Inhibition of MMP-1 expression by antisense RNA decreases invasiveness of human chondrosarcoma

~

ELSEVIER

~

~

~

Journal of Orthopaedic Research

Journal of Orthopaedic Research 21 (2003) 1063-1070

www.elsevier.com/locate/orthres

Inhibition of MMP-1 expression by antisense RNA decreases invasiveness of human chondrosarcoma Xiaoling Jiang a, Char 1 M. Dutton a, Wen-ning Qi b, Joel A. Block ‘, Pnina Brodt Margaret Durko d,e, Sean P. Scully a,*

{

Department of Orthopaedic Research, Mayo Clinic, 200 First Street S W, Rochester, M N 55905, USA Division of’ Orthopaedic Surgery, Duke University Medical Center, Durhum, NC 27710, USA ‘ Depurtment of Rheumutology, Rush-Presbyterian St. Luke’s Medical Center, Chicugo. IL 60612, USA Dcyrtment of Surgery. Division of’ Surgical Resetrrch, McGill University and Royal Victoria Ho.spituI, Montreal, Que.. Cunudu H3A I A1 Montreul Ncwologicul Institute, Montreul, Que., Cunudci H 3 A 2 B4 Accepted 20 March 2003

Abstract

We previously reported that an elevated level of matrix metalloproteinase-1 (MMP-I) gene expression in patients with chondrosarcoma has a strong statistical correlation with recurrence and in vitro invasion. In the present study, we used a n antisense RNA strategy for MMP-1 inhibition to determine if this would affect the invasive characteristics of the cells. We transfected a human chondrosarcoma cell line with a retroviral plasmid expressing a 770 bp genomic fragment of the human MMP-I gene in the sense or antisense orientation. The results show that cells transfected with the MMP-1 antisense fragment had a significant decrease in both MMP-1 protein and enzyme activity (p < 0.05) as compared to cells transfected with an empty plasmid or the parental cells. Cells transfected with the MMP-I antisense fragment demonstrated a significant decrease in their ability to invade the collagen I barrier (p < 0.05). The gene expression for MMP-8 and MMP-13 were unaffected in cells transfected with the MMP-I antisense fragment, MMP-1 sense fragment, or empty plasmid. These results support the hypothesis that MMP-I facilitates tumor cell egress from chondrosarcoma tissue and demonstrate the potential of MMP-1 as a promising target for a novel biologic therapy in chondrosarcoma. 0 2003 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved Abhrcviations: MMPs, matrix metalloproteinase; MMP-1, interstitial collagenase; (3418, geneticin; 35012, a human chondrosarcoma cell line; dsRNA, double-stranded RNA K ~ y i . o r d ~Invasion; : Chondrosarcoma; MMP-1; Metastasis; Antisense RNA

Introduction

One characteristic of cancer is the ability of tumor cells to invade normal tissue and spread to distant sites. This process, termed metastasis, is the major cause of death in cancer patients and the major impediment to curative treatment [ 1,5,7]. In some mesenchymal malignancies, such as Ewing’s sarcoma and osteosarcoma, systemic chemotherapy and radiation therapy have increased survival several fold in the past few decades [27,3 1,35,42]. In chondrosarcoma, tumor cells are re-

* Corresponding author. Tel.: + 1-507-284-9457: fax: + 1-507-2845075. E-rnuil address: [email protected] (S.P. Scully).

sistant to existing adjuvant therapies and a novel approach to treatment is critically needed. Invasion and metastasis involves detachment of cells from the tumor, attachment of cancer cells to the basement membrane, degradation of the local connective tissue and invasion of tumor cells into circulation, followed by extravasation from circulation and the growth of tumor cells in a secondary site [8,17,20,25]. The detachment of tumor cells from solid tumors is incompletely understood but involves the degradation of the extracellular matrix including induced biologic signaling [44], escape from apoptosis, and locomotion in the extracellular matrix [ 5 ] . During tumor cell growth, both in primary and secondary sites, neoangiogenesis in the tumor is necessary for tumor cells to acquire nutrients and to metastasize. In tumor cell invasion and

0736-0266/$ - see front matter 0 2003 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved doi: 10.1016/S0736-0266(03)00079-2

~

1064

X . Jiung et

ill.

