of June 18, 2018.This information is current as

Role of Vascular Endothelial Growth FactorT Lymphocytes of Mammary Tumor Bearers: Up-Regulation of Matrix Metalloproteinase-9 in

LopezGunja-Smith, Lynn M. Herbert, Joseph F. Grosso and Diana M. Jennifer L. Owen, Vijaya Iragavarapu-Charyulu, Zeenat

http://www.jimmunol.org/content/171/8/4340doi: 10.4049/jimmunol.171.8.4340

2003; 171:4340-4351; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/171/8/4340.full#ref-list-1

, 27 of which you can access for free at: cites 83 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2003 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on June 18, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Up-Regulation of Matrix Metalloproteinase-9 in TLymphocytes of Mammary Tumor Bearers: Role of VascularEndothelial Growth Factor1

Jennifer L. Owen,* Vijaya Iragavarapu-Charyulu,§ Zeenat Gunja-Smith,† Lynn M. Herbert,*Joseph F. Grosso,* and Diana M. Lopez2*‡

Matrix metalloproteinase-9 (MMP-9), a matrix-degrading enzyme, is crucial in tumor invasion and metastasis and is implicatedin leukocyte extravasation. In this report, we demonstrate that during growth of the D1–7,12-dimethylbenzanthracene-3 mam-mary tumor in BALB/c mice, there is progressive up-regulation of MMP-9 in splenic T cells at both the transcriptional andtranslational levels. Our previous work has identified several factors produced by this tumor, including PGE2, GM-CSF, andphosphatidyl serine; however, none of these agents induces increased production of MMP-9 by normal splenic T cells. Althoughnot produced by the tumor, TNF-� and IL-6 are up-regulated in both macrophages and B cells in tumor-bearing mice. Exposureof normal T cells to these two cytokines, however, also fails to up-regulate MMP-9 production. Vascular endothelial growth factor(VEGF) is produced by many tumors, and we determined that the mammary tumor used in our studies expresses high levels ofthis angiogenic growth factor. Importantly, splenic T cells from tumor bearers constitutively produce increased amounts of VEGF,and treatment of normal T cells with VEGF results in up-regulated MMP-9 production. Of crucial importance is the finding thattumor-infiltrating T cells also produce high levels of VEGF and MMP-9. Our studies indicate that VEGF can act directly on Tlymphocytes and that elevated VEGF levels may contribute to the aberrant MMP-9 secretion by mammary tumor bearers’ Tcells. The Journal of Immunology, 2003, 171: 4340–4351.

M atrix metalloproteinases (MMPs)3 are a family ofmore than 20 zinc-dependent endopeptidases that candegrade every component of the extracellular matrix

(ECM), enabling cells to invade tissues. Because of their key rolesin matrix degradation, activated members of the MMP family areinvolved in many physiological processes such as ovulation, em-bryogenesis, bone remodeling, tooth eruption, wound healing, andangiogenesis. In contrast, excessive or inappropriate expression ofMMPs contributes to tissue-destructive diseases including arthri-tis, multiple sclerosis, atherosclerosis, gastric ulcers, fibrotic lungdisease, and tumor progression (1). Consequently, these proteasesplay a crucial role in the pathological processes of tumor invasionand metastasis (2–6) and contribute to chronic inflammatory dis-eases by aiding the migration of inflammatory cells into varioustissues (7, 8).

The metalloproteinases have elevated expression in advancedtumors, and a number of studies outline roles for these enzymes

during multiple stages of tumorigenesis, including establishmentand growth, angiogenesis, invasion, migration, and metastasis (9).The pluripotent effects of MMPs are due to the ability of theseproteases to regulate the tumor microenvironment by cleaving avariety of substrates, which include not only structural componentsof the ECM, but also growth factor-binding proteins, growth factorprecursors, receptor tyrosine kinases, cell-adhesion molecules, andother proteinases (10, 11). Accordingly, the gene expression ofMMPs is tightly regulated to maintain normal tissue function. Atthe protein level, regulation is achieved by the two-pronged actionof proteolytic activation of the latent enzyme and inhibition ofactive enzymes by the tissue inhibitors of metalloproteinases(TIMPs) (12, 13).

Several MMPs, including the gelatinases, MMP-2 (gelatinaseA) and MMP-9 (gelatinase B), are expressed at very high levels intumors of various origins and in various species (14), and serumand tissue levels of both, especially MMP-9, correlate with ma-lignant tumor grade (15–18). Although many studies had evaluatedthe production of proteases by tumor cells, immunohistochemicalanalyses revealed that MMPs are detected not only in the tumorcells themselves, but also in the stromal cells and, in some cases,preferentially in the latter (19, 20). Initially it was thought thatfibroblast-like cells were the source of the stromal cell-derivedMMPs. However, a series of reports from several laboratories havesuggested that inflammatory cells invading the ECM are importantproducers of these enzymes. In a squamous cell carcinoma model,MMP-9 was primarily expressed in neutrophils, macrophages, andmast cells rather than in the neoplastic cells (21), whereas hostmacrophages were concluded to be the main source of MMP-9 inhuman ovarian carcinoma in mice (22). Additionally, it has beenreported that, in epithelial cancer invasion, most MMPs are madeby stromal cells of the host, and their expression is regulated bychemokines and cytokines produced by the tumor (23).

Departments of *Microbiology and Immunology and †Medicine, University of MiamiSchool of Medicine, Miami, FL 33101; ‡Sylvester Cancer Center, Miami, FL 33101;and §Department of Biomedical Sciences, Florida Atlantic University, Boca Raton,FL 33431

Received for publication March 17, 2003. Accepted for publication August 13, 2003.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by Grant R01 CA25583 from the National Institutes ofHealth.2 Address correspondence and reprint requests to Dr. Diana M. Lopez, Department ofMicrobiology and Immunology, University of Miami School of Medicine, P.O. Box016960 (R-138), Miami, FL 33101. E-mail address: dlopez@med.miami.edu3 Abbreviations used in this paper: MMP, matrix metalloproteinase; ECM, extracel-lular matrix; TIMP, tissue inhibitor of metalloproteinases; DMBA, 7,12-dimethyl-benzanthracene; Renca, renal carcinoma; VEGF, vascular endothelial growth factor;m, murine; Flt-1, fms-like tyrosine kinase-1; Flk-1, fetal liver kinase-1; PS, phos-phatidyl serine.

The Journal of Immunology

Copyright © 2003 by The American Association of Immunologists, Inc. 0022-1767/03/$02.00

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Due to their role in immunosurveillance, lymphocytes normallyhave an invasive phenotype, and findings support a role forMMP-9 in normal leukocyte migration from the vascular compart-ment to sites of inflammation (24). It has also been shown that Tlymphocytes constitutively produce small amounts of MMP-9 thatare up-regulated by phorbol esters and IL-2 stimulation (25) andthat increased production of MMP-9 in human PBMCs enhancesinvasion through reconstituted basement membrane (26). Becauselymphocyte invasion of a tumor mass is of crucial importance inthe host’s defense against the developing neoplasm, we evaluatedMMP production by the T cells from normal and mammary tumor-bearing mice. This study shows that although mammary tumorcells produce MMP-9, the T cells from tumor-bearing animals con-stitutively produce higher levels of this enzyme. The possiblecauses of enhanced production of MMP-9 in the T cell compart-ment of tumor bearers were also investigated. The findings pre-sented in this study may have important implications in elucidatingthe effects of tumor-derived and/or -induced factors on T cellsduring tumor development.

Materials and MethodsMice and cell lines

BALB/c mice used in these studies were 8 –12 wk of age and were bredin our animal facility at the University of Miami. Animal care and usewas according to the guidelines of the National Institutes of Health. TheD1–7,12-dimethylbenzanthracene-3 (D1-DMBA-3) tumor, syngeneic toBALB/c mice, is a transplantable mammary adenocarcinoma derivedfrom a nonviral, noncarcinogen-induced preneoplastic nodule aftertreatment with DMBA (27). The D1-DMBA-3 tumor is immunogenic tothe host of origin and nonmetastatic to the spleen or other organs, butminute metastases to the lung do occasionally occur. The DA-3 mam-mary tumor cell line was derived in our laboratory from the D1-DMBA-3 tumor and maintained in DMEM/high glucose, 10% charac-terized heat-inactivate FCS (HyClone Laboratories, Logan, UT), 100U/ml penicillin, and 100 �g/ml streptomycin with OPI medium sup-plement (Sigma-Aldrich, St. Louis, MO). The renal carcinoma (Renca)cells were maintained in RPMI 1640 supplemented with 10% heat-inactivated FCS, 100 U/ml penicillin, 100 �g/ml streptomycin, 1% L-glutamine, 1% sodium pyruvate, 1% nonessential amino acids, and 50mM 2-ME. The K-7 osteosarcoma cells were grown in DMEM/highglucose with 10% heat-inactivated FCS, 100 U/ml penicillin, and 100�g/ml streptomycin. Both the Renca and K-7 cell lines were a gift fromEduardo Sotomayor (Moffit Cancer Center, Tampa, FL). Tumors wereimplanted in BALB/c mice by s.c. injection of 1 � 106 tumor cells,resulting in a measurable tumor 7–10 days postimplantation.

