Purpose: Hepatocyte growth factor/scatter factor (HGF/SF) is known to increase the invasiveness and migration of cancer cells in vitro and induce angiogenesis. This study examined if inhibition of HGF/SF receptor expression by cancer cells and HGF/SF expression by stromal fibroblasts affects the growth of mammary cancer.

Experimental Design: Transgenes encoding ribozymes to specifically target human HGF/SF receptor (pLXSN-MET) or HGF/SF (pLXSN-HGF) were constructed using a pLXSN retroviral vector. Human mammary cancer cells MDA MB 231 was transduced with pLXSN-MET (MDA+/+). A human fibroblast cell line MRC5, which produces bioactive HGF/SF, was transduced with pLXSN-HGF (MRC5+/+). These cells were used in a nude mice breast tumor model.

Results: HGF receptor in MDA+/+ cells and HGF in MRC5+/+cells were successfully removed with respective ribozymes as shown by reverse transcription-PCR and Western blotting, respectively. MDA+/+ was found to have reduced invasiveness when stimulated with HGF/SF. MRC5+/+ exhibited a significant reduction in HGF/SF production. When injected into athymic nude mice, MDA+/+ exhibited a slower rate of growth, compared with the wild type (MDA−/−), and the cells transduced with control viral vector (MDA+/−). The growth of MDA−/− tumor was significantly enhanced when coimplanted with wild-type MRC5 (MRC5−/−), and the stimulatory effect was reduced when MRC5+/+ cells were coimplanted instead of MRC5−/−. The reduction of tumor growth was accompanied by reduction of angiogenesis, as demonstrated by the staining of VE-cadherin in primary tumor tissues.

Conclusions: Retroviral ribozyme transgenes targeting HGF/SF in fibroblasts or its receptor cMET in mammary cancer cells can reduce the growth of mammary cancer and associated angiogenesis by inhibiting paracrine stromal–tumor cell interactions.

The cytokine HGF,3 also known as SF, is primarily expressed and produced by stromal cells, such as fibroblasts in mammary tissues (1, 2). It plays multiple roles in mammary cancer and is known to stimulate the dissociation, migration, motility, adhesion to, and invasion of extracellular matrix, via mechanisms such as induction of the phosphorylation of focal adhesion kinase and paxillin (3, 4, 5). HGF/SF is one of the most potent angiogenic factors, as demonstrated by both in vitro and in vivo studies (6, 7, 8). The action of HGF/SF is mediated by its specific receptor, c-MET, a transmembrane protein encoded by the proto-oncogene, c-MET, which has an intrinsic kinase domain (9, 10). On stimulation of HGF/SF, c-MET becomes phosphorylated and initiates a range of intracellular signals that lead to activation of cellular behaviors.

The clinical significance of HGF and its receptor in human mammary cancer has begun to emerge in the last few years. There is an overexpression of HGF and the HGF/SF receptor in human mammary tumors, compared with normal mammary tissues. In a series of 258 primary mammary cancers, the tissue level of immunoreactive HGF was found to increase and inversely correlate with both relapse-free and overall survival (8). The clinical importance of HGF and its receptor is further demonstrated in recent studies (9, 10), showing that the levels of c-MET in mammary cancer tissues and levels of circulating HGF in patients with mammary cancer are associated with a lower survival and development of distant metastasis (11, 12). Patients with multiple metastatic sites and liver metastases exhibited the highest HGF levels. Furthermore, the increase in serum HGF levels was significantly associated with tumor size, nodal status, and histological evidence of venous invasion (13). The levels of c-MET in the circulating mammary cancer cells correlate significantly with the size and grade of mammary cancer (14), and the levels of c-MET, as visualized by confocal microscopy, are linked to shorter survival of patients with mammary cancer (15, 16). Thus, suppression of the action of HGF and its receptor may be of therapeutic value.

Recently, it has been reported that hammerhead ribozymes that target c-MET and HGF are able to reverse the malignancy of glioblastoma cells both in vitro and in vivo(17), and we found that anti-cMET U1/ribozyme reduces invasiveness and migration of mammary cancer cells in vitro(18). Here, we report the development of retroviral U1/ribozyme transgenes that target both c-MET and HGF, and their effects on mammary cancer and particularly on fibroblasts in vitro, and in a mammary tumor model in vivo.

Human mammary cancer cells, MDA MB 231, and an immortalized nontumorigenic human fibroblast cell line MRC5 were from the European Collection of Animal Cell Culture (Salisbury, England) and were routinely maintained in DMEM:F12 with 10% FCS. Recombinant human HGF/SF was a gift from Dr. T. Nakamura, Osaka University Medical School (Osaka, Japan). Matrigel (reconstituted basement membrane) was purchased from Collaborative Research Products (Bedford, MA). A rabbit anti-α chain of human HGF; a rabbit antihuman c-MET (h-MET); a monoclonal antityrosine (PY99); and monoclonal antibodies to human VE-cadherin, antivimentin, anti-CK19, anti-Ki67; and peroxidase-conjugated anti-IgG were from Santa Cruz Biotechnology (Santa Cruz, CA) and Sigma (Poole, Dorset, United Kingdom), respectively. A chemiluminescence detection kit for Western blotting and protein A/G conjugate were from Santa Cruz Biotechnology. A transwell plate equipped with a porous insert (pore size of 8 μm) was from Becton Dickinson Labware (Oxford, United Kingdom). DNA restriction enzymes and T4 DNA ligase were obtained from New England Biological Laboratories (Hertfordshire, United Kingdom). DNA gel extraction and plasmid extraction kits were from Qiagen (Crawley, United Kingdom).

Construction of Retroviral Hammerhead Ribozyme Transgenes Targeting Human cMET HGF/SF and Generation of Active Viral Hammerhead Ribozymes.

A hammerhead ribozyme that specifically targets a GUC site of human HGF/SF (GenBank accession no. M29145, position 701) and a ribozyme which targets a GUC site of the HGF/SF receptor (GenBank accession no. J02958, position 560) have been described previously (17, 18). Plasmids with the U1 ribozyme transgenes were digested using BamH1. The correct inserts were purified and recloned into a retroviral vector, pLXSN (Clontech Laboratories, Palo Alto, CA), at a matched site. An irrelevant sequence was also cloned into pU1 as control empty vector and was referred to as pU1 in the text. The successfully ligated plasmids were amplified in Escherichia coli strain JM109 (Clontech). The direction and sequence were verified using a plasmid specific primer LXSNF-5′cccttgaacctcctcgttcgacc′3 and U1-specific primer 5′ggatccgccaaccgaaagt3′ (UBAMHF) and 5′gtacgattaacaactaaga′3 (UBAMHR).