/ Journal of 01-thr~puerlic Raseurch 21 !2003) 1063-1070

metastasis, the two critical events, extracellular matrix degradation and angiogenesis, involve matrix metalloproteinase (MMP) activity [3,5,10,30]. MMP-1 is an interstitial collagenase and is produced by a variety of normal cells, e.g., stromal fibroblasts, macrophages, endothelial cells and epithelial cells, as well as numerous tumors. Physiological expression of MMP-1 by most cells is low but readily induced by phorbol esters, growth factors, and inflammatory cytokines. Recently, MMP-I has been implicated in a variety of advanced cancers, and in nearly all instances, there was a significant link between expression of MMP-1 and survival [2,22,23]. In previous studies, we found that MMP-I could serve as a prognostic indicator for local recurrence, metastasis, and survival in chondrosarcoma patients [4,38,39]. As more is learned about the role MMP-1 plays in chondrosarcoma metastasis, its potential for prognostication and as a target for novel biologic therapy becomes clearer. We hypothesized that MMP-1 facilitated tumor spread by degrading the extracellular collagen matrix and permitting cell dissociation from the tumor. This may not be via a strictly mechanical constraint to migration but may involve induction of biological signaling. To test the hypothesis that MMP-1 facilitates cell egress and further understand the mechanisms involved in chondrosarcoma metastasis, we sought to inhibit MMP-1 and determine the effects of MMP-I inhibition on invasiveness. Currently, there are three conceptual approaches to MMP inhibition [I I]. The native inhibitors, such as the tissue inhibitor of metalloproteinase, are composed of four molecular species classified by the substrate specificity and tissue distribution. This type of inhibitor works in a reversible manner: it promotes inhibition when it binds to MMPs and loses inhibition when it dissociates from MMPs. The second subgroup contains synthetic inhibitors of MMPs that occupy the active site on metalloproteinases by competing with native collagen molecules [37]. Because of the configuration of the active site of MMP-I, there are currently no specific MMP-I inhibitors. The third approach used is to downregulate MMP expression [21]. Due to the absence of a specific protein inhibitor for MMP-I, we used an antisense approach to inhibit MMP-1 activity in JJ012 cells, and demonstrate that by decreasing the level of MMP-I in human chondrosarcoma cells, there is a reduction in in vitro invasiveness.

Materials and methods C'ell culturc IJ012 cells were cultured iis a monolayer in 10% FBS complete media which consists of40'%,DMEM (high glucose), 40% MEM (alpha medium) and 10% F12 ( H A M ) plus 1.2 pg human insulin, 28.1 pg

hydrocortisone and 25 mg gentamicin in 500 ml of media. The JJ-sense cells, JJ-antisense cells, and JJ-empty cells are JJOl2 cells transfected respectively with MMP-1 sense plasmid, MMP-1 antisense plasmid, or an empty plasmid. These cells were cultured in lo'%) FBS complete media plus selective agent G41 8 (Geneticin, Life Technologies. Inc., Rockville. M D ) at a concentration of 600 pg/ml of media. The cells were grown to 90% and passaged at a dilution of 1.6. Construction of M M P - I smse. urztisense unrl iwpt,y plii~snii~l.\ A 770 bp genomic DNA fragment of human MMP-I was generated by polymerase chain reaction (PCR) performed on the crude lysate of normal lung fibroblasts. The PCR prodnct was inserted into the Eco RI site of the pSVK3 plasmid (Pharmacia, Peapack, NJ) in the sense and antisense orientations relative to the SV40 origin/early promoter region as was described in detail elsewhere [I41 (Figs. 1 and 21. The neomycin resistance gene is also expressed in this plasmid. The empty plasmid was constructed by digesting the sense plasmid with Eco RI, separating the large fragments on an 0.8'%,agarose gel, then ligating at the Eco RI site. Trunsfiction with M M P-1 .scwsc, unriserise and cnipt,v plusinid\ Cell transfection was performed with Cytofectene (Bio-Rad, Hercules, CA). The day prior to transfection, cells were cultured in a 6-well plate at a concentration of 5 x lo5 cells per well. Two microgram of sense, antisense, or empty vector plasmid were diluted with 100 111 of serum-free complete media. Six microliters of the transfection agent Cytofectene was added. mixed gently, and incubated at room temperature for 20 min. Ten percent FBS complete media was added to the above mixture of Cytofectene/plasmid for a final volume of 1 ml. The media was removed from the cells in the 6-well plate, replaced with 1 ml of' the transfection complex of Cytofectenelplasmid/l0'%, FBS

--

Construction of Expression Plasmids PCR Fragment

*

4

exonl

exon2 exon3

exon 10

MMP-1 Gene Structure

1 SV40 promoter sense fragment

neo gene

MMP-1 Sense Plasmid

m SV40 promoter antisense fragment neo gene

MMP-1 Antisense Plasmid

Empty Plasmid SV40 promoter

neo gene

Fig. 1. The construction of the MMP-I sense, antisense and empty plasmids. A 770 bp DNA fragment of MMP-1 was produced by PCR performed on the crude lysate of normal lung fibroblasts. This PCR product encompassed a portion of human genomic MMP-I DNA, which extended from the 5' UTR to include a portion of exon 2. The fragment was inserted into the Eco RI site of a pSVK3 plasmid in both sense (MMP-1 sense plasmid) and antisense orientation (MMP-1 antisense plasmid) relative to the SV40 originiearly promoter. Both MMP-I sense plasmid and antisense plasmid were checked for fidelity by restriction enzyme mapping and fragment DNA sequencing. The empty plasmid was constructed by digesting the sense plasmid with Eco RI, separating the large fragments on a 0.8%)agarose gel, then ligating. Included in all three plasmids is a selective marker gene (neo, aminoglycoside phosphotransferase) positioned downstream of the Eco RI site under the control of SV40 early promoter for cell selection.

