Increased matrix metalloproteinase-1 (MMP-1) expression is associated with advanced stages of breast cancer and may be a predictive marker for the development of invasive disease. In this report, we used short hairpin RNA (shRNA) molecules to investigate whether MMP-1 production in MDA-231 breast cancer cells contributed to the degradation of a collagen matrix or tumor formation in nude mice. We created two groups of MDA-231 cell lines. MDA-231 cells containing a vector producing shRNA specific for MMP-1 had a >90% decrease in MMP-1 mRNA and protein compared with cells containing an empty vector, and blocking MMP-1 expression inhibited the in vitro collagenolytic activity of MDA-231 cells. When the cells were injected into the mammary fat pad, there was no difference in the frequency of tumor formation in mice. However, the average tumor size was larger in mice injected with cells containing the empty vector (1,216 ± 334 mm3) than in mice injected with cells expressing the MMP-1 shRNA (272 ± 117 mm3; P = 0.027). We conclude that MMP-1 expression is essential for the ability of MDA-231 cells to invade and destroy a collagen matrix and in vivo experiments suggest an important role for MMP-1 in breast tumor growth.

Proteins comprising the extracellular matrix (ECM) provide the biological scaffold for support and protection of all cells in the body, act as a pathway for cellular communication, facilitate cell migration, help to regulate cell growth and division, and serve as a barrier to microbial infection (1, 2). Whereas the structural integrity of the ECM gives form and protection, it is the ability of the ECM to be remodeled that is the source of its versatility. ECM remodeling occurs in normal physiology and in many pathologic processes, including arthritis, periodontal disease, and tumor growth and metastasis (3, 4). Because the structural and functional roles of the ECM are so important, its remodeling is a tightly regulated balance between the production of matrix proteins and their degradation.

Type I collagen is a major component of the ECM and it is degraded by enzymes called matrix metalloproteinases (MMP). MMPs with enzyme activity directed against type I collagen include the interstitial collagenases MMP-1 and MMP-13 and the membrane-bound MMP-14 (MT1‐MMP; ref. 1). In normal physiology, basal expression of most MMPs, including the interstitial collagenases, is low; however, high expression has been linked to several pathologies, including cancer cell invasiveness. In clinical studies, increased MMP-1 expression has been associated with the incidence or invasiveness of several cancers: colorectal, esophageal, pancreatic, gastric, breast, and malignant melanoma (510). Furthermore, elevated MMP-1 expression in atypical ductal hyperplastic tissues may serve as a marker for predicting which patients will develop invasive breast cancer (11).

To begin to understand the contributions of MMP-1 to tumor growth and progression, MMP-1 has been overexpressed in various experimental systems. In a differentiated epithelial cell line (Madin-Darby canine kidney cells), overexpressing proMMP-1 did not confer an invasive phenotype in an in vitro peritoneal invasion model or in a three-dimensional collagen matrix, perhaps because these cells did not express the proteinases needed to fully activate the latent enzyme (3, 1214). However, when MMP-1 was overexpressed in the skin of transgenic mice, it increased their susceptibility to skin carcinogenesis (15).

An alternative to overexpression has been blocking MMP-1 expression. Durko et al. (16) developed a 777-nucleotide antisense RNA molecule capable of specifically silencing MMP-1 expression and showed that blocking this expression in melanoma cells blocked in vitro invasion of a collagen matrix. These findings were extended to chondrosarcoma cells using the same antisense RNA plasmid, and more recently, the experiments with chondrosarcoma cells were repeated using transiently transfected short interfering RNA (siRNA) molecules (17, 18). Both approaches resulted in a substantial reduction in MMP-1 protein levels, which was accompanied by a significant decrease in the ability of the cells to invade through a type I collagen gel.

A plasmid-based system that stably expresses short hairpin RNA (shRNA) molecules, which silence gene expression by RNA interference (1921), has been developed, thereby providing a mechanism for long-term gene silencing. In this study, we developed a shRNA sequence which, when stably integrated into cellular DNA, selectively targets MMP-1 expression. In MDA-231 breast cancer cells that have constitutively high levels of MMP-1, this shRNA molecule specifically blocks MMP-1 expression and MMP-1-mediated destruction of collagen in an in vitro assay and the MMP-1 shRNA molecule significantly inhibits the growth of breast cancer cells in nude mice.

Cell culture. NIH 3T3 cells from American Type Culture Collection (ATCC, Manassas, VA) were cultured in DMEM containing 10% FCS, penicillin (100 units/mL)/streptomycin (100 μg/mL), and 2 mmol/L glutamine. MDA-231 cells (ATCC) were cultured in DMEM/F12 50:50 medium with 10% FCS, penicillin (100 units/mL)/streptomycin (100 μg/mL), and 2 mmol/L glutamine. For experiments, MDA-231 cells were washed thrice in HBSS to remove traces of serum and placed in serum-free medium.

Matrix metalloproteinase-1 and enhanced green fluorescent protein expression plasmids. The pCMV-MMP-1 expression construct was prepared with NotI and SalI restriction enzymes (Invitrogen, Carlsbad, CA), used to excise the complete MMP-1 cDNA from pSP64-MMP-1 (ATCC) and to digest the pCMV-Tag4c vector (Stratagene, La Jolla, CA). The purified digestion products and the MMP-1 cDNA fragment were ligated into the pCMV-Tag 4c vector to create the pCMV-MMP-1 vector. The enhanced green fluorescent protein expression vector (peGFP) is commercially available from BD Biosciences (Franklin Lakes, NJ).

Short hairpin RNA expression plasmids. Circular pSuper vectors (pSuper and pSUPER.Retro.Neo.GFP, Oligoengine) were linearized with BglII and HindIII restriction enzymes (Invitrogen) and dephosphorylated with calf intestinal phosphatase (Life Technologies, Inc., Gaithersburg, MD). DNA oligonucleotides (see below) specifically designed for use with the pSuper plasmids and containing either MMP-1 shRNA sequence or scrambled shRNA sequence were annealed to create double-stranded oligos to be cloned into the pSuper vectors. Annealed oligonucleotides were phosphorylated with T4 polynucleotide kinase (Promega, Madison, WI), ligated to the pSUPER plasmids, and then transformed into bacteria. The newly created pSuper-MMP-1 shRNA, pSuper-retro-MMP-1 shRNA, and pSuper-retro-scrambled shRNA vectors were prepared from individual bacterial colonies. Correct orientation and location of the oligonucleotide cloning were confirmed by sequencing the plasmids with T3 and T7 primers (T3, 5′-AATTAACCCTCACTAAAGGG-3′; T7, 5′-TAATACGACTCACTATAGGG-3′) and the Big Dye sequencing reagent (Applied Biosystems, Foster City, CA). MMP-1 shRNA oligonucleotides (MMP-1 sense, 5′-GATCCCCACCAGATGCTGAAACCCTGTTCAAGAGACAGGGTTTCAGCATCTGGTTTTTTGGAAA-3′; MMP-1 antisense, 5′-AGCTTTTCCAAAAAACCAGATGCTGAAACCCTGTCTCTTGAACAGGGTTTCAGCATCTGGTGGG-3′; patent pending) contained a region specific to bases 234 to 252 of MMP-1 mRNA (bold), a hairpin loop region (italic), and 5′ and 3′ linker sequences for subcloning into the BglII and HindIII sites of the pSUPER vectors. The scrambled shRNA oligonucleotides contained identical hairpin loop and linker sequence but contained a sequence of DNA that was not complementary to any human gene (AAGTGGAGGGACGTATGCA).