Retroviral Packaging and Transduction of Cells.

Plasmid, extracted and purified using a plasmid extraction kit (Qiafilter; Qiagen, Crawley, United Kingdom), was introduced to a packaging cell line, PT67, using electroporation as described previously (18), followed by selection with G418-containing medium for >3 weeks. Viral titers from stably transfected PT67 cells were tested using NIH3T3 cells and found to be on average 8 × 105 cfu/ml. Active viral stocks were used to transduce MDA MB 231 mammary cancer cells, or MRC5 cells, in the presence of Polybrene (8 μg/ml final concentration). Each transduction lasted 24 h, and three consecutive tranductions were carried out. Transduced cells were subject to selection with G418 (Calbiochem, Nottingham, United Kingdom), at 100 μg/ml for >3 weeks to obtain stably transduced strains. These stably transduced and subsequently verified cells are designated the following names and are used through the text: MDA−/−: MDA MB 231 wild type; MDA+/−: MDA MB 231 transduced with pLXSN-PU1 empty vector; MDA+/+: MDA MB 231 transduced with pLXSN-MET560 transgene; MRC5−/−: wild-type MRC5 cells; MRC5+/−: MRC5 cells transduced with control pLXSN-PU1; and MRC5+/+: MRC5 cells transduced with pLXSN-HGF701.

RNA Preparation and RT-PCR.

RNA from cells and tissues (including bones and lungs after homogenization) was extracted using an RNA extraction kit (AbGene Ltd., Surrey, United Kingdom) and quantified using spectrophotometer (Wolf Laboratories). RNA from blood was extracted using a TRIZol method (Sigma). cDNA was synthesized using a first strand synthesis with an oligo dt primer (AbGene). PCR primers were: for cMET, 5′gtccaggcagtgcagcatgta′3 and 5′actatagtattctttatcatacatgtc′3; human HGF, ′atgccagcactgaagataaaaacc′3, 5′ctgtggtgtggatgttatcag′3; CK19 primers, 5′atgacttcctacagctatcgccagt′3, catggttagcttctcgtt′3. The PCR was performed using sets of primers with the following conditions: 5 min at 95°C and then 20 s at 94°C to 25 s at 56°C for HGF (60°C for cMET and 52°C for vector primers), 50 s at 72°C for 36 cycles, and finally 72°C for 7 min. β-actin was amplified simultaneously using the following primers: 5′gctgatttgatggagttgga3′ and 5′tcagctacttgttcttgagtgaa3′. PCR products were then separated on a 0.8% agarose gel, visualized under UV light, photographed using a Unisave camera (Wolf Laboratories), and documented with PhotoShop software.

Immunoprecipitation and Western Blotting.

Cells were extracted using a lysis buffer that contained 2.4 mg/ml Tris, 4.4 mg/ml NaCl, 5 mg/ml sodium deoxycholate, 20 μg/ml sodium azide, 1.5% Triton, 100 μg/ml phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin, and 1 μg/ml aprotinin for 30 min (19). They were then boiled at 100°C for 5 min before clarification at 13,000 × g for 10 min. Protein concentrations were measured using fluorescamine and quantified by using a multifluoroscanner (Denly, Sussex, United Kingdom). Equal amounts of protein from each cell sample (controls and treated) were added onto an 8% poly-acrylamide gel. For immunoprecipitation and HGF receptor activation, MDA MB 231 cells were subject to serum hunger for 4 h, before stimulation with either serum-free medium, HGF, or serum-free conditioned medium from MRC5−/−, MRC5+/−, or MRC5+/+. Positive control was cells treated with 2 mm sodium orthovanadate with 0.3% hydrogen peroxide. After 30 min, cells were pelleted and lysed in the same buffer with 1 mm sodium orthovanadate. To equal amounts of protein was added anti-cMET antibody (4 μg, for 1 h with constant agitation), followed by the addition of protein A/G agarose for an additional hour and washing with the same lysis buffer. The immunoprecipitate was solubilized with a sample buffer that contained 5% 2-mercaptolethanol. Protein blot was probed with anti-cMET (for MDA MB 231 total lysates and precipitates), antityrosine antibody (for immunoprecipitates), and anti-HGF antibody (for MRC5 lysates), followed by washing and subsequent probing with appropriate peroxidase-conjugated secondary antibodies. Protein bands were visualized with an enhanced chemiluminescence kit.

In Vitro Invasion Analysis and Cell Growth Assay.

This was performed as reported previously and modified in our laboratory (19, 20). Briefly, transwell inserts (top chamber) with 8-μm pore size were coated with 50 μg/insert of Matrigel and air dried, before being rehydrated. A total of 20,000 cells was added to each well with or without HGF/SF. In selected wells, MRC5 cells were preseeded in the bottom chamber as a source of bioactive HGF/SF. After 72 h, cells that had migrated through the matrix and adhered to the other side of the insert were fixed and stained with 0.5% (w/v) crystal violet. Cells that have invaded and stained with crystal violet were extracted with 10% (volume for volume) of acetic acid, and absorbance was obtained using a multiplate reader.

For cell growth assay, MRC5−/−, MRC5+/−, MRC5+/+, MDA−/−, MDA+/−, or MDA+/+ were plated into a 96-well plate at 2500 cells/well. Cells were fixed in 10% formaldehyde at the day of plating, day 1, 2, 3, 4, 5, and 6 after plating, and then stained with 0.5% (w/v) crystal violet. After washing, stained crystal was extracted with 10% (volume for volume) acetic acid, and absorbance was determined using a multiplate reader. The growth of cells is shown here as absorbance (mean ± SD).

HGF/SF Bioassay.

The quantity of bioactive HGF/SF secreted by MRC5 was determined using a MDCK bioassay, as described previously (21). rhHGF/SF was used as an internal standard. In the current study, 1 unit/ml of bioactivity equals 0.5 ng/ml rhHGF.

In Vivo Development of Mammary Tumor.

