ispanovic and Tara L. HaasSchool of Kinesiology and Health Sciences, York University, Toronto, Ontario, Canada

Submitted 17 June 2005; accepted in final form 28 April 2006

Ispanovic, Eric, and Tara L. Haas. JNK and PI3K differentiallyregulate MMP-2 and MT1-MMP mRNA and protein in response toactin cytoskeleton reorganization in endothelial cells. Am J PhysiolCell Physiol 291: C579–C588, 2006. First published May 3, 2006;doi:10.1152/ajpcell.00300.2005.—Increased production and activa-tion of matrix metalloproteinase-2 (MMP-2) are critical events inskeletal muscle angiogenesis and are known to occur in response tomechanical stresses. We hypothesized that reorganization of the actincytoskeleton would increase endothelial cell production and activationof MMP-2 and that this increase would require a MAPK-dependentsignaling pathway in endothelial cells. The pharmacological actindepolymerization agent cytochalasin D increased expression ofMMP-2 and membrane type 1-matrix metalloproteinase (MT1-MMP)mRNA, and this was reduced significantly in the presence of the JNKinhibitor SP600125. Activation of JNK by anisomycin was sufficientto induce expression of both MMP-2 and MT1-MMP mRNA inquiescent cells. Downregulation of c-Jun, a downstream target ofJNK, with small interference (si)RNA inhibited MMP-2 expression inresponse to anisomycin. Inhibition of phosphoinositide 3-kinase(PI3K), but not JNK, significantly decreased the amount of activeMMP-2 following cytochalasin D stimulation with a concurrent de-crease in MT1-MMP protein. Physiological reorganization of actinoccurs during VEGF stimulation. VEGF-induced MMP-2 proteinproduction and activation, as well as MT1-MMP protein production,depended on PI3K activity. VEGF-induced MMP-2 mRNA expres-sion was reduced by inhibition of JNK or by treatment with c-JunsiRNA. In summary, our results provide novel insight into thesignaling cascades initiated in the early stages of angiogenesisthrough the reorganization of the actin cytoskeleton and demonstratea critical role for JNK in regulating MMP-2 and MT1-MMP mRNAexpression, whereas PI3K regulates protein levels of both MMP-2 andMT1-MMP.

angiogenesis; mechanotransduction; vascular endothelial growth fac-tor; c-Jun; phosphoinositide 3-kinase; membrane type 1-matrix met-alloproteinase

ANGIOGENESIS IS THE GROWTH of new capillaries from preexistingmature ones and occurs through a cascade of events in whichdisruption of the endothelial adherens junctions by way of actincytoskeleton reorganization and proteolysis of the basementmembrane and interstitial matrix are critical steps. Remodelingof both the basement membrane and interstitial matrix isfacilitated by the expression and activation of matrix metallo-proteinases (MMPs), a family of zinc- and calcium-dependentenzymes (7, 54). Each MMP has substrate specificity for alimited set of extracellular matrix proteins (for review, see Ref.53). MMPs are key participants in several steps of the angio-genic response, including regulation of endothelial cell perme-ability, migration, invasion, and tubule formation (16, 23).

MMP-2 and membrane type 1 (MT1)-MMP are produced byendothelial cells and can degrade type I and IV collagen (2,41). MMP-2-deficient mice have reduced tumor, corneal, andretinal angiogenesis (27, 29, 43), whereas MT1-MMP-deficientmice fail to gain weight, have deficient connective tissuemetabolism, and die 3–4 wk after birth (26, 61). CombinedMMP-2 and MT1-MMP deficiency causes embryonic lethality,highlighting the potential of these MMPs as key targets incontrolling angiogenesis (42).

The actin cytoskeleton creates a scaffold that provides struc-tural stability and the organization of signaling molecules. Inendothelial cells, shear stress and mechanical stretch causeactin stress fiber reorganization in the direction of flow andperpendicular to the axis of stretch, respectively (8, 11, 17).This reorganization of the actin cytoskeleton increases MMP-2and MT1-MMP expression (20, 48). Growth factors such asVEGF induce changes in endothelial cell permeability byremodeling the actin cytoskeleton (47). These changes in theactin cytoskeleton may contribute significantly to the initialsignaling events during angiogenesis. However, the signalingevents that connect actin reorganization to MMP-2 productionhave yet to be determined.

Numerous signaling cascades may be activated with re-organization of the actin cytoskeleton. These include themitogen-activated protein kinases (MAPK), specifically ex-tracellular signal-regulated kinase (ERK) and c-Jun NH2-terminal kinase (JNK), focal adhesion kinase (FAK), Rho-family GTPases (Rho, Rac, and Cdc42), and their down-stream effector, p21-activated kinase (PAK), as well asphosphoinositide 3-kinase (PI3K) (1, 5, 8, 30, 33, 56). TheMAPKs are implicated in controlling a number of angio-genic processes (migration, proliferation) with ERK andJNK specifically implicated in the production and activationof MMP-2 (9, 13, 38, 40). The MAPKs can activate numeroustranscription factors, including the AP-1 family, which theninitiate gene expression.

We hypothesized that expression of endothelial MMP-2 andMT1-MMP in response to depolymerization of the actin cy-toskeleton requires MAPKs. We have shown that in primarycultures of rat skeletal muscle endothelial cells (SMEC), treat-ment with cytochalasin D activates JNK and increases bothMMP-2 and MT1-MMP mRNA, which results in increasedMMP-2 production and activation. These increases inMMP-2 and MT1-MMP involve JNK- and PI3K-dependentpathways. Downstream of JNK activation, we have identi-fied c-Jun as a transcriptional activator regulating MMP-2.Furthermore, the potent angiogenic factor VEGF increases

Address for reprint requests and other correspondence: T. L. Haas, Schoolof Kinesiology and Health Sciences, York Univ., 4700 Keele St., Toronto,Ontario, Canada M3J 1P3 (e-mail: [email protected]).

The costs of publication of this article were defrayed in part by the paymentof page charges. The article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Am J Physiol Cell Physiol 291: C579–C588, 2006.First published May 3, 2006; doi:10.1152/ajpcell.00300.2005.

0363-6143/06 $8.00 Copyright © 2006 the American Physiological Societyhttp://www.ajpcell.org C579

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

MMP-2 mRNA and MMP-2 and MT1-MMP protein pro-duction, also through JNK- and PI3K-dependent pathways.Together, these data implicate JNK/c-Jun, in conjunctionwith PI3K, as physiological regulators of MMP-2 and MT1-MMP in endothelial cells.

