We determined whether tumor cells consistently generating granulocyte/macrophage colony– stimulating factor (GM-CSF) can recruit and activate macrophages to generate angiostatin and, hence, inhibit the growth of distant metastasis. Two murine melanoma lines, B16-F10 (syngeneic to C57BL/6 mice) and K-1735 (syngeneic to C3H/HeN mice), were engineered to produce GM-CSF. High GM-CSF (>1 ng/106 cells)– and low GM-CSF (<10 pg/106 cells)–producing clones were identified. Parental, low, and high GM-CSF–producing cells were injected subcutaneously into syngeneic and into nude mice. Parental and low-producing cells produced rapidly growing tumors, whereas the high-producing cells produced slow-growing tumors. Macrophage density inversely correlated with tumorigenicity and directly correlated with steady state levels of macrophage metalloelastase (MME) mRNA. B16 and K-1735 subcutaneous (s.c.) tumors producing high levels of GM-CSF significantly suppressed lung metastasis of 3LL, UV-2237 fibrosarcoma, K-1735 M2, and B16-F10 cells, but parental or low-producing tumors did not. The level of angiostatin in the serum directly correlated with the production of GM-CSF by the s.c. tumors. Macrophages incubated with medium conditioned by GM-CSF– producing B16 or K-1735 cells had higher MME activity and generated fourfold more angiostatin than control counterparts. These data provide direct evidence that GM-CSF released from a primary tumor can upregulate angiostatin production and suppress growth of metastases.

The progressive growth and spread of neoplasms are dependent on the formation of adequate vasculature, i.e., angiogenesis (1). The extent of angiogenesis depends on the balance between positive and negative regulating molecules released by both tumor cells and host cells, e.g., lymphoid cells (26). Angiogenesis begins with local degradation of the basement membrane of capillaries, followed by invasion of the stroma by underlying endothelial cells in the direction of an angiogenic stimulus. Subsequent to migration, endothelial cells proliferate at the leading edge of a migrating column and then organize into three-dimensional structures to form new capillary tubes (26). Interruption of any one of these steps can inhibit angiogenesis (7) and thus metastasis.

Angiostatin is a specific inhibitor of endothelial cell proliferation originally purified from the serum and urine of mice bearing a murine Lewis lung carcinoma (3LL; reference 8). It is an internal fragment of plasminogen containing the first four triple loop disulfide-linked structures, with an apparent molecule mass of 38 kD (8). Angiostatin isolated from serum and urine of mice or from the proteolytic degradation of human plasminogen is a potent inhibitor of angiogenesis and, hence, of tumor growth in vivo (8, 9). Proteolytic degradation of plasminogen by elastase (10), urokinase plasminogen activator in PC3 human prostate carcinoma cells (11, 12), human matrilysin, and human gelatinase B (13) results in the production of angiostatin.

We have reported recently that the production of angiostatin by 3LL tumors growing subcutaneously (s.c. tumors) is associated with the upregulation of murine metalloelastase (MME)1 in tumor-infiltrating macrophages (10). Further, we concluded that MME induced by GM-CSF can cleave plasminogen to angiostatin (10). Since MME expression in macrophages is upregulated by GM-CSF (14), we wished to determine whether we could engineer tumor cells to produce high levels of GM-CSF to recruit macrophages and upregulate expression of MME, which would lead to production of angiostatin and thus to inhibition of the outgrowth of cancer metastases.

Mice.

Athymic Ncr-nu/nu male mice were purchased from the Animal Production Area, National Cancer Institute-Frederick Cancer Research Facility (Frederick, MD). Specific pathogen– free female C57BL/6 and C3H/HeN mice were purchased from The Jackson Laboratory (Bar Harbor, ME). All mice were maintained under specific pathogen–free conditions in a facility approved by the American Association for Accreditation of Laboratory Animal Care. The care and experimental procedures were in accordance with institutional guidelines and current regulations and standards of the United States Department of Agriculture, the Department of Health and Human Services, and the National Institutes of Health. The mice were used when they were 8–12 wk old.

Tumor Lines.