Athymnic nude mice (Nude CD-1) of 4–6 weeks old were purchased from Charles River Laboratories (Kent, England, United Kingdom) and maintained in filter-toped units. One-hundred μl of cell suspension (in 0.5 mg/ml Matrigel) was injected s.c. at the left scapula area. The following groups were included: group 1, no cells; group 2, wild-type MDA MB 231 cells (MDA−/−; 1 × 106); group 3, MDA+/+ (1 × 106); group 4, MDA+/− (1 × 106); group 5, MDA−/− (1 × 106) and wild MRC5 (MRC5−/−; 0.3 × 106); group 6, MDA+/+ (1 × 106) and MRC5−/− (0.3 × 106); and group 7, MDA−/− (1 × 106) and MRC5+/+ (0.3 × 106); and finally, group 8, MDA+/+ (1 × 106) and MRC5+/+ (0.3 × 106) in a 100-μl mixture. Each group had six mice.

Mice were weighed, and tumor sizes were measured twice weekly for 4 weeks. Mice with weight loss over 25% and tumor size >1 cm in any dimension were terminated according to the United Kingdom Home Office and UKCCCR guideline. The volume of the tumor was determined using the formula: tumor volume = 0.523 × width2 × length. At the conclusion of the experiment, animals were terminally anesthetized, and primary tumors, lungs, liver, and both hind legs were dissected, weighed, and frozen at −80°C. Blood was withdrawn by cardiac puncture. Part of the primary tumors and lungs were fixed for histological examination. RNA from the bone and blood was extracted after homogenization in RNAzol (bone and lung) or TRIzol (blood). RT-PCR was subsequently carried out to detect the existence of cancer cells using primers set for human CK19 and human cMET.

Immunohistochemical Analysis of Tumor and Lung Tissues.

Primary tumors and lungs were immediately immersed in a fixation solution of neutral buffered formalin for 48 h, followed by processing on a Reichert-Jung automatic tissue processor (24 h cycle) before embedding in pastillated paraffin wax (congealing point at 57°C–60°C). Seven-μm sections were cut and placed on Super Frost Plus slides and incubated overnight at 60°C.

After antigen retrieval treatment and pretreating with blocking reagent, the tumor tissue sections (four sequential sections) were stained first with a monoclonal antibody to VE-cadherin, then with biotinylated secondary antibody and avidin-biotin complex. Each step was followed by extensive washing in buffer. 3,3′-diaminobenzidine was used as a chromogen to develop visible color. The number of positive vessels was assessed independently by two assessors, from randomly chosen fields using the following criteria: positively stained cells with lumen structure. In addition, tumor tissue sections were also similarly stained with antivimentin antibody, to identify fibroblasts (positively stained cells). The number of positively stained cells and total number of cells in a fixed frame were independently counted by two researchers (22). The proportion of fibroblasts in tissues is shown as a percentage of vimentin-positive cells (fibroblasts)/total number of cells. The proliferation of cells was assessed after staining tumor tissues with anti-Ki67 antibody and is shown as the number of cells stained positive for Ki67 in 100 randomly chosen cells in a randomly chosen field. Lung tissues were sequentially sectioned (>sections) and stained with anti-CK19 to assess the presence of possible micrometastatic lesions.

Statistical analysis was carried out using Mann-Whitney U test, and significant difference was taken at P < 0.05.

Retroviral Ribozymes Reduced Expression of cMET by MDA Cells and HGF/SF by Fibroblasts.

Three retroviral hammerhead ribozyme transgenes were constructed using the retroviral vector pLXSN. pLXSN-MET targets a GUC sequence of human cMET (GenBank accession no. J02958, position 560). pLXSN-HGF/SF701 targets a GUC sequence of human HGF/SF (GenBank accession no. M29145, position 701) and control vector with an irrelevant control sequence, the direction and sequence of which were verified using plasmid/ribozyme specific primers and direct DNA sequencing. Live viral stocks generated from the packaging cell line, PT67, were used directly to transduce the respective cells, i.e., MET transgene for mammary cancer cells and HGF transgene for MRC5 cells.

RT-PCR of the stably transduced MDA MB 231 cells (MDA+/+) showed the presence of U1/MET ribozyme transcript (Fig. 1). Furthermore, MDA+/+ cells showed markedly reduced levels of met transcript, whereas in the wild-type MDA MB 231 (MDA−/−), met transcript existed at a high level (Fig. 1). Transduction with control viral stocks (MDA−/− cells) did not significantly affect the level of met expression. Successful expression of U1/HGF ribozyme transgene was also obtained in MRC5 human fibroblast cells. MRC5−/− expressed high levels of HGF mRNA, which was effectively inhibited by U1/HGF ribozyme in the stably transduced cells (MRC5+/+).

Western blotting from these cell extracts revealed almost removal of cMET from MDA MB 231 cells (MDA+/+; Fig. 2,A, top) and removal of HGF protein from MRC5+/+ cells (Fig. 2,A, bottom). The reduction of HGF protein was also reflected in the HGF bioassay using MDCK cells. Fig. 2 B shows that the U1/HGF viral ribozyme (MRC5+/+) significantly reduced the amount of bioactive HGF/SF secreted by the fibroblast cells, when compared with supernatants from wild-type MRC5 and control viral transductant MRC5+/−.

The effects of the viral ribozymes on HGF and HGF receptor were further demonstrated by immunoprecipitation analysis. Immunoprecipitation with anti-cMET antibody resulted in the appearance of a high level of cMET from both MDA−/− and MDA+/− but not MDA+/+ (Fig. 2,C, bottom, left, middle, and right, respectively). In addition, cMET in MDA−/− (Fig. 2,C, top left) and MDA+/− (Fig. 2,C, top middle) cells can be activated (phosphorylated) by recombinant HGF, supernatant from MRC5−/− and MRC5+/−. Interestingly, supernatant from MRC5+/+ lost its ability to induce the tyrosine phosphorylation of cMET (Fig. 2,C, top). In contrast, MDA+/+ cells, which lost the expression of cMET because of the ribozyme transgene, no longer exhibited any receptor phosphorylation in response to HGF/SF or MRC5-conditioned medium (Fig. 2 C, top right). This indicates that MDA MB 231 wild-type and MDA+/− expressed functioning HGF receptors, which can be activated by HGF and bioactive HGF produced by MRC5 cells. Specific retroviral ribozymes successfully reduced both HGF receptor from MDA MB 231 cells and HGF from MRC5 fibroblasts.