X . Jirrrig e l ul. I JoLiwu1 o f Orthopueu'ic Restwcli 21 (2003) 1063-1070 Qurmtiturirv cell riiigrution

Fig. 2. Eco RI restriction enzyme digest of MMP-I antisense, sense and empty plasmids. The Eco RI digest pattern showed that the MMP-I sense insert (760 bp) was ten bp less than the MMP-I antisense insert (770 bp). This is due to terminal modifications of the sense and antisense MMP-I fragments, which ensure their proper orientation into the pSVK3 plasmid. complete media. and incubated at 37 "C, 5% COz for 4 h. One ml of 10%)FBS complete media was added, followed by overnight incubation. The following day, the transfection complex was removed from the cells. replaced with fresh 10% FBS complete media. and incubated overnight. The media was replaced with 10% FBS complete media plus 600 pglml of selective agent (3418. The G418-resistant cells. which had successfully taken up the MMP-I sense plasmid, MMP-I antisense plasmid. or the empty plasmid was selected under the presence of G418 for 4 weeks. Prcyxrrtrtion of i~otiditionc~tl riicJiiirr coiicentrute

Cells wcrc grown to 8 0 X confluency, then washed and cultured with serum-free complete medium. The supernatant was collected every 24 h for 3 days. Cells and debris were removed from the supcrnatants by centrifugation, concentrated by freeze-drying at -60 "C (Mayo Protein Core Facility). and stored at -80 "C. Wc.stc,m hlot u ~ ~ u l ~ * . ~ i s

SDS-PAGE and Western blotting were performed as previously described [26,41]. Briefly, the serum-free conditioned media supernatant concentrates were separated on a 10%)SDS-polyacrylamide gel (20 pg of protein/lane) and transferred onto an Immobilon-P transfer membrane (Millipore. Bedford, MA). The blots were probed with a mixture of mouse monoclonal antibody (MMP-I Ab-2 and Ab-3, NeoMarkers, Inc., Fremont, CA) at a dilution of 1:3300. Horseradish peroxidase-conjugated rabbit anti-mouse IgG was used a s the secondary antibody at a dilution of 1: 10.000. A commercially available MMP-I standard was used as a positive control (Sigma, St. Louis, MO).

T i p

I cdltrgeriusr uctii~i/j* u s q

A collagenaae activity assay was performed according to the inanufxturcr's instruction (Type I Collagenase Activity Assay Kit, Chemicon International. Inc.. Temecula, CA) [39]. BrieHy, serum-free conditioned media concentrates were activated with 20 niM APMA (paniinophenylmercuric acetate, Sigma, St. Louis, MO) and incubated for 2 h at 37 "C with biotinylated collagenase substrate. Enhancer was added and the samples were incubated for an additional 30 min. The above mixture was transferred to a biotin-binding plate, incubated for 30 min and washed with assay buffer. Streptavidin-enzyme conjugate was added then washed, followed by the addition of a substrate solution. The activity of the sample was calculated by comparing the degradation of biotinylated collagen to the MMP-I standard.

I065

LISSUJ

A quantitative cell migration assay was performed per the manufacturer's instructions (QCM'" Collagen I Quantitative Cell Migration Assay, Chemicon International, Inc. Temecula, CA) [39]. Briefly, cells were grown in 10'%1 FBS complete media until they reached 80'%1 confluency. They were then washed. the media replaced with a serumfree complete media, and incubated for 24 h. The cells were dispersed with a 0.25% trypsin-EDTA solution (Sigma, St. Louis, MO). counted and re-seeded at 2.0 x 10' cellsl300 pI onto a Boyden chamber precoated with Type I collagen. After incubating for 48 h at 37 "C. 5'!4 COz, the cells that had invaded through the matrix barrier of collagen I were stained. The eluent from the stained cells was collected and measured at 570 nm. The chamber pre-coated with bovine serum albumin was used as a negative control. The cell migration was quantitated by subtracting the optical absorbance from cells passed through the chamber coated with bovine serum albumin from that of the chamber coated with collagen 1. Nortliern blot

Total RNA was isolated following the manupacturer's instructions (RNAqeuousTMMidi Total RNA Isolation Kit, Ambion, Austin, TX). Briefly, cells were harvested from culture flasks with 0.25'%,trypsinl EDTA and re-suspended in lysis solution at -2 x 10' cells/ml. The cell lysate was passed through a glass fiber column and total RNA was eluted from the column. Samples of total RNA were treated with DNase I (DNA-free, Ambion, Austin, TX), resolved on ii denaturing 1% agarose gel ( I 5 pg of total RNAilane) at 5 Vlcni. and transferred to a positively charged nylon membrane (Bright Star"' Plus, Ambion, Austin, TX). The human GAPDH (ATCC # 5303919) and MMP-I probes (ATCC ## 6122647) were generated separately from pOTB7 plasmids. The bacteria strains containing each plasmid were grown at 30 "C with chloramphenicol per the instructions (American Type Culture Collection, Manassas, VA). Plasmid DNA was extracted following the instructions of the DNA purification kit (Wizard Plus Midipreps. Promega, Madison, WI). The inserts of human GAPDH or MMP-I were recovered from a l'!4 electrophesis gel after treatment of the plasmid DNA with Xho IlEco RI. The probes were labeled with [z-~'P] dCTP. according to the manufacturer's instructions (Random Primed DNA Labeling Kit, Roche, Indianapolis, IN). The radiolabeled probes were hybridized to the total RNA blot membrane. Membranes were probed with MMP-I probe or GAPDH probe, and then exposed to film. The film was scanned and analyzed by Scion Image program (NIH, Bethesda. M D ) with normalization of the MMP-1 sample band to GAPDH. Sou t l i i w hlo /