Metastatic murine Lewis lung carcinoma syngeneic to C57BL/6 mice (3LL-met; obtained from Dr. M. O'Reilly, Harvard Medical School, Boston, MA [8]), B16-F10 GM-CSF, a variant of the metastatic B16-F10 melanoma syngeneic to C57BL/6 mice (15) transduced using an MFG retroviral vector encoding murine GM-CSF (obtained from Dr. D. Pardoll, Johns Hopkins University, Baltimore, MD), metastatic K-1735 M2 melanoma syngeneic to C3H/HeN mice (16), UV-2237M fibrosarcoma syngeneic to C3H/HeN mice (17), and murine RENCA renal carcinoma syngeneic to BALB/c mice (18) were maintained as monolayer cultures in Eagle's MEM supplemented with sodium pyruvate, nonessential amino acids, l-glutamine, vitamin solution (GIBCO BRL, Gaithersburg, MD), and 5% (vol/ vol) heat-inactivated fetal bovine serum (FBS) at 10% CO2/95% air at 37°C. The K-1735 M2 melanoma cells were transfected with the plasmid pcDNA3-GM-CSF encoding for murine GM-CSF driven by human cytomegalovirus promoter (a gift of Dr. W. Yang, Academia Sinica, Taiwan) to derive K-1735 M2 GM-CSF cells. K-1735 M2 cells transfected with pcDNA3 encoding for neomycin resistance gene (K-1735 M2-Neo) served as a control. Cells resistant to 3 wk of continuous exposure to 800 μg/ml G418 were selected.

B16-F10 GM-CSF and K-1735 M2 GM-CSF cells were subcloned by limited dilution. Clones producing low (L) or high (H) levels of GM-CSF were isolated as described below. Bovine capillary endothelial cells (BCE; obtained from Dr. M. O'Reilly [8]) were cultured on gelatinized plastic in DMEM/10% FBS supplemented with 3 ng/ml human basic fibroblast growth factor (bFGF; Genzyme Corp., Cambridge, MA). All cultures were free of mycoplasma and the pathogenic murine viruses reovirus type 3, pneumonia virus of mice, mouse adenovirus, murine hepatitis virus, lymphocyte choriomeningitis virus, ectromelia virus, and lactate dehydrogenase virus (assayed by Program Resources, Inc., NCI-Frederick Cancer Research and Development Center, Frederick, MD).

Measurement of GM-CSF Protein.

Tumor cells (105) were plated into 30-mm2 plates in medium containing 5% FBS. After 12 h, the cultures were washed with serum-free medium and then cultured with 2 ml of medium containing 5% FBS for 24 h. The level of GM-CSF in the culture supernatants was determined using an ELISA kit according to the manufacturer's instructions (Quantikine; R&D Systems, Inc., Minneapolis, MN).

In Vivo Studies.

Cultures of wild-type and transfected tumor cells in their exponential growth phase were harvested by a brief trypsinization, washed in medium containing 10% FBS, and resuspended in HBSS. Different numbers of cells (0.5–4 × 105 in 0.1 ml HBSS) were injected into the subcutis of the right flank of syngeneic or nude mice. When tumors reached 8–10 mm in diameter, unanesthetized syngeneic or nude mice were injected intravenously with 105 K-1735 M2 melanoma cells or 105 UV-2273M cells. In another set of experiments, nude mice bearing s.c. B16-F10 tumors were anesthetized with methoxyflurane. The tumors in one group of mice were surgically excised, and the area was closed with metal wound clips. The other group underwent a sham surgical procedure. 1 d later, these mice were injected intravenously with 105 K-1735 M2 cells. The mice with no s.c. tumors received intravenous injections of 3LL-met; K-1735 M2 cells served as additional controls. The mice were monitored daily, killed 3 wk later, and necropsied. The lungs were weighed and fixed in Bouin's solution. The experimental lung metastases were counted using a dissecting microscope.

To determine whether s.c. tumors that produce GM-CSF can influence the take and growth of tumor cells implanted at another subcutaneous site, we injected B16-F10 GM-CSF cells from 8–10-mm tumors into the subcutis on the right flank of nude mice. K-1735 M2 cells (105) were injected into the subcutis of the contralateral flank. Tumor size was determined twice weekly using a caliper. Tumor volume was calculated by the formula, V = (A × B2)/2, where V = volume, A = long diameter, and B = short diameter.

Collection and Cultivation of Mouse Peritoneal Exudate Macrophages.

Peritoneal exudate macrophages (PEM) were collected by peritoneal lavage with HBSS of mice injected intraperitoneally with 1.5 ml thioglycollate broth 4 d previously. The PEM were plated in serum-free medium for 2 h on a 24-well dish at a density of 2 × 105 cells/well. 2 h later, nonadherent cells were removed. The resultant adherent population was 98% pure according to immunologic, morphologic, and phagocytic criteria (19).

Sample Preparations for Elastase and Angiostatin Assays.