MATERIALS AND METHODS

Cell culture. SMEC were isolated from extensor digitorum longusmuscles and cultured as previously described (21), using protocolsapproved by the York University Animal Care Committee and incompliance with the American Physiological Society’s “GuidingPrinciples in the Care and Use of Animals.” Cells were used forexperiments between passages 4 and 11 and plated on type I collagen(12.5 �g collagen/ml coating buffer)-coated culture dishes. For inhib-itor studies, SMEC were pretreated for 3 h with either 50 �MPD-98059, 10 �M U0126, 50 or 100 �M SP600125, 40 �M SB-203580, or 50 �M LY-294002 and then treated with either 1 or 10 �Mcytochalasin D for 24 h, 25 ng/ml VEGF or 1 �g/ml anisomycin(Sigma) for 7 h. All inhibitors were purchased from Calbiochem.

Immunofluorescence staining. Cells were plated on collagen-coatedglass coverslips and cultured in the presence or absence of 25 ng/mlVEGF for 24 h at 37°C before staining. Cells were fixed with 3.75%paraformaldehyde and then blocked and permeabilized in PBS plus5% normal goat serum and 0.05% Triton X-100. Cells were thenstained with primary phospho-JNK antibody (1:300 dilution; Upstate)and then with secondary goat anti-rabbit Alexa-568 (1:400 dilution;Molecular Probes). Nuclei were counterstained with 4�,6�-diamin-odino-2-phenylindole (1:1,500 dilution; Molecular Probes). Cellswere visualized using fluorescence microscopy (Zeiss Axiovert200M). Images were captured using a cooled digital charge-coupleddevice camera (Quantix 57) and imaging software (Metamorph;Universal Imaging).

Gelatin zymography. Cells were lysed in 120 mM Tris �HCl (pH8.7), 0.1% Triton X-100, and 5% glycerol supplemented with proteaseinhibitors (Sigma) and sodium orthovandate (lysis buffer). Protein (10�g), as determined using bicinchoninic acid assay (BCA; Pierce), wasseparated on an 8% SDS-polyacrylamide gel containing 0.02% gelatin(20). The gel was incubated at 37°C for 20–24 h in buffer containing50 mM Tris �HCl buffer (pH 7.6) with 5 mM CaCl2, after which thegels were fixed with 50% methanol and 10% acetic acid and, finally,stained with 0.25% Coomassie blue protein stain. Gels were visual-ized and imaged using the Fluorchem gel doc system and analyzedusing Alphaease software (Alpha Innotech). Total MMP-2 proteinexpression was calculated as the sum of the latent (72 kDa) and active(62 kDa) bands, whereas the amount of active MMP-2 was calculatedas the percentage of active compared with total MMP-2 protein. ForPI3K inhibitor treatments, the amount of active MMP-2 (62 kDa) wasexpressed as the increase relative to cytochalasin D.

Northern blot. Total RNA was isolated by lysing cells in TRIzolreagent (GIBCO) and quantified using spectrophotometry at 260 nm.RNA (10 �g) was separated by formaldehyde-containing 1% agarosegel electrophoresis, and gels were stained with SYBR green (Sigma)to visualize 28S/18S ribosomal RNA bands. Gels were transferredovernight to a nylon membrane (GeneScreen Plus; NEN Life ScienceProducts) and probed with 32P-labeled MMP-2 or MT1-MMP cDNA(26). The membrane was exposed to film for 1 h to 7 days at �80°C.The film was developed and scanned, and densitometry was per-formed using Alphaease software. Loading was normalized to the 28Sband.

Western blot. SMEC were lysed using lysis buffer or 1� loadingbuffer (37.75 mM Tris �HCl, 65 mM DTT, 3.75% SDS, 5% glycerol,and 0.003% bromphenol blue), and 10 �g of protein or 30 �l wereseparated by SDS-PAGE, respectively. Proteins were transferred toPVDF membrane (Millipore) using the semidry transfer method.Primary antibodies (phospho-c-Jun, �-actin, and phospho-ERK/ERK,Cell Signaling; phospho-p38/p38, Santa Cruz Biotechnology; MT1-

MMP, Chemicon) were incubated overnight at 4°C with gentle agi-tation and secondary antibodies (Amersham) for 1 h at room temper-ature. Bound antibodies were detected using chemiluminescence (Su-perSignal West Pico Chemiluminescent; Pierce). The film wasdeveloped and scanned, and densitometry was performed using Al-phaease software. Phosphorylated values were normalized to totalvalues to account for variability in loading.

JNK kinase assay. JNK was immunoprecipitated by incubating150–200 �g of protein from total cell lysates with 2 �g of JNKantibody (Upstate) for 1 h at 4°C with gentle rocking. ProteinA-agarose beads (Pierce) were then added to the lysates and allowedto bind to the JNK antibody for 1 h at 4°C with gentle rocking. Theantibody-protein A-agarose complex was collected by centrifugation,and the pellet was washed. The lysates were then incubated in thepresence of 1� assay buffer (20 mM MOPS, 25 mM �-glycerolphos-phate, pH 7.2, 1 mM EGTA, 1 mM sodium orthovanadate, and 1 mMdithiothreitol), magnesium/ATP cocktail (10 �M nonradioactive ATPand 75 mM MgCl2 in assay buffer), 1 �Ci/�l �-[32P]ATP (Amer-sham) and 2 �g/�l glutathione S-transferase-c-Jun (substrate; CellSignaling) at 37°C for 30 min with gentle shaking. The reaction wasterminated by the addition of denaturing loading buffer, and theproteins were separated by 10% SDS-PAGE as previously described.The gels were fixed with 50% methanol, 10% acetic acid, and 3%glycerol, transferred to Wattman paper, and vacuum dried over-night. Dry gels were then exposed to X-ray film for 1–7 days at�80°C and then developed, scanned, and analyzed using Alphaeasesoftware.

Surface biotinylation. SMEC were cultured on type I collagen for24 h and then stimulated with 25 ng/ml VEGF for 60 min. Cells werewashed with ice-cold PBS and incubated with 1 mg/ml Sulfo-NHS-biotin in PBS for 30 min on ice. The reaction was terminated bywashing the cells with 100 mM glycine for 20 min. Cells were thenlysed as previously described, and 75 �g of protein from total celllysates were incubated with streptavidin-agarose beads (Pierce) over-night at 4°C with gentle rocking. The streptavidin-agarose complexwas collected by centrifugation, and the pellet was washed with PBScontaining 0.1% Nonidet P-40. Next, 50 �l of 1� loading buffer wasadded, the samples were boiled, and the proteins were separated by10% SDS-PAGE as previously described.