Transfection of NIH 3T3 and MDA-231 cells. NIH 3T3 cells were transiently transfected with Lipofectamine 2000 (Invitrogen) along with 0.5 μg peGFP, 0.5 μg pCMV-MMP-1, and 2 μg pSUPER-MMP-1 shRNA or pSuper with no insert. Transfection efficiency was monitored using eGFP signal. Total cellular RNA was harvested 48 hours after transfection and analyzed for MMP-1 mRNA by real-time reverse transcription-PCR (RT-PCR; see below). MDA-231 cells were stably transfected in triplicate with Lipofectamine 2000 (Invitrogen) and 2 μg of the pSUPER.Retro.Neo.GFP plasmid containing the MMP-1 shRNA, the scrambled sequence, or the empty vector. Stable cell lines were selected by growth in the presence of 1 mg/mL G418 (Stratagene) and individual cell lines were isolated with cloning discs (PGC Scientific, Frederick, MD).

RNA isolation and real-time reverse transcription-PCR. MDA-231 stable cell lines were cultured in serum-free medium for 24 hours, RNA was harvested using the RNeasy RNA isolation kit (Qiagen, Valencia, CA), and DNA contamination was removed with RNase-free DNase (Qiagen). For real-time RT-PCR, reverse transcription was done using protocols and reagents from Applied Biosystems TaqMan reverse transcription reagent kit. Briefly, 2 μg of DNase treated RNA were reverse transcribed in a 20-μL reaction containing 5.5 mmol/L MgCl2, 500 μmol/L each deoxynucleotide triphosphate (dNTP), 2.5 μmol/L oligo d(T)16, 0.4 units/μL RNase inhibitor, and 1.25 units/μL Multiscribe reverse transcriptase. The reactions were incubated at 25°C for 5 minutes, 48° for 30 minutes, then 95°C for 5 minutes.

Real-time PCR was done with reagents and protocols from the Applied Biosystems SYBR Green master mix kit. Five microliters of each reverse transcription reaction were used to amplify MMP cDNA in triplicate real-time PCR reactions and 2 μL of each reverse transcription reaction were used to amplify glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or eGFP cDNA in duplicate reactions. For quantitative comparisons between PCR assays, standard curves were generated with every assay (see standards outlined below). The PCRs contained 200 nmol/L of each primer and were incubated on an M.J. Research Opticon (Waltham, MA) real-time detection thermal cycler at 95°C for 10 minutes, followed by 50 PCR cycles of 95° for 15 seconds, and 60°C for 1 minute, and plate read. The PCR cycles were followed by a SYBR Green melting curve from 55°C to 90°C. MMP mRNA levels are reported as copies of the target gene per picogram of GAPDH or eGFP.

Real-time PCR primer sequences. The MMP-2 and MMP-9 primer sequences were generous gifts from Dr. Ian Clark (University of East Anglia, Norwich, United Kingdom). The remaining primers were designed using Oligo primer analysis software (Molecular Biology Insights, Cascade, CO). Forward and reverse real-time PCR primers for MMP-1, MMP-2, MMP-9, MMP-14 (MT1-MMP), GAPDH, and peGFP were MMP-1 forward 5′-AGCTAGCTCAGGATGACATTGATG-3′, MMP-1 reverse 5′-GCCGATGGGCTGGACAG-3′; MMP-2 forward 5′-TGGCGATGGATACCCCTTT-3′, MMP-2 reverse 5′-TTCTCCCAAGGTCCATAGCTCAT-3′; MMP-9 forward 5′-CCTGGGCAGATTCCAAACCT-3′, MMP-9 reverse 5′-GCAAGTCTTCCGAGTAGTTTTGGAT-3′; MMP-13 forward 5′-TGGCATTGCTGACATCATGA-3′, MMP-13 reverse 5′-GCCAGAGGGCCCATCAA-3′; MMP-14 forward 5′-CCCCGAAGCCTGGCTACA-3′, MMP-14 reverse 5′-GCATCAGCTTTGCCTGTTACT-3′; GAPDH forward 5′-CGACAGTCAGCCGCATCTT-3′, GAPDH reverse 5′-CCCCATGGTGTCTGAGCG-3′; and eGFP forward 5′-TATCATGGCCGACAAGCAGAAGAAC-3′, eGFP reverse 5′-TTTGCTCAGGGCGGACTGGGTGCTC-3′.

Real-time PCR standards. The real-time standards for MMP-1, MMP-2, MMP-9, GAPDH, and peGFP were plasmids containing either a portion or all of the cDNA of the target gene. The standard for MMP-14 was a PCR fragment generated with the real-time primers. Plasmid dilutions were serial log dilutions from 1 ng to 10 fg. The MMP-14 standards were diluted from 10 pg to 1 fg. Numbers for converting picograms to copies of mRNA were 360,000 copies/pg MMP-1, 480,000 copies/pg MMP-2, 340,000 copies/pg MMP-9, 270,000 copies/pg MMP-13, and 10 × 106 copies/pg MMP-14 (MT1-MMP).

IFN response reverse transcription-PCR assay. Three micrograms of RNA pooled from MDA-231 cells containing the empty vector or the MMP-1 shRNA were reverse transcribed with Moloney murine leukemia virus (MMLV) reverse transcriptase (Invitrogen) according to the instructions of the manufacturer. Briefly, 3 μg of RNA, 99 ng of random hexamers, and 10 nmol of dNTP were combined to a total volume of 12 μL. The mixture was then incubated at 65°C for 5 minutes and cooled to 4°C. Four microliters of 5× reverse strand buffer, 2 μL of 0.1 mol/L DTT, and 1 μL of RNase inhibitor A (40 units/μL) were added to the reaction and incubated at 37°C for 2 minutes. One microliter of MMLV reverse transcriptase was added to the reaction, which was incubated at 25°C for 10 minutes, 37°C for 50 minutes, and 70°C for 15 minutes. Expression of 11 IFN-responsive genes and GAPDH was estimated using semiquantitative PCR and the human IFN response MultiGene-12 RT-PCR Profiling Kit (#PH-009B) from Superarray (Frederick, MD).