Purification of splenic T cells

Spleens were compressed in Teflon tissue homogenizers, and the resultingsingle cell suspension was pelletted at 300 � g, subjected to hypotonicshock for red cell removal, washed, and counted. Macrophages were re-moved from the cell suspension by plastic adherence in prewarmed RPMI1640, 5% FCS at 37°C for 1 h in CO2 incubator. The nonadherent Tlymphocytes were purified on nylon wool columns according to the methodof Julius et al. (28) and by positive selection using the MACS magneticseparation system (Miltenyi Biotec, Auburn, CA), according to the man-ufacturer’s instructions. Briefly, single-cell suspensions in cold PEB buffer(PBS supplemented with 2 mM EDTA and 0.5% BSA) were incubatedwith supermagnetic microbeads conjugated to anti-mouse CD90 (Thy1.2),anti-mouse CD4, or anti-mouse CD8 mAb at 4°C for 15 min. Cells werewashed twice and loaded onto the magnetic separation columns. The col-umns were washed three times with cold PEB buffer, and the positivelyselected Thy1.2�, CD4�, or CD8� T cells were then eluted. After purifi-cation, the cells were routinely �95% viable, as assessed by trypan blueexclusion. FACS analysis using an LSR analyzer and anti-mouse FITC-CD90, anti-mouse FITC-CD4, and anti-mouse PE-CD8 Abs (BD PharM-ingen, San Diego, CA) confirmed the populations to be �93% Thy1.2�,�94% CD4�, and �90% CD8� T lymphocytes.

Cell culture

Purified splenic T cells used in zymography assays were first washed threetimes with RPMI 1640 to remove all residual serum before overnight cul-ture (2 � 106 cells/ml) in DMEM/F12 medium with 100 U/ml penicillin,

100 �g/ml streptomycin, 1 mM L-glutamine, 1 mM sodium pyruvate, and1� nonessential amino acids (all from Life Technologies, Grand Island,NY). At the end of the incubation period, the supernatants were removedand stored at �80°C. After purification, splenic T cells cultured forELISAs were resuspended in complete medium consisting of RPMI 1640,10% FCS, 100 U/ml penicillin, 100 �g/ml streptomycin, and 50 mM 2-MEand were cultured overnight. Supernatants were removed and stored at�80°C until use. The murine DA-3 and D1-DMBA-3 adenocarcinomacells were cultured in complete medium as described above. After 1, 2, 3,and 4 days in culture, supernatants were removed and stored at �80°C untilanalysis in the vascular endothelial growth factor (VEGF) ELISA.

The following reagents were used in experiments as noted: Con A,PMA, and PGE2 (Sigma-Aldrich); recombinant murine TNF-� (rmTNF-�), GM-CSF, and IL-6, (PeproTech, Rocky Hill, NJ); murine VEGF andanti-mouse VEGF Ab (R&D Systems, Minneapolis, MN); and phospha-dityl serine (Avanti Polar Lipids, Alabaster, AL).

Zymography

Gelatin zymography followed a modified procedure of Heussen and Dow-dle (29) for detecting picogram amounts of MMP-2 and MMP-9. Identicalamounts of supernatant were electrophoresed under nonreducing condi-tions using 10% SDS polyacrylamide gels containing 0.33 mg/ml gelatin.After electrophoresis, the gels were washed twice in 2.5% Triton X-100 for15 min to remove SDS. After overnight incubation at 37°C in assay/incu-bation buffer (50 mM Tris-HCl (pH 7.5), 200 mM NaCl, 1�M ZnCl2,0.02% NaN3, and 0.005% Brij 35), the gels were stained for 2 h withCoomassie blue R 250 and destained with 7% acetic acid. Both latent andactive forms of gelatinases produce clear areas in the gel.

Western blot analysis

Equal amounts of T cell supernatants were separated on 10% SDS poly-acrylamide gels under reducing conditions and then were transferred ontoProtran nitrocellulose membranes (0.45-�m pore size; Schleicher &Schuell, Keene, NH) using a Trans-Blot electrophoretic cell (Bio-Rad, Her-cules, CA). Membranes were blocked for 1 h at room temperature in 5%nonfat dry milk in 1� TBS-0.1% Tween 20 followed by 1-h incubationwith an anti-MMP-9 Ab (Chemicon International, Temecula, CA) at roomtemperature. Blots were washed for 30 min with five changes of 1� TBS-0.1% Tween 20 solution followed by 1-h incubation at room temperaturewith the HRP-conjugated anti-rabbit IgG Ab (Chemicon International).Blots were washed again for 30 min and incubated for 3 min with Super-signal West Pico chemiluminescent substrate (Pierce, Rockford, IL). Theresults were visualized by exposing blots to BioMax autoradiographic film(Kodak, Rochester, NY).

Reverse zymography

TIMPs secreted into the culture medium were detected using a modifica-tion of the method of Herron et al. (30). Briefly, conditioned medium fromcultured DA-3 or splenic T cells was electrophoresed under nonreducingconditions using 12% SDS polyacrylamide gels containing 0.75 mg/mlgelatin and 4-aminophenylmercuric acetate-2-activated MMP-2 (1–2 U/ml,purified from human skin fibroblasts). The gel was rinsed twice in 50 mMTris-HCl, 2.5% Triton X-100 solution for 20 min and incubated overnightwith zymography assay/incubation buffer at 37°C. Undigested gelatin inthe gel was visualized by staining with 0.25% Coomassie blue R 250 anddestaining in 7% acetic acid. Areas of the gel containing TIMPs appearedas dark blue bands against a paler blue background.

RNA analyses

Three types of RNA analyses were performed on 2- to 4-h splenic T cellcultures in this study: RT-PCR, Northern blotting, and RNA gene arrays.For the RT-PCR analysis, total RNA was isolated using TriReagent (Mo-lecular Research Center, Cincinnati, OH). One microgram of total RNAfrom each sample was reverse-transcribed using a Primus thermocycler(MWG Biotech, High Point, NC) and the GeneAmp RNA PCR kit (Ap-plied Biosystems, Foster City, CA) with oligo d(T)16 primers, according tothe manufacturer’s instructions. The reaction was incubated for 60 min at42°C, followed by inactivation of the murine leukemia virus reverse tran-scriptase at 99°C for 5 min. The cDNA products obtained were subjectedto PCR for 35 cycles, after which 15 �l of the amplified DNA fragmentswere electrophoresed on a 1.6% agarose gel stained with ethidium bromideand visualized by UV transillumination. The gels were analyzed with ScionImage � 4.02 densitometry software (Scion, Frederick, MD) and normal-ized with �-actin band intensity. Percentages were calculated relative to themaximum for each experiment. The PCR for MMP-9 amplification wasconducted by incubating the samples at 94°C for 10 min, followed by 35

4341The Journal of Immunology

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min, with a finalextension for 10 min at 72°C. For fms-like tyrosine kinase-1 (Flt-1) andfetal liver kinase-1 (Flk-1) amplification, the samples were incubated at94°C for 10 min, followed by 35 cycles of 90°C for 30 s, 65°C for 45 s, and72°C for 75 s, with a final extension for 10 min at 72°C. The primers wereas follows: murine MMP-9 sense, 5�-GCGACCACATCGAACTTCGACACT-3�; antisense, 5�-TCAGGAACTTCCAGTACCAACCGT-3� (31);murine Flt-1 sense, 5�-CCTGATTTCCTACAGTTTCCA-3�; antisense, 5�-TCCGGGGTTCTCATCCGCATG-3� (32); murine Flk-1 sense, 5�-CTTAGGTGCCTCCCCATACCCTGGG-3�; antisense, 5�-TGGCCGGCTCTTTCGCTTACTGTTC-3� (33); murine VEGF sense, 5�-CGAGACCCTGGTGGACATCT-3�; antisense, 5�-CACCGCCTCGGCTTGTCAC-3�(34); murine �-actin sense, 5�-TCTGGCACCACACCTTCTAC-3�; and anti-sense, 5�-GAAGGAAGGCTGGAAGAGTG-3� (35).

Standard Northern blot technique was followed using the Northern Maxformaldehyde-based system (Ambion, Austin, TX). Briefly, total RNA (10�g) was electrophoresed through agarose-formaldehyde gels, blotted ontonylon membranes by capillary electrophoresis, followed by prehybridiza-tion of the membrane, and hybridized with the appropriate radiolabeledprobe. Blots were preincubated in the ULTRAhyb hybridization solutionfor 3 h at 42°C and then hybridized at 42°C overnight in ULTRAhyb withan [�-32P]dATP-radiolabeled DNA probe. Probes were prepared by ran-dom priming with either �-actin cDNA or a PCR fragment of MMP-9(Decaprime II kit; Ambion). Blots were exposed to BioMax autoradiogra-phy film overnight.

The relative mRNA expression of mouse ECM and adhesion moleculeswas analyzed with the chemiluminescent GEArray Q Series (SuperArray,Bethesda, MD) according to the manufacturer’s protocol. Briefly, totalRNA was isolated and 5 �g from each sample was reverse transcribed intocDNA with Moloney murine leukemia virus reverse transcriptase (Pro-mega, Madison, WI) in the presence of Biotin-16-dUTP (Roche, Indianap-olis, IN). The resulting cDNA probes were hybridized to cDNA fragmentsthat were spotted on the GEArray nylon membranes. The relative expres-sion level of each gene was analyzed using GEArray Analyzer and ScanA-lyze software.

Cytokine ELISA

The amounts of VEGF present in the supernatants of tumor cells and Tlymphocytes were measured by Quantikine M murine VEGF ELISA (R&DSystems) according to the manufacturer’s instructions. Absorbance at 450nm was read on a Tecan SLT Rainbow Reader (Labinstruments, ResearchTriangle Park, NC), and OD values of samples were converted to pico-grams against a standard curve of known quantities of rmVEGF.

Immunohistochemistry

Tissue from 3-wk D1-DMBA-3 tumors was formalin-fixed, embeded inparaffin, cut, and processed according to standard procedures. Serial sec-tions were stained with goat IgG anti-CD3� (Santa Cruz Biotechnology,Santa Cruz, CA) and detected with a biotinylated HRP complex as a de-tection reagent (Vector Laboratories, Burlingame, CA) per the manufac-turer’s directions. Omission of the primary Ab was used as a negativecontrol.