Reverse transcription and quantitative real-time PCR. cDNA fromSMEC was produced without RNA isolation using the Cells-to-cDNAkit (Ambion, TX). Briefly, SMEC cells were washed three times insterilized PBS at 4°C and then heated (75°C for 15 min) in cell lysisbuffer II. Next, the cell lysate was treated with DNase 1 (0.004 U/�l)to degrade genomic DNA (37°C for 15 min), followed by inactivationof DNase (75°C for 5 min). The cell lysate was stored at �20°C untilit was used for the reverse transcription reaction. Cell lysate (10 �l)was reverse transcribed in a 20-�l reaction by using reagents from theCells-to-cDNA kit according to the manufacturer’s protocols. ThecDNA was diluted fourfold with RNase-free water. Quantitativereal-time PCR (Q-PCR) was performed using the ABI PRISM 7700sequence detection system. VIC-labeled control rRNA and a FAM-labeled MT1-MMP probe and primers set were purchased fromApplied Biosystems (catalog nos. P/N4308329 and Mm00485954-m1, respectively). Primers and TaqMan FAM-labeled probes forMMP-2 were designed using PrimerExpress 1.0 software(PerkinElmer Life Sciences): MMP-2 probe, 6FAM-caa tgc tga tggaca gcc ctg ca-MGBNFQ; forward primer, CCA TGA AGC CTTGTT TAC CA; reverse primer, CTG GAA GCG GAA CGG AAA(for siRNA experiments, FAM-labeled MMP-2 probe and primersset was purchased from Applied Biosystems, catalog no.Rn01538174_m1). A 25-�l reaction mixture contained 12.5 �l ofTaqMan universal PCR master mix (PCR Mix; Applied Biosystems),4 �l of cDNA template, and the appropriate concentrations of gene-specific primers and probe sets. PCR was performed with thermalconditions as follows: 50°C for 2 min, 95°C for 10 min, followed by40 cycles of 95°C for 15 s and 60°C for 1 min. Cycle threshold (Ct)

C580 JNK AND PI3K REGULATE MMP-2 AND MT1-MMP

AJP-Cell Physiol • VOL 291 • OCTOBER 2006 • www.ajpcell.org

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

values were used to determine the amount of MMP-2 and MT1-MMPmRNAs and 18S rRNA for all groups. The mean Ct values oftriplicate samples from each group were determined, and then �Ct(sample) was calculated according to the equation �Ct(sample) �average Ct(rRNA) � average Ct(sample). Changes in MMP-2 and MT1-MMP mRNA expression following VEGF or anisomycin treatmentwere calculated using the ��Ct method as described in the AppliedBiosystems manual (as described and validated in Ref. 39). ThemRNA expression levels of target genes were expressed relative to theappropriate untreated control, which was set to 1.0.

siRNA. c-Jun siRNA (30 nM; Ambion) or equal amounts ofnegative control were incubated in 50 �l of Opti-MEM containingsiPORT-neoFX (3 �l/ml; Ambion) for 15 min at room temperatureper the manufacturer’s instructions. The complexes were then addedto 45,000 SMEC plated in 24-well plates or 90,000 SMEC plated in12-well plates and incubated at 37°C for 48 or 72 h for mRNA orprotein analysis, respectively.

Statistics. Data were normalized to control values and are presentedas means SE relative to controls. Student’s t-test or one-wayANOVA, followed by Tukey’s post hoc tests, was applied to deter-mine statistical significance (P 0.05).

RESULTS

Depolymerization of the actin cytoskeleton induces MMP-2and MT1-MMP expression. Treatment of endothelial cells for24 h with 10 �M cytochalasin D resulted in cell rounding,consistent with complete depolymerization of the actin cytoskel-eton (data not shown). Previous studies showed that cytochalasinD treatment increases production and activation of MMP-2 infibroblasts, and we verified this finding in microvascular endothe-lial cells (18). Treatment of cells with cytochalasin D increased

Fig. 1. Cytochalasin D increases JNK activity, and JNKregulates matrix metalloproteinase-2 (MMP-2) and mem-brane type 1-matrix metalloproteinase (MT1-MMP) mRNAexpression. Rat skeletal muscle endothelial cells (SMEC)treated with cytochalasin D showed a significant increase inJNK activity as assessed by kinase assay (A). SMEC pre-treated with 100 �M SP600125 were treated with 10 �Mcytochalasin D for 24 h. MMP-2 production and activationwas measured using gelatin zymography (B). MMP-2 (C)and MT1-MMP mRNA (D) were quantified using Northernblot analysis. C, control; CD, cytochalasin D; SP,SP600125. Values are means SE, n � 3. *P 0.05 vs.control. #P 0.05 vs. cytochalasin D.

Fig. 2. Inhibition of p38 increases MMP-2 production and activation. SMECpretreated with SB-203580 were treated with 1 �M cytochalasin D for 24 h.MMP-2 production and activation were measured using gelatin zymography.SB, SB-203580. Values are means SE, n � 3. *P 0.05 vs. control. #P 0.05 vs. cytochalasin D.

C581JNK AND PI3K REGULATE MMP-2 AND MT1-MMP

AJP-Cell Physiol • VOL 291 • OCTOBER 2006 • www.ajpcell.org

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

total MMP-2 protein (1.96 0.29-fold vs. control, P � 0.05)and the percentage of active MMP-2 protein (42.87 5.6vs. 5. 28 4.42%, P � 0.05) as detected by gelatinzymography. Cytochalasin D treatment also increased MMP-2and MT1-MMP mRNA expression (1.59 0.06- and 1.90 0.19-fold vs. control, respectively, P � 0.05).

JNK, but not ERK1/2 or p38, regulates MMP-2 mRNA inresponse to actin cytoskeleton reorganization. The MAPKs areknown to be key regulators of MMP-2 and MT1-MMP inresponse to various stimuli (1, 4, 45). To determine the in-volvement of the MAPKs in controlling MMP-2 and MT1-MMP production induced by actin cytoskeleton rearrangement,

we pretreated endothelial cells with specific inhibitors to JNK,ERK1/2, or p38 before cytochalasin D treatment. CytochalasinD induced JNK activity in SMEC, with a trend for increasedJNK activity seen as early as 30 min posttreatment and asignificant increase at 4 h (Fig. 1A). Treatment of SMEC withSP600125 resulted in a nonsignificant attenuation of cytocha-lasin D-induced MMP-2 protein expression (P � 0.37) and hadno effect on active MMP-2 (cytochalasin D: 45.54 1.67% vs.cytochalasin D � SP600125: 45.80 1.69%) (Fig. 1B).Inhibition of JNK significantly attenuated MMP-2 mRNA andMT1-MMP mRNA (Fig. 1, C and D). Inhibition of ERK1/2with either PD-98059 or U0126 did not affect MMP-2 produc-

Fig. 3. Activation of JNK is sufficient to induce MMP-2 and MT1-MMP mRNA expression and requires c-Jun. SMEC were treated with 250 ng/ml or 1 �g/mlanisomycin to induce JNK activation. MMP-2 (A) and MT1-MMP mRNA (B) were measured using TaqMan real-time RT-PCR. C: SMEC were treated with30 nM c-Jun or negative control small interference RNA (siRNA) and then stimulated with 1 �g/ml anisomycin. aniso, Anisomycin; si, siRNA. Values aremeans SE, n � 3. *P 0.05 vs. control. #P 0.05 vs. anisomycin.