ELISA for matrix metalloproteinase-1 protein. MMP-1 protein in serum-free medium of stably transfected MDA-231 cells was quantified with the human MMP-1 biotrak ELISA system for latent and active enzyme (Amersham, Piscataway, NJ). Cells were grown in six-well culture dishes with 1 mL of serum-free medium for 24 hours and 100 μL of a 1:30 dilution of the medium were used for the ELISA. Cells were lysed with 350 μL of passive lysis buffer (Promega) and total protein was determined with 70 μL in a Bradford assay (Bio-Rad, Hercules, CA). Nanograms of MMP-1 protein from the ELISA were normalized to micrograms of total protein.

Western blotting. MDA-231 cells were cultured in 1 mL serum-free DMEM/F12 medium in six-well plates in the presence or absence of a collagen gel, with or without aprotinin (50 units/mL; Sigma, St. Louis, MO). After 24 hours, medium was precipitated with cold 10% trichloroacetic acid and resuspended in 50 μL SDS loading buffer (Promega). Following SDS-PAGE and transfer to a polyvinylidene difluoride membrane (Immobilon-P, Millipore, Billerica, MA), the membrane was incubated in a 5% milk, TBS solution with 0.1% Tween 20 (milk TBST) for 1 hour. The membrane was probed with a MMP-1 polyclonal antibody (Chemicon, Temecula, CA) diluted 1:5,000 in milk TBST overnight at 4°C, washed thrice with TBST, and a 1:2,000 dilution of goat anti-rabbit horseradish peroxidase (HRP)–conjugated antibody (Cell Signaling, Beverly, MA) in milk TBST was added for 60 minutes. The blot was washed thrice with TBST and HRP activity was detected with the Western Lightning chemiluminescence reagent (Perkin-Elmer, Boston, MA).

Cell proliferation assay. The proliferation rate of the stably transfected MDA-231 cell lines at 24 and 48 hours was determined with a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (ATCC). The assay was optimized to be sure signal obtained was within the linear range for the reagent. For each cell line, the optimum cell count was 10,000 cells per well of a 96-well tissue culture dish.

In vitro collagen gel destruction assay. Collagen preparations were carried out on ice to prevent premature jelling. A solution of purified type I bovine collagen (Cohesion Technologies, Palo Alto, CA) was neutralized with a sterile 10× PBS (pH 7.4) and 0.1 mol/L NaOH. Neutralized collagen (2 mg/mL) was mixed with an equivalent volume of cells suspended in serum-free DMEM to yield a final concentration of 2 × 105 cells/mL in 1 mg/mL fibrillar collagen, and 1 mL of the collagen/cell mixture was added to each well of a six-well plate. After the collagen jelled (∼1 hour), 1 mL of serum-free medium was added on top of the gel. For some experiments, 25 μmol/L Ilomastat (Chemicon), 50 units/mL aprotinin (Sigma), 1 μg/mL neutralizing MMP-1 antibody (MMP-1 Ab-5, Oncogene, Cambridge, MA), 1 μg/mL monoclonal FLAG antibody (Sigma), or 1% DMSO was added to the collagen before it jelled (1 μg antibody/mL collagen). After 36 hours at 37°C, the medium was removed and weighed. The weight of medium added at the start of the 37°C incubation (1 g) was subtracted from the total medium weight to reveal the amount liberated by collagen destruction.

Tumorigenesis studies in mice. Pooled MDA-231 stable cell lines containing the empty vector or the MMP-1 shRNA (1 × 106 cells/100 μL HBSS) were injected into the fourth inguinal fat pad of 51 female, 6-week-old nude mice (nu/nu, Charles River, Wilmington, MA) using 28 gauge × 1/2 in. single-use insulin needles. Mice were examined weekly until tumors were grossly apparent; then they were examined several times a week. Tumors were measured twice with Vernier calipers and tumor volume was calculated using the formula (4/3)πr3. When the two measurements differed, the smaller radius measurement was squared and multiplied by the largest radius measurement. This number was then substituted for the r3 portion of the formula.

Statistical methods. Student's t test was used for calculating the statistical significance of measurements of MMP levels, collagen destruction, cell proliferation, and tumor size. Fisher's exact test was used to determine whether there was a correlation between tumor incidence and the presence of the MMP-1 shRNA.

Matrix metalloproteinase-1 short hairpin RNA design. The shRNA sequence was homologous only to MMP-1 mRNA as confirmed with National Center for Biotechnology Information (NCBI) Blast search engine. Because RNA interference is thought to be a cytoplasmic process, the shRNA was complementary to an exonic sequence of the mRNA (22, 23). Additional design considerations made the shRNA sequence compatible with the pSUPER plasmid-based systems of shRNA expression. These plasmids contain an H1 RNA polymerase III promoter that requires an adenine at the +1 nucleotide position. Furthermore, four or more thymine residues in a row are the termination sequence for the H1 RNA polymerase, and thus tangent thymines were avoided (20).

Matrix metalloproteinase-1 short hairpin RNA efficacy: blocking ectopic gene expression by transient transfections. To test the efficacy of the selected shRNA sequence, we used two pSUPER plasmids: one with the MMP-1 shRNA sequence to block MMP-1 gene expression and one with no shRNA insert. NIH 3T3 mouse embryonic fibroblasts were used because they do not contain a homologue to MMP-1 and because they are easily transfectable. MMP-1 mRNA production was supplied to the NIH 3T3 cells by a plasmid producing the mRNA from a CMV promoter (pCMV-MMP-1-tag) and transfection efficiency was tracked by cotransfecting with peGFP. The eGFP mRNA also served as a control gene to normalize the MMP-1 mRNA levels. Forty-eight hours after transfection, MMP-1 mRNA levels of cells transfected with the MMP-1 shRNA vector were 73% lower than the MMP-1 levels of cells transfected with the empty vector (P = 0.015; Fig. 1). These findings indicated that the MMP-1 shRNA reduced MMP-1 mRNA levels; consequently, the shRNA oligonucleotides were cloned into the stably integrating pSUPER-retro plasmid to create MDA-231 cell lines.

Figure 1.

Efficacy of MMP-1 shRNA oligonucleotide. NIH 3T3 murine fibroblasts were cotransfected in triplicate with 500 ng pCMV MMP-1, 500 ng peGFP, and 2 μg pSUPER plasmid containing either the MMP-1 shRNA insert or no insert. Cellular RNA was harvested 48 hours after transfection and MMP-1 and eGFP mRNAs were measured using quantitative RT-PCR. Copies of MMP-1 mRNA were normalized to picograms of eGFP mRNA. Columns, mean; bars, SD. Student's t test was used for comparison of MMP-1 mRNA levels between the two groups of cells (P = 0.015).