LXSN-MET Reduced Invasiveness of MDA MB 231 Mammary Cancer Cells in Response to HGF/SF and Fibroblasts.

MDA MB 231 cells were tested for their in vitro invasiveness in response to either recombinant human HGF/SF or to MRC5 cells in a coculture system. HGF/SF, as expected, significantly increased the invasiveness of the wild-type cancer cells (MDA−/−) through Matrigel (Fig. 3). Similarly, the effect of fibroblasts on the invasiveness of mammary cancer cells was examined in a coculture invasion model. In this analysis, MRC5 cells sat in the bottom chamber as the source of bioactive HGF/SF, whereas mammary cancer cells sat in the top chamber. Fig. 3 shows that wild-type MRC5−/− facilitated invasiveness of the mammary cancer cells, as a result of bioactive HGF/SF produced by MRC5 (Fig. 2).

The study also demonstrated that MDA+/+, which has lost the expression of cMET mRNA, exhibited a diminished invasive response to either recombinant HGF/SF or stromal fibroblasts MRC5 (Fig. 3).

In a cell growth assay, retroviral vectors (control and pLXSN-MET) have marginally increased the growth of MDA MD 231 cells (Table 1). Interestingly, control vector and pLXSN-HGF resulted in a small but insignificant increase in the growth of MRC5 cells (Table 1).

LXSN-MET Inhibits the Growth of MDA MB 231 Cells in Nude Mice Model.

Palpable s.c. tumors were measurable 1 week after inoculation. The size of tumors was continuously monitored twice weekly until tumors grew to the limit permitted by the UKCCCR (1 cm for weight of the mice used in the current study). MDA+/+ cells displayed a reduced growth compared with the wild-type MDA−/− and the cells transduced by control virus, MDA+/− (Fig. 4, left). The final weight of dissected tumors supported the observation based on tumor volume (Fig. 4, right). The correlation coefficient between tumor weight and final volume measurement was 0.943 (Spearman correlation test). The reduction of tumor volume and weight was reflected by the proliferation index as assessed by Ki67 staining of tumor tissues, in that the proliferation index of tumors derived from MET560 transduced MDA+/+ cells was significantly reduced compared with that of the wild-type MDA−/− and control MDA+/− tumors (P < 0.05; Table 2).

MRC5+/+ Exerted Less Stimulation to the Growth of Mammary Cancer in Animal Model.

Our second objective in the in vivo study was to examine the paracrine effect of fibroblasts in the growth of MDA MB 231 mammary tumor. Wild-type mammary cancer cells (MDA−/−), when coinjected with wild type MRC5−/−, exhibited a faster growth rate, compared with MDA MB 231 alone. The increased growth rate in fibroblast-injected group was unlikely from the coinjected fibroblasts, because the proportion of fibroblasts, as demonstrated by vimentin staining, was very small (<5% of total cells; Table 2). Although there was a small increase in the number of fibroblasts in tumors coinjected with MRC5 (3.08 ± 2.4%) compared with the same cell with no MRC5 (2.88 ± 1.57%), the difference is nonetheless insignificant (Table 2). This enhancement of tumor growth by MRC5 was reduced when MRC5+/+ cells were coinjected, instead of wild-type MRC5−/− (Fig. 5). In separate groups, when both MDA+/+ and MRC5+/+ were coinjected, the growth of the tumor was similar to that seen with transduced MDA+/+ and wild-type MRC5−/− (Fig. 6). The difference in tumor volume and weight in respective groups was supported by the changes in the tumor proliferation index (Ki67 staining), in that wild-type MRC5−/− significantly increased the proliferation index (33 ± 4.4%) compared with tumor cell alone (14.13 ± 10.6%, P < 0.01) and that this increase in proliferation index was reduced when cells were coinjected (MRC5+/+; 17.3 ± 6.94%, P < 0.01; Table 2).

Changes of Angiogenesis in Tumor Model.

The tumor vasculature was assessed using an anti-VE-cadherin antibody. There were ∼13 stained lumen per low-power field in tumors derived from with wild-type MDA−/−. When coinjected with MRC5−/−, the number increased to 18 (P = 0.049 versus MDA MB 231 only). The number of vessels was significantly lower in tumors derived from tumors coinjected with MRC5+/+ cells (P < 0.01 versus MDA−/− only and MDA−/− + MRC5−/− wild type). Interestingly, MDA+/+ tumors exhibited a significant reduction in angiogenesis compared with MDA−/− tumors (P < 0.01 versus wild-type MDA). Tumors derived from the combination of MDA+/+ and MRC5+/+ also exhibited significantly fewer vessels compared with their wild-type cells (P < 0.01 versus wild type; Fig. 7).

We failed to detect signs of micrometastasis from liver, lungs, and bone marrow using RT-PCR. Sequential sections of the lungs did not show any CK19-positive tumor cells in lung tissues, suggesting that in the current model and specified study period (4 weeks), metastatic disease did not occur.

This study has demonstrated that a retroviral U1/ribozyme transgene targeting the human HGF receptor, cMET, can effectively eliminate the expression of the HGF receptor mRNA and cMET protein in mammary cancer cells. Our data further demonstrate that mammary cancer cells transduced with the U1/MET transgene have reduced invasiveness in response to HGF and HGF-producing fibroblasts. These cells also exhibited a significant reduction in their rate of growth in vivo. The reduced growth was proposed to be associated with the loss of response to HGF, generated by fibroblasts and possibly other stromal cells.

To determine the effects of HGF produced by fibroblasts on tumor growth, we further constructed a retroviral ribozyme transgene which targets human HGF, namely pLXSN-HGF/SF701. The viral stock of the transgene is very effective in eliminating the expression of HGF mRNA and significantly reduced the level of HGF protein from MRC5, a nontumorigenic human fibroblast cell line known to produce a large quantity of bioactive HGF. This effect was further reflected by the reduction of bioactive HGF from the fibroblasts and reduction of stimulating invasiveness after transduction. Although conditioned medium from wild-type MRC5 and control viral stock transduced MRC5 were able to induce a phosphorylation of cMET in MDA−/− and MDA5+/− cells, conditioned medium from MRC5+/+ failed to elicit any activation on cMET (Fig. 2 C). This indicates that pLXSN-HGF has effectively eliminated HGF transcript and protein from MRC5 fibroblasts and the subsequent activation of HGF receptor of breast cancer cells.