DNA was extracted per the manufacturer's instructions (Genomic DNA Purification Kit, Gentra, Minneapolis. MN). Cells were harvested from flasks with 0.25% trypsin/EDTA, and re-suspended in cell lysis solution at approximately 2 x lo7 ceH/ml. After treatment with RNase A and precipitation of protein with protein precipitation solution, the genomic DNA was precipitated with isopropanol, and redissolved in DNA hydration solution. The genomic DNA samples were digested separately with Eco RllXho I, Hap 11, Hpa 11, or Msp 1 at 37 "C for at least 2 h, resolved on a 1% agarose gel at 1.5 Vlcm overnight, transferred to a positively charged nylon membrane and crosslinked. The MMP-I probe preparation and labeling with [%-"PI dCTP was the same as that used for the Northern blot. The DNA blot membrane was pre-hybridized ExpressHyb solution (CloneTech, Palo Alto, CA) at I0 mV100 cm' and incubated at 60 "C overnight. For hybridization, new ExpressHyb solution containing the MMP-I probe at lo6 cpm/ml was added and incubated at 60 "C overnight. The membrane was washed and exposed to X-ray film. Seiiii-~irurititrrti~.e RT-PCR

Semi-quantitative reverse transcriptase polymerase chain reaction (RT-PCR) was used to evaluate MMP-I, MMP-8, and MMP-13 mRNA expression levels. Sequences of the primers used for RT-PCR are as follows: MMP-IF, 5' GGAC CAAC AATT TCAC AGAG 3',

1066

X . Jrang et aI. I Journal

of

Orthopmdic Reseurch 21 (2003) 1063-1070

MMP-IR, 5' GCGC ATGT AGAA TCTG TCTT 3', MMP-8F, 5' GCAT CCTT CCAT GCAG GACA C 3', MMP-8R, 5' C G G G TCAG ACCA TGAC TAGA C 3', MMP-I3F, 5' T A C T AGAT G T G G CCTT TGAA 3'. MMP-t3R, 5' CCAT GGCC TTAG ACTA TTTT 3', GAPDH-F, 5' ACCA CAGT CCAT GCCA TCAC 3', GAPDH-4, 5' TCCA CCAC CCTG TTGC TGTA 3'. Total RNA from cells transfected with antisense, sense o r empty vectors was isolated following the instructions of the RNeasy Mini Kit (Qiagen. Valencia, CA). Briefly, cells were harvested from culture flasks with 0.25'%1trypsin1EDTA and re-suspended in lysis solution at approximately 1.5 x 10' cellslml. The total RNA was eluted from the RNeasy mini column and treated with DNase I (DNA-free, Ambion, Austin, TX). The cDNA was synthesized from 1 pg of total R N A using the Omniscript R T Kit (Qiagen. Valencia, CA). Two microliters of the cDNA synthesis reaction were amplified in a total volume of 25 p1 containing I x PCR buffer, 0.2 pM dNTPs, 1.5 m M MgCL, 0.2 pM of each primer, and 1.25 U of thermostable Taq DNA polymerase (Roche, Indianapolis, IN). After an initial denaturation at 94 "C for 5 min, cycling conditions were 94 "C for 1 min, 60 "C for 1 min, and 72 "C for 1 min, followed by a final extension and incubation at 72 "C for 7 min. The appropriate number of cycles that provided linear amplification of the PCR product was determined for each primer pair. If no bands were amplified using 2 111of the cDNA synthesis reaction, the amount of template was increased up to 4 p1 in the PCR reaction. The PCR products were resolved on 1.2% agarose gel, and the corresponding individual gene band densities were analyzed using Scion Image software (NIH, Bethesda, M D ) and normalized relative to GAPDH.

MMP-1 Western Blot

r li

I I

Fig. 3 . Western blot of MMP-I. The samples of conditioned media were resolved o n a 10% SDS-PAGE. Twenty micrograms of protein for each sample was loaded in each lane. The 46 kDa MMP-I protein was detected with a mixture of monoclonal antibodies (MMP-I Ah-2 and MMP-I Ah-3) at a dilution of 1:3300. The JJ-sense cells, JJ-antisense cells, and JJ-empty cells are JJ012 cells transfected respectively with MMP-I sense plasmid, MMP-I antisense plasmid, or an empty plasmid. MMP-I protein expression was significantly lower in JJantisense cells ( 9 0 % p~ < 0.05) and JJ-sense cells (80'%1,p < 0.05) than that of 55012 cells. There was no significant difference in MMP-I protein expression between JJOl2 cells and JJ-empty cells.