Isolation of Thy1.2� tumor-infiltrating lymphocytes

Thy1.2� tumor-infiltrating lymphocytes were isolated based on the methodof Rosenberg and coworkers (36). Briefly, 3–4 wk after implantation, tu-mors were aseptically removed, minced into 1- to 2-mm3 pieces, and en-zymatically digested in 1% collagenase and 325 U/ml DNaseI (both fromRoche) at 37°C for 2 h. The resulting single cell suspension was washedtwice and resuspended in RPMI 1640 medium, 10% FCS, 100 U/ml pen-icillin, 100 �g/ml streptomycin, and 3000 IU/ml rmIL-2 (PeproTech). Af-ter overnight incubation at 37°C and 5% CO2 in plastic culture dishes, thenonadherent population was removed and counted. Nonadherent cells werestained with anti-Thy1.2 (BD PharMingen) and sorted on a FACSVantageSE (BD Biosciences, Mountain View, CA). Population purity was analyzedon a BD PharMingen LSR Analyzer. The positive population was washedthree times to remove any serum and then was cultured overnight.

ResultsEvaluation of gelatinase production in splenic T cells fromnormal and tumor-bearing mice

The gelatinase activity produced by splenic T cells from normalmice and from animals bearing the D1-DMBA-3 mammary tumorwas evaluated by zymography using gelatin as a substrate (Fig.1A). Splenic T cells from normal and mice bearing tumors for 1, 2,

and 3 wk were cultured in serum-free medium for 18 h, and thesupernatants were collected to assay by zymography for enzymaticactivity. No detectable MMP-2 (72 kDa) was observed in the T cellcultures, but MMP-9 activity increased in parallel with tumor bur-den from the basal levels seen in normal T cells. Importantly, theseenhanced MMP-9 levels were not only of the latent pro-enzymeform observed as a single band at 105 kDa, but they were also ofthe active form, as indicated by the bands of activation at 95 kDa.Proteolytic cleavage of the proenzyme domain of MMP-9 resultsin the active enzyme that is �10 kDa less in molecular mass. It issuggested that the bands at 125 and 215 kDa are due to the bindingof the associated 25-kDa proteins, microglobulin or neutrophil gela-tinase-associated lipocalin, to monomers and dimers of MMP-9 (37).

FIGURE 1. MMP-9 production by T cells from D1-DMBA-3 tumorbearers increases in parallel to tumor burden. Splenic T cells from normaland 1-, 2-, and 3-wk tumor bearers were purified and washed three timesto remove serum-derived gelatinase activity as described in Materials andMethods, and then they were cultured overnight in serum-free medium. A,Supernatants were collected and equal amounts were assayed by gelatinzymography for enzyme activity. B, Duplicate supernatants were subjectedto 10% SDS-PAGE and immunoblotting analysis for MMP-9 protein ex-pression using a rabbit anti-rat MMP-9 polyclonal Ab. Data are represen-tative of at least four independent experiments with two mice per condition.C, Splenic T cells from normal and tumor bearers were separated usingMiltenyi Biotec magnetic beads and Abs. The purified populations werewashed three times and cultured overnight. Equal amounts of supernatantwere assayed by zymography. Data are representative of three independentexperiments.

4342 UP-REGULATION OF MMP-9 IN TUMOR BEARERS’ T LYMPHOCYTES

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Because the level of enzymatic activity of MMP-9 is determinedonly in part by the amount of MMP-9 secreted, we next assessedthe protein levels of this metalloproteinase. Duplicate aliquots ofthe cell culture supernatants tested for enzymatic activity weresubjected to SDS-PAGE and immunoblotting analysis for proteinexpression (Fig. 1B). The results indicated that the increasedMMP-9 enzymatic activity paralleling tumor burden correlateswith increased MMP-9 protein synthesis. To determine whetherCD4� and CD8� T lymphocytes vary in their secretion ofMMP-9, these populations were positively selected using MiltenyiBiotec magnetic beads conjugated to anti-CD4 and anti-CD8 Abs.Thy1.2�, CD4�, and CD8� splenic T cells from normal and tu-mor-bearing mice were cultured overnight, and the supernatantswere assayed by zymography. CD4� and CD8� splenic T cellsdemonstrated similar levels of MMP-9 activity (Fig. 1C).

This metalloproteinase up-regulation is not solely a phenome-non of the D1-DMBA-3 tumor. The zymogram in Fig. 2A showsthat T cells from animals bearing both D1-DMBA-3 (lanes 4, 5,and 6) or DA-3 tumors (lanes 7, 8, and 9), an in vitro-derived cellline from the D1-DMBA-3 in vivo tumor, produced enhanced lev-els of MMP-9 compared with T cells from normal mice (lanes 10,11, and 12). It was also determined that the protein kinase C ac-tivator PMA (25 ng/ml) failed to elevate MMP-9 in the BALB/c Tlymphocytes, in contrast with its reported induction capabilities for

this metalloproteinase in various T cell lines and human peripheralblood cells (25, 38). Likewise, Con A did not increase the levels ofMMP-9 in normal or tumor-bearing mouse lymphocytes. MouseMMP-9 is 105 kDa because of an additional 24-aa insert (39, 40)and is located higher in the gel than is the human MMP-9 (92 kDa)positive control. To ascertain whether this up-regulation was onlyseen in a mammary adenocarcinoma animal model, T cells fromBALB/c mice bearing either the K-7 osteosarcoma or the Rencatumors were also tested for MMP-9 activity (Fig. 2B). Splenic Tcells isolated from these animals also demonstrated increasedMMP-9 activity with tumor growth.

MMP-9 mRNA expression parallels enzyme production

To determine whether the increased levels of MMP-9 secretion inthe T cells from tumor bearers were due to altered transcription,MMP-9 gene expression in normal and tumor-bearing mice wasanalyzed by RT-PCR and Northern blots. A progressive increase inthe levels of MMP-9 mRNA was seen in 4-h cultures of unstimu-lated splenic T cells from mice bearing 1-, 2-, or 3-wk tumors, asdetected by RT-PCR (Fig. 3A). Additionally, Northern blot anal-ysis using a murine-specific cDNA probe confirmed that steadystate levels of MMP-9 mRNA increase with tumor burden (Fig.3B). Splenic T cells from mice bearing 3-wk tumors expressedhigher levels of each of the two different murine MMP-9 mRNAs(2.5 and 3.2 kb), in comparison with T cells from normal mice.

FIGURE 2. Splenic T cells from various solid tumors secrete elevatedlevels of MMP-9. A, MMP-9 activity is increased in T cells from bothD1-DMBA-3 and DA-3 tumor-bearing mice. Equal amounts of supernatantfrom splenic T cell cultures from animals bearing either D1-DMBA-3(lanes 4, 5, and 6) or DA-3 tumors for 4 wk (lanes 7, 8, and 9) and fromnormal mice (lanes 10, 11, and 12) were concentrated 10� and assayed bygelatin zymography. Some T cell cultures were incubated with Con A (10�g/ml) or PMA (25 ng/ml). Lane 2 contains a molecular mass marker, andlane 3 contains the 92-kDa human MMP-9 standard. Data are representa-tive of three independent experiments using three mice per group. B,MMP-9 activity is increased in T cells from mice bearing Renca and K-7tumors for 4 wk. Equal amounts of supernatants from T cell cultures wereassayed by gelatin zymography for enzyme activity. Data are representa-tive of three independent experiments using two mice per group.

FIGURE 3. Increased MMP-9 mRNA expression in splenic T cellsfrom tumor-bearing mice. A, Splenic T cells from normal and tumor-bear-ing mice were cultured without stimulation for 4 h. Semiquantitative RT-PCR was performed on total RNA (1 �g) isolated from the cells. A 100-bpDNA ladder was used as molecular mass marker. �-Actin band intensitywas used for normalization, and integration intensity percentages were cal-culated relative to the maximum for each experiment. The data are repre-sentative of four independent experiments. B, Total RNA (10 �g) fromunstimulated splenic T cells from normal and 3-wk tumor-bearing micewas subjected to Northern blot analysis by hybridization with a specificcDNA probe for MMP-9. The two different murine MMP-9 mRNAs (2.5and 3.2 kb) were detected. Equivalence of RNA loading was confirmed byrehybridization to a �-actin probe. Data are representative of three inde-pendent experiments.

4343The Journal of Immunology

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Equivalence of RNA loading in different lanes was ascertained byrehybridization to a �-actin probe. Results indicated that splenic Tcells from tumor-bearing animals express high levels of MMP-9mRNA, compared with normal animals. These results corroboratethat the transcriptional activity of MMP-9 correlates with its se-creted enzyme activity.

TIMP secretion by mammary tumor cells and T lymphocytes

The enzymatic activity of MMPs is dependent upon equilibriumamong the production of the proenzyme, activation of the latentenzyme, and production of the naturally occurring inhibitors, es-pecially TIMPs (41). Currently, the TIMP family consists of fourmembers: TIMP-1, -2, -3, and -4. The gelatinase activities of bothhuman and mouse MMP-2 and -9 are regulated by TIMP-2 andTIMP-1, respectively, which inhibit protease activity and canblock cell invasion in vitro and metastasis in vivo (42, 43). Theseinhibitors are often produced by the same cells that produce thegelatinases and are usually found complexed noncovalently withthe enzymes at the carboxy-terminal domain (44). Therefore,TIMP activity was assayed by reverse zymography using the samecell cultures tested for MMPs. As seen in Fig. 4A, DA-3 tumorcells secreted very low levels of TIMP-2 (21 kDa), and their levelsof secreted TIMP-1 (29 kDa), the natural inhibitor of MMP-9,were barely detectable. No TIMP-3 (24 kDa) or TIMP-4 (23 kDa)

was detected. Although these inhibitors are often produced by thesame cells that secrete MMPs, no TIMPs could be detected in thesupernatants of cultured splenic T cells from either normal or tu-mor-bearing mice (Fig. 4B). Furthermore, no TIMP-1 mRNAcould be detected in cultured splenic T cells, and minimal levels ofTIMP-2 mRNA expression were seen only in normal T cells usinggene expression array analysis (Fig. 4C). Results were confirmedby RT-PCR (data not shown). The high levels of MMP-9 mRNAseen only in RNA isolated from tumor bearers’ splenic T cells alsoconfirmed the results of Fig. 3.