The strongest evidence that the ribozyme and fibroblasts play an important role in tumor development comes from in vivo studies, in which wild-type fibroblasts significantly increase the rate of tumor growth. When HGF mRNA was eliminated, however, the transduced fibroblasts lost their ability to promote tumor growth. The reduction of tumor volume was reflected in the tumor weight at the end of the experiment. We view these data to be of particular importance, because it is well documented that mammary stromal fibroblasts are the largest donors for increased levels of HGF in tumor tissues and in the circulation of patients with mammary cancer. These levels are strongly associated with prognosis in these patients (23, 24). HGF produced by tumor-associated fibroblasts is therefore a key factor in the stimulation of mammary cancer growth and invasion (25, 26, 27, 28, 29). Our data provide evidence for the first time that targeting HGF in stromal fibroblasts using U1/ribozyme transgene is effective in reducing its stimulation to tumor growth.

HGF is a known powerful angiogenic factor (6, 7, 8, 30, 31). In the current study, we first demonstrated that MRC5, when coinjected with mammary cancer cells, increased the number of VE-cadherin-positive vessels in tumors. In line with the antitumor effect, MRC5 transduced with pLXSN-HGF/SF701 (MRC5−/−) lost its capability to stimulate the formation of blood vessels, compared with its wild type. This, together with the in vitro bioassay data, indicates that fibroblasts are a rich donor of HGF to mammary cancer cells and that this HGF is a powerful stimulus to angiogenesis. The reduction of tumor growth in groups with transduced fibroblasts can therefore be partly contributed to by a reduction of angiogenesis in these settings.

Given the potential effects of HGF on invasiveness of mammary cancer cells, we anticipated that cancer cells, after coinjection with MRC5 cells, may have increased its micro and/or macrometastases. However, the current study failed to detect any existence of metastases in the lungs, circulation, and bone marrow. This indicates that the current model and cells used here are not suitable for assessing metastasis, as indicated in recent studies (22). The experimental length (≤4 weeks) is probably not long enough for metastasis to occur. We are currently using a systemic metastasis model to examine the antimetastasis function of these transgenes.

This study chose a retroviral vector as the delivery vehicle for the transgenes, by taking advantage of its ability to produce long-term integration of retroviral DNA in target cells, thus creating stably transduced cells. However, retroviral vectors may not be ideal for a therapeutic approach, because of the possible hazardous effect of the vector. This is supported partly by the current study in that cells transduced with control vectors have a marginal, although not statistically significant, increase in secretion of bioactive HGF (MRC5+/−) in vitro (Fig. 2) and tumor growth (MDA+/−) in vivo (Fig. 4). From this point of view, adenoviral delivery of such transgenes may be more attractive, although adenoviral vectors are known for their transient integration into genomic DNA and transient expression. The development of adenoviral vectors for the ribozyme transgenes has been recently reported (32). The adenoviral transgenes exhibited transient inhibition of the expression of HGF and its receptor in glioma cells. Intratumor and systemic injections of active adenoviral transgene stock have been shown to reduce the growth of glioma tumors and increase the survival rate of experimental mice (32).

Taken together, the current study and recent studies provide evidence that viral vectors are effective tools in delivering ribozyme-associated transgenes. Governed by the administrative regulations on animal studies, the current in vivo study was limited to an end point using tumor growth of certain size. However, the study still provides direct evidence that mammary cancer cells whose HGF receptor is eliminated by the retroviral ribozyme transgene have significantly reduced tumor growth. In addition, fibroblasts, which have lost the expression of HGF mRNA and HGF protein as a result of the HGF transgene, have impaired stimulation on mammary tumor growth. Furthermore, as the role of HGF and its receptor in mammary cancer progression in clinical settings (8, 9, 10, 11, 12, 13, 14, 33) becomes clear, we would argue that using the ribozyme transgenic approach as described here and in other studies (17, 18, 32) may be of therapeutic value.

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.

1

Supported by the Breast Cancer Campaign and the Susan G. Komen Breast Cancer Foundation.

3

The abbreviations used are: HGF, hepatocyte growth factor; SF, scatter factor; rhHGF, recombinant human hepatocyte growth factor; RT-PCR, reverse transcription-PCR.

Fig. 1.

The effects of retroviral transgenes on the expression of HGF in MRC5 cells (top) and the HGF receptor, c-MET, in MDA MD 231 (bottom). Cells were transduced with active viral stocks. RNA was extracted from stably transduced cells. Wild-type MRC5 (MRC5−/−) expressed high levels of HGF mRNA. After transduction with HGF ribozyme transgene (MRC5+/+), the cell significantly reduced expression of HGF mRNA, compared with those transduced with control transgene (MRC5+/−). MDA MB 231 mammary cancer cells expressed HGF receptor mRNA. A reduction of the expression from transduced MDA cells (MDA+/+) was seen. However, mammary cancer cells transduced with control vector (MDA+/−) did not display a significant effect.

Fig. 1.

The effects of retroviral transgenes on the expression of HGF in MRC5 cells (top) and the HGF receptor, c-MET, in MDA MD 231 (bottom). Cells were transduced with active viral stocks. RNA was extracted from stably transduced cells. Wild-type MRC5 (MRC5−/−) expressed high levels of HGF mRNA. After transduction with HGF ribozyme transgene (MRC5+/+), the cell significantly reduced expression of HGF mRNA, compared with those transduced with control transgene (MRC5+/−). MDA MB 231 mammary cancer cells expressed HGF receptor mRNA. A reduction of the expression from transduced MDA cells (MDA+/+) was seen. However, mammary cancer cells transduced with control vector (MDA+/−) did not display a significant effect.

Close modal
Fig. 2.