Stutisticul onrrlj~sis

All experiments were performed in triplicate. Data analysis was performed using one-way ANOVA and unpaired t-test (GraphPad Prism, GraphPad Software, Inc., San Diego, CA). A p-value of less than 0.05 was considered statistically significant.

MMP-1 Enzymatic Activity Assay

m

0

-e .-.-7

0

al

.-

Results

z

We transfected the human chondrosarcoma cell line 55012 with MMP-1 antisense, sense or empty plasmids to test the hypothesis that inhibition of MMP-1 synthesis would affect the in vitro invasiveness and metastatic potential of these cells. Decreased MMP-1 protein expression in JJ-antisense and JJ-sense cells A 46 kDa protein band representing MMP-1 was detected by Western blotting in lyophilized samples of conditioned supernatant media and compared to a commercial MMP-I standard. The protein level of both JJantisense and JJ-sense cells was significantly lower than that from parental 55012 cells and those transfected with empty plasmid (Fig. 3, p < 0.05). MMP-1 activity decreased in JJ-antisense and JJ-sense cells

We quantitated the MMP-I activity using a fluorochrome-labeled Type I collagenase assay. MMP-I activity in JJ-empty cells was not significantly different from their parental cell line. The MMP-1 activity in both JJ-antisense cells and JJ-sense cells was significantly less

> 0

r

ti

I I

Fig. 4. MMP-1 enzymatic activity assay. The samples of conditioned media were activated with 20 mM APMA and incubated with substrate. The MMP-1 activity was evaluated by calculating the degradation of the substrate in samples from 53012, JJ-antisense. JJ-sense and JJ-empty cells, with reference to the MMP-I enzymatic activity standard curve. The JJ-sense cells, JJ-antisense cells, and JJ-empty cells are 53012 cells transfected respectively with MMP-I sense plasmid, MMP-1 antisense plasmid, or an empty plasmid. Shown are the means of three experiments, each performed in triplicate. The MMP-I enzymatic activity was significantly lower in JJ-antisense (80%1, p < 0.05) and JJ-sense cells (63.3% p < 0.05) compared to 55012 cells. The MMP-I enzymatic activity in JJ-empty cells was not significantly different than 53012 cells.

than that of their parental cells or cells transfected with empty plasmid (Fig. 4, p < 0.05). Cell migration ability decreased in JJ-antisense and JJsense cells We used the Boyden chamber as an in vitro index of cell invasiveness using a matrix barrier of Type I collagen. We observed no significant difference in cell

X . J i m g et (11. I Journal

of

Orrlropcrerlic Reseurch 21 (2003) 1063-1070

1067

Northern Blot

Cell Migration Assay

=JJ Cell

I 1 JJ Cell 2 JJ Empty 3 JJ Sense 4 JJ Antisense

JJ ErnDtv . _

T

0 JJ Sense UJJ Antisense

p<0.05

p
1

2

3

4

MMP-1

Fig. 5. Cell migration assay. In the cell migration assay, Boyden chambers were coated with Type I collagen, which served as an artificial matrix barrier and was designed to mimic to the matrix barrier present in vivo. Control Boyden chambers were coated with bovine serum albumin. The cell's invasive ability was evaluated by normalizing the optical absorbance of stained cells that had invaded through the artificial matrix barrier of collagen I against the optical absorbance of stained cells that had invaded through the chamber coated with bovine serum albumin in control piates. Shown are the means standard deviation of three experiments, each performed in triplicate. The JJ sense cells, JJ-antisense cells, and JJ-empty cells are 53012 cells transfected respectively with MMP-1 sense plasmid, MMP-I antisense plasmid, or an empty plasmid. The cell migration rate was significantly decreased in JJ-antisense cells (86'%1,p < 0.05) and JJ-sense cells (72'%1, p < 0.05) when compared to 53012 cells. There was no significant difference between detected 55012 and JJ-empty cells.

GAPDH Fig. 6. Northern blot. Northern blotting was performed on total RNA with a MMP-1 probe produced from the pOTB7 plasmid (ATCC # 6122647). MMP-1 mRNA expression was quantitated by densitometry and normalized to GAPDH mRNA expression. The data demonstrate a significant reduction in MMP-I tnRNA expression levels i n JJantisense (75'%1,p < 0.05) and JJ-sense cells (60%. p < 0.05) as conipared to 55012 cells alone. There was no significant change i n MMP-I mRNA expression between 35012 cells and JJ-empty cells.

*

significant change in MMP- 13 expression between JJantisense, JJ-sense, JJ-empty and 55012 cells (Fig. 7).

migration between the 33012 parental cells and the cells transfected with empty plasmid. It was observed that the cell migration of JJ-antisense and JJ-sense cells were each significantly lower than that of the parental JJ012 cells (Fig. 5 , p < 0.05).