Fig. 4. MMP-2 protein production is dependent onphosphoinositide 3-kinase (PI3K). SMEC pre-treated with the PI3K inhibitor LY-294002 weretreated with 1 �M cytochalasin D for 24 h.MMP-2 (A) and MT1-MMP mRNA (B) werequantified using Northern blot analysis. Total (C)and active MMP-2 (D) was measured using gelatinzymography. LY, LY-294002. Values aremeans SE, n � 3. *P 0.05 vs. control. #P 0.05 vs. cytochalasin D.

C582 JNK AND PI3K REGULATE MMP-2 AND MT1-MMP

AJP-Cell Physiol • VOL 291 • OCTOBER 2006 • www.ajpcell.org

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

tion or activation in response to cytochalasin D treatment (datanot shown). Inhibition of p38 with SB203580 increasedMMP-2 protein expression and induced MMP-2 activationwithout altering endothelial cell morphology (Fig. 2). Additionof cytochalasin D to endothelial cells pretreated with SBfurther augmented MMP-2 protein production, suggesting thatinhibition of p38 stimulates MMP-2 protein production by apathway distinct from that initiated by cytochalasin D (Fig. 2).

Activation of JNK using anisomycin increased both MMP-2and MT1-MMP mRNA expression as evidenced by RT-PCR(Fig. 3, A and B). On the basis of data supporting a role forJNK in the regulation of MMP-2 and MT1-MMP, we exam-ined the involvement of c-Jun as a transcriptional regulator ofMMP-2 and MT1-MMP. Transfection of c-Jun siRNA intoSMEC decreased basal levels of c-Jun protein by 40% (data notshown). c-Jun siRNA inhibited the anisomycin-inducedMMP-2 mRNA expression (Fig. 3C).

PI3K is required for MMP-2 protein production and acti-vation. We considered that PI3K could be involved in theregulation of MMP-2 and MT1-MMP in response to cytoskel-etal remodeling, either by directly activating JNK or by acti-vating a parallel pathway by way of Akt (49). Inhibition ofPI3K with LY-294002 did not affect cytochalasin D-inducedMMP-2 mRNA expression (Fig. 4A) or cytochalasin D-in-duced MT1-MMP mRNA expression (Fig. 4B). On the otherhand, total MMP-2 protein was decreased with PI3K inhibition(Fig. 4C), as was the amount of active MMP-2 protein (Fig. 4D).

Simultaneous inhibition of PI3K and JNK (100 �MSP600125 and 50 �M LY-294002) abolished the cytochalasinD-induced increase in total MMP-2 protein expression (Fig.5A) with a significant decrease in the amount of active MMP-2protein (Fig. 5B). The decrease in active MMP-2 protein(shown in Figs. 4D and 5B) suggested that there was aconcurrent decrease in MT1-MMP protein, and this was veri-fied with Western blotting (Fig. 5C). Similar to the effect ofJNK inhibition alone (Fig. 1), dual inhibition of JNK and PI3Kprevented a cytochalasin D-mediated increase in MMP-2 (cy-tochalasin D: 1.81 0.12 above control vs. cytochalasin D �LY-294002/SP600125: 0.94 0.23 relative to control) orMT1-MMP mRNA expression (cytochalasin D: 1.46 0.13above control vs. cytochalasin D � LY-294002/SP600125:1.1 0.23 above control).

VEGF induces production and activation of MMP-2 via JNKand PI3K. VEGF, which plays a significant role as a growthfactor in activity-induced angiogenesis in skeletal muscle,induces modification of the actin cytoskeleton and influencesendothelial cell permeability and migration during angiogene-sis (15, 24, 58). Because VEGF activates JNK (47), wehypothesized that VEGF mediates increased MMP-2 and MT1-MMP expression in SMEC via JNK. Stimulation of SMECwith VEGF (25 ng/ml) caused nuclear translocation of phos-pho-JNK and increased total cellular phospho-JNK staining(Fig. 6, A and B). MMP-2 protein secretion, production, andactivation were increased in response to VEGF treatment for24 h (Fig. 6, C–E). Similar to the cytochalasin D-inducedMMP-2 response, VEGF-induced protein production and acti-vation of MMP-2 were inhibited by pretreatment with 10 �MLY-294002 (Fig. 6, D and E). RT-PCR analysis showed thatVEGF increased MMP-2 mRNA after 24 h, and JNK inhibitionwith SP600125 blocked the VEGF-induced increase in MMP-2mRNA expression (Fig. 6F). c-Jun siRNA treatment also

significantly attenuated the VEGF-induced increase in MMP-2mRNA expression (Fig. 6G).

VEGF increases MT1-MMP protein, but not mRNA, expres-sion. Because VEGF induced activation of MMP-2, we testedthe effects of VEGF on MT1-MMP. Interestingly, VEGF did notinduce an upregulation in MT1-MMP mRNA (Fig. 7A). TotalMT1-MMP was increased after 24 h of VEGF stimulation andwas abrogated by pretreatment with 50 �M LY-294002 (Fig. 7B).VEGF is known to increase the amount of MT1-MMP on the cellsurface and to cause its localization to caveolae (34, 35). Consis-tent with these reports, we found that short exposure (60 min) toVEGF increased the amount of MT1-MMP on the cell surface asmeasured by surface biotinylation, with no change in the amountof MT1-MMP protein in total cell lysates (Fig. 7C).

Fig. 5. Increased MMP-2 and MT1-MMP protein production in response tocytochalasin D is mediated by PI3K/JNK. SMEC pretreated with 50 �MSP600125 and 50 �M LY-294002 were treated with 1 �M cytochalasin D for24 h. MMP-2 production (A) and activation (B) were measured using gelatinzymography. MT1-MMP protein level (C) was measured by Western blotting.Values are means SE, n � 3. *P 0.05 vs. control. #P 0.05 vs.cytochalasin D.