The effect of viral ribozymes on the level of HGF and cMET and activation of cMET. A, Western blotting of cMET (top) and HGF (bottom) from MDA MB 231 and MRC5 cells, respectively. pLXSN-MET- and pLXSN-HGF-transduced cells have lost the protein expression (MDA+/+ and MRC5+/+). B, HGF bioassay of MRC5 supernatant. Wild-type MRC5 (MRC5−/−) secreted high levels of bioactive HGF. However, after transduction with HGF transgene (MRC5+/+), the secretion was significantly reduced. Cells transduced with control vector (MRC5+/−) exhibited a small, but insignificant effect on the secretion. C, cMET activation (tyrosine phosphorylation) by HGF and serum-free conditioned medium from MRC5 cells. Anti-cMET antibody was used for immunoprecipitation. The immunoprecipitates were probed with an antiphosphotyrosine antibody (PY99, top panel) or anti-cMET antibody (bottom panel). Although both MDA−/− (bottom left) and MDA+/− (bottom middle) exhibited good levels of cMET protein, MDA+/+ has almost completely lost cMET (bottom right). The figure further showed that cMET in MDA−/− (top left) and MDA+/− (bottom middle) can be activated by rhHGF and conditioned medium from MRC5−/− and MRC5+/− cells. In contrast, conditioned medium from MRC5+/+ did not display any effect on cMET phosphorylation.

Fig. 2.

The effect of viral ribozymes on the level of HGF and cMET and activation of cMET. A, Western blotting of cMET (top) and HGF (bottom) from MDA MB 231 and MRC5 cells, respectively. pLXSN-MET- and pLXSN-HGF-transduced cells have lost the protein expression (MDA+/+ and MRC5+/+). B, HGF bioassay of MRC5 supernatant. Wild-type MRC5 (MRC5−/−) secreted high levels of bioactive HGF. However, after transduction with HGF transgene (MRC5+/+), the secretion was significantly reduced. Cells transduced with control vector (MRC5+/−) exhibited a small, but insignificant effect on the secretion. C, cMET activation (tyrosine phosphorylation) by HGF and serum-free conditioned medium from MRC5 cells. Anti-cMET antibody was used for immunoprecipitation. The immunoprecipitates were probed with an antiphosphotyrosine antibody (PY99, top panel) or anti-cMET antibody (bottom panel). Although both MDA−/− (bottom left) and MDA+/− (bottom middle) exhibited good levels of cMET protein, MDA+/+ has almost completely lost cMET (bottom right). The figure further showed that cMET in MDA−/− (top left) and MDA+/− (bottom middle) can be activated by rhHGF and conditioned medium from MRC5−/− and MRC5+/− cells. In contrast, conditioned medium from MRC5+/+ did not display any effect on cMET phosphorylation.

Close modal
Fig. 3.

The effect of HGF and MRC5 cells on the in vitro invasiveness of mammary cancer cells in a Matrigel invasion assay. Recombinant HGF and wild-type MRC5 (MRC5−/−) significantly increased the invasiveness of wild-type mammary cancer cells. However, MDA+/+ cells have lost their response to HGF and MRC5. In addition, MRC5+/+ cells have reduced its stimulation to mammary cancer cells. ∗, P < 0.05 versus control.

Fig. 3.

The effect of HGF and MRC5 cells on the in vitro invasiveness of mammary cancer cells in a Matrigel invasion assay. Recombinant HGF and wild-type MRC5 (MRC5−/−) significantly increased the invasiveness of wild-type mammary cancer cells. However, MDA+/+ cells have lost their response to HGF and MRC5. In addition, MRC5+/+ cells have reduced its stimulation to mammary cancer cells. ∗, P < 0.05 versus control.

Close modal
Fig. 4.

The growth of mammary tumor. MDA MB 231 cells exhibited a rapid growth in mice tumor model. Tumors from cells transduced with MET ribozyme transgene (MDA+/+) displayed a significantly slower growth, compared with that from MDA−/− and MDA+/−. Left, tumor volume; right, tumor weight at the conclusion of the experiments.

Fig. 4.

The growth of mammary tumor. MDA MB 231 cells exhibited a rapid growth in mice tumor model. Tumors from cells transduced with MET ribozyme transgene (MDA+/+) displayed a significantly slower growth, compared with that from MDA−/− and MDA+/−. Left, tumor volume; right, tumor weight at the conclusion of the experiments.

Close modal
Fig. 5.

The effects of MRC5 on the growth of wild-type mammary cancer cells. Coinjection of wild-type MRC5 (MRC5−/−) accelerated the growth of mammary tumor over the period. However, in the group which received coinjection of MRC5+/+, the growth was retarded, compared with that of MRC5−/−. Dashed line, MDA−/− only. Right, tumor weight.

Fig. 5.

The effects of MRC5 on the growth of wild-type mammary cancer cells. Coinjection of wild-type MRC5 (MRC5−/−) accelerated the growth of mammary tumor over the period. However, in the group which received coinjection of MRC5+/+, the growth was retarded, compared with that of MRC5−/−. Dashed line, MDA−/− only. Right, tumor weight.

Close modal
Fig. 6.

Elimination of the HGF receptor from MDA MB 231 cells rendered tumors to lose response to fibroblasts. MDA+/+ cells displayed a similar response to MRC5+/+ cells and MRC−/− cells.

Fig. 6.

Elimination of the HGF receptor from MDA MB 231 cells rendered tumors to lose response to fibroblasts. MDA+/+ cells displayed a similar response to MRC5+/+ cells and MRC−/− cells.

Close modal
Fig. 7.

Quantitation of angiogenesis. Coinjection with MRC5−/− cells increased the number of vessels in tumor (∗, P = 0.049, versus MDA−/−). The number of vessels was significantly reduced when MRC5+/+ cells were injected [P < 0.01 versus MDA−/− only (∗) and MDA−/− + MRC5−/− (∗∗)]. MDA+/+ exhibited a significant reduction of angiogenesis compared with wild-type MDA MB 231 (ε, P < 0.01 versus MDA−/−). The combination of MDA+/+ and MRC5+/+ exhibited a lower number of vessels (δ and δδ, P < 0.01 versus respective wild types).

Fig. 7.

Quantitation of angiogenesis. Coinjection with MRC5−/− cells increased the number of vessels in tumor (∗, P = 0.049, versus MDA−/−). The number of vessels was significantly reduced when MRC5+/+ cells were injected [P < 0.01 versus MDA−/− only (∗) and MDA−/− + MRC5−/− (∗∗)]. MDA+/+ exhibited a significant reduction of angiogenesis compared with wild-type MDA MB 231 (ε, P < 0.01 versus MDA−/−). The combination of MDA+/+ and MRC5+/+ exhibited a lower number of vessels (δ and δδ, P < 0.01 versus respective wild types).

Close modal
Table 1

Growth of MDA MB 231 and MRC5 cells

Shown are absorbance after staining cells with crystal violet.