South ern hlot Genomic DNA from JJ-antisense, JJ-sense, JJ-empty and 55012 cells was digested with Xho I/Eco RI and detected with an MMP-1 probe. In the Xho I/Eco RI group, the genomic DNA from JJ-antisense cells and JJsense cells showed a band correlating with the exogenous MMP-I gene fragment that was not present in JJ012 cells or those transfected with empty plasmid. This indicates that the exogenous MMP-1 fragment was being integrated into the cell chromosome. The Southern blot patterns from individual genomic DNA samples treated separately with Msp I, Hap I1 and Hpa 11 restriction endonucleases did not show significant changes

Northern blot f o r MMP-I and RT-PCR f o r MMP-I, MMP-8, MMP-I3 The MMP-1 mRNA level was evaluated using RTPCR and Northern blotting and subsequently normalized to GAPDH. The densities of MMP-1 bands were each significantly decreased in JJ-antisense or in JJ-sense cells compared to the parental cells or those transfected with empty plasmid (Fig. 6, p < 0.05). Semi-quantitative RT-PCR detected no MMP-8 mRNA, and there was no Semi-Quantitative RT-PCR MMP-8

1

2

MMP-1

MMP-13

3

4

1

2

3

4

1

2

3

4

Fig. 7. RT-PCR for MMP-8 and MMP-13. Semi-quantitative RT-PCR was used to amplify MMP-8 (338 bp), MMP-13 (198 bp), and MMP-I (306 bp). While MMP-I levels in JJ-sense and JJ-antisense decreased, the MMP-13 mRNA levels showed no significant change when compared to 35012 o r JJ-empty cells. Note the MMP-8 PCR product should be 338 bp, but it could not be amplified in 55012, J-empty, JJ-sense, or JJ-antisense cells. The
X . Jitmg ef nl. I Journtrl of Orthopuedic RC.wtvch 21 (2003) 1063-1070

I068

1. JJ Cell 2. JJ Empty 3. JJ Sense 4. JJ Antisense

Southern Blot Hap II

Hpa II

Msp I

Xho IlEco RI

4-5100

bp

4- 3500 bp

+Antisense Insert (770 bp) 4- Sense Insert (760 bp)

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

Fig. 8. Southern blot. In the Xho I/Eco RI digestion group, JJ-sense and JJ-antisense cells showed extra small bands corresponding to the inserts of their individual transfected plasmids. indicating the exogenous MMP-I DNA fragments had integrated themselves into cell chromosomes. Hap 11. Hpa 11. and Msp 1 restriction enzyme digests showed similar genomic DNA patterns in JJ-sense and JJ-antisense cells which indicated there was no difference in methylation modifications to explain the alteration in MMP-I expression. If the internal C of the CCGG restriction enzyme site was rnethylated, Msp I would be able to digest the DNA, while Hpa 11 would not.

indicating that methylation was not responsible for the changes in expression associated with the transfection of the sense oligonucleotide (Fig. 8).

Discussion

MMP-1 has been implicated in various advanced cancers and thought to mediate several mechanisms including tumor invasion of basement membranes and angiogenesis [8,9,16,24,29,32,33,36]. In previous studies, our lab has reported that elevated MMP-I levels correlate with an increased frequency of recurrence and metastasis, and poor survival rate in chondrosarcoma patients. It was hypothesized that MMP-1 served to facilitate cell egress from the tumor matrix [4,38,39]. We selected an antisense approach as a means of inhibiting MMP- 1 because of the molecular specificity that this technique affords. It was hypothesized that the 55012 cells would exhibit decreased invasion capacity as the MMP- 1 gene expression and collagenase activity were i 11 hi bi ted . I n this study, JJ012 cells transfected with empty plasmid demonstrated no significant difference in enzyme activity and invasion from the parental line, indicating that the plasmid without the target fragment did not affect endogenous MMP-1 expression and cell migration behavior. Transfection with MMP- 1 antisense plasmid or sense plasmid demonstrated a significant decrease in MMP-I protein and enzymatic activity ( p < 0.05), and also in invasion capabilities (p < 0.05). By Southern blot the exogenous MMP-1 fragment was demonstrated to have integrated itself into the chromosomes of JJ-antisense and JJ-sense cells. Semi-