Figure 1.

Efficacy of MMP-1 shRNA oligonucleotide. NIH 3T3 murine fibroblasts were cotransfected in triplicate with 500 ng pCMV MMP-1, 500 ng peGFP, and 2 μg pSUPER plasmid containing either the MMP-1 shRNA insert or no insert. Cellular RNA was harvested 48 hours after transfection and MMP-1 and eGFP mRNAs were measured using quantitative RT-PCR. Copies of MMP-1 mRNA were normalized to picograms of eGFP mRNA. Columns, mean; bars, SD. Student's t test was used for comparison of MMP-1 mRNA levels between the two groups of cells (P = 0.015).

Close modal

MDA-231 stable cell lines: blocking endogenous gene expression. We created three groups of stable MDA-231 cell lines: one contained the pSUPER-retro plasmid, a second contained a pSUPER-retro plasmid with the scrambled shRNA sequence, and the third contained a pSUPER-retro plasmid with the MMP-1 shRNA oligonucleotide. Cells with the scrambled shRNA did not grow well and did not survive subcloning (see Discussion). Analysis of the scrambled shRNA sequence using NCBI BLAST search did not reveal any homology to mammalian genes (data not shown).

After neomycin selection and isolation of individual stable cell lines, MMP-1 mRNA and protein levels were determined using quantitative RT-PCR and ELISA. The ELISA measured total (active and zymogen) MMP-1 protein. Three cell lines producing MMP-1 shRNA had a >90% reduction in MMP-1 mRNA and protein levels compared with the cell lines containing the empty vector (Table 1).

Table 1.

MMP-1 mRNA and protein levels in stably transfected MDA-231 cell lines

pSUPER-retro insertMDA-231 cell lineCopies of MMP-1 mRNA/pg of GAPDH mRNAng of MMP-1 protein/μg of total protein
Empty vector A1 380,000 ± 210,000 19.10 ± 6.67 
 A11 550,000 ± 320,000 30.70 ± 6.94 
 A12 680,000 ± 230,000 24.63 ± 7.26 
 Group 540,000 ± 260,000 24.81 ± 5.81 
MMP-1 shRNA D1 8,900 ± 3,600 0.49 ± 0.03 
 Y2 4,600 ± 2,500 0.47 ± 0.01 
 Y7 5,000 ± 2,400 0.42 ± 0.04 
 Group 6,200 ± 3,200 0.46 ± 0.04 
 P 0.004 0.002 
pSUPER-retro insertMDA-231 cell lineCopies of MMP-1 mRNA/pg of GAPDH mRNAng of MMP-1 protein/μg of total protein
Empty vector A1 380,000 ± 210,000 19.10 ± 6.67 
 A11 550,000 ± 320,000 30.70 ± 6.94 
 A12 680,000 ± 230,000 24.63 ± 7.26 
 Group 540,000 ± 260,000 24.81 ± 5.81 
MMP-1 shRNA D1 8,900 ± 3,600 0.49 ± 0.03 
 Y2 4,600 ± 2,500 0.47 ± 0.01 
 Y7 5,000 ± 2,400 0.42 ± 0.04 
 Group 6,200 ± 3,200 0.46 ± 0.04 
 P 0.004 0.002 

NOTE: MMP-1 mRNA levels were measured in RNA harvested from stably transfected cell lines after a 24-hour incubation in serum-free medium and were normalized to GAPDH mRNA levels. MMP-1 protein levels were measured in the serum-free media of the cell lines after 24 hours of incubation and were normalized to total protein content of the cells. Data for the individual cell lines represent nine measurements; mRNA levels were determined in triplicate on three separate occasions. Data are presented as average ± SD. Statistical comparisons were made with Student's t test.

We verified the specificity of the MMP-1 shRNA by measuring the mRNA levels of the gelatinases (MMP-2 and MMP-9), the interstitial collagenase MMP-13, and MT1‐MMP. The MMP expression data for each cell line were averaged to obtain mean MMP expression levels for each group of cells. These MMPs were chosen because they are widely studied in cancer research and because MMP-13 and MT1‐MMP have similar substrate specificity as MMP-1 (3, 24). We found no significant difference in MMP mRNA levels between cells harboring the empty vector and those with the MMP-1 shRNA (Table 2). One cell line harboring the empty vector had elevated MMP-9 expression levels, thus affecting the average of the group. However, this clone was not removed from the pool because the relative levels of MMP-9 expression were low compared with MMP-1 expression (Tables 1 and 2). Additionally, including the third cell line provides a further control for variations in the expression of unmeasured genes.

Table 2.

MMP mRNA levels in stably transfected MDA-231 cell lines

pSUPER-retro insertCopies of MMP mRNA/pg of GAPDH mRNA
MMP-2MMP-9MMP-13MT1-MMP
Empty vector 11.2 ± 7.3 1,000.4 ± 645.4 1.5 ± 1.4 145,000 ± 46,638 
MMP-1 shRNA 8.8 ± 7.8 197.3 ± 106 0.5 ± 0.2 107,000 ± 46,000 
P 0.3 0.1 0.3 0.4 
pSUPER-retro insertCopies of MMP mRNA/pg of GAPDH mRNA
MMP-2MMP-9MMP-13MT1-MMP
Empty vector 11.2 ± 7.3 1,000.4 ± 645.4 1.5 ± 1.4 145,000 ± 46,638 
MMP-1 shRNA 8.8 ± 7.8 197.3 ± 106 0.5 ± 0.2 107,000 ± 46,000 
P 0.3 0.1 0.3 0.4 

NOTE: MMP mRNA levels were measured in RNA harvested from stable cell lines after a 24-hour incubation in serum-free medium and were normalized to GAPDH mRNA levels. mRNA levels were determined in triplicate on three separate occasions. Data are presented as average ± SD. Statistical comparisons were made with Student's t test.

To be certain that the MMP-1 shRNA did not affect cell growth, we measured the increase in metabolic activity of the cells from 24 to 48 hours as an estimation of their proliferation rate. The fold increase in metabolic activity of MDA-231 cells with the empty vector (1.63 ± 0.359 in 24 hours) did not differ significantly from that of the cells producing MMP-1 shRNA molecules (2.03 ± 0.06 in 24 hours; Table 3). Therefore, the only measured differences between the two groups of cell lines were the levels of MMP-1 mRNA and protein.

Table 3.

Metabolic activity of stably transfected MDA-231 cell lines

pSUPER-retro insertFold increase in metabolic activity
Empty vector 1.64 ± 0.36 
MMP-1 shRNA 2.03 ± 0.07 
P 0.13 
pSUPER-retro insertFold increase in metabolic activity
Empty vector 1.64 ± 0.36 
MMP-1 shRNA 2.03 ± 0.07 
P 0.13 

NOTE: The metabolic activity of 10,000 cells plated in a 96-well plate was determined after 24 and 48 hours. The proliferation rates of the two groups of cell lines were estimated by calculating the fold increase in metabolic activity from 24 to 48 hours. Data are presented as average ± SD. Statistical comparisons were made with Student's paired t test.