Cells testedDay 0Day 1Day 3Day 5
MDA−/− 0.03 ± 19.6 13.3 ± 25.8 93.3 ± 35.6 486 ± 153 
MDA+/− 0.01 ± 20 30.0 ± 33 98.0 ± 22.2 592.7 ± 100.9 
MDA+/+ 0.00 ± 12.4 47.3 ± 53.7 96.4 ± 48.0 663.5 ± 171.2 
MRC5−/− 0.00 ± 18.0 36.9 ± 13.1 87.5 ± 47.5 589.8 ± 130 
MRC5+/− 0.03 ± 14.2 3.7 ± 13.3 51.3 ± 2.2 529.7 ± 137.6 
MRC5+/+ 0.00 ± 3.8 7.0 ± 5.8 60.0 ± 11.8 577.5 ± 143.2 
Cells testedDay 0Day 1Day 3Day 5
MDA−/− 0.03 ± 19.6 13.3 ± 25.8 93.3 ± 35.6 486 ± 153 
MDA+/− 0.01 ± 20 30.0 ± 33 98.0 ± 22.2 592.7 ± 100.9 
MDA+/+ 0.00 ± 12.4 47.3 ± 53.7 96.4 ± 48.0 663.5 ± 171.2 
MRC5−/− 0.00 ± 18.0 36.9 ± 13.1 87.5 ± 47.5 589.8 ± 130 
MRC5+/− 0.03 ± 14.2 3.7 ± 13.3 51.3 ± 2.2 529.7 ± 137.6 
MRC5+/+ 0.00 ± 3.8 7.0 ± 5.8 60.0 ± 11.8 577.5 ± 143.2 
Table 2

Proportion of fibroblasts and proliferation index in tumor tissues

Shown are percentage of fibroblasts (vimentin-positive cells) of all cells assessed and percentage of Ki67 cells, respectively.

MDA−/− onlyMDA+/− onlyMDA+/+ onlyMDA−/− plus MRC5−/−MDA−/− plus MRC5+/+
Fibroblasts 2.88 ± 1.57% 2.77 ± 1.42% 1.79 ± 1.86% 3.08 ± 2.4% 2.13 ± 1.35% 
Ki67 staining 14.13 ± 10.6% 11.0 ± 3.9% 5.13 ± 1.36%a 33.0 ± 4.4%b 17.3 ± 6.94%c 
MDA−/− onlyMDA+/− onlyMDA+/+ onlyMDA−/− plus MRC5−/−MDA−/− plus MRC5+/+
Fibroblasts 2.88 ± 1.57% 2.77 ± 1.42% 1.79 ± 1.86% 3.08 ± 2.4% 2.13 ± 1.35% 
Ki67 staining 14.13 ± 10.6% 11.0 ± 3.9% 5.13 ± 1.36%a 33.0 ± 4.4%b 17.3 ± 6.94%c 
a

P < 0.05 vs. MDA−/− and MDA+/−.

b

P < 0.01 vs. MDA−/−.

c

P < 0.01 vs. MDA−/− plus MRC5−/−.