quantitative RT-PCR demonstrated that the observed decrease of MMP- 1 enzymatic activity in JJ-antisense and JJ-sense cells was not associated with a significant change in MMP-8 or MMP-13 mRNA expression. RTPCR results showed no significant changes in MMP-8 or MMP-13 mRNA expression levels. This suggests invasive inhibition is specifically mediated through MMP-I expression [24,43]. The present findings support our hypothesis that MMP-1 facilitates cell egress from the tumor matrix, and suggest that MMP-1 may be a potential target of novel biologic therapy for chondrosarcoma. An intriguing observation that transfection with MMP-1 sense plasmid demonstrates inhibition, when an empty plasmid does not, suggests distinct molecular mechanisms underlying interference with endogenous MMP-I in JJ-sense cells. Based on reports from other laboratories, we postulate that in the JJ-antisense cells, the inhibition was due to the expression of the exogenous MMP-1 antisense fragment integrated in cell chromosomes, which bound to endogenous MMP-I mRNA via base-pairing leading to accelerated targeted mRNA degradation, incorrect pre-mature splicing, mRNA translation handicap, and mRNA traffic disorder [19,28,34]. In JJ-sense cells, we postulate that the transcripts from the exogenous MMP-1 sense fragment integrated in cell chromosomes could function as interfering RNA [15]. Msp I, Hap 11, and Hpa I1 restriction enzymes recognize the same sequence (CCGG), but differ in their sensitivities to methylation of C residues. Since all three restriction enzymes exhibited similar DNA digestion patterns on the Southern blot, it is unlikely that MMP- 1 transcription is being blocked by DNA methylation in JJ-sense cells. RNA interference

X.

Jiaii,o

i d I JouniuI of Orthopurdic Rewrrcli 21 (2003) 1063-1070

phenomena, the functional silencing of specific genes, can be induced in a wide variety of organisms either by introduction of homologous transgenes or by delivery of double-stranded RNA molecules homologous to the target locus, and have been observed in plant, fungi, invertebrates, vertebrates and human cells [6,12,13, 15,18,40]. These observations point to the existence of incompletely understood biological processes. In summary, we inhibited MMP-I utilizing an antisense RNA approach in order to test the hypothesis that this gene process facilitated tumor cell egress for chondrosarcoma tissue. Our results indicate that we can indeed change a cell's invasive behavior by decreasing MMP-1 activity. This evidence supports our hypothesis that MMP-1 functions to facilitate cell egress from chondrosarcoma tissue, an early step in the metastatic cascade. This also demonstrates the potential of MMP- 1 as a promising target for novel biologic therapy in chondrosarcoma. Further study is needed to completely understand the metastatic mechanisms involved as well as the molecular alterations associated with the introduction of MMP-I gene fragment expression plasmids.

Acknowledgements

We wish to acknowledge the support of the National Institutes of Health R29 (NIAMS R29AR 42863-6) and National Institutes of Health (NIAMS ROlAR48612-1).

References [ I ] Ahniad A. Hart IR. Mechanisms of metastasis. Crit Rev Oncol Hematol 1997;26:163-73. [2] Airola K, Karonen T. Vaalamo M. et al. Expression of collagenases-l and -3 and their inhibitors TIMP-I and -3 correlates with the level of invasion in malignant melanomas. Br J Cancer 1999; 80:73343. [3] Arii S. Mori A, Uchida S, et al. Implication of vascular endothelial growth factor in the development and metastasis of human cancers. Hum Cell 1999;12:25-30. [4] Berend KR. Toth AP, Harrelson JM. et al. Association between ratio of matrix metalloproteinase-l to tissue inhibitor of metalloproteinase- 1 and local recurrence, metastasis, and survival in human chondrosarcoma. J Bone Joint Surg Am 1998;XO:1 1-7. [ S ] Bohle AS, Kalthoff H. Molecular mechanisms of tumor metastasis and angiogenesis. Langenbecks Arch Surg 1999;384: 13340. [6] Catalanotto C, Azzalin G, Macino G, Cogoni C. Gene silencing in worms and fungi. Nature 2000;404:245. [7] Chambers AF. The metastatic process: basic research and clinical implications. Oncol Res 1999;11:161-8. [XI Chambers AF. Matrisian LM. Changing views of the role of matrix metalloproteinases in metastasis. J Natl Cancer lnst 1997; 89: 1260-70. [9] Curran S. Murray GI. Matrix metalloproteinases in tumour invasion and metastasis. J Pathol 1999;189:300-8. [lo] Curran S, Murray GI. Matrix metalloproteinases: molecular aspects of their roles in tumour invasion and metastasis. Eur J Cancer 200036: 1621-30.