To collect tumor cell culture supernatants, tumor cells were plated at a density of 2 × 106/75 cm2, flasked, and cultured for 48 h in medium containing 5% FBS. The culture supernatant fluids were collected, centrifuged at 3,000 rpm for 10 min to remove intact cells, aliquoted, and stored at −80°C until use. For the elastase and angiostatin assays, PEM were incubated for 24 h with tumor cell culture supernatants diluted in medium containing 5% FBS. The PEM were washed and incubated for another 72 h in 0.5 ml/well of serum-free DMEM/F12 medium in the absence or presence of 100–200 μg/ml human plasminogen purified from human plasma using a lysine–Sepharose 4B column (Pharmacia Biotech AB, Uppsala, Sweden). The culture supernatants were centrifuged at 3,000 g for 30 min at 4°C. Elastase or angiostatin activities were assessed as described below.

Elastase Assay.

Elastinolytic activity in macrophage-conditioned medium was determined by a method described previously using [3H]NaBH4-labeled elastin as a substrate (20). In brief, the samples were incubated for 16 h at 37°C with 600 μg of the radiolabeled elastin in a reaction buffer (100 mM Tris-HCl, 5 mM CaCl2, 0.2 mg/ml SDS, and 0.006% NaH3). Released (free-form) 3H-peptide was harvested by centrifugation, and radioactivity was monitored by scintillation counting. Enzyme activity was expressed as cpm per reaction.

Angiostatin Assay.

Angiostatin activity was assessed by inhibition of BCE proliferation as described previously (8). In brief, BCE seeded at a density of 1.25 × 104 cells/gelatinized well/0.5 ml DMEM containing 10% FBS were allowed to adhere overnight, then were rinsed and incubated for 20 min with 0.25 ml/well DMEM/5% FBS or test samples. Additional medium containing bFGF was added to a final concentration of 1 ng/ml. 72 h later, the BCE were harvested by trypsinization and counted using a hemacytometer. Angiostatin activity was determined from the formula for inhibition of BCE proliferation and was expressed as a percentage.

Northern Blot Analysis.

Poly(A)+ mRNA was extracted from 107 tumor cells at 70% confluence or from tumors growing in vivo using the FastTrack mRNA isolation kit (Invitrogen Corp., San Diego, CA). mRNA was electrophoresed on a 1% denaturing formaldehyde/agarose gel, electrotransferred at 0.6 A to GeneScreen nylon membrane (DuPont-NEN, Boston, MA), and UV cross-linked with 120,000 μJ/cm2 using a UV Stratalinker 1800 (Stratagene Inc., La Jolla, CA). Hybridization was performed as described previously (21). The nylon filters were washed three times at 55–60°C with 30 mM NaCl/3 mM sodium citrate (pH 7.2)/ 0.1% SDS (wt/vol). The cDNA probes used in this analysis were a 1.3-kb PstI cDNA fragment corresponding to rat glyceraldehyde 3-phosphate dehydrogenase (GAPDH; reference 22), a 1.7-kb bamboo cDNA fragment of MME (provided by Dr. S.D. Shapiro, Jewish Hospital at Washington University School of Medicine, St. Louis, MO [23]), and a 1.0-kb PstI cDNA fragment of murine GM-CSF. The cDNA fragments were radiolabeled using random primer with α-32P–labeled deoxyribonucleotide triphosphates (Amersham Corp., Arlington Heights, IL). mRNA expression was quantified on a laser densitometer (Ultrascan XL; LKB, Uppsala, Sweden). Each sample measurement was calculated as the ratio of the average area of the GM-CSF or MME to that of the GAPDH transcript and then standardized using the smallest ratio in each group of samples.

Western Blot Analysis.

Samples isolated from culture supernatants were mixed with sample buffer (62.5 mM Tris-HCl [pH 6.8], 2.3% SDS, 100 mM dithiothreitol, and 0.05% bromphenol blue), boiled, and separated on 10% SDS-PAGE. The protein was transferred onto 0.45-μm nitrocellulose membranes. The filter was blocked with 3% BSA in Tris-buffered saline (20 mM Tris-HCl [pH 7.5], 150 mM NaCl), probed with antibody against plasminogen lysine-binding site (LBS)-1 (1 μg/ml) in Tris-buffered saline containing 0.1% Tween 20, incubated with a second antibody in the buffer, and visualized by the enhanced chemiluminescence Western blotting detection system (24).

Immunohistochemistry.