1
Nakamura T., Nishizawa T., Hagiya M., Seki T., Shimonishi M., Sugimura A., Tashiro K., Shimizu S. Molecular cloning and expression of human hepatocyte growth factor.
Nature (Lond.)
,
342
:
440
-443,  
1989
.
2
Gherardi E., Stoker M. Hepatocyte and scatter factor.
Nature (Lond.)
,
346
:
228
1990
.
3
Jiang W. G., Lloyd D., Puntis M. C. A., Nakamura T., Hallett M. B. Regulation of spreading and growth of human colon cancer cells by hepatocyte growth factor.
Clin. Exp. Metastasis
,
11
:
235
-242,  
1993
.
4
Matsumoto K., Nakamura T., Kramer R. H. Hepatocyte growth factor/scatter factor induces tyrosine phosphorylation of focal adhesion kinase (p125FAK) and promotes migration and invasion by oral squamous cel carcinoma cells.
J. Biol. Chem.
,
269
:
31807
-31813,  
1994
.
5
Jiang W. G., Hiscox S., Matsumoto K., Nakamura T. Hepatocyte growth factor/scatter factor, its molecular and cellular and clinical implications in cancer.
Crit. Rev. Oncol.-Hematol.
,
29
:
209
-248,  
1999
.
6
Bussolino F., DiRenzo M. F., Ziche M., Bocchietto E., Olivero M., Naldini L., Gaudino G., Tamagnone L., Coffer A., Comoglio P. M. Hepatocyte growth factor is a potent angiogenic factor which stimulates endothelial motility and growth.
J. Cell Biol.
,
119
:
629
-641,  
1992
.
7
Grant D. S., Kleinman H. K., Goldberg I. D., Bhargava M. M., Nickoloff B. J., Kinsella J. L., Polverini P., Rosen E. M. Scatter factor induces blood vessel formation in vivo.
Proc. Natl. Acad. Sci. USA
,
90
:
1937
-1941,  
1993
.
8
Yamashita J., Ogawa M., Yamashita S., Nomura K., Kuramoto M., Saishoji T., Shin S. Immunoreactive hepatocyte growth factor is a strong and independent predictor of recurrence and survival in human mammary cancer.
Cancer Res.
,
54
:
1630
-1633,  
1994
.
9
Ghoussoub R. A. D., Dillon D. A., D’Aquila T., Rimm E. B., Fearon E. R., Rimm D. L. Expression of c-met is a strong independent prognostic factor in mammary carcinoma.
Cancer (Phila.)
,
82
:
1513
-1520,  
1998
.
10
Camp R. L., Rimm E. B., Rimm D. L. Met expression is associated with poor outcome in patients with axillary lymph node negative mammary carcinoma.
Cancer (Phila.)
,
86
:
2259
-2265,  
1999
.
11
Ueno T., Toi M., Taniguchi T., Tominaga T. Serum hepatocyte growth factor levels in mammary cancer patients with distant metastases.
Cancer J.
,
11
:
147
-152,  
1998
.
12
Maemura M., Lino Y., Yokoe T., Horiguchi J., Takei H., Koibuchi Y., Horii Y., Takeyoshi I., Ohwada S., Morishita Y. Serum concentration of hepatocyte growth factor in patients with metastatic mammary cancer.
Cancer Lett.
,
126
:
215
-220,  
1998
.
13
Toi M., Taniguchi T., Ueno T., Asano M., Funata N., Sekiguchi K., Iwanari H., Tominaga T. Significance of circulating hepatocyte growth factor level as a prognostic indicator in primary mammary cancer.
Clin. Cancer Res.
,
4
:
659
-664,  
1998
.
14
Taback B., Chan A. D., Kuo C. T., Bostick P. J., Wang H. J., Giuliano A. E., Hoon D. S. B. Detection of occult metastatic mammary cancer cells in blood by a multimolecular marker assay: correlation with clinical stage of disease.
Cancer Res.
,
61
:
8845
-8850,  
2001
.
15
Tsarfaty I., Alvord W. G., Resau J. H., Altstock R. T., Lidereau R., Bieche I., Bertrand F., Horev J., Klabansky R. L., Keydar I., Vande Woude G. F. Alteration of Met protooncogene product expression and prognosis in mammary carcinomas.
Analyt. Quant. Cytology Histology
,
21
:
397
-408,  
1999
.
16
Altstock R. T., Stein G. Y., Resau J. H., Tsarfaty I. Algorithms for quantitation of protein expression variation in normal versus tumor tissue as a prognostic factor in cancer: Met oncogene expression, and mammary cancer as a model.
Cytometry
,
41
:
155
-156,  
2000
.
17
Abounader R., Ranganathan S., Lal B., Fielding K., Book A., Dietz H., Burger P., Laterra J. Reversion of human glioblastoma malignancy by U1 small nuclear RNA/ribozyme targeting of scatter factor/hepatocyte growth factor and c-met expression.
J. Natl. Cancer Inst. (Bethesda)
,
91
:
1548
-1556,  
1999
.
18
Jiang W. G., Grimshaw D., Leterra J., Abounader R., Lane J., Mansel R. E. Supression of invasiveness and cellular motility of mammary cancer cells by a hammerhead ribozyme encoding cMET antisense.
Clin. Cancer Res.
,
7
:
2555
-2562,  
2001
.
19
Jiang W. G., Hiscox S., Hallett M. B., Horrobin D. F., Mansel R. E., Puntis M. C. A. Regulation of the expression of E-cadherin on human cancer cells by gamma linolenic acid.
Cancer Res.
,
55
:
5043
-5048,  
1995
.
20
Albini A., Iwanoto Y., Kleinman H. K., Martin G. R., Aaronson S. A., Kozlowski J. M., McEwan R. N. A rapid in vitro assay for quantitating the invasive potential of tumor cells.
Cancer Res.
,
47
:
3239
-3245,  
1987
.
21
Jiang W. G., Puntis M. C. A., Hallett M. B. Monocyte conditioned media possess a novel factor which increases motility of cancer cells.
Int. J. Cancer
,
53
:
426
-431,  
1993
.
22
Martin T. A., Parr C., Davies G., Watkins G., Matsumot K., Nakamura T., Mansel R. E., Jiang W. G. NK4 reduced HGF/SF and fibroblast induced breast tumour growth and angiogenesis.
Carcinogenesis
,
24
:
1317
-1325,  
2003
.
23
Nakamura T., Matsumoto K., Kiritoshi A., Tano Y., Nakamura T. Induction of hepatocyte growth factor in fibroblasts by tumor-derived factors affects invasive growth of tumor cells: in vitro analysis of tumor-stromal interactions.
Cancer Res.
,
57
:
3305
-3313,  
1997
.
24
Seslar S., Namakura T., Byers S. Regulation of fibroblast hepatocyte growth factor/scatter factor expression by human mammary carcinoma cell lines and peptide growth factors.
Cancer Res.
,
53
:
1233
-1238,  
1993
.
25
Moinfar F., Man Y. G., Arnould L., Bratthauer G. L., Ratschek M., Tavassoli F. A. Concurrent and independent genetic alterations in the stromal and epithelial cells of mammary carcinoma: Implications for tumorigenesis.
Cancer Res.
,
60
:
2562
-2566,  
2000
.
26
Rosen E. M., Knesel J., Goldberg I. D., Jin L., Bhargava M., Joseph A., Zitnik R., Wines J., Kelley M., Rockwell S. Scatter factor modulates the metastatic phenotype of the emt6 mouse mammary-tumor.
Int. J. Cancer
,
57
:
706
-714,  
1994
.
27
Sasaki M., Nishio M., Tsukada Y., Enami J. Mammary fibroblast-derived hepatocyte growth factor stimulates growth and morphogenesis of mouse mammary tumor cells in primary culture.
Jpn. J. Cancer Res.
,
89
:
1134
-1143,  
1998
.
28
Sasaki M., Enami J. Mammary fibroblast-derived hepatocyte growth factor and mammogenic hormones stimulate the growth of mouse mammary epithelial cells in primary culture.
Endocr. J.
,
46
:
359
-366,  
1999
.
29
Matsumoto-Taniura N., Matsumoto K., Nakamura T. Prostaglandin production in mouse mammary tumor cells confers invasive growth potential by inducing hepatocyte growth factor in stromal fibroblasts.
Br. J. Cancer
,
81
:
194
-202,  
1999
.
30
Jiang W. G., Hiscox S., Cai J., Martin T., Matsumoto K., Nakamura T., Mansel R. E. Antagonistic effects of NK4, an novel HGF variant, on the in vitro angiogenesis of human vascular endothelial cells.
Clin. Cancer Res.
,
5
:
3695
-3703,  
1999
.
31
Nagy J., Curry G. W., Hillan K. J., Mallon E., Purushotham A. D., George W. D. Hepatocyte growth factor scatter factor, angiogenesis and tumor cell proliferation in primary mammary cancer.
Mammary
,
5
:
105
-109,  
1996
.
32
Abounader R., Lal B., Luddy C., Koe G., Davidson B., Rosen E. M., Laterra J. In vivo targeting of SF/HGF and c-met expression via U1snRNA/ribozymes inhibits glioma growth and angiogenesis and promotes apoptosis.
FASEB J.
,
13
:
22
-37,  
2001
.
33
Taniguchi T., Toi M., Inada K., Imazawa T., Yamamoto Y., Tominaga T. Serum concentration of hepatocyte growth factor in mammary cancer patients.
Clin. Cancer Res.
,
1
:
1031
-1034,  
1995
.