I069

Denis LJ, Verweij J. Matrix metalloproteinase inhibitors: present achievements and future prospects. Invest New DrLigs 1997; 15: 175-85. Dernburg AF, Zalevsky J, Colaiacovo MP, Villeneuve AM. Transgene-mediated cosuppression i n the C. r k q c r r i s germ line. Genes Dev 2000; 14: 1578 83. Domeier ME. Morse DP, Knight SW. et al. A link between RNA interference and nonsense-mediated decay in C~rc~norlrrrhrlitis elc,,orm.s. Science 2000;289: 1928-3 1. Durko M, Navab R. Shibata HR, Brodt P. Suppression of basement membrane type IV collagen degradation and cell invasion in human melanoma cells expressing an antisense RNA for MMP-I. Biochim Biophys Acta 1997;I3S6:271 80. Elbashir SM, Harborth J, Lendeckel W. et al. Duplexes of 21nucleotide RNAs mediate RNA interference in cultured niamnialian cells. Nature 2001;41 1:49&8. Ellerbroek SM, Stack MS. Membrane associated matrix metalloproteinases in metastasis. Bioessays I9992 I :940 9. Eiigers R. Gabbert HE. Mechanisms of tumor metastasis: cell biological aspects and clinical implications. J Cancer Res Clin Oncol 2000;126:682 92. Fagard M. Boutet S, Morel JB. et al. AGOI, QDE-2. and RDE-I are related proteins required for post-transcriptional gene silencing in plants, quelling in fungi, and RNA interference i n animals. Proc Natl Acad Sci USA 2000;97:116504. Fang K. Chen MH. Transfection of anti-sense complementary DNA of human epidermal-growth-factor receptor attenuates the proliferation of human non-small-cell-lung-cancer cells. I n t J Cancer 199923I :47 1-8. Fidler IJ. Critical determinants of cancer metastasis: rationale for therapy. Cancer Chemother Pharniacol I999;43:S3- 10. Hua J. Muschel RJ. Inhibition of matrix metalloproteinase 9 expression by a ribozyme blocks metastasis in a rat sarcoma model system. Cancer Res 1996;56:5279-84. lnoue T. Yashiro M, Nishimura S. et al. Matrix metalloproteinase-l expression is a prognostic factor for patients with advanced gastric cancer. Int J Mol Med 1999;4:73-7. Kanamori Y, Matsushima M, Minaguchi T, et al. Correlation between expression of the matrix metalloproteinase-l gene i n ovarian cancers and an insertionldeletion polymorphism in its promoter region. Cancer Res 1999:59:4225 7. Kleiner DE, Stetler-Stevenson WG. Matrix metalloproteinases and metastasis. Cancer Chemother Pharmacol 1999;43:S4251. Kurschat P. Mauch C. Mechanisms of met Dermatol 2000;25:482-9. Laemnili UK. Cleavage of structural proteins during the assembly of the head of bacteriophage 14. Nature 1970;227:680 5. Laws HJ. Burdach S. van Kaick B, et al. Multimodality diagnostics and megatherapy in poor prognosis Ewing's tumor patients. A single-center report. Strahlenther Onkol 1999;175: 488-94. Marcusson EG, Yacyshyn BR. Shanahan Jr WR, Dean NM. Preclinical and clinical pharmacology of antisense oligonucleotides. Mol Biotechnol 1999:12:1--1I . McCawley LJ, Matrisian LM. Matrix metalloproteinases: multifunctional contributors to tumor progression. Mol Med Today 2000;6: 149-56. McMahon G. VEGF receptor signaling in tumor angiogenesis. Oncologist 2000;5:3-10. Meyer W, Malawer M. Osteosarcoma: clinical features and evolving surgical and chemotherapeutic strategies. Pediatr Clin North Am 1991:38:317-48. Murray GI. Duncan ME, O'Neil P, et al. Matrix metalloproteinase-l is associated with poor prognosis in colorectal cancer. Nat Med 1996;2:461--2.

1070

X Jiang et a[. I JournuI of Orthopurdic Riwurcli 21 (2003) 1063-1070

(331 Murray GI, Duncan ME, ONeil P, et al. Matrix metalloproteinase-I is associated with poor prognosis in oesophageal cancer. J Pathol 1998;185:256-61. [34] Pawlak W. Zolnierek J , Sarosiek T, Szczylik C. Antisense therapy in cancer. Cancer Treat Rev 2000;26:333-50. [35] Saeter G. Wiebe T, Wiklund T, et al. Chemotherapy in osteosarcoma-the Scandinavian sarcoma group experience. Acta Orthop Scand 1999;70:7482. [36] Sato H. Seiki M. Membrane-type matrix metalloproteinases (MT-MMPS) in tumor metastasis. J Biochem (Tokyo) 1996;119: 209-15. [37] Schwartz MA, Venkatardman S, Ghaffari MA, et al. Inhibition of' human collagenases by sulfur-based substrate analogs. Biochem Biophys Res Commun 1991;176:173-9. [38] Scully SP, Berend K R , Qi WN, Harrelson JM. Collagenase specificity in chondrosarcoma metastasis. Braz J Med Biol Res 1999;32:885-9.

[39] Scully SP, Berend KR, Toth A, et al. Marshall Urist Award. Interstitial collagenase gene expression correlates with in vitro invasion in human chondrosarcoma. Clin Orthop 2000:29 1- 303. [40] Sharp PA. R N A interference--2001. Genes Dev 2001.15:485-90. [41] Towbin H, Staehelin T. Gordon J. Electrophoretic transfer of proteins from polyacrylainide gels t o nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 1979; 76:43504. [42] Uchida A, Myoui A, Araki N, et al. Neoadjuvant chemotherapy for pediatric osteosarcoma patients. Cancer 1997;79:411-5. [43] Ueda Y, Oda Y, Tsuchiya H, et al. Immunohistological study on collagenous proteins of benign and malignant human cartilaginous tumours of bone. Virchows Arch A Pathol Anat Histopathol 1990;4 17:291-7. [44] Ziober BL, Chen YQ, Ramos DM, et al. Expression of the alpha7betal laminin receptor suppresses melanoma growth and metastatic potential. Cell Growth Differ 1999; I0:479-90.