C583JNK AND PI3K REGULATE MMP-2 AND MT1-MMP

AJP-Cell Physiol • VOL 291 • OCTOBER 2006 • www.ajpcell.org

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

DISCUSSION

This study provides novel evidence that JNK regulatesboth MMP-2 and MT1-MMP mRNA expression in responseto actin cytoskeleton reorganization in microvascular endo-

thelial cells. Downstream of JNK, c-Jun is a regulator ofMMP-2 transcription. VEGF, a potent vascular permeabilityfactor and angiogenic mediator, induces MMP-2 mRNAexpression via JNK and increases MMP-2 protein produc-tion and activation. PI3K appears to regulate MMP-2 and

Fig. 6. VEGF-induced MMP-2 mRNA expres-sion is dependent on JNK. SMEC in the absence(A) or presence (B) of 25 ng/ml VEGF werestained for phospho-JNK (red) and 4�,6�-diami-nodino-2-phenylindole (blue, nuclei). Arrowsdenote nuclear localization of phospho-JNK.Media were collected from SMEC treated with25 ng/ml VEGF, and MMP-2 production wasanalyzed using gelatin zymography (C). SMECpretreated with 50 �M SP600125 and/or 50 �MLY-294002 were treated with 25 ng/ml VEGFfor 24 h. MMP-2 production (D) and activation(E) were measured using gelatin zymography.SMEC pretreated with 50 �M SP600125 (F) or30 nM c-Jun or negative control siRNA (G)were then stimulated with 25 ng/ml VEGF for24 h, and MMP-2 mRNA was measured usingTaqMan real-time RT-PCR. Values aremeans SE, n � 3 (n � 5 for C). *P 0.05 vs.control. #P 0.05 vs. VEGF.

C584 JNK AND PI3K REGULATE MMP-2 AND MT1-MMP

AJP-Cell Physiol • VOL 291 • OCTOBER 2006 • www.ajpcell.org

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

MT1-MMP protein processing independently of mRNA syn-thesis.

Reorganization of the endothelial actin cytoskeleton occursvery early in the initiation of angiogenesis, in response togrowth factors, cytokines, and inflammatory mediators or as aresult of mechanotransduction (22, 31, 47, 55). This reorgani-zation may contribute to the process by which quiescentendothelial cells become activated and make the “angiogenicswitch.” Initially, we avoided the pleiotropic signals initiatedby growth factors by using cytochalasin D to initiate cytoskel-etal alterations independently of a specific external stimulus.Our results point to JNK as a transcriptional regulator ofMMP-2 and MT1-MMP in response to reorganization of theactin cytoskeleton. JNK regulates transcription through activa-tion of members of the AP-1 protein family including c-Jun,Jun B, Jun D, NFAT, egr-1, and ATF-2 (14, 37). Utilizingc-Jun siRNA, we identified c-Jun as an important transcrip-tional regulator of MMP-2 expression in SMEC.

A key role for JNK in mediating numerous steps of theangiogenic process is gaining recognition. JNK has been im-plicated in endothelial cell migration and proliferation as wellas matrix invasion and network formation (38, 60). Recently,Zhang et al. (60), using DNAzymes to inhibit c-Jun, were ableto inhibit aspects of the angiogenic response following VEGFstimulation, including migration, chemoinvasion, and tubuleformation. We have shown in the present study that inhibitionof JNK or c-Jun attenuates VEGF-induced MMP-2 mRNAexpression, thus extending our knowledge of the roles playedby JNK in these angiogenesis assays and physiological angio-genesis.

PI3K activates the Akt signaling pathway but also isreported to be upstream of JNK in a number of cellularprocesses, including endothelial cell migration (49). Cy-tochalasin D has been shown to either increase or decreasephospho-Akt levels in different cell lines (32, 51). Ourresults suggest that PI3K regulates MMP-2 and MT1-MMPprotein independently of mRNA synthesis in response toboth Cytochalasin D and VEGF. Both total and activeMMP-2 protein were partially, but significantly, decreasedwith LY-294002 without a concurrent inhibition of MMP-2mRNA. Likewise, MT1-MMP protein decreased indepen-dently of changes in MT1-MMP mRNA, and similar effectsof LY-294002 on MT1-MMP translation were observed in

smooth muscle cells in response to balloon injury (59).Several possibilities exist to explain the results observedwith PI3K inhibition: 1) inhibition of MMP-2 and MT1-MMP mRNA translation, 2) inhibition of glycogen synthasekinase-3� (GSK-3�)-mediated effects, and 3) PI3K-depen-dent protein trafficking.

PI3K modulates mRNA translation by modifying the 70-kDa ribosomal S6 kinase (S6K) through phosphoinositide-dependent kinase 1 (PDK) or 4E-binding protein 1 (4E-BP1)via the mTOR (mammalian target of rapamycin) pathway (57).Activation of S6K by PI3K has been shown to be required forMMP-2 translation, whereas the translation factor e1F-4B (andits inhibitory binding protein, 4E-BP1) regulates MT1-MMPtranslation in smooth muscle cells (10). These multiple effectson mRNA translation by PI3K may explain the effects ob-served in MMP-2 and MT1-MMP protein in response to PI3Kinhibition.

A second possible explanation for the effect of PI3Kinhibition is that a decrease in PI3K-dependent phosphory-lation of GSK-3� alters GSK-3� signals within the cell.GSK-3� has a broad range of cellular functions, includingregulation of protein transport, mRNA translation (throughthe activation of eIF-2B) (46), and suppression of c-Jun(44). It is possible that altered GSK-3� activity is respon-sible for the observed changes in active MMP-2 by alteringMT1-MMP translation and its localization within the cell.Under quiescent conditions, much of the cellular MT1-MMP is localized to endosomal compartments and, uponstimulation, rapidly moves to the cell surface to areas ofactive angiogenesis (18). PI3K, through the activation ofPIKfyve, causes redistribution of the glucose transporterGLUT-4 receptor from endosomes in response to insulin,thereby coordinating its trafficking and sorting (6, 50). It ispossible that a similar PI3K-dependent transport mechanismexists for MT1-MMP trafficking. This would be consistentwith our results, because we measured a decrease in theamount of active MMP-2 with PI3K inhibition without adecrease in MT1-MMP mRNA, suggesting that fewer MT1-MMP molecules were on the cell surface after treatmentwith LY-294009. Ultimately, our data support differingroles of JNK and PI3K in the regulation of MMP-2 andMT1-MMP gene products, through transcriptional regula-tion and posttranscriptional protein processing, respectively,

Fig. 7. VEGF-induced MT1-MMP protein production is dependent on PI3K. SMEC pretreated with 50 �M SP600125 were then stimulated with 25 ng/ml VEGFfor 24 h. MT1-MMP mRNA was measured using TaqMan real-time RT-PCR (A). SMEC pretreated with 50 �M SP600125 and/or 50 �M LY-294002 weretreated with 25 ng/ml VEGF for 24 h. MT1-MMP protein level was measured using Western blotting (B). Cell surface MT1-MMP protein in response toshort-term VEGF treatment (25 ng/ml for 60 min) was measured using surface biotinylation (C). Values are means SE, n � 3. *P 0.05 vs. control.

C585JNK AND PI3K REGULATE MMP-2 AND MT1-MMP

AJP-Cell Physiol • VOL 291 • OCTOBER 2006 • www.ajpcell.org

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

rather than the two acting in series through a direct signalcascade as originally hypothesized.