The presence of macrophages in tumors was identified by immunohistochemistry using the macrophage-specific scavenger receptor antibody (25). At necropsy, tumors were cut into 5-mm3 pieces, placed in OCT compound (Miles Inc., Elkhart, IN), and snap frozen in liquid nitrogen. Frozen sections (8–10 μm) were air-dried, fixed with cold acetone, rinsed with PBS, and treated with 3% hydrogen peroxide in methanol (vol/vol). The treated slides were incubated in a blocking solution (5% normal human serum/1% normal goat serum in PBS [vol/vol]) and then for 18 h with a rat polyclonal antibody against scavenger receptor (Serotec Ltd., Kidlington, Oxford, UK) at a 1:70 dilution at 4°C in a humidified chamber. The sections were then rinsed and incubated, first with the blocking solution and then with a peroxidase-conjugated anti–rat antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) at a 1:100 dilution. Positive reactions were visualized by incubating the slides for 1–5 min with Stable DAB (Research Genetics, Huntsville, AL), followed by counterstaining with Mayer's hematoxylin (Research Genetics). The slides were dried and mounted with Universal mount (Research Genetics). Images were digitized using a color video camera (3CCD; Sony Corp., Tokyo, Japan) and a personal computer equipped with image analysis software (Optimas Corp., Bothell, WA).

Purification of Angiostatin from Murine Serum.

Mice were bled from the lateral tail vein. Serum diluted 1:2 to 1:10 with PBS was incubated overnight with lysine–Sepharose 4B (Pharmacia Biotech AB) at 4°C. After washing with PBS and 0.3 M phosphate buffer containing 3 mM EDTA (pH 7.4), lysine-binding proteins were eluted with 0.2 M aminocaproic acid (pH 7.4). The aminocaproic acid in the eluates was removed using a spin column (Microcon 10; Amicon, Inc., Beverly, MA), and the proteins were resuspended in 20 mM Tris-HCl (pH 7.6).

Statistical Analysis.

The significance of the in vitro data, lung weight, and tumor volume was analyzed by the Student's t test (two-tailed). The difference in the number of lung metastases between groups was compared using the Mann-Whitney U test.

Establishment of GM-CSF–producing Murine Melanoma Cell Lines.

In the first set of experiments, we cloned GM-CSF–transduced B16-F10 cells (20). Several clones producing 60–75 ng GM-CSF/106 cells/24 h were combined, and the line was designated B16-H (high). Three clones producing <10 pg GM-CSF/106 cells/24 h were combined to yield the B16-L (low) line. GM-CSF–transfected K-1735 M2 cells were also cloned (see Materials and Methods). Several clones producing 1–2 ng GM-CSF/106 cells/24 h were combined to yield the K-1735-H line. Several clones producing <10 pg GM-CSF/106 cells/24 h were combined to yield the K-1735-L line. The level of GM-CSF mRNA directly correlated with the level of protein (Fig. 1, top and bottom). Southern blot analysis indicated that recombinant GM-CSF DNA was intact in all the cell lines (data not shown).

The production of GM-CSF did not alter the in vitro growth of the tumor cells. In contrast, tumorigenicity and rate of growth of s.c. tumors (in syngeneic or nude mice) inversely correlated with production of GM-CSF. B16-F10 parental or B16-L cells produced 1,800-mm3 tumors by day 25, whereas the B16-H tumors measured 321 mm3. Similar results were obtained with the K-1735 melanoma. 3 wk after subcutaneous inoculation, K-1735 M2 parental, K-1735-L, and K-1735-H cells produced 900-, 900-, and 310-mm3 tumors, respectively.

Macrophage Infiltration into s.c. Tumors and Expression of MME.

Immunohistochemical staining of cryostat sections using an mAb against the murine macrophage-specific scavenger receptor revealed that the density of tumor-infiltrating macrophages in B16-H (Fig. 2, top right) and K-1735-H (Fig. 2, bottom right) tumors was threefold that in parental (GM-CSF–negative) tumors. Specifically, the average number of macrophages/100× field in the B16-F10, B16-L, and B16-H tumors was 23, 21, and 85, respectively. In the K-1735 M2, K-1735-L, and K-1735-H s.c. tumors, the average number of macrophages/100× field was 11, 31, and 65, respectively. Northern blot analysis indicated that tumors produced by B16-H and K-1735-H cells, but not B16-L, K-1735-L, or their parental cells, expressed high levels of GM-CSF and MME mRNA (Fig. 1, top and bottom, in vivo). These results suggest that GM-CSF released by tumor cells recruits macrophages and induces MME gene expression in the tumor-infiltrating macrophages.

Suppression of Distant Tumor Growth by GM-CSF–producing s.c. Tumors.