Finally, to determine whether the MMP-1 shRNA was producing a shRNA-associated IFN response, we compared the expression of 11 IFN-responsive genes between pools of stable cell lines (25). Semiquantitative RT-PCR revealed two genes (MxA and GBP1) with a slight (∼50%) increase in expression in the MMP-1 shRNA cell lines when compared with the empty vector cell lines and three genes (SCYB10/IP-10, MxB, and OAS1) with a modest (∼35-50%) decrease in expression (Fig. 2). Overall, there is no consistent up-regulation of IFN-responsive genes to suggest that the MMP-1 shRNA induced an IFN response.

Figure 2.

IFN-responsive genes. RNA from individual cell lines was pooled into two groups: MMP-1 shRNA–producing cells and empty vector cells. RT-PCR was done on GAPDH and 11 IFN-responsive genes. The PCR products were then run on an agarose gel and the bands were quantified with gel imaging software. There was no consistent increase in the expression of IFN-responsive genes in either group of cells.

Figure 2.

IFN-responsive genes. RNA from individual cell lines was pooled into two groups: MMP-1 shRNA–producing cells and empty vector cells. RT-PCR was done on GAPDH and 11 IFN-responsive genes. The PCR products were then run on an agarose gel and the bands were quantified with gel imaging software. There was no consistent increase in the expression of IFN-responsive genes in either group of cells.

Close modal

In vitro collagen destruction assay. Next we developed an in vitro collagen destruction assay to determine whether the MMP-1 levels in MDA-231 cells with the empty vector could degrade a collagen matrix and whether the shRNA-mediated reduction in MMP-1 was sufficient to change the invasive behavior of the cells. MDA-231 cells harboring the empty vector or the MMP-1 shRNA were suspended in a solution of type 1 collagen (1 mL). The collagen was allowed to solidify at 37°C and 1 mL of serum-free medium was layered on top of each gel. After 36 hours at 37°C, the medium was removed from each well and photographs were taken (Fig. 3A). No collagen was visible in wells containing MDA-231 cells with the empty vector. However, the collagen in wells containing cells producing the MMP-1 shRNA appeared intact. The medium from each well was weighed to quantitate the amount liberated as the gel was destroyed. After 36 hours, ∼1 mL (0.8 ± 0.06 g) of medium was liberated from the gels containing the MDA-231 cells with the empty vector (Fig. 3B). This volume was comparable to the volume of the gel at the beginning of the experiments (1 mL), indicating that these cells had completely degraded the collagen matrix. Conversely, no medium (−0.01 ± 0.03 g) was liberated from collagen containing MDA-231 cells producing the MMP-1 shRNA. The slight decrease in the volume of medium recovered may be due to evaporation during the incubation period. These results show that MMP-1 produced by cells containing the empty vector can destroy a collagen gel and that shRNA-mediated reduction of MMP-1 expression blocked the collagenolytic activity of the cells.

Figure 3.

In vitro collagen destruction assay. MDA-231 cells (2 × 105) containing either the empty pSUPER-retro vector or MMP-1 shRNA were embedded in 1 mL of collagen gels and overlaid with 1 mL of serum-free medium. A, after 36 hours of incubation, the medium was removed from the top of the gels and the gels were photographed. B, the medium removed from the top of each gel was weighed to quantitate the extent of collagen gel destruction. MDA-231 cells containing the empty vector were also embedded with several treatments: Ilomastat, DMSO vehicle control, aprotinin, MMP-1 neutralizing antibody, or anti-FLAG mouse monoclonal antibody. *, P < 0.005, in liberated medium. C, Western blot of MMP-1 protein in the culture medium of MDA-231 stable cell lines. Lane 1, MMP-1 protein detected when cells containing the MMP-1 shRNA were grown on plastic. Lane 2, MMP-1 protein detected when cell lines containing the empty vector were grown on plastic. Lane 3, MMP-1 protein when cells with the empty vector were grown embedded in collagen. Lane 4, MMP-protein when cells with the empty vector were grown in collagen with 50 units/mL aprotinin. Upper bands, MMP-1 proenzyme (∼54 kDa); lower bands, shorter, activated forms of the enzyme (∼24 kDa). Cleavage of the propeptide domain typically changes the 54-kDa proenzyme to a 44-kDa active form; however, smaller 24-kDa MMP-1 Western blot bands have previously been reported (26).

Figure 3.

In vitro collagen destruction assay. MDA-231 cells (2 × 105) containing either the empty pSUPER-retro vector or MMP-1 shRNA were embedded in 1 mL of collagen gels and overlaid with 1 mL of serum-free medium. A, after 36 hours of incubation, the medium was removed from the top of the gels and the gels were photographed. B, the medium removed from the top of each gel was weighed to quantitate the extent of collagen gel destruction. MDA-231 cells containing the empty vector were also embedded with several treatments: Ilomastat, DMSO vehicle control, aprotinin, MMP-1 neutralizing antibody, or anti-FLAG mouse monoclonal antibody. *, P < 0.005, in liberated medium. C, Western blot of MMP-1 protein in the culture medium of MDA-231 stable cell lines. Lane 1, MMP-1 protein detected when cells containing the MMP-1 shRNA were grown on plastic. Lane 2, MMP-1 protein detected when cell lines containing the empty vector were grown on plastic. Lane 3, MMP-1 protein when cells with the empty vector were grown embedded in collagen. Lane 4, MMP-protein when cells with the empty vector were grown in collagen with 50 units/mL aprotinin. Upper bands, MMP-1 proenzyme (∼54 kDa); lower bands, shorter, activated forms of the enzyme (∼24 kDa). Cleavage of the propeptide domain typically changes the 54-kDa proenzyme to a 44-kDa active form; however, smaller 24-kDa MMP-1 Western blot bands have previously been reported (26).

Close modal

To confirm that MMP expression was responsible for destruction of the collagen, we embedded the pan-MMP chemical inhibitor Ilomastat (25 μmol/L per gel) or vehicle control (1% DMSO) in the collagen together with cells containing the empty vector. After 36 hours, no medium (−0.1 ± 0.08 g) had been liberated from the Ilomastat-treated collagen gels whereas cells incubated with the DMSO vehicle had liberated ∼1 mL of medium (0.8 ± 0.05 g; Fig. 3B).