The observed increase in percent active MMP-2 proteinindependent of an increase in MT-MMP mRNA when JNKwas inhibited (Fig. 1, B and D) is consistent with previousreports that actin cytoskeleton depolymerization increasesthe number of MT1-MMP molecules on the cell surface(63). Increased MMP-2 activation was detected in cellextracts as early as 2 h after cytochalasin D treatment (datanot shown). Endocytosis of cell surface localized MT1-MMP occurs by both clathrin- and caveolae-dependent in-ternalization (3, 19, 28, 52). Both of these processes rely onintact actin scaffolding; thus disruption of the actin cy-toskeleton may result in an increase in MT1-MMP mole-cules on the cell surface. Zucker et al. (63) showed that actindepolymerization increases the number of cell surface re-ceptors for tissue inhibitor of MMP (TIMP)-2 (i.e., MT1-MMP) without altering the binding affinity for TIMP-2. Inour study, VEGF-induced activation of MMP-2 without aconcurrent increase in MT1-MMP mRNA correlates with anincreased amount of cell surface MT1-MMP. This finding isin line with recent observations by Labrecque et al. (34),who showed that VEGF increases cell surface MT1-MMPthrough a Src-dependent mechanism.

Our laboratory (48) previously reported that VEGF didnot increase MMP-2 protein production in rat endothelialcells isolated from the epididymal fat pad. We have ob-served several differences in MAPK signaling that mayunderlie this differential responsiveness to VEGF. Mostnotably, in rat endothelial cells, the MMP-2 and MT1-MMPresponse to three-dimensional collagen occurs via ERK1/2and is not affected by JNK inhibition (9). This suggests thatendothelial cells of different origin utilize unique combina-tions of signal molecules, which may account for the dif-ferences in VEGF responsiveness. Notably, VEGF-inducedMMP-2 secretion has been observed in human cells, imply-ing that the signal pathways are conserved across species(25, 36).

Upstream activators of PI3K and JNK may include Rac1,which is involved in actin cytoskeleton signaling throughstimulation of lamellipodia and regulation of cell-cell junc-tion during migration (12, 40). VEGF activates Rac1 (15),and Rac1 has been linked to increased MMP-2 productionand activation in fibrocarcinoma cells (62). Future work isneeded to further elucidate the VEGF-dependent signals thatmediate MMP-2 production in endothelial cells.

In summary, endothelial cell production of both MMP-2 andMT1-MMP is induced by actin cytoskeleton destabilization.MMP-2 and MT1-MMP mRNA expression requires activationof JNK, whereas MMP-2 protein production and activationare dependent on PI3K. The dependence of VEGF-inducedMMP-2 mRNA expression on JNK/c-Jun activity demonstratesthat this signal pathway may be a component of physiologicaland pathological angiogenesis. Together, these data illustratethat initial remodeling of the actin cytoskeleton may coordinatethe angiogenesis process, linking early angiogenic responses tosubsequent stages that require the production of proteasesresponsible for the degradation of the basement membrane andinterstitial matrix.

ACKNOWLEDGMENTS

We thank G. Dhanota and S. Pallan for technical assistance.

GRANTS

Funding for this project was granted to T. L. Haas from the NationalSciences and Engineering Research Council and the Canadian Institutes ofHealth Research.

REFERENCES

1. Ailenberg M and Silverman M. Cytochalasin D disruption of actinfilaments in 3T3 cells produces an anti-apoptotic response by activatinggelatinase A extracellularly and initiating intracellular survival signals.Biochim Biophys Acta 1593: 249–258, 2003.

2. Aimes RT and Quigley JP. Matrix metalloproteinase-2 is an interstitialcollagenase. J Biol Chem 270: 5872–5876, 1995.

3. Annabi B, Lachambre M, Bousquet-Gagnon N, Page M, Gingras D,and Beliveau R. Localization of membrane-type 1 matrix metalloprotein-ase in caveolae membrane domains. Biochem J 353: 547–553, 2001.

4. Ara T, Deyama Y, Yoshimura Y, Higashino F, Shindoh M, Matsu-moto A, and Fukuda H. Membrane type 1-matrix metalloproteinaseexpression is regulated by E-cadherin through the suppression of mitogen-activated protein kinase cascade. Cancer Lett 157: 115–121, 2000.

5. Barberis L, Wary KK, Fiucci G, Liu F, Hirsch E, Brancaccio M,Altruda F, Tarone G, and Giancotti FG. Distinct roles of the adaptorprotein Shc and focal adhesion kinase in integrin signaling to ERK. J BiolChem 275: 36532–36540, 2000.

6. Berwick DC, Dell GC, Welsh GI, Heesom KJ, Hers I, Fletcher LM,Cooke FT, and Tavare JM. Protein kinase B phosphorylation of PIKfyveregulates the trafficking of GLUT4 vesicles. J Cell Sci 117: 5985–5993,2004.

7. Birkedal-Hansen H. Proteolytic remodeling of extracellular matrix. CurrOpin Cell Biol 7: 728–735, 1995.

8. Birukov KG, Birukova AA, Dudek SM, Verin AD, Crow MT, Zhan X,DePaola N, and Garcia JGN. Shear stress-mediated cytoskeletal remod-eling and cortactin translocation in pulmonary endothelial cells. Am JRespir Cell Mol Biol 26: 453–464, 2002.

9. Boyd PJ, Doyle J, Gee E, Pallan S, and Haas TL. Mitogen-activatedprotein kinase signaling regulates endothelial cell assembly into networksand the expression of MT1-MMP and MMP-2. Am J Physiol Cell Physiol288: C659–C668, 2005.

10. Bradley JM, Kelley MJ, Rose A, and Acott TS. Signaling pathwaysused in trabecular matrix metalloproteinase response to mechanicalstretch. Invest Ophthalmol Vis Sci 44: 5174–5181, 2003.

11. Civelekoglu G, Tardy Y, and Meister JJ. Modeling actin filamentreorganization in endothelial cells subjected to cyclic stretch. Bull MathBiol 60: 1017–1037, 1998.

12. Connolly JO, Simpson N, Hewlett L, and Hall A. Rac regulatesendothelial morphogenesis and capillary assembly. Mol Biol Cell 13:2474–2485, 2002.

13. D’Angelo G, Struman I, Martial J, and Weiner RI. Activation ofmitogen-activated protein kinases by vascular endothelial growth factorand basic fibroblast growth factor in capillary endothelial cells is inhibitedby the antiangiogenic factor 16-kDa N-terminal fragment of prolactin.Proc Natl Acad Sci USA 92: 6374–6378, 1995.