We next determined whether the cells engineered to produce GM-CSF could affect the growth of distant tumors. Because we used tumors of different origins for the local (s.c.) and secondary (metastases) challenges, we could rule out the contribution of T cell–mediated specific immune response in suppression of secondary tumors (26– 28). Specifically, B16-L and B16-H (syngeneic to C57BL/6 mice) or K-1735-L and K-1735-H cells (syngeneic to C3H mice) were inoculated subcutaneously into syngeneic recipients. When the tumors reached 8–10 mm in diameter, 3LL-met (syngeneic to C57BL/6 mice) or UV-2237M (syngeneic to C3H mice) cells were injected intravenously into the mice bearing B16 or K-1735 tumors, respectively. In both tumor systems, the s.c. tumors produced by cells releasing high amounts of GM-CSF (but not parental or low GM-CSF–producing cells) significantly suppressed (P <0.01) the growth of experimental lung metastasis by 3LL or UV-2237M cells (Table 1).

Next, we injected B16-L or B16-H cells into the subcutis of nude mice. When the tumors reached 10 mm in diameter, the mice were anesthetized. One group (n = 10) underwent resection of the s.c. tumors, and the other (n = 10) underwent sham surgery. 1 d later, the mice were injected intravenously with 3LL-met K-1735 M2 cells. The mice were killed 21 d later and necropsied, and the lung metastases were counted under a microscope. The data shown in Table 2 demonstrate that an intact s.c. B16-H (but not B16-L) tumor significantly reduced the median number of lung metastases and their size (lung weight) produced by both 3LL-met and K-1735 M2 cells. In contrast, the median number or size of lung metastases in mice whose primary/s.c. tumor was resected did not differ from control mice.

The B16-H tumors in the subcutis also suppressed the growth of K-1735 M2 cells implanted into the contralateral flank of nude mice (Fig. 3). By day 20 after inoculation of the second tumor, the volume of the K-1735 M2 tumors was 936 (132, 724 ± 143, and 288 ± 12 mm2 in mice bearing a primary B16-P, B16-L, and B16-H s.c. tumor, respectively; P <0.01).

Angiostatin in the Serum of Tumor-bearing Mice.

A previous report from our laboratory concluded that MME induced by GM-CSF can cleave plasminogen to angiostatin (10). To correlate the presence of angiostatin with suppression of metastases, we collected serum from normal mice and mice with B16-L and B16-H s.c. tumors. The sera were enriched for lysine-binding protein by passing through a lysine–Sepharose 4B column, and the eluents were separated on 10% SDS-PAGE, blotted, and probed with an mAb against human LBS-1. Western blotting analysis demonstrated that serum from both normal mice and mice bearing B16-L tumors contained 38-kD protein(s) that reacted with the antibody (Fig. 4). The same band with significantly higher intensity was found in the serum of mice bearing B16-H tumors. These results suggest that angiostatin is present in normal mice and that its production is upregulated in mice bearing B16-H but not B16-L tumors.

Generation of Angiostatin by Cultured Macrophages.

PEM were treated for 24 h with media conditioned by wild-type or GM-CSF gene–engineered B16 or K-1735 cells. The cells were then washed and incubated for an additional 72 h in serum-free medium in the absence or presence of human plasminogen. The resultant supernatants were assessed for elastase (Fig. 5,A) and angiostatin (Fig. 5,B) activity. The level of elastinolytic activity constitutively secreted by mouse PEM was not altered by treatment with media conditioned by B16-P, K-1735-P, B16-L, or K-1735-L cells. In contrast, PEM elastinolytic activity reached 1.5–2-fold the control levels with PEM incubated with media conditioned by B16-H or K-1735-H (Fig. 5,A). Significantly higher angiostatin activity (determined by a bioassay using growth inhibition of BCE) was found in supernatants of PEM with higher MME activity and plasminogen. Specifically, the mixture derived from control PEM or PEM conditioned by medium of control tumor cells inhibited growth of BCE by 10–15%, whereas the mixture from PEM incubated with medium conditioned by cells releasing high amounts of GM-CSF inhibited BCE growth by 40–55% (Fig. 5,B). The growth inhibition of BCE was neutralized by treatment of the supernatants with antibodies against LBS-1 but not by an isotype-matched control immunoglobulin (Fig. 5,C). Western blot analysis using the antibody to LBS-1 confirmed that the mixtures derived from PEM treated by medium conditioned by B16-H or K-1735-H cells contained significantly higher levels of the 38-kD protein (Fig. 5 D).