Assessment of the role of tumor-derived factors and tumor-induced cytokines in the elevation of MMP-9 in T cells fromtumor bearers

The mammary tumors used in our studies, D1-DMBA-3 and itscloned cell line DA-3, have been shown in previous work to con-stitutively produce several factors, such as PGE2 (45), GM-CSF(46), phosphatidyl serine (PS) (47), and noncharacterized angio-genic molecules (45). To investigate whether one or more of thesetumor-derived factors was causing the enhanced levels of activeMMP-9 observed in T cells from D1-DMBA-3 mammary tumorbearers, we treated normal lymphocytes with PGE2, PS, and GM-CSF at concentrations known to be in circulation in tumor-bearingmice (Fig. 5A). The cell supernatants of normal T cells cultured

FIGURE 4. Analysis of TIMP secretion. A, Aliquots of cell-free supernatants from DA-3 cells cultured serum-free overnight were assayed by reversezymography to detect TIMP secretion. Data are representative of three independent experiments. B, Aliquots of cell-free supernantants from splenic T cellsfrom normal (lanes 4–9) and tumor-bearing animals (lanes 11–13) were assayed by reverse zymography to detect TIMP secretion. Lane 1, Purifiedrecombinant TIMP-1 and TIMP-2; lane 2, purified recombinant TIMP-3; lane 3, molecular mass marker. Data are representative of three independentexperiments. C, Representative gene expression array analysis of splenic T cell mRNA from normal and tumor-bearing animals. TIMP-1, TIMP-2, andMMP-9 mRNA expression are shown in the black boxes, whereas the �-actin mRNA expression is shown in the last two tetra-spots.

4344 UP-REGULATION OF MMP-9 IN TUMOR BEARERS’ T LYMPHOCYTES

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

overnight with 50 U/ml GM-CSF, 30 and 60 �g/ml phosphatidylserine, or 10�6 M PGE2 were analyzed by zymography. Titrationsof these tumor-derived factors, either alone or in combination (datanot shown), failed to increase the levels of proenzyme or to acti-vate MMP-9 from normal T lymphocytes. Supernatants from un-treated tumor bearers’ T cells were included in the gel as a com-parison of MMP-9 activity.

After eliminating these characterized tumor-derived factors asmediators of MMP-9 up-regulation, we looked at known tumor-induced factors in this model system. The proinflammatory cyto-kines TNF-� and IL-6 are up-regulated in the B cells (48, 49) andmacrophages (50, 51) of tumor bearers, leading to elevated levelsof these cytokines in their sera compared with normal BALB/c

mice. Therefore, these two molecules were possible candidates forsoluble factor(s) regulating MMP-9 expression. Splenic T cellsfrom normal mice were cultured overnight with 60 U/ml IL-6 or 50U/ml TNF-�, and MMP-9 enzyme activity in the cell supernatantswas assessed by zymography. Again, treatment of normal T cellswith the proinflammatory cytokines had no effect on MMP-9 se-cretion (Fig. 5B). Identical MMP-9 secretion was seen in controland cytokine-treated normal T cells with titrations of both of therecombinant cytokines, either alone or in combination (data notshown). Supernatants from untreated tumor bearers’ T cells wereincluded in the gel as a comparison of MMP-9 activity. It shouldbe noted that the efficacy of the recombinant cytokines and otherfactors was ascertained by their ability to modulate cytokine/che-mokine production in parallel cultures of T cells and macrophages(data not shown).

VEGF production in the mammary tumor model and its effectson MMP-9

Because none of the previously described tumor-derived factors orup-regulated cytokines in our mammary tumor model was capableof modulating MMP-9 in the T cells, we looked for other tumor-derived and/or -induced factors that may contribute to elevatedMMP-9 production. Many solid tumors express VEGF, and thisexpression directly correlates with regions of angiogenesis andhigh vascular density (52). Furthermore, the newly formed bloodvessels are inherently leaky, enhancing the likelihood of cell mi-gration and metastasis (53). Based on the knowledge that mostsolid tumors express VEGF and that the supernatants of DA-3 cellshave angiogenic activity, we measured DA-3 tumor cell produc-tion of VEGF by ELISA. The results depicted in Fig. 6A confirmthat cultured DA-3 cells express physiologically relevant concen-trations of VEGF, because this growth factor is known to affectbiologic activity at concentrations below 1 pM (54). D1-DMBA-3cells cultured under identical conditions secreted 2364 � 37 pg/mlVEGF after 4 days in culture.

Of the few published reports of VEGF production by T cells,most cite studies of human PBLs (55, 56). We sought to investi-gate whether murine T lymphocytes produced detectable levels ofVEGF and whether such a production is affected by tumor pro-gression. Splenic T cells from normal and 4-wk tumor-bearingmice were cultured overnight in the absence of any stimulus andwere tested by ELISA for the production of VEGF (Fig. 6B). Lowlevels of VEGF were constitutively expressed by normal BALB/cT lymphocytes, with higher concentrations of this growth factordetected in the culture supernatants of T lymphocytes from 4-wktumor-bearing mice. Stimulation with PHA (2 �g/ml) or IL-2 (50U/ml) did not increase the levels of VEGF in T cells from normalor tumor-bearing mice. Interestingly, PMA (10 ng/ml) enhancedthe VEGF production by T cells from tumor bearers but not fromnormal mice.

The physiological levels of VEGF in our system were deter-mined by ELISA testing of sera from normal mice and animalsbearing 2- and 4-wk D1-DMBA-3 tumors. The mean serum levelof VEGF nearly doubles after 4 wk of tumor burden (Fig. 6C).Because macrophages are also known producers of VEGF, peri-toneal elicited macrophages were analyzed for their production ofVEGF, but no significant differences were seen between the levelsproduced by cells from normal and tumor-bearing mice (data notshown).

To determine whether VEGF could act directly on the T lym-phocytes, it was important to demonstrate the expression of VEGFreceptors by splenic T cells. There are no reports in the literaturedemonstrating the expression of such receptors in primary murineT cells. The two tyrosine kinases of the VEGF receptor family,

FIGURE 5. No effect of various tumor-derived and tumor-induced fac-tors on the production of metalloproteinases by T cells. A, Splenic T cellswere cultured overnight alone or in the presence of factors produced by thetumor, rmGM-CSF (lanes 3 and 7), PS (lanes 4, 5, 7, and 8), and PGE2

(lanes 6 and 8). After 18 h, supernatants were concentrated 10� and as-sayed by gelatin zymography. Lane 1 contains the molecular mass markerand lane 14 contains human MMP-9 used as a standard. Each experimentalgroup consisted of at least four mice. Similar results were obtained in threeindependent experiments. B, Normal T lymphocytes were cultured alone orwere treated with rmIL-6 (lanes 4 and 9) or rmTNF-� (lanes 5 and 10).After 18 h, supernatants from the cultures were concentrated 10� andassayed by gelatin zymography. Lane 1 contains the molecular massmarker, and lane 14 contains the human MMP-9 standard. Supernatantsfrom untreated tumor bearers’ T cells were included in the gel as a com-parison of MMP-9 activity. Each experimental group consisted of at leastfour mice. Similar results were obtained in three independent experiments.

4345The Journal of Immunology

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Flt-1 (VEGFR-1) and Flk-1 (kinase domain region/VEGFR-2),were the likely candidates for the putative receptors of this growthfactor because they are expressed on endothelium, skeletal muscle,hematopoietic stem cells, megakaryocytes, macrophages, and hu-man CD3� cells (53). The expression of both receptors has alsobeen detected in some leukemia cell lines (57). RT-PCR was per-formed on total RNA isolated from 2-h cultures of T cells fromnormal and tumor-bearing animals. The expression of Flt-1 andFlk-1 mRNA was seen in both normal and tumor bearers’ splenicT cells. Actin mRNA levels served as controls to normalize RNAquantity. Interestingly, both Flt-1 and Flk-1 mRNA levels weregreater in the T cells from mice bearing mammary tumors than inthe normal BALB/c animals (Fig. 7A).

In other cells expressing VEGF receptors, VEGF has beenshown to up-regulate its production in an autocrine manner (57).To determine whether VEGF up-regulates it own production inmurine splenic T cells, T lymphocytes from normal and tumor-bearing mice were cultured with rVEGF, and total RNA was iso-lated and analyzed by RT-PCR (Fig. 7B). In contrast with othercell types, VEGF failed to up-regulate its expression in splenic Tcells from either normal or tumor-bearing mice. Two different mR-NAs were detected in both normal and tumor-bearing animals,corresponding with VEGF165 and VEGF120. VEGF165 is the pre-dominant isoform and has the optimal characteristics of bioavail-ability and biological potency (57). No difference in intensity ofbands corresponding with the two isoforms VEGF165 andVEGF120 was detected in those cells cultured with rVEGF, com-pared with untreated cells. There was increased expression ofVEGF165 in tumor bearers’ T cells compared with T cells fromnormal animals, corroborating data at the protein level (Fig. 6B).

The up-regulation of the expression of MMP-9 in the tumorbearers’ T cells appears to be induced by the developing cancer,possibly via one or more of the tumor-derived factors. Therefore,we investigated whether VEGF plays a role in the elevated secre-tion of MMP-9 by the T cells of tumor bearers. Normal splenic Tcells were cultured overnight with a titration of rmVEGF. Fig. 8Ashows an up-regulation of MMP-9 in T cells treated with rVEGF.Because VEGF is known to be a potent mitogen of endothelialcells, 5�-bromo-2�-deoxyuridine incorporation assays were per-formed to confirm that this increased production of MMP-9 wasnot due to proliferation of the T cells cultured with VEGF (data notshown). We also see increased levels of MMP-9 mRNA in T cellscultured with rVEGF compared with normal T cells cultured with-out treatment (Fig. 8B). Furthermore, an anti-VEGF neutralizingAb reduced the up-regulated levels of MMP-9 in normal T cellscultured with rVEGF (Fig. 8C). These results strongly suggest thatalthough there may be other yet to be identified modulatory fac-tors, VEGF certainly plays a role in the up-regulation of MMP-9in T lymphocytes of mammary tumor bearers.