14. Davis RJ. Signal transduction by the JNK group of MAP kinases. Cell103: 239–252, 2000.

15. Eriksson A, Cao RH, Roy J, Tritsaris K, Wahlestedt C, Dissing S,Thyberg J, and Cao YH. Small GTP-binding protein Rac is an essentialmediator of vascular endothelial growth factor-induced endothelial fenes-trations and vascular permeability. Circulation 107: I532–I538, 2003.

16. Fang J, Shing Y, Wiederschain D, Yan L, Butterfield C, Jackson G,Harper J, Tamvakopoulos G, and Moses MA. Matrix metalloprotein-ase-2 is required for the switch to the angiogenic phenotype in a tumormodel. Proc Natl Acad Sci USA 97: 3884–3889, 2000.

17. Franke RP, Grafe M, Schnittler H, Seiffge D, Mittermayer C, andDrenckhahn D. Induction of human vascular endothelial stress fibers byfluid shear-stress. Nature 307: 648–649, 1984.

18. Galvez BG, Matias-Roman S, Yanez-Mo M, Sanchez-Madrid F, andArroyo AG. ECM regulates MT1-MMP localization with �1 or �v�3integrins at distinct cell compartments modulating its internalization andactivity on human endothelial cells. J Cell Biol 159: 509–521, 2002.

19. Galvez BG, Matias-Roman S, Yanez-Mo M, Vicente-Manzanares M,Sanchez-Madrid F, and Arroyo AG. Caveolae are a novel pathway for

C586 JNK AND PI3K REGULATE MMP-2 AND MT1-MMP

AJP-Cell Physiol • VOL 291 • OCTOBER 2006 • www.ajpcell.org

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

membrane-type 1 matrix metalloproteinase traffic in human endothelialcells. Mol Biol Cell 15: 678–687, 2004.

20. Haas TL, Davis SJ, and Madri JA. Three dimensional type I collagenlattices induce coordinate expression of matrix metalloproteinases MT1-MMP and MMP-2 in microvascular endothelial cells. J Biol Chem 273:3604–3610, 1998.

21. Han XY, Boyd PJ, Colgan S, Madri JA, and Haas TL. Transcriptionalup-regulation of endothelial cell matrix metalloproteinase-2 in response toextracellular cues involves GATA-2. J Biol Chem 278: 47785–47791,2003.

22. Haralabopoulos GC, Grant DS, Kleinman HK, and MaragoudakisME. Thrombin promotes endothelial cell alignment in Matrigel in vitroand angiogenesis in vivo. Am J Physiol Cell Physiol 273: C239–C245,1997.

23. Herren B, Levkau B, Raines EW, and Ross R. Cleavage of �-cateninand plakoglobin and shedding of VE-cadherin during endothelial apopto-sis: evidence for a role for caspases and metalloproteinases. Mol Biol Cell9: 1589–1601, 1998.

24. Hoeben A, Landuyt B, Highley MS, Wildiers H, van Oosterom AT,and De Bruijn EA. Vascular endothelial growth factor and angiogenesis.Pharmacol Rev 56: 549–580, 2004.

25. Hoffmann S, He S, Ehren M, Ryan SJ, Wiedemann P, and Hinton DR.MMP-2 and MMP-9 secretion by RPE is stimulated by angiogenicmolecules found in choroidal neovascular membranes. Retina 26: 454–461, 2006.

26. Holmbeck K, Bianco P, Caterina J, Yamada S, Kromer M, KuznetsovSA, Mankani M, Robey PG, Poole AR, Pidoux I, Ward JM, andBirkedal-Hansen H. MT1-MMP-deficient mice develop dwarfism, os-teopenia, arthritis, and connective tissue disease due to inadequate colla-gen turnover. Cell 99: 81–92, 1999.

27. Itoh T, Tanioka M, Yoshida H, Yoshioka T, Nishimoto H, and ItoharaS. Reduced angiogenesis and tumor progression in gelatinase A-deficientmice. Cancer Res 58: 1048–1051, 1998.

28. Jiang A, Lehti K, Wang X, Weiss SJ, Keski-Oja J, and Pei D.Regulation of membrane-type matrix metalloproteinase 1 activity bydynamin-mediated endocytosis. Proc Natl Acad Sci USA 98: 13693–13698, 2001.

29. Kato T, Kure T, Chang JH, Gabison EE, Itoh T, Itohara S, and AzarDT. Diminished corneal angiogenesis in gelatinase A-deficient mice.FEBS Lett 508: 187–190, 2001.

30. Keely PJ, Westwick JK, Whitehead IP, Der CJ, and Parise LV. Cdc42and Rac1 induce integrin-mediated cell motility and invasiveness throughPI(3)K. Nature 390: 632–636, 1997.

31. Kilarski WW, Jura N, and Gerwins P. Inactivation of Src familykinases inhibits angiogenesis in vivo: implications for a mechanisminvolving organization of the actin cytoskeleton. Exp Cell Res 291:70 – 82, 2003.

32. Kim SJ, Hwang SG, Kim IC, and Chun JS. Actin cytoskeletal archi-tecture regulates nitric oxide-induced apoptosis, dedifferentiation, andcyclooxygenase-2 expression in articular chondrocytes via mitogen-acti-vated protein kinase and protein kinase C pathways. J Biol Chem 278:42448–42456, 2003.

33. Kiosses WB, Cheresh DA, Yang S, Gerritsen ME, Cheresh DA,Alderson N, and Schwartz MA. A dominant negative p65 PAK peptideinhibits angiogenesis. Circ Res 90: 697–702, 2002.

34. Labrecque L, Nyalendo C, Langlois S, Durocher Y, Roghi C, MurphyG, Gingras D, and Beliveau R. Src-mediated tyrosine phosphorylation ofcaveolin-1 induces its association with membrane type 1 matrix metallo-proteinase. J Biol Chem 279: 52132–52140, 2004.

35. Labrecque L, Royal I, Surprenant DS, Patterson C, Gingras D, andBeliveau R. Regulation of vascular endothelial growth factor receptor-2activity by caveolin-1 and plasma membrane cholesterol. Mol Biol Cell14: 334–347, 2003.

36. Lamoreaux WJ, Fitzgerald ME, Reiner A, Hasty KA, and CharlesST. Vascular endothelial growth factor increases release of gelatinaseA and decreases release of tissue inhibitor of metalloproteinases bymicrovascular endothelial cells in vitro. Microvasc Res 55: 29 – 42,1998.

37. Lim CP, Jain N, and Cao X. Stress-induced immediate-early gene, egr-1,involves activation of p38/JNK1. Oncogene 16: 2915–2926, 1998.

38. Meadows KN, Bryant P, Vincent PA, and Pumiglia KM. Activated Rasinduces a proangiogenic phenotype in primary endothelial cells. Oncogene23: 192–200, 2004.