The growth of metastases can accelerate subsequent to resection of some primary neoplasms (29, 30). This accelerated growth can be found in both autochthonous human neoplasms and experimental rodent tumors, and is independent of specific immune response (3135). Sustained tumor growth depends on an adequate supply of nutrients and thus on angiogenesis (2, 3638), which has led to efforts to block the angiogenic pathway. Recent studies have concluded that some primary neoplasms can induce the production of a breakdown product from plasminogen called angiostatin (8) that specifically inhibits proliferation of vascular endothelial cells (8, 10). By thus inhibiting the development of adequate vasculature, circulating angiostatin can inhibit the growth of metastases and contribute to dormancy of metastatic tumor cells (9, 12, 3941). One mechanism for the in vivo production of angiostatin involves the upregulation of MME in macrophages that infiltrate tumor cells that produce GM-CSF (10). Whether GM-CSF–negative tumors engineered to produce GM-CSF could thus generate production of angiostatin by infiltrating macrophages remained unclear.

The present results demonstrate that s.c. tumors of B16 melanoma and K-1735 melanoma cells engineered to produce high amounts of GM-CSF were infiltrated by macrophages, whereas parental (wild-type) or cells producing low levels of GM-CSF were not. The production of GM-CSF directly correlated with infiltrating macrophages' MME activity and production of circulating angiostatin and, hence, growth inhibition of tumors implanted at a distant subcutaneous site or growing as lung metastases. In vitro analysis indicated that treatment of macrophages with media conditioned by cells producing a high level of GM-CSF increased the secretion of MME and production of angiostatin from plasminogen. These data confirmed our previous findings (10, 14), and provide direct evidence that GM-CSF can augment angiostatin production by upregulating expression of MME in tumor-infiltrating macrophages.

Tumor cells producing a high level of GM-CSF had decreased tumorigenicity in both syngeneic and nude mice compared with control cells or cells producing low levels of GM-CSF. The reduced tumorigenicity was not restricted to that produced by tumors injected in the subcutaneous site. B16 or K-1735 cells producing high levels of GM-CSF also had a reduced incidence of experimental lung metastasis (data not shown). The decrease in tumorigenicity could well be due to tumor-infiltrating macrophages. First, the density of tumor-infiltrating macrophages directly correlated with production of GM-CSF. Second, preliminary analysis indicates that tumor cells producing high levels of GM-CSF were more susceptible to lysis mediated by macrophages than low GM-CSF–producing counterparts (data not shown).

The production of GM-CSF by tumor cells has been correlated with stimulation of antigen presenting cells and induction of tumor-specific T cell–mediated immunity (2628). However, specific immunity cannot account for the present results. We base this conclusion on results of two major experiments. First, in many of the studies, s.c. B16 tumors (producing GM-CSF) inhibited the outgrowth of syngeneic but non–cross-reactive 3LL carcinoma cells. Second, lung metastasis was inhibited in athymic nude mice when GM-CSF–producing s.c. tumors were coinjected with metastatic allogeneic tumor cells.

In summary, we provide direct evidence that upregulation of MME expression in macrophages by GM-CSF produced by local tumors can lead to generation of angiostatin and, hence, growth suppression of distant metastases. Recent data indicate that angiostatin can also be generated by reduction and proteolysis of plasmin (11, 12) and by cleavage of plasminogen by gelatinase B and matrilysin (13). Regardless of the biochemical mechanism, our data suggest that chronic administration of GM-CSF or transduction of the GM-CSF gene into tumor or normal tissue can be used to treat cancer metastasis.

The authors thank Walter Pagel for his critical editorial review, and Lola López for expert assistance in the preparation of this manuscript. We also thank Dr. Corazon D. Bucana and Mrs. Donna Reynolds for technical assistance with immunohistochemical staining.

This work was supported in part by grant CA16672 from the Cancer Center Support Core, and grant R35-CA-42107 from the National Cancer Institute, National Institutes of Health.

BCE

bovine capillary endothelial cell(s)

bFGF

basic fibroblast growth factor

FBS

fetal bovine serum

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

H

high (levels of GM-CSF– producing)

LBS-1

plasminogen lysine–binding site 1

L

low (levels of GM-CSF–producing)

MME

macrophage metalloelastase

PEM

peritoneal exudate macrophage(s)