Production of MMP-9 in the tumor

The mammary tumor cells used in our studies were assessed fortheir production of MMP-9. In Fig. 9A, it can be seen that theDA-3 tumor cells produce measurable levels of MMP-9, as de-tected by zymography. Inflammatory cells have been shown to

supernatants. Representative data of three independent experiments arepresented as mean � SD of samples in duplicate. C, Serum levels of VEGFin normal and tumor-bearing animals. Blood from normal and tumor-bear-ing animals was collected and allowed to clot for 2 h at room temperature.After centrifugation, serum was removed and stored at �70°C until assayby ELISA. Normal animals, n � 20; 2-wk tumor bearers, n � 18; 4-wktumor bearers, n � 12.

FIGURE 6. VEGF production in the tumor model. A, VEGF productionby the mammary tumor cells. Supernatants of 1-, 2-, 3-, and 4-day DA-3cell cultures were collected to test for VEFG production by specific ELISA.The number of viable cells at each day was determined in parallel by trypanblue exclusion. Representative data of three independent experiments arepresented as the mean of samples in duplicate. B, VEGF production by Tcells is increased with tumor burden. Splenic T cells from normal andtumor-bearing mice were cultured in vitro. The supernatants from T cellscultured in growth medium alone (Untreated) or stimulated with PHA (2�g/ml), IL-2 (50 U/ml), or PMA (10 ng/ml) were collected after 18 h. TheVEGF levels were determined by specific ELISA using the undiluted

4346 UP-REGULATION OF MMP-9 IN TUMOR BEARERS’ T LYMPHOCYTES

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

secrete several MMPs, including MMP-9, -12, and -14, potentiallypromoting cancer progression by the release of these proteases(21). To assess the contributions of T cells to the levels of pro-

teases secreted within the tumor microenvironment, immunohis-tochemistry was first performed using an anti-CD3� Ab to evaluatethe presence of T lymphocytes in sections of tumors obtained 3–4

FIGURE 7. Expression of VEGF receptors by splenic T cells. A,Splenic T cells from normal mice (N) and mice bearing tumors (T) werecultured without stimulation for 2 h. Total RNA was isolated for analysisof Flt-1 and Flk-1 expression by RT-PCR. Amplified products were sub-jected to electrophoresis on 1.6% agarose gels and visualized by ethidiumbromide staining. A 100-bp DNA ladder was used as molecular massmarker. �-Actin band intensity was used for normalization, and integrationintensity percentages were calculated relative to the maximum for eachexperiment. Gels shown are representative of four experiments. B, Purifiedsplenic T cells from normal (N) and tumor-bearing animals (T) were cul-tured untreated or with 10 ng/ml rmVEGF for 4 h. Total RNA was isolatedfor analysis of VEGF expression by semiquantitative RT-PCR. Gel shownis representative of three experiments.

FIGURE 8. VEGF up-regulates the production of MMP-9 in splenic Tcells. A, Normal splenic T cells were either untreated (lane 3) or treatedwith 0.05 ng/ml VEGF (lane 4), 0.1 ng/ml VEGF (lane 5), 1 ng/ml VEGF(lane 6), 5 ng/ml VEGF (lane 7), 10 ng/ml VEGF (lane 8), 20 ng/ml VEGF(lane 9), 50 ng/ml VEGF (lane 10), or 100 ng/ml VEGF (lane 11). Lane 1contains a molecular mass marker. Equal amounts of supernatant wereassayed by gelatin zymography. Data are representative of at least fourindependent experiments. B, Normal splenic T cells were cultured un-treated or with 1 ng/ml rmVEGF for 4 h. Total RNA was isolated foranalysis of MMP-9 expression by semiquantitative RT-PCR. A 100-bpDNA ladder was used as molecular mass marker. Amplified products weresubjected to electrophoresis on 1.6% agarose gels and visualized byethidium bromide staining. �-Actin band intensity was used for normal-ization, and integration intensity percentages were calculated relative to themaximum for each experiment. The gel shown is representative of fourexperiments. C, Anti-VEGF neutralizing Ab reduces the up-regulation ofMMP-9 by recombinant VEGF. Recombinant murine VEGF (1 ng/ml) wasincubated with 0.1 �g/ml of the Ab (R&D Systems) for 1 h. After thispreincubation period, normal splenic T cells were added and incubatedovernight. Equal amounts of supernatant were assayed by gelatin zymog-raphy. Data are representative of three independent experiments.

4347The Journal of Immunology

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

wk postimplantation. As seen in Fig. 9B, tumor-infiltrating T lym-phocytes could be detected within the tumor bed. In further stud-ies, T cells were isolated from the tumors after enzymatic diges-tion, as described in Materials and Methods. After overnightculture, the nonadherent cells were sorted based on their expres-sion of Thy1.2. As seen in Fig. 9C (left panel), the preparationswere greatly enriched for Thy1.2 after sorting, from 10.8% to97.8%. The purified tumor-infiltrating T cells were then analyzedfor their production of MMP-9 by zymography. The Thy1.2� cellsproduced very high levels of this proenzyme and its active form(Fig. 9C, right panel). As was the case with the splenic T lym-phocytes from tumor bearers, when duplicate cultures of the tu-

mor-infiltrating Thy1.2� cells were assayed for VEGF production,a high constitutive level of this angiogenic factor was also ob-served (72 pg/ml).

DiscussionProgressive tumor growth is associated with multiple changes inthe immune compartment of the host. In this study, it was foundthat MMP-9 expression and VEGF production are up-regulated inlymphocytes of animals with developing neoplasia. The up-regu-lated expression of MMP-9 is found not only in splenic T cells but,more importantly, in the T cells infiltrating the tumor microenvi-ronment, and this increased expression correlates with tumor

FIGURE 9. Production of MMP-9 in the tumor. A, Aliquots of cell-free supernatants from DA-3 cells cultured serum-free overnight (2 � 106 cells/ml)were assayed by zymography to detect MMP-9 secretion. Data are representative of three independent experiments. B, Paraffin tumor sections were stainedwith anti-CD3�. Figures are representative of several tumors from at least three separate experiments. C, Left and middle panels, Flow cytometry profiles,before and after sorting, of Thy1.2� lymphocytes obtained from the tumors as described in Materials and Methods. C, Right panel, Thy1.2� tumor-infiltrating lymphocytes were cultured serum-free overnight (2 � 106 cells/ml), and the cell-free supernatants were assayed by gelatin zymography. Dataare representative of three independent experiments.

4348 UP-REGULATION OF MMP-9 IN TUMOR BEARERS’ T LYMPHOCYTES

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

growth. Our previous studies (58) and those of many other investi-gators (19, 59) have shown the infiltration of the tumor microenvi-ronment by lymphocytes through a process involving adhesion mol-ecules (60). The presence of MMP-9 in tumor bearers’ lymphocytesmay also play a role, because this metalloproteinase has the potentialto degrade matrix protein and to promote leukocyte traffic.

Although there are many reports about the regulation of metal-loproteinases in tumor cells, less attention has been focused ontheir expression in lymphocytes. The production of MMP-9 hasbeen studied by several investigators using cultured macrophages andmacrophage and T cell lines (61, 62), and most studies have usedstimuli to induce secretion of MMPs (37). Conversely, the higherlevels of MMP-9 observed in our mammary tumor bearers’ T cells isa constitutive event, in that it occurs in the absence of added stimuli.

Another aspect of this study investigated the production ofTIMPs by the mammary tumor and the splenic T cells. An impor-tant mechanism for the regulation of MMP proteolysis is via theactions of the TIMPs that bind to the active site of both latent andactive MMPs, forming stable, but inactive, enzyme-inhibitor com-plexes (63). TIMPs are considered the key inhibitors of MMPs intissue and, like MMPs, the expression of these inhibitors is con-trolled to maintain a balance in the integrity of the ECM (64). TheTIMPs differ in their inhibitory properties. Thus, TIMP-2 binds tothe proMMP-2, whereas TIMP-1 forms a complex with proMMP-9(44). Investigators have seen coregulation of MMP-9 and its specificinhibitor, TIMP-1, in a malignant form of non-Hodgkin’s lymphoma,and it has been suggested that this coregulation of gelatinase andinhibitor plays an important role in controlling the aggressiveness ofthis cancer (65). We observed very low levels of TIMP-1 secretion bythe mammary tumor cells and no expression of this molecule by thesplenic T cells of either normal or tumor-bearing mice. Therefore, theup-regulation of MMP-9 does not seem to correlate with increasedTIMP production in our system.

The up-regulation of the expression of MMP-9 in the tumorbearers’ T cells appears to be induced by the developing cancer,possibly via one or more of the tumor-derived or tumor-inducedfactors. Experiments in which normal T cells were cultured withPGE2, GM-CSF, and PS were performed to evaluate the potentialcontribution(s) of tumor-derived factors to the induction ofMMP-9. We have previously shown that these factors are pro-duced by the mammary tumors used in our studies and that theycan exert powerful regulatory actions on the cells of the immunesystem during tumor growth (46, 66, 67). Moreover, Leppert et al.(68) have seen that 10�6 M PGE2 increased the secretion ofMMP-2, -3, and -9 by the human lymphoblastoma cell line Tsup-1.In addition, TNF-� and IL-6, although not produced by the tumorcells, are up-regulated in the macrophages (50, 51) and the B cells(48, 49) of mammary tumor bearers. Therefore, the effects of thesecytokines were also evaluated in these experiments. The autocrineregulation of MMP-9 expression by TNF-� has been described inthe human promyelocytic cell line, HL-60 (69), and IL-6 has beenshown to up-regulate MMP-9 in malignant non-Hodgkin’s lym-phomas (70). Thus, it was of special interest to investigate theeffects of these cytokines on T cells. However, our results indi-cated that treatment with these cytokines and tumor products hadno effect on the levels of MMP-9 secreted by murine splenic Tcells. This is not surprising, because MMPs are stimulated by dif-ferent factors, and different cell types have varying responses to arange of stimuli in the induction of these proteases.