39. Milkiewicz M and Haas TL. Effect of mechanical stretch on HIF-1� andMMP-2 expression in capillaries isolated from overloaded skeletal mus-cles: laser capture microdissection study. Am J Physiol Heart Circ Physiol289: H1315–H1320, 2005.

40. Minden A, Lin AN, Claret FX, Abo A, and Karin M. Selectiveactivation of the JNK signaling cascade and c-Jun transcriptional activityby the small GTPases Rac and Cdc42Hs. Cell 81: 1147–1157, 1995.

41. Murphy G, Nguyen Q, Cockett MI, Atkinson SJ, Allan JA, KnightCG, Willenbrock F, and Docherty AJP. Assessment of the role of thefibronectin-like domain of gelatinase-A by analysis of a deletion mutant.J Biol Chem 269: 6632–6636, 1994.

42. Oh J, Takahashi R, Adachi E, Kondo S, Kuratomi S, Noma A,Alexander DB, Motoda H, Okada A, Seiki M, Itoh T, Itohara S,Takahashi C, and Noda M. Mutations in two matrix metalloproteinasegenes, MMP-2 and MT1-MMP, are synthetic lethal in mice. Oncogene 23:5041–5048, 2004.

43. Ohno-Matsui K, Uetama T, Yoshida T, Hayano M, Itoh T, Morita I,and Mochizuki M. Reduced retinal angiogenesis in MMP-2-deficientmice. Invest Ophthalmol Vis Sci 44: 5370–5375, 2003.

44. Oudit GY, Sun H, Kerfant BG, Crackower MA, Penninger JM, andBackx PH. The role of phosphoinositide-3 kinase and PTEN incardiovascular physiology and disease. J Mol Cell Cardiol 37: 449 –471, 2004.

45. Pan MR and Hung WC. Nonsteroidal anti-inflammatory drugs inhibitmatrix metalloproteinase-2 via suppression of the ERK/Sp1-mediatedtranscription. J Biol Chem 277: 32775–32780, 2002.

46. Pap M and Cooper GM. Role of translation initiation factor 2B incontrol of cell survival by the phosphatidylinositol 3-kinase/Akt/glycogen synthase kinase 3� signaling pathway. Mol Cell Biol 22:578 –586, 2002.

47. Pedram A, Razandi M, and Levin ER. Deciphering vascular endothelialcell growth factor/vascular permeability factor signaling to vascular per-meability—Inhibition by atrial natriuretic peptide. J Biol Chem 277:44385–44398, 2002.

48. Rivilis I, Milkiewicz M, Boyd P, Goldstein J, Brown MD, EggintonS, Hansen FM, Hudlicka O, and Haas TL. Differential involvementof MMP-2 and VEGF during muscle stretch- versus shear stress-induced angiogenesis. Am J Physiol Heart Circ Physiol 283: H1430 –H1438, 2002.

49. Steinle JJ, Meininger CJ, Chowdhury U, Wu G, and Granger HJ.Role of ephrin B2 in human retinal cell proliferation and migration. CellSignal 15: 1011–1017, 2003.

50. Tanti JF, Grillo S, Gremeaux T, Coffer PJ, Van Obberghen E, andMarchand-Brustel Y. Potential role of protein kinase B in glucosetransporter 4 translocation in adipocytes. Endocrinology 138: 2005–2010,1997.

51. Tran NL, Adams DG, Vaillancourt RR, and Heimark RL. Signaltransduction from N-cadherin increases Bcl-2. Regulation of the phospha-tidylinositol 3-kinase/Akt pathway by homophilic adhesion and actincytoskeletal organization. J Biol Chem 277: 32905–32914, 2002.

52. Uekita T, Itoh Y, Yana I, Ohno H, and Seiki M. Cytoplasmic tail-dependent internalization of membrane-type 1 matrix metalloproteinase isimportant for its invasion-promoting activity. J Cell Biol 155: 1345–1356,2001.

53. Werb Z. ECM and cell surface proteolysis: regulating cellular ecology.Cell 91: 439–442, 1997.

54. Woessner JF Jr. The family of matrix metalloproteinases. Ann NY AcadSci 732: 11–21, 1994.

55. Wojciak-Stothard B, Potempa S, Eichholtz T, and Ridley AJ. Rho andRac but not Cdc42 regulate endothelial cell permeability. J Cell Sci 114:1343–1355, 2001.

56. Wojciak-Stothard B and Ridley AJ. Shear stress-induced endothelialcell polarization is mediated by Rho and Rac but not Cdc42 or PI3-kinases. J Cell Biol 161: 429–439, 2003.

57. Wymann MP and Marone R. Phosphoinositide 3-kinase in disease:timing, location, and scaffolding. Curr Opin Cell Biol 17: 141–149,2005.

58. Zachary I. VEGF signalling: integration and multi-tasking in endothelialcell biology. Biochem Soc Trans 31: 1171–1177, 2003.

59. Zahradka P, Harding G, Litchie B, Thomas S, Werner JP, Wilson DP,and Yurkova N. Activation of MMP-2 in response to vascular injury ismediated by phosphatidylinositol 3-kinase-dependent expression of MT1-MMP. Am J Physiol Heart Circ Physiol 287: H2861–H2870, 2004.

C587JNK AND PI3K REGULATE MMP-2 AND MT1-MMP

AJP-Cell Physiol • VOL 291 • OCTOBER 2006 • www.ajpcell.org

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from

60. Zhang GS, Dass CR, Sumithran E, Di Girolamo N, Sun LQ, andKhachigian LM. Effect of deoxyribozymes targeting c-Jun on solid tumorgrowth and angiogenesis in rodents. J Natl Cancer Inst 96: 683–696, 2004.

61. Zhou Z, Apte SS, Soininen R, Cao R, Baaklini GY, Rauser RW, WangJ, Cao Y, and Tryggvason K. Impaired endochondral ossification andangiogenesis in mice deficient in membrane-type matrix metalloproteinaseI. Proc Natl Acad Sci USA 97: 4052–4057, 2000.

62. Zhuge Y and Xu J. Rac1 mediates type I collagen-dependent MMP-2activation. Role in cell invasion across collagen barrier. J Biol Chem 276:16248–16256, 2001.

63. Zucker S, Hymowitz M, Conner CE, DiYanni EA, and Cao J. Rapidtrafficking of membrane type 1-matrix metalloproteinase to the cellsurface regulates progelatinase a activation. Lab Invest 82: 1673–1684,2002.

C588 JNK AND PI3K REGULATE MMP-2 AND MT1-MMP

AJP-Cell Physiol • VOL 291 • OCTOBER 2006 • www.ajpcell.org

on Septem

ber 14, 2006 ajpcell.physiology.org

Dow

nloaded from