1
Folkman
J
Angiogenesis in cancer, vascular, rheumatoid and other disease
Nat Med
1995
1
27
31
[PubMed]
2
Folkman
J
Clinical application of research on angiogenesis
N Engl J Med
1995
333
1753
1763
3
Auerbach
W
,
Auerbach
R
Angiogenesis inhibition: a review
Pharmacol Ther
1994
63
265
311
[PubMed]
4
Folkman
J
,
Klagsburn
M
Angiogenic factors
Science
1987
235
444
447
5
Cockerill
GW
,
Gamble
JR
,
Vadas
MA
Angiogenesis: models and modulators
Int Rev Cytol
1995
159
113
160
[PubMed]
6
Hanahan
D
,
Folkman
J
Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis
Cell
1996
86
353
364
[PubMed]
7
Fidler
IJ
,
Ellis
LM
The implications of angiogenesis to the biology and therapy of cancer metastasis
Cell
1994
79
185
188
[PubMed]
8
O'Reilly
MS
,
Holmgren
L
,
Shing
Y
,
Chen
C
,
Rosenthal
RA
,
Moses
M
,
Lane
WS
,
Cao
Y
,
Sage
EH
,
Folkman
J
Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma
Cell
1994
79
315
328
[PubMed]
9
O'Reilly
MS
,
Holmgren
L
,
Chen
C
,
Folkman
J
Angiostatin induces and sustains dormancy of human primary tumors in mice
Nat Med
1996
2
689
692
[PubMed]
10
Dong
Z
,
Kumar
R
,
Yang
X
,
Fidler
IJ
Macrophage-derived metalloelastase is responsible for the generation of angiostatin in Lewis lung carcinoma
Cell
1997
88
801
810
[PubMed]
11
Gately
S
,
Twardowski
P
,
Stack
MS
,
Patrick
M
,
Boggio
L
,
Cundiff
DL
,
Schnaper
HW
,
Madison
L
,
Volpert
O
,
Bouck
N
et al
Human prostate carcinoma cells express enzymatic activity that converts human plasminogen to the angiogenesis inhibitor, angiostatin
Cancer Res
1996
56
4887
4890
[PubMed]
12
Gately
S
,
Twardowski
P
,
Stack
MS
,
Cundiff
DL
,
Grella
D
,
Castellino
FJ
,
Enghild
J
,
Kwaan
HC
,
Lee
F
,
Kramer
RA
et al
The mechanism of cancer-mediated conversion of plasminogen to the angiogenesis inhibitor angiostatin
Proc Natl Acad Sci USA
1997
94
10868
10872
[PubMed]
13
Patterson
BC
,
Sang
QA
Angiostatin-converting enzyme activities of human matrilysin (MMP-7) and gelatinase B/type IV collagenase (MMP-9)
J Biol Chem
1997
272
28823
28825
[PubMed]
14
Kumar
R
,
Dong
Z
,
Fidler
IJ
Differential regulation of metalloelastase activity in murine peritoneal macrophages by granulocyte-macrophage colony-stimulating factor and macrophage colony-stimulating factor
J Immunol
1996
157
5104
5111
[PubMed]
15
Raz
A
,
Fogler
WE
,
Fidler
IJ
The effects of experimental conditions on expression of in vitromediated tumor cytotoxicity by murine macrophages
Cancer Immunol Immunother
1979
7
157
163
16
Staroselsky
AH
,
Pathak
S
,
Chernajovsky
Y
,
Tucker
SL
,
Fidler
IJ
Predominance of the metastatic phenotype in somatic cell hybrids of the K-1735 murine melanoma
Cancer Res
1991
51
6292
6298
[PubMed]
17
Killion
JJ
,
Beltran
PJ
,
O'Brian
CA
,
Yoon
S-S
,
Fan
D
,
Wilson
MR
,
Fidler
IJ
The antitumor activity of doxorubicin against drug-resistant murine carcinoma is enhanced by oral administration of a synthetic staurosporine analogue, CGP 41251
Oncol Res
1995
7
453
459
[PubMed]
18
Dinney
CPN
,
Bucana
CD
,
Utsugi
T
,
Fidler
IJ
,
von Eschenbach
AC
,
Killion
JJ
Therapy of spontaneous lung metastasis of murine renal adenocarcinoma by systemic administration of liposomes containing the macrophage activator CGP 31362
Cancer Res
1991
51
3741
3747
[PubMed]
19
Dong
Z
,
Qi
X
,
Xie
K
,
Fidler
IJ
Protein tyrosine kinase inhibitors decrease induction of nitric oxide synthase activity in lipopolysaccharide-responsive and lipopolysaccharide-nonresponsive murine macrophages
J Immunol
1993
151
2717
2724
[PubMed]
20
Werb
Z
,
Banda
MJ
,
McKerrow
JH
,
Sandhaus
RA
Elastases and elastin degradation
J Invest Dermatol
1982
79
154S
159S
[PubMed]
21
Dong
Z
,
Radinsky
R
,
Fan
D
,
Tsan
R
,
Bucana
CD
,
Wilmanns
C
,
Fidler
IJ