Previously we demonstrated that supernatants of the DA-3mammary tumor cells had angiogenic activity (45). VEGF is pro-duced by a variety of tumors and tissues, and in fact we have foundthat our tumors are high producers of this growth factor. Thus, themammary tumor cell-derived VEGF could be responsible for the

angiogenic activity detected in the supernatants of these cells.VEGF plays a central role in neovascularization, because it is anendothelial cell mitogen in vitro and a potent inducer of angiogen-esis in vivo (52). The tumor cells were found to produce substan-tial quantities of VEGF, and increased circulating levels of thisgrowth factor were measured in the sera of tumor bearers.

Interestingly, splenic T lymphocytes from mice bearing tumorsalso secrete VEGF. These constitutive levels were not significantlyaltered by incubation of the T cells with either PHA or IL-2, butPMA did up-regulate production in tumor bearers’ T cells. Tumor-infiltrating Thy1.2� cells were also found to secrete VEGF. Hu-man lymphocytes are known to produce VEGF; negligible VEGFmRNA expression was detected in resting human peripheral Tcells, whereas PMA activation markedly increased the expressionof this growth factor (55). More relevant for our studies is theobservation that immunohistochemically identifiable T lympho-cytes infiltrating human prostate and bladder cancers were identi-fied by in situ hybridization as expressing VEGF mRNA (56). Thisfinding, as well as our observations, suggests that T cell infiltra-tions, regarded as evidence of a protective immune responseagainst tumors, may be promoting vasculogenesis and microvas-cular remodeling. To our knowledge, the only other demonstrationof VEGF production by murine T lymphocytes is during the pro-cess of collateral vessel growth in an ApoE�/� mouse model (71).

Importantly, we also present the novel finding of the expressionof the VEGF receptors Flt-1 and Flk-1/kinase domain region inmurine splenic T cells and the increased expression of these re-ceptors paralleling that of VEGF and tumor burden. Previous in-vestigations have shown that Flt-1 and Flk-1 are normally ex-pressed at low levels in endothelial cells and in inflamed jointtissue, but that they are up-regulated at sites where there is a con-comitant up-regulation of VEGF (72, 73). Our mammary tumorcells produce VEGF, and there are increased circulating levels ofthis growth factor post-tumor implantation. In this study, we dem-onstrate that VEGF can up-regulate MMP-9 expression in the mu-rine T lymphocytes. Because splenic T cells from tumor-bearingmice produce higher levels of endogenous VEGF, we examinedthe possibility that there may be an autocrine effect on T cell-derived MMP-9. To this end, we blocked the in vitro VEGF se-cretion by tumor bearers’ T cells and measured their MMP-9 pro-duction (data not shown). There were no differences between theanti-VEGF-treated and -untreated cells’ MMP-9 levels. Becausethe T lymphocytes are exposed to high serum levels of VEGF forseveral weeks in vivo, it appears that it may be too late to revert theexaggerated production of MMP-9 in tumor bearers’ T cells.

VEGF signaling is mediated through two transmembrane ty-rosine kinase receptors, Flt-1/VEGFR-1 and Flk-1/VEGFR-2.VEGFR-2 is the critical receptor for the proliferation and differ-entiation of endothelial cells, whereas VEGFR-1 may be moreimportant for vascular remodeling (74). Preliminary data using anti-mouse VEGFR-1 and -2 neutralizing Abs indicate that both recep-tors are involved in the VEGF-mediated up-regulation of MMP-9in splenic T cells (data not shown).

In the literature, there are only a few reports of MMP-9 up-regulation by this growth factor in human aortic smooth musclecells (75), human brain tumor cells (76), and human myelo-mono-cytic leukemia cell lines (74). However, there are no reports dem-onstrating the up-regulation of MMP-9 in T cells by this growthfactor. Of all the tumor-derived and tumor-induced factors tested,VEGF was the only one that exerted an effect in the splenic T cells.However, it should be emphasized that we do not believe that thisis the only factor present either in the tumor cells or in the host,which is responsible for the observed up-regulation of MMP-9.Indeed, we continue our investigations for different molecules

4349The Journal of Immunology

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

present in our model system that may be contributing, as is VEGF,to the modulation of this metalloproteinase in the T lymphocytecompartment of tumor bearers.

Our work further supports the previous observations that inflam-matory cells produce significant amounts of proteases within thetumor bed. Coussens et al. have implicated monocytes, neutro-phils, and mast cells as predominant producers of MMP-9 in skincarcinogenesis in K14-HPV16 transgenic mice (21). Recently, thisgroup reported that CD4� T cells enhance skin carcinogenesis byrecruiting MMP-secreting Gr-1�/Mac-1� cells into the neoplasticlesions, but that they do not themselves produce MMP-9 (77). Inthis analysis, we have demonstrated that T cells appear to be keycontributors of MMP-9 in a mammary tumor model, both in theperiphery and within the tumor microenvironment. MMPs are pos-itive regulators of angiogenesis, in that both endogenous and syn-thetic MMP inhibitors reduce tumor angiogenesis in tumor models(10). Angiogenesis is crucial for expansive tumor growth and dis-semination. Tumors do not grow beyond a few mm3 in size with-out vascularization for proper nourishment and removal of metabolicwastes from the site of the tumor (78, 79). MMPs are important mod-ulators of these events because they are responsible for the proteolysisof connective tissue barriers necessary for new vessel formation. Fur-thermore, there is evidence that MMP-9 renders the angiogenic factor,VEGF, more available to its receptors on endothelial cells during pan-creatic islet carcinogenesis (80), adding another dimension to the pro-angiogenic potential of MMP-9.

Based on our findings that MMP-9 and VEGF are up-regulatedin the T lymphocytes from tumor bearers and the role of VEGF inthe elevated MMP-9 production, it is tempting to speculate thatthese two factors may be interacting to promote neoplastic pro-gression in this mammary tumor model. Within the tumor micro-environment of our mammary adenocarcinoma model, T cells aresupplying MMP-9 and VEGF. In addition, several investigationshave shown that, via cellular contact, T cells can induce MMP-9expression in a variety of cell lines, including monocytic and fi-broblastic cells in coculture (81, 82), suggesting that the T cells’secretion of MMP-9 could further potentiate protease productionby other inflammatory cells within the tumor bed. In light of recentexperiments, Coussens et al. have suggested that inflammatory celltypes may become an important class of tumor-associated cellstargeted by anticancer drugs (21). Our studies indicate that VEGF actsdirectly on T lymphocytes and that elevated VEGF levels may beassociated with aberrant MMP-9 secretion, suggesting that modula-tion of MMP-9 expression by anti-VEGF therapeutics might be usefulas a strategy to attenuate protease activity in developing cancers.

VEGF has also been shown to inhibit the functional maturationof dendritic cells, which could result in a decreased ability of theimmune system to generate effective antitumor responses (83). Inaddition, VEGF has been recently been shown to cause thymicatrophy and the inhibition of T cell development (84). In light ofthe demonstration of VEGF receptor expression in murine splenicT cells, it should be determined whether VEGF is capable of mod-ulating T cell function, including cytokine secretion, proliferation,and/or cytotoxicity, in addition to its induction of MMP-9. Indeed,in our model tumor system, we have observed thymic atrophyaccompanying tumor progression (85), as well as a decrease inIFN-� production by splenic T cells (67). Perhaps VEGF has ad-ditional effects on the T cells of our mammary tumor bearers be-yond that of protease induction.

AcknowledgmentsWe are grateful for the excellent technical assistance of Mantley Dorsey,Jr., and Yun Qi Liu. We thank Pat Washington and Michelle Perez for theirhelp in preparing the figures and the manuscript.

References1. Shapiro, S. D. 1998. Matrix metalloproteinase degradation of extracellular ma-

trix: biological consequences. Curr. Opin. Cell Biol. 10:602.2. Basbaum, C. B., and Z. Werb. 1996. Focalized proteolysis: spatial and temporal

regulation of extracellular matrix degradation at the cell surface. Curr. Opin. CellBiol. 8:731.

3. Birkedal-Hansen, H. 1995. Proteolytic remodeling of extracellular matrix. Curr.Opin. Cell Biol. 7:728. .

4. Sato, H., T. Takino, Y. Okada, J. Cao, A. Shinagawa, E. Yamamoto, andM. Seiki. 1994. A matrix metalloproteinase expressed on the surface of invasivetumour cells. Nature 370:61.

5. Liotta, L. A., and W. G. Stetler-Stevenson. 1991. Tumor invasion and metastasis:an imbalance of positive and negative regulation. Cancer Res. 51:S5054.

6. Kim, J., W. Yu, K. Kovalski, and L. Ossowski. 1998. Requirements for specificproteases in cancer cell intravasation as revealed by a novel semiquantitativePCR-based assay. Cell 94:353.

7. Ahrens, D., A. E. Koch, R. M. Pope, M. Stein-Picarella, and M. J. Niedbala.1996. Expression of matrix metalloproteinase-9 (96-kD gelatinase B) in humanrheumatoid arthritis. Arthritis Rheum. 39:1576.

8. Jovanovic, D. V., J. Martel-Pelletier, J. A. Di Battista, F. Mineau, F. C. Joliocoeur,M. Benderdour, and J. P. Pelletier. 2000. Stimulation of 92-kD gelatinase (matrixmetalloproteinase 9) production by interleukin-17 in human monocyte/macro-phages: a possible role in rheumatoid arthritis. Arthritis Rheum. 43:1134.

9. McCawley, L. J., and L. M. Matrisian. 2000. Matrix metalloproteinases: multi-functional contributors to tumor progression. Mol. Med. Today 6:149.