The three MMPs primarily responsible for degrading type 1 collagen are MMP-1, MMP-13, and the membrane-bound MT1-MMP (1). MMP-1 and MMP-13 are secreted as proenzymes and require stepwise cleavage of the propeptide (often by serine proteases) before the enzymes become active (3). MT1-MMP is embedded in the cell membrane as an active enzyme. Therefore, to distinguish between MT1-MMP–mediated destruction of the collagen versus that mediated by MMP-1 and MMP-13, the serine protease inhibitor aprotinin (50 units/mL collagen) was embedded in collagen with MDA-231 cells containing the empty vector. After 36 hours, no medium had been liberated from collagen containing cells and aprotinin (0.01 ± 0.1 g; Fig. 3B), implying that MMP-1 or even MMP-13 (despite its low level of expression; Table 2) was responsible for the collagenolytic activity of the MDA-231 cells. Western blot analysis of the culture medium confirmed that aprotinin blocked activation of MMP-1. We found no detectable MMP-1 protein in the medium of MDA-231 cells producing the MMP-1 shRNA (Fig. 3C,, lane 1), confirming the ELISA and quantitative RT-PCR findings. However, a band representing cleaved and activated MMP-1 was visible when cells containing the empty vector were grown on plastic (Fig. 3C,, lane 2). When these cells were embedded in collagen, there were two MMP-1 protein bands representing the 54 kDa zymogen and the smaller activated form of the enzyme (Fig. 3C,, lane 3). Finally, medium from aprotinin-treated cells contained only the higher molecular weight zymogen band (Fig. 3C , lane 4). The presence of more MMP-1 protein in the medium of cells grown in collagen agrees with our findings that embedding MDA-231 cells in a collagen matrix stimulates a 2- to 3-fold increase in MMP-1 expression (data not shown). Cleavage of the propeptide domain of the 54-kDa MMP-1 zymogen produces a 44-kDa active form; however, smaller 24-kDa MMP-1 proteins represent active MMP-1 (26, 27). The Western blot confirms that MDA-231 cells produce proteinases that activate proMMP-1 (28) and that aprotinin can block this activation.

Finally, to definitively prove that the collagenolytic activity of the MDA-231 cells is due to MMP-1, MMP-1-neutralizing antibody was embedded in the collagen along with MDA-231 cells containing the empty vector. A monoclonal antibody against the FLAG epitope was used as the negative control. At 24 hours, there was a noticeable difference in the thickness of the collagen gels (data not shown). After 40 hours, the weight of medium liberated from gels containing FLAG antibody and cells was comparable to that of the medium liberated from wells containing cells alone (0.78 ± 0.03 g) whereas very little medium was liberated from collagen gels containing the MMP-1 antibody and cells (0.14 ± 0.04 g; Fig. 3B). The small volume of medium liberated from collagen by MDA-231 cells incubated with the MMP-1-neutralizing antibody may be attributed to the limited effectiveness of the MMP-1-neutralizing antibody in blocking MMP-1 activity for a long period of time. Indeed, by 50 hours, the collagen was completely destroyed (data not shown), indicating that MMP-1 production by these cells could eventually overwhelm the neutralizing antibody. Conversely, when the MDA-231 cell lines producing MMP-1 shRNA were incubated in collagen for as long as 5 days, there was no collagen destruction (data not shown). These results confirm that the MDA-231 cell–mediated destruction of collagen is a result of their MMP-1 expression and supports the hypothesis that blocking MMP-1 expression with a stably integrated MMP-1 shRNA vector blocks the ability of MDA-231 cells to degrade a type 1 collagen matrix.

In vivo tumorigenesis. We next determined whether the shRNA-mediated reduction of MMP-1 in MDA-231 cells would affect the growth of the cells in an in vivo model of breast tumor growth. Two separate groups of nude mice were injected in the fourth inguinal mammary fat pad with MDA-231 cells containing the empty vector or the MMP-1 shRNA. At week 6, tumors began to appear and at week 11, the mice were sacrificed (Fig. 4). The frequency of tumor formation in mice injected with the empty vector cell lines (57%, n = 23) was not significantly greater than that in mice injected with MMP-1 shRNA–producing cells (36%, n = 28; P = 0.11, Fisher's exact test; Table 4). However, at the time of sacrifice, the average tumor size in mice injected with cells containing the empty vector (1,217 ± 334 mm3, mean ± SE; n = 13) was significantly greater (P = 0.027, Student's t test) than the average tumor size in mice injected with cells producing the MMP-1 shRNA (272 ± 117 mm3, mean ± SE; n = 10; Fig. 4A). Furthermore, exponential curve fits of the average weekly tumor sizes generated by the Excel program reveal that the tumor cells producing MMP-1 were growing at a faster rate (tumor volume = 2.4874e0.5431week, R2 = 0.9412) than tumor cells not producing MMP-1 (tumor volume = 21.802e0.1892week, R2 = 0.2883; Fig. 4B). Examination at autopsy revealed no macroscopic metastases to any organs in the peritoneal cavity, the lungs, or the brain (data not shown). Furthermore, histologic slides of these organs revealed no microscopic metastases (data not shown). These data suggest that MMP-1 is not necessary for tumor formation but that it contributes to the growth of primary tumors in this mouse model.

Figure 4.

Tumor volumes in nude mice. Mice were injected in the fourth inguinal fat pad with 1 × 106 pooled MDA-231 cells. A, the mice were observed for 11 weeks and average tumor size ± SE was recorded. Tumors in mice injected with cells containing the empty pSuper-retro vector grew more quickly than tumors in mice injected with cells producing MMP-1 shRNA. B, MDA-231 cells containing the empty pSuper-retro vector (○) formed larger tumors (1,217 ± 334 mm3, average ± SE; n = 13) than MDA-231 cells containing the MMP-1 shRNA (+; 272 + 117 mm3, average ± SE; n = 10; P = 0.027, Student's t test). Superficial tumor measurements were used to calculate the tumor volumes [(4/3)πr3].

Figure 4.

Tumor volumes in nude mice. Mice were injected in the fourth inguinal fat pad with 1 × 106 pooled MDA-231 cells. A, the mice were observed for 11 weeks and average tumor size ± SE was recorded. Tumors in mice injected with cells containing the empty pSuper-retro vector grew more quickly than tumors in mice injected with cells producing MMP-1 shRNA. B, MDA-231 cells containing the empty pSuper-retro vector (○) formed larger tumors (1,217 ± 334 mm3, average ± SE; n = 13) than MDA-231 cells containing the MMP-1 shRNA (+; 272 + 117 mm3, average ± SE; n = 10; P = 0.027, Student's t test). Superficial tumor measurements were used to calculate the tumor volumes [(4/3)πr3].

Close modal
Table 4.