Organ-specific modulation of steady-state mdr gene expression and drug resistance in murine colon cancer cells
J Natl Cancer Inst
1994
86
913
920
[PubMed]
22
Fort
P
,
Marty
L
,
Liechaczyk
M
,
Sabrouty
SE
,
Dani
C
,
Blanchard
JM
Various rat adult tissues express only one major mRNA species from the glyceraldehyde- 3-phosphate-dehydrogenase multigenic family
Nucleic Acids Res
1985
13
1431
1442
[PubMed]
23
Shapiro
SD
,
Griffin
GL
,
Gilbert
DJ
,
Jenkins
NA
,
Copeland
NG
,
Welgus
HG
,
Senior
RM
,
Ley
TJ
Molecular cloning, chromosomal localization, and bacterial expression of a murine macrophage metalloelastase
J Biol Chem
1992
267
4664
4671
[PubMed]
24
Dong
Z
,
Qi
X
,
Fidler
IJ
Tyrosine phosphorylation of mitogen-activated protein kinases is necessary for activation of murine macrophages by natural and synthetic bacterial products
J Exp Med
1993
177
1071
1077
[PubMed]
25
Hughes
DA
,
Fraser
IP
,
Gordon
S
Murine M phi scavenger receptor: adhesion function and expression
Immunol Lett
1994
43
7
14
[PubMed]
26
Ju
DW
,
Cao
X
,
Acres
B
Active specific immunotherapy of pulmonary metastasis with vaccinia melanoma oncolysate prepared from granulocyte/macrophage-colony-stimulating-factor-gene-encoded vaccinia virus
J Cancer Res Clin Oncol
1996
122
716
722
[PubMed]
27
Qin
H
,
Chatterjee
SK
Cancer gene therapy using tumor cells infected with recombinant vaccinia virus expressing GM-CSF
Hum Gene Ther
1996
7
1853
1860
[PubMed]
28
Prehn
RT
The inhibition of tumor growth by tumor mass
Cancer Res
1991
51
2
4
[PubMed]
29
Prehn
RT
Two competing influences that may explain concomitant tumor resistance
Cancer Res
1993
53
3266
3269
[PubMed]
30
Fish
B
,
Gunduz
N
,
Coyle
J
,
Ruddock
C
,
Saffer
E
Presence of a growth-stimulating factor in serum following primary tumor removal in mice
Cancer Res
1989
49
1996
2001
[PubMed]
31
Gorelik
E
Concomitant tumor immunity and the resistance to a second tumor challenge
Adv Cancer Res
1983
39
71
120
[PubMed]
32
Gorelik
E
Resistance of tumor-bearing mice to a second tumor challenge
Cancer Res
1983
43
138
145
[PubMed]
33
Sugarbaker, E.V., J. Thornthwaite, and A.S. Ketcham. 1977. Inhibitory effect of a primary tumor on metastasis. In Progress in Cancer Research and Therapy. S.B. Day, W.P.L. Myers, P. Stansly, S. Garrattini, and M.G. Lewis, editors. Raven Press, Ltd., New York. 227–240.
34
Tyzzer
EE
Factors in the production and growth of tumor metastases
J Med Res
1913
28
309
333
35
Folkman
J
Tumor angiogenesis: therapeutic implications
N Engl J Med
1971
285
1182
1186
[PubMed]
36
Folkman
J
,
Cotran
R
Relation of vascular proliferation to tumor growth
Int Rev Exp Pathol
1976
16
207
248
[PubMed]
37
Dvorak
HF
Tumors: wounds that do not heal. Similarities between tumor stroma generation and wound healing
N Engl J Med
1986
315
1650
1659
[PubMed]
38
Wu
Z
,
O'Reilly
MS
,
Folkman
J
,
Shing
Y
Suppression of tumor growth with recombinant murine angiostatin
Biochem Biophys Res Commun
1997
236
651
654
[PubMed]
39
Sim
BK
,
O'Reilly
MS
,
Liang
H
,
Fortier
AH
,
He
W
,
Madsen
JW
,
Lapcevick
R
,
Nacy
CA
A recombinant human angiostatin protein inhibits experimental primary and metastatic cancer
Cancer Res
1997
57
1329
1334
[PubMed]
40
O'Reilly
MS
,
Holmgren
L
,
Chen
C
,
Folkman
J
Angiostatin induces and sustains dormancy of human primary tumors in mice
Nat Med
1996
2
689
692
[PubMed]
41
Boehm
T
,
Folkman
J
,
Browder
T
,
O'Reilly
MS
Antiangiogenic therapy of experimental cancer does not induce acquired drug resistance
Nature
1997
390
404
407
[PubMed]

Author notes

Address correspondence to Zhongyun Dong, Department of Cell Biology, Box 173, University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030. Phone: 713-792-8523; Fax: 713-792-8747; E-mail: zdong@notes.mdacc.tmc.edu