10. Egeblad, M., and Z. Werb. 2002. New functions for the matrix metalloproteinasesin cancer progression. Nat. Rev. 2:161.

11. McCawley, L. J., and L. M. Matrisian. 2001. Matrix metalloproteinases: they’renot just for matrix anymore! Curr. Opin. Cell Biol. 13:534.

12. Nagase, H., and J. F. Woessner, Jr. 1999. Matrix metalloproteinases. J. Biol.Chem. 274: 21491.

13. Parsons, S. L., S. A. Watson, P. D. Brown, H. M. Collins, and R. J. Steele. 1997.Matrix metalloproteinases. Br. J. Surg. 84:160.

14. Nagase, H. 1994. Matrix metalloproteinases. Cancer Metastasis Rev. 4:85.15. Liotta, L. A., K. Tryggvason, S. Garbisa, I. Hart, M. Foltz, and S. Shafie. 1980.

Metastatic potential correlates with enzymatic degradation of basement mem-brane collagen. Nature 284:67.

16. Rao, J. S., M. Yamamoto, S. Mohaman, Z. L. Gokaslan, G. N. Fuller,W. G. Stetler-Stevenson, V. H. Rao, L. A. Liotta, G. L. Nicolson, andR. E. Sawaya. 1996. Expression and localization of 92 kDa type IV collagenase/gelatinase B (MMP-9) in human gliomas. Clin. Exp. Metastasis 14:12.

17. Levy, A., V. Cioce, M. E. Sobel, S. Garbisa, W. F. Griogioni, L. A. Liotta, andW. G. Stetler-Stevenson. 1991. Increased expression of the 72 kDa type IV col-lagenase in human colonic adenocarcinoma. Cancer Res. 51:439.

18. Monteagudo, C., M. Merino, J. San-Juan, L. A. Liotta, and W. G. Stetler-Stevenson.1990. Immunohistologic distribution of type IV collagen in normal, benign, andmalignant breast tissue. Am. J. Pathol. 136:585.

19. Nielson, B. S., S. Timshel, L. Kjeldsen, M. Sehested, C. Pyke, N. Borregaard, andK. Dan. 1996. 92 kDa type IV collagenase (MMP-9) is expressed in neutrophilsand macrophages but not in malignant epithelial cells in human colon cancer. Int.J. Cancer 65:57.

20. Heppner, K. J., L. M. Matrisian, R. A. Jensen, and W. H. Rodgers. 1996. Ex-pression of most matrix metalloproteinase family members in breast cancer rep-resents a tumor-induced host response. Am. J. Pathol. 149:273.

21. Coussens, L. M., C. L. Tinkle, D. Hanahan, and Z. Werb. 2000. MMP-9 suppliedby bone marrow-derived cells contributes to skin carcinogenesis. Cell 103:481.

22. Huang, S., M. Van Arsdall, S. Tedjarati, M. McCarty, W. Wu, R. Langley, andI. J. Fidler. 2002. Contributions of stromal metalloproteinase-9 to angiogenesisand growth of human ovarian carcinoma in mice. J. Natl. Cancer Inst. 94:1134.

23. Chang, C., and Z. Werb. 2001. The many faces of metalloproteinases: cellgrowth, invasion, angiogenesis, and metastasis. Trends Cell Biol. 11:S37.

24. Stetler-Stevenson, M., A. Mansoor, M. Lim, P. Fukushima, J. Kehrl, G. Marti,K. Ptaszynski, J. Wang, and W. G. Stetler-Stevenson. 1997. Expression of matrixmetalloproteinases and tissue inhibitors of metalloproteinases in reactive and neo-plastic lymphoid cells. Blood 89:1708.

25. Weeks, B. S., H. W. Schnaper, M. Handy, E. Holloway, and H. K. Kleinman.1993. Human T lymphocytes synthesize the 92 kDa type IV collagenase (gela-tinase B). J. Cell. Physiol. 157:644.

26. Leppert, D., E. Waubant, R. Galardy, N. W. Bunnett, and S. L. Hauser. 1995. Tcell gelatinases mediate basement membrane transmigration in vitro. J. Immunol.154:4379.

27. Medina, D., and K. B. DeOme. 1969. Response of hyperplastic alveolar noduleoutgrowth line D1 to mammary tumor virus, nodule-inducing virus, and pro-longed hormonal stimulation acting singly and in combination. J. Natl. CancerInst. 42:303.

28. Julius, M. H., E. Simpson, and L. A. Herzenberg. 1973. A rapid method for theisolation of functional thymus directed murine lymphocytes. Eur. J. Immunol.147:645.

29. Heussen, C., and E. B. Dowdle. 1980. Electrophoretic analysis of plasminogenactivators in polyacrylamide gels containing sodium dodecyl sulfate and copo-lymerized substrates. Anal. Biochem. 102:196.

30. Herron, G. S., Z. Werb, K. Dwyer, and M. J. Banda. 1986. Secretion of metal-loproteinases by stimulated capillary endothelial cells. II. Expression of collage-nase and stromelysin activities is regulated by endogenous inhibitors. J. Biol.Chem. 261:2814.

4350 UP-REGULATION OF MMP-9 IN TUMOR BEARERS’ T LYMPHOCYTES

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

31. Roach, J., S. J. Choi, R. L. Schaub, R. J. Leach, G. D. Roodman, and S. V. Reddy.1998. Further characterization of the murine collagenase (type IVB) gene pro-moter and analysis of MRNA expression in murine tissues. Gene 208:117.

32. Finnerty, H., K. Kelleher, G. E. Morris, K. Bean, D. M. Merberg, R. Kriz,J. C. Morris, H. Sookdeo, K. J. Turner, and C. R. Wood. 1993. Molecular cloningof murine FLT and FLT4. Oncogene 8:2293.

33. Quinn, T. P., K. G. Peters, C. de Vries, N. Ferrara, and L. T. Williams. 1993. Fetalliver kinase 1 is a receptor for vascular endothelial growth factor and is selec-tively expressed in vascular endothelium. Proc. Natl. Acad. Sci. USA 90:7533.

34. Sugihara, T., S. C. Kaul, Y. Mitsui, and R. Wadhwa. 1994. Enhanced expressionof multiple forms of VEGF is associated with spontaneous immortalization ofmurine fibroblasts. Biochim. Biophys. Acta 1224:365.

35. Muta, H., L. H. Boise, L. Fang, and E. R. Podack. 2000. CD30 signals integrateexpression of cytotoxic effector molecules, lymphocyte trafficking signals, andsignals for proliferation and apoptosis. J. Immunol. 165:5105.

36. Topalian, S. L., D. Solomon, and S. A. Rosenberg. 1989. Tumor-specific cytol-ysis by lymphocytes infiltrating human melanomas. J. Immunol. 142:3714.

37. Johnatty, R. N., D. D. Taub, S. P. Reeder, S. M. Turcovski-Corrales,D. W. Cottam, T. J. Stephenson, and R. C. Rees. 1997. Cytokine and chemokineregulation of proMMP-9 and TIMP-1 by human peripheral blood lymphocytes.J. Immunol. 158:2327.

38. Zhou, H., E. J. Bernhard, F. E. Fox, and P. C. Billings. 1993. Induction ofmetalloproteinase activity in human T-lymphocytes. Biochim. Biophys. Acta1177:174.

39. Masure, S., G. Nys, P. Fiften, J. Van Damme, and G. Opdenakker. 1993. Mousegelatinase B: cDNA cloning, regulation of expression and glycosylation inWEHI-3 macrophages. Eur. J. Biochem. 218:129.

40. Tanaka, H., K. Hojo, H. Yoshida, Y. Takayuki, and K. Sugita. 1993. Molecularcloning and expression of the mouse 105-kDa gelatinase cDNA. Biochem. Bio-phys. Res. Commun. 190:732.

41. Stetler-Stevenson, W. G., H. C. Krutzsch, and L. A. Liotta. 1989. Tissue inhibitorof metalloproteinase (TIMP-2), a new member of the metalloproteinase inhibitorfamily. J. Biol. Chem. 262:17374.

42. Stetler-Stevenson, W. G., P. D. Brown, M. Onisto, A. T. Levy, and L. A. Liotta.1990. Tissue inhibitor of metalloproteinase-2 (TIMP-2) mRNA expression intumor cell lines and human tumor tissues. J. Biol. Chem. 265:13933.

43. Albini, A., A. Melchiori, L. Santi, L. A. Liotta, P. D. Brown, andW. G. Stetler-Stevenson. 1991. Tumor cell invasion inhibited by TIMP-2. J. Natl.Cancer Inst. 83:775.

44. Goldberg, G. I., A. Strongin, I. E. Collier, L. T. Genrich, and B. L. Marmer. 1992.Interaction of the 92-kDa type IV collagenase with the tissue inhibitor of metallo-proteinases prevents dimerization, complex formation with interstitial collagenase,and activation of the proenzyme with stromelysin. J. Biol. Chem. 267:4583.

45. Lopez, D. M., M. Lopez-Cepero, G. A. Watson, A. Ganju, E. Sotomayor, andY.-X. Fu. 1991. Modulation of the immune system by mammary tumor-derivedfactors. Cancer Invest. 9:643.

46. Fu, Y.-X., G. Watson, J. J. Jimenez, Y. Wang, and D. M. Lopez. 1990. Expansionof immunoregulatory macrophages by granulocyte-macrophage colony-stimulat-ing factor derived from a murine mammary tumor. Cancer Res. 50:227.

47. Calderon, C., Z.-H. Huang, D. A. Gage, E. M. Sotomayor, and D. M. Lopez.1994. Isolation of a nitric oxide inhibitor from mammary tumor cells and itscharacterization as phosphatidyl serine. J. Exp. Med. 180:945.

48. Lopez-Cepero, M., J. A. Garcia-Sanz, L. Herbert, R. Riley, M. E. Handel,E. R. Podack, and D. M. Lopez. 1994. Soluble and membrane bound TNF-� areinvolved in the cytotoxic activity of B cells from tumor bearing mice againsttumor targets. J. Immunol. 152:3333.