Frequency of tumor formation in mice injected with stably transfected MDA-231 cell lines

MDA stable cell lineNo. miceNo. mice with tumorsFrequency of tumor formation (%)Fisher's exact test
Empty vector 23 13 57 P = 0.11 
MMP-1 shRNA 28 10 36  
MDA stable cell lineNo. miceNo. mice with tumorsFrequency of tumor formation (%)Fisher's exact test
Empty vector 23 13 57 P = 0.11 
MMP-1 shRNA 28 10 36  

NOTE: MDA-231 cells were injected into the mammary fat pad of nude mice and the mice were observed for 11 weeks.

In previous studies, transgenic expression of MMP-1 in mouse skin resulted in increased susceptibility to tumorigenesis (15). Conversely, our data suggest that MMP-1 does not contribute to the frequency of tumor formation but does contribute to tumor growth. A possible explanation for this difference is that in our system, MMP-1 was expressed in cancer cells only and transgenic expression of MMP-1 resulted in expression from both cancer cells and normal stromal cells. Although our data suggest that MMP-1 contributes to the growth of tumors in mice, the mechanism by which MMP-1 influences tumor growth is not entirely clear.

MMP-1 activity may be required to break down the fibrous stroma within the mammary fat pad, removing a physical barrier to growth; however, degradation of ECM molecules does more than facilitate cell mobility by removing a physical barrier. The ECM contains growth factors that are liberated by breakdown of matrix proteins. MMP-1, in particular, is capable of cleaving the proteoglycan perlecan, which can release basic fibroblast growth factor (29). MMP-1 can also cleave insulin-like growth factor (IGF) binding proteins that are present in the ECM, thereby increasing the availability of IGF and increasing cell proliferation (30). Other MMP-1 substrates within the ECM include proteins that modulate immune responses, such as monocyte chemoattractant proteins (31). When MMP-1 cleaves monocyte chemoattractant protein, it produces a CC chemokine receptor antagonist which may block phagocytic cell recruitment to the site of tumor cell injection. Although nude mice are athymic, they have some components of their immune system, including monocyte/macrophages. Thus, MMP-1 expression may have contributed to tumor growth in these mice through mechanisms beyond the traditional collagen destruction model of tumor invasion.

Our findings are similar to those describing decreased collagenolytic activity of chondrosarcoma cells transiently transfected with MMP-1 siRNA molecules (18). Transient transfection of three siRNA molecules targeted at MMP-1 resulted in a 50% to 80% inhibition of MMP-1 between 48 and 72 hours after transfections with a recovery to pretransfection levels by 120 hours. By using a stably integrating plasmid to express our shRNA molecule, we obtained a >90% reduction in MMP-1 expression that persisted in the cells for several months (8 months to date). The persistent inhibition effected by stably integrating the shRNA vector into cellular DNA may account for the increased effectiveness of our shRNA molecule compared with transient transfection.

The specificity of the MMP-1 shRNA for its target gene agrees with previously published reports on the specificity of RNA interference molecules (23, 32). It is unclear why the cells lines with the scrambled shRNA vector did not grow because the sequence used did not show homology to any human gene. RNA interference molecules can induce IFN responses in a sequence-specific and cell type–specific manner (25, 33). Therefore, the scrambled shRNA may have triggered an IFN response in the MDA-231 cells. Indeed, our recent findings indicate that the same scrambled shRNA sequence induced IFN-responsive genes in melanoma cell lines (data not shown). siRNAs may cause changes in untargeted proteins in mammalian cells; however, the mechanism is unknown (34, 35). One possibility is that transiently transfected siRNAs engage toll-like receptors on the target cells (36). shRNA stably expressed from an integrated vector circumvents this potential difficulty. Additionally, to minimize the possibility of off-target effects, we selected one shRNA sequence that showed efficacy in reducing the levels of both an ectopically and endogenously expressed gene (Fig. 1; Table 1). Furthermore, we measured closely related genes to verify the specificity of the shRNA and found no consistent differences in expression. The effects of blocking MMP-1 expression on cancer cell invasion and tumor growth described here are consistent with published findings on blocking MMP expression in vitro and in vivo (17, 18, 37).

A recent report describes the ability of RNA interference targeted to two other proteinases, MMP-9 and cathepsin B, to suppress tumor cell invasion, tumor growth, and angiogenesis in a glioblastoma model (37). Similarly, RNA interference blocked MT1-MMP expression in fibrosarcoma and gastric cancer cells, resulting in decreased in vitro invasion (38). Our results show that MMP-1 is another protease that can affect the growth of tumors formed by cancer cells introduced into an animal model. Furthermore, another recent article hints at the importance of MMP-1 expression in the development of invasive breast cancer and suggests that MMP-1 expression can be used as a predictive marker for the development of this disease in patients with atypical ductal hyperplasia (11). Our findings confirm that siRNA/shRNA-mediated suppression of proteinases represents a powerful new approach to understanding the role of these enzymes in cancer and provides a potential avenue for new antiproteinase cancer therapies.

Grant support: NIH grant AR-26599, Department of Defense grant 17-00-1-0221 (C.E. Brinckerhoff), and Department of Defense predoctoral fellowship 17-01-1-0225 and NIH grant T32 AR07576 (C.A. Wyatt).

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

We thank the biostatistics center at the Norris Cotton Cancer Center and the pathology department at Dartmouth Hitchcock Medical Center for support services.