49. Charyulu, V., and D. M. Lopez. 2000. Elevated GM-CSF levels in tumor bearingmice upregulates IL-6 production by B cells via a mechanism independent ofTNF-�. Int. J. Oncol. 16:161.

50. Handel-Fernandez, M. E., and D. M. Lopez. 1998. Altered binding of the NF-GMa transcription factor is involved in the upregulated production of TNF-� bymacrophages of mammary tumor bearing mice. Int. J. Mol. Med. 1:227.

51. Lopez, D. M., M. E. Handel-Fernandez, X. Cheng, V. Charyulu, L. M. Herbert,M. R. DiNapoli, and C. L. Calderon. 1996. Cytokine production by lymphore-ticular cells from mammary tumor bearing mice: the role of tumor-derived fac-tors. Anticancer Res. 16:3923.

52. Ferrara, N., and T. Davis-Smyth. 1997. The biology of vascular endothelialgrowth factor. Endocr. Rev. 18:4.

53. Robinson, C. J., and S. E. Stringer. 2001. The splice variants of vascular endo-thelial growth factor (VEGF) and their receptors. J. Cell Sci. 114:853.

54. Senger, D. R., L. Van De Water, L. F. Brown, J. A. Nagy, K.-T. Yeo, T.-K. Yeo,B. Berse, R. W. Jackman, A. M. Dvorak, and H. F. Dvorak. 1993. Vascular perme-ability factor (VPF, VEGF) in tumor biology. Cancer Metastasis Rev. 12:303.

55. Iijima, K., N. Yoshikawa, and H. Nakamura. 1996. Activation-induced expres-sion of vascular permeability factor by human peripheral T cells: a non-radio-isotopic semiquantitative reverse transcription-polymerase chain reaction assay.J. Immunol. Methods 196:199.

56. Freeman, M. R., F. X. Schneck, M. L. Gagnon, C. Corless, S. Soker, K. Niknejad,G. E. Peoples, and M. Klagsbrun. 1995. Peripheral blood T lymphocytes andlymphocytes infiltrating human cancers express vascular endothelial growth fac-tor: a potential role for T cells in angiogenesis. Cancer Res. 55:4140.

57. Ferrara, N., H.-P. Gerber, and J. LeCouter. 2003. The biology of VEGF and itsreceptors. Nat. Med. 9:669.

58. Buessow, S. C., R. D. Paul, A. M. Miller, and D. M. Lopez. 1984. Lymphore-ticular cells isolated by centrifugal elutriation from a mammary adenocarcinoma.I. Characterization of an in situ lymphocyte suppressor population by surfacemarkers and functional reactivity. Int. J. Cancer 33:79.

59. Topalian, S. L., and S. A. Rosenburg. 1990. Tumor-infiltrating lymphocytes:evidence for specific immune reactions against growing cancers in mice andhumans. In Important Advances in Oncology. V. T. DeVita, S. Hellman, andS. A. Rosenburg, eds. JB Lippincott, Philadelphia, PA, p. 19.

60. Imhof, B. A., and D. Dunon. 1995. Leukocyte migration and adhesion. Adv.Immunol. 58:345.

61. Xie, B., Z. Dong, and I. J. Fidler. 1993. Regulatory mechanisms for the expres-sion of type IV collagenases/gelatinases in murine macrophages. J. Immunol.152:3637.

62. Goetzl, E. J., M. J. Banda, and D. Leppert. 1996. Matrix metalloproteinases inimmunity. J. Immunol. 156:1.

63. Brew, K., D. Dinakarpandian, and H. Nagase. 2000. Tissue inhibitors of metallopro-teinases: evolution, structure and function. Biochim. Biophys. Acta 1477:267.

64. Gomez, D. E., D. F. Alonso, H. Yoshiji, and U. P. Thorgeirsson. 1997. Tissueinhibitors of metalloproteinases: structure, regulation and biological functions.Eur. J. Cell Biol. 74:111.

65. Kossakowska, A. E., S. J. Urbanski, and A. Janowska-Wieczorek. 2000. Matrixmetalloproteinases and their tissue inhibitors: expression, role and regulation inhuman malignant non-Hodgkin’s lymphomas. Leuk. Lymphoma 39:485.

66. Handel-Fernandez, M. E., X. Cheng, L. M. Herbert, and D. M. Lopez. 1997.Downregulation of IL-12, not a shift from a T helper-1 to T helper-2 phenotype,is responsible for impaired IFN-� production in tumor bearing mice. J. Immunol.158:280.

67. Cheng, X.-F., and D. M. Lopez. 1998. CD4� but not CD8� T cells from mam-mary tumor bearing mice have a downregulated production of IFN-�: role ofphosphatidyl serine. J. Immunol. 160:2735.

68. Leppert, D., S. L. Hauser, J. L. Kishiyama, S. An, L. Zheng, and E. J. Goetzl.1995. Stimulation of matrix metalloproteinase-dependent migration of T cells byeicosanoids. FASEB J. 9:1473.

69. Ries, C., H. Kolb, and P. E. Petrides. 1994. Regulation of 92-kD gelatinaserelease in HL-60 leukemia cells: tumor necrosis factor-� as an autocrine stimulusfor basal- and phorbol ester-induced secretion. Blood 83:3638.

70. Kossakowska, A. E., D. R. Edwards, C. Prusinkiewicz, M. C. Zhang, D. Guo,S. J. Urbanski, T. Grogan, L. A. Marquez, and A. Janowska-Wieczorek. 1999.Interleukin-6 regulation of matrix metalloproteinase (MMP-2 and MMP-9) andtissue inhibitor of metalloproteinase (TIMP-1) expression in malignant non-Hodgkin’s lymphomas. Blood 94:2080.

71. Couffinhal, T., M. Silver, M. Kearney, A. Sullivan, B. Witzenbichler, M. Magner,B. Annex, K. Peters, and J. M. Isner. 1999. Impaired collateral vessel develop-ment associated with reduced expression of vascular endothelial growth factor inApoE�/� mice. Circulation 24:3188.

72. Brown, L. F., B. Berse, K. Tognazzi, E. J. Manseau, L. van De Water,D. R. Senger, H. F. Dvorak, and S. Rosen. 1992. Vascular permeability factormRNA and protein expression in human kidney. Kidney Int. 42:1457.

73. Lu, J., T. Kasama, K. Kobayashi, Y. Yoda, F. Shiozawa, M. Hanyuda,M. Negishi, H. Ide, and M. Adachi. 2000. Vascular endothelial growth factorexpression and regulation of murine collagen-induced arthritis. J. Immunol.164:5922.

74. Dias, S., K. Hattori, Z. Zhu, B. Heissig, M. Choy, W. Lane, Y. Wu, A. Chadburn,E. Hyjek, M. Gill, et al. 2000. Autocrine stimulation of VEGFR-2 activates hu-man leukemic cell growth and migration. J. Clin. Invest. 106:511.

75. Wang, H., and J. A. Keiser. 1998. Vascular endothelial growth factor upregulatesthe expression of matrix metalloproteinases in vascular smooth muscle cells: roleof flt-1. Circ. Res. 83:832.

76. Rooprai, H. K., G. J. Rucklidge, C. Panou, and G. J. Pilkington. 2000. The effectsof exogenous growth factors on matrix metalloproteinase secretion by humanbrain tumour cells. Brit. J. Cancer 82:52.

77. Daniel, D., N. Meyer-Morse, E. K. Bergsland, K. Dehne, L. M. Coussens, andD. Hanahan. 2003. Immune enhancement of skin carcinogenesis by CD4� Tcells. J. Exp. Med. 197:1017.

78. Folkman, J. 1971. Tumor angiogenesis: therapeutic implications. N. Engl. J. Med.285:1182.

79. Hanahan, D., and J. Folkman. 1996. Patterns and emerging mechanisms of theangiogenic switch during tumourigenesis. Cell 86:353.

80. Bergers, G., R. Brekken, G. McMahon, T. H. Vu, T. Itoh, K. Tamaki,K. Tanzawa, P. Thorpe, S. Itohara, Z. Werb, and D. Hanahan. 2000. Matrixmetalloproteinase-9 triggers the angiogenic switch during carcinogenesis. Nat.Cell Biol. 2:737.

81. Miltenburg, A. M., S. Lacraz, H. G. Welgus, and J. M. Dayer. 1995. Immobilizedanti-CD3 antibody activates T cell clones to induce the production of interstitialcollagenase, but not tissue inhibitor of metalloproteinases, in monocytic THP-1cells and dermal fibroblasts. J. Immunol. 154:2655.

82. Aoudjit, F., P. O. Esteve, M. Desrosiers, E. F. Potworoski, and Y. St-Pierre. 1997.Gelatinse B (MMP-9) production and expression by stromal cells in the normaland adult thymus and experimental thymic lymphoma. Int. J. Cancer. 71:71.

83. Gabrilovich, D. I., H. L. Chen, K. R. Girgis, H. T. Cunningham, G. M. Meny,S. Nadaf, D. Kavanaugh, and D. P. Carbone. 1996. Production of vascular en-dothelial growth factor by human tumors inhibits the functional maturation ofdendritic cells. Nat. Med. 2:1096.

84. Ohm, J. E., D. I. Gabrilovich, G. D. Sempowski, E. Kisseleva, K. S. Parman,S. Nadaf, and D. P. Carbone. VEGF inhibits T-cell development and may con-tribute to tumor-induced immune suppression. Blood 101:4878.

85. Adkins, B., V. Charyulu, Q.-L. Sun, D. Lobo, and D. M. Lopez. 2000. Earlyblock in maturation is associated with thymic involution in mammary tumor-bearing mice. J. Immunol. 164:5635.

4351The Journal of Immunology

by guest on June 18, 2018http://w

ww

.jimm

unol.org/D

ownloaded from