1
Stamenkovic I. Extracellular matrix remodelling: the role of matrix metalloproteinases.
J Pathol
2003
;
200
:
448
–64.
2
Junqueira LC, Carneiro J, Kelley RO. Connective tissue. Basic histology. 9 ed. Stamford (CT): Appleton & Lange; 1998. p. 89–120.
3
Visse R, Nagase H. Matrix metalloproteinases and tissue inhibitors of metalloproteinases: structure, function, and biochemistry.
Circ Res
2003
;
92
:
827
–39.
4
Brinckerhoff CE, Matrisian LM. Matrix metalloproteinases: a tail of a frog that became a prince.
Nat Rev Mol Cell Biol
2002
;
3
:
207
–14.
5
Murray GI, Duncan ME, O'Neil P, Melvin WT, Fothergill JE. Matrix metalloproteinase-1 is associated with poor prognosis in colorectal cancer.
Nat Med
1996
;
2
:
461
–2.
6
Airola K, Karonen T, Vaalamo M, et al. Expression of collagenases-1 and -3 and their inhibitors TIMP-1 and -3 correlates with the level of invasion in malignant melanomas.
Br J Cancer
1999
;
80
:
733
–43.
7
Murray GI, Duncan ME, O'Neil P, McKay JA, Melvin WT, Fothergill JE. Matrix metalloproteinase-1 is associated with poor prognosis in oesophageal cancer.
J Pathol
1998
;
185
:
256
–61.
8
Ito T, Ito M, Shiozawa J, Naito S, Kanematsu T, Sekine I. Expression of the MMP-1 in human pancreatic carcinoma: relationship with prognostic factor.
Mod Pathol
1999
;
12
:
669
–74.
9
Inoue T, Yashiro M, Nishimura S, et al. Matrix metalloproteinase-1 expression is a prognostic factor for patients with advanced gastric cancer.
Int J Mol Med
1999
;
4
:
73
–7.
10
Nakopoulou L, Giannopoulou I, Gakiopoulou H, Liapis H, Tzonou A, Davaris PS. Matrix metalloproteinase-1 and -3 in breast cancer: correlation with progesterone receptors and other clinicopathologic features.
Hum Pathol
1999
;
30
:
436
–42.
11
Poola I, Dewitty RL, Marshalleck JJ, Bhatnagar R, Abraham J, Leffall LD. Identification of MMP-1 as a putative breast cancer predictive marker by global gene expression analysis.
Nat Med
2005
;
11
:
481
–3.
12
Hotary K, Allen E, Punturieri A, Yana I, Weiss SJ. Regulation of cell invasion and morphogenesis in a three-dimensional type I collagen matrix by membrane-type matrix metalloproteinases 1, 2, and 3.
J Cell Biol
2000
;
149
:
1309
–23.
13
Hotary KB, Allen ED, Brooks PC, Datta NS, Long MW, Weiss SJ. Membrane type I matrix metalloproteinase usurps tumor growth control imposed by the three-dimensional extracellular matrix.
Cell
2003
;
114
:
33
–45.
14
Benbow U, Schoenermark MP, Mitchell TI, et al. A novel host/tumor cell interaction activates matrix metalloproteinase 1 and mediates invasion through type I collagen.
J Biol Chem
1999
;
274
:
25371
–8.
15
D'Armiento J, DiColandrea T, Dalal SS, et al. Collagenase expression in transgenic mouse skin causes hyperkeratosis and acanthosis and increases susceptibility to tumorigenesis.
Mol Cell Biol
1995
;
15
:
5732
–9.
16
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-1.
Biochim Biophys Acta
1997
;
1356
:
271
–80.
17
Jiang X, Dutton CM, Qi W, et al. Inhibition of MMP-1 expression by antisense RNA decreases invasiveness of human chondrosarcoma.
J Orthop Res
2003
;
21
:
1063
–70.
18
Jiang X, Dutton CM, Qi WN, Block JA, Garamszegi N, Scully SP. siRNA mediated inhibition of MMP-1 reduces invasive potential of a human chondrosarcoma cell line.
J Cell Physiol
2005
;
202
:
723
–30.
19
Paddison PJ, Caudy AA, Bernstein E, Hannon GJ, Conklin DS. Short hairpin RNAs (shRNAs) induce sequence-specific silencing in mammalian cells.
Genes Dev
2002
;
16
:
948
–58.
20
Brummelkamp TR, Bernards R, Agami R. A system for stable expression of short interfering RNAs in mammalian cells.
Science
2002
;
296
:
550
–3.
21
Brummelkamp TR, Bernards R, Agami R. Stable suppression of tumorigenicity by virus-mediated RNA interference.
Cancer Cell
2002
;
2
:
243
–7.
22
Tuschl T. Expanding small RNA interference.
Nat Biotechnol
2002
;
20
:
446
–8.
23
Dykxhoorn DM, Novina CD, Sharp PA. Killing the messenger: short RNAs that silence gene expression.
Nat Rev Mol Cell Biol
2003
;
4
:
457
–67.
24
Stamenkovic I. Matrix metalloproteinases in tumor invasion and metastasis.
Semin Cancer Biol
2000
;
10
:
415
–33.
25
Bridge AJ, Pebernard S, Ducraux A, Nicoulaz AL, Iggo R. Induction of an interferon response by RNAi vectors in mammalian cells.
Nat Genet
2003
;
34
:
263
–4.
26
Clark IM, Cawston TE. Fragments of human fibroblast collagenase. Purification and characterization.
Biochem J
1989
;
263
:
201
–6.
27
Clark IM, Cawston TE. Fragments of human fibroblast collagenase: purification and characterisation.
Matrix Suppl
1992
;
1
:
73
.
28
Benbow U, Schoenermark MP, Orndorff KA, Givan AL, Brinckerhoff CE. Human breast cancer cells activate procollagenase-1 and invade type I collagen: invasion is inhibited by all-trans retinoic acid.
Clin Exp Metastasis
1999
;
17
:
231
–8.
29
Whitelock JM, Murdoch AD, Iozzo RV, Underwood PA. The degradation of human endothelial cell-derived perlecan and release of bound basic fibroblast growth factor by stromelysin, collagenase, plasmin, and heparanases.
J Biol Chem
1996
;
271
:
10079
–86.
30
Fowlkes JL, Enghild JJ, Suzuki K, Nagase H. Matrix metalloproteinases degrade insulin-like growth factor-binding protein-3 in dermal fibroblast cultures.
J Biol Chem
1994
;
269
:
25742
–6.
31
McQuibban GA, Gong JH, Wong JP, Wallace JL, Clark-Lewis I, Overall CM. Matrix metalloproteinase processing of monocyte chemoattractant proteins generates CC chemokine receptor antagonists with anti-inflammatory properties in vivo.
Blood
2002
;
100
:
1160
–7.
32
Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K, Tuschl T. Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells.
Nature
2001
;
411
:
494
–8.
33
Sledz CA, Holko M, de Veer MJ, Silverman RH, Williams BR. Activation of the interferon system by short-interfering RNAs.
Nat Cell Biol
2003
;
5
:
834
–9.
34
Jackson AL, Linsley PS. Noise amidst the silence: off-target effects of siRNAs?
Trends Genet
2004
;
20
:
521
–4.
35
Scacheri PC, Rozenblatt-Rosen O, Caplen NJ, et al. Short interfering RNAs can induce unexpected and divergent changes in the levels of untargeted proteins in mammalian cells.
Proc Natl Acad Sci U S A
2004
;
101
:
1892
–7.
36
Kariko K, Bhuyan P, Capodici J, et al. Exogenous siRNA mediates sequence-independent gene suppression by signaling through toll-like receptor 3.
Cells Tissues Organs
2004
;
177
:
132
–8.
37
Lakka SS, Gondi CS, Yanamandra N, et al. Inhibition of cathepsin B and MMP-9 gene expression in glioblastoma cell line via RNA interference reduces tumor cell invasion, tumor growth and angiogenesis.
Oncogene
2004
;
23
:
4681
–9.
38
Ueda J, Kajita M, Suenaga N, Fujii K, Seiki M. Sequence-specific silencing of MT1-MMP expression suppresses tumor cell migration and invasion: importance of MT1-MMP as a therapeutic target for invasive tumors.
Oncogene
2003
;
22
:
8